kernel-vserver-2.3.patch

   1 diff -NurpP --minimal linux-3.0/.config.inet linux-3.0-vs2.3.1-pre7/.config.inet
   2 --- linux-3.0/.config.inet      1970-01-01 01:00:00.000000000 +0100
   3 +++ linux-3.0-vs2.3.1-pre7/.config.inet 2011-06-13 14:55:18.000000000 +0200
   4 @@ -0,0 +1,1620 @@
   5 +#
   6 +# Automatically generated make config: don't edit
   7 +# Linux/x86_64 3.0.0-rc2-vs2.3.x Kernel Configuration
   8 +#
   9 +CONFIG_64BIT=y
  10 +# CONFIG_X86_32 is not set
  11 +CONFIG_X86_64=y
  12 +CONFIG_X86=y
  13 +CONFIG_INSTRUCTION_DECODER=y
  14 +CONFIG_OUTPUT_FORMAT="elf64-x86-64"
  15 +CONFIG_ARCH_DEFCONFIG="arch/x86/configs/x86_64_defconfig"
  16 +CONFIG_GENERIC_CMOS_UPDATE=y
  17 +CONFIG_CLOCKSOURCE_WATCHDOG=y
  18 +CONFIG_GENERIC_CLOCKEVENTS=y
  19 +CONFIG_GENERIC_CLOCKEVENTS_BROADCAST=y
  20 +CONFIG_LOCKDEP_SUPPORT=y
  21 +CONFIG_STACKTRACE_SUPPORT=y
  22 +CONFIG_HAVE_LATENCYTOP_SUPPORT=y
  23 +CONFIG_MMU=y
  24 +CONFIG_ZONE_DMA=y
  25 +CONFIG_NEED_DMA_MAP_STATE=y
  26 +CONFIG_NEED_SG_DMA_LENGTH=y
  27 +CONFIG_GENERIC_ISA_DMA=y
  28 +CONFIG_GENERIC_IOMAP=y
  29 +CONFIG_GENERIC_BUG=y
  30 +CONFIG_GENERIC_BUG_RELATIVE_POINTERS=y
  31 +CONFIG_GENERIC_HWEIGHT=y
  32 +CONFIG_ARCH_MAY_HAVE_PC_FDC=y
  33 +# CONFIG_RWSEM_GENERIC_SPINLOCK is not set
  34 +CONFIG_RWSEM_XCHGADD_ALGORITHM=y
  35 +CONFIG_ARCH_HAS_CPU_IDLE_WAIT=y
  36 +CONFIG_GENERIC_CALIBRATE_DELAY=y
  37 +CONFIG_GENERIC_TIME_VSYSCALL=y
  38 +CONFIG_ARCH_HAS_CPU_RELAX=y
  39 +CONFIG_ARCH_HAS_DEFAULT_IDLE=y
  40 +CONFIG_ARCH_HAS_CACHE_LINE_SIZE=y
  41 +CONFIG_HAVE_SETUP_PER_CPU_AREA=y
  42 +CONFIG_NEED_PER_CPU_EMBED_FIRST_CHUNK=y
  43 +CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK=y
  44 +CONFIG_HAVE_CPUMASK_OF_CPU_MAP=y
  45 +CONFIG_ARCH_HIBERNATION_POSSIBLE=y
  46 +CONFIG_ARCH_SUSPEND_POSSIBLE=y
  47 +CONFIG_ZONE_DMA32=y
  48 +CONFIG_ARCH_POPULATES_NODE_MAP=y
  49 +CONFIG_AUDIT_ARCH=y
  50 +CONFIG_ARCH_SUPPORTS_OPTIMIZED_INLINING=y
  51 +CONFIG_ARCH_SUPPORTS_DEBUG_PAGEALLOC=y
  52 +CONFIG_X86_64_SMP=y
  53 +CONFIG_X86_HT=y
  54 +CONFIG_ARCH_HWEIGHT_CFLAGS="-fcall-saved-rdi -fcall-saved-rsi -fcall-saved-rdx -fcall-saved-rcx -fcall-saved-r8 -fcall-saved-r9 -fcall-saved-r10 -fcall-saved-r11"
  55 +# CONFIG_KTIME_SCALAR is not set
  56 +CONFIG_ARCH_CPU_PROBE_RELEASE=y
  57 +CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
  58 +CONFIG_CONSTRUCTORS=y
  59 +CONFIG_HAVE_IRQ_WORK=y
  60 +CONFIG_IRQ_WORK=y
  61 +
  62 +#
  63 +# General setup
  64 +#
  65 +CONFIG_EXPERIMENTAL=y
  66 +CONFIG_INIT_ENV_ARG_LIMIT=32
  67 +CONFIG_CROSS_COMPILE=""
  68 +CONFIG_LOCALVERSION=""
  69 +# CONFIG_LOCALVERSION_AUTO is not set
  70 +CONFIG_HAVE_KERNEL_GZIP=y
  71 +CONFIG_HAVE_KERNEL_BZIP2=y
  72 +CONFIG_HAVE_KERNEL_LZMA=y
  73 +CONFIG_HAVE_KERNEL_XZ=y
  74 +CONFIG_HAVE_KERNEL_LZO=y
  75 +# CONFIG_KERNEL_GZIP is not set
  76 +# CONFIG_KERNEL_BZIP2 is not set
  77 +CONFIG_KERNEL_LZMA=y
  78 +# CONFIG_KERNEL_XZ is not set
  79 +# CONFIG_KERNEL_LZO is not set
  80 +CONFIG_SWAP=y
  81 +CONFIG_SYSVIPC=y
  82 +CONFIG_SYSVIPC_SYSCTL=y
  83 +CONFIG_POSIX_MQUEUE=y
  84 +CONFIG_POSIX_MQUEUE_SYSCTL=y
  85 +CONFIG_BSD_PROCESS_ACCT=y
  86 +# CONFIG_BSD_PROCESS_ACCT_V3 is not set
  87 +# CONFIG_FHANDLE is not set
  88 +CONFIG_TASKSTATS=y
  89 +CONFIG_TASK_DELAY_ACCT=y
  90 +CONFIG_TASK_XACCT=y
  91 +CONFIG_TASK_IO_ACCOUNTING=y
  92 +# CONFIG_AUDIT is not set
  93 +CONFIG_HAVE_GENERIC_HARDIRQS=y
  94 +
  95 +#
  96 +# IRQ subsystem
  97 +#
  98 +CONFIG_GENERIC_HARDIRQS=y
  99 +CONFIG_HAVE_SPARSE_IRQ=y
 100 +CONFIG_GENERIC_IRQ_PROBE=y
 101 +CONFIG_GENERIC_IRQ_SHOW=y
 102 +CONFIG_GENERIC_PENDING_IRQ=y
 103 +CONFIG_IRQ_FORCED_THREADING=y
 104 +# CONFIG_SPARSE_IRQ is not set
 105 +
 106 +#
 107 +# RCU Subsystem
 108 +#
 109 +CONFIG_TREE_RCU=y
 110 +# CONFIG_PREEMPT_RCU is not set
 111 +# CONFIG_RCU_TRACE is not set
 112 +CONFIG_RCU_FANOUT=64
 113 +# CONFIG_RCU_FANOUT_EXACT is not set
 114 +# CONFIG_TREE_RCU_TRACE is not set
 115 +CONFIG_IKCONFIG=y
 116 +CONFIG_IKCONFIG_PROC=y
 117 +CONFIG_LOG_BUF_SHIFT=18
 118 +CONFIG_HAVE_UNSTABLE_SCHED_CLOCK=y
 119 +CONFIG_CGROUPS=y
 120 +# CONFIG_CGROUP_DEBUG is not set
 121 +# CONFIG_CGROUP_FREEZER is not set
 122 +# CONFIG_CGROUP_DEVICE is not set
 123 +# CONFIG_CPUSETS is not set
 124 +# CONFIG_CGROUP_CPUACCT is not set
 125 +CONFIG_RESOURCE_COUNTERS=y
 126 +CONFIG_CGROUP_MEM_RES_CTLR=y
 127 +CONFIG_CGROUP_MEM_RES_CTLR_SWAP=y
 128 +CONFIG_CGROUP_MEM_RES_CTLR_SWAP_ENABLED=y
 129 +# CONFIG_CGROUP_PERF is not set
 130 +CONFIG_CGROUP_SCHED=y
 131 +CONFIG_FAIR_GROUP_SCHED=y
 132 +# CONFIG_RT_GROUP_SCHED is not set
 133 +# CONFIG_BLK_CGROUP is not set
 134 +CONFIG_NAMESPACES=y
 135 +CONFIG_UTS_NS=y
 136 +CONFIG_IPC_NS=y
 137 +CONFIG_USER_NS=y
 138 +# CONFIG_PID_NS is not set
 139 +# CONFIG_NET_NS is not set
 140 +CONFIG_SCHED_AUTOGROUP=y
 141 +CONFIG_MM_OWNER=y
 142 +# CONFIG_SYSFS_DEPRECATED is not set
 143 +# CONFIG_RELAY is not set
 144 +# CONFIG_BLK_DEV_INITRD is not set
 145 +CONFIG_CC_OPTIMIZE_FOR_SIZE=y
 146 +CONFIG_SYSCTL=y
 147 +CONFIG_ANON_INODES=y
 148 +CONFIG_EXPERT=y
 149 +CONFIG_UID16=y
 150 +CONFIG_SYSCTL_SYSCALL=y
 151 +CONFIG_KALLSYMS=y
 152 +CONFIG_KALLSYMS_ALL=y
 153 +CONFIG_HOTPLUG=y
 154 +CONFIG_PRINTK=y
 155 +CONFIG_BUG=y
 156 +CONFIG_ELF_CORE=y
 157 +CONFIG_PCSPKR_PLATFORM=y
 158 +CONFIG_BASE_FULL=y
 159 +CONFIG_FUTEX=y
 160 +CONFIG_EPOLL=y
 161 +CONFIG_SIGNALFD=y
 162 +CONFIG_TIMERFD=y
 163 +CONFIG_EVENTFD=y
 164 +CONFIG_SHMEM=y
 165 +CONFIG_AIO=y
 166 +# CONFIG_EMBEDDED is not set
 167 +CONFIG_HAVE_PERF_EVENTS=y
 168 +
 169 +#
 170 +# Kernel Performance Events And Counters
 171 +#
 172 +CONFIG_PERF_EVENTS=y
 173 +# CONFIG_PERF_COUNTERS is not set
 174 +# CONFIG_DEBUG_PERF_USE_VMALLOC is not set
 175 +CONFIG_VM_EVENT_COUNTERS=y
 176 +CONFIG_PCI_QUIRKS=y
 177 +CONFIG_SLUB_DEBUG=y
 178 +# CONFIG_COMPAT_BRK is not set
 179 +# CONFIG_SLAB is not set
 180 +CONFIG_SLUB=y
 181 +# CONFIG_SLOB is not set
 182 +# CONFIG_PROFILING is not set
 183 +CONFIG_HAVE_OPROFILE=y
 184 +# CONFIG_JUMP_LABEL is not set
 185 +CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y
 186 +CONFIG_HAVE_IOREMAP_PROT=y
 187 +CONFIG_HAVE_KPROBES=y
 188 +CONFIG_HAVE_KRETPROBES=y
 189 +CONFIG_HAVE_OPTPROBES=y
 190 +CONFIG_HAVE_ARCH_TRACEHOOK=y
 191 +CONFIG_HAVE_DMA_ATTRS=y
 192 +CONFIG_USE_GENERIC_SMP_HELPERS=y
 193 +CONFIG_HAVE_REGS_AND_STACK_ACCESS_API=y
 194 +CONFIG_HAVE_DMA_API_DEBUG=y
 195 +CONFIG_HAVE_HW_BREAKPOINT=y
 196 +CONFIG_HAVE_MIXED_BREAKPOINTS_REGS=y
 197 +CONFIG_HAVE_USER_RETURN_NOTIFIER=y
 198 +CONFIG_HAVE_PERF_EVENTS_NMI=y
 199 +CONFIG_HAVE_ARCH_JUMP_LABEL=y
 200 +
 201 +#
 202 +# GCOV-based kernel profiling
 203 +#
 204 +# CONFIG_HAVE_GENERIC_DMA_COHERENT is not set
 205 +CONFIG_SLABINFO=y
 206 +CONFIG_RT_MUTEXES=y
 207 +CONFIG_BASE_SMALL=0
 208 +# CONFIG_MODULES is not set
 209 +CONFIG_STOP_MACHINE=y
 210 +CONFIG_BLOCK=y
 211 +# CONFIG_BLK_DEV_BSG is not set
 212 +# CONFIG_BLK_DEV_INTEGRITY is not set
 213 +CONFIG_BLOCK_COMPAT=y
 214 +
 215 +#
 216 +# IO Schedulers
 217 +#
 218 +CONFIG_IOSCHED_NOOP=y
 219 +# CONFIG_IOSCHED_DEADLINE is not set
 220 +CONFIG_IOSCHED_CFQ=y
 221 +CONFIG_DEFAULT_CFQ=y
 222 +# CONFIG_DEFAULT_NOOP is not set
 223 +CONFIG_DEFAULT_IOSCHED="cfq"
 224 +# CONFIG_INLINE_SPIN_TRYLOCK is not set
 225 +# CONFIG_INLINE_SPIN_TRYLOCK_BH is not set
 226 +# CONFIG_INLINE_SPIN_LOCK is not set
 227 +# CONFIG_INLINE_SPIN_LOCK_BH is not set
 228 +# CONFIG_INLINE_SPIN_LOCK_IRQ is not set
 229 +# CONFIG_INLINE_SPIN_LOCK_IRQSAVE is not set
 230 +CONFIG_INLINE_SPIN_UNLOCK=y
 231 +# CONFIG_INLINE_SPIN_UNLOCK_BH is not set
 232 +CONFIG_INLINE_SPIN_UNLOCK_IRQ=y
 233 +# CONFIG_INLINE_SPIN_UNLOCK_IRQRESTORE is not set
 234 +# CONFIG_INLINE_READ_TRYLOCK is not set
 235 +# CONFIG_INLINE_READ_LOCK is not set
 236 +# CONFIG_INLINE_READ_LOCK_BH is not set
 237 +# CONFIG_INLINE_READ_LOCK_IRQ is not set
 238 +# CONFIG_INLINE_READ_LOCK_IRQSAVE is not set
 239 +CONFIG_INLINE_READ_UNLOCK=y
 240 +# CONFIG_INLINE_READ_UNLOCK_BH is not set
 241 +CONFIG_INLINE_READ_UNLOCK_IRQ=y
 242 +# CONFIG_INLINE_READ_UNLOCK_IRQRESTORE is not set
 243 +# CONFIG_INLINE_WRITE_TRYLOCK is not set
 244 +# CONFIG_INLINE_WRITE_LOCK is not set
 245 +# CONFIG_INLINE_WRITE_LOCK_BH is not set
 246 +# CONFIG_INLINE_WRITE_LOCK_IRQ is not set
 247 +# CONFIG_INLINE_WRITE_LOCK_IRQSAVE is not set
 248 +CONFIG_INLINE_WRITE_UNLOCK=y
 249 +# CONFIG_INLINE_WRITE_UNLOCK_BH is not set
 250 +CONFIG_INLINE_WRITE_UNLOCK_IRQ=y
 251 +# CONFIG_INLINE_WRITE_UNLOCK_IRQRESTORE is not set
 252 +CONFIG_MUTEX_SPIN_ON_OWNER=y
 253 +CONFIG_FREEZER=y
 254 +
 255 +#
 256 +# Processor type and features
 257 +#
 258 +CONFIG_TICK_ONESHOT=y
 259 +# CONFIG_NO_HZ is not set
 260 +CONFIG_HIGH_RES_TIMERS=y
 261 +CONFIG_GENERIC_CLOCKEVENTS_BUILD=y
 262 +CONFIG_SMP=y
 263 +CONFIG_X86_MPPARSE=y
 264 +# CONFIG_X86_EXTENDED_PLATFORM is not set
 265 +CONFIG_X86_SUPPORTS_MEMORY_FAILURE=y
 266 +CONFIG_SCHED_OMIT_FRAME_POINTER=y
 267 +# CONFIG_PARAVIRT_GUEST is not set
 268 +CONFIG_NO_BOOTMEM=y
 269 +# CONFIG_MEMTEST is not set
 270 +# CONFIG_MK8 is not set
 271 +# CONFIG_MPSC is not set
 272 +CONFIG_MCORE2=y
 273 +# CONFIG_MATOM is not set
 274 +# CONFIG_GENERIC_CPU is not set
 275 +CONFIG_X86_INTERNODE_CACHE_SHIFT=6
 276 +CONFIG_X86_CMPXCHG=y
 277 +CONFIG_CMPXCHG_LOCAL=y
 278 +CONFIG_X86_L1_CACHE_SHIFT=6
 279 +CONFIG_X86_XADD=y
 280 +CONFIG_X86_WP_WORKS_OK=y
 281 +CONFIG_X86_INTEL_USERCOPY=y
 282 +CONFIG_X86_USE_PPRO_CHECKSUM=y
 283 +CONFIG_X86_P6_NOP=y
 284 +CONFIG_X86_TSC=y
 285 +CONFIG_X86_CMPXCHG64=y
 286 +CONFIG_X86_CMOV=y
 287 +CONFIG_X86_MINIMUM_CPU_FAMILY=64
 288 +CONFIG_X86_DEBUGCTLMSR=y
 289 +# CONFIG_PROCESSOR_SELECT is not set
 290 +CONFIG_CPU_SUP_INTEL=y
 291 +CONFIG_CPU_SUP_AMD=y
 292 +CONFIG_CPU_SUP_CENTAUR=y
 293 +CONFIG_HPET_TIMER=y
 294 +CONFIG_HPET_EMULATE_RTC=y
 295 +CONFIG_DMI=y
 296 +CONFIG_GART_IOMMU=y
 297 +# CONFIG_CALGARY_IOMMU is not set
 298 +# CONFIG_AMD_IOMMU is not set
 299 +CONFIG_SWIOTLB=y
 300 +CONFIG_IOMMU_HELPER=y
 301 +# CONFIG_IOMMU_API is not set
 302 +# CONFIG_MAXSMP is not set
 303 +CONFIG_NR_CPUS=32
 304 +CONFIG_SCHED_SMT=y
 305 +CONFIG_SCHED_MC=y
 306 +# CONFIG_IRQ_TIME_ACCOUNTING is not set
 307 +CONFIG_PREEMPT_NONE=y
 308 +# CONFIG_PREEMPT_VOLUNTARY is not set
 309 +# CONFIG_PREEMPT is not set
 310 +CONFIG_X86_LOCAL_APIC=y
 311 +CONFIG_X86_IO_APIC=y
 312 +# CONFIG_X86_REROUTE_FOR_BROKEN_BOOT_IRQS is not set
 313 +CONFIG_X86_MCE=y
 314 +CONFIG_X86_MCE_INTEL=y
 315 +CONFIG_X86_MCE_AMD=y
 316 +CONFIG_X86_MCE_THRESHOLD=y
 317 +# CONFIG_X86_MCE_INJECT is not set
 318 +CONFIG_X86_THERMAL_VECTOR=y
 319 +# CONFIG_I8K is not set
 320 +# CONFIG_MICROCODE is not set
 321 +CONFIG_X86_MSR=y
 322 +CONFIG_X86_CPUID=y
 323 +CONFIG_ARCH_PHYS_ADDR_T_64BIT=y
 324 +CONFIG_ARCH_DMA_ADDR_T_64BIT=y
 325 +CONFIG_DIRECT_GBPAGES=y
 326 +# CONFIG_NUMA is not set
 327 +CONFIG_ARCH_SPARSEMEM_ENABLE=y
 328 +CONFIG_ARCH_SPARSEMEM_DEFAULT=y
 329 +CONFIG_ARCH_SELECT_MEMORY_MODEL=y
 330 +CONFIG_ILLEGAL_POINTER_VALUE=0xdead000000000000
 331 +CONFIG_SELECT_MEMORY_MODEL=y
 332 +CONFIG_SPARSEMEM_MANUAL=y
 333 +CONFIG_SPARSEMEM=y
 334 +CONFIG_HAVE_MEMORY_PRESENT=y
 335 +CONFIG_SPARSEMEM_EXTREME=y
 336 +CONFIG_SPARSEMEM_VMEMMAP_ENABLE=y
 337 +CONFIG_SPARSEMEM_ALLOC_MEM_MAP_TOGETHER=y
 338 +CONFIG_SPARSEMEM_VMEMMAP=y
 339 +CONFIG_HAVE_MEMBLOCK=y
 340 +# CONFIG_MEMORY_HOTPLUG is not set
 341 +CONFIG_PAGEFLAGS_EXTENDED=y
 342 +CONFIG_SPLIT_PTLOCK_CPUS=4
 343 +CONFIG_COMPACTION=y
 344 +CONFIG_MIGRATION=y
 345 +CONFIG_PHYS_ADDR_T_64BIT=y
 346 +CONFIG_ZONE_DMA_FLAG=1
 347 +CONFIG_BOUNCE=y
 348 +CONFIG_VIRT_TO_BUS=y
 349 +# CONFIG_KSM is not set
 350 +CONFIG_DEFAULT_MMAP_MIN_ADDR=4096
 351 +CONFIG_ARCH_SUPPORTS_MEMORY_FAILURE=y
 352 +# CONFIG_MEMORY_FAILURE is not set
 353 +CONFIG_TRANSPARENT_HUGEPAGE=y
 354 +CONFIG_TRANSPARENT_HUGEPAGE_ALWAYS=y
 355 +# CONFIG_TRANSPARENT_HUGEPAGE_MADVISE is not set
 356 +CONFIG_CLEANCACHE=y
 357 +# CONFIG_X86_CHECK_BIOS_CORRUPTION is not set
 358 +CONFIG_X86_RESERVE_LOW=4
 359 +CONFIG_MTRR=y
 360 +CONFIG_MTRR_SANITIZER=y
 361 +CONFIG_MTRR_SANITIZER_ENABLE_DEFAULT=0
 362 +CONFIG_MTRR_SANITIZER_SPARE_REG_NR_DEFAULT=1
 363 +CONFIG_X86_PAT=y
 364 +CONFIG_ARCH_USES_PG_UNCACHED=y
 365 +# CONFIG_EFI is not set
 366 +CONFIG_SECCOMP=y
 367 +# CONFIG_CC_STACKPROTECTOR is not set
 368 +CONFIG_HZ_100=y
 369 +# CONFIG_HZ_250 is not set
 370 +# CONFIG_HZ_300 is not set
 371 +# CONFIG_HZ_1000 is not set
 372 +CONFIG_HZ=100
 373 +CONFIG_SCHED_HRTICK=y
 374 +# CONFIG_KEXEC is not set
 375 +# CONFIG_CRASH_DUMP is not set
 376 +CONFIG_PHYSICAL_START=0x1000000
 377 +# CONFIG_RELOCATABLE is not set
 378 +CONFIG_PHYSICAL_ALIGN=0x1000000
 379 +CONFIG_HOTPLUG_CPU=y
 380 +# CONFIG_COMPAT_VDSO is not set
 381 +# CONFIG_CMDLINE_BOOL is not set
 382 +CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
 383 +
 384 +#
 385 +# Power management and ACPI options
 386 +#
 387 +CONFIG_SUSPEND=y
 388 +CONFIG_SUSPEND_FREEZER=y
 389 +# CONFIG_HIBERNATION is not set
 390 +CONFIG_PM_SLEEP=y
 391 +CONFIG_PM_SLEEP_SMP=y
 392 +# CONFIG_PM_RUNTIME is not set
 393 +CONFIG_PM=y
 394 +# CONFIG_PM_DEBUG is not set
 395 +CONFIG_ACPI=y
 396 +CONFIG_ACPI_SLEEP=y
 397 +# CONFIG_ACPI_PROCFS is not set
 398 +# CONFIG_ACPI_PROCFS_POWER is not set
 399 +# CONFIG_ACPI_EC_DEBUGFS is not set
 400 +CONFIG_ACPI_PROC_EVENT=y
 401 +CONFIG_ACPI_AC=y
 402 +CONFIG_ACPI_BATTERY=y
 403 +CONFIG_ACPI_BUTTON=y
 404 +CONFIG_ACPI_FAN=y
 405 +# CONFIG_ACPI_DOCK is not set
 406 +CONFIG_ACPI_PROCESSOR=y
 407 +CONFIG_ACPI_HOTPLUG_CPU=y
 408 +# CONFIG_ACPI_PROCESSOR_AGGREGATOR is not set
 409 +CONFIG_ACPI_THERMAL=y
 410 +# CONFIG_ACPI_CUSTOM_DSDT is not set
 411 +CONFIG_ACPI_BLACKLIST_YEAR=0
 412 +# CONFIG_ACPI_DEBUG is not set
 413 +# CONFIG_ACPI_PCI_SLOT is not set
 414 +CONFIG_X86_PM_TIMER=y
 415 +CONFIG_ACPI_CONTAINER=y
 416 +# CONFIG_ACPI_SBS is not set
 417 +# CONFIG_ACPI_HED is not set
 418 +# CONFIG_ACPI_APEI is not set
 419 +# CONFIG_SFI is not set
 420 +
 421 +#
 422 +# CPU Frequency scaling
 423 +#
 424 +# CONFIG_CPU_FREQ is not set
 425 +CONFIG_CPU_IDLE=y
 426 +CONFIG_CPU_IDLE_GOV_LADDER=y
 427 +# CONFIG_INTEL_IDLE is not set
 428 +
 429 +#
 430 +# Memory power savings
 431 +#
 432 +# CONFIG_I7300_IDLE is not set
 433 +
 434 +#
 435 +# Bus options (PCI etc.)
 436 +#
 437 +CONFIG_PCI=y
 438 +CONFIG_PCI_DIRECT=y
 439 +# CONFIG_PCI_MMCONFIG is not set
 440 +CONFIG_PCI_DOMAINS=y
 441 +# CONFIG_PCI_CNB20LE_QUIRK is not set
 442 +CONFIG_PCIEPORTBUS=y
 443 +# CONFIG_PCIEAER is not set
 444 +CONFIG_PCIEASPM=y
 445 +# CONFIG_PCIEASPM_DEBUG is not set
 446 +CONFIG_ARCH_SUPPORTS_MSI=y
 447 +# CONFIG_PCI_MSI is not set
 448 +# CONFIG_PCI_DEBUG is not set
 449 +# CONFIG_PCI_STUB is not set
 450 +CONFIG_HT_IRQ=y
 451 +# CONFIG_PCI_IOV is not set
 452 +CONFIG_PCI_IOAPIC=y
 453 +CONFIG_PCI_LABEL=y
 454 +CONFIG_ISA_DMA_API=y
 455 +CONFIG_AMD_NB=y
 456 +# CONFIG_PCCARD is not set
 457 +# CONFIG_HOTPLUG_PCI is not set
 458 +# CONFIG_RAPIDIO is not set
 459 +
 460 +#
 461 +# Executable file formats / Emulations
 462 +#
 463 +CONFIG_BINFMT_ELF=y
 464 +CONFIG_COMPAT_BINFMT_ELF=y
 465 +# CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS is not set
 466 +# CONFIG_HAVE_AOUT is not set
 467 +# CONFIG_BINFMT_MISC is not set
 468 +CONFIG_IA32_EMULATION=y
 469 +# CONFIG_IA32_AOUT is not set
 470 +CONFIG_COMPAT=y
 471 +CONFIG_COMPAT_FOR_U64_ALIGNMENT=y
 472 +CONFIG_SYSVIPC_COMPAT=y
 473 +CONFIG_HAVE_TEXT_POKE_SMP=y
 474 +CONFIG_NET=y
 475 +
 476 +#
 477 +# Networking options
 478 +#
 479 +CONFIG_PACKET=y
 480 +CONFIG_UNIX=y
 481 +CONFIG_XFRM=y
 482 +CONFIG_XFRM_USER=y
 483 +# CONFIG_XFRM_SUB_POLICY is not set
 484 +# CONFIG_XFRM_MIGRATE is not set
 485 +# CONFIG_XFRM_STATISTICS is not set
 486 +CONFIG_XFRM_IPCOMP=y
 487 +# CONFIG_NET_KEY is not set
 488 +CONFIG_IP_MULTICAST=y
 489 +# CONFIG_IP_ADVANCED_ROUTER is not set
 490 +# CONFIG_IP_PNP is not set
 491 +# CONFIG_NET_IPIP is not set
 492 +# CONFIG_NET_IPGRE_DEMUX is not set
 493 +# CONFIG_IP_MROUTE is not set
 494 +# CONFIG_ARPD is not set
 495 +CONFIG_SYN_COOKIES=y
 496 +# CONFIG_INET_AH is not set
 497 +# CONFIG_INET_ESP is not set
 498 +# CONFIG_INET_IPCOMP is not set
 499 +# CONFIG_INET_XFRM_TUNNEL is not set
 500 +CONFIG_INET_TUNNEL=y
 501 +# CONFIG_INET_XFRM_MODE_TRANSPORT is not set
 502 +# CONFIG_INET_XFRM_MODE_TUNNEL is not set
 503 +# CONFIG_INET_XFRM_MODE_BEET is not set
 504 +CONFIG_INET_LRO=y
 505 +CONFIG_INET_DIAG=y
 506 +CONFIG_INET_TCP_DIAG=y
 507 +# CONFIG_TCP_CONG_ADVANCED is not set
 508 +CONFIG_TCP_CONG_CUBIC=y
 509 +CONFIG_DEFAULT_TCP_CONG="cubic"
 510 +# CONFIG_TCP_MD5SIG is not set
 511 +CONFIG_IPV6=y
 512 +# CONFIG_IPV6_PRIVACY is not set
 513 +# CONFIG_IPV6_ROUTER_PREF is not set
 514 +# CONFIG_IPV6_OPTIMISTIC_DAD is not set
 515 +CONFIG_INET6_AH=y
 516 +CONFIG_INET6_ESP=y
 517 +CONFIG_INET6_IPCOMP=y
 518 +# CONFIG_IPV6_MIP6 is not set
 519 +CONFIG_INET6_XFRM_TUNNEL=y
 520 +CONFIG_INET6_TUNNEL=y
 521 +CONFIG_INET6_XFRM_MODE_TRANSPORT=y
 522 +CONFIG_INET6_XFRM_MODE_TUNNEL=y
 523 +CONFIG_INET6_XFRM_MODE_BEET=y
 524 +# CONFIG_INET6_XFRM_MODE_ROUTEOPTIMIZATION is not set
 525 +CONFIG_IPV6_SIT=y
 526 +# CONFIG_IPV6_SIT_6RD is not set
 527 +CONFIG_IPV6_NDISC_NODETYPE=y
 528 +# CONFIG_IPV6_TUNNEL is not set
 529 +# CONFIG_IPV6_MULTIPLE_TABLES is not set
 530 +# CONFIG_IPV6_MROUTE is not set
 531 +# CONFIG_NETWORK_SECMARK is not set
 532 +# CONFIG_NETWORK_PHY_TIMESTAMPING is not set
 533 +CONFIG_NETFILTER=y
 534 +# CONFIG_NETFILTER_DEBUG is not set
 535 +# CONFIG_NETFILTER_ADVANCED is not set
 536 +
 537 +#
 538 +# Core Netfilter Configuration
 539 +#
 540 +CONFIG_NETFILTER_NETLINK=y
 541 +CONFIG_NETFILTER_NETLINK_LOG=y
 542 +CONFIG_NF_CONNTRACK=y
 543 +CONFIG_NF_CONNTRACK_FTP=y
 544 +CONFIG_NF_CONNTRACK_IRC=y
 545 +CONFIG_NF_CONNTRACK_SIP=y
 546 +CONFIG_NF_CT_NETLINK=y
 547 +CONFIG_NETFILTER_XTABLES=y
 548 +
 549 +#
 550 +# Xtables combined modules
 551 +#
 552 +CONFIG_NETFILTER_XT_MARK=y
 553 +
 554 +#
 555 +# Xtables targets
 556 +#
 557 +CONFIG_NETFILTER_XT_TARGET_NFLOG=y
 558 +CONFIG_NETFILTER_XT_TARGET_TCPMSS=y
 559 +
 560 +#
 561 +# Xtables matches
 562 +#
 563 +CONFIG_NETFILTER_XT_MATCH_CONNTRACK=y
 564 +CONFIG_NETFILTER_XT_MATCH_POLICY=y
 565 +CONFIG_NETFILTER_XT_MATCH_STATE=y
 566 +# CONFIG_IP_SET is not set
 567 +CONFIG_IP_VS=y
 568 +CONFIG_IP_VS_IPV6=y
 569 +# CONFIG_IP_VS_DEBUG is not set
 570 +CONFIG_IP_VS_TAB_BITS=12
 571 +
 572 +#
 573 +# IPVS transport protocol load balancing support
 574 +#
 575 +# CONFIG_IP_VS_PROTO_TCP is not set
 576 +# CONFIG_IP_VS_PROTO_UDP is not set
 577 +# CONFIG_IP_VS_PROTO_AH_ESP is not set
 578 +# CONFIG_IP_VS_PROTO_ESP is not set
 579 +# CONFIG_IP_VS_PROTO_AH is not set
 580 +# CONFIG_IP_VS_PROTO_SCTP is not set
 581 +
 582 +#
 583 +# IPVS scheduler
 584 +#
 585 +# CONFIG_IP_VS_RR is not set
 586 +# CONFIG_IP_VS_WRR is not set
 587 +# CONFIG_IP_VS_LC is not set
 588 +# CONFIG_IP_VS_WLC is not set
 589 +# CONFIG_IP_VS_LBLC is not set
 590 +# CONFIG_IP_VS_LBLCR is not set
 591 +# CONFIG_IP_VS_DH is not set
 592 +# CONFIG_IP_VS_SH is not set
 593 +# CONFIG_IP_VS_SED is not set
 594 +# CONFIG_IP_VS_NQ is not set
 595 +
 596 +#
 597 +# IPVS application helper
 598 +#
 599 +# CONFIG_IP_VS_NFCT is not set
 600 +
 601 +#
 602 +# IP: Netfilter Configuration
 603 +#
 604 +CONFIG_NF_DEFRAG_IPV4=y
 605 +CONFIG_NF_CONNTRACK_IPV4=y
 606 +# CONFIG_NF_CONNTRACK_PROC_COMPAT is not set
 607 +CONFIG_IP_NF_IPTABLES=y
 608 +CONFIG_IP_NF_FILTER=y
 609 +CONFIG_IP_NF_TARGET_REJECT=y
 610 +CONFIG_IP_NF_TARGET_LOG=y
 611 +CONFIG_IP_NF_TARGET_ULOG=y
 612 +CONFIG_NF_NAT=y
 613 +CONFIG_NF_NAT_NEEDED=y
 614 +CONFIG_IP_NF_TARGET_MASQUERADE=y
 615 +CONFIG_NF_NAT_FTP=y
 616 +CONFIG_NF_NAT_IRC=y
 617 +# CONFIG_NF_NAT_TFTP is not set
 618 +# CONFIG_NF_NAT_AMANDA is not set
 619 +# CONFIG_NF_NAT_PPTP is not set
 620 +# CONFIG_NF_NAT_H323 is not set
 621 +CONFIG_NF_NAT_SIP=y
 622 +CONFIG_IP_NF_MANGLE=y
 623 +
 624 +#
 625 +# IPv6: Netfilter Configuration
 626 +#
 627 +# CONFIG_NF_DEFRAG_IPV6 is not set
 628 +# CONFIG_NF_CONNTRACK_IPV6 is not set
 629 +CONFIG_IP6_NF_IPTABLES=y
 630 +CONFIG_IP6_NF_MATCH_IPV6HEADER=y
 631 +CONFIG_IP6_NF_TARGET_LOG=y
 632 +CONFIG_IP6_NF_FILTER=y
 633 +CONFIG_IP6_NF_TARGET_REJECT=y
 634 +CONFIG_IP6_NF_MANGLE=y
 635 +# CONFIG_IP_DCCP is not set
 636 +CONFIG_IP_SCTP=y
 637 +# CONFIG_SCTP_DBG_MSG is not set
 638 +# CONFIG_SCTP_DBG_OBJCNT is not set
 639 +# CONFIG_SCTP_HMAC_NONE is not set
 640 +# CONFIG_SCTP_HMAC_SHA1 is not set
 641 +CONFIG_SCTP_HMAC_MD5=y
 642 +# CONFIG_RDS is not set
 643 +# CONFIG_TIPC is not set
 644 +# CONFIG_ATM is not set
 645 +# CONFIG_L2TP is not set
 646 +# CONFIG_BRIDGE is not set
 647 +# CONFIG_NET_DSA is not set
 648 +# CONFIG_VLAN_8021Q is not set
 649 +# CONFIG_DECNET is not set
 650 +# CONFIG_LLC2 is not set
 651 +# CONFIG_IPX is not set
 652 +# CONFIG_ATALK is not set
 653 +CONFIG_X25=y
 654 +# CONFIG_LAPB is not set
 655 +# CONFIG_ECONET is not set
 656 +# CONFIG_WAN_ROUTER is not set
 657 +# CONFIG_PHONET is not set
 658 +# CONFIG_IEEE802154 is not set
 659 +# CONFIG_NET_SCHED is not set
 660 +# CONFIG_DCB is not set
 661 +# CONFIG_BATMAN_ADV is not set
 662 +CONFIG_RPS=y
 663 +CONFIG_RFS_ACCEL=y
 664 +CONFIG_XPS=y
 665 +CONFIG_HAVE_BPF_JIT=y
 666 +
 667 +#
 668 +# Network testing
 669 +#
 670 +# CONFIG_NET_PKTGEN is not set
 671 +# CONFIG_HAMRADIO is not set
 672 +# CONFIG_CAN is not set
 673 +# CONFIG_IRDA is not set
 674 +# CONFIG_BT is not set
 675 +# CONFIG_AF_RXRPC is not set
 676 +# CONFIG_WIRELESS is not set
 677 +# CONFIG_WIMAX is not set
 678 +# CONFIG_RFKILL is not set
 679 +# CONFIG_NET_9P is not set
 680 +# CONFIG_CAIF is not set
 681 +# CONFIG_CEPH_LIB is not set
 682 +
 683 +#
 684 +# Device Drivers
 685 +#
 686 +
 687 +#
 688 +# Generic Driver Options
 689 +#
 690 +CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
 691 +CONFIG_DEVTMPFS=y
 692 +# CONFIG_DEVTMPFS_MOUNT is not set
 693 +CONFIG_STANDALONE=y
 694 +CONFIG_PREVENT_FIRMWARE_BUILD=y
 695 +CONFIG_FW_LOADER=y
 696 +CONFIG_FIRMWARE_IN_KERNEL=y
 697 +CONFIG_EXTRA_FIRMWARE=""
 698 +# CONFIG_DEBUG_DRIVER is not set
 699 +# CONFIG_DEBUG_DEVRES is not set
 700 +# CONFIG_SYS_HYPERVISOR is not set
 701 +# CONFIG_CONNECTOR is not set
 702 +# CONFIG_MTD is not set
 703 +# CONFIG_PARPORT is not set
 704 +CONFIG_PNP=y
 705 +CONFIG_PNP_DEBUG_MESSAGES=y
 706 +
 707 +#
 708 +# Protocols
 709 +#
 710 +CONFIG_PNPACPI=y
 711 +CONFIG_BLK_DEV=y
 712 +# CONFIG_BLK_DEV_FD is not set
 713 +# CONFIG_BLK_CPQ_DA is not set
 714 +# CONFIG_BLK_CPQ_CISS_DA is not set
 715 +# CONFIG_BLK_DEV_DAC960 is not set
 716 +# CONFIG_BLK_DEV_UMEM is not set
 717 +# CONFIG_BLK_DEV_COW_COMMON is not set
 718 +CONFIG_BLK_DEV_LOOP=y
 719 +# CONFIG_BLK_DEV_CRYPTOLOOP is not set
 720 +
 721 +#
 722 +# DRBD disabled because PROC_FS, INET or CONNECTOR not selected
 723 +#
 724 +CONFIG_BLK_DEV_VROOT=y
 725 +# CONFIG_BLK_DEV_NBD is not set
 726 +# CONFIG_BLK_DEV_SX8 is not set
 727 +# CONFIG_BLK_DEV_RAM is not set
 728 +# CONFIG_CDROM_PKTCDVD is not set
 729 +# CONFIG_ATA_OVER_ETH is not set
 730 +# CONFIG_BLK_DEV_HD is not set
 731 +# CONFIG_BLK_DEV_RBD is not set
 732 +# CONFIG_SENSORS_LIS3LV02D is not set
 733 +# CONFIG_MISC_DEVICES is not set
 734 +CONFIG_HAVE_IDE=y
 735 +CONFIG_IDE=y
 736 +
 737 +#
 738 +# Please see Documentation/ide/ide.txt for help/info on IDE drives
 739 +#
 740 +CONFIG_IDE_XFER_MODE=y
 741 +# CONFIG_BLK_DEV_IDE_SATA is not set
 742 +CONFIG_IDE_GD=y
 743 +CONFIG_IDE_GD_ATA=y
 744 +# CONFIG_IDE_GD_ATAPI is not set
 745 +# CONFIG_BLK_DEV_IDECD is not set
 746 +# CONFIG_BLK_DEV_IDETAPE is not set
 747 +# CONFIG_BLK_DEV_IDEACPI is not set
 748 +CONFIG_IDE_TASK_IOCTL=y
 749 +CONFIG_IDE_PROC_FS=y
 750 +
 751 +#
 752 +# IDE chipset support/bugfixes
 753 +#
 754 +CONFIG_IDE_GENERIC=y
 755 +CONFIG_BLK_DEV_PLATFORM=y
 756 +# CONFIG_BLK_DEV_CMD640 is not set
 757 +# CONFIG_BLK_DEV_IDEPNP is not set
 758 +CONFIG_BLK_DEV_IDEDMA_SFF=y
 759 +
 760 +#
 761 +# PCI IDE chipsets support
 762 +#
 763 +CONFIG_BLK_DEV_IDEPCI=y
 764 +CONFIG_IDEPCI_PCIBUS_ORDER=y
 765 +# CONFIG_BLK_DEV_OFFBOARD is not set
 766 +CONFIG_BLK_DEV_GENERIC=y
 767 +# CONFIG_BLK_DEV_OPTI621 is not set
 768 +# CONFIG_BLK_DEV_RZ1000 is not set
 769 +CONFIG_BLK_DEV_IDEDMA_PCI=y
 770 +# CONFIG_BLK_DEV_AEC62XX is not set
 771 +# CONFIG_BLK_DEV_ALI15X3 is not set
 772 +# CONFIG_BLK_DEV_AMD74XX is not set
 773 +# CONFIG_BLK_DEV_ATIIXP is not set
 774 +# CONFIG_BLK_DEV_CMD64X is not set
 775 +# CONFIG_BLK_DEV_TRIFLEX is not set
 776 +# CONFIG_BLK_DEV_CS5520 is not set
 777 +# CONFIG_BLK_DEV_CS5530 is not set
 778 +# CONFIG_BLK_DEV_HPT366 is not set
 779 +# CONFIG_BLK_DEV_JMICRON is not set
 780 +# CONFIG_BLK_DEV_SC1200 is not set
 781 +CONFIG_BLK_DEV_PIIX=y
 782 +# CONFIG_BLK_DEV_IT8172 is not set
 783 +# CONFIG_BLK_DEV_IT8213 is not set
 784 +# CONFIG_BLK_DEV_IT821X is not set
 785 +# CONFIG_BLK_DEV_NS87415 is not set
 786 +# CONFIG_BLK_DEV_PDC202XX_OLD is not set
 787 +# CONFIG_BLK_DEV_PDC202XX_NEW is not set
 788 +# CONFIG_BLK_DEV_SVWKS is not set
 789 +# CONFIG_BLK_DEV_SIIMAGE is not set
 790 +# CONFIG_BLK_DEV_SIS5513 is not set
 791 +# CONFIG_BLK_DEV_SLC90E66 is not set
 792 +# CONFIG_BLK_DEV_TRM290 is not set
 793 +# CONFIG_BLK_DEV_VIA82CXXX is not set
 794 +# CONFIG_BLK_DEV_TC86C001 is not set
 795 +CONFIG_BLK_DEV_IDEDMA=y
 796 +
 797 +#
 798 +# SCSI device support
 799 +#
 800 +CONFIG_SCSI_MOD=y
 801 +# CONFIG_RAID_ATTRS is not set
 802 +CONFIG_SCSI=y
 803 +CONFIG_SCSI_DMA=y
 804 +# CONFIG_SCSI_TGT is not set
 805 +# CONFIG_SCSI_NETLINK is not set
 806 +# CONFIG_SCSI_PROC_FS is not set
 807 +
 808 +#
 809 +# SCSI support type (disk, tape, CD-ROM)
 810 +#
 811 +CONFIG_BLK_DEV_SD=y
 812 +# CONFIG_CHR_DEV_ST is not set
 813 +# CONFIG_CHR_DEV_OSST is not set
 814 +# CONFIG_BLK_DEV_SR is not set
 815 +# CONFIG_CHR_DEV_SG is not set
 816 +# CONFIG_CHR_DEV_SCH is not set
 817 +# CONFIG_SCSI_MULTI_LUN is not set
 818 +# CONFIG_SCSI_CONSTANTS is not set
 819 +# CONFIG_SCSI_LOGGING is not set
 820 +# CONFIG_SCSI_SCAN_ASYNC is not set
 821 +
 822 +#
 823 +# SCSI Transports
 824 +#
 825 +# CONFIG_SCSI_SPI_ATTRS is not set
 826 +# CONFIG_SCSI_FC_ATTRS is not set
 827 +# CONFIG_SCSI_ISCSI_ATTRS is not set
 828 +# CONFIG_SCSI_SAS_ATTRS is not set
 829 +# CONFIG_SCSI_SAS_LIBSAS is not set
 830 +# CONFIG_SCSI_SRP_ATTRS is not set
 831 +# CONFIG_SCSI_LOWLEVEL is not set
 832 +# CONFIG_SCSI_DH is not set
 833 +# CONFIG_SCSI_OSD_INITIATOR is not set
 834 +CONFIG_ATA=y
 835 +# CONFIG_ATA_NONSTANDARD is not set
 836 +CONFIG_ATA_VERBOSE_ERROR=y
 837 +CONFIG_ATA_ACPI=y
 838 +CONFIG_SATA_PMP=y
 839 +
 840 +#
 841 +# Controllers with non-SFF native interface
 842 +#
 843 +CONFIG_SATA_AHCI=y
 844 +# CONFIG_SATA_AHCI_PLATFORM is not set
 845 +# CONFIG_SATA_INIC162X is not set
 846 +# CONFIG_SATA_ACARD_AHCI is not set
 847 +# CONFIG_SATA_SIL24 is not set
 848 +# CONFIG_ATA_SFF is not set
 849 +CONFIG_MD=y
 850 +CONFIG_BLK_DEV_MD=y
 851 +CONFIG_MD_AUTODETECT=y
 852 +# CONFIG_MD_LINEAR is not set
 853 +CONFIG_MD_RAID0=y
 854 +CONFIG_MD_RAID1=y
 855 +CONFIG_MD_RAID10=y
 856 +CONFIG_MD_RAID456=y
 857 +# CONFIG_MULTICORE_RAID456 is not set
 858 +# CONFIG_MD_MULTIPATH is not set
 859 +# CONFIG_MD_FAULTY is not set
 860 +CONFIG_BLK_DEV_DM=y
 861 +# CONFIG_DM_DEBUG is not set
 862 +# CONFIG_DM_CRYPT is not set
 863 +CONFIG_DM_SNAPSHOT=y
 864 +CONFIG_DM_MIRROR=y
 865 +CONFIG_DM_RAID=y
 866 +# CONFIG_DM_LOG_USERSPACE is not set
 867 +# CONFIG_DM_ZERO is not set
 868 +# CONFIG_DM_MULTIPATH is not set
 869 +# CONFIG_DM_DELAY is not set
 870 +# CONFIG_DM_UEVENT is not set
 871 +# CONFIG_DM_FLAKEY is not set
 872 +# CONFIG_TARGET_CORE is not set
 873 +# CONFIG_FUSION is not set
 874 +
 875 +#
 876 +# IEEE 1394 (FireWire) support
 877 +#
 878 +# CONFIG_FIREWIRE is not set
 879 +# CONFIG_FIREWIRE_NOSY is not set
 880 +# CONFIG_I2O is not set
 881 +# CONFIG_MACINTOSH_DRIVERS is not set
 882 +CONFIG_NETDEVICES=y
 883 +# CONFIG_DUMMY is not set
 884 +# CONFIG_BONDING is not set
 885 +# CONFIG_MACVLAN is not set
 886 +# CONFIG_EQUALIZER is not set
 887 +CONFIG_TUN=y
 888 +# CONFIG_VETH is not set
 889 +# CONFIG_NET_SB1000 is not set
 890 +# CONFIG_ARCNET is not set
 891 +CONFIG_MII=y
 892 +# CONFIG_PHYLIB is not set
 893 +# CONFIG_NET_ETHERNET is not set
 894 +CONFIG_NETDEV_1000=y
 895 +# CONFIG_ACENIC is not set
 896 +# CONFIG_DL2K is not set
 897 +# CONFIG_E1000 is not set
 898 +# CONFIG_E1000E is not set
 899 +# CONFIG_IP1000 is not set
 900 +# CONFIG_IGB is not set
 901 +# CONFIG_IGBVF is not set
 902 +# CONFIG_NS83820 is not set
 903 +# CONFIG_HAMACHI is not set
 904 +# CONFIG_YELLOWFIN is not set
 905 +CONFIG_R8169=y
 906 +# CONFIG_SIS190 is not set
 907 +# CONFIG_SKGE is not set
 908 +# CONFIG_SKY2 is not set
 909 +# CONFIG_VIA_VELOCITY is not set
 910 +# CONFIG_TIGON3 is not set
 911 +# CONFIG_BNX2 is not set
 912 +# CONFIG_CNIC is not set
 913 +# CONFIG_QLA3XXX is not set
 914 +# CONFIG_ATL1 is not set
 915 +# CONFIG_ATL1E is not set
 916 +# CONFIG_ATL1C is not set
 917 +# CONFIG_JME is not set
 918 +# CONFIG_STMMAC_ETH is not set
 919 +# CONFIG_PCH_GBE is not set
 920 +# CONFIG_NETDEV_10000 is not set
 921 +# CONFIG_TR is not set
 922 +# CONFIG_WLAN is not set
 923 +
 924 +#
 925 +# Enable WiMAX (Networking options) to see the WiMAX drivers
 926 +#
 927 +# CONFIG_WAN is not set
 928 +
 929 +#
 930 +# CAIF transport drivers
 931 +#
 932 +# CONFIG_FDDI is not set
 933 +# CONFIG_HIPPI is not set
 934 +# CONFIG_PPP is not set
 935 +# CONFIG_SLIP is not set
 936 +# CONFIG_NET_FC is not set
 937 +# CONFIG_NETCONSOLE is not set
 938 +# CONFIG_NETPOLL is not set
 939 +# CONFIG_NET_POLL_CONTROLLER is not set
 940 +# CONFIG_VMXNET3 is not set
 941 +# CONFIG_ISDN is not set
 942 +# CONFIG_PHONE is not set
 943 +
 944 +#
 945 +# Input device support
 946 +#
 947 +CONFIG_INPUT=y
 948 +# CONFIG_INPUT_FF_MEMLESS is not set
 949 +# CONFIG_INPUT_POLLDEV is not set
 950 +# CONFIG_INPUT_SPARSEKMAP is not set
 951 +
 952 +#
 953 +# Userland interfaces
 954 +#
 955 +CONFIG_INPUT_MOUSEDEV=y
 956 +# CONFIG_INPUT_MOUSEDEV_PSAUX is not set
 957 +CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
 958 +CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
 959 +# CONFIG_INPUT_JOYDEV is not set
 960 +# CONFIG_INPUT_EVDEV is not set
 961 +# CONFIG_INPUT_EVBUG is not set
 962 +
 963 +#
 964 +# Input Device Drivers
 965 +#
 966 +CONFIG_INPUT_KEYBOARD=y
 967 +CONFIG_KEYBOARD_ATKBD=y
 968 +# CONFIG_KEYBOARD_LKKBD is not set
 969 +# CONFIG_KEYBOARD_NEWTON is not set
 970 +# CONFIG_KEYBOARD_OPENCORES is not set
 971 +# CONFIG_KEYBOARD_STOWAWAY is not set
 972 +# CONFIG_KEYBOARD_SUNKBD is not set
 973 +# CONFIG_KEYBOARD_XTKBD is not set
 974 +# CONFIG_INPUT_MOUSE is not set
 975 +# CONFIG_INPUT_JOYSTICK is not set
 976 +# CONFIG_INPUT_TABLET is not set
 977 +# CONFIG_INPUT_TOUCHSCREEN is not set
 978 +# CONFIG_INPUT_MISC is not set
 979 +
 980 +#
 981 +# Hardware I/O ports
 982 +#
 983 +CONFIG_SERIO=y
 984 +CONFIG_SERIO_I8042=y
 985 +# CONFIG_SERIO_SERPORT is not set
 986 +# CONFIG_SERIO_CT82C710 is not set
 987 +# CONFIG_SERIO_PCIPS2 is not set
 988 +CONFIG_SERIO_LIBPS2=y
 989 +# CONFIG_SERIO_RAW is not set
 990 +# CONFIG_SERIO_ALTERA_PS2 is not set
 991 +# CONFIG_SERIO_PS2MULT is not set
 992 +# CONFIG_GAMEPORT is not set
 993 +
 994 +#
 995 +# Character devices
 996 +#
 997 +CONFIG_VT=y
 998 +CONFIG_CONSOLE_TRANSLATIONS=y
 999 +CONFIG_VT_CONSOLE=y
1000 +CONFIG_HW_CONSOLE=y
1001 +# CONFIG_VT_HW_CONSOLE_BINDING is not set
1002 +CONFIG_UNIX98_PTYS=y
1003 +CONFIG_DEVPTS_MULTIPLE_INSTANCES=y
1004 +CONFIG_LEGACY_PTYS=y
1005 +CONFIG_LEGACY_PTY_COUNT=0
1006 +# CONFIG_SERIAL_NONSTANDARD is not set
1007 +# CONFIG_NOZOMI is not set
1008 +# CONFIG_N_GSM is not set
1009 +# CONFIG_TRACE_SINK is not set
1010 +# CONFIG_DEVKMEM is not set
1011 +
1012 +#
1013 +# Serial drivers
1014 +#
1015 +CONFIG_SERIAL_8250=y
1016 +CONFIG_SERIAL_8250_CONSOLE=y
1017 +CONFIG_FIX_EARLYCON_MEM=y
1018 +CONFIG_SERIAL_8250_PCI=y
1019 +CONFIG_SERIAL_8250_PNP=y
1020 +CONFIG_SERIAL_8250_NR_UARTS=4
1021 +CONFIG_SERIAL_8250_RUNTIME_UARTS=4
1022 +# CONFIG_SERIAL_8250_EXTENDED is not set
1023 +
1024 +#
1025 +# Non-8250 serial port support
1026 +#
1027 +# CONFIG_SERIAL_MFD_HSU is not set
1028 +CONFIG_SERIAL_CORE=y
1029 +CONFIG_SERIAL_CORE_CONSOLE=y
1030 +# CONFIG_SERIAL_JSM is not set
1031 +# CONFIG_SERIAL_TIMBERDALE is not set
1032 +# CONFIG_SERIAL_ALTERA_JTAGUART is not set
1033 +# CONFIG_SERIAL_ALTERA_UART is not set
1034 +# CONFIG_SERIAL_PCH_UART is not set
1035 +# CONFIG_SERIAL_XILINX_PS_UART is not set
1036 +CONFIG_TTY_PRINTK=y
1037 +# CONFIG_IPMI_HANDLER is not set
1038 +# CONFIG_HW_RANDOM is not set
1039 +# CONFIG_NVRAM is not set
1040 +CONFIG_RTC=y
1041 +# CONFIG_R3964 is not set
1042 +# CONFIG_APPLICOM is not set
1043 +# CONFIG_MWAVE is not set
1044 +# CONFIG_RAW_DRIVER is not set
1045 +# CONFIG_HPET is not set
1046 +# CONFIG_HANGCHECK_TIMER is not set
1047 +# CONFIG_TCG_TPM is not set
1048 +# CONFIG_TELCLOCK is not set
1049 +CONFIG_DEVPORT=y
1050 +CONFIG_RAMOOPS=y
1051 +# CONFIG_I2C is not set
1052 +# CONFIG_SPI is not set
1053 +
1054 +#
1055 +# PPS support
1056 +#
1057 +# CONFIG_PPS is not set
1058 +
1059 +#
1060 +# PPS generators support
1061 +#
1062 +
1063 +#
1064 +# PTP clock support
1065 +#
1066 +
1067 +#
1068 +# Enable Device Drivers -> PPS to see the PTP clock options.
1069 +#
1070 +CONFIG_ARCH_WANT_OPTIONAL_GPIOLIB=y
1071 +# CONFIG_GPIOLIB is not set
1072 +# CONFIG_W1 is not set
1073 +CONFIG_POWER_SUPPLY=y
1074 +# CONFIG_POWER_SUPPLY_DEBUG is not set
1075 +# CONFIG_PDA_POWER is not set
1076 +# CONFIG_TEST_POWER is not set
1077 +# CONFIG_BATTERY_DS2780 is not set
1078 +# CONFIG_BATTERY_BQ27x00 is not set
1079 +# CONFIG_CHARGER_MAX8903 is not set
1080 +# CONFIG_HWMON is not set
1081 +CONFIG_THERMAL=y
1082 +# CONFIG_WATCHDOG is not set
1083 +CONFIG_SSB_POSSIBLE=y
1084 +
1085 +#
1086 +# Sonics Silicon Backplane
1087 +#
1088 +# CONFIG_SSB is not set
1089 +CONFIG_BCMA_POSSIBLE=y
1090 +
1091 +#
1092 +# Broadcom specific AMBA
1093 +#
1094 +# CONFIG_BCMA is not set
1095 +# CONFIG_MFD_SUPPORT is not set
1096 +# CONFIG_REGULATOR is not set
1097 +# CONFIG_MEDIA_SUPPORT is not set
1098 +
1099 +#
1100 +# Graphics support
1101 +#
1102 +# CONFIG_AGP is not set
1103 +CONFIG_VGA_ARB=y
1104 +CONFIG_VGA_ARB_MAX_GPUS=2
1105 +# CONFIG_VGA_SWITCHEROO is not set
1106 +# CONFIG_DRM is not set
1107 +# CONFIG_STUB_POULSBO is not set
1108 +# CONFIG_VGASTATE is not set
1109 +# CONFIG_VIDEO_OUTPUT_CONTROL is not set
1110 +# CONFIG_FB is not set
1111 +# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
1112 +
1113 +#
1114 +# Display device support
1115 +#
1116 +# CONFIG_DISPLAY_SUPPORT is not set
1117 +
1118 +#
1119 +# Console display driver support
1120 +#
1121 +CONFIG_VGA_CONSOLE=y
1122 +# CONFIG_VGACON_SOFT_SCROLLBACK is not set
1123 +CONFIG_DUMMY_CONSOLE=y
1124 +# CONFIG_SOUND is not set
1125 +# CONFIG_HID_SUPPORT is not set
1126 +# CONFIG_USB_SUPPORT is not set
1127 +# CONFIG_UWB is not set
1128 +# CONFIG_MMC is not set
1129 +# CONFIG_MEMSTICK is not set
1130 +# CONFIG_NEW_LEDS is not set
1131 +# CONFIG_NFC_DEVICES is not set
1132 +# CONFIG_ACCESSIBILITY is not set
1133 +# CONFIG_INFINIBAND is not set
1134 +# CONFIG_EDAC is not set
1135 +# CONFIG_RTC_CLASS is not set
1136 +# CONFIG_DMADEVICES is not set
1137 +# CONFIG_AUXDISPLAY is not set
1138 +# CONFIG_UIO is not set
1139 +# CONFIG_STAGING is not set
1140 +# CONFIG_X86_PLATFORM_DEVICES is not set
1141 +
1142 +#
1143 +# Firmware Drivers
1144 +#
1145 +# CONFIG_EDD is not set
1146 +CONFIG_FIRMWARE_MEMMAP=y
1147 +# CONFIG_DELL_RBU is not set
1148 +# CONFIG_DCDBAS is not set
1149 +# CONFIG_DMIID is not set
1150 +# CONFIG_DMI_SYSFS is not set
1151 +# CONFIG_ISCSI_IBFT_FIND is not set
1152 +# CONFIG_GOOGLE_FIRMWARE is not set
1153 +
1154 +#
1155 +# File systems
1156 +#
1157 +CONFIG_EXT2_FS=y
1158 +CONFIG_EXT2_FS_XATTR=y
1159 +CONFIG_EXT2_FS_POSIX_ACL=y
1160 +CONFIG_EXT2_FS_SECURITY=y
1161 +CONFIG_EXT2_FS_XIP=y
1162 +CONFIG_EXT3_FS=y
1163 +CONFIG_EXT3_DEFAULTS_TO_ORDERED=y
1164 +CONFIG_EXT3_FS_XATTR=y
1165 +CONFIG_EXT3_FS_POSIX_ACL=y
1166 +CONFIG_EXT3_FS_SECURITY=y
1167 +CONFIG_EXT4_FS=y
1168 +CONFIG_EXT4_FS_XATTR=y
1169 +CONFIG_EXT4_FS_POSIX_ACL=y
1170 +CONFIG_EXT4_FS_SECURITY=y
1171 +# CONFIG_EXT4_DEBUG is not set
1172 +CONFIG_FS_XIP=y
1173 +CONFIG_JBD=y
1174 +CONFIG_JBD2=y
1175 +CONFIG_FS_MBCACHE=y
1176 +CONFIG_REISERFS_FS=y
1177 +CONFIG_REISERFS_CHECK=y
1178 +CONFIG_REISERFS_PROC_INFO=y
1179 +CONFIG_REISERFS_FS_XATTR=y
1180 +CONFIG_REISERFS_FS_POSIX_ACL=y
1181 +CONFIG_REISERFS_FS_SECURITY=y
1182 +CONFIG_JFS_FS=y
1183 +CONFIG_JFS_POSIX_ACL=y
1184 +CONFIG_JFS_SECURITY=y
1185 +CONFIG_JFS_DEBUG=y
1186 +CONFIG_JFS_STATISTICS=y
1187 +CONFIG_XFS_FS=y
1188 +CONFIG_XFS_QUOTA=y
1189 +CONFIG_XFS_POSIX_ACL=y
1190 +CONFIG_XFS_RT=y
1191 +# CONFIG_XFS_DEBUG is not set
1192 +CONFIG_GFS2_FS=y
1193 +CONFIG_GFS2_FS_LOCKING_DLM=y
1194 +CONFIG_OCFS2_FS=y
1195 +CONFIG_OCFS2_FS_O2CB=y
1196 +CONFIG_OCFS2_FS_USERSPACE_CLUSTER=y
1197 +CONFIG_OCFS2_DEBUG_MASKLOG=y
1198 +CONFIG_OCFS2_DEBUG_FS=y
1199 +CONFIG_BTRFS_FS=y
1200 +CONFIG_BTRFS_FS_POSIX_ACL=y
1201 +# CONFIG_NILFS2_FS is not set
1202 +CONFIG_FS_POSIX_ACL=y
1203 +CONFIG_EXPORTFS=y
1204 +CONFIG_FILE_LOCKING=y
1205 +CONFIG_FSNOTIFY=y
1206 +CONFIG_DNOTIFY=y
1207 +CONFIG_INOTIFY_USER=y
1208 +CONFIG_FANOTIFY=y
1209 +CONFIG_QUOTA=y
1210 +CONFIG_QUOTA_NETLINK_INTERFACE=y
1211 +CONFIG_PRINT_QUOTA_WARNING=y
1212 +# CONFIG_QUOTA_DEBUG is not set
1213 +CONFIG_QUOTA_TREE=y
1214 +CONFIG_QFMT_V1=y
1215 +CONFIG_QFMT_V2=y
1216 +CONFIG_QUOTACTL=y
1217 +CONFIG_QUOTACTL_COMPAT=y
1218 +# CONFIG_AUTOFS4_FS is not set
1219 +# CONFIG_FUSE_FS is not set
1220 +CONFIG_GENERIC_ACL=y
1221 +
1222 +#
1223 +# Caches
1224 +#
1225 +# CONFIG_FSCACHE is not set
1226 +
1227 +#
1228 +# CD-ROM/DVD Filesystems
1229 +#
1230 +# CONFIG_ISO9660_FS is not set
1231 +# CONFIG_UDF_FS is not set
1232 +
1233 +#
1234 +# DOS/FAT/NT Filesystems
1235 +#
1236 +# CONFIG_MSDOS_FS is not set
1237 +# CONFIG_VFAT_FS is not set
1238 +# CONFIG_NTFS_FS is not set
1239 +
1240 +#
1241 +# Pseudo filesystems
1242 +#
1243 +CONFIG_PROC_FS=y
1244 +# CONFIG_PROC_KCORE is not set
1245 +CONFIG_PROC_SYSCTL=y
1246 +CONFIG_PROC_PAGE_MONITOR=y
1247 +CONFIG_SYSFS=y
1248 +CONFIG_TMPFS=y
1249 +CONFIG_TMPFS_POSIX_ACL=y
1250 +CONFIG_TMPFS_XATTR=y
1251 +# CONFIG_HUGETLBFS is not set
1252 +# CONFIG_HUGETLB_PAGE is not set
1253 +CONFIG_CONFIGFS_FS=y
1254 +# CONFIG_MISC_FILESYSTEMS is not set
1255 +CONFIG_NETWORK_FILESYSTEMS=y
1256 +CONFIG_NFS_FS=y
1257 +CONFIG_NFS_V3=y
1258 +# CONFIG_NFS_V3_ACL is not set
1259 +# CONFIG_NFS_V4 is not set
1260 +CONFIG_NFSD=y
1261 +CONFIG_NFSD_DEPRECATED=y
1262 +CONFIG_NFSD_V3=y
1263 +# CONFIG_NFSD_V3_ACL is not set
1264 +# CONFIG_NFSD_V4 is not set
1265 +CONFIG_LOCKD=y
1266 +CONFIG_LOCKD_V4=y
1267 +CONFIG_NFS_COMMON=y
1268 +CONFIG_SUNRPC=y
1269 +# CONFIG_CEPH_FS is not set
1270 +# CONFIG_CIFS is not set
1271 +# CONFIG_NCP_FS is not set
1272 +# CONFIG_CODA_FS is not set
1273 +# CONFIG_AFS_FS is not set
1274 +
1275 +#
1276 +# Partition Types
1277 +#
1278 +# CONFIG_PARTITION_ADVANCED is not set
1279 +CONFIG_MSDOS_PARTITION=y
1280 +CONFIG_NLS=y
1281 +CONFIG_NLS_DEFAULT="iso8859-1"
1282 +# CONFIG_NLS_CODEPAGE_437 is not set
1283 +# CONFIG_NLS_CODEPAGE_737 is not set
1284 +# CONFIG_NLS_CODEPAGE_775 is not set
1285 +# CONFIG_NLS_CODEPAGE_850 is not set
1286 +# CONFIG_NLS_CODEPAGE_852 is not set
1287 +# CONFIG_NLS_CODEPAGE_855 is not set
1288 +# CONFIG_NLS_CODEPAGE_857 is not set
1289 +# CONFIG_NLS_CODEPAGE_860 is not set
1290 +# CONFIG_NLS_CODEPAGE_861 is not set
1291 +# CONFIG_NLS_CODEPAGE_862 is not set
1292 +# CONFIG_NLS_CODEPAGE_863 is not set
1293 +# CONFIG_NLS_CODEPAGE_864 is not set
1294 +# CONFIG_NLS_CODEPAGE_865 is not set
1295 +# CONFIG_NLS_CODEPAGE_866 is not set
1296 +# CONFIG_NLS_CODEPAGE_869 is not set
1297 +# CONFIG_NLS_CODEPAGE_936 is not set
1298 +# CONFIG_NLS_CODEPAGE_950 is not set
1299 +# CONFIG_NLS_CODEPAGE_932 is not set
1300 +# CONFIG_NLS_CODEPAGE_949 is not set
1301 +# CONFIG_NLS_CODEPAGE_874 is not set
1302 +# CONFIG_NLS_ISO8859_8 is not set
1303 +# CONFIG_NLS_CODEPAGE_1250 is not set
1304 +# CONFIG_NLS_CODEPAGE_1251 is not set
1305 +# CONFIG_NLS_ASCII is not set
1306 +# CONFIG_NLS_ISO8859_1 is not set
1307 +# CONFIG_NLS_ISO8859_2 is not set
1308 +# CONFIG_NLS_ISO8859_3 is not set
1309 +# CONFIG_NLS_ISO8859_4 is not set
1310 +# CONFIG_NLS_ISO8859_5 is not set
1311 +# CONFIG_NLS_ISO8859_6 is not set
1312 +# CONFIG_NLS_ISO8859_7 is not set
1313 +# CONFIG_NLS_ISO8859_9 is not set
1314 +# CONFIG_NLS_ISO8859_13 is not set
1315 +# CONFIG_NLS_ISO8859_14 is not set
1316 +# CONFIG_NLS_ISO8859_15 is not set
1317 +# CONFIG_NLS_KOI8_R is not set
1318 +# CONFIG_NLS_KOI8_U is not set
1319 +# CONFIG_NLS_UTF8 is not set
1320 +CONFIG_DLM=y
1321 +# CONFIG_DLM_DEBUG is not set
1322 +
1323 +#
1324 +# Kernel hacking
1325 +#
1326 +CONFIG_TRACE_IRQFLAGS_SUPPORT=y
1327 +CONFIG_PRINTK_TIME=y
1328 +CONFIG_DEFAULT_MESSAGE_LOGLEVEL=4
1329 +# CONFIG_ENABLE_WARN_DEPRECATED is not set
1330 +# CONFIG_ENABLE_MUST_CHECK is not set
1331 +CONFIG_FRAME_WARN=2048
1332 +CONFIG_MAGIC_SYSRQ=y
1333 +# CONFIG_STRIP_ASM_SYMS is not set
1334 +# CONFIG_UNUSED_SYMBOLS is not set
1335 +# CONFIG_DEBUG_FS is not set
1336 +# CONFIG_HEADERS_CHECK is not set
1337 +# CONFIG_DEBUG_SECTION_MISMATCH is not set
1338 +CONFIG_DEBUG_KERNEL=y
1339 +# CONFIG_DEBUG_SHIRQ is not set
1340 +# CONFIG_LOCKUP_DETECTOR is not set
1341 +# CONFIG_HARDLOCKUP_DETECTOR is not set
1342 +CONFIG_DETECT_HUNG_TASK=y
1343 +CONFIG_DEFAULT_HUNG_TASK_TIMEOUT=120
1344 +# CONFIG_BOOTPARAM_HUNG_TASK_PANIC is not set
1345 +CONFIG_BOOTPARAM_HUNG_TASK_PANIC_VALUE=0
1346 +# CONFIG_SCHED_DEBUG is not set
1347 +# CONFIG_SCHEDSTATS is not set
1348 +# CONFIG_TIMER_STATS is not set
1349 +# CONFIG_DEBUG_OBJECTS is not set
1350 +# CONFIG_SLUB_DEBUG_ON is not set
1351 +# CONFIG_SLUB_STATS is not set
1352 +# CONFIG_DEBUG_KMEMLEAK is not set
1353 +# CONFIG_DEBUG_RT_MUTEXES is not set
1354 +# CONFIG_RT_MUTEX_TESTER is not set
1355 +# CONFIG_DEBUG_SPINLOCK is not set
1356 +# CONFIG_DEBUG_MUTEXES is not set
1357 +# CONFIG_DEBUG_LOCK_ALLOC is not set
1358 +# CONFIG_PROVE_LOCKING is not set
1359 +# CONFIG_SPARSE_RCU_POINTER is not set
1360 +# CONFIG_LOCK_STAT is not set
1361 +# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
1362 +# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
1363 +# CONFIG_DEBUG_STACK_USAGE is not set
1364 +# CONFIG_DEBUG_KOBJECT is not set
1365 +CONFIG_DEBUG_BUGVERBOSE=y
1366 +CONFIG_DEBUG_INFO=y
1367 +# CONFIG_DEBUG_INFO_REDUCED is not set
1368 +# CONFIG_DEBUG_VM is not set
1369 +# CONFIG_DEBUG_VIRTUAL is not set
1370 +# CONFIG_DEBUG_WRITECOUNT is not set
1371 +CONFIG_DEBUG_MEMORY_INIT=y
1372 +# CONFIG_DEBUG_LIST is not set
1373 +# CONFIG_TEST_LIST_SORT is not set
1374 +# CONFIG_DEBUG_SG is not set
1375 +# CONFIG_DEBUG_NOTIFIERS is not set
1376 +# CONFIG_DEBUG_CREDENTIALS is not set
1377 +CONFIG_ARCH_WANT_FRAME_POINTERS=y
1378 +# CONFIG_FRAME_POINTER is not set
1379 +# CONFIG_BOOT_PRINTK_DELAY is not set
1380 +# CONFIG_RCU_TORTURE_TEST is not set
1381 +CONFIG_RCU_CPU_STALL_TIMEOUT=60
1382 +# CONFIG_BACKTRACE_SELF_TEST is not set
1383 +# CONFIG_DEBUG_BLOCK_EXT_DEVT is not set
1384 +# CONFIG_DEBUG_FORCE_WEAK_PER_CPU is not set
1385 +# CONFIG_DEBUG_PER_CPU_MAPS is not set
1386 +# CONFIG_CPU_NOTIFIER_ERROR_INJECT is not set
1387 +# CONFIG_FAULT_INJECTION is not set
1388 +# CONFIG_LATENCYTOP is not set
1389 +# CONFIG_SYSCTL_SYSCALL_CHECK is not set
1390 +# CONFIG_DEBUG_PAGEALLOC is not set
1391 +CONFIG_USER_STACKTRACE_SUPPORT=y
1392 +CONFIG_HAVE_FUNCTION_TRACER=y
1393 +CONFIG_HAVE_FUNCTION_GRAPH_TRACER=y
1394 +CONFIG_HAVE_FUNCTION_GRAPH_FP_TEST=y
1395 +CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST=y
1396 +CONFIG_HAVE_DYNAMIC_FTRACE=y
1397 +CONFIG_HAVE_FTRACE_MCOUNT_RECORD=y
1398 +CONFIG_HAVE_SYSCALL_TRACEPOINTS=y
1399 +CONFIG_HAVE_C_RECORDMCOUNT=y
1400 +CONFIG_TRACING_SUPPORT=y
1401 +# CONFIG_FTRACE is not set
1402 +# CONFIG_PROVIDE_OHCI1394_DMA_INIT is not set
1403 +# CONFIG_DMA_API_DEBUG is not set
1404 +CONFIG_ATOMIC64_SELFTEST=y
1405 +# CONFIG_ASYNC_RAID6_TEST is not set
1406 +# CONFIG_SAMPLES is not set
1407 +CONFIG_HAVE_ARCH_KGDB=y
1408 +# CONFIG_KGDB is not set
1409 +CONFIG_HAVE_ARCH_KMEMCHECK=y
1410 +# CONFIG_TEST_KSTRTOX is not set
1411 +# CONFIG_STRICT_DEVMEM is not set
1412 +# CONFIG_X86_VERBOSE_BOOTUP is not set
1413 +CONFIG_EARLY_PRINTK=y
1414 +# CONFIG_EARLY_PRINTK_DBGP is not set
1415 +# CONFIG_DEBUG_STACKOVERFLOW is not set
1416 +# CONFIG_X86_PTDUMP is not set
1417 +CONFIG_DEBUG_RODATA=y
1418 +CONFIG_DEBUG_RODATA_TEST=y
1419 +# CONFIG_IOMMU_DEBUG is not set
1420 +# CONFIG_IOMMU_STRESS is not set
1421 +CONFIG_HAVE_MMIOTRACE_SUPPORT=y
1422 +CONFIG_IO_DELAY_TYPE_0X80=0
1423 +CONFIG_IO_DELAY_TYPE_0XED=1
1424 +CONFIG_IO_DELAY_TYPE_UDELAY=2
1425 +CONFIG_IO_DELAY_TYPE_NONE=3
1426 +CONFIG_IO_DELAY_0X80=y
1427 +# CONFIG_IO_DELAY_0XED is not set
1428 +# CONFIG_IO_DELAY_UDELAY is not set
1429 +# CONFIG_IO_DELAY_NONE is not set
1430 +CONFIG_DEFAULT_IO_DELAY_TYPE=0
1431 +# CONFIG_CPA_DEBUG is not set
1432 +# CONFIG_OPTIMIZE_INLINING is not set
1433 +# CONFIG_DEBUG_STRICT_USER_COPY_CHECKS is not set
1434 +
1435 +#
1436 +# Linux VServer
1437 +#
1438 +CONFIG_VSERVER_AUTO_LBACK=y
1439 +# CONFIG_VSERVER_AUTO_SINGLE is not set
1440 +CONFIG_VSERVER_COWBL=y
1441 +CONFIG_VSERVER_VTIME=y
1442 +# CONFIG_VSERVER_DEVICE is not set
1443 +CONFIG_VSERVER_PROC_SECURE=y
1444 +# CONFIG_TAGGING_NONE is not set
1445 +# CONFIG_TAGGING_UID16 is not set
1446 +# CONFIG_TAGGING_GID16 is not set
1447 +CONFIG_TAGGING_ID24=y
1448 +# CONFIG_TAGGING_INTERN is not set
1449 +# CONFIG_TAG_NFSD is not set
1450 +# CONFIG_VSERVER_PRIVACY is not set
1451 +CONFIG_VSERVER_CONTEXTS=768
1452 +CONFIG_VSERVER_WARN=y
1453 +CONFIG_VSERVER_WARN_DEVPTS=y
1454 +CONFIG_VSERVER_DEBUG=y
1455 +CONFIG_VSERVER_HISTORY=y
1456 +CONFIG_VSERVER_HISTORY_SIZE=64
1457 +CONFIG_VSERVER_LEGACY_MEM=y
1458 +# CONFIG_QUOTES_ISO8859 is not set
1459 +# CONFIG_QUOTES_UTF8 is not set
1460 +CONFIG_QUOTES_ASCII=y
1461 +CONFIG_VSERVER=y
1462 +
1463 +#
1464 +# Security options
1465 +#
1466 +# CONFIG_KEYS is not set
1467 +CONFIG_SECURITY_DMESG_RESTRICT=y
1468 +# CONFIG_SECURITY is not set
1469 +# CONFIG_SECURITYFS is not set
1470 +CONFIG_DEFAULT_SECURITY_DAC=y
1471 +CONFIG_DEFAULT_SECURITY=""
1472 +CONFIG_XOR_BLOCKS=y
1473 +CONFIG_ASYNC_CORE=y
1474 +CONFIG_ASYNC_MEMCPY=y
1475 +CONFIG_ASYNC_XOR=y
1476 +CONFIG_ASYNC_PQ=y
1477 +CONFIG_ASYNC_RAID6_RECOV=y
1478 +CONFIG_CRYPTO=y
1479 +
1480 +#
1481 +# Crypto core or helper
1482 +#
1483 +CONFIG_CRYPTO_ALGAPI=y
1484 +CONFIG_CRYPTO_ALGAPI2=y
1485 +CONFIG_CRYPTO_AEAD=y
1486 +CONFIG_CRYPTO_AEAD2=y
1487 +CONFIG_CRYPTO_BLKCIPHER=y
1488 +CONFIG_CRYPTO_BLKCIPHER2=y
1489 +CONFIG_CRYPTO_HASH=y
1490 +CONFIG_CRYPTO_HASH2=y
1491 +CONFIG_CRYPTO_RNG=y
1492 +CONFIG_CRYPTO_RNG2=y
1493 +CONFIG_CRYPTO_PCOMP2=y
1494 +CONFIG_CRYPTO_MANAGER=y
1495 +CONFIG_CRYPTO_MANAGER2=y
1496 +CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
1497 +# CONFIG_CRYPTO_GF128MUL is not set
1498 +# CONFIG_CRYPTO_NULL is not set
1499 +# CONFIG_CRYPTO_PCRYPT is not set
1500 +CONFIG_CRYPTO_WORKQUEUE=y
1501 +# CONFIG_CRYPTO_CRYPTD is not set
1502 +CONFIG_CRYPTO_AUTHENC=y
1503 +
1504 +#
1505 +# Authenticated Encryption with Associated Data
1506 +#
1507 +# CONFIG_CRYPTO_CCM is not set
1508 +# CONFIG_CRYPTO_GCM is not set
1509 +# CONFIG_CRYPTO_SEQIV is not set
1510 +
1511 +#
1512 +# Block modes
1513 +#
1514 +CONFIG_CRYPTO_CBC=y
1515 +# CONFIG_CRYPTO_CTR is not set
1516 +# CONFIG_CRYPTO_CTS is not set
1517 +# CONFIG_CRYPTO_ECB is not set
1518 +# CONFIG_CRYPTO_LRW is not set
1519 +# CONFIG_CRYPTO_PCBC is not set
1520 +# CONFIG_CRYPTO_XTS is not set
1521 +
1522 +#
1523 +# Hash modes
1524 +#
1525 +CONFIG_CRYPTO_HMAC=y
1526 +# CONFIG_CRYPTO_XCBC is not set
1527 +# CONFIG_CRYPTO_VMAC is not set
1528 +
1529 +#
1530 +# Digest
1531 +#
1532 +CONFIG_CRYPTO_CRC32C=y
1533 +# CONFIG_CRYPTO_CRC32C_INTEL is not set
1534 +# CONFIG_CRYPTO_GHASH is not set
1535 +# CONFIG_CRYPTO_MD4 is not set
1536 +CONFIG_CRYPTO_MD5=y
1537 +# CONFIG_CRYPTO_MICHAEL_MIC is not set
1538 +# CONFIG_CRYPTO_RMD128 is not set
1539 +# CONFIG_CRYPTO_RMD160 is not set
1540 +# CONFIG_CRYPTO_RMD256 is not set
1541 +# CONFIG_CRYPTO_RMD320 is not set
1542 +CONFIG_CRYPTO_SHA1=y
1543 +# CONFIG_CRYPTO_SHA256 is not set
1544 +# CONFIG_CRYPTO_SHA512 is not set
1545 +# CONFIG_CRYPTO_TGR192 is not set
1546 +# CONFIG_CRYPTO_WP512 is not set
1547 +# CONFIG_CRYPTO_GHASH_CLMUL_NI_INTEL is not set
1548 +
1549 +#
1550 +# Ciphers
1551 +#
1552 +CONFIG_CRYPTO_AES=y
1553 +# CONFIG_CRYPTO_AES_X86_64 is not set
1554 +# CONFIG_CRYPTO_AES_NI_INTEL is not set
1555 +# CONFIG_CRYPTO_ANUBIS is not set
1556 +# CONFIG_CRYPTO_ARC4 is not set
1557 +# CONFIG_CRYPTO_BLOWFISH is not set
1558 +# CONFIG_CRYPTO_CAMELLIA is not set
1559 +# CONFIG_CRYPTO_CAST5 is not set
1560 +# CONFIG_CRYPTO_CAST6 is not set
1561 +CONFIG_CRYPTO_DES=y
1562 +# CONFIG_CRYPTO_FCRYPT is not set
1563 +# CONFIG_CRYPTO_KHAZAD is not set
1564 +# CONFIG_CRYPTO_SALSA20 is not set
1565 +# CONFIG_CRYPTO_SALSA20_X86_64 is not set
1566 +# CONFIG_CRYPTO_SEED is not set
1567 +# CONFIG_CRYPTO_SERPENT is not set
1568 +# CONFIG_CRYPTO_TEA is not set
1569 +# CONFIG_CRYPTO_TWOFISH is not set
1570 +# CONFIG_CRYPTO_TWOFISH_X86_64 is not set
1571 +
1572 +#
1573 +# Compression
1574 +#
1575 +CONFIG_CRYPTO_DEFLATE=y
1576 +# CONFIG_CRYPTO_ZLIB is not set
1577 +# CONFIG_CRYPTO_LZO is not set
1578 +
1579 +#
1580 +# Random Number Generation
1581 +#
1582 +CONFIG_CRYPTO_ANSI_CPRNG=y
1583 +CONFIG_CRYPTO_USER_API=y
1584 +CONFIG_CRYPTO_USER_API_HASH=y
1585 +CONFIG_CRYPTO_USER_API_SKCIPHER=y
1586 +CONFIG_CRYPTO_HW=y
1587 +# CONFIG_CRYPTO_DEV_PADLOCK is not set
1588 +# CONFIG_CRYPTO_DEV_HIFN_795X is not set
1589 +CONFIG_HAVE_KVM=y
1590 +# CONFIG_VIRTUALIZATION is not set
1591 +# CONFIG_BINARY_PRINTF is not set
1592 +
1593 +#
1594 +# Library routines
1595 +#
1596 +CONFIG_RAID6_PQ=y
1597 +CONFIG_BITREVERSE=y
1598 +CONFIG_GENERIC_FIND_FIRST_BIT=y
1599 +# CONFIG_CRC_CCITT is not set
1600 +CONFIG_CRC16=y
1601 +# CONFIG_CRC_T10DIF is not set
1602 +# CONFIG_CRC_ITU_T is not set
1603 +CONFIG_CRC32=y
1604 +# CONFIG_CRC7 is not set
1605 +CONFIG_LIBCRC32C=y
1606 +CONFIG_ZLIB_INFLATE=y
1607 +CONFIG_ZLIB_DEFLATE=y
1608 +CONFIG_LZO_COMPRESS=y
1609 +CONFIG_LZO_DECOMPRESS=y
1610 +CONFIG_XZ_DEC=y
1611 +CONFIG_XZ_DEC_X86=y
1612 +# CONFIG_XZ_DEC_POWERPC is not set
1613 +# CONFIG_XZ_DEC_IA64 is not set
1614 +# CONFIG_XZ_DEC_ARM is not set
1615 +# CONFIG_XZ_DEC_ARMTHUMB is not set
1616 +# CONFIG_XZ_DEC_SPARC is not set
1617 +CONFIG_XZ_DEC_BCJ=y
1618 +# CONFIG_XZ_DEC_TEST is not set
1619 +CONFIG_HAS_IOMEM=y
1620 +CONFIG_HAS_IOPORT=y
1621 +CONFIG_HAS_DMA=y
1622 +CONFIG_CPU_RMAP=y
1623 +CONFIG_NLATTR=y
1624 +# CONFIG_AVERAGE is not set
1625 diff -NurpP --minimal linux-3.0/Documentation/vserver/debug.txt linux-3.0-vs2.3.1-pre7/Documentation/vserver/debug.txt
1626 --- linux-3.0/Documentation/vserver/debug.txt   1970-01-01 01:00:00.000000000 +0100
1627 +++ linux-3.0-vs2.3.1-pre7/Documentation/vserver/debug.txt      2011-06-10 22:11:24.000000000 +0200
1628 @@ -0,0 +1,154 @@
1629 +
1630 +debug_cvirt:
1631 +
1632 + 2   4 "vx_map_tgid: %p/%llx: %d -> %d"
1633 +       "vx_rmap_tgid: %p/%llx: %d -> %d"
1634 +
1635 +debug_dlim:
1636 +
1637 + 0   1 "ALLOC (%p,#%d)%c inode (%d)"
1638 +       "FREE  (%p,#%d)%c inode"
1639 + 1   2 "ALLOC (%p,#%d)%c %lld bytes (%d)"
1640 +       "FREE  (%p,#%d)%c %lld bytes"
1641 + 2   4 "ADJUST: %lld,%lld on %ld,%ld [mult=%d]"
1642 + 3   8 "ext3_has_free_blocks(%p): %lu<%lu+1, %c, %u!=%u r=%d"
1643 +       "ext3_has_free_blocks(%p): free=%lu, root=%lu"
1644 +       "rcu_free_dl_info(%p)"
1645 + 4  10 "alloc_dl_info(%p,%d) = %p"
1646 +       "dealloc_dl_info(%p)"
1647 +       "get_dl_info(%p[#%d.%d])"
1648 +       "put_dl_info(%p[#%d.%d])"
1649 + 5  20 "alloc_dl_info(%p,%d)*"
1650 + 6  40 "__hash_dl_info: %p[#%d]"
1651 +       "__unhash_dl_info: %p[#%d]"
1652 + 7  80 "locate_dl_info(%p,#%d) = %p"
1653 +
1654 +debug_misc:
1655 +
1656 + 0   1 "destroy_dqhash: %p [#0x%08x] c=%d"
1657 +       "new_dqhash: %p [#0x%08x]"
1658 +       "vroot[%d]_clr_dev: dev=%p[%lu,%d:%d]"
1659 +       "vroot[%d]_get_real_bdev: dev=%p[%lu,%d:%d]"
1660 +       "vroot[%d]_set_dev: dev=%p[%lu,%d:%d]"
1661 +       "vroot_get_real_bdev not set"
1662 + 1   2 "cow_break_link(»%s«)"
1663 +       "temp copy »%s«"
1664 + 2   4 "dentry_open(new): %p"
1665 +       "dentry_open(old): %p"
1666 +       "lookup_create(new): %p"
1667 +       "old path »%s«"
1668 +       "path_lookup(old): %d"
1669 +       "vfs_create(new): %d"
1670 +       "vfs_rename: %d"
1671 +       "vfs_sendfile: %d"
1672 + 3   8 "fput(new_file=%p[#%d])"
1673 +       "fput(old_file=%p[#%d])"
1674 + 4  10 "vx_info_kill(%p[#%d],%d,%d) = %d"
1675 +       "vx_info_kill(%p[#%d],%d,%d)*"
1676 + 5  20 "vs_reboot(%p[#%d],%d)"
1677 + 6  40 "dropping task %p[#%u,%u] for %p[#%u,%u]"
1678 +
1679 +debug_net:
1680 +
1681 + 2   4 "nx_addr_conflict(%p,%p) %d.%d,%d.%d"
1682 + 3   8 "inet_bind(%p) %d.%d.%d.%d, %d.%d.%d.%d, %d.%d.%d.%d"
1683 +       "inet_bind(%p)* %p,%p;%lx %d.%d.%d.%d"
1684 + 4  10 "ip_route_connect(%p) %p,%p;%lx"
1685 + 5  20 "__addr_in_socket(%p,%d.%d.%d.%d) %p:%d.%d.%d.%d %p;%lx"
1686 + 6  40 "sk,egf: %p [#%d] (from %d)"
1687 +       "sk,egn: %p [#%d] (from %d)"
1688 +       "sk,req: %p [#%d] (from %d)"
1689 +       "sk: %p [#%d] (from %d)"
1690 +       "tw: %p [#%d] (from %d)"
1691 + 7  80 "__sock_recvmsg: %p[%p,%p,%p;%d]:%d/%d"
1692 +       "__sock_sendmsg: %p[%p,%p,%p;%d]:%d/%d"
1693 +
1694 +debug_nid:
1695 +
1696 + 0   1 "__lookup_nx_info(#%u): %p[#%u]"
1697 +       "alloc_nx_info(%d) = %p"
1698 +       "create_nx_info(%d) (dynamic rejected)"
1699 +       "create_nx_info(%d) = %p (already there)"
1700 +       "create_nx_info(%d) = %p (new)"
1701 +       "dealloc_nx_info(%p)"
1702 + 1   2 "alloc_nx_info(%d)*"
1703 +       "create_nx_info(%d)*"
1704 + 2   4 "get_nx_info(%p[#%d.%d])"
1705 +       "put_nx_info(%p[#%d.%d])"
1706 + 3   8 "claim_nx_info(%p[#%d.%d.%d]) %p"
1707 +       "clr_nx_info(%p[#%d.%d])"
1708 +       "init_nx_info(%p[#%d.%d])"
1709 +       "release_nx_info(%p[#%d.%d.%d]) %p"
1710 +       "set_nx_info(%p[#%d.%d])"
1711 + 4  10 "__hash_nx_info: %p[#%d]"
1712 +       "__nx_dynamic_id: [#%d]"
1713 +       "__unhash_nx_info: %p[#%d.%d.%d]"
1714 + 5  20 "moved task %p into nxi:%p[#%d]"
1715 +       "nx_migrate_task(%p,%p[#%d.%d.%d])"
1716 +       "task_get_nx_info(%p)"
1717 + 6  40 "nx_clear_persistent(%p[#%d])"
1718 +
1719 +debug_quota:
1720 +
1721 + 0   1 "quota_sync_dqh(%p,%d) discard inode %p"
1722 + 1   2 "quota_sync_dqh(%p,%d)"
1723 +       "sync_dquots(%p,%d)"
1724 +       "sync_dquots_dqh(%p,%d)"
1725 + 3   8 "do_quotactl(%p,%d,cmd=%d,id=%d,%p)"
1726 +
1727 +debug_switch:
1728 +
1729 + 0   1 "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]"
1730 + 1   2 "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]"
1731 + 4  10 "%s: (%s %s) returned %s with %d"
1732 +
1733 +debug_tag:
1734 +
1735 + 7  80 "dx_parse_tag(»%s«): %d:#%d"
1736 +       "dx_propagate_tag(%p[#%lu.%d]): %d,%d"
1737 +
1738 +debug_xid:
1739 +
1740 + 0   1 "__lookup_vx_info(#%u): %p[#%u]"
1741 +       "alloc_vx_info(%d) = %p"
1742 +       "alloc_vx_info(%d)*"
1743 +       "create_vx_info(%d) (dynamic rejected)"
1744 +       "create_vx_info(%d) = %p (already there)"
1745 +       "create_vx_info(%d) = %p (new)"
1746 +       "dealloc_vx_info(%p)"
1747 +       "loc_vx_info(%d) = %p (found)"
1748 +       "loc_vx_info(%d) = %p (new)"
1749 +       "loc_vx_info(%d) = %p (not available)"
1750 + 1   2 "create_vx_info(%d)*"
1751 +       "loc_vx_info(%d)*"
1752 + 2   4 "get_vx_info(%p[#%d.%d])"
1753 +       "put_vx_info(%p[#%d.%d])"
1754 + 3   8 "claim_vx_info(%p[#%d.%d.%d]) %p"
1755 +       "clr_vx_info(%p[#%d.%d])"
1756 +       "init_vx_info(%p[#%d.%d])"
1757 +       "release_vx_info(%p[#%d.%d.%d]) %p"
1758 +       "set_vx_info(%p[#%d.%d])"
1759 + 4  10 "__hash_vx_info: %p[#%d]"
1760 +       "__unhash_vx_info: %p[#%d.%d.%d]"
1761 +       "__vx_dynamic_id: [#%d]"
1762 + 5  20 "enter_vx_info(%p[#%d],%p) %p[#%d,%p]"
1763 +       "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]"
1764 +       "moved task %p into vxi:%p[#%d]"
1765 +       "task_get_vx_info(%p)"
1766 +       "vx_migrate_task(%p,%p[#%d.%d])"
1767 + 6  40 "vx_clear_persistent(%p[#%d])"
1768 +       "vx_exit_init(%p[#%d],%p[#%d,%d,%d])"
1769 +       "vx_set_init(%p[#%d],%p[#%d,%d,%d])"
1770 +       "vx_set_persistent(%p[#%d])"
1771 +       "vx_set_reaper(%p[#%d],%p[#%d,%d])"
1772 + 7  80 "vx_child_reaper(%p[#%u,%u]) = %p[#%u,%u]"
1773 +
1774 +
1775 +debug_limit:
1776 +
1777 + n 2^n "vx_acc_cres[%5d,%s,%2d]: %5d%s"
1778 +       "vx_cres_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
1779 +
1780 + m 2^m "vx_acc_page[%5d,%s,%2d]: %5d%s"
1781 +       "vx_acc_pages[%5d,%s,%2d]: %5d += %5d"
1782 +       "vx_pages_avail[%5d,%s,%2d]: %5ld > %5d + %5d"
1783 diff -NurpP --minimal linux-3.0/arch/alpha/Kconfig linux-3.0-vs2.3.1-pre7/arch/alpha/Kconfig
1784 --- linux-3.0/arch/alpha/Kconfig        2011-07-22 11:17:32.000000000 +0200
1785 +++ linux-3.0-vs2.3.1-pre7/arch/alpha/Kconfig   2011-06-10 22:11:24.000000000 +0200
1786 @@ -668,6 +668,8 @@ config DUMMY_CONSOLE
1787         depends on VGA_HOSE
1788         default y
1789
1790 +source "kernel/vserver/Kconfig"
1791 +
1792  source "security/Kconfig"
1793
1794  source "crypto/Kconfig"
1795 diff -NurpP --minimal linux-3.0/arch/alpha/kernel/entry.S linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/entry.S
1796 --- linux-3.0/arch/alpha/kernel/entry.S 2010-10-21 13:06:45.000000000 +0200
1797 +++ linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/entry.S    2011-06-10 22:11:24.000000000 +0200
1798 @@ -860,24 +860,15 @@ sys_getxgid:
1799         .globl  sys_getxpid
1800         .ent    sys_getxpid
1801  sys_getxpid:
1802 +       lda     $sp, -16($sp)
1803 +       stq     $26, 0($sp)
1804         .prologue 0
1805 -       ldq     $2, TI_TASK($8)
1806
1807 -       /* See linux/kernel/timer.c sys_getppid for discussion
1808 -          about this loop.  */
1809 -       ldq     $3, TASK_GROUP_LEADER($2)
1810 -       ldq     $4, TASK_REAL_PARENT($3)
1811 -       ldl     $0, TASK_TGID($2)
1812 -1:     ldl     $1, TASK_TGID($4)
1813 -#ifdef CONFIG_SMP
1814 -       mov     $4, $5
1815 -       mb
1816 -       ldq     $3, TASK_GROUP_LEADER($2)
1817 -       ldq     $4, TASK_REAL_PARENT($3)
1818 -       cmpeq   $4, $5, $5
1819 -       beq     $5, 1b
1820 -#endif
1821 -       stq     $1, 80($sp)
1822 +       lda     $16, 96($sp)
1823 +       jsr     $26, do_getxpid
1824 +       ldq     $26, 0($sp)
1825 +
1826 +       lda     $sp, 16($sp)
1827         ret
1828  .end sys_getxpid
1829
1830 diff -NurpP --minimal linux-3.0/arch/alpha/kernel/ptrace.c linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/ptrace.c
1831 --- linux-3.0/arch/alpha/kernel/ptrace.c        2011-01-05 21:48:40.000000000 +0100
1832 +++ linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/ptrace.c   2011-06-10 22:11:24.000000000 +0200
1833 @@ -13,6 +13,7 @@
1834  #include <linux/user.h>
1835  #include <linux/security.h>
1836  #include <linux/signal.h>
1837 +#include <linux/vs_base.h>
1838
1839  #include <asm/uaccess.h>
1840  #include <asm/pgtable.h>
1841 diff -NurpP --minimal linux-3.0/arch/alpha/kernel/systbls.S linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/systbls.S
1842 --- linux-3.0/arch/alpha/kernel/systbls.S       2011-07-22 11:17:32.000000000 +0200
1843 +++ linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/systbls.S  2011-06-10 22:11:24.000000000 +0200
1844 @@ -446,7 +446,7 @@ sys_call_table:
1845         .quad sys_stat64                        /* 425 */
1846         .quad sys_lstat64
1847         .quad sys_fstat64
1848 -       .quad sys_ni_syscall                    /* sys_vserver */
1849 +       .quad sys_vserver                       /* sys_vserver */
1850         .quad sys_ni_syscall                    /* sys_mbind */
1851         .quad sys_ni_syscall                    /* sys_get_mempolicy */
1852         .quad sys_ni_syscall                    /* sys_set_mempolicy */
1853 diff -NurpP --minimal linux-3.0/arch/alpha/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/traps.c
1854 --- linux-3.0/arch/alpha/kernel/traps.c 2010-10-21 13:06:46.000000000 +0200
1855 +++ linux-3.0-vs2.3.1-pre7/arch/alpha/kernel/traps.c    2011-06-10 22:11:24.000000000 +0200
1856 @@ -183,7 +183,8 @@ die_if_kernel(char * str, struct pt_regs
1857  #ifdef CONFIG_SMP
1858         printk("CPU %d ", hard_smp_processor_id());
1859  #endif
1860 -       printk("%s(%d): %s %ld\n", current->comm, task_pid_nr(current), str, err);
1861 +       printk("%s(%d[#%u]): %s %ld\n", current->comm,
1862 +               task_pid_nr(current), current->xid, str, err);
1863         dik_show_regs(regs, r9_15);
1864         add_taint(TAINT_DIE);
1865         dik_show_trace((unsigned long *)(regs+1));
1866 diff -NurpP --minimal linux-3.0/arch/arm/Kconfig linux-3.0-vs2.3.1-pre7/arch/arm/Kconfig
1867 --- linux-3.0/arch/arm/Kconfig  2011-07-22 11:17:32.000000000 +0200
1868 +++ linux-3.0-vs2.3.1-pre7/arch/arm/Kconfig     2011-06-10 22:11:24.000000000 +0200
1869 @@ -2035,6 +2035,8 @@ source "fs/Kconfig"
1870
1871  source "arch/arm/Kconfig.debug"
1872
1873 +source "kernel/vserver/Kconfig"
1874 +
1875  source "security/Kconfig"
1876
1877  source "crypto/Kconfig"
1878 diff -NurpP --minimal linux-3.0/arch/arm/kernel/calls.S linux-3.0-vs2.3.1-pre7/arch/arm/kernel/calls.S
1879 --- linux-3.0/arch/arm/kernel/calls.S   2011-07-22 11:17:32.000000000 +0200
1880 +++ linux-3.0-vs2.3.1-pre7/arch/arm/kernel/calls.S      2011-06-10 22:11:24.000000000 +0200
1881 @@ -322,7 +322,7 @@
1882  /* 310 */      CALL(sys_request_key)
1883                 CALL(sys_keyctl)
1884                 CALL(ABI(sys_semtimedop, sys_oabi_semtimedop))
1885 -/* vserver */  CALL(sys_ni_syscall)
1886 +               CALL(sys_vserver)
1887                 CALL(sys_ioprio_set)
1888  /* 315 */      CALL(sys_ioprio_get)
1889                 CALL(sys_inotify_init)
1890 diff -NurpP --minimal linux-3.0/arch/arm/kernel/process.c linux-3.0-vs2.3.1-pre7/arch/arm/kernel/process.c
1891 --- linux-3.0/arch/arm/kernel/process.c 2011-05-22 16:16:47.000000000 +0200
1892 +++ linux-3.0-vs2.3.1-pre7/arch/arm/kernel/process.c    2011-06-10 22:11:24.000000000 +0200
1893 @@ -315,7 +315,8 @@ void __show_regs(struct pt_regs *regs)
1894  void show_regs(struct pt_regs * regs)
1895  {
1896         printk("\n");
1897 -       printk("Pid: %d, comm: %20s\n", task_pid_nr(current), current->comm);
1898 +       printk("Pid: %d[#%u], comm: %20s\n",
1899 +               task_pid_nr(current), current->xid, current->comm);
1900         __show_regs(regs);
1901         __backtrace();
1902  }
1903 diff -NurpP --minimal linux-3.0/arch/arm/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/arm/kernel/traps.c
1904 --- linux-3.0/arch/arm/kernel/traps.c   2011-07-22 11:17:32.000000000 +0200
1905 +++ linux-3.0-vs2.3.1-pre7/arch/arm/kernel/traps.c      2011-06-22 12:39:12.000000000 +0200
1906 @@ -242,8 +242,8 @@ static int __die(const char *str, int er
1907
1908         print_modules();
1909         __show_regs(regs);
1910 -       printk(KERN_EMERG "Process %.*s (pid: %d, stack limit = 0x%p)\n",
1911 -               TASK_COMM_LEN, tsk->comm, task_pid_nr(tsk), thread + 1);
1912 +       printk(KERN_EMERG "Process %.*s (pid: %d:#%u, stack limit = 0x%p)\n",
1913 +               TASK_COMM_LEN, tsk->comm, task_pid_nr(tsk), tsk->xid, thread + 1);
1914
1915         if (!user_mode(regs) || in_interrupt()) {
1916                 dump_mem(KERN_EMERG, "Stack: ", regs->ARM_sp,
1917 diff -NurpP --minimal linux-3.0/arch/cris/Kconfig linux-3.0-vs2.3.1-pre7/arch/cris/Kconfig
1918 --- linux-3.0/arch/cris/Kconfig 2011-07-22 11:17:35.000000000 +0200
1919 +++ linux-3.0-vs2.3.1-pre7/arch/cris/Kconfig    2011-06-10 22:11:24.000000000 +0200
1920 @@ -678,6 +678,8 @@ source "drivers/staging/Kconfig"
1921
1922  source "arch/cris/Kconfig.debug"
1923
1924 +source "kernel/vserver/Kconfig"
1925 +
1926  source "security/Kconfig"
1927
1928  source "crypto/Kconfig"
1929 diff -NurpP --minimal linux-3.0/arch/frv/kernel/kernel_thread.S linux-3.0-vs2.3.1-pre7/arch/frv/kernel/kernel_thread.S
1930 --- linux-3.0/arch/frv/kernel/kernel_thread.S   2008-12-25 00:26:37.000000000 +0100
1931 +++ linux-3.0-vs2.3.1-pre7/arch/frv/kernel/kernel_thread.S      2011-06-10 22:11:24.000000000 +0200
1932 @@ -37,7 +37,7 @@ kernel_thread:
1933
1934         # start by forking the current process, but with shared VM
1935         setlos.p        #__NR_clone,gr7         ; syscall number
1936 -       ori             gr10,#CLONE_VM,gr8      ; first syscall arg     [clone_flags]
1937 +       ori             gr10,#CLONE_KT,gr8      ; first syscall arg     [clone_flags]
1938         sethi.p         #0xe4e4,gr9             ; second syscall arg    [newsp]
1939         setlo           #0xe4e4,gr9
1940         setlos.p        #0,gr10                 ; third syscall arg     [parent_tidptr]
1941 diff -NurpP --minimal linux-3.0/arch/h8300/Kconfig linux-3.0-vs2.3.1-pre7/arch/h8300/Kconfig
1942 --- linux-3.0/arch/h8300/Kconfig        2011-07-22 11:17:35.000000000 +0200
1943 +++ linux-3.0-vs2.3.1-pre7/arch/h8300/Kconfig   2011-06-10 22:11:24.000000000 +0200
1944 @@ -213,6 +213,8 @@ source "fs/Kconfig"
1945
1946  source "arch/h8300/Kconfig.debug"
1947
1948 +source "kernel/vserver/Kconfig"
1949 +
1950  source "security/Kconfig"
1951
1952  source "crypto/Kconfig"
1953 diff -NurpP --minimal linux-3.0/arch/ia64/Kconfig linux-3.0-vs2.3.1-pre7/arch/ia64/Kconfig
1954 --- linux-3.0/arch/ia64/Kconfig 2011-07-22 11:17:35.000000000 +0200
1955 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/Kconfig    2011-06-10 22:11:24.000000000 +0200
1956 @@ -671,6 +671,8 @@ source "fs/Kconfig"
1957
1958  source "arch/ia64/Kconfig.debug"
1959
1960 +source "kernel/vserver/Kconfig"
1961 +
1962  source "security/Kconfig"
1963
1964  source "crypto/Kconfig"
1965 diff -NurpP --minimal linux-3.0/arch/ia64/include/asm/tlb.h linux-3.0-vs2.3.1-pre7/arch/ia64/include/asm/tlb.h
1966 --- linux-3.0/arch/ia64/include/asm/tlb.h       2011-07-22 11:17:35.000000000 +0200
1967 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/include/asm/tlb.h  2011-06-10 22:11:24.000000000 +0200
1968 @@ -40,6 +40,7 @@
1969  #include <linux/mm.h>
1970  #include <linux/pagemap.h>
1971  #include <linux/swap.h>
1972 +#include <linux/vs_memory.h>
1973
1974  #include <asm/pgalloc.h>
1975  #include <asm/processor.h>
1976 diff -NurpP --minimal linux-3.0/arch/ia64/kernel/entry.S linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/entry.S
1977 --- linux-3.0/arch/ia64/kernel/entry.S  2011-07-22 11:17:35.000000000 +0200
1978 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/entry.S     2011-06-10 22:11:24.000000000 +0200
1979 @@ -1714,7 +1714,7 @@ sys_call_table:
1980         data8 sys_mq_notify
1981         data8 sys_mq_getsetattr
1982         data8 sys_kexec_load
1983 -       data8 sys_ni_syscall                    // reserved for vserver
1984 +       data8 sys_vserver
1985         data8 sys_waitid                        // 1270
1986         data8 sys_add_key
1987         data8 sys_request_key
1988 diff -NurpP --minimal linux-3.0/arch/ia64/kernel/perfmon.c linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/perfmon.c
1989 --- linux-3.0/arch/ia64/kernel/perfmon.c        2011-03-15 18:06:39.000000000 +0100
1990 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/perfmon.c   2011-06-10 22:11:24.000000000 +0200
1991 @@ -42,6 +42,7 @@
1992  #include <linux/completion.h>
1993  #include <linux/tracehook.h>
1994  #include <linux/slab.h>
1995 +#include <linux/vs_memory.h>
1996
1997  #include <asm/errno.h>
1998  #include <asm/intrinsics.h>
1999 diff -NurpP --minimal linux-3.0/arch/ia64/kernel/process.c linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/process.c
2000 --- linux-3.0/arch/ia64/kernel/process.c        2011-03-15 18:06:39.000000000 +0100
2001 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/process.c   2011-06-10 22:11:24.000000000 +0200
2002 @@ -109,8 +109,8 @@ show_regs (struct pt_regs *regs)
2003         unsigned long ip = regs->cr_iip + ia64_psr(regs)->ri;
2004
2005         print_modules();
2006 -       printk("\nPid: %d, CPU %d, comm: %20s\n", task_pid_nr(current),
2007 -                       smp_processor_id(), current->comm);
2008 +       printk("\nPid: %d[#%u], CPU %d, comm: %20s\n", task_pid_nr(current),
2009 +                       current->xid, smp_processor_id(), current->comm);
2010         printk("psr : %016lx ifs : %016lx ip  : [<%016lx>]    %s (%s)\n",
2011                regs->cr_ipsr, regs->cr_ifs, ip, print_tainted(),
2012                init_utsname()->release);
2013 diff -NurpP --minimal linux-3.0/arch/ia64/kernel/ptrace.c linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/ptrace.c
2014 --- linux-3.0/arch/ia64/kernel/ptrace.c 2011-01-05 21:48:59.000000000 +0100
2015 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/ptrace.c    2011-06-10 22:11:24.000000000 +0200
2016 @@ -21,6 +21,7 @@
2017  #include <linux/regset.h>
2018  #include <linux/elf.h>
2019  #include <linux/tracehook.h>
2020 +#include <linux/vs_base.h>
2021
2022  #include <asm/pgtable.h>
2023  #include <asm/processor.h>
2024 diff -NurpP --minimal linux-3.0/arch/ia64/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/traps.c
2025 --- linux-3.0/arch/ia64/kernel/traps.c  2010-07-07 18:31:01.000000000 +0200
2026 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/kernel/traps.c     2011-06-10 22:11:24.000000000 +0200
2027 @@ -59,8 +59,9 @@ die (const char *str, struct pt_regs *re
2028         put_cpu();
2029
2030         if (++die.lock_owner_depth < 3) {
2031 -               printk("%s[%d]: %s %ld [%d]\n",
2032 -               current->comm, task_pid_nr(current), str, err, ++die_counter);
2033 +               printk("%s[%d[#%u]]: %s %ld [%d]\n",
2034 +                       current->comm, task_pid_nr(current), current->xid,
2035 +                       str, err, ++die_counter);
2036                 if (notify_die(DIE_OOPS, str, regs, err, 255, SIGSEGV)
2037                     != NOTIFY_STOP)
2038                         show_regs(regs);
2039 @@ -323,8 +324,9 @@ handle_fpu_swa (int fp_fault, struct pt_
2040                         if ((last.count & 15) < 5 && (ia64_fetchadd(1, &last.count, acq) & 15) < 5) {
2041                                 last.time = current_jiffies + 5 * HZ;
2042                                 printk(KERN_WARNING
2043 -                                       "%s(%d): floating-point assist fault at ip %016lx, isr %016lx\n",
2044 -                                       current->comm, task_pid_nr(current), regs->cr_iip + ia64_psr(regs)->ri, isr);
2045 +                                       "%s(%d[#%u]): floating-point assist fault at ip %016lx, isr %016lx\n",
2046 +                                       current->comm, task_pid_nr(current), current->xid,
2047 +                                       regs->cr_iip + ia64_psr(regs)->ri, isr);
2048                         }
2049                 }
2050         }
2051 diff -NurpP --minimal linux-3.0/arch/ia64/mm/fault.c linux-3.0-vs2.3.1-pre7/arch/ia64/mm/fault.c
2052 --- linux-3.0/arch/ia64/mm/fault.c      2011-07-22 11:17:35.000000000 +0200
2053 +++ linux-3.0-vs2.3.1-pre7/arch/ia64/mm/fault.c 2011-06-10 22:28:23.000000000 +0200
2054 @@ -11,6 +11,7 @@
2055  #include <linux/kprobes.h>
2056  #include <linux/kdebug.h>
2057  #include <linux/prefetch.h>
2058 +#include <linux/vs_memory.h>
2059
2060  #include <asm/pgtable.h>
2061  #include <asm/processor.h>
2062 diff -NurpP --minimal linux-3.0/arch/m32r/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/m32r/kernel/traps.c
2063 --- linux-3.0/arch/m32r/kernel/traps.c  2009-12-03 20:01:57.000000000 +0100
2064 +++ linux-3.0-vs2.3.1-pre7/arch/m32r/kernel/traps.c     2011-06-10 22:11:24.000000000 +0200
2065 @@ -196,8 +196,9 @@ static void show_registers(struct pt_reg
2066         } else {
2067                 printk("SPI: %08lx\n", sp);
2068         }
2069 -       printk("Process %s (pid: %d, process nr: %d, stackpage=%08lx)",
2070 -               current->comm, task_pid_nr(current), 0xffff & i, 4096+(unsigned long)current);
2071 +       printk("Process %s (pid: %d[#%u], process nr: %d, stackpage=%08lx)",
2072 +               current->comm, task_pid_nr(current), current->xid,
2073 +               0xffff & i, 4096+(unsigned long)current);
2074
2075         /*
2076          * When in-kernel, we also print out the stack and code at the
2077 diff -NurpP --minimal linux-3.0/arch/m68k/Kconfig linux-3.0-vs2.3.1-pre7/arch/m68k/Kconfig
2078 --- linux-3.0/arch/m68k/Kconfig 2011-07-22 11:17:35.000000000 +0200
2079 +++ linux-3.0-vs2.3.1-pre7/arch/m68k/Kconfig    2011-06-10 22:11:24.000000000 +0200
2080 @@ -241,6 +241,8 @@ source "fs/Kconfig"
2081
2082  source "arch/m68k/Kconfig.debug"
2083
2084 +source "kernel/vserver/Kconfig"
2085 +
2086  source "security/Kconfig"
2087
2088  source "crypto/Kconfig"
2089 diff -NurpP --minimal linux-3.0/arch/mips/Kconfig linux-3.0-vs2.3.1-pre7/arch/mips/Kconfig
2090 --- linux-3.0/arch/mips/Kconfig 2011-07-22 11:17:35.000000000 +0200
2091 +++ linux-3.0-vs2.3.1-pre7/arch/mips/Kconfig    2011-06-10 22:11:24.000000000 +0200
2092 @@ -2485,6 +2485,8 @@ source "fs/Kconfig"
2093
2094  source "arch/mips/Kconfig.debug"
2095
2096 +source "kernel/vserver/Kconfig"
2097 +
2098  source "security/Kconfig"
2099
2100  source "crypto/Kconfig"
2101 diff -NurpP --minimal linux-3.0/arch/mips/kernel/ptrace.c linux-3.0-vs2.3.1-pre7/arch/mips/kernel/ptrace.c
2102 --- linux-3.0/arch/mips/kernel/ptrace.c 2011-07-22 11:17:36.000000000 +0200
2103 +++ linux-3.0-vs2.3.1-pre7/arch/mips/kernel/ptrace.c    2011-06-10 22:11:24.000000000 +0200
2104 @@ -25,6 +25,7 @@
2105  #include <linux/security.h>
2106  #include <linux/audit.h>
2107  #include <linux/seccomp.h>
2108 +#include <linux/vs_base.h>
2109
2110  #include <asm/byteorder.h>
2111  #include <asm/cpu.h>
2112 @@ -263,6 +264,9 @@ long arch_ptrace(struct task_struct *chi
2113         void __user *datavp = (void __user *) data;
2114         unsigned long __user *datalp = (void __user *) data;
2115
2116 +       if (!vx_check(vx_task_xid(child), VS_WATCH_P | VS_IDENT))
2117 +               goto out;
2118 +
2119         switch (request) {
2120         /* when I and D space are separate, these will need to be fixed. */
2121         case PTRACE_PEEKTEXT: /* read word at location addr. */
2122 diff -NurpP --minimal linux-3.0/arch/mips/kernel/scall32-o32.S linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall32-o32.S
2123 --- linux-3.0/arch/mips/kernel/scall32-o32.S    2011-07-22 11:17:36.000000000 +0200
2124 +++ linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall32-o32.S       2011-06-10 22:11:24.000000000 +0200
2125 @@ -523,7 +523,7 @@ einval:     li      v0, -ENOSYS
2126         sys     sys_mq_timedreceive     5
2127         sys     sys_mq_notify           2       /* 4275 */
2128         sys     sys_mq_getsetattr       3
2129 -       sys     sys_ni_syscall          0       /* sys_vserver */
2130 +       sys     sys_vserver             3
2131         sys     sys_waitid              5
2132         sys     sys_ni_syscall          0       /* available, was setaltroot */
2133         sys     sys_add_key             5       /* 4280 */
2134 diff -NurpP --minimal linux-3.0/arch/mips/kernel/scall64-64.S linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall64-64.S
2135 --- linux-3.0/arch/mips/kernel/scall64-64.S     2011-07-22 11:17:36.000000000 +0200
2136 +++ linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall64-64.S        2011-06-10 22:11:24.000000000 +0200
2137 @@ -362,7 +362,7 @@ sys_call_table:
2138         PTR     sys_mq_timedreceive
2139         PTR     sys_mq_notify
2140         PTR     sys_mq_getsetattr               /* 5235 */
2141 -       PTR     sys_ni_syscall                  /* sys_vserver */
2142 +       PTR     sys_vserver
2143         PTR     sys_waitid
2144         PTR     sys_ni_syscall                  /* available, was setaltroot */
2145         PTR     sys_add_key
2146 diff -NurpP --minimal linux-3.0/arch/mips/kernel/scall64-n32.S linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall64-n32.S
2147 --- linux-3.0/arch/mips/kernel/scall64-n32.S    2011-07-22 11:17:36.000000000 +0200
2148 +++ linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall64-n32.S       2011-06-10 22:11:24.000000000 +0200
2149 @@ -361,7 +361,7 @@ EXPORT(sysn32_call_table)
2150         PTR     compat_sys_mq_timedreceive
2151         PTR     compat_sys_mq_notify
2152         PTR     compat_sys_mq_getsetattr
2153 -       PTR     sys_ni_syscall                  /* 6240, sys_vserver */
2154 +       PTR     sys32_vserver                   /* 6240 */
2155         PTR     compat_sys_waitid
2156         PTR     sys_ni_syscall                  /* available, was setaltroot */
2157         PTR     sys_add_key
2158 diff -NurpP --minimal linux-3.0/arch/mips/kernel/scall64-o32.S linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall64-o32.S
2159 --- linux-3.0/arch/mips/kernel/scall64-o32.S    2011-07-22 11:17:36.000000000 +0200
2160 +++ linux-3.0-vs2.3.1-pre7/arch/mips/kernel/scall64-o32.S       2011-06-10 22:11:24.000000000 +0200
2161 @@ -480,7 +480,7 @@ sys_call_table:
2162         PTR     compat_sys_mq_timedreceive
2163         PTR     compat_sys_mq_notify            /* 4275 */
2164         PTR     compat_sys_mq_getsetattr
2165 -       PTR     sys_ni_syscall                  /* sys_vserver */
2166 +       PTR     sys32_vserver
2167         PTR     sys_32_waitid
2168         PTR     sys_ni_syscall                  /* available, was setaltroot */
2169         PTR     sys_add_key                     /* 4280 */
2170 diff -NurpP --minimal linux-3.0/arch/mips/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/mips/kernel/traps.c
2171 --- linux-3.0/arch/mips/kernel/traps.c  2011-05-22 16:17:00.000000000 +0200
2172 +++ linux-3.0-vs2.3.1-pre7/arch/mips/kernel/traps.c     2011-06-10 22:11:24.000000000 +0200
2173 @@ -343,9 +343,10 @@ void show_registers(struct pt_regs *regs
2174
2175         __show_regs(regs);
2176         print_modules();
2177 -       printk("Process %s (pid: %d, threadinfo=%p, task=%p, tls=%0*lx)\n",
2178 -              current->comm, current->pid, current_thread_info(), current,
2179 -             field, current_thread_info()->tp_value);
2180 +       printk("Process %s (pid: %d:#%u, threadinfo=%p, task=%p, tls=%0*lx)\n",
2181 +               current->comm, task_pid_nr(current), current->xid,
2182 +               current_thread_info(), current,
2183 +               field, current_thread_info()->tp_value);
2184         if (cpu_has_userlocal) {
2185                 unsigned long tls;
2186
2187 diff -NurpP --minimal linux-3.0/arch/parisc/Kconfig linux-3.0-vs2.3.1-pre7/arch/parisc/Kconfig
2188 --- linux-3.0/arch/parisc/Kconfig       2011-07-22 11:17:36.000000000 +0200
2189 +++ linux-3.0-vs2.3.1-pre7/arch/parisc/Kconfig  2011-06-10 22:11:24.000000000 +0200
2190 @@ -279,6 +279,8 @@ source "fs/Kconfig"
2191
2192  source "arch/parisc/Kconfig.debug"
2193
2194 +source "kernel/vserver/Kconfig"
2195 +
2196  source "security/Kconfig"
2197
2198  source "crypto/Kconfig"
2199 diff -NurpP --minimal linux-3.0/arch/parisc/kernel/syscall_table.S linux-3.0-vs2.3.1-pre7/arch/parisc/kernel/syscall_table.S
2200 --- linux-3.0/arch/parisc/kernel/syscall_table.S        2011-07-22 11:17:36.000000000 +0200
2201 +++ linux-3.0-vs2.3.1-pre7/arch/parisc/kernel/syscall_table.S   2011-06-10 22:11:24.000000000 +0200
2202 @@ -361,7 +361,7 @@
2203         ENTRY_COMP(mbind)               /* 260 */
2204         ENTRY_COMP(get_mempolicy)
2205         ENTRY_COMP(set_mempolicy)
2206 -       ENTRY_SAME(ni_syscall)  /* 263: reserved for vserver */
2207 +       ENTRY_DIFF(vserver)
2208         ENTRY_SAME(add_key)
2209         ENTRY_SAME(request_key)         /* 265 */
2210         ENTRY_SAME(keyctl)
2211 diff -NurpP --minimal linux-3.0/arch/parisc/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/parisc/kernel/traps.c
2212 --- linux-3.0/arch/parisc/kernel/traps.c        2009-09-10 15:25:40.000000000 +0200
2213 +++ linux-3.0-vs2.3.1-pre7/arch/parisc/kernel/traps.c   2011-06-10 22:11:24.000000000 +0200
2214 @@ -236,8 +236,9 @@ void die_if_kernel(char *str, struct pt_
2215                 if (err == 0)
2216                         return; /* STFU */
2217
2218 -               printk(KERN_CRIT "%s (pid %d): %s (code %ld) at " RFMT "\n",
2219 -                       current->comm, task_pid_nr(current), str, err, regs->iaoq[0]);
2220 +               printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld) at " RFMT "\n",
2221 +                       current->comm, task_pid_nr(current), current->xid,
2222 +                       str, err, regs->iaoq[0]);
2223  #ifdef PRINT_USER_FAULTS
2224                 /* XXX for debugging only */
2225                 show_regs(regs);
2226 @@ -270,8 +271,8 @@ void die_if_kernel(char *str, struct pt_
2227                 pdc_console_restart();
2228
2229         if (err)
2230 -               printk(KERN_CRIT "%s (pid %d): %s (code %ld)\n",
2231 -                       current->comm, task_pid_nr(current), str, err);
2232 +               printk(KERN_CRIT "%s (pid %d:#%u): %s (code %ld)\n",
2233 +                       current->comm, task_pid_nr(current), current->xid, str, err);
2234
2235         /* Wot's wrong wif bein' racy? */
2236         if (current->thread.flags & PARISC_KERNEL_DEATH) {
2237 diff -NurpP --minimal linux-3.0/arch/parisc/mm/fault.c linux-3.0-vs2.3.1-pre7/arch/parisc/mm/fault.c
2238 --- linux-3.0/arch/parisc/mm/fault.c    2010-08-02 16:52:06.000000000 +0200
2239 +++ linux-3.0-vs2.3.1-pre7/arch/parisc/mm/fault.c       2011-06-10 22:11:24.000000000 +0200
2240 @@ -237,8 +237,9 @@ bad_area:
2241
2242  #ifdef PRINT_USER_FAULTS
2243                 printk(KERN_DEBUG "\n");
2244 -               printk(KERN_DEBUG "do_page_fault() pid=%d command='%s' type=%lu address=0x%08lx\n",
2245 -                   task_pid_nr(tsk), tsk->comm, code, address);
2246 +               printk(KERN_DEBUG "do_page_fault() pid=%d:#%u "
2247 +                   "command='%s' type=%lu address=0x%08lx\n",
2248 +                   task_pid_nr(tsk), tsk->xid, tsk->comm, code, address);
2249                 if (vma) {
2250                         printk(KERN_DEBUG "vm_start = 0x%08lx, vm_end = 0x%08lx\n",
2251                                         vma->vm_start, vma->vm_end);
2252 diff -NurpP --minimal linux-3.0/arch/powerpc/Kconfig linux-3.0-vs2.3.1-pre7/arch/powerpc/Kconfig
2253 --- linux-3.0/arch/powerpc/Kconfig      2011-07-22 11:17:36.000000000 +0200
2254 +++ linux-3.0-vs2.3.1-pre7/arch/powerpc/Kconfig 2011-06-10 22:11:24.000000000 +0200
2255 @@ -978,6 +978,8 @@ source "lib/Kconfig"
2256
2257  source "arch/powerpc/Kconfig.debug"
2258
2259 +source "kernel/vserver/Kconfig"
2260 +
2261  source "security/Kconfig"
2262
2263  config KEYS_COMPAT
2264 diff -NurpP --minimal linux-3.0/arch/powerpc/include/asm/unistd.h linux-3.0-vs2.3.1-pre7/arch/powerpc/include/asm/unistd.h
2265 --- linux-3.0/arch/powerpc/include/asm/unistd.h 2011-07-22 11:17:40.000000000 +0200
2266 +++ linux-3.0-vs2.3.1-pre7/arch/powerpc/include/asm/unistd.h    2011-06-10 22:11:24.000000000 +0200
2267 @@ -275,7 +275,7 @@
2268  #endif
2269  #define __NR_rtas              255
2270  #define __NR_sys_debug_setcontext 256
2271 -/* Number 257 is reserved for vserver */
2272 +#define __NR_vserver           257
2273  #define __NR_migrate_pages     258
2274  #define __NR_mbind             259
2275  #define __NR_get_mempolicy     260
2276 diff -NurpP --minimal linux-3.0/arch/powerpc/kernel/process.c linux-3.0-vs2.3.1-pre7/arch/powerpc/kernel/process.c
2277 --- linux-3.0/arch/powerpc/kernel/process.c     2011-07-22 11:17:40.000000000 +0200
2278 +++ linux-3.0-vs2.3.1-pre7/arch/powerpc/kernel/process.c        2011-06-10 22:11:24.000000000 +0200
2279 @@ -656,8 +656,9 @@ void show_regs(struct pt_regs * regs)
2280  #else
2281                 printk("DAR: "REG", DSISR: %08lx\n", regs->dar, regs->dsisr);
2282  #endif
2283 -       printk("TASK = %p[%d] '%s' THREAD: %p",
2284 -              current, task_pid_nr(current), current->comm, task_thread_info(current));
2285 +       printk("TASK = %p[%d,#%u] '%s' THREAD: %p",
2286 +              current, task_pid_nr(current), current->xid,
2287 +              current->comm, task_thread_info(current));
2288
2289  #ifdef CONFIG_SMP
2290         printk(" CPU: %d", raw_smp_processor_id());
2291 diff -NurpP --minimal linux-3.0/arch/powerpc/kernel/traps.c linux-3.0-vs2.3.1-pre7/arch/powerpc/kernel/traps.c
2292 --- linux-3.0/arch/powerpc/kernel/traps.c       2011-07-22 11:17:40.000000000 +0200
2293 +++ linux-3.0-vs2.3.1-pre7/arch/powerpc/kernel/traps.c  2011-07-19 00:44:39.000000000 +0200
2294 @@ -1075,8 +1075,9 @@ void nonrecoverable_exception(struct pt_
2295
2296  void trace_syscall(struct pt_regs *regs)
2297  {
2298 -       printk("Task: %p(%d), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
2299 -              current, task_pid_nr(current), regs->nip, regs->link, regs->gpr[0],
2300 +       printk("Task: %p(%d[#%u]), PC: %08lX/%08lX, Syscall: %3ld, Result: %s%ld    %s\n",
2301 +              current, task_pid_nr(current), current->xid,
2302 +              regs->nip, regs->link, regs->gpr[0],
2303                regs->ccr&0x10000000?"Error=":"", regs->gpr[3], print_tainted());
2304  }
2305
2306 diff -NurpP --minimal linux-3.0/arch/powerpc/kernel/vdso.c linux-3.0-vs2.3.1-pre7/arch/powerpc/kernel/vdso.c
2307 --- linux-3.0/arch/powerpc/kernel/vdso.c        2011-05-22 16:17:02.000000000 +0200
2308 +++ linux-3.0-vs2.3.1-pre7/arch/powerpc/kernel/vdso.c   2011-06-10 22:11:24.000000000 +0200
2309 @@ -23,6 +23,7 @@
2310  #include <linux/security.h>
2311  #include <linux/bootmem.h>
2312  #include <linux/memblock.h>
2313 +#include <linux/vs_memory.h>
2314
2315  #include <asm/pgtable.h>
2316  #include <asm/system.h>
2317 diff -NurpP --minimal linux-3.0/arch/s390/Kconfig linux-3.0-vs2.3.1-pre7/arch/s390/Kconfig
2318 --- linux-3.0/arch/s390/Kconfig 2011-07-22 11:17:41.000000000 +0200
2319 +++ linux-3.0-vs2.3.1-pre7/arch/s390/Kconfig    2011-07-01 11:35:34.000000000 +0200
2320 @@ -628,6 +628,8 @@ source "fs/Kconfig"
2321
2322  source "arch/s390/Kconfig.debug"
2323
2324 +source "kernel/vserver/Kconfig"
2325 +
2326  source "security/Kconfig"
2327
2328  source "crypto/Kconfig"
2329 diff -NurpP --minimal linux-3.0/arch/s390/include/asm/tlb.h linux-3.0-vs2.3.1-pre7/arch/s390/include/asm/tlb.h
2330 --- linux-3.0/arch/s390/include/asm/tlb.h       2011-07-22 11:17:41.000000000 +0200
2331 +++ linux-3.0-vs2.3.1-pre7/arch/s390/include/asm/tlb.h  2011-06-15 02:40:14.000000000 +0200
2332 @@ -24,6 +24,8 @@
2333  #include <linux/mm.h>
2334  #include <linux/pagemap.h>
2335  #include <linux/swap.h>
2336 +#include <linux/vs_memory.h>
2337 +
2338  #include <asm/processor.h>
2339  #include <asm/pgalloc.h>
2340  #include <asm/tlbflush.h>
2341 diff -NurpP --minimal linux-3.0/arch/s390/include/asm/unistd.h linux-3.0-vs2.3.1-pre7/arch/s390/include/asm/unistd.h
2342 --- linux-3.0/arch/s390/include/asm/unistd.h    2011-07-22 11:17:41.000000000 +0200
2343 +++ linux-3.0-vs2.3.1-pre7/arch/s390/include/asm/unistd.h       2011-06-10 22:11:24.000000000 +0200
2344 @@ -202,7 +202,7 @@
2345  #define __NR_clock_gettime     (__NR_timer_create+6)
2346  #define __NR_clock_getres      (__NR_timer_create+7)
2347  #define __NR_clock_nanosleep   (__NR_timer_create+8)
2348 -/* Number 263 is reserved for vserver */
2349 +#define __NR_vserver           263
2350  #define __NR_statfs64          265
2351  #define __NR_fstatfs64         266
2352  #define __NR_remap_file_pages  267
2353 diff -NurpP --minimal linux-3.0/arch/s390/kernel/ptrace.c linux-3.0-vs2.3.1-pre7/arch/s390/kernel/ptrace.c
2354 --- linux-3.0/arch/s390/kernel/ptrace.c 2011-03-15 18:06:45.000000000 +0100
2355 +++ linux-3.0-vs2.3.1-pre7/arch/s390/kernel/ptrace.c    2011-06-10 22:11:24.000000000 +0200
2356 @@ -20,6 +20,7 @@
2357  #include <linux/regset.h>
2358  #include <linux/tracehook.h>
2359  #include <linux/seccomp.h>
2360 +#include <linux/vs_base.h>
2361  #include <trace/syscall.h>
2362  #include <asm/compat.h>
2363  #include <asm/segment.h>
2364 diff -NurpP --minimal linux-3.0/arch/s390/kernel/syscalls.S linux-3.0-vs2.3.1-pre7/arch/s390/kernel/syscalls.S
2365 --- linux-3.0/arch/s390/kernel/syscalls.S       2011-07-22 11:17:41.000000000 +0200
2366 +++ linux-3.0-vs2.3.1-pre7/arch/s390/kernel/syscalls.S  2011-06-10 22:11:24.000000000 +0200
2367 @@ -271,7 +271,7 @@ SYSCALL(sys_clock_settime,sys_clock_sett
2368  SYSCALL(sys_clock_gettime,sys_clock_gettime,sys32_clock_gettime_wrapper)       /* 260 */
2369  SYSCALL(sys_clock_getres,sys_clock_getres,sys32_clock_getres_wrapper)
2370  SYSCALL(sys_clock_nanosleep,sys_clock_nanosleep,sys32_clock_nanosleep_wrapper)
2371 -NI_SYSCALL                                                     /* reserved for vserver */
2372 +SYSCALL(sys_vserver,sys_vserver,sys32_vserver)
2373  SYSCALL(sys_s390_fadvise64_64,sys_ni_syscall,sys32_fadvise64_64_wrapper)
2374  SYSCALL(sys_statfs64,sys_statfs64,compat_sys_statfs64_wrapper)
2375  SYSCALL(sys_fstatfs64,sys_fstatfs64,compat_sys_fstatfs64_wrapper)
2376 diff -NurpP --minimal linux-3.0/arch/sh/Kconfig linux-3.0-vs2.3.1-pre7/arch/sh/Kconfig
2377 --- linux-3.0/arch/sh/Kconfig   2011-07-22 11:17:41.000000000 +0200
2378 +++ linux-3.0-vs2.3.1-pre7/arch/sh/Kconfig      2011-07-19 00:44:39.000000000 +0200
2379 @@ -893,6 +893,8 @@ source "fs/Kconfig"
2380
2381  source "arch/sh/Kconfig.debug"
2382
2383 +source "kernel/vserver/Kconfig"
2384 +
2385  source "security/Kconfig"
2386
2387  source "crypto/Kconfig"
2388 diff -NurpP --minimal linux-3.0/arch/sh/kernel/irq.c linux-3.0-vs2.3.1-pre7/arch/sh/kernel/irq.c
2389 --- linux-3.0/arch/sh/kernel/irq.c      2011-07-22 11:17:41.000000000 +0200
2390 +++ linux-3.0-vs2.3.1-pre7/arch/sh/kernel/irq.c 2011-07-19 00:45:06.000000000 +0200
2391 @@ -14,6 +14,7 @@
2392  #include <linux/ftrace.h>
2393  #include <linux/delay.h>
2394  #include <linux/ratelimit.h>
2395 +// #include <linux/vs_context.h>
2396  #include <asm/processor.h>
2397  #include <asm/machvec.h>
2398  #include <asm/uaccess.h>
2399 diff -NurpP --minimal linux-3.0/arch/sh/kernel/vsyscall/vsyscall.c linux-3.0-vs2.3.1-pre7/arch/sh/kernel/vsyscall/vsyscall.c
2400 --- linux-3.0/arch/sh/kernel/vsyscall/vsyscall.c        2011-05-22 16:17:07.000000000 +0200
2401 +++ linux-3.0-vs2.3.1-pre7/arch/sh/kernel/vsyscall/vsyscall.c   2011-06-10 22:11:24.000000000 +0200
2402 @@ -18,6 +18,7 @@
2403  #include <linux/elf.h>
2404  #include <linux/sched.h>
2405  #include <linux/err.h>
2406 +#include <linux/vs_memory.h>
2407
2408  /*
2409   * Should the kernel map a VDSO page into processes and pass its
2410 diff -NurpP --minimal linux-3.0/arch/sparc/Kconfig linux-3.0-vs2.3.1-pre7/arch/sparc/Kconfig
2411 --- linux-3.0/arch/sparc/Kconfig        2011-07-22 11:17:41.000000000 +0200
2412 +++ linux-3.0-vs2.3.1-pre7/arch/sparc/Kconfig   2011-06-15 02:40:14.000000000 +0200
2413 @@ -601,6 +601,8 @@ source "fs/Kconfig"
2414
2415  source "arch/sparc/Kconfig.debug"
2416
2417 +source "kernel/vserver/Kconfig"
2418 +
2419  source "security/Kconfig"
2420
2421  source "crypto/Kconfig"
2422 diff -NurpP --minimal linux-3.0/arch/sparc/include/asm/tlb_64.h linux-3.0-vs2.3.1-pre7/arch/sparc/include/asm/tlb_64.h
2423 --- linux-3.0/arch/sparc/include/asm/tlb_64.h   2011-07-22 11:17:42.000000000 +0200
2424 +++ linux-3.0-vs2.3.1-pre7/arch/sparc/include/asm/tlb_64.h      2011-06-10 22:11:24.000000000 +0200
2425 @@ -3,6 +3,7 @@
2426
2427  #include <linux/swap.h>
2428  #include <linux/pagemap.h>
2429 +#include <linux/vs_memory.h>
2430  #include <asm/pgalloc.h>
2431  #include <asm/tlbflush.h>
2432  #include <asm/mmu_context.h>
2433 diff -NurpP --minimal linux-3.0/arch/sparc/include/asm/unistd.h linux-3.0-vs2.3.1-pre7/arch/sparc/include/asm/unistd.h
2434 --- linux-3.0/arch/sparc/include/asm/unistd.h   2011-07-22 11:17:42.000000000 +0200
2435 +++ linux-3.0-vs2.3.1-pre7/arch/sparc/include/asm/unistd.h      2011-06-10 22:11:24.000000000 +0200
2436 @@ -335,7 +335,7 @@
2437  #define __NR_timer_getoverrun  264
2438  #define __NR_timer_delete      265
2439  #define __NR_timer_create      266
2440 -/* #define __NR_vserver                267 Reserved for VSERVER */
2441 +#define __NR_vserver           267
2442  #define __NR_io_setup          268
2443  #define __NR_io_destroy                269
2444  #define __NR_io_submit         270
2445 diff -NurpP --minimal linux-3.0/arch/sparc/kernel/systbls_32.S linux-3.0-vs2.3.1-pre7/arch/sparc/kernel/systbls_32.S
2446 --- linux-3.0/arch/sparc/kernel/systbls_32.S    2011-07-22 11:17:42.000000000 +0200
2447 +++ linux-3.0-vs2.3.1-pre7/arch/sparc/kernel/systbls_32.S       2011-06-10 22:11:24.000000000 +0200
2448 @@ -70,7 +70,7 @@ sys_call_table:
2449  /*250*/        .long sys_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
2450  /*255*/        .long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
2451  /*260*/        .long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
2452 -/*265*/        .long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy
2453 +/*265*/        .long sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
2454  /*270*/        .long sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
2455  /*275*/        .long sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
2456  /*280*/        .long sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
2457 diff -NurpP --minimal linux-3.0/arch/sparc/kernel/systbls_64.S linux-3.0-vs2.3.1-pre7/arch/sparc/kernel/systbls_64.S
2458 --- linux-3.0/arch/sparc/kernel/systbls_64.S    2011-07-22 11:17:42.000000000 +0200
2459 +++ linux-3.0-vs2.3.1-pre7/arch/sparc/kernel/systbls_64.S       2011-06-10 22:11:24.000000000 +0200
2460 @@ -71,7 +71,7 @@ sys_call_table32:
2461  /*250*/        .word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl
2462         .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep
2463  /*260*/        .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun
2464 -       .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy
2465 +       .word sys_timer_delete, compat_sys_timer_create, sys32_vserver, compat_sys_io_setup, sys_io_destroy
2466  /*270*/        .word sys32_io_submit, sys_io_cancel, compat_sys_io_getevents, sys32_mq_open, sys_mq_unlink
2467         .word compat_sys_mq_timedsend, compat_sys_mq_timedreceive, compat_sys_mq_notify, compat_sys_mq_getsetattr, compat_sys_waitid
2468  /*280*/        .word sys32_tee, sys_add_key, sys_request_key, sys_keyctl, compat_sys_openat
2469 @@ -148,7 +148,7 @@ sys_call_table:
2470  /*250*/        .word sys_64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl
2471         .word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep
2472  /*260*/        .word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun
2473 -       .word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy
2474 +       .word sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy
2475  /*270*/        .word sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink
2476         .word sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid
2477  /*280*/        .word sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat
2478 diff -NurpP --minimal linux-3.0/arch/um/Kconfig.rest linux-3.0-vs2.3.1-pre7/arch/um/Kconfig.rest
2479 --- linux-3.0/arch/um/Kconfig.rest      2009-06-11 17:12:19.000000000 +0200
2480 +++ linux-3.0-vs2.3.1-pre7/arch/um/Kconfig.rest 2011-06-10 22:11:24.000000000 +0200
2481 @@ -18,6 +18,8 @@ source "drivers/connector/Kconfig"
2482
2483  source "fs/Kconfig"
2484
2485 +source "kernel/vserver/Kconfig"
2486 +
2487  source "security/Kconfig"
2488
2489  source "crypto/Kconfig"
2490 diff -NurpP --minimal linux-3.0/arch/um/include/asm/tlb.h linux-3.0-vs2.3.1-pre7/arch/um/include/asm/tlb.h
2491 --- linux-3.0/arch/um/include/asm/tlb.h 2011-07-22 11:17:42.000000000 +0200
2492 +++ linux-3.0-vs2.3.1-pre7/arch/um/include/asm/tlb.h    2011-06-10 22:11:24.000000000 +0200
2493 @@ -3,6 +3,7 @@
2494
2495  #include <linux/pagemap.h>
2496  #include <linux/swap.h>
2497 +#include <linux/vs_memory.h>
2498  #include <asm/percpu.h>
2499  #include <asm/pgalloc.h>
2500  #include <asm/tlbflush.h>
2501 diff -NurpP --minimal linux-3.0/arch/um/include/shared/kern_constants.h linux-3.0-vs2.3.1-pre7/arch/um/include/shared/kern_constants.h
2502 --- linux-3.0/arch/um/include/shared/kern_constants.h   1970-01-01 01:00:00.000000000 +0100
2503 +++ linux-3.0-vs2.3.1-pre7/arch/um/include/shared/kern_constants.h      2011-06-10 22:11:24.000000000 +0200
2504 @@ -0,0 +1 @@
2505 +#include "../../../../include/generated/asm-offsets.h"
2506 diff -NurpP --minimal linux-3.0/arch/um/include/shared/user_constants.h linux-3.0-vs2.3.1-pre7/arch/um/include/shared/user_constants.h
2507 --- linux-3.0/arch/um/include/shared/user_constants.h   1970-01-01 01:00:00.000000000 +0100
2508 +++ linux-3.0-vs2.3.1-pre7/arch/um/include/shared/user_constants.h      2011-06-10 22:11:24.000000000 +0200
2509 @@ -0,0 +1,40 @@
2510 +/*
2511 + * DO NOT MODIFY.
2512 + *
2513 + * This file was generated by arch/um/Makefile
2514 + *
2515 + */
2516 +
2517 +#define HOST_SC_CR2 176 /* offsetof(struct sigcontext, cr2)    # */
2518 +#define HOST_SC_ERR 152 /* offsetof(struct sigcontext, err)    # */
2519 +#define HOST_SC_TRAPNO 160 /* offsetof(struct sigcontext, trapno)      # */
2520 +#define HOST_FP_SIZE 64 /* sizeof(struct _fpstate) / sizeof(unsigned long)     # */
2521 +#define HOST_RBX 5 /* RBX      # */
2522 +#define HOST_RCX 11 /* RCX     # */
2523 +#define HOST_RDI 14 /* RDI     # */
2524 +#define HOST_RSI 13 /* RSI     # */
2525 +#define HOST_RDX 12 /* RDX     # */
2526 +#define HOST_RBP 4 /* RBP      # */
2527 +#define HOST_RAX 10 /* RAX     # */
2528 +#define HOST_R8 9 /* R8        # */
2529 +#define HOST_R9 8 /* R9        # */
2530 +#define HOST_R10 7 /* R10      # */
2531 +#define HOST_R11 6 /* R11      # */
2532 +#define HOST_R12 3 /* R12      # */
2533 +#define HOST_R13 2 /* R13      # */
2534 +#define HOST_R14 1 /* R14      # */
2535 +#define HOST_R15 0 /* R15      # */
2536 +#define HOST_ORIG_RAX 15 /* ORIG_RAX   # */
2537 +#define HOST_CS 17 /* CS       # */
2538 +#define HOST_SS 20 /* SS       # */
2539 +#define HOST_EFLAGS 18 /* EFLAGS       # */
2540 +#define HOST_IP 16 /* RIP      # */
2541 +#define HOST_SP 19 /* RSP      # */
2542 +#define UM_FRAME_SIZE 216 /* sizeof(struct user_regs_struct)   # */
2543 +#define UM_POLLIN 1 /* POLLIN  # */
2544 +#define UM_POLLPRI 2 /* POLLPRI        # */
2545 +#define UM_POLLOUT 4 /* POLLOUT        # */
2546 +#define UM_PROT_READ 1 /* PROT_READ    # */
2547 +#define UM_PROT_WRITE 2 /* PROT_WRITE  # */
2548 +#define UM_PROT_EXEC 4 /* PROT_EXEC    # */
2549 +
2550 diff -NurpP --minimal linux-3.0/arch/x86/Kconfig linux-3.0-vs2.3.1-pre7/arch/x86/Kconfig
2551 --- linux-3.0/arch/x86/Kconfig  2011-07-22 11:17:42.000000000 +0200
2552 +++ linux-3.0-vs2.3.1-pre7/arch/x86/Kconfig     2011-07-22 11:20:39.000000000 +0200
2553 @@ -2159,6 +2159,8 @@ source "fs/Kconfig"
2554
2555  source "arch/x86/Kconfig.debug"
2556
2557 +source "kernel/vserver/Kconfig"
2558 +
2559  source "security/Kconfig"
2560
2561  source "crypto/Kconfig"
2562 diff -NurpP --minimal linux-3.0/arch/x86/ia32/ia32entry.S linux-3.0-vs2.3.1-pre7/arch/x86/ia32/ia32entry.S
2563 --- linux-3.0/arch/x86/ia32/ia32entry.S 2011-07-22 11:17:42.000000000 +0200
2564 +++ linux-3.0-vs2.3.1-pre7/arch/x86/ia32/ia32entry.S    2011-06-10 22:11:24.000000000 +0200
2565 @@ -776,7 +776,7 @@ ia32_sys_call_table:
2566         .quad sys_tgkill                /* 270 */
2567         .quad compat_sys_utimes
2568         .quad sys32_fadvise64_64
2569 -       .quad quiet_ni_syscall  /* sys_vserver */
2570 +       .quad sys32_vserver
2571         .quad sys_mbind
2572         .quad compat_sys_get_mempolicy  /* 275 */
2573         .quad sys_set_mempolicy
2574 diff -NurpP --minimal linux-3.0/arch/x86/include/asm/unistd_64.h linux-3.0-vs2.3.1-pre7/arch/x86/include/asm/unistd_64.h
2575 --- linux-3.0/arch/x86/include/asm/unistd_64.h  2011-07-22 11:17:43.000000000 +0200
2576 +++ linux-3.0-vs2.3.1-pre7/arch/x86/include/asm/unistd_64.h     2011-06-10 22:11:24.000000000 +0200
2577 @@ -535,7 +535,7 @@ __SYSCALL(__NR_tgkill, sys_tgkill)
2578  #define __NR_utimes                            235
2579  __SYSCALL(__NR_utimes, sys_utimes)
2580  #define __NR_vserver                           236
2581 -__SYSCALL(__NR_vserver, sys_ni_syscall)
2582 +__SYSCALL(__NR_vserver, sys_vserver)
2583  #define __NR_mbind                             237
2584  __SYSCALL(__NR_mbind, sys_mbind)
2585  #define __NR_set_mempolicy                     238
2586 diff -NurpP --minimal linux-3.0/arch/x86/kernel/syscall_table_32.S linux-3.0-vs2.3.1-pre7/arch/x86/kernel/syscall_table_32.S
2587 --- linux-3.0/arch/x86/kernel/syscall_table_32.S        2011-07-22 11:17:43.000000000 +0200
2588 +++ linux-3.0-vs2.3.1-pre7/arch/x86/kernel/syscall_table_32.S   2011-06-10 22:11:24.000000000 +0200
2589 @@ -272,7 +272,7 @@ ENTRY(sys_call_table)
2590         .long sys_tgkill        /* 270 */
2591         .long sys_utimes
2592         .long sys_fadvise64_64
2593 -       .long sys_ni_syscall    /* sys_vserver */
2594 +       .long sys_vserver
2595         .long sys_mbind
2596         .long sys_get_mempolicy
2597         .long sys_set_mempolicy
2598 diff -NurpP --minimal linux-3.0/block/genhd.c linux-3.0-vs2.3.1-pre7/block/genhd.c
2599 --- linux-3.0/block/genhd.c     2011-07-22 11:17:43.000000000 +0200
2600 +++ linux-3.0-vs2.3.1-pre7/block/genhd.c        2011-07-01 11:35:34.000000000 +0200
2601 @@ -1154,17 +1154,17 @@ static int diskstats_show(struct seq_fil
2602                 cpu = part_stat_lock();
2603                 part_round_stats(cpu, hd);
2604                 part_stat_unlock();
2605 -               seq_printf(seqf, "%4d %7d %s %lu %lu %llu "
2606 -                          "%u %lu %lu %llu %u %u %u %u\n",
2607 +               seq_printf(seqf, "%4d %7d %s %lu %lu %lu "
2608 +                          "%u %lu %lu %lu %u %u %u %u\n",
2609                            MAJOR(part_devt(hd)), MINOR(part_devt(hd)),
2610                            disk_name(gp, hd->partno, buf),
2611                            part_stat_read(hd, ios[READ]),
2612                            part_stat_read(hd, merges[READ]),
2613 -                          (unsigned long long)part_stat_read(hd, sectors[READ]),
2614 +                          part_stat_read(hd, sectors[READ]),
2615                            jiffies_to_msecs(part_stat_read(hd, ticks[READ])),
2616                            part_stat_read(hd, ios[WRITE]),
2617                            part_stat_read(hd, merges[WRITE]),
2618 -                          (unsigned long long)part_stat_read(hd, sectors[WRITE]),
2619 +                          part_stat_read(hd, sectors[WRITE]),
2620                            jiffies_to_msecs(part_stat_read(hd, ticks[WRITE])),
2621                            part_in_flight(hd),
2622                            jiffies_to_msecs(part_stat_read(hd, io_ticks)),
2623 diff -NurpP --minimal linux-3.0/drivers/block/Kconfig linux-3.0-vs2.3.1-pre7/drivers/block/Kconfig
2624 --- linux-3.0/drivers/block/Kconfig     2011-07-22 11:17:44.000000000 +0200
2625 +++ linux-3.0-vs2.3.1-pre7/drivers/block/Kconfig        2011-06-10 22:11:24.000000000 +0200
2626 @@ -273,6 +273,13 @@ config BLK_DEV_CRYPTOLOOP
2627
2628  source "drivers/block/drbd/Kconfig"
2629
2630 +config BLK_DEV_VROOT
2631 +       tristate "Virtual Root device support"
2632 +       depends on QUOTACTL
2633 +       ---help---
2634 +         Saying Y here will allow you to use quota/fs ioctls on a shared
2635 +         partition within a virtual server without compromising security.
2636 +
2637  config BLK_DEV_NBD
2638         tristate "Network block device support"
2639         depends on NET
2640 diff -NurpP --minimal linux-3.0/drivers/block/Makefile linux-3.0-vs2.3.1-pre7/drivers/block/Makefile
2641 --- linux-3.0/drivers/block/Makefile    2011-07-22 11:17:44.000000000 +0200
2642 +++ linux-3.0-vs2.3.1-pre7/drivers/block/Makefile       2011-06-10 22:11:24.000000000 +0200
2643 @@ -34,6 +34,7 @@ obj-$(CONFIG_VIODASD)         += viodasd.o
2644  obj-$(CONFIG_BLK_DEV_SX8)      += sx8.o
2645  obj-$(CONFIG_BLK_DEV_UB)       += ub.o
2646  obj-$(CONFIG_BLK_DEV_HD)       += hd.o
2647 +obj-$(CONFIG_BLK_DEV_VROOT)    += vroot.o
2648
2649  obj-$(CONFIG_XEN_BLKDEV_FRONTEND)      += xen-blkfront.o
2650  obj-$(CONFIG_XEN_BLKDEV_BACKEND)       += xen-blkback/
2651 diff -NurpP --minimal linux-3.0/drivers/block/loop.c linux-3.0-vs2.3.1-pre7/drivers/block/loop.c
2652 --- linux-3.0/drivers/block/loop.c      2011-07-22 11:17:44.000000000 +0200
2653 +++ linux-3.0-vs2.3.1-pre7/drivers/block/loop.c 2011-06-10 22:11:24.000000000 +0200
2654 @@ -75,6 +75,7 @@
2655  #include <linux/kthread.h>
2656  #include <linux/splice.h>
2657  #include <linux/sysfs.h>
2658 +#include <linux/vs_context.h>
2659
2660  #include <asm/uaccess.h>
2661
2662 @@ -891,6 +892,7 @@ static int loop_set_fd(struct loop_devic
2663         lo->lo_blocksize = lo_blocksize;
2664         lo->lo_device = bdev;
2665         lo->lo_flags = lo_flags;
2666 +       lo->lo_xid = vx_current_xid();
2667         lo->lo_backing_file = file;
2668         lo->transfer = transfer_none;
2669         lo->ioctl = NULL;
2670 @@ -1019,6 +1021,7 @@ static int loop_clr_fd(struct loop_devic
2671         lo->lo_encrypt_key_size = 0;
2672         lo->lo_flags = 0;
2673         lo->lo_thread = NULL;
2674 +       lo->lo_xid = 0;
2675         memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE);
2676         memset(lo->lo_crypt_name, 0, LO_NAME_SIZE);
2677         memset(lo->lo_file_name, 0, LO_NAME_SIZE);
2678 @@ -1057,7 +1060,7 @@ loop_set_status(struct loop_device *lo,
2679
2680         if (lo->lo_encrypt_key_size &&
2681             lo->lo_key_owner != uid &&
2682 -           !capable(CAP_SYS_ADMIN))
2683 +           !vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP))
2684                 return -EPERM;
2685         if (lo->lo_state != Lo_bound)
2686                 return -ENXIO;
2687 @@ -1141,7 +1144,8 @@ loop_get_status(struct loop_device *lo,
2688         memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE);
2689         info->lo_encrypt_type =
2690                 lo->lo_encryption ? lo->lo_encryption->number : 0;
2691 -       if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) {
2692 +       if (lo->lo_encrypt_key_size &&
2693 +               vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_CLOOP)) {
2694                 info->lo_encrypt_key_size = lo->lo_encrypt_key_size;
2695                 memcpy(info->lo_encrypt_key, lo->lo_encrypt_key,
2696                        lo->lo_encrypt_key_size);
2697 @@ -1487,6 +1491,9 @@ static int lo_open(struct block_device *
2698  {
2699         struct loop_device *lo = bdev->bd_disk->private_data;
2700
2701 +       if (!vx_check(lo->lo_xid, VS_IDENT|VS_HOSTID|VS_ADMIN_P))
2702 +               return -EACCES;
2703 +
2704         mutex_lock(&lo->lo_ctl_mutex);
2705         lo->lo_refcnt++;
2706         mutex_unlock(&lo->lo_ctl_mutex);
2707 diff -NurpP --minimal linux-3.0/drivers/block/vroot.c linux-3.0-vs2.3.1-pre7/drivers/block/vroot.c
2708 --- linux-3.0/drivers/block/vroot.c     1970-01-01 01:00:00.000000000 +0100
2709 +++ linux-3.0-vs2.3.1-pre7/drivers/block/vroot.c        2011-06-10 22:11:24.000000000 +0200
2710 @@ -0,0 +1,292 @@
2711 +/*
2712 + *  linux/drivers/block/vroot.c
2713 + *
2714 + *  written by Herbert Pötzl, 9/11/2002
2715 + *  ported to 2.6.10 by Herbert Pötzl, 30/12/2004
2716 + *
2717 + *  based on the loop.c code by Theodore Ts'o.
2718 + *
2719 + * Copyright (C) 2002-2007 by Herbert Pötzl.
2720 + * Redistribution of this file is permitted under the
2721 + * GNU General Public License.
2722 + *
2723 + */
2724 +
2725 +#include <linux/module.h>
2726 +#include <linux/moduleparam.h>
2727 +#include <linux/file.h>
2728 +#include <linux/major.h>
2729 +#include <linux/blkdev.h>
2730 +#include <linux/slab.h>
2731 +
2732 +#include <linux/vroot.h>
2733 +#include <linux/vs_context.h>
2734 +
2735 +
2736 +static int max_vroot = 8;
2737 +
2738 +static struct vroot_device *vroot_dev;
2739 +static struct gendisk **disks;
2740 +
2741 +
2742 +static int vroot_set_dev(
2743 +       struct vroot_device *vr,
2744 +       struct block_device *bdev,
2745 +       unsigned int arg)
2746 +{
2747 +       struct block_device *real_bdev;
2748 +       struct file *file;
2749 +       struct inode *inode;
2750 +       int error;
2751 +
2752 +       error = -EBUSY;
2753 +       if (vr->vr_state != Vr_unbound)
2754 +               goto out;
2755 +
2756 +       error = -EBADF;
2757 +       file = fget(arg);
2758 +       if (!file)
2759 +               goto out;
2760 +
2761 +       error = -EINVAL;
2762 +       inode = file->f_dentry->d_inode;
2763 +
2764 +
2765 +       if (S_ISBLK(inode->i_mode)) {
2766 +               real_bdev = inode->i_bdev;
2767 +               vr->vr_device = real_bdev;
2768 +               __iget(real_bdev->bd_inode);
2769 +       } else
2770 +               goto out_fput;
2771 +
2772 +       vxdprintk(VXD_CBIT(misc, 0),
2773 +               "vroot[%d]_set_dev: dev=" VXF_DEV,
2774 +               vr->vr_number, VXD_DEV(real_bdev));
2775 +
2776 +       vr->vr_state = Vr_bound;
2777 +       error = 0;
2778 +
2779 + out_fput:
2780 +       fput(file);
2781 + out:
2782 +       return error;
2783 +}
2784 +
2785 +static int vroot_clr_dev(
2786 +       struct vroot_device *vr,
2787 +       struct block_device *bdev)
2788 +{
2789 +       struct block_device *real_bdev;
2790 +
2791 +       if (vr->vr_state != Vr_bound)
2792 +               return -ENXIO;
2793 +       if (vr->vr_refcnt > 1)  /* we needed one fd for the ioctl */
2794 +               return -EBUSY;
2795 +
2796 +       real_bdev = vr->vr_device;
2797 +
2798 +       vxdprintk(VXD_CBIT(misc, 0),
2799 +               "vroot[%d]_clr_dev: dev=" VXF_DEV,
2800 +               vr->vr_number, VXD_DEV(real_bdev));
2801 +
2802 +       bdput(real_bdev);
2803 +       vr->vr_state = Vr_unbound;
2804 +       vr->vr_device = NULL;
2805 +       return 0;
2806 +}
2807 +
2808 +
2809 +static int vr_ioctl(struct block_device *bdev, fmode_t mode,
2810 +       unsigned int cmd, unsigned long arg)
2811 +{
2812 +       struct vroot_device *vr = bdev->bd_disk->private_data;
2813 +       int err;
2814 +
2815 +       down(&vr->vr_ctl_mutex);
2816 +       switch (cmd) {
2817 +       case VROOT_SET_DEV:
2818 +               err = vroot_set_dev(vr, bdev, arg);
2819 +               break;
2820 +       case VROOT_CLR_DEV:
2821 +               err = vroot_clr_dev(vr, bdev);
2822 +               break;
2823 +       default:
2824 +               err = -EINVAL;
2825 +               break;
2826 +       }
2827 +       up(&vr->vr_ctl_mutex);
2828 +       return err;
2829 +}
2830 +
2831 +static int vr_open(struct block_device *bdev, fmode_t mode)
2832 +{
2833 +       struct vroot_device *vr = bdev->bd_disk->private_data;
2834 +
2835 +       down(&vr->vr_ctl_mutex);
2836 +       vr->vr_refcnt++;
2837 +       up(&vr->vr_ctl_mutex);
2838 +       return 0;
2839 +}
2840 +
2841 +static int vr_release(struct gendisk *disk, fmode_t mode)
2842 +{
2843 +       struct vroot_device *vr = disk->private_data;
2844 +
2845 +       down(&vr->vr_ctl_mutex);
2846 +       --vr->vr_refcnt;
2847 +       up(&vr->vr_ctl_mutex);
2848 +       return 0;
2849 +}
2850 +
2851 +static struct block_device_operations vr_fops = {
2852 +       .owner =        THIS_MODULE,
2853 +       .open =         vr_open,
2854 +       .release =      vr_release,
2855 +       .ioctl =        vr_ioctl,
2856 +};
2857 +
2858 +static int vroot_make_request(struct request_queue *q, struct bio *bio)
2859 +{
2860 +       printk("vroot_make_request %p, %p\n", q, bio);
2861 +       bio_io_error(bio);
2862 +       return 0;
2863 +}
2864 +
2865 +struct block_device *__vroot_get_real_bdev(struct block_device *bdev)
2866 +{
2867 +       struct inode *inode = bdev->bd_inode;
2868 +       struct vroot_device *vr;
2869 +       struct block_device *real_bdev;
2870 +       int minor = iminor(inode);
2871 +
2872 +       vr = &vroot_dev[minor];
2873 +       real_bdev = vr->vr_device;
2874 +
2875 +       vxdprintk(VXD_CBIT(misc, 0),
2876 +               "vroot[%d]_get_real_bdev: dev=" VXF_DEV,
2877 +               vr->vr_number, VXD_DEV(real_bdev));
2878 +
2879 +       if (vr->vr_state != Vr_bound)
2880 +               return ERR_PTR(-ENXIO);
2881 +
2882 +       __iget(real_bdev->bd_inode);
2883 +       return real_bdev;
2884 +}
2885 +
2886 +
2887 +
2888 +/*
2889 + * And now the modules code and kernel interface.
2890 + */
2891 +
2892 +module_param(max_vroot, int, 0);
2893 +
2894 +MODULE_PARM_DESC(max_vroot, "Maximum number of vroot devices (1-256)");
2895 +MODULE_LICENSE("GPL");
2896 +MODULE_ALIAS_BLOCKDEV_MAJOR(VROOT_MAJOR);
2897 +
2898 +MODULE_AUTHOR ("Herbert Pötzl");
2899 +MODULE_DESCRIPTION ("Virtual Root Device Mapper");
2900 +
2901 +
2902 +int __init vroot_init(void)
2903 +{
2904 +       int err, i;
2905 +
2906 +       if (max_vroot < 1 || max_vroot > 256) {
2907 +               max_vroot = MAX_VROOT_DEFAULT;
2908 +               printk(KERN_WARNING "vroot: invalid max_vroot "
2909 +                       "(must be between 1 and 256), "
2910 +                       "using default (%d)\n", max_vroot);
2911 +       }
2912 +
2913 +       if (register_blkdev(VROOT_MAJOR, "vroot"))
2914 +               return -EIO;
2915 +
2916 +       err = -ENOMEM;
2917 +       vroot_dev = kmalloc(max_vroot * sizeof(struct vroot_device), GFP_KERNEL);
2918 +       if (!vroot_dev)
2919 +               goto out_mem1;
2920 +       memset(vroot_dev, 0, max_vroot * sizeof(struct vroot_device));
2921 +
2922 +       disks = kmalloc(max_vroot * sizeof(struct gendisk *), GFP_KERNEL);
2923 +       if (!disks)
2924 +               goto out_mem2;
2925 +
2926 +       for (i = 0; i < max_vroot; i++) {
2927 +               disks[i] = alloc_disk(1);
2928 +               if (!disks[i])
2929 +                       goto out_mem3;
2930 +               disks[i]->queue = blk_alloc_queue(GFP_KERNEL);
2931 +               if (!disks[i]->queue)
2932 +                       goto out_mem3;
2933 +               blk_queue_make_request(disks[i]->queue, vroot_make_request);
2934 +       }
2935 +
2936 +       for (i = 0; i < max_vroot; i++) {
2937 +               struct vroot_device *vr = &vroot_dev[i];
2938 +               struct gendisk *disk = disks[i];
2939 +
2940 +               memset(vr, 0, sizeof(*vr));
2941 +               sema_init(&vr->vr_ctl_mutex, 1);
2942 +               vr->vr_number = i;
2943 +               disk->major = VROOT_MAJOR;
2944 +               disk->first_minor = i;
2945 +               disk->fops = &vr_fops;
2946 +               sprintf(disk->disk_name, "vroot%d", i);
2947 +               disk->private_data = vr;
2948 +       }
2949 +
2950 +       err = register_vroot_grb(&__vroot_get_real_bdev);
2951 +       if (err)
2952 +               goto out_mem3;
2953 +
2954 +       for (i = 0; i < max_vroot; i++)
2955 +               add_disk(disks[i]);
2956 +       printk(KERN_INFO "vroot: loaded (max %d devices)\n", max_vroot);
2957 +       return 0;
2958 +
2959 +out_mem3:
2960 +       while (i--)
2961 +               put_disk(disks[i]);
2962 +       kfree(disks);
2963 +out_mem2:
2964 +       kfree(vroot_dev);
2965 +out_mem1:
2966 +       unregister_blkdev(VROOT_MAJOR, "vroot");
2967 +       printk(KERN_ERR "vroot: ran out of memory\n");
2968 +       return err;
2969 +}
2970 +
2971 +void vroot_exit(void)
2972 +{
2973 +       int i;
2974 +
2975 +       if (unregister_vroot_grb(&__vroot_get_real_bdev))
2976 +               printk(KERN_WARNING "vroot: cannot unregister grb\n");
2977 +
2978 +       for (i = 0; i < max_vroot; i++) {
2979 +               del_gendisk(disks[i]);
2980 +               put_disk(disks[i]);
2981 +       }
2982 +       unregister_blkdev(VROOT_MAJOR, "vroot");
2983 +
2984 +       kfree(disks);
2985 +       kfree(vroot_dev);
2986 +}
2987 +
2988 +module_init(vroot_init);
2989 +module_exit(vroot_exit);
2990 +
2991 +#ifndef MODULE
2992 +
2993 +static int __init max_vroot_setup(char *str)
2994 +{
2995 +       max_vroot = simple_strtol(str, NULL, 0);
2996 +       return 1;
2997 +}
2998 +
2999 +__setup("max_vroot=", max_vroot_setup);
3000 +
3001 +#endif
3002 +
3003 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/r100_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r100_reg_safe.h
3004 --- linux-3.0/drivers/gpu/drm/radeon/r100_reg_safe.h    1970-01-01 01:00:00.000000000 +0100
3005 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r100_reg_safe.h       2011-06-10 22:11:24.000000000 +0200
3006 @@ -0,0 +1,28 @@
3007 +static const unsigned r100_reg_safe_bm[102] = {
3008 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3009 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3010 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3011 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3012 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3013 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3014 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3015 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3016 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3017 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3018 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3019 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3020 +       0xFFFFFFCF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3021 +       0xFFFFFF9F, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3022 +       0x38E7FE1F, 0xFFC3FF8E, 0x7FF8FFFF, 0xFFFF803C,
3023 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3024 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFEFFFF, 0xFFFFFFFF,
3025 +       0x00000000, 0xFFFFFFFD, 0xFFFFFFFF, 0xFFFFFFFF,
3026 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3027 +       0xFFFFFFFF, 0xFFFCFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3028 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3029 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3030 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3031 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3032 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3033 +       0xFFFFFFFF, 0xFFFFFFEF,
3034 +};
3035 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/r200_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r200_reg_safe.h
3036 --- linux-3.0/drivers/gpu/drm/radeon/r200_reg_safe.h    1970-01-01 01:00:00.000000000 +0100
3037 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r200_reg_safe.h       2011-06-10 22:11:24.000000000 +0200
3038 @@ -0,0 +1,28 @@
3039 +static const unsigned r200_reg_safe_bm[102] = {
3040 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3041 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3042 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3043 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3044 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3045 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3046 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3047 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3048 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3049 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3050 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3051 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3052 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3053 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3054 +       0xFFE7FE1F, 0xF003FFFF, 0x7EFFFFFF, 0xFFFF803C,
3055 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3056 +       0xFFFFFFFF, 0xFFFFEFCE, 0xFFFEFFFF, 0xFFFFFFFE,
3057 +       0x020E0FF0, 0xFFCC83FD, 0xFFFFFFFF, 0xFFFFFFFF,
3058 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3059 +       0xFFFBFFFF, 0xEFFCFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3060 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3061 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3062 +       0xDFDFDFDF, 0x3FFDDFDF, 0xFFFFFFFF, 0xFFFFFF7F,
3063 +       0xFFFFFFFF, 0x00FFFFFF, 0x00000000, 0x00000000,
3064 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3065 +       0xFFFFFE3F, 0xFFFFFFEF,
3066 +};
3067 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/r300_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r300_reg_safe.h
3068 --- linux-3.0/drivers/gpu/drm/radeon/r300_reg_safe.h    1970-01-01 01:00:00.000000000 +0100
3069 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r300_reg_safe.h       2011-06-10 22:11:24.000000000 +0200
3070 @@ -0,0 +1,42 @@
3071 +static const unsigned r300_reg_safe_bm[159] = {
3072 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3073 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3074 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3075 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3076 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3077 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3078 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3079 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3080 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3081 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3082 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3083 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3084 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3085 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3086 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
3087 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3088 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
3089 +       0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
3090 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3091 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3092 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3093 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3094 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3095 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3096 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3097 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3098 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3099 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3100 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3101 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3102 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3103 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3104 +       0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
3105 +       0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
3106 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
3107 +       0x00000000, 0x0000C100, 0x00000000, 0x00000000,
3108 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3109 +       0x00000000, 0xFFFF0000, 0xFFFFFFFF, 0xFF80FFFF,
3110 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3111 +       0x0003FC01, 0xFFFFFCF8, 0xFF800B19,
3112 +};
3113 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/r420_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r420_reg_safe.h
3114 --- linux-3.0/drivers/gpu/drm/radeon/r420_reg_safe.h    1970-01-01 01:00:00.000000000 +0100
3115 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/r420_reg_safe.h       2011-06-10 22:11:24.000000000 +0200
3116 @@ -0,0 +1,42 @@
3117 +static const unsigned r420_reg_safe_bm[159] = {
3118 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3119 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3120 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3121 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3122 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3123 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3124 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3125 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3126 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3127 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3128 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3129 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3130 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3131 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3132 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
3133 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3134 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
3135 +       0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
3136 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3137 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3138 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3139 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3140 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3141 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3142 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3143 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3144 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3145 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3146 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3147 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3148 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3149 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3150 +       0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
3151 +       0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
3152 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
3153 +       0x00000000, 0x00000100, 0x00000000, 0x00000000,
3154 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3155 +       0x00000000, 0x00000000, 0x00000000, 0xFF800000,
3156 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3157 +       0x0003FC01, 0xFFFFFCF8, 0xFF800B19,
3158 +};
3159 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/rn50_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/rn50_reg_safe.h
3160 --- linux-3.0/drivers/gpu/drm/radeon/rn50_reg_safe.h    1970-01-01 01:00:00.000000000 +0100
3161 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/rn50_reg_safe.h       2011-06-10 22:11:24.000000000 +0200
3162 @@ -0,0 +1,28 @@
3163 +static const unsigned rn50_reg_safe_bm[102] = {
3164 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3165 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3166 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3167 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3168 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3169 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3170 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3171 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3172 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3173 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3174 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3175 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3176 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3177 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3178 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3179 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3180 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3181 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3182 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3183 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3184 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3185 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3186 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3187 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3188 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3189 +       0xFFFFFFFF, 0xFFFFFFFF,
3190 +};
3191 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/rs600_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/rs600_reg_safe.h
3192 --- linux-3.0/drivers/gpu/drm/radeon/rs600_reg_safe.h   1970-01-01 01:00:00.000000000 +0100
3193 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/rs600_reg_safe.h      2011-06-10 22:11:24.000000000 +0200
3194 @@ -0,0 +1,57 @@
3195 +static const unsigned rs600_reg_safe_bm[219] = {
3196 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3197 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3198 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3199 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3200 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3201 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3202 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3203 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3204 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3205 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3206 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3207 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3208 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3209 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3210 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
3211 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3212 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
3213 +       0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
3214 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3215 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3216 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3217 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3218 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3219 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3220 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3221 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3222 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3223 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3224 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3225 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3226 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3227 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3228 +       0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF,
3229 +       0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF,
3230 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
3231 +       0x00000000, 0x00000100, 0x00000000, 0x00000000,
3232 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3233 +       0x00000000, 0x00000000, 0x00000000, 0xFF800000,
3234 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3235 +       0x0003FC01, 0xFFFFFCF8, 0xFF800B19, 0xFFFFFFFF,
3236 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3237 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3238 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3239 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3240 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3241 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3242 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3243 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3244 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3245 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3246 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3247 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3248 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3249 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3250 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3251 +};
3252 diff -NurpP --minimal linux-3.0/drivers/gpu/drm/radeon/rv515_reg_safe.h linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/rv515_reg_safe.h
3253 --- linux-3.0/drivers/gpu/drm/radeon/rv515_reg_safe.h   1970-01-01 01:00:00.000000000 +0100
3254 +++ linux-3.0-vs2.3.1-pre7/drivers/gpu/drm/radeon/rv515_reg_safe.h      2011-06-10 22:11:24.000000000 +0200
3255 @@ -0,0 +1,57 @@
3256 +static const unsigned rv515_reg_safe_bm[219] = {
3257 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3258 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3259 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3260 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3261 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3262 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3263 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3264 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3265 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3266 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3267 +       0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF,
3268 +       0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF,
3269 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3270 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3271 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F,
3272 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3273 +       0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000,
3274 +       0xF0000038, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF,
3275 +       0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, 0xFFFFFFFF,
3276 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3277 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3278 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3279 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3280 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3281 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3282 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3283 +       0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3284 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3285 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3286 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3287 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3288 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3289 +       0x1FFFF878, 0xFFFFE000, 0xFFFFFE1E, 0xFFFFFFFF,
3290 +       0x388F8F50, 0xFFF88082, 0xFF0000FC, 0xFAE009FF,
3291 +       0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000,
3292 +       0xFFFF8CFC, 0xFFFFC1FF, 0xFFFFFFFF, 0xFFFFFFFF,
3293 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3294 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFF80FFFF,
3295 +       0x00000000, 0x00000000, 0x00000000, 0x00000000,
3296 +       0x0003FC01, 0x3FFFFCF8, 0xFF800B19, 0xFFDFFFFF,
3297 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3298 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3299 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3300 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3301 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3302 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3303 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3304 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3305 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3306 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3307 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3308 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3309 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3310 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3311 +       0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF,
3312 +};
3313 diff -NurpP --minimal linux-3.0/drivers/infiniband/core/addr.c linux-3.0-vs2.3.1-pre7/drivers/infiniband/core/addr.c
3314 --- linux-3.0/drivers/infiniband/core/addr.c    2011-07-22 11:17:45.000000000 +0200
3315 +++ linux-3.0-vs2.3.1-pre7/drivers/infiniband/core/addr.c       2011-06-16 14:16:51.000000000 +0200
3316 @@ -252,7 +252,7 @@ static int addr6_resolve(struct sockaddr
3317
3318         if (ipv6_addr_any(&fl6.saddr)) {
3319                 ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev,
3320 -                                        &fl6.daddr, 0, &fl6.saddr);
3321 +                                        &fl6.daddr, 0, &fl6.saddr, NULL);
3322                 if (ret)
3323                         goto put;
3324
3325 diff -NurpP --minimal linux-3.0/drivers/infiniband/hw/ipath/ipath_user_pages.c linux-3.0-vs2.3.1-pre7/drivers/infiniband/hw/ipath/ipath_user_pages.c
3326 --- linux-3.0/drivers/infiniband/hw/ipath/ipath_user_pages.c    2011-05-22 16:17:16.000000000 +0200
3327 +++ linux-3.0-vs2.3.1-pre7/drivers/infiniband/hw/ipath/ipath_user_pages.c       2011-06-10 22:11:24.000000000 +0200
3328 @@ -35,6 +35,7 @@
3329  #include <linux/device.h>
3330  #include <linux/slab.h>
3331  #include <linux/sched.h>
3332 +#include <linux/vs_memory.h>
3333
3334  #include "ipath_kernel.h"
3335
3336 diff -NurpP --minimal linux-3.0/drivers/md/dm-ioctl.c linux-3.0-vs2.3.1-pre7/drivers/md/dm-ioctl.c
3337 --- linux-3.0/drivers/md/dm-ioctl.c     2011-05-22 16:17:18.000000000 +0200
3338 +++ linux-3.0-vs2.3.1-pre7/drivers/md/dm-ioctl.c        2011-06-10 22:11:24.000000000 +0200
3339 @@ -16,6 +16,7 @@
3340  #include <linux/dm-ioctl.h>
3341  #include <linux/hdreg.h>
3342  #include <linux/compat.h>
3343 +#include <linux/vs_context.h>
3344
3345  #include <asm/uaccess.h>
3346
3347 @@ -106,7 +107,8 @@ static struct hash_cell *__get_name_cell
3348         unsigned int h = hash_str(str);
3349
3350         list_for_each_entry (hc, _name_buckets + h, name_list)
3351 -               if (!strcmp(hc->name, str)) {
3352 +               if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
3353 +                       !strcmp(hc->name, str)) {
3354                         dm_get(hc->md);
3355                         return hc;
3356                 }
3357 @@ -120,7 +122,8 @@ static struct hash_cell *__get_uuid_cell
3358         unsigned int h = hash_str(str);
3359
3360         list_for_each_entry (hc, _uuid_buckets + h, uuid_list)
3361 -               if (!strcmp(hc->uuid, str)) {
3362 +               if (vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT) &&
3363 +                       !strcmp(hc->uuid, str)) {
3364                         dm_get(hc->md);
3365                         return hc;
3366                 }
3367 @@ -427,6 +430,9 @@ typedef int (*ioctl_fn)(struct dm_ioctl
3368
3369  static int remove_all(struct dm_ioctl *param, size_t param_size)
3370  {
3371 +       if (!vx_check(0, VS_ADMIN))
3372 +               return -EPERM;
3373 +
3374         dm_hash_remove_all(1);
3375         param->data_size = 0;
3376         return 0;
3377 @@ -474,6 +480,8 @@ static int list_devices(struct dm_ioctl
3378          */
3379         for (i = 0; i < NUM_BUCKETS; i++) {
3380                 list_for_each_entry (hc, _name_buckets + i, name_list) {
3381 +                       if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
3382 +                               continue;
3383                         needed += sizeof(struct dm_name_list);
3384                         needed += strlen(hc->name) + 1;
3385                         needed += ALIGN_MASK;
3386 @@ -497,6 +505,8 @@ static int list_devices(struct dm_ioctl
3387          */
3388         for (i = 0; i < NUM_BUCKETS; i++) {
3389                 list_for_each_entry (hc, _name_buckets + i, name_list) {
3390 +                       if (!vx_check(dm_get_xid(hc->md), VS_WATCH_P | VS_IDENT))
3391 +                               continue;
3392                         if (old_nl)
3393                                 old_nl->next = (uint32_t) ((void *) nl -
3394                                                            (void *) old_nl);
3395 @@ -731,10 +741,11 @@ static struct hash_cell *__find_device_h
3396         if (!md)
3397                 goto out;
3398
3399 -       mdptr = dm_get_mdptr(md);
3400 +       if (vx_check(dm_get_xid(md), VS_WATCH_P | VS_IDENT))
3401 +               mdptr = dm_get_mdptr(md);
3402 +
3403         if (!mdptr)
3404                 dm_put(md);
3405 -
3406  out:
3407         return mdptr;
3408  }
3409 @@ -1577,8 +1588,8 @@ static int ctl_ioctl(uint command, struc
3410         ioctl_fn fn = NULL;
3411         size_t input_param_size;
3412
3413 -       /* only root can play with this */
3414 -       if (!capable(CAP_SYS_ADMIN))
3415 +       /* only root and certain contexts can play with this */
3416 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_ADMIN_MAPPER))
3417                 return -EACCES;
3418
3419         if (_IOC_TYPE(command) != DM_IOCTL)
3420 diff -NurpP --minimal linux-3.0/drivers/md/dm.c linux-3.0-vs2.3.1-pre7/drivers/md/dm.c
3421 --- linux-3.0/drivers/md/dm.c   2011-05-22 16:17:18.000000000 +0200
3422 +++ linux-3.0-vs2.3.1-pre7/drivers/md/dm.c      2011-06-10 22:11:24.000000000 +0200
3423 @@ -20,6 +20,7 @@
3424  #include <linux/idr.h>
3425  #include <linux/hdreg.h>
3426  #include <linux/delay.h>
3427 +#include <linux/vs_base.h>
3428
3429  #include <trace/events/block.h>
3430
3431 @@ -119,6 +120,7 @@ struct mapped_device {
3432         rwlock_t map_lock;
3433         atomic_t holders;
3434         atomic_t open_count;
3435 +       xid_t xid;
3436
3437         unsigned long flags;
3438
3439 @@ -326,6 +328,7 @@ int dm_deleting_md(struct mapped_device
3440  static int dm_blk_open(struct block_device *bdev, fmode_t mode)
3441  {
3442         struct mapped_device *md;
3443 +       int ret = -ENXIO;
3444
3445         spin_lock(&_minor_lock);
3446
3447 @@ -334,18 +337,19 @@ static int dm_blk_open(struct block_devi
3448                 goto out;
3449
3450         if (test_bit(DMF_FREEING, &md->flags) ||
3451 -           dm_deleting_md(md)) {
3452 -               md = NULL;
3453 +           dm_deleting_md(md))
3454 +               goto out;
3455 +
3456 +       ret = -EACCES;
3457 +       if (!vx_check(md->xid, VS_IDENT|VS_HOSTID))
3458                 goto out;
3459 -       }
3460
3461         dm_get(md);
3462         atomic_inc(&md->open_count);
3463 -
3464 +       ret = 0;
3465  out:
3466         spin_unlock(&_minor_lock);
3467 -
3468 -       return md ? 0 : -ENXIO;
3469 +       return ret;
3470  }
3471
3472  static int dm_blk_close(struct gendisk *disk, fmode_t mode)
3473 @@ -566,6 +570,14 @@ int dm_set_geometry(struct mapped_device
3474         return 0;
3475  }
3476
3477 +/*
3478 + * Get the xid associated with a dm device
3479 + */
3480 +xid_t dm_get_xid(struct mapped_device *md)
3481 +{
3482 +       return md->xid;
3483 +}
3484 +
3485  /*-----------------------------------------------------------------
3486   * CRUD START:
3487   *   A more elegant soln is in the works that uses the queue
3488 @@ -1841,6 +1853,7 @@ static struct mapped_device *alloc_dev(i
3489         INIT_LIST_HEAD(&md->uevent_list);
3490         spin_lock_init(&md->uevent_lock);
3491
3492 +       md->xid = vx_current_xid();
3493         md->queue = blk_alloc_queue(GFP_KERNEL);
3494         if (!md->queue)
3495                 goto bad_queue;
3496 diff -NurpP --minimal linux-3.0/drivers/md/dm.h linux-3.0-vs2.3.1-pre7/drivers/md/dm.h
3497 --- linux-3.0/drivers/md/dm.h   2011-05-22 16:17:18.000000000 +0200
3498 +++ linux-3.0-vs2.3.1-pre7/drivers/md/dm.h      2011-06-10 22:11:24.000000000 +0200
3499 @@ -41,6 +41,8 @@ struct dm_dev_internal {
3500  struct dm_table;
3501  struct dm_md_mempools;
3502
3503 +xid_t dm_get_xid(struct mapped_device *md);
3504 +
3505  /*-----------------------------------------------------------------
3506   * Internal table functions.
3507   *---------------------------------------------------------------*/
3508 diff -NurpP --minimal linux-3.0/drivers/net/tun.c linux-3.0-vs2.3.1-pre7/drivers/net/tun.c
3509 --- linux-3.0/drivers/net/tun.c 2011-07-22 11:17:54.000000000 +0200
3510 +++ linux-3.0-vs2.3.1-pre7/drivers/net/tun.c    2011-06-22 12:39:14.000000000 +0200
3511 @@ -64,6 +64,7 @@
3512  #include <linux/nsproxy.h>
3513  #include <linux/virtio_net.h>
3514  #include <linux/rcupdate.h>
3515 +#include <linux/vs_network.h>
3516  #include <net/net_namespace.h>
3517  #include <net/netns/generic.h>
3518  #include <net/rtnetlink.h>
3519 @@ -121,6 +122,7 @@ struct tun_struct {
3520         unsigned int            flags;
3521         uid_t                   owner;
3522         gid_t                   group;
3523 +       nid_t                   nid;
3524
3525         struct net_device       *dev;
3526         u32                     set_features;
3527 @@ -904,6 +906,7 @@ static void tun_setup(struct net_device
3528
3529         tun->owner = -1;
3530         tun->group = -1;
3531 +       tun->nid = current->nid;
3532
3533         dev->ethtool_ops = &tun_ethtool_ops;
3534         dev->destructor = tun_free_netdev;
3535 @@ -1054,7 +1057,7 @@ static int tun_set_iff(struct net *net,
3536
3537                 if (((tun->owner != -1 && cred->euid != tun->owner) ||
3538                      (tun->group != -1 && !in_egroup_p(tun->group))) &&
3539 -                   !capable(CAP_NET_ADMIN))
3540 +               !cap_raised(current_cap(), CAP_NET_ADMIN))
3541                         return -EPERM;
3542                 err = security_tun_dev_attach(tun->socket.sk);
3543                 if (err < 0)
3544 @@ -1068,7 +1071,7 @@ static int tun_set_iff(struct net *net,
3545                 char *name;
3546                 unsigned long flags = 0;
3547
3548 -               if (!capable(CAP_NET_ADMIN))
3549 +               if (!nx_capable(CAP_NET_ADMIN, NXC_TUN_CREATE))
3550                         return -EPERM;
3551                 err = security_tun_dev_create();
3552                 if (err < 0)
3553 @@ -1136,6 +1139,9 @@ static int tun_set_iff(struct net *net,
3554
3555                 sk->sk_destruct = tun_sock_destruct;
3556
3557 +               if (!nx_check(tun->nid, VS_IDENT | VS_HOSTID | VS_ADMIN_P))
3558 +                       return -EPERM;
3559 +
3560                 err = tun_attach(tun, file);
3561                 if (err < 0)
3562                         goto failed;
3563 @@ -1317,6 +1323,16 @@ static long __tun_chr_ioctl(struct file
3564                 tun_debug(KERN_INFO, tun, "group set to %d\n", tun->group);
3565                 break;
3566
3567 +       case TUNSETNID:
3568 +               if (!capable(CAP_CONTEXT))
3569 +                       return -EPERM;
3570 +
3571 +               /* Set nid owner of the device */
3572 +               tun->nid = (nid_t) arg;
3573 +
3574 +               tun_debug(KERN_INFO, tun, "nid owner set to %u\n", tun->nid);
3575 +               break;
3576 +
3577         case TUNSETLINK:
3578                 /* Only allow setting the type when the interface is down */
3579                 if (tun->dev->flags & IFF_UP) {
3580 diff -NurpP --minimal linux-3.0/drivers/tty/sysrq.c linux-3.0-vs2.3.1-pre7/drivers/tty/sysrq.c
3581 --- linux-3.0/drivers/tty/sysrq.c       2011-05-22 16:17:44.000000000 +0200
3582 +++ linux-3.0-vs2.3.1-pre7/drivers/tty/sysrq.c  2011-06-10 22:11:24.000000000 +0200
3583 @@ -41,6 +41,7 @@
3584  #include <linux/oom.h>
3585  #include <linux/slab.h>
3586  #include <linux/input.h>
3587 +#include <linux/vserver/debug.h>
3588
3589  #include <asm/ptrace.h>
3590  #include <asm/irq_regs.h>
3591 @@ -395,6 +396,21 @@ static struct sysrq_key_op sysrq_unrt_op
3592         .enable_mask    = SYSRQ_ENABLE_RTNICE,
3593  };
3594
3595 +
3596 +#ifdef CONFIG_VSERVER_DEBUG
3597 +static void sysrq_handle_vxinfo(int key)
3598 +{
3599 +       dump_vx_info_inactive((key == 'x') ? 0 : 1);
3600 +}
3601 +
3602 +static struct sysrq_key_op sysrq_showvxinfo_op = {
3603 +       .handler        = sysrq_handle_vxinfo,
3604 +       .help_msg       = "conteXt",
3605 +       .action_msg     = "Show Context Info",
3606 +       .enable_mask    = SYSRQ_ENABLE_DUMP,
3607 +};
3608 +#endif
3609 +
3610  /* Key Operations table and lock */
3611  static DEFINE_SPINLOCK(sysrq_key_table_lock);
3612
3613 @@ -449,7 +465,11 @@ static struct sysrq_key_op *sysrq_key_ta
3614         NULL,                           /* v */
3615         &sysrq_showstate_blocked_op,    /* w */
3616         /* x: May be registered on ppc/powerpc for xmon */
3617 +#ifdef CONFIG_VSERVER_DEBUG
3618 +       &sysrq_showvxinfo_op,           /* x */
3619 +#else
3620         NULL,                           /* x */
3621 +#endif
3622         /* y: May be registered on sparc64 for global register dump */
3623         NULL,                           /* y */
3624         &sysrq_ftrace_dump_op,          /* z */
3625 @@ -464,6 +484,8 @@ static int sysrq_key_table_key2index(int
3626                 retval = key - '0';
3627         else if ((key >= 'a') && (key <= 'z'))
3628                 retval = key + 10 - 'a';
3629 +       else if ((key >= 'A') && (key <= 'Z'))
3630 +               retval = key + 10 - 'A';
3631         else
3632                 retval = -1;
3633         return retval;
3634 diff -NurpP --minimal linux-3.0/drivers/tty/tty_io.c linux-3.0-vs2.3.1-pre7/drivers/tty/tty_io.c
3635 --- linux-3.0/drivers/tty/tty_io.c      2011-07-22 11:18:03.000000000 +0200
3636 +++ linux-3.0-vs2.3.1-pre7/drivers/tty/tty_io.c 2011-06-10 22:11:24.000000000 +0200
3637 @@ -104,6 +104,7 @@
3638
3639  #include <linux/kmod.h>
3640  #include <linux/nsproxy.h>
3641 +#include <linux/vs_pid.h>
3642
3643  #undef TTY_DEBUG_HANGUP
3644
3645 @@ -2057,7 +2058,8 @@ static int tiocsti(struct tty_struct *tt
3646         char ch, mbz = 0;
3647         struct tty_ldisc *ld;
3648
3649 -       if ((current->signal->tty != tty) && !capable(CAP_SYS_ADMIN))
3650 +       if (((current->signal->tty != tty) &&
3651 +               !vx_capable(CAP_SYS_ADMIN, VXC_TIOCSTI)))
3652                 return -EPERM;
3653         if (get_user(ch, p))
3654                 return -EFAULT;
3655 @@ -2345,6 +2347,7 @@ static int tiocspgrp(struct tty_struct *
3656                 return -ENOTTY;
3657         if (get_user(pgrp_nr, p))
3658                 return -EFAULT;
3659 +       pgrp_nr = vx_rmap_pid(pgrp_nr);
3660         if (pgrp_nr < 0)
3661                 return -EINVAL;
3662         rcu_read_lock();
3663 diff -NurpP --minimal linux-3.0/fs/attr.c linux-3.0-vs2.3.1-pre7/fs/attr.c
3664 --- linux-3.0/fs/attr.c 2011-07-22 11:18:05.000000000 +0200
3665 +++ linux-3.0-vs2.3.1-pre7/fs/attr.c    2011-06-10 22:11:24.000000000 +0200
3666 @@ -13,6 +13,9 @@
3667  #include <linux/fsnotify.h>
3668  #include <linux/fcntl.h>
3669  #include <linux/security.h>
3670 +#include <linux/proc_fs.h>
3671 +#include <linux/devpts_fs.h>
3672 +#include <linux/vs_tag.h>
3673
3674  /**
3675   * inode_change_ok - check if attribute changes to an inode are allowed
3676 @@ -73,6 +76,10 @@ int inode_change_ok(const struct inode *
3677                         return -EPERM;
3678         }
3679
3680 +       /* check for inode tag permission */
3681 +       if (dx_permission(inode, MAY_WRITE))
3682 +               return -EACCES;
3683 +
3684         return 0;
3685  }
3686  EXPORT_SYMBOL(inode_change_ok);
3687 @@ -143,6 +150,8 @@ void setattr_copy(struct inode *inode, c
3688                 inode->i_uid = attr->ia_uid;
3689         if (ia_valid & ATTR_GID)
3690                 inode->i_gid = attr->ia_gid;
3691 +       if ((ia_valid & ATTR_TAG) && IS_TAGGED(inode))
3692 +               inode->i_tag = attr->ia_tag;
3693         if (ia_valid & ATTR_ATIME)
3694                 inode->i_atime = timespec_trunc(attr->ia_atime,
3695                                                 inode->i_sb->s_time_gran);
3696 @@ -170,7 +179,8 @@ int notify_change(struct dentry * dentry
3697         struct timespec now;
3698         unsigned int ia_valid = attr->ia_valid;
3699
3700 -       if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID | ATTR_TIMES_SET)) {
3701 +       if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID |
3702 +               ATTR_TAG | ATTR_TIMES_SET)) {
3703                 if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
3704                         return -EPERM;
3705         }
3706 diff -NurpP --minimal linux-3.0/fs/binfmt_aout.c linux-3.0-vs2.3.1-pre7/fs/binfmt_aout.c
3707 --- linux-3.0/fs/binfmt_aout.c  2010-10-21 13:07:47.000000000 +0200
3708 +++ linux-3.0-vs2.3.1-pre7/fs/binfmt_aout.c     2011-06-10 22:11:24.000000000 +0200
3709 @@ -25,6 +25,7 @@
3710  #include <linux/init.h>
3711  #include <linux/coredump.h>
3712  #include <linux/slab.h>
3713 +#include <linux/vs_memory.h>
3714
3715  #include <asm/system.h>
3716  #include <asm/uaccess.h>
3717 diff -NurpP --minimal linux-3.0/fs/binfmt_elf.c linux-3.0-vs2.3.1-pre7/fs/binfmt_elf.c
3718 --- linux-3.0/fs/binfmt_elf.c   2011-05-22 16:17:48.000000000 +0200
3719 +++ linux-3.0-vs2.3.1-pre7/fs/binfmt_elf.c      2011-06-10 22:11:24.000000000 +0200
3720 @@ -32,6 +32,7 @@
3721  #include <linux/elf.h>
3722  #include <linux/utsname.h>
3723  #include <linux/coredump.h>
3724 +#include <linux/vs_memory.h>
3725  #include <asm/uaccess.h>
3726  #include <asm/param.h>
3727  #include <asm/page.h>
3728 diff -NurpP --minimal linux-3.0/fs/binfmt_flat.c linux-3.0-vs2.3.1-pre7/fs/binfmt_flat.c
3729 --- linux-3.0/fs/binfmt_flat.c  2011-07-22 11:18:05.000000000 +0200
3730 +++ linux-3.0-vs2.3.1-pre7/fs/binfmt_flat.c     2011-06-10 22:11:24.000000000 +0200
3731 @@ -35,6 +35,7 @@
3732  #include <linux/init.h>
3733  #include <linux/flat.h>
3734  #include <linux/syscalls.h>
3735 +#include <linux/vs_memory.h>
3736
3737  #include <asm/byteorder.h>
3738  #include <asm/system.h>
3739 diff -NurpP --minimal linux-3.0/fs/binfmt_som.c linux-3.0-vs2.3.1-pre7/fs/binfmt_som.c
3740 --- linux-3.0/fs/binfmt_som.c   2010-02-25 11:52:04.000000000 +0100
3741 +++ linux-3.0-vs2.3.1-pre7/fs/binfmt_som.c      2011-06-10 22:11:24.000000000 +0200
3742 @@ -28,6 +28,7 @@
3743  #include <linux/shm.h>
3744  #include <linux/personality.h>
3745  #include <linux/init.h>
3746 +#include <linux/vs_memory.h>
3747
3748  #include <asm/uaccess.h>
3749  #include <asm/pgtable.h>
3750 diff -NurpP --minimal linux-3.0/fs/block_dev.c linux-3.0-vs2.3.1-pre7/fs/block_dev.c
3751 --- linux-3.0/fs/block_dev.c    2011-07-22 11:18:05.000000000 +0200
3752 +++ linux-3.0-vs2.3.1-pre7/fs/block_dev.c       2011-07-01 11:35:34.000000000 +0200
3753 @@ -25,6 +25,7 @@
3754  #include <linux/namei.h>
3755  #include <linux/log2.h>
3756  #include <linux/kmemleak.h>
3757 +#include <linux/vs_device.h>
3758  #include <asm/uaccess.h>
3759  #include "internal.h"
3760
3761 @@ -553,6 +554,7 @@ struct block_device *bdget(dev_t dev)
3762                 bdev->bd_invalidated = 0;
3763                 inode->i_mode = S_IFBLK;
3764                 inode->i_rdev = dev;
3765 +               inode->i_mdev = dev;
3766                 inode->i_bdev = bdev;
3767                 inode->i_data.a_ops = &def_blk_aops;
3768                 mapping_set_gfp_mask(&inode->i_data, GFP_USER);
3769 @@ -599,6 +601,11 @@ EXPORT_SYMBOL(bdput);
3770  static struct block_device *bd_acquire(struct inode *inode)
3771  {
3772         struct block_device *bdev;
3773 +       dev_t mdev;
3774 +
3775 +       if (!vs_map_blkdev(inode->i_rdev, &mdev, DATTR_OPEN))
3776 +               return NULL;
3777 +       inode->i_mdev = mdev;
3778
3779         spin_lock(&bdev_lock);
3780         bdev = inode->i_bdev;
3781 @@ -609,7 +616,7 @@ static struct block_device *bd_acquire(s
3782         }
3783         spin_unlock(&bdev_lock);
3784
3785 -       bdev = bdget(inode->i_rdev);
3786 +       bdev = bdget(mdev);
3787         if (bdev) {
3788                 spin_lock(&bdev_lock);
3789                 if (!inode->i_bdev) {
3790 diff -NurpP --minimal linux-3.0/fs/btrfs/ctree.h linux-3.0-vs2.3.1-pre7/fs/btrfs/ctree.h
3791 --- linux-3.0/fs/btrfs/ctree.h  2011-07-22 11:18:05.000000000 +0200
3792 +++ linux-3.0-vs2.3.1-pre7/fs/btrfs/ctree.h     2011-07-19 00:44:39.000000000 +0200
3793 @@ -600,11 +600,14 @@ struct btrfs_inode_item {
3794         /* modification sequence number for NFS */
3795         __le64 sequence;
3796
3797 +       __le16 tag;
3798         /*
3799          * a little future expansion, for more than this we can
3800          * just grow the inode item and version it
3801          */
3802 -       __le64 reserved[4];
3803 +       __le16 reserved16;
3804 +       __le32 reserved32;
3805 +       __le64 reserved[3];
3806         struct btrfs_timespec atime;
3807         struct btrfs_timespec ctime;
3808         struct btrfs_timespec mtime;
3809 @@ -1359,6 +1362,8 @@ struct btrfs_ioctl_defrag_range_args {
3810  #define BTRFS_MOUNT_AUTO_DEFRAG                (1 << 16)
3811  #define BTRFS_MOUNT_INODE_MAP_CACHE    (1 << 17)
3812
3813 +#define BTRFS_MOUNT_TAGGED             (1 << 24)
3814 +
3815  #define btrfs_clear_opt(o, opt)                ((o) &= ~BTRFS_MOUNT_##opt)
3816  #define btrfs_set_opt(o, opt)          ((o) |= BTRFS_MOUNT_##opt)
3817  #define btrfs_test_opt(root, opt)      ((root)->fs_info->mount_opt & \
3818 @@ -1568,6 +1573,7 @@ BTRFS_SETGET_FUNCS(inode_block_group, st
3819  BTRFS_SETGET_FUNCS(inode_nlink, struct btrfs_inode_item, nlink, 32);
3820  BTRFS_SETGET_FUNCS(inode_uid, struct btrfs_inode_item, uid, 32);
3821  BTRFS_SETGET_FUNCS(inode_gid, struct btrfs_inode_item, gid, 32);
3822 +BTRFS_SETGET_FUNCS(inode_tag, struct btrfs_inode_item, tag, 16);
3823  BTRFS_SETGET_FUNCS(inode_mode, struct btrfs_inode_item, mode, 32);
3824  BTRFS_SETGET_FUNCS(inode_rdev, struct btrfs_inode_item, rdev, 64);
3825  BTRFS_SETGET_FUNCS(inode_flags, struct btrfs_inode_item, flags, 64);
3826 @@ -1621,6 +1627,10 @@ BTRFS_SETGET_FUNCS(extent_flags, struct
3827
3828  BTRFS_SETGET_FUNCS(extent_refs_v0, struct btrfs_extent_item_v0, refs, 32);
3829
3830 +#define BTRFS_INODE_IXUNLINK           (1 << 24)
3831 +#define BTRFS_INODE_BARRIER            (1 << 25)
3832 +#define BTRFS_INODE_COW                        (1 << 26)
3833 +
3834
3835  BTRFS_SETGET_FUNCS(tree_block_level, struct btrfs_tree_block_info, level, 8);
3836
3837 @@ -2595,6 +2605,7 @@ extern const struct dentry_operations bt
3838  long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
3839  void btrfs_update_iflags(struct inode *inode);
3840  void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
3841 +int btrfs_sync_flags(struct inode *inode, int, int);
3842  int btrfs_defrag_file(struct inode *inode, struct file *file,
3843                       struct btrfs_ioctl_defrag_range_args *range,
3844                       u64 newer_than, unsigned long max_pages);
3845 diff -NurpP --minimal linux-3.0/fs/btrfs/disk-io.c linux-3.0-vs2.3.1-pre7/fs/btrfs/disk-io.c
3846 --- linux-3.0/fs/btrfs/disk-io.c        2011-07-22 11:18:05.000000000 +0200
3847 +++ linux-3.0-vs2.3.1-pre7/fs/btrfs/disk-io.c   2011-06-22 12:39:15.000000000 +0200
3848 @@ -1765,6 +1765,9 @@ struct btrfs_root *open_ctree(struct sup
3849                 goto fail_alloc;
3850         }
3851
3852 +       if (btrfs_test_opt(tree_root, TAGGED))
3853 +               sb->s_flags |= MS_TAGGED;
3854 +
3855         features = btrfs_super_incompat_flags(disk_super) &
3856                 ~BTRFS_FEATURE_INCOMPAT_SUPP;
3857         if (features) {
3858 diff -NurpP --minimal linux-3.0/fs/btrfs/inode.c linux-3.0-vs2.3.1-pre7/fs/btrfs/inode.c
3859 --- linux-3.0/fs/btrfs/inode.c  2011-07-22 11:18:05.000000000 +0200
3860 +++ linux-3.0-vs2.3.1-pre7/fs/btrfs/inode.c     2011-07-19 00:44:39.000000000 +0200
3861 @@ -38,6 +38,7 @@
3862  #include <linux/falloc.h>
3863  #include <linux/slab.h>
3864  #include <linux/ratelimit.h>
3865 +#include <linux/vs_tag.h>
3866  #include "compat.h"
3867  #include "ctree.h"
3868  #include "disk-io.h"
3869 @@ -2508,6 +2509,8 @@ static void btrfs_read_locked_inode(stru
3870         struct btrfs_key location;
3871         int maybe_acls;
3872         u32 rdev;
3873 +       uid_t uid;
3874 +       gid_t gid;
3875         int ret;
3876         bool filled = false;
3877
3878 @@ -2540,8 +2543,13 @@ static void btrfs_read_locked_inode(stru
3879
3880         inode->i_mode = btrfs_inode_mode(leaf, inode_item);
3881         inode->i_nlink = btrfs_inode_nlink(leaf, inode_item);
3882 -       inode->i_uid = btrfs_inode_uid(leaf, inode_item);
3883 -       inode->i_gid = btrfs_inode_gid(leaf, inode_item);
3884 +
3885 +       uid = btrfs_inode_uid(leaf, inode_item);
3886 +       gid = btrfs_inode_gid(leaf, inode_item);
3887 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
3888 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
3889 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
3890 +               btrfs_inode_tag(leaf, inode_item));
3891         btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item));
3892
3893         tspec = btrfs_inode_atime(inode_item);
3894 @@ -2624,6 +2632,9 @@ static void fill_inode_item(struct btrfs
3895                             struct btrfs_inode_item *item,
3896                             struct inode *inode)
3897  {
3898 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
3899 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
3900 +
3901         if (!leaf->map_token)
3902                 map_private_extent_buffer(leaf, (unsigned long)item,
3903                                           sizeof(struct btrfs_inode_item),
3904 @@ -2631,8 +2642,11 @@ static void fill_inode_item(struct btrfs
3905                                           &leaf->map_start, &leaf->map_len,
3906                                           KM_USER1);
3907
3908 -       btrfs_set_inode_uid(leaf, item, inode->i_uid);
3909 -       btrfs_set_inode_gid(leaf, item, inode->i_gid);
3910 +       btrfs_set_inode_uid(leaf, item, uid);
3911 +       btrfs_set_inode_gid(leaf, item, gid);
3912 +#ifdef CONFIG_TAGGING_INTERN
3913 +       btrfs_set_inode_tag(leaf, item, inode->i_tag);
3914 +#endif
3915         btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size);
3916         btrfs_set_inode_mode(leaf, item, inode->i_mode);
3917         btrfs_set_inode_nlink(leaf, item, inode->i_nlink);
3918 @@ -7359,6 +7373,7 @@ static const struct inode_operations btr
3919         .listxattr      = btrfs_listxattr,
3920         .removexattr    = btrfs_removexattr,
3921         .permission     = btrfs_permission,
3922 +       .sync_flags     = btrfs_sync_flags,
3923  };
3924  static const struct inode_operations btrfs_dir_ro_inode_operations = {
3925         .lookup         = btrfs_lookup,
3926 @@ -7431,6 +7446,7 @@ static const struct inode_operations btr
3927         .removexattr    = btrfs_removexattr,
3928         .permission     = btrfs_permission,
3929         .fiemap         = btrfs_fiemap,
3930 +       .sync_flags     = btrfs_sync_flags,
3931  };
3932  static const struct inode_operations btrfs_special_inode_operations = {
3933         .getattr        = btrfs_getattr,
3934 diff -NurpP --minimal linux-3.0/fs/btrfs/ioctl.c linux-3.0-vs2.3.1-pre7/fs/btrfs/ioctl.c
3935 --- linux-3.0/fs/btrfs/ioctl.c  2011-07-22 11:18:05.000000000 +0200
3936 +++ linux-3.0-vs2.3.1-pre7/fs/btrfs/ioctl.c     2011-06-22 12:39:15.000000000 +0200
3937 @@ -70,10 +70,13 @@ static unsigned int btrfs_flags_to_ioctl
3938  {
3939         unsigned int iflags = 0;
3940
3941 -       if (flags & BTRFS_INODE_SYNC)
3942 -               iflags |= FS_SYNC_FL;
3943         if (flags & BTRFS_INODE_IMMUTABLE)
3944                 iflags |= FS_IMMUTABLE_FL;
3945 +       if (flags & BTRFS_INODE_IXUNLINK)
3946 +               iflags |= FS_IXUNLINK_FL;
3947 +
3948 +       if (flags & BTRFS_INODE_SYNC)
3949 +               iflags |= FS_SYNC_FL;
3950         if (flags & BTRFS_INODE_APPEND)
3951                 iflags |= FS_APPEND_FL;
3952         if (flags & BTRFS_INODE_NODUMP)
3953 @@ -90,28 +93,78 @@ static unsigned int btrfs_flags_to_ioctl
3954         else if (flags & BTRFS_INODE_NOCOMPRESS)
3955                 iflags |= FS_NOCOMP_FL;
3956
3957 +       if (flags & BTRFS_INODE_BARRIER)
3958 +               iflags |= FS_BARRIER_FL;
3959 +       if (flags & BTRFS_INODE_COW)
3960 +               iflags |= FS_COW_FL;
3961         return iflags;
3962  }
3963
3964  /*
3965 - * Update inode->i_flags based on the btrfs internal flags.
3966 + * Update inode->i_(v)flags based on the btrfs internal flags.
3967   */
3968  void btrfs_update_iflags(struct inode *inode)
3969  {
3970         struct btrfs_inode *ip = BTRFS_I(inode);
3971
3972 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
3973 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
3974 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
3975
3976 -       if (ip->flags & BTRFS_INODE_SYNC)
3977 -               inode->i_flags |= S_SYNC;
3978         if (ip->flags & BTRFS_INODE_IMMUTABLE)
3979                 inode->i_flags |= S_IMMUTABLE;
3980 +       if (ip->flags & BTRFS_INODE_IXUNLINK)
3981 +               inode->i_flags |= S_IXUNLINK;
3982 +
3983 +       if (ip->flags & BTRFS_INODE_SYNC)
3984 +               inode->i_flags |= S_SYNC;
3985         if (ip->flags & BTRFS_INODE_APPEND)
3986                 inode->i_flags |= S_APPEND;
3987         if (ip->flags & BTRFS_INODE_NOATIME)
3988                 inode->i_flags |= S_NOATIME;
3989         if (ip->flags & BTRFS_INODE_DIRSYNC)
3990                 inode->i_flags |= S_DIRSYNC;
3991 +
3992 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
3993 +
3994 +       if (ip->flags & BTRFS_INODE_BARRIER)
3995 +               inode->i_vflags |= V_BARRIER;
3996 +       if (ip->flags & BTRFS_INODE_COW)
3997 +               inode->i_vflags |= V_COW;
3998 +}
3999 +
4000 +/*
4001 + * Update btrfs internal flags from inode->i_(v)flags.
4002 + */
4003 +void btrfs_update_flags(struct inode *inode)
4004 +{
4005 +       struct btrfs_inode *ip = BTRFS_I(inode);
4006 +
4007 +       unsigned int flags = inode->i_flags;
4008 +       unsigned int vflags = inode->i_vflags;
4009 +
4010 +       ip->flags &= ~(BTRFS_INODE_SYNC | BTRFS_INODE_APPEND |
4011 +                       BTRFS_INODE_IMMUTABLE | BTRFS_INODE_IXUNLINK |
4012 +                       BTRFS_INODE_NOATIME | BTRFS_INODE_DIRSYNC |
4013 +                       BTRFS_INODE_BARRIER | BTRFS_INODE_COW);
4014 +
4015 +       if (flags & S_IMMUTABLE)
4016 +               ip->flags |= BTRFS_INODE_IMMUTABLE;
4017 +       if (flags & S_IXUNLINK)
4018 +               ip->flags |= BTRFS_INODE_IXUNLINK;
4019 +
4020 +       if (flags & S_SYNC)
4021 +               ip->flags |= BTRFS_INODE_SYNC;
4022 +       if (flags & S_APPEND)
4023 +               ip->flags |= BTRFS_INODE_APPEND;
4024 +       if (flags & S_NOATIME)
4025 +               ip->flags |= BTRFS_INODE_NOATIME;
4026 +       if (flags & S_DIRSYNC)
4027 +               ip->flags |= BTRFS_INODE_DIRSYNC;
4028 +
4029 +       if (vflags & V_BARRIER)
4030 +               ip->flags |= BTRFS_INODE_BARRIER;
4031 +       if (vflags & V_COW)
4032 +               ip->flags |= BTRFS_INODE_COW;
4033  }
4034
4035  /*
4036 @@ -129,7 +182,7 @@ void btrfs_inherit_iflags(struct inode *
4037         flags = BTRFS_I(dir)->flags;
4038
4039         if (S_ISREG(inode->i_mode))
4040 -               flags &= ~BTRFS_INODE_DIRSYNC;
4041 +               flags &= ~(BTRFS_INODE_DIRSYNC | BTRFS_INODE_BARRIER);
4042         else if (!S_ISDIR(inode->i_mode))
4043                 flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME);
4044
4045 @@ -137,6 +190,30 @@ void btrfs_inherit_iflags(struct inode *
4046         btrfs_update_iflags(inode);
4047  }
4048
4049 +int btrfs_sync_flags(struct inode *inode, int flags, int vflags)
4050 +{
4051 +       struct btrfs_inode *ip = BTRFS_I(inode);
4052 +       struct btrfs_root *root = ip->root;
4053 +       struct btrfs_trans_handle *trans;
4054 +       int ret;
4055 +
4056 +       trans = btrfs_join_transaction(root);
4057 +       BUG_ON(!trans);
4058 +
4059 +       inode->i_flags = flags;
4060 +       inode->i_vflags = vflags;
4061 +       btrfs_update_flags(inode);
4062 +
4063 +       ret = btrfs_update_inode(trans, root, inode);
4064 +       BUG_ON(ret);
4065 +
4066 +       btrfs_update_iflags(inode);
4067 +       inode->i_ctime = CURRENT_TIME;
4068 +       btrfs_end_transaction(trans, root);
4069 +
4070 +       return 0;
4071 +}
4072 +
4073  static int btrfs_ioctl_getflags(struct file *file, void __user *arg)
4074  {
4075         struct btrfs_inode *ip = BTRFS_I(file->f_path.dentry->d_inode);
4076 @@ -188,7 +265,8 @@ static int btrfs_ioctl_setflags(struct f
4077
4078         flags = btrfs_mask_flags(inode->i_mode, flags);
4079         oldflags = btrfs_flags_to_ioctl(ip->flags);
4080 -       if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) {
4081 +       if ((flags ^ oldflags) & (FS_APPEND_FL |
4082 +               FS_IMMUTABLE_FL | FS_IXUNLINK_FL)) {
4083                 if (!capable(CAP_LINUX_IMMUTABLE)) {
4084                         ret = -EPERM;
4085                         goto out_unlock;
4086 @@ -199,14 +277,19 @@ static int btrfs_ioctl_setflags(struct f
4087         if (ret)
4088                 goto out_unlock;
4089
4090 -       if (flags & FS_SYNC_FL)
4091 -               ip->flags |= BTRFS_INODE_SYNC;
4092 -       else
4093 -               ip->flags &= ~BTRFS_INODE_SYNC;
4094         if (flags & FS_IMMUTABLE_FL)
4095                 ip->flags |= BTRFS_INODE_IMMUTABLE;
4096         else
4097                 ip->flags &= ~BTRFS_INODE_IMMUTABLE;
4098 +       if (flags & FS_IXUNLINK_FL)
4099 +               ip->flags |= BTRFS_INODE_IXUNLINK;
4100 +       else
4101 +               ip->flags &= ~BTRFS_INODE_IXUNLINK;
4102 +
4103 +       if (flags & FS_SYNC_FL)
4104 +               ip->flags |= BTRFS_INODE_SYNC;
4105 +       else
4106 +               ip->flags &= ~BTRFS_INODE_SYNC;
4107         if (flags & FS_APPEND_FL)
4108                 ip->flags |= BTRFS_INODE_APPEND;
4109         else
4110 diff -NurpP --minimal linux-3.0/fs/btrfs/super.c linux-3.0-vs2.3.1-pre7/fs/btrfs/super.c
4111 --- linux-3.0/fs/btrfs/super.c  2011-07-22 11:18:05.000000000 +0200
4112 +++ linux-3.0-vs2.3.1-pre7/fs/btrfs/super.c     2011-07-19 00:44:39.000000000 +0200
4113 @@ -162,7 +162,7 @@ enum {
4114         Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
4115         Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
4116         Opt_enospc_debug, Opt_subvolrootid, Opt_defrag,
4117 -       Opt_inode_cache, Opt_err,
4118 +       Opt_inode_cache, Opt_tag, Opt_notag, Opt_tagid, Opt_err,
4119  };
4120
4121  static match_table_t tokens = {
4122 @@ -195,6 +195,9 @@ static match_table_t tokens = {
4123         {Opt_subvolrootid, "subvolrootid=%d"},
4124         {Opt_defrag, "autodefrag"},
4125         {Opt_inode_cache, "inode_cache"},
4126 +       {Opt_tag, "tag"},
4127 +       {Opt_notag, "notag"},
4128 +       {Opt_tagid, "tagid=%u"},
4129         {Opt_err, NULL},
4130  };
4131
4132 @@ -381,6 +384,22 @@ int btrfs_parse_options(struct btrfs_roo
4133                         printk(KERN_INFO "btrfs: enabling auto defrag");
4134                         btrfs_set_opt(info->mount_opt, AUTO_DEFRAG);
4135                         break;
4136 +#ifndef CONFIG_TAGGING_NONE
4137 +               case Opt_tag:
4138 +                       printk(KERN_INFO "btrfs: use tagging\n");
4139 +                       btrfs_set_opt(info->mount_opt, TAGGED);
4140 +                       break;
4141 +               case Opt_notag:
4142 +                       printk(KERN_INFO "btrfs: disabled tagging\n");
4143 +                       btrfs_clear_opt(info->mount_opt, TAGGED);
4144 +                       break;
4145 +#endif
4146 +#ifdef CONFIG_PROPAGATE
4147 +               case Opt_tagid:
4148 +                       /* use args[0] */
4149 +                       btrfs_set_opt(info->mount_opt, TAGGED);
4150 +                       break;
4151 +#endif
4152                 case Opt_err:
4153                         printk(KERN_INFO "btrfs: unrecognized mount option "
4154                                "'%s'\n", p);
4155 @@ -907,6 +926,12 @@ static int btrfs_remount(struct super_bl
4156         if (ret)
4157                 return -EINVAL;
4158
4159 +       if (btrfs_test_opt(root, TAGGED) && !(sb->s_flags & MS_TAGGED)) {
4160 +               printk("btrfs: %s: tagging not permitted on remount.\n",
4161 +                       sb->s_id);
4162 +               return -EINVAL;
4163 +       }
4164 +
4165         if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
4166                 return 0;
4167
4168 diff -NurpP --minimal linux-3.0/fs/char_dev.c linux-3.0-vs2.3.1-pre7/fs/char_dev.c
4169 --- linux-3.0/fs/char_dev.c     2011-03-15 18:07:31.000000000 +0100
4170 +++ linux-3.0-vs2.3.1-pre7/fs/char_dev.c        2011-06-10 22:11:24.000000000 +0200
4171 @@ -21,6 +21,8 @@
4172  #include <linux/mutex.h>
4173  #include <linux/backing-dev.h>
4174  #include <linux/tty.h>
4175 +#include <linux/vs_context.h>
4176 +#include <linux/vs_device.h>
4177
4178  #include "internal.h"
4179
4180 @@ -371,14 +373,21 @@ static int chrdev_open(struct inode *ino
4181         struct cdev *p;
4182         struct cdev *new = NULL;
4183         int ret = 0;
4184 +       dev_t mdev;
4185 +
4186 +       if (!vs_map_chrdev(inode->i_rdev, &mdev, DATTR_OPEN))
4187 +               return -EPERM;
4188 +       inode->i_mdev = mdev;
4189
4190         spin_lock(&cdev_lock);
4191         p = inode->i_cdev;
4192         if (!p) {
4193                 struct kobject *kobj;
4194                 int idx;
4195 +
4196                 spin_unlock(&cdev_lock);
4197 -               kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx);
4198 +
4199 +               kobj = kobj_lookup(cdev_map, mdev, &idx);
4200                 if (!kobj)
4201                         return -ENXIO;
4202                 new = container_of(kobj, struct cdev, kobj);
4203 diff -NurpP --minimal linux-3.0/fs/dcache.c linux-3.0-vs2.3.1-pre7/fs/dcache.c
4204 --- linux-3.0/fs/dcache.c       2011-07-22 11:18:05.000000000 +0200
4205 +++ linux-3.0-vs2.3.1-pre7/fs/dcache.c  2011-07-22 11:20:39.000000000 +0200
4206 @@ -36,6 +36,7 @@
4207  #include <linux/bit_spinlock.h>
4208  #include <linux/rculist_bl.h>
4209  #include <linux/prefetch.h>
4210 +#include <linux/vs_limit.h>
4211  #include "internal.h"
4212
4213  /*
4214 @@ -479,6 +480,8 @@ int d_invalidate(struct dentry * dentry)
4215                 spin_lock(&dentry->d_lock);
4216         }
4217
4218 +       vx_dentry_dec(dentry);
4219 +
4220         /*
4221          * Somebody else still using it?
4222          *
4223 @@ -506,6 +509,7 @@ EXPORT_SYMBOL(d_invalidate);
4224  static inline void __dget_dlock(struct dentry *dentry)
4225  {
4226         dentry->d_count++;
4227 +       vx_dentry_inc(dentry);
4228  }
4229
4230  static inline void __dget(struct dentry *dentry)
4231 @@ -1266,6 +1270,9 @@ struct dentry *d_alloc(struct dentry * p
4232         struct dentry *dentry;
4233         char *dname;
4234
4235 +       if (!vx_dentry_avail(1))
4236 +               return NULL;
4237 +
4238         dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL);
4239         if (!dentry)
4240                 return NULL;
4241 @@ -1288,6 +1295,7 @@ struct dentry *d_alloc(struct dentry * p
4242
4243         dentry->d_count = 1;
4244         dentry->d_flags = 0;
4245 +       vx_dentry_inc(dentry);
4246         spin_lock_init(&dentry->d_lock);
4247         seqcount_init(&dentry->d_seq);
4248         dentry->d_inode = NULL;
4249 @@ -1945,6 +1953,7 @@ struct dentry *__d_lookup(struct dentry
4250                 }
4251
4252                 dentry->d_count++;
4253 +               vx_dentry_inc(dentry);
4254                 found = dentry;
4255                 spin_unlock(&dentry->d_lock);
4256                 break;
4257 diff -NurpP --minimal linux-3.0/fs/devpts/inode.c linux-3.0-vs2.3.1-pre7/fs/devpts/inode.c
4258 --- linux-3.0/fs/devpts/inode.c 2011-05-22 16:17:50.000000000 +0200
4259 +++ linux-3.0-vs2.3.1-pre7/fs/devpts/inode.c    2011-06-10 22:11:24.000000000 +0200
4260 @@ -25,6 +25,7 @@
4261  #include <linux/parser.h>
4262  #include <linux/fsnotify.h>
4263  #include <linux/seq_file.h>
4264 +#include <linux/vs_base.h>
4265
4266  #define DEVPTS_DEFAULT_MODE 0600
4267  /*
4268 @@ -36,6 +37,20 @@
4269  #define DEVPTS_DEFAULT_PTMX_MODE 0000
4270  #define PTMX_MINOR     2
4271
4272 +static int devpts_permission(struct inode *inode, int mask, unsigned int flags)
4273 +{
4274 +       int ret = -EACCES;
4275 +
4276 +       /* devpts is xid tagged */
4277 +       if (vx_check((xid_t)inode->i_tag, VS_WATCH_P | VS_IDENT))
4278 +               ret = generic_permission(inode, mask, flags, NULL);
4279 +       return ret;
4280 +}
4281 +
4282 +static struct inode_operations devpts_file_inode_operations = {
4283 +       .permission     = devpts_permission,
4284 +};
4285 +
4286  extern int pty_limit;                  /* Config limit on Unix98 ptys */
4287  static DEFINE_MUTEX(allocated_ptys_lock);
4288
4289 @@ -263,6 +278,34 @@ static int devpts_show_options(struct se
4290         return 0;
4291  }
4292
4293 +static int devpts_filter(struct dentry *de)
4294 +{
4295 +       xid_t xid = 0;
4296 +
4297 +       /* devpts is xid tagged */
4298 +       if (de && de->d_inode)
4299 +               xid = (xid_t)de->d_inode->i_tag;
4300 +#ifdef CONFIG_VSERVER_WARN_DEVPTS
4301 +       else
4302 +               vxwprintk_task(1, "devpts " VS_Q("%.*s") " without inode.",
4303 +                       de->d_name.len, de->d_name.name);
4304 +#endif
4305 +       return vx_check(xid, VS_WATCH_P | VS_IDENT);
4306 +}
4307 +
4308 +static int devpts_readdir(struct file * filp, void * dirent, filldir_t filldir)
4309 +{
4310 +       return dcache_readdir_filter(filp, dirent, filldir, devpts_filter);
4311 +}
4312 +
4313 +static struct file_operations devpts_dir_operations = {
4314 +       .open           = dcache_dir_open,
4315 +       .release        = dcache_dir_close,
4316 +       .llseek         = dcache_dir_lseek,
4317 +       .read           = generic_read_dir,
4318 +       .readdir        = devpts_readdir,
4319 +};
4320 +
4321  static const struct super_operations devpts_sops = {
4322         .statfs         = simple_statfs,
4323         .remount_fs     = devpts_remount,
4324 @@ -302,12 +345,15 @@ devpts_fill_super(struct super_block *s,
4325         inode = new_inode(s);
4326         if (!inode)
4327                 goto free_fsi;
4328 +
4329         inode->i_ino = 1;
4330         inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
4331         inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
4332         inode->i_op = &simple_dir_inode_operations;
4333 -       inode->i_fop = &simple_dir_operations;
4334 +       inode->i_fop = &devpts_dir_operations;
4335         inode->i_nlink = 2;
4336 +       /* devpts is xid tagged */
4337 +       inode->i_tag = (tag_t)vx_current_xid();
4338
4339         s->s_root = d_alloc_root(inode);
4340         if (s->s_root)
4341 @@ -494,6 +540,9 @@ int devpts_pty_new(struct inode *ptmx_in
4342         inode->i_gid = opts->setgid ? opts->gid : current_fsgid();
4343         inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
4344         init_special_inode(inode, S_IFCHR|opts->mode, device);
4345 +       /* devpts is xid tagged */
4346 +       inode->i_tag = (tag_t)vx_current_xid();
4347 +       inode->i_op = &devpts_file_inode_operations;
4348         inode->i_private = tty;
4349         tty->driver_data = inode;
4350
4351 diff -NurpP --minimal linux-3.0/fs/ext2/balloc.c linux-3.0-vs2.3.1-pre7/fs/ext2/balloc.c
4352 --- linux-3.0/fs/ext2/balloc.c  2011-05-22 16:17:51.000000000 +0200
4353 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/balloc.c     2011-06-10 22:11:24.000000000 +0200
4354 @@ -701,7 +701,6 @@ ext2_try_to_allocate(struct super_block
4355                         start = 0;
4356                 end = EXT2_BLOCKS_PER_GROUP(sb);
4357         }
4358 -
4359         BUG_ON(start > EXT2_BLOCKS_PER_GROUP(sb));
4360
4361  repeat:
4362 diff -NurpP --minimal linux-3.0/fs/ext2/ext2.h linux-3.0-vs2.3.1-pre7/fs/ext2/ext2.h
4363 --- linux-3.0/fs/ext2/ext2.h    2011-05-22 16:17:51.000000000 +0200
4364 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/ext2.h       2011-06-10 22:11:24.000000000 +0200
4365 @@ -126,6 +126,7 @@ extern void ext2_set_inode_flags(struct
4366  extern void ext2_get_inode_flags(struct ext2_inode_info *);
4367  extern int ext2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
4368                        u64 start, u64 len);
4369 +extern int ext2_sync_flags(struct inode *, int, int);
4370
4371  /* ioctl.c */
4372  extern long ext2_ioctl(struct file *, unsigned int, unsigned long);
4373 diff -NurpP --minimal linux-3.0/fs/ext2/file.c linux-3.0-vs2.3.1-pre7/fs/ext2/file.c
4374 --- linux-3.0/fs/ext2/file.c    2010-08-02 16:52:48.000000000 +0200
4375 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/file.c       2011-06-10 22:11:24.000000000 +0200
4376 @@ -104,4 +104,5 @@ const struct inode_operations ext2_file_
4377         .setattr        = ext2_setattr,
4378         .check_acl      = ext2_check_acl,
4379         .fiemap         = ext2_fiemap,
4380 +       .sync_flags     = ext2_sync_flags,
4381  };
4382 diff -NurpP --minimal linux-3.0/fs/ext2/ialloc.c linux-3.0-vs2.3.1-pre7/fs/ext2/ialloc.c
4383 --- linux-3.0/fs/ext2/ialloc.c  2011-05-22 16:17:51.000000000 +0200
4384 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/ialloc.c     2011-06-10 22:11:24.000000000 +0200
4385 @@ -17,6 +17,7 @@
4386  #include <linux/backing-dev.h>
4387  #include <linux/buffer_head.h>
4388  #include <linux/random.h>
4389 +#include <linux/vs_tag.h>
4390  #include "ext2.h"
4391  #include "xattr.h"
4392  #include "acl.h"
4393 @@ -549,6 +550,7 @@ got:
4394                 inode->i_mode = mode;
4395                 inode->i_uid = current_fsuid();
4396                 inode->i_gid = dir->i_gid;
4397 +               inode->i_tag = dx_current_fstag(sb);
4398         } else
4399                 inode_init_owner(inode, dir, mode);
4400
4401 diff -NurpP --minimal linux-3.0/fs/ext2/inode.c linux-3.0-vs2.3.1-pre7/fs/ext2/inode.c
4402 --- linux-3.0/fs/ext2/inode.c   2011-05-22 16:17:51.000000000 +0200
4403 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/inode.c      2011-06-10 22:11:24.000000000 +0200
4404 @@ -32,6 +32,7 @@
4405  #include <linux/mpage.h>
4406  #include <linux/fiemap.h>
4407  #include <linux/namei.h>
4408 +#include <linux/vs_tag.h>
4409  #include "ext2.h"
4410  #include "acl.h"
4411  #include "xip.h"
4412 @@ -1167,7 +1168,7 @@ static void ext2_truncate_blocks(struct
4413                 return;
4414         if (ext2_inode_is_fast_symlink(inode))
4415                 return;
4416 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
4417 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
4418                 return;
4419         __ext2_truncate_blocks(inode, offset);
4420  }
4421 @@ -1256,36 +1257,61 @@ void ext2_set_inode_flags(struct inode *
4422  {
4423         unsigned int flags = EXT2_I(inode)->i_flags;
4424
4425 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
4426 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4427 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4428 +
4429 +
4430 +       if (flags & EXT2_IMMUTABLE_FL)
4431 +               inode->i_flags |= S_IMMUTABLE;
4432 +       if (flags & EXT2_IXUNLINK_FL)
4433 +               inode->i_flags |= S_IXUNLINK;
4434 +
4435         if (flags & EXT2_SYNC_FL)
4436                 inode->i_flags |= S_SYNC;
4437         if (flags & EXT2_APPEND_FL)
4438                 inode->i_flags |= S_APPEND;
4439 -       if (flags & EXT2_IMMUTABLE_FL)
4440 -               inode->i_flags |= S_IMMUTABLE;
4441         if (flags & EXT2_NOATIME_FL)
4442                 inode->i_flags |= S_NOATIME;
4443         if (flags & EXT2_DIRSYNC_FL)
4444                 inode->i_flags |= S_DIRSYNC;
4445 +
4446 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
4447 +
4448 +       if (flags & EXT2_BARRIER_FL)
4449 +               inode->i_vflags |= V_BARRIER;
4450 +       if (flags & EXT2_COW_FL)
4451 +               inode->i_vflags |= V_COW;
4452  }
4453
4454  /* Propagate flags from i_flags to EXT2_I(inode)->i_flags */
4455  void ext2_get_inode_flags(struct ext2_inode_info *ei)
4456  {
4457         unsigned int flags = ei->vfs_inode.i_flags;
4458 +       unsigned int vflags = ei->vfs_inode.i_vflags;
4459 +
4460 +       ei->i_flags &= ~(EXT2_SYNC_FL | EXT2_APPEND_FL |
4461 +                       EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL |
4462 +                       EXT2_NOATIME_FL | EXT2_DIRSYNC_FL |
4463 +                       EXT2_BARRIER_FL | EXT2_COW_FL);
4464 +
4465 +       if (flags & S_IMMUTABLE)
4466 +               ei->i_flags |= EXT2_IMMUTABLE_FL;
4467 +       if (flags & S_IXUNLINK)
4468 +               ei->i_flags |= EXT2_IXUNLINK_FL;
4469
4470 -       ei->i_flags &= ~(EXT2_SYNC_FL|EXT2_APPEND_FL|
4471 -                       EXT2_IMMUTABLE_FL|EXT2_NOATIME_FL|EXT2_DIRSYNC_FL);
4472         if (flags & S_SYNC)
4473                 ei->i_flags |= EXT2_SYNC_FL;
4474         if (flags & S_APPEND)
4475                 ei->i_flags |= EXT2_APPEND_FL;
4476 -       if (flags & S_IMMUTABLE)
4477 -               ei->i_flags |= EXT2_IMMUTABLE_FL;
4478         if (flags & S_NOATIME)
4479                 ei->i_flags |= EXT2_NOATIME_FL;
4480         if (flags & S_DIRSYNC)
4481                 ei->i_flags |= EXT2_DIRSYNC_FL;
4482 +
4483 +       if (vflags & V_BARRIER)
4484 +               ei->i_flags |= EXT2_BARRIER_FL;
4485 +       if (vflags & V_COW)
4486 +               ei->i_flags |= EXT2_COW_FL;
4487  }
4488
4489  struct inode *ext2_iget (struct super_block *sb, unsigned long ino)
4490 @@ -1295,6 +1321,8 @@ struct inode *ext2_iget (struct super_bl
4491         struct ext2_inode *raw_inode;
4492         struct inode *inode;
4493         long ret = -EIO;
4494 +       uid_t uid;
4495 +       gid_t gid;
4496         int n;
4497
4498         inode = iget_locked(sb, ino);
4499 @@ -1313,12 +1341,17 @@ struct inode *ext2_iget (struct super_bl
4500         }
4501
4502         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
4503 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
4504 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
4505 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
4506 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
4507         if (!(test_opt (inode->i_sb, NO_UID32))) {
4508 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
4509 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
4510 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
4511 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
4512         }
4513 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
4514 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
4515 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
4516 +               le16_to_cpu(raw_inode->i_raw_tag));
4517 +
4518         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
4519         inode->i_size = le32_to_cpu(raw_inode->i_size);
4520         inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
4521 @@ -1416,8 +1449,8 @@ static int __ext2_write_inode(struct ino
4522         struct ext2_inode_info *ei = EXT2_I(inode);
4523         struct super_block *sb = inode->i_sb;
4524         ino_t ino = inode->i_ino;
4525 -       uid_t uid = inode->i_uid;
4526 -       gid_t gid = inode->i_gid;
4527 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
4528 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
4529         struct buffer_head * bh;
4530         struct ext2_inode * raw_inode = ext2_get_inode(sb, ino, &bh);
4531         int n;
4532 @@ -1453,6 +1486,9 @@ static int __ext2_write_inode(struct ino
4533                 raw_inode->i_uid_high = 0;
4534                 raw_inode->i_gid_high = 0;
4535         }
4536 +#ifdef CONFIG_TAGGING_INTERN
4537 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
4538 +#endif
4539         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
4540         raw_inode->i_size = cpu_to_le32(inode->i_size);
4541         raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
4542 @@ -1533,7 +1569,8 @@ int ext2_setattr(struct dentry *dentry,
4543         if (is_quota_modification(inode, iattr))
4544                 dquot_initialize(inode);
4545         if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
4546 -           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
4547 +           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
4548 +           (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
4549                 error = dquot_transfer(inode, iattr);
4550                 if (error)
4551                         return error;
4552 diff -NurpP --minimal linux-3.0/fs/ext2/ioctl.c linux-3.0-vs2.3.1-pre7/fs/ext2/ioctl.c
4553 --- linux-3.0/fs/ext2/ioctl.c   2011-05-22 16:17:51.000000000 +0200
4554 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/ioctl.c      2011-06-10 22:11:24.000000000 +0200
4555 @@ -17,6 +17,16 @@
4556  #include <asm/uaccess.h>
4557
4558
4559 +int ext2_sync_flags(struct inode *inode, int flags, int vflags)
4560 +{
4561 +       inode->i_flags = flags;
4562 +       inode->i_vflags = vflags;
4563 +       ext2_get_inode_flags(EXT2_I(inode));
4564 +       inode->i_ctime = CURRENT_TIME_SEC;
4565 +       mark_inode_dirty(inode);
4566 +       return 0;
4567 +}
4568 +
4569  long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
4570  {
4571         struct inode *inode = filp->f_dentry->d_inode;
4572 @@ -51,6 +61,11 @@ long ext2_ioctl(struct file *filp, unsig
4573
4574                 flags = ext2_mask_flags(inode->i_mode, flags);
4575
4576 +               if (IS_BARRIER(inode)) {
4577 +                       vxwprintk_task(1, "messing with the barrier.");
4578 +                       return -EACCES;
4579 +               }
4580 +
4581                 mutex_lock(&inode->i_mutex);
4582                 /* Is it quota file? Do not allow user to mess with it */
4583                 if (IS_NOQUOTA(inode)) {
4584 @@ -66,7 +81,9 @@ long ext2_ioctl(struct file *filp, unsig
4585                  *
4586                  * This test looks nicer. Thanks to Pauline Middelink
4587                  */
4588 -               if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) {
4589 +               if ((oldflags & EXT2_IMMUTABLE_FL) ||
4590 +                       ((flags ^ oldflags) & (EXT2_APPEND_FL |
4591 +                       EXT2_IMMUTABLE_FL | EXT2_IXUNLINK_FL))) {
4592                         if (!capable(CAP_LINUX_IMMUTABLE)) {
4593                                 mutex_unlock(&inode->i_mutex);
4594                                 ret = -EPERM;
4595 @@ -74,7 +91,7 @@ long ext2_ioctl(struct file *filp, unsig
4596                         }
4597                 }
4598
4599 -               flags = flags & EXT2_FL_USER_MODIFIABLE;
4600 +               flags &= EXT2_FL_USER_MODIFIABLE;
4601                 flags |= oldflags & ~EXT2_FL_USER_MODIFIABLE;
4602                 ei->i_flags = flags;
4603                 mutex_unlock(&inode->i_mutex);
4604 diff -NurpP --minimal linux-3.0/fs/ext2/namei.c linux-3.0-vs2.3.1-pre7/fs/ext2/namei.c
4605 --- linux-3.0/fs/ext2/namei.c   2011-05-22 16:17:51.000000000 +0200
4606 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/namei.c      2011-06-10 22:11:24.000000000 +0200
4607 @@ -32,6 +32,7 @@
4608
4609  #include <linux/pagemap.h>
4610  #include <linux/quotaops.h>
4611 +#include <linux/vs_tag.h>
4612  #include "ext2.h"
4613  #include "xattr.h"
4614  #include "acl.h"
4615 @@ -75,6 +76,7 @@ static struct dentry *ext2_lookup(struct
4616                                 return ERR_PTR(-EIO);
4617                         } else {
4618                                 return ERR_CAST(inode);
4619 +               dx_propagate_tag(nd, inode);
4620                         }
4621                 }
4622         }
4623 @@ -413,6 +415,7 @@ const struct inode_operations ext2_dir_i
4624  #endif
4625         .setattr        = ext2_setattr,
4626         .check_acl      = ext2_check_acl,
4627 +       .sync_flags     = ext2_sync_flags,
4628  };
4629
4630  const struct inode_operations ext2_special_inode_operations = {
4631 diff -NurpP --minimal linux-3.0/fs/ext2/super.c linux-3.0-vs2.3.1-pre7/fs/ext2/super.c
4632 --- linux-3.0/fs/ext2/super.c   2011-07-22 11:18:05.000000000 +0200
4633 +++ linux-3.0-vs2.3.1-pre7/fs/ext2/super.c      2011-06-10 22:11:24.000000000 +0200
4634 @@ -394,7 +394,8 @@ enum {
4635         Opt_err_ro, Opt_nouid32, Opt_nocheck, Opt_debug,
4636         Opt_oldalloc, Opt_orlov, Opt_nobh, Opt_user_xattr, Opt_nouser_xattr,
4637         Opt_acl, Opt_noacl, Opt_xip, Opt_ignore, Opt_err, Opt_quota,
4638 -       Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation
4639 +       Opt_usrquota, Opt_grpquota, Opt_reservation, Opt_noreservation,
4640 +       Opt_tag, Opt_notag, Opt_tagid
4641  };
4642
4643  static const match_table_t tokens = {
4644 @@ -422,6 +423,9 @@ static const match_table_t tokens = {
4645         {Opt_acl, "acl"},
4646         {Opt_noacl, "noacl"},
4647         {Opt_xip, "xip"},
4648 +       {Opt_tag, "tag"},
4649 +       {Opt_notag, "notag"},
4650 +       {Opt_tagid, "tagid=%u"},
4651         {Opt_grpquota, "grpquota"},
4652         {Opt_ignore, "noquota"},
4653         {Opt_quota, "quota"},
4654 @@ -492,6 +496,20 @@ static int parse_options(char *options,
4655                 case Opt_nouid32:
4656                         set_opt (sbi->s_mount_opt, NO_UID32);
4657                         break;
4658 +#ifndef CONFIG_TAGGING_NONE
4659 +               case Opt_tag:
4660 +                       set_opt (sbi->s_mount_opt, TAGGED);
4661 +                       break;
4662 +               case Opt_notag:
4663 +                       clear_opt (sbi->s_mount_opt, TAGGED);
4664 +                       break;
4665 +#endif
4666 +#ifdef CONFIG_PROPAGATE
4667 +               case Opt_tagid:
4668 +                       /* use args[0] */
4669 +                       set_opt (sbi->s_mount_opt, TAGGED);
4670 +                       break;
4671 +#endif
4672                 case Opt_nocheck:
4673                         clear_opt (sbi->s_mount_opt, CHECK);
4674                         break;
4675 @@ -850,6 +868,8 @@ static int ext2_fill_super(struct super_
4676         if (!parse_options((char *) data, sb))
4677                 goto failed_mount;
4678
4679 +       if (EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_TAGGED)
4680 +               sb->s_flags |= MS_TAGGED;
4681         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
4682                 ((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ?
4683                  MS_POSIXACL : 0);
4684 @@ -1224,6 +1244,14 @@ static int ext2_remount (struct super_bl
4685                 goto restore_opts;
4686         }
4687
4688 +       if ((sbi->s_mount_opt & EXT2_MOUNT_TAGGED) &&
4689 +               !(sb->s_flags & MS_TAGGED)) {
4690 +               printk("EXT2-fs: %s: tagging not permitted on remount.\n",
4691 +                      sb->s_id);
4692 +               err = -EINVAL;
4693 +               goto restore_opts;
4694 +       }
4695 +
4696         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
4697                 ((sbi->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
4698
4699 diff -NurpP --minimal linux-3.0/fs/ext3/file.c linux-3.0-vs2.3.1-pre7/fs/ext3/file.c
4700 --- linux-3.0/fs/ext3/file.c    2010-07-07 18:31:51.000000000 +0200
4701 +++ linux-3.0-vs2.3.1-pre7/fs/ext3/file.c       2011-06-10 22:11:24.000000000 +0200
4702 @@ -81,5 +81,6 @@ const struct inode_operations ext3_file_
4703  #endif
4704         .check_acl      = ext3_check_acl,
4705         .fiemap         = ext3_fiemap,
4706 +       .sync_flags     = ext3_sync_flags,
4707  };
4708
4709 diff -NurpP --minimal linux-3.0/fs/ext3/ialloc.c linux-3.0-vs2.3.1-pre7/fs/ext3/ialloc.c
4710 --- linux-3.0/fs/ext3/ialloc.c  2011-05-22 16:17:52.000000000 +0200
4711 +++ linux-3.0-vs2.3.1-pre7/fs/ext3/ialloc.c     2011-06-10 22:11:24.000000000 +0200
4712 @@ -23,6 +23,7 @@
4713  #include <linux/buffer_head.h>
4714  #include <linux/random.h>
4715  #include <linux/bitops.h>
4716 +#include <linux/vs_tag.h>
4717
4718  #include <asm/byteorder.h>
4719
4720 @@ -532,6 +533,7 @@ got:
4721                 inode->i_mode = mode;
4722                 inode->i_uid = current_fsuid();
4723                 inode->i_gid = dir->i_gid;
4724 +               inode->i_tag = dx_current_fstag(sb);
4725         } else
4726                 inode_init_owner(inode, dir, mode);
4727
4728 diff -NurpP --minimal linux-3.0/fs/ext3/inode.c linux-3.0-vs2.3.1-pre7/fs/ext3/inode.c
4729 --- linux-3.0/fs/ext3/inode.c   2011-07-22 11:18:05.000000000 +0200
4730 +++ linux-3.0-vs2.3.1-pre7/fs/ext3/inode.c      2011-06-10 22:11:24.000000000 +0200
4731 @@ -38,6 +38,7 @@
4732  #include <linux/bio.h>
4733  #include <linux/fiemap.h>
4734  #include <linux/namei.h>
4735 +#include <linux/vs_tag.h>
4736  #include "xattr.h"
4737  #include "acl.h"
4738
4739 @@ -2391,7 +2392,7 @@ static void ext3_free_branches(handle_t
4740
4741  int ext3_can_truncate(struct inode *inode)
4742  {
4743 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
4744 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
4745                 return 0;
4746         if (S_ISREG(inode->i_mode))
4747                 return 1;
4748 @@ -2775,36 +2776,60 @@ void ext3_set_inode_flags(struct inode *
4749  {
4750         unsigned int flags = EXT3_I(inode)->i_flags;
4751
4752 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
4753 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
4754 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
4755 +
4756 +       if (flags & EXT3_IMMUTABLE_FL)
4757 +               inode->i_flags |= S_IMMUTABLE;
4758 +       if (flags & EXT3_IXUNLINK_FL)
4759 +               inode->i_flags |= S_IXUNLINK;
4760 +
4761         if (flags & EXT3_SYNC_FL)
4762                 inode->i_flags |= S_SYNC;
4763         if (flags & EXT3_APPEND_FL)
4764                 inode->i_flags |= S_APPEND;
4765 -       if (flags & EXT3_IMMUTABLE_FL)
4766 -               inode->i_flags |= S_IMMUTABLE;
4767         if (flags & EXT3_NOATIME_FL)
4768                 inode->i_flags |= S_NOATIME;
4769         if (flags & EXT3_DIRSYNC_FL)
4770                 inode->i_flags |= S_DIRSYNC;
4771 +
4772 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
4773 +
4774 +       if (flags & EXT3_BARRIER_FL)
4775 +               inode->i_vflags |= V_BARRIER;
4776 +       if (flags & EXT3_COW_FL)
4777 +               inode->i_vflags |= V_COW;
4778  }
4779
4780  /* Propagate flags from i_flags to EXT3_I(inode)->i_flags */
4781  void ext3_get_inode_flags(struct ext3_inode_info *ei)
4782  {
4783         unsigned int flags = ei->vfs_inode.i_flags;
4784 +       unsigned int vflags = ei->vfs_inode.i_vflags;
4785 +
4786 +       ei->i_flags &= ~(EXT3_SYNC_FL | EXT3_APPEND_FL |
4787 +                       EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL |
4788 +                       EXT3_NOATIME_FL | EXT3_DIRSYNC_FL |
4789 +                       EXT3_BARRIER_FL | EXT3_COW_FL);
4790 +
4791 +       if (flags & S_IMMUTABLE)
4792 +               ei->i_flags |= EXT3_IMMUTABLE_FL;
4793 +       if (flags & S_IXUNLINK)
4794 +               ei->i_flags |= EXT3_IXUNLINK_FL;
4795
4796 -       ei->i_flags &= ~(EXT3_SYNC_FL|EXT3_APPEND_FL|
4797 -                       EXT3_IMMUTABLE_FL|EXT3_NOATIME_FL|EXT3_DIRSYNC_FL);
4798         if (flags & S_SYNC)
4799                 ei->i_flags |= EXT3_SYNC_FL;
4800         if (flags & S_APPEND)
4801                 ei->i_flags |= EXT3_APPEND_FL;
4802 -       if (flags & S_IMMUTABLE)
4803 -               ei->i_flags |= EXT3_IMMUTABLE_FL;
4804         if (flags & S_NOATIME)
4805                 ei->i_flags |= EXT3_NOATIME_FL;
4806         if (flags & S_DIRSYNC)
4807                 ei->i_flags |= EXT3_DIRSYNC_FL;
4808 +
4809 +       if (vflags & V_BARRIER)
4810 +               ei->i_flags |= EXT3_BARRIER_FL;
4811 +       if (vflags & V_COW)
4812 +               ei->i_flags |= EXT3_COW_FL;
4813  }
4814
4815  struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
4816 @@ -2818,6 +2843,8 @@ struct inode *ext3_iget(struct super_blo
4817         transaction_t *transaction;
4818         long ret;
4819         int block;
4820 +       uid_t uid;
4821 +       gid_t gid;
4822
4823         inode = iget_locked(sb, ino);
4824         if (!inode)
4825 @@ -2834,12 +2861,17 @@ struct inode *ext3_iget(struct super_blo
4826         bh = iloc.bh;
4827         raw_inode = ext3_raw_inode(&iloc);
4828         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
4829 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
4830 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
4831 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
4832 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
4833         if(!(test_opt (inode->i_sb, NO_UID32))) {
4834 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
4835 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
4836 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
4837 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
4838         }
4839 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
4840 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
4841 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
4842 +               le16_to_cpu(raw_inode->i_raw_tag));
4843 +
4844         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
4845         inode->i_size = le32_to_cpu(raw_inode->i_size);
4846         inode->i_atime.tv_sec = (signed)le32_to_cpu(raw_inode->i_atime);
4847 @@ -2994,6 +3026,8 @@ static int ext3_do_update_inode(handle_t
4848         struct ext3_inode *raw_inode = ext3_raw_inode(iloc);
4849         struct ext3_inode_info *ei = EXT3_I(inode);
4850         struct buffer_head *bh = iloc->bh;
4851 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
4852 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
4853         int err = 0, rc, block;
4854
4855  again:
4856 @@ -3008,29 +3042,32 @@ again:
4857         ext3_get_inode_flags(ei);
4858         raw_inode->i_mode = cpu_to_le16(inode->i_mode);
4859         if(!(test_opt(inode->i_sb, NO_UID32))) {
4860 -               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
4861 -               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
4862 +               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
4863 +               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
4864  /*
4865   * Fix up interoperability with old kernels. Otherwise, old inodes get
4866   * re-used with the upper 16 bits of the uid/gid intact
4867   */
4868                 if(!ei->i_dtime) {
4869                         raw_inode->i_uid_high =
4870 -                               cpu_to_le16(high_16_bits(inode->i_uid));
4871 +                               cpu_to_le16(high_16_bits(uid));
4872                         raw_inode->i_gid_high =
4873 -                               cpu_to_le16(high_16_bits(inode->i_gid));
4874 +                               cpu_to_le16(high_16_bits(gid));
4875                 } else {
4876                         raw_inode->i_uid_high = 0;
4877                         raw_inode->i_gid_high = 0;
4878                 }
4879         } else {
4880                 raw_inode->i_uid_low =
4881 -                       cpu_to_le16(fs_high2lowuid(inode->i_uid));
4882 +                       cpu_to_le16(fs_high2lowuid(uid));
4883                 raw_inode->i_gid_low =
4884 -                       cpu_to_le16(fs_high2lowgid(inode->i_gid));
4885 +                       cpu_to_le16(fs_high2lowgid(gid));
4886                 raw_inode->i_uid_high = 0;
4887                 raw_inode->i_gid_high = 0;
4888         }
4889 +#ifdef CONFIG_TAGGING_INTERN
4890 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
4891 +#endif
4892         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
4893         raw_inode->i_size = cpu_to_le32(ei->i_disksize);
4894         raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
4895 @@ -3190,7 +3227,8 @@ int ext3_setattr(struct dentry *dentry,
4896         if (is_quota_modification(inode, attr))
4897                 dquot_initialize(inode);
4898         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
4899 -               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
4900 +               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
4901 +               (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
4902                 handle_t *handle;
4903
4904                 /* (user+group)*(old+new) structure, inode write (sb,
4905 @@ -3212,6 +3250,8 @@ int ext3_setattr(struct dentry *dentry,
4906                         inode->i_uid = attr->ia_uid;
4907                 if (attr->ia_valid & ATTR_GID)
4908                         inode->i_gid = attr->ia_gid;
4909 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
4910 +                       inode->i_tag = attr->ia_tag;
4911                 error = ext3_mark_inode_dirty(handle, inode);
4912                 ext3_journal_stop(handle);
4913         }
4914 diff -NurpP --minimal linux-3.0/fs/ext3/ioctl.c linux-3.0-vs2.3.1-pre7/fs/ext3/ioctl.c
4915 --- linux-3.0/fs/ext3/ioctl.c   2011-05-22 16:17:52.000000000 +0200
4916 +++ linux-3.0-vs2.3.1-pre7/fs/ext3/ioctl.c      2011-06-10 22:11:24.000000000 +0200
4917 @@ -8,6 +8,7 @@
4918   */
4919
4920  #include <linux/fs.h>
4921 +#include <linux/mount.h>
4922  #include <linux/jbd.h>
4923  #include <linux/capability.h>
4924  #include <linux/ext3_fs.h>
4925 @@ -17,6 +18,34 @@
4926  #include <linux/compat.h>
4927  #include <asm/uaccess.h>
4928
4929 +
4930 +int ext3_sync_flags(struct inode *inode, int flags, int vflags)
4931 +{
4932 +       handle_t *handle = NULL;
4933 +       struct ext3_iloc iloc;
4934 +       int err;
4935 +
4936 +       handle = ext3_journal_start(inode, 1);
4937 +       if (IS_ERR(handle))
4938 +               return PTR_ERR(handle);
4939 +
4940 +       if (IS_SYNC(inode))
4941 +               handle->h_sync = 1;
4942 +       err = ext3_reserve_inode_write(handle, inode, &iloc);
4943 +       if (err)
4944 +               goto flags_err;
4945 +
4946 +       inode->i_flags = flags;
4947 +       inode->i_vflags = vflags;
4948 +       ext3_get_inode_flags(EXT3_I(inode));
4949 +       inode->i_ctime = CURRENT_TIME_SEC;
4950 +
4951 +       err = ext3_mark_iloc_dirty(handle, inode, &iloc);
4952 +flags_err:
4953 +       ext3_journal_stop(handle);
4954 +       return err;
4955 +}
4956 +
4957  long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
4958  {
4959         struct inode *inode = filp->f_dentry->d_inode;
4960 @@ -50,6 +79,11 @@ long ext3_ioctl(struct file *filp, unsig
4961
4962                 flags = ext3_mask_flags(inode->i_mode, flags);
4963
4964 +               if (IS_BARRIER(inode)) {
4965 +                       vxwprintk_task(1, "messing with the barrier.");
4966 +                       return -EACCES;
4967 +               }
4968 +
4969                 mutex_lock(&inode->i_mutex);
4970
4971                 /* Is it quota file? Do not allow user to mess with it */
4972 @@ -68,7 +102,9 @@ long ext3_ioctl(struct file *filp, unsig
4973                  *
4974                  * This test looks nicer. Thanks to Pauline Middelink
4975                  */
4976 -               if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) {
4977 +               if ((oldflags & EXT3_IMMUTABLE_FL) ||
4978 +                       ((flags ^ oldflags) & (EXT3_APPEND_FL |
4979 +                       EXT3_IMMUTABLE_FL | EXT3_IXUNLINK_FL))) {
4980                         if (!capable(CAP_LINUX_IMMUTABLE))
4981                                 goto flags_out;
4982                 }
4983 @@ -93,7 +129,7 @@ long ext3_ioctl(struct file *filp, unsig
4984                 if (err)
4985                         goto flags_err;
4986
4987 -               flags = flags & EXT3_FL_USER_MODIFIABLE;
4988 +               flags &= EXT3_FL_USER_MODIFIABLE;
4989                 flags |= oldflags & ~EXT3_FL_USER_MODIFIABLE;
4990                 ei->i_flags = flags;
4991
4992 diff -NurpP --minimal linux-3.0/fs/ext3/namei.c linux-3.0-vs2.3.1-pre7/fs/ext3/namei.c
4993 --- linux-3.0/fs/ext3/namei.c   2011-07-22 11:18:05.000000000 +0200
4994 +++ linux-3.0-vs2.3.1-pre7/fs/ext3/namei.c      2011-06-10 22:11:24.000000000 +0200
4995 @@ -36,6 +36,7 @@
4996  #include <linux/quotaops.h>
4997  #include <linux/buffer_head.h>
4998  #include <linux/bio.h>
4999 +#include <linux/vs_tag.h>
5000
5001  #include "namei.h"
5002  #include "xattr.h"
5003 @@ -923,6 +924,7 @@ restart:
5004                                 if (bh)
5005                                         ll_rw_block(READ_META, 1, &bh);
5006                         }
5007 +               dx_propagate_tag(nd, inode);
5008                 }
5009                 if ((bh = bh_use[ra_ptr++]) == NULL)
5010                         goto next;
5011 @@ -2534,6 +2536,7 @@ const struct inode_operations ext3_dir_i
5012         .removexattr    = generic_removexattr,
5013  #endif
5014         .check_acl      = ext3_check_acl,
5015 +       .sync_flags     = ext3_sync_flags,
5016  };
5017
5018  const struct inode_operations ext3_special_inode_operations = {
5019 diff -NurpP --minimal linux-3.0/fs/ext3/super.c linux-3.0-vs2.3.1-pre7/fs/ext3/super.c
5020 --- linux-3.0/fs/ext3/super.c   2011-07-22 11:18:05.000000000 +0200
5021 +++ linux-3.0-vs2.3.1-pre7/fs/ext3/super.c      2011-06-10 22:11:24.000000000 +0200
5022 @@ -821,7 +821,8 @@ enum {
5023         Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
5024         Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
5025         Opt_noquota, Opt_ignore, Opt_barrier, Opt_nobarrier, Opt_err,
5026 -       Opt_resize, Opt_usrquota, Opt_grpquota
5027 +       Opt_resize, Opt_usrquota, Opt_grpquota,
5028 +       Opt_tag, Opt_notag, Opt_tagid
5029  };
5030
5031  static const match_table_t tokens = {
5032 @@ -878,6 +879,9 @@ static const match_table_t tokens = {
5033         {Opt_barrier, "barrier"},
5034         {Opt_nobarrier, "nobarrier"},
5035         {Opt_resize, "resize"},
5036 +       {Opt_tag, "tag"},
5037 +       {Opt_notag, "notag"},
5038 +       {Opt_tagid, "tagid=%u"},
5039         {Opt_err, NULL},
5040  };
5041
5042 @@ -1030,6 +1034,20 @@ static int parse_options (char *options,
5043                 case Opt_nouid32:
5044                         set_opt (sbi->s_mount_opt, NO_UID32);
5045                         break;
5046 +#ifndef CONFIG_TAGGING_NONE
5047 +               case Opt_tag:
5048 +                       set_opt (sbi->s_mount_opt, TAGGED);
5049 +                       break;
5050 +               case Opt_notag:
5051 +                       clear_opt (sbi->s_mount_opt, TAGGED);
5052 +                       break;
5053 +#endif
5054 +#ifdef CONFIG_PROPAGATE
5055 +               case Opt_tagid:
5056 +                       /* use args[0] */
5057 +                       set_opt (sbi->s_mount_opt, TAGGED);
5058 +                       break;
5059 +#endif
5060                 case Opt_nocheck:
5061                         clear_opt (sbi->s_mount_opt, CHECK);
5062                         break;
5063 @@ -1724,6 +1742,9 @@ static int ext3_fill_super (struct super
5064                             NULL, 0))
5065                 goto failed_mount;
5066
5067 +       if (EXT3_SB(sb)->s_mount_opt & EXT3_MOUNT_TAGGED)
5068 +               sb->s_flags |= MS_TAGGED;
5069 +
5070         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
5071                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
5072
5073 @@ -2604,6 +2625,14 @@ static int ext3_remount (struct super_bl
5074         if (test_opt(sb, ABORT))
5075                 ext3_abort(sb, __func__, "Abort forced by user");
5076
5077 +       if ((sbi->s_mount_opt & EXT3_MOUNT_TAGGED) &&
5078 +               !(sb->s_flags & MS_TAGGED)) {
5079 +               printk("EXT3-fs: %s: tagging not permitted on remount.\n",
5080 +                       sb->s_id);
5081 +               err = -EINVAL;
5082 +               goto restore_opts;
5083 +       }
5084 +
5085         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
5086                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
5087
5088 diff -NurpP --minimal linux-3.0/fs/ext4/ext4.h linux-3.0-vs2.3.1-pre7/fs/ext4/ext4.h
5089 --- linux-3.0/fs/ext4/ext4.h    2011-07-22 11:18:05.000000000 +0200
5090 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/ext4.h       2011-06-10 22:11:24.000000000 +0200
5091 @@ -350,8 +350,12 @@ struct flex_groups {
5092  #define EXT4_EXTENTS_FL                        0x00080000 /* Inode uses extents */
5093  #define EXT4_EA_INODE_FL               0x00200000 /* Inode used for large EA */
5094  #define EXT4_EOFBLOCKS_FL              0x00400000 /* Blocks allocated beyond EOF */
5095 +#define EXT4_IXUNLINK_FL               0x08000000 /* Immutable invert on unlink */
5096  #define EXT4_RESERVED_FL               0x80000000 /* reserved for ext4 lib */
5097
5098 +#define EXT4_BARRIER_FL                        0x04000000 /* Barrier for chroot() */
5099 +#define EXT4_COW_FL                    0x20000000 /* Copy on Write marker */
5100 +
5101  #define EXT4_FL_USER_VISIBLE           0x004BDFFF /* User visible flags */
5102  #define EXT4_FL_USER_MODIFIABLE                0x004B80FF /* User modifiable flags */
5103
5104 @@ -608,7 +612,8 @@ struct ext4_inode {
5105                         __le16  l_i_file_acl_high;
5106                         __le16  l_i_uid_high;   /* these 2 fields */
5107                         __le16  l_i_gid_high;   /* were reserved2[0] */
5108 -                       __u32   l_i_reserved2;
5109 +                       __le16  l_i_tag;        /* Context Tag */
5110 +                       __u16   l_i_reserved2;
5111                 } linux2;
5112                 struct {
5113                         __le16  h_i_reserved1;  /* Obsoleted fragment number/size which are removed in ext4 */
5114 @@ -726,6 +731,7 @@ do {                                                                               \
5115  #define i_gid_low      i_gid
5116  #define i_uid_high     osd2.linux2.l_i_uid_high
5117  #define i_gid_high     osd2.linux2.l_i_gid_high
5118 +#define i_raw_tag      osd2.linux2.l_i_tag
5119  #define i_reserved2    osd2.linux2.l_i_reserved2
5120
5121  #elif defined(__GNU__)
5122 @@ -902,6 +908,7 @@ struct ext4_inode_info {
5123  #define EXT4_MOUNT_POSIX_ACL           0x08000 /* POSIX Access Control Lists */
5124  #define EXT4_MOUNT_NO_AUTO_DA_ALLOC    0x10000 /* No auto delalloc mapping */
5125  #define EXT4_MOUNT_BARRIER             0x20000 /* Use block barriers */
5126 +#define EXT4_MOUNT_TAGGED              0x40000 /* Enable Context Tags */
5127  #define EXT4_MOUNT_QUOTA               0x80000 /* Some quota option set */
5128  #define EXT4_MOUNT_USRQUOTA            0x100000 /* "old" user quota */
5129  #define EXT4_MOUNT_GRPQUOTA            0x200000 /* "old" group quota */
5130 @@ -2172,6 +2179,7 @@ extern int ext4_map_blocks(handle_t *han
5131                            struct ext4_map_blocks *map, int flags);
5132  extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
5133                         __u64 start, __u64 len);
5134 +extern int ext4_sync_flags(struct inode *, int, int);
5135  /* move_extent.c */
5136  extern int ext4_move_extents(struct file *o_filp, struct file *d_filp,
5137                              __u64 start_orig, __u64 start_donor,
5138 diff -NurpP --minimal linux-3.0/fs/ext4/file.c linux-3.0-vs2.3.1-pre7/fs/ext4/file.c
5139 --- linux-3.0/fs/ext4/file.c    2011-07-22 11:18:05.000000000 +0200
5140 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/file.c       2011-06-10 22:11:24.000000000 +0200
5141 @@ -282,5 +282,6 @@ const struct inode_operations ext4_file_
5142  #endif
5143         .check_acl      = ext4_check_acl,
5144         .fiemap         = ext4_fiemap,
5145 +       .sync_flags     = ext4_sync_flags,
5146  };
5147
5148 diff -NurpP --minimal linux-3.0/fs/ext4/ialloc.c linux-3.0-vs2.3.1-pre7/fs/ext4/ialloc.c
5149 --- linux-3.0/fs/ext4/ialloc.c  2011-05-22 16:17:52.000000000 +0200
5150 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/ialloc.c     2011-06-10 22:11:24.000000000 +0200
5151 @@ -22,6 +22,7 @@
5152  #include <linux/random.h>
5153  #include <linux/bitops.h>
5154  #include <linux/blkdev.h>
5155 +#include <linux/vs_tag.h>
5156  #include <asm/byteorder.h>
5157
5158  #include "ext4.h"
5159 @@ -992,6 +993,7 @@ got:
5160                 inode->i_mode = mode;
5161                 inode->i_uid = current_fsuid();
5162                 inode->i_gid = dir->i_gid;
5163 +               inode->i_tag = dx_current_fstag(sb);
5164         } else
5165                 inode_init_owner(inode, dir, mode);
5166
5167 diff -NurpP --minimal linux-3.0/fs/ext4/inode.c linux-3.0-vs2.3.1-pre7/fs/ext4/inode.c
5168 --- linux-3.0/fs/ext4/inode.c   2011-07-22 11:18:05.000000000 +0200
5169 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/inode.c      2011-07-01 11:35:34.000000000 +0200
5170 @@ -42,6 +42,7 @@
5171  #include <linux/printk.h>
5172  #include <linux/slab.h>
5173  #include <linux/ratelimit.h>
5174 +#include <linux/vs_tag.h>
5175
5176  #include "ext4_jbd2.h"
5177  #include "xattr.h"
5178 @@ -4797,41 +4798,64 @@ void ext4_set_inode_flags(struct inode *
5179  {
5180         unsigned int flags = EXT4_I(inode)->i_flags;
5181
5182 -       inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
5183 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
5184 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
5185 +
5186 +       if (flags & EXT4_IMMUTABLE_FL)
5187 +               inode->i_flags |= S_IMMUTABLE;
5188 +       if (flags & EXT4_IXUNLINK_FL)
5189 +               inode->i_flags |= S_IXUNLINK;
5190 +
5191         if (flags & EXT4_SYNC_FL)
5192                 inode->i_flags |= S_SYNC;
5193         if (flags & EXT4_APPEND_FL)
5194                 inode->i_flags |= S_APPEND;
5195 -       if (flags & EXT4_IMMUTABLE_FL)
5196 -               inode->i_flags |= S_IMMUTABLE;
5197         if (flags & EXT4_NOATIME_FL)
5198                 inode->i_flags |= S_NOATIME;
5199         if (flags & EXT4_DIRSYNC_FL)
5200                 inode->i_flags |= S_DIRSYNC;
5201 +
5202 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
5203 +
5204 +       if (flags & EXT4_BARRIER_FL)
5205 +               inode->i_vflags |= V_BARRIER;
5206 +       if (flags & EXT4_COW_FL)
5207 +               inode->i_vflags |= V_COW;
5208  }
5209
5210  /* Propagate flags from i_flags to EXT4_I(inode)->i_flags */
5211  void ext4_get_inode_flags(struct ext4_inode_info *ei)
5212  {
5213 -       unsigned int vfs_fl;
5214 +       unsigned int vfs_fl, vfs_vf;
5215         unsigned long old_fl, new_fl;
5216
5217         do {
5218                 vfs_fl = ei->vfs_inode.i_flags;
5219 +               vfs_vf = ei->vfs_inode.i_vflags;
5220                 old_fl = ei->i_flags;
5221                 new_fl = old_fl & ~(EXT4_SYNC_FL|EXT4_APPEND_FL|
5222                                 EXT4_IMMUTABLE_FL|EXT4_NOATIME_FL|
5223 -                               EXT4_DIRSYNC_FL);
5224 +                               EXT4_DIRSYNC_FL|EXT4_BARRIER_FL|
5225 +                               EXT4_COW_FL);
5226 +
5227 +               if (vfs_fl & S_IMMUTABLE)
5228 +                       new_fl |= EXT4_IMMUTABLE_FL;
5229 +               if (vfs_fl & S_IXUNLINK)
5230 +                       new_fl |= EXT4_IXUNLINK_FL;
5231 +
5232                 if (vfs_fl & S_SYNC)
5233                         new_fl |= EXT4_SYNC_FL;
5234                 if (vfs_fl & S_APPEND)
5235                         new_fl |= EXT4_APPEND_FL;
5236 -               if (vfs_fl & S_IMMUTABLE)
5237 -                       new_fl |= EXT4_IMMUTABLE_FL;
5238                 if (vfs_fl & S_NOATIME)
5239                         new_fl |= EXT4_NOATIME_FL;
5240                 if (vfs_fl & S_DIRSYNC)
5241                         new_fl |= EXT4_DIRSYNC_FL;
5242 +
5243 +               if (vfs_vf & V_BARRIER)
5244 +                       new_fl |= EXT4_BARRIER_FL;
5245 +               if (vfs_vf & V_COW)
5246 +                       new_fl |= EXT4_COW_FL;
5247         } while (cmpxchg(&ei->i_flags, old_fl, new_fl) != old_fl);
5248  }
5249
5250 @@ -4867,6 +4891,8 @@ struct inode *ext4_iget(struct super_blo
5251         journal_t *journal = EXT4_SB(sb)->s_journal;
5252         long ret;
5253         int block;
5254 +       uid_t uid;
5255 +       gid_t gid;
5256
5257         inode = iget_locked(sb, ino);
5258         if (!inode)
5259 @@ -4882,12 +4908,16 @@ struct inode *ext4_iget(struct super_blo
5260                 goto bad_inode;
5261         raw_inode = ext4_raw_inode(&iloc);
5262         inode->i_mode = le16_to_cpu(raw_inode->i_mode);
5263 -       inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
5264 -       inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
5265 +       uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
5266 +       gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
5267         if (!(test_opt(inode->i_sb, NO_UID32))) {
5268 -               inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
5269 -               inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
5270 +               uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
5271 +               gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
5272         }
5273 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
5274 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
5275 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
5276 +               le16_to_cpu(raw_inode->i_raw_tag));
5277         inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
5278
5279         ext4_clear_state_flags(ei);     /* Only relevant on 32-bit archs */
5280 @@ -5106,6 +5136,8 @@ static int ext4_do_update_inode(handle_t
5281         struct ext4_inode *raw_inode = ext4_raw_inode(iloc);
5282         struct ext4_inode_info *ei = EXT4_I(inode);
5283         struct buffer_head *bh = iloc->bh;
5284 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
5285 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
5286         int err = 0, rc, block;
5287
5288         /* For fields not not tracking in the in-memory inode,
5289 @@ -5116,29 +5148,32 @@ static int ext4_do_update_inode(handle_t
5290         ext4_get_inode_flags(ei);
5291         raw_inode->i_mode = cpu_to_le16(inode->i_mode);
5292         if (!(test_opt(inode->i_sb, NO_UID32))) {
5293 -               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid));
5294 -               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid));
5295 +               raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid));
5296 +               raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid));
5297  /*
5298   * Fix up interoperability with old kernels. Otherwise, old inodes get
5299   * re-used with the upper 16 bits of the uid/gid intact
5300   */
5301                 if (!ei->i_dtime) {
5302                         raw_inode->i_uid_high =
5303 -                               cpu_to_le16(high_16_bits(inode->i_uid));
5304 +                               cpu_to_le16(high_16_bits(uid));
5305                         raw_inode->i_gid_high =
5306 -                               cpu_to_le16(high_16_bits(inode->i_gid));
5307 +                               cpu_to_le16(high_16_bits(gid));
5308                 } else {
5309                         raw_inode->i_uid_high = 0;
5310                         raw_inode->i_gid_high = 0;
5311                 }
5312         } else {
5313                 raw_inode->i_uid_low =
5314 -                       cpu_to_le16(fs_high2lowuid(inode->i_uid));
5315 +                       cpu_to_le16(fs_high2lowuid(uid));
5316                 raw_inode->i_gid_low =
5317 -                       cpu_to_le16(fs_high2lowgid(inode->i_gid));
5318 +                       cpu_to_le16(fs_high2lowgid(gid));
5319                 raw_inode->i_uid_high = 0;
5320                 raw_inode->i_gid_high = 0;
5321         }
5322 +#ifdef CONFIG_TAGGING_INTERN
5323 +       raw_inode->i_raw_tag = cpu_to_le16(inode->i_tag);
5324 +#endif
5325         raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
5326
5327         EXT4_INODE_SET_XTIME(i_ctime, inode, raw_inode);
5328 @@ -5324,7 +5359,8 @@ int ext4_setattr(struct dentry *dentry,
5329         if (is_quota_modification(inode, attr))
5330                 dquot_initialize(inode);
5331         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
5332 -               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
5333 +               (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
5334 +               (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
5335                 handle_t *handle;
5336
5337                 /* (user+group)*(old+new) structure, inode write (sb,
5338 @@ -5346,6 +5382,8 @@ int ext4_setattr(struct dentry *dentry,
5339                         inode->i_uid = attr->ia_uid;
5340                 if (attr->ia_valid & ATTR_GID)
5341                         inode->i_gid = attr->ia_gid;
5342 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
5343 +                       inode->i_tag = attr->ia_tag;
5344                 error = ext4_mark_inode_dirty(handle, inode);
5345                 ext4_journal_stop(handle);
5346         }
5347 diff -NurpP --minimal linux-3.0/fs/ext4/ioctl.c linux-3.0-vs2.3.1-pre7/fs/ext4/ioctl.c
5348 --- linux-3.0/fs/ext4/ioctl.c   2011-05-22 16:17:52.000000000 +0200
5349 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/ioctl.c      2011-06-10 22:11:24.000000000 +0200
5350 @@ -14,10 +14,39 @@
5351  #include <linux/compat.h>
5352  #include <linux/mount.h>
5353  #include <linux/file.h>
5354 +#include <linux/vs_tag.h>
5355  #include <asm/uaccess.h>
5356  #include "ext4_jbd2.h"
5357  #include "ext4.h"
5358
5359 +
5360 +int ext4_sync_flags(struct inode *inode, int flags, int vflags)
5361 +{
5362 +       handle_t *handle = NULL;
5363 +       struct ext4_iloc iloc;
5364 +       int err;
5365 +
5366 +       handle = ext4_journal_start(inode, 1);
5367 +       if (IS_ERR(handle))
5368 +               return PTR_ERR(handle);
5369 +
5370 +       if (IS_SYNC(inode))
5371 +               ext4_handle_sync(handle);
5372 +       err = ext4_reserve_inode_write(handle, inode, &iloc);
5373 +       if (err)
5374 +               goto flags_err;
5375 +
5376 +       inode->i_flags = flags;
5377 +       inode->i_vflags = vflags;
5378 +       ext4_get_inode_flags(EXT4_I(inode));
5379 +       inode->i_ctime = ext4_current_time(inode);
5380 +
5381 +       err = ext4_mark_iloc_dirty(handle, inode, &iloc);
5382 +flags_err:
5383 +       ext4_journal_stop(handle);
5384 +       return err;
5385 +}
5386 +
5387  long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
5388  {
5389         struct inode *inode = filp->f_dentry->d_inode;
5390 @@ -50,6 +79,11 @@ long ext4_ioctl(struct file *filp, unsig
5391
5392                 flags = ext4_mask_flags(inode->i_mode, flags);
5393
5394 +               if (IS_BARRIER(inode)) {
5395 +                       vxwprintk_task(1, "messing with the barrier.");
5396 +                       return -EACCES;
5397 +               }
5398 +
5399                 err = -EPERM;
5400                 mutex_lock(&inode->i_mutex);
5401                 /* Is it quota file? Do not allow user to mess with it */
5402 @@ -67,7 +101,9 @@ long ext4_ioctl(struct file *filp, unsig
5403                  *
5404                  * This test looks nicer. Thanks to Pauline Middelink
5405                  */
5406 -               if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) {
5407 +               if ((oldflags & EXT4_IMMUTABLE_FL) ||
5408 +                       ((flags ^ oldflags) & (EXT4_APPEND_FL |
5409 +                       EXT4_IMMUTABLE_FL | EXT4_IXUNLINK_FL))) {
5410                         if (!capable(CAP_LINUX_IMMUTABLE))
5411                                 goto flags_out;
5412                 }
5413 diff -NurpP --minimal linux-3.0/fs/ext4/namei.c linux-3.0-vs2.3.1-pre7/fs/ext4/namei.c
5414 --- linux-3.0/fs/ext4/namei.c   2011-07-22 11:18:05.000000000 +0200
5415 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/namei.c      2011-06-10 22:11:24.000000000 +0200
5416 @@ -34,6 +34,7 @@
5417  #include <linux/quotaops.h>
5418  #include <linux/buffer_head.h>
5419  #include <linux/bio.h>
5420 +#include <linux/vs_tag.h>
5421  #include "ext4.h"
5422  #include "ext4_jbd2.h"
5423
5424 @@ -924,6 +925,7 @@ restart:
5425                                 if (bh)
5426                                         ll_rw_block(READ_META, 1, &bh);
5427                         }
5428 +               dx_propagate_tag(nd, inode);
5429                 }
5430                 if ((bh = bh_use[ra_ptr++]) == NULL)
5431                         goto next;
5432 @@ -2596,6 +2598,7 @@ const struct inode_operations ext4_dir_i
5433  #endif
5434         .check_acl      = ext4_check_acl,
5435         .fiemap         = ext4_fiemap,
5436 +       .sync_flags     = ext4_sync_flags,
5437  };
5438
5439  const struct inode_operations ext4_special_inode_operations = {
5440 diff -NurpP --minimal linux-3.0/fs/ext4/super.c linux-3.0-vs2.3.1-pre7/fs/ext4/super.c
5441 --- linux-3.0/fs/ext4/super.c   2011-07-22 11:18:05.000000000 +0200
5442 +++ linux-3.0-vs2.3.1-pre7/fs/ext4/super.c      2011-07-01 11:35:34.000000000 +0200
5443 @@ -1294,6 +1294,7 @@ enum {
5444         Opt_dioread_nolock, Opt_dioread_lock,
5445         Opt_discard, Opt_nodiscard,
5446         Opt_init_inode_table, Opt_noinit_inode_table,
5447 +       Opt_tag, Opt_notag, Opt_tagid
5448  };
5449
5450  static const match_table_t tokens = {
5451 @@ -1369,6 +1370,9 @@ static const match_table_t tokens = {
5452         {Opt_init_inode_table, "init_itable=%u"},
5453         {Opt_init_inode_table, "init_itable"},
5454         {Opt_noinit_inode_table, "noinit_itable"},
5455 +       {Opt_tag, "tag"},
5456 +       {Opt_notag, "notag"},
5457 +       {Opt_tagid, "tagid=%u"},
5458         {Opt_err, NULL},
5459  };
5460
5461 @@ -1537,6 +1541,20 @@ static int parse_options(char *options,
5462                 case Opt_nouid32:
5463                         set_opt(sb, NO_UID32);
5464                         break;
5465 +#ifndef CONFIG_TAGGING_NONE
5466 +               case Opt_tag:
5467 +                       set_opt(sb, TAGGED);
5468 +                       break;
5469 +               case Opt_notag:
5470 +                       clear_opt(sb, TAGGED);
5471 +                       break;
5472 +#endif
5473 +#ifdef CONFIG_PROPAGATE
5474 +               case Opt_tagid:
5475 +                       /* use args[0] */
5476 +                       set_opt(sb, TAGGED);
5477 +                       break;
5478 +#endif
5479                 case Opt_debug:
5480                         set_opt(sb, DEBUG);
5481                         break;
5482 @@ -3194,6 +3212,9 @@ static int ext4_fill_super(struct super_
5483                            &journal_ioprio, NULL, 0))
5484                 goto failed_mount;
5485
5486 +       if (EXT4_SB(sb)->s_mount_opt & EXT4_MOUNT_TAGGED)
5487 +               sb->s_flags |= MS_TAGGED;
5488 +
5489         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
5490                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
5491
5492 @@ -4324,6 +4345,14 @@ static int ext4_remount(struct super_blo
5493         if (sbi->s_mount_flags & EXT4_MF_FS_ABORTED)
5494                 ext4_abort(sb, "Abort forced by user");
5495
5496 +       if ((sbi->s_mount_opt & EXT4_MOUNT_TAGGED) &&
5497 +               !(sb->s_flags & MS_TAGGED)) {
5498 +               printk("EXT4-fs: %s: tagging not permitted on remount.\n",
5499 +                       sb->s_id);
5500 +               err = -EINVAL;
5501 +               goto restore_opts;
5502 +       }
5503 +
5504         sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
5505                 (test_opt(sb, POSIX_ACL) ? MS_POSIXACL : 0);
5506
5507 diff -NurpP --minimal linux-3.0/fs/fcntl.c linux-3.0-vs2.3.1-pre7/fs/fcntl.c
5508 --- linux-3.0/fs/fcntl.c        2011-05-22 16:17:52.000000000 +0200
5509 +++ linux-3.0-vs2.3.1-pre7/fs/fcntl.c   2011-06-10 22:11:24.000000000 +0200
5510 @@ -20,6 +20,7 @@
5511  #include <linux/signal.h>
5512  #include <linux/rcupdate.h>
5513  #include <linux/pid_namespace.h>
5514 +#include <linux/vs_limit.h>
5515
5516  #include <asm/poll.h>
5517  #include <asm/siginfo.h>
5518 @@ -103,6 +104,8 @@ SYSCALL_DEFINE3(dup3, unsigned int, oldf
5519
5520         if (tofree)
5521                 filp_close(tofree, files);
5522 +       else
5523 +               vx_openfd_inc(newfd);   /* fd was unused */
5524
5525         return newfd;
5526
5527 @@ -447,6 +450,8 @@ SYSCALL_DEFINE3(fcntl, unsigned int, fd,
5528         filp = fget_raw(fd);
5529         if (!filp)
5530                 goto out;
5531 +       if (!vx_files_avail(1))
5532 +               goto out;
5533
5534         if (unlikely(filp->f_mode & FMODE_PATH)) {
5535                 if (!check_fcntl_cmd(cmd)) {
5536 diff -NurpP --minimal linux-3.0/fs/file.c linux-3.0-vs2.3.1-pre7/fs/file.c
5537 --- linux-3.0/fs/file.c 2011-05-22 16:17:52.000000000 +0200
5538 +++ linux-3.0-vs2.3.1-pre7/fs/file.c    2011-06-10 22:11:24.000000000 +0200
5539 @@ -21,6 +21,7 @@
5540  #include <linux/spinlock.h>
5541  #include <linux/rcupdate.h>
5542  #include <linux/workqueue.h>
5543 +#include <linux/vs_limit.h>
5544
5545  struct fdtable_defer {
5546         spinlock_t lock;
5547 @@ -359,6 +360,8 @@ struct files_struct *dup_fd(struct files
5548                 struct file *f = *old_fds++;
5549                 if (f) {
5550                         get_file(f);
5551 +                       /* TODO: sum it first for check and performance */
5552 +                       vx_openfd_inc(open_files - i);
5553                 } else {
5554                         /*
5555                          * The fd may be claimed in the fd bitmap but not yet
5556 @@ -466,6 +469,7 @@ repeat:
5557         else
5558                 FD_CLR(fd, fdt->close_on_exec);
5559         error = fd;
5560 +       vx_openfd_inc(fd);
5561  #if 1
5562         /* Sanity check */
5563         if (rcu_dereference_raw(fdt->fd[fd]) != NULL) {
5564 diff -NurpP --minimal linux-3.0/fs/file_table.c linux-3.0-vs2.3.1-pre7/fs/file_table.c
5565 --- linux-3.0/fs/file_table.c   2011-05-22 16:17:52.000000000 +0200
5566 +++ linux-3.0-vs2.3.1-pre7/fs/file_table.c      2011-06-10 22:11:24.000000000 +0200
5567 @@ -24,6 +24,8 @@
5568  #include <linux/percpu_counter.h>
5569  #include <linux/percpu.h>
5570  #include <linux/ima.h>
5571 +#include <linux/vs_limit.h>
5572 +#include <linux/vs_context.h>
5573
5574  #include <asm/atomic.h>
5575
5576 @@ -135,6 +137,8 @@ struct file *get_empty_filp(void)
5577         spin_lock_init(&f->f_lock);
5578         eventpoll_init_file(f);
5579         /* f->f_version: 0 */
5580 +       f->f_xid = vx_current_xid();
5581 +       vx_files_inc(f);
5582         return f;
5583
5584  over:
5585 @@ -253,6 +257,8 @@ static void __fput(struct file *file)
5586         }
5587         fops_put(file->f_op);
5588         put_pid(file->f_owner.pid);
5589 +       vx_files_dec(file);
5590 +       file->f_xid = 0;
5591         file_sb_list_del(file);
5592         if ((file->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
5593                 i_readcount_dec(inode);
5594 @@ -383,6 +389,8 @@ void put_filp(struct file *file)
5595  {
5596         if (atomic_long_dec_and_test(&file->f_count)) {
5597                 security_file_free(file);
5598 +               vx_files_dec(file);
5599 +               file->f_xid = 0;
5600                 file_sb_list_del(file);
5601                 file_free(file);
5602         }
5603 diff -NurpP --minimal linux-3.0/fs/fs_struct.c linux-3.0-vs2.3.1-pre7/fs/fs_struct.c
5604 --- linux-3.0/fs/fs_struct.c    2011-03-15 18:07:31.000000000 +0100
5605 +++ linux-3.0-vs2.3.1-pre7/fs/fs_struct.c       2011-06-10 22:11:24.000000000 +0200
5606 @@ -4,6 +4,7 @@
5607  #include <linux/path.h>
5608  #include <linux/slab.h>
5609  #include <linux/fs_struct.h>
5610 +#include <linux/vserver/global.h>
5611  #include "internal.h"
5612
5613  static inline void path_get_longterm(struct path *path)
5614 @@ -96,6 +97,7 @@ void free_fs_struct(struct fs_struct *fs
5615  {
5616         path_put_longterm(&fs->root);
5617         path_put_longterm(&fs->pwd);
5618 +       atomic_dec(&vs_global_fs);
5619         kmem_cache_free(fs_cachep, fs);
5620  }
5621
5622 @@ -135,6 +137,7 @@ struct fs_struct *copy_fs_struct(struct
5623                 fs->pwd = old->pwd;
5624                 path_get_longterm(&fs->pwd);
5625                 spin_unlock(&old->lock);
5626 +               atomic_inc(&vs_global_fs);
5627         }
5628         return fs;
5629  }
5630 diff -NurpP --minimal linux-3.0/fs/gfs2/file.c linux-3.0-vs2.3.1-pre7/fs/gfs2/file.c
5631 --- linux-3.0/fs/gfs2/file.c    2011-07-22 11:18:05.000000000 +0200
5632 +++ linux-3.0-vs2.3.1-pre7/fs/gfs2/file.c       2011-06-10 22:11:24.000000000 +0200
5633 @@ -134,6 +134,9 @@ static const u32 fsflags_to_gfs2[32] = {
5634         [7] = GFS2_DIF_NOATIME,
5635         [12] = GFS2_DIF_EXHASH,
5636         [14] = GFS2_DIF_INHERIT_JDATA,
5637 +       [27] = GFS2_DIF_IXUNLINK,
5638 +       [26] = GFS2_DIF_BARRIER,
5639 +       [29] = GFS2_DIF_COW,
5640  };
5641
5642  static const u32 gfs2_to_fsflags[32] = {
5643 @@ -143,6 +146,9 @@ static const u32 gfs2_to_fsflags[32] = {
5644         [gfs2fl_NoAtime] = FS_NOATIME_FL,
5645         [gfs2fl_ExHash] = FS_INDEX_FL,
5646         [gfs2fl_InheritJdata] = FS_JOURNAL_DATA_FL,
5647 +       [gfs2fl_IXUnlink] = FS_IXUNLINK_FL,
5648 +       [gfs2fl_Barrier] = FS_BARRIER_FL,
5649 +       [gfs2fl_Cow] = FS_COW_FL,
5650  };
5651
5652  static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
5653 @@ -173,10 +179,16 @@ void gfs2_set_inode_flags(struct inode *
5654  {
5655         struct gfs2_inode *ip = GFS2_I(inode);
5656         unsigned int flags = inode->i_flags;
5657 +       unsigned int vflags = inode->i_vflags;
5658 +
5659 +       flags &= ~(S_IMMUTABLE | S_IXUNLINK |
5660 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
5661
5662 -       flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
5663         if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
5664                 flags |= S_IMMUTABLE;
5665 +       if (ip->i_diskflags & GFS2_DIF_IXUNLINK)
5666 +               flags |= S_IXUNLINK;
5667 +
5668         if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
5669                 flags |= S_APPEND;
5670         if (ip->i_diskflags & GFS2_DIF_NOATIME)
5671 @@ -184,6 +196,43 @@ void gfs2_set_inode_flags(struct inode *
5672         if (ip->i_diskflags & GFS2_DIF_SYNC)
5673                 flags |= S_SYNC;
5674         inode->i_flags = flags;
5675 +
5676 +       vflags &= ~(V_BARRIER | V_COW);
5677 +
5678 +       if (ip->i_diskflags & GFS2_DIF_BARRIER)
5679 +               vflags |= V_BARRIER;
5680 +       if (ip->i_diskflags & GFS2_DIF_COW)
5681 +               vflags |= V_COW;
5682 +       inode->i_vflags = vflags;
5683 +}
5684 +
5685 +void gfs2_get_inode_flags(struct inode *inode)
5686 +{
5687 +       struct gfs2_inode *ip = GFS2_I(inode);
5688 +       unsigned int flags = inode->i_flags;
5689 +       unsigned int vflags = inode->i_vflags;
5690 +
5691 +       ip->i_diskflags &= ~(GFS2_DIF_APPENDONLY |
5692 +                       GFS2_DIF_NOATIME | GFS2_DIF_SYNC |
5693 +                       GFS2_DIF_IMMUTABLE | GFS2_DIF_IXUNLINK |
5694 +                       GFS2_DIF_BARRIER | GFS2_DIF_COW);
5695 +
5696 +       if (flags & S_IMMUTABLE)
5697 +               ip->i_diskflags |= GFS2_DIF_IMMUTABLE;
5698 +       if (flags & S_IXUNLINK)
5699 +               ip->i_diskflags |= GFS2_DIF_IXUNLINK;
5700 +
5701 +       if (flags & S_APPEND)
5702 +               ip->i_diskflags |= GFS2_DIF_APPENDONLY;
5703 +       if (flags & S_NOATIME)
5704 +               ip->i_diskflags |= GFS2_DIF_NOATIME;
5705 +       if (flags & S_SYNC)
5706 +               ip->i_diskflags |= GFS2_DIF_SYNC;
5707 +
5708 +       if (vflags & V_BARRIER)
5709 +               ip->i_diskflags |= GFS2_DIF_BARRIER;
5710 +       if (vflags & V_COW)
5711 +               ip->i_diskflags |= GFS2_DIF_COW;
5712  }
5713
5714  /* Flags that can be set by user space */
5715 @@ -295,6 +344,37 @@ static int gfs2_set_flags(struct file *f
5716         return do_gfs2_set_flags(filp, gfsflags, ~GFS2_DIF_JDATA);
5717  }
5718
5719 +int gfs2_sync_flags(struct inode *inode, int flags, int vflags)
5720 +{
5721 +       struct gfs2_inode *ip = GFS2_I(inode);
5722 +       struct gfs2_sbd *sdp = GFS2_SB(inode);
5723 +       struct buffer_head *bh;
5724 +       struct gfs2_holder gh;
5725 +       int error;
5726 +
5727 +       error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
5728 +       if (error)
5729 +               return error;
5730 +       error = gfs2_trans_begin(sdp, RES_DINODE, 0);
5731 +       if (error)
5732 +               goto out;
5733 +       error = gfs2_meta_inode_buffer(ip, &bh);
5734 +       if (error)
5735 +               goto out_trans_end;
5736 +       gfs2_trans_add_bh(ip->i_gl, bh, 1);
5737 +       inode->i_flags = flags;
5738 +       inode->i_vflags = vflags;
5739 +       gfs2_get_inode_flags(inode);
5740 +       gfs2_dinode_out(ip, bh->b_data);
5741 +       brelse(bh);
5742 +       gfs2_set_aops(inode);
5743 +out_trans_end:
5744 +       gfs2_trans_end(sdp);
5745 +out:
5746 +       gfs2_glock_dq_uninit(&gh);
5747 +       return error;
5748 +}
5749 +
5750  static long gfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
5751  {
5752         switch(cmd) {
5753 diff -NurpP --minimal linux-3.0/fs/gfs2/inode.h linux-3.0-vs2.3.1-pre7/fs/gfs2/inode.h
5754 --- linux-3.0/fs/gfs2/inode.h   2011-07-22 11:18:05.000000000 +0200
5755 +++ linux-3.0-vs2.3.1-pre7/fs/gfs2/inode.h      2011-06-10 22:11:24.000000000 +0200
5756 @@ -120,6 +120,7 @@ extern const struct file_operations gfs2
5757  extern const struct file_operations gfs2_dir_fops_nolock;
5758
5759  extern void gfs2_set_inode_flags(struct inode *inode);
5760 +extern int gfs2_sync_flags(struct inode *inode, int flags, int vflags);
5761
5762  #ifdef CONFIG_GFS2_FS_LOCKING_DLM
5763  extern const struct file_operations gfs2_file_fops;
5764 diff -NurpP --minimal linux-3.0/fs/inode.c linux-3.0-vs2.3.1-pre7/fs/inode.c
5765 --- linux-3.0/fs/inode.c        2011-07-22 11:18:05.000000000 +0200
5766 +++ linux-3.0-vs2.3.1-pre7/fs/inode.c   2011-07-01 11:35:34.000000000 +0200
5767 @@ -26,6 +26,7 @@
5768  #include <linux/ima.h>
5769  #include <linux/cred.h>
5770  #include <linux/buffer_head.h> /* for inode_has_buffers */
5771 +#include <linux/vs_tag.h>
5772  #include "internal.h"
5773
5774  /*
5775 @@ -146,6 +147,9 @@ int inode_init_always(struct super_block
5776         struct address_space *const mapping = &inode->i_data;
5777
5778         inode->i_sb = sb;
5779 +
5780 +       /* essential because of inode slab reuse */
5781 +       inode->i_tag = 0;
5782         inode->i_blkbits = sb->s_blocksize_bits;
5783         inode->i_flags = 0;
5784         atomic_set(&inode->i_count, 1);
5785 @@ -166,6 +170,7 @@ int inode_init_always(struct super_block
5786         inode->i_bdev = NULL;
5787         inode->i_cdev = NULL;
5788         inode->i_rdev = 0;
5789 +       inode->i_mdev = 0;
5790         inode->dirtied_when = 0;
5791
5792         if (security_inode_alloc(inode))
5793 @@ -404,6 +409,8 @@ void __insert_inode_hash(struct inode *i
5794  }
5795  EXPORT_SYMBOL(__insert_inode_hash);
5796
5797 +EXPORT_SYMBOL_GPL(__iget);
5798 +
5799  /**
5800   *     remove_inode_hash - remove an inode from the hash
5801   *     @inode: inode to unhash
5802 @@ -1643,9 +1650,11 @@ void init_special_inode(struct inode *in
5803         if (S_ISCHR(mode)) {
5804                 inode->i_fop = &def_chr_fops;
5805                 inode->i_rdev = rdev;
5806 +               inode->i_mdev = rdev;
5807         } else if (S_ISBLK(mode)) {
5808                 inode->i_fop = &def_blk_fops;
5809                 inode->i_rdev = rdev;
5810 +               inode->i_mdev = rdev;
5811         } else if (S_ISFIFO(mode))
5812                 inode->i_fop = &def_fifo_fops;
5813         else if (S_ISSOCK(mode))
5814 @@ -1674,6 +1683,7 @@ void inode_init_owner(struct inode *inod
5815         } else
5816                 inode->i_gid = current_fsgid();
5817         inode->i_mode = mode;
5818 +       inode->i_tag = dx_current_fstag(inode->i_sb);
5819  }
5820  EXPORT_SYMBOL(inode_init_owner);
5821
5822 diff -NurpP --minimal linux-3.0/fs/ioctl.c linux-3.0-vs2.3.1-pre7/fs/ioctl.c
5823 --- linux-3.0/fs/ioctl.c        2011-05-22 16:17:52.000000000 +0200
5824 +++ linux-3.0-vs2.3.1-pre7/fs/ioctl.c   2011-06-10 22:11:24.000000000 +0200
5825 @@ -15,6 +15,9 @@
5826  #include <linux/writeback.h>
5827  #include <linux/buffer_head.h>
5828  #include <linux/falloc.h>
5829 +#include <linux/proc_fs.h>
5830 +#include <linux/vserver/inode.h>
5831 +#include <linux/vs_tag.h>
5832
5833  #include <asm/ioctls.h>
5834
5835 diff -NurpP --minimal linux-3.0/fs/ioprio.c linux-3.0-vs2.3.1-pre7/fs/ioprio.c
5836 --- linux-3.0/fs/ioprio.c       2011-01-05 21:50:24.000000000 +0100
5837 +++ linux-3.0-vs2.3.1-pre7/fs/ioprio.c  2011-06-10 22:11:24.000000000 +0200
5838 @@ -27,6 +27,7 @@
5839  #include <linux/syscalls.h>
5840  #include <linux/security.h>
5841  #include <linux/pid_namespace.h>
5842 +#include <linux/vs_base.h>
5843
5844  int set_task_ioprio(struct task_struct *task, int ioprio)
5845  {
5846 @@ -119,6 +120,8 @@ SYSCALL_DEFINE3(ioprio_set, int, which,
5847                         else
5848                                 pgrp = find_vpid(who);
5849                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
5850 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
5851 +                                       continue;
5852                                 ret = set_task_ioprio(p, ioprio);
5853                                 if (ret)
5854                                         break;
5855 @@ -208,6 +211,8 @@ SYSCALL_DEFINE2(ioprio_get, int, which,
5856                         else
5857                                 pgrp = find_vpid(who);
5858                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
5859 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
5860 +                                       continue;
5861                                 tmpio = get_task_ioprio(p);
5862                                 if (tmpio < 0)
5863                                         continue;
5864 diff -NurpP --minimal linux-3.0/fs/jfs/file.c linux-3.0-vs2.3.1-pre7/fs/jfs/file.c
5865 --- linux-3.0/fs/jfs/file.c     2011-07-22 11:18:05.000000000 +0200
5866 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/file.c        2011-07-01 11:35:34.000000000 +0200
5867 @@ -102,7 +102,8 @@ int jfs_setattr(struct dentry *dentry, s
5868         if (is_quota_modification(inode, iattr))
5869                 dquot_initialize(inode);
5870         if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) ||
5871 -           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) {
5872 +           (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) ||
5873 +           (iattr->ia_valid & ATTR_TAG && iattr->ia_tag != inode->i_tag)) {
5874                 rc = dquot_transfer(inode, iattr);
5875                 if (rc)
5876                         return rc;
5877 @@ -133,6 +134,7 @@ const struct inode_operations jfs_file_i
5878  #ifdef CONFIG_JFS_POSIX_ACL
5879         .check_acl      = jfs_check_acl,
5880  #endif
5881 +       .sync_flags     = jfs_sync_flags,
5882  };
5883
5884  const struct file_operations jfs_file_operations = {
5885 diff -NurpP --minimal linux-3.0/fs/jfs/ioctl.c linux-3.0-vs2.3.1-pre7/fs/jfs/ioctl.c
5886 --- linux-3.0/fs/jfs/ioctl.c    2011-05-22 16:17:52.000000000 +0200
5887 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/ioctl.c       2011-06-10 22:11:24.000000000 +0200
5888 @@ -11,6 +11,7 @@
5889  #include <linux/mount.h>
5890  #include <linux/time.h>
5891  #include <linux/sched.h>
5892 +#include <linux/mount.h>
5893  #include <asm/current.h>
5894  #include <asm/uaccess.h>
5895
5896 @@ -52,6 +53,16 @@ static long jfs_map_ext2(unsigned long f
5897  }
5898
5899
5900 +int jfs_sync_flags(struct inode *inode, int flags, int vflags)
5901 +{
5902 +       inode->i_flags = flags;
5903 +       inode->i_vflags = vflags;
5904 +       jfs_get_inode_flags(JFS_IP(inode));
5905 +       inode->i_ctime = CURRENT_TIME_SEC;
5906 +       mark_inode_dirty(inode);
5907 +       return 0;
5908 +}
5909 +
5910  long jfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
5911  {
5912         struct inode *inode = filp->f_dentry->d_inode;
5913 @@ -85,6 +96,11 @@ long jfs_ioctl(struct file *filp, unsign
5914                 if (!S_ISDIR(inode->i_mode))
5915                         flags &= ~JFS_DIRSYNC_FL;
5916
5917 +               if (IS_BARRIER(inode)) {
5918 +                       vxwprintk_task(1, "messing with the barrier.");
5919 +                       return -EACCES;
5920 +               }
5921 +
5922                 /* Is it quota file? Do not allow user to mess with it */
5923                 if (IS_NOQUOTA(inode)) {
5924                         err = -EPERM;
5925 @@ -102,8 +118,8 @@ long jfs_ioctl(struct file *filp, unsign
5926                  * the relevant capability.
5927                  */
5928                 if ((oldflags & JFS_IMMUTABLE_FL) ||
5929 -                       ((flags ^ oldflags) &
5930 -                       (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) {
5931 +                       ((flags ^ oldflags) & (JFS_APPEND_FL |
5932 +                       JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL))) {
5933                         if (!capable(CAP_LINUX_IMMUTABLE)) {
5934                                 mutex_unlock(&inode->i_mutex);
5935                                 err = -EPERM;
5936 @@ -111,7 +127,7 @@ long jfs_ioctl(struct file *filp, unsign
5937                         }
5938                 }
5939
5940 -               flags = flags & JFS_FL_USER_MODIFIABLE;
5941 +               flags &= JFS_FL_USER_MODIFIABLE;
5942                 flags |= oldflags & ~JFS_FL_USER_MODIFIABLE;
5943                 jfs_inode->mode2 = flags;
5944
5945 diff -NurpP --minimal linux-3.0/fs/jfs/jfs_dinode.h linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_dinode.h
5946 --- linux-3.0/fs/jfs/jfs_dinode.h       2008-12-25 00:26:37.000000000 +0100
5947 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_dinode.h  2011-06-10 22:11:24.000000000 +0200
5948 @@ -161,9 +161,13 @@ struct dinode {
5949
5950  #define JFS_APPEND_FL          0x01000000 /* writes to file may only append */
5951  #define JFS_IMMUTABLE_FL       0x02000000 /* Immutable file */
5952 +#define JFS_IXUNLINK_FL                0x08000000 /* Immutable invert on unlink */
5953
5954 -#define JFS_FL_USER_VISIBLE    0x03F80000
5955 -#define JFS_FL_USER_MODIFIABLE 0x03F80000
5956 +#define JFS_BARRIER_FL         0x04000000 /* Barrier for chroot() */
5957 +#define JFS_COW_FL             0x20000000 /* Copy on Write marker */
5958 +
5959 +#define JFS_FL_USER_VISIBLE    0x07F80000
5960 +#define JFS_FL_USER_MODIFIABLE 0x07F80000
5961  #define JFS_FL_INHERIT         0x03C80000
5962
5963  /* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */
5964 diff -NurpP --minimal linux-3.0/fs/jfs/jfs_filsys.h linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_filsys.h
5965 --- linux-3.0/fs/jfs/jfs_filsys.h       2008-12-25 00:26:37.000000000 +0100
5966 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_filsys.h  2011-06-10 22:11:24.000000000 +0200
5967 @@ -263,6 +263,7 @@
5968  #define JFS_NAME_MAX   255
5969  #define JFS_PATH_MAX   BPSIZE
5970
5971 +#define JFS_TAGGED             0x00800000      /* Context Tagging */
5972
5973  /*
5974   *     file system state (superblock state)
5975 diff -NurpP --minimal linux-3.0/fs/jfs/jfs_imap.c linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_imap.c
5976 --- linux-3.0/fs/jfs/jfs_imap.c 2011-07-22 11:18:05.000000000 +0200
5977 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_imap.c    2011-07-01 11:35:34.000000000 +0200
5978 @@ -46,6 +46,7 @@
5979  #include <linux/pagemap.h>
5980  #include <linux/quotaops.h>
5981  #include <linux/slab.h>
5982 +#include <linux/vs_tag.h>
5983
5984  #include "jfs_incore.h"
5985  #include "jfs_inode.h"
5986 @@ -3058,6 +3059,8 @@ static int copy_from_dinode(struct dinod
5987  {
5988         struct jfs_inode_info *jfs_ip = JFS_IP(ip);
5989         struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
5990 +       uid_t uid;
5991 +       gid_t gid;
5992
5993         jfs_ip->fileset = le32_to_cpu(dip->di_fileset);
5994         jfs_ip->mode2 = le32_to_cpu(dip->di_mode);
5995 @@ -3078,14 +3081,18 @@ static int copy_from_dinode(struct dinod
5996         }
5997         ip->i_nlink = le32_to_cpu(dip->di_nlink);
5998
5999 -       jfs_ip->saved_uid = le32_to_cpu(dip->di_uid);
6000 +       uid = le32_to_cpu(dip->di_uid);
6001 +       gid = le32_to_cpu(dip->di_gid);
6002 +       ip->i_tag = INOTAG_TAG(DX_TAG(ip), uid, gid, 0);
6003 +
6004 +       jfs_ip->saved_uid = INOTAG_UID(DX_TAG(ip), uid, gid);
6005         if (sbi->uid == -1)
6006                 ip->i_uid = jfs_ip->saved_uid;
6007         else {
6008                 ip->i_uid = sbi->uid;
6009         }
6010
6011 -       jfs_ip->saved_gid = le32_to_cpu(dip->di_gid);
6012 +       jfs_ip->saved_gid = INOTAG_GID(DX_TAG(ip), uid, gid);
6013         if (sbi->gid == -1)
6014                 ip->i_gid = jfs_ip->saved_gid;
6015         else {
6016 @@ -3150,14 +3157,12 @@ static void copy_to_dinode(struct dinode
6017         dip->di_size = cpu_to_le64(ip->i_size);
6018         dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks));
6019         dip->di_nlink = cpu_to_le32(ip->i_nlink);
6020 -       if (sbi->uid == -1)
6021 -               dip->di_uid = cpu_to_le32(ip->i_uid);
6022 -       else
6023 -               dip->di_uid = cpu_to_le32(jfs_ip->saved_uid);
6024 -       if (sbi->gid == -1)
6025 -               dip->di_gid = cpu_to_le32(ip->i_gid);
6026 -       else
6027 -               dip->di_gid = cpu_to_le32(jfs_ip->saved_gid);
6028 +
6029 +       dip->di_uid = cpu_to_le32(TAGINO_UID(DX_TAG(ip),
6030 +               (sbi->uid == -1) ? ip->i_uid : jfs_ip->saved_uid, ip->i_tag));
6031 +       dip->di_gid = cpu_to_le32(TAGINO_GID(DX_TAG(ip),
6032 +               (sbi->gid == -1) ? ip->i_gid : jfs_ip->saved_gid, ip->i_tag));
6033 +
6034         jfs_get_inode_flags(jfs_ip);
6035         /*
6036          * mode2 is only needed for storing the higher order bits.
6037 diff -NurpP --minimal linux-3.0/fs/jfs/jfs_inode.c linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_inode.c
6038 --- linux-3.0/fs/jfs/jfs_inode.c        2010-08-02 16:52:49.000000000 +0200
6039 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_inode.c   2011-06-10 22:11:24.000000000 +0200
6040 @@ -18,6 +18,7 @@
6041
6042  #include <linux/fs.h>
6043  #include <linux/quotaops.h>
6044 +#include <linux/vs_tag.h>
6045  #include "jfs_incore.h"
6046  #include "jfs_inode.h"
6047  #include "jfs_filsys.h"
6048 @@ -30,29 +31,46 @@ void jfs_set_inode_flags(struct inode *i
6049  {
6050         unsigned int flags = JFS_IP(inode)->mode2;
6051
6052 -       inode->i_flags &= ~(S_IMMUTABLE | S_APPEND |
6053 -               S_NOATIME | S_DIRSYNC | S_SYNC);
6054 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
6055 +               S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
6056
6057         if (flags & JFS_IMMUTABLE_FL)
6058                 inode->i_flags |= S_IMMUTABLE;
6059 +       if (flags & JFS_IXUNLINK_FL)
6060 +               inode->i_flags |= S_IXUNLINK;
6061 +
6062 +       if (flags & JFS_SYNC_FL)
6063 +               inode->i_flags |= S_SYNC;
6064         if (flags & JFS_APPEND_FL)
6065                 inode->i_flags |= S_APPEND;
6066         if (flags & JFS_NOATIME_FL)
6067                 inode->i_flags |= S_NOATIME;
6068         if (flags & JFS_DIRSYNC_FL)
6069                 inode->i_flags |= S_DIRSYNC;
6070 -       if (flags & JFS_SYNC_FL)
6071 -               inode->i_flags |= S_SYNC;
6072 +
6073 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
6074 +
6075 +       if (flags & JFS_BARRIER_FL)
6076 +               inode->i_vflags |= V_BARRIER;
6077 +       if (flags & JFS_COW_FL)
6078 +               inode->i_vflags |= V_COW;
6079  }
6080
6081  void jfs_get_inode_flags(struct jfs_inode_info *jfs_ip)
6082  {
6083         unsigned int flags = jfs_ip->vfs_inode.i_flags;
6084 +       unsigned int vflags = jfs_ip->vfs_inode.i_vflags;
6085 +
6086 +       jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_IXUNLINK_FL |
6087 +                          JFS_APPEND_FL | JFS_NOATIME_FL |
6088 +                          JFS_DIRSYNC_FL | JFS_SYNC_FL |
6089 +                          JFS_BARRIER_FL | JFS_COW_FL);
6090
6091 -       jfs_ip->mode2 &= ~(JFS_IMMUTABLE_FL | JFS_APPEND_FL | JFS_NOATIME_FL |
6092 -                          JFS_DIRSYNC_FL | JFS_SYNC_FL);
6093         if (flags & S_IMMUTABLE)
6094                 jfs_ip->mode2 |= JFS_IMMUTABLE_FL;
6095 +       if (flags & S_IXUNLINK)
6096 +               jfs_ip->mode2 |= JFS_IXUNLINK_FL;
6097 +
6098         if (flags & S_APPEND)
6099                 jfs_ip->mode2 |= JFS_APPEND_FL;
6100         if (flags & S_NOATIME)
6101 @@ -61,6 +79,11 @@ void jfs_get_inode_flags(struct jfs_inod
6102                 jfs_ip->mode2 |= JFS_DIRSYNC_FL;
6103         if (flags & S_SYNC)
6104                 jfs_ip->mode2 |= JFS_SYNC_FL;
6105 +
6106 +       if (vflags & V_BARRIER)
6107 +               jfs_ip->mode2 |= JFS_BARRIER_FL;
6108 +       if (vflags & V_COW)
6109 +               jfs_ip->mode2 |= JFS_COW_FL;
6110  }
6111
6112  /*
6113 diff -NurpP --minimal linux-3.0/fs/jfs/jfs_inode.h linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_inode.h
6114 --- linux-3.0/fs/jfs/jfs_inode.h        2011-07-22 11:18:05.000000000 +0200
6115 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/jfs_inode.h   2011-06-10 22:11:24.000000000 +0200
6116 @@ -39,6 +39,7 @@ extern struct dentry *jfs_fh_to_dentry(s
6117  extern struct dentry *jfs_fh_to_parent(struct super_block *sb, struct fid *fid,
6118         int fh_len, int fh_type);
6119  extern void jfs_set_inode_flags(struct inode *);
6120 +extern int jfs_sync_flags(struct inode *, int, int);
6121  extern int jfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
6122  extern int jfs_setattr(struct dentry *, struct iattr *);
6123
6124 diff -NurpP --minimal linux-3.0/fs/jfs/namei.c linux-3.0-vs2.3.1-pre7/fs/jfs/namei.c
6125 --- linux-3.0/fs/jfs/namei.c    2011-05-22 16:17:53.000000000 +0200
6126 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/namei.c       2011-06-10 22:11:24.000000000 +0200
6127 @@ -22,6 +22,7 @@
6128  #include <linux/ctype.h>
6129  #include <linux/quotaops.h>
6130  #include <linux/exportfs.h>
6131 +#include <linux/vs_tag.h>
6132  #include "jfs_incore.h"
6133  #include "jfs_superblock.h"
6134  #include "jfs_inode.h"
6135 @@ -1486,6 +1487,7 @@ static struct dentry *jfs_lookup(struct
6136                 return ERR_CAST(ip);
6137         }
6138
6139 +       dx_propagate_tag(nd, ip);
6140         return d_splice_alias(ip, dentry);
6141  }
6142
6143 @@ -1550,6 +1552,7 @@ const struct inode_operations jfs_dir_in
6144  #ifdef CONFIG_JFS_POSIX_ACL
6145         .check_acl      = jfs_check_acl,
6146  #endif
6147 +       .sync_flags     = jfs_sync_flags,
6148  };
6149
6150  const struct file_operations jfs_dir_operations = {
6151 diff -NurpP --minimal linux-3.0/fs/jfs/super.c linux-3.0-vs2.3.1-pre7/fs/jfs/super.c
6152 --- linux-3.0/fs/jfs/super.c    2011-05-22 16:17:53.000000000 +0200
6153 +++ linux-3.0-vs2.3.1-pre7/fs/jfs/super.c       2011-06-10 22:11:24.000000000 +0200
6154 @@ -198,7 +198,8 @@ static void jfs_put_super(struct super_b
6155  enum {
6156         Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize,
6157         Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota,
6158 -       Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask
6159 +       Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask,
6160 +       Opt_tag, Opt_notag, Opt_tagid
6161  };
6162
6163  static const match_table_t tokens = {
6164 @@ -208,6 +209,10 @@ static const match_table_t tokens = {
6165         {Opt_resize, "resize=%u"},
6166         {Opt_resize_nosize, "resize"},
6167         {Opt_errors, "errors=%s"},
6168 +       {Opt_tag, "tag"},
6169 +       {Opt_notag, "notag"},
6170 +       {Opt_tagid, "tagid=%u"},
6171 +       {Opt_tag, "tagxid"},
6172         {Opt_ignore, "noquota"},
6173         {Opt_ignore, "quota"},
6174         {Opt_usrquota, "usrquota"},
6175 @@ -342,6 +347,20 @@ static int parse_options(char *options,
6176                         }
6177                         break;
6178                 }
6179 +#ifndef CONFIG_TAGGING_NONE
6180 +               case Opt_tag:
6181 +                       *flag |= JFS_TAGGED;
6182 +                       break;
6183 +               case Opt_notag:
6184 +                       *flag &= JFS_TAGGED;
6185 +                       break;
6186 +#endif
6187 +#ifdef CONFIG_PROPAGATE
6188 +               case Opt_tagid:
6189 +                       /* use args[0] */
6190 +                       *flag |= JFS_TAGGED;
6191 +                       break;
6192 +#endif
6193                 default:
6194                         printk("jfs: Unrecognized mount option \"%s\" "
6195                                         " or missing value\n", p);
6196 @@ -373,6 +392,12 @@ static int jfs_remount(struct super_bloc
6197                 return -EINVAL;
6198         }
6199
6200 +       if ((flag & JFS_TAGGED) && !(sb->s_flags & MS_TAGGED)) {
6201 +               printk(KERN_ERR "JFS: %s: tagging not permitted on remount.\n",
6202 +                       sb->s_id);
6203 +               return -EINVAL;
6204 +       }
6205 +
6206         if (newLVSize) {
6207                 if (sb->s_flags & MS_RDONLY) {
6208                         printk(KERN_ERR
6209 @@ -455,6 +480,9 @@ static int jfs_fill_super(struct super_b
6210  #ifdef CONFIG_JFS_POSIX_ACL
6211         sb->s_flags |= MS_POSIXACL;
6212  #endif
6213 +       /* map mount option tagxid */
6214 +       if (sbi->flag & JFS_TAGGED)
6215 +               sb->s_flags |= MS_TAGGED;
6216
6217         if (newLVSize) {
6218                 printk(KERN_ERR "resize option for remount only\n");
6219 diff -NurpP --minimal linux-3.0/fs/libfs.c linux-3.0-vs2.3.1-pre7/fs/libfs.c
6220 --- linux-3.0/fs/libfs.c        2011-07-22 11:18:05.000000000 +0200
6221 +++ linux-3.0-vs2.3.1-pre7/fs/libfs.c   2011-07-22 11:20:39.000000000 +0200
6222 @@ -133,7 +133,8 @@ static inline unsigned char dt_type(stru
6223   * both impossible due to the lock on directory.
6224   */
6225
6226 -int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir)
6227 +static inline int do_dcache_readdir_filter(struct file *filp,
6228 +       void *dirent, filldir_t filldir, int (*filter)(struct dentry *dentry))
6229  {
6230         struct dentry *dentry = filp->f_path.dentry;
6231         struct dentry *cursor = filp->private_data;
6232 @@ -164,6 +165,8 @@ int dcache_readdir(struct file * filp, v
6233                         for (p=q->next; p != &dentry->d_subdirs; p=p->next) {
6234                                 struct dentry *next;
6235                                 next = list_entry(p, struct dentry, d_u.d_child);
6236 +                               if (filter && !filter(next))
6237 +                                       continue;
6238                                 spin_lock_nested(&next->d_lock, DENTRY_D_LOCK_NESTED);
6239                                 if (!simple_positive(next)) {
6240                                         spin_unlock(&next->d_lock);
6241 @@ -190,6 +193,17 @@ int dcache_readdir(struct file * filp, v
6242         return 0;
6243  }
6244
6245 +int dcache_readdir(struct file *filp, void *dirent, filldir_t filldir)
6246 +{
6247 +       return do_dcache_readdir_filter(filp, dirent, filldir, NULL);
6248 +}
6249 +
6250 +int dcache_readdir_filter(struct file *filp, void *dirent, filldir_t filldir,
6251 +       int (*filter)(struct dentry *))
6252 +{
6253 +       return do_dcache_readdir_filter(filp, dirent, filldir, filter);
6254 +}
6255 +
6256  ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos)
6257  {
6258         return -EISDIR;
6259 @@ -965,6 +979,7 @@ EXPORT_SYMBOL(dcache_dir_close);
6260  EXPORT_SYMBOL(dcache_dir_lseek);
6261  EXPORT_SYMBOL(dcache_dir_open);
6262  EXPORT_SYMBOL(dcache_readdir);
6263 +EXPORT_SYMBOL(dcache_readdir_filter);
6264  EXPORT_SYMBOL(generic_read_dir);
6265  EXPORT_SYMBOL(mount_pseudo);
6266  EXPORT_SYMBOL(simple_write_begin);
6267 diff -NurpP --minimal linux-3.0/fs/locks.c linux-3.0-vs2.3.1-pre7/fs/locks.c
6268 --- linux-3.0/fs/locks.c        2011-07-22 11:18:05.000000000 +0200
6269 +++ linux-3.0-vs2.3.1-pre7/fs/locks.c   2011-07-19 00:51:58.000000000 +0200
6270 @@ -126,6 +126,8 @@
6271  #include <linux/time.h>
6272  #include <linux/rcupdate.h>
6273  #include <linux/pid_namespace.h>
6274 +#include <linux/vs_base.h>
6275 +#include <linux/vs_limit.h>
6276
6277  #include <asm/uaccess.h>
6278
6279 @@ -171,13 +173,19 @@ static void locks_init_lock_always(struc
6280         fl->fl_flags = 0;
6281         fl->fl_type = 0;
6282         fl->fl_start = fl->fl_end = 0;
6283 +       fl->fl_xid = -1;
6284  }
6285
6286 +
6287  /* Allocate an empty lock structure. */
6288  struct file_lock *locks_alloc_lock(void)
6289  {
6290 -       struct file_lock *fl = kmem_cache_alloc(filelock_cache, GFP_KERNEL);
6291 +       struct file_lock *fl;
6292 +
6293 +       if (!vx_locks_avail(1))
6294 +               return NULL;
6295
6296 +       fl = kmem_cache_alloc(filelock_cache, GFP_KERNEL);
6297         if (fl)
6298                 locks_init_lock_always(fl);
6299
6300 @@ -208,6 +216,7 @@ void locks_free_lock(struct file_lock *f
6301         BUG_ON(!list_empty(&fl->fl_block));
6302         BUG_ON(!list_empty(&fl->fl_link));
6303
6304 +       vx_locks_dec(fl);
6305         locks_release_private(fl);
6306         kmem_cache_free(filelock_cache, fl);
6307  }
6308 @@ -272,6 +281,7 @@ void locks_copy_lock(struct file_lock *n
6309         new->fl_file = fl->fl_file;
6310         new->fl_ops = fl->fl_ops;
6311         new->fl_lmops = fl->fl_lmops;
6312 +       new->fl_xid = fl->fl_xid;
6313
6314         locks_copy_private(new, fl);
6315  }
6316 @@ -310,6 +320,11 @@ static int flock_make_lock(struct file *
6317         fl->fl_flags = FL_FLOCK;
6318         fl->fl_type = type;
6319         fl->fl_end = OFFSET_MAX;
6320 +
6321 +       vxd_assert(filp->f_xid == vx_current_xid(),
6322 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
6323 +       fl->fl_xid = filp->f_xid;
6324 +       vx_locks_inc(fl);
6325
6326         *lock = fl;
6327         return 0;
6328 @@ -459,6 +474,7 @@ static int lease_init(struct file *filp,
6329
6330         fl->fl_owner = current->files;
6331         fl->fl_pid = current->tgid;
6332 +       fl->fl_xid = vx_current_xid();
6333
6334         fl->fl_file = filp;
6335         fl->fl_flags = FL_LEASE;
6336 @@ -478,6 +494,11 @@ static struct file_lock *lease_alloc(str
6337         if (fl == NULL)
6338                 return ERR_PTR(error);
6339
6340 +       fl->fl_xid = vx_current_xid();
6341 +       if (filp)
6342 +               vxd_assert(filp->f_xid == fl->fl_xid,
6343 +                       "f_xid(%d) == fl_xid(%d)", filp->f_xid, fl->fl_xid);
6344 +       vx_locks_inc(fl);
6345         error = lease_init(filp, type, fl);
6346         if (error) {
6347                 locks_free_lock(fl);
6348 @@ -779,6 +800,7 @@ static int flock_lock_file(struct file *
6349                 lock_flocks();
6350         }
6351
6352 +       new_fl->fl_xid = -1;
6353  find_conflict:
6354         for_each_lock(inode, before) {
6355                 struct file_lock *fl = *before;
6356 @@ -799,6 +821,7 @@ find_conflict:
6357                 goto out;
6358         locks_copy_lock(new_fl, request);
6359         locks_insert_lock(before, new_fl);
6360 +       vx_locks_inc(new_fl);
6361         new_fl = NULL;
6362         error = 0;
6363
6364 @@ -809,7 +832,8 @@ out:
6365         return error;
6366  }
6367
6368 -static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
6369 +static int __posix_lock_file(struct inode *inode, struct file_lock *request,
6370 +       struct file_lock *conflock, xid_t xid)
6371  {
6372         struct file_lock *fl;
6373         struct file_lock *new_fl = NULL;
6374 @@ -819,6 +843,8 @@ static int __posix_lock_file(struct inod
6375         struct file_lock **before;
6376         int error, added = 0;
6377
6378 +       vxd_assert(xid == vx_current_xid(),
6379 +               "xid(%d) == current(%d)", xid, vx_current_xid());
6380         /*
6381          * We may need two file_lock structures for this operation,
6382          * so we get them in advance to avoid races.
6383 @@ -829,7 +855,11 @@ static int __posix_lock_file(struct inod
6384             (request->fl_type != F_UNLCK ||
6385              request->fl_start != 0 || request->fl_end != OFFSET_MAX)) {
6386                 new_fl = locks_alloc_lock();
6387 +               new_fl->fl_xid = xid;
6388 +               vx_locks_inc(new_fl);
6389                 new_fl2 = locks_alloc_lock();
6390 +               new_fl2->fl_xid = xid;
6391 +               vx_locks_inc(new_fl2);
6392         }
6393
6394         lock_flocks();
6395 @@ -1028,7 +1058,8 @@ static int __posix_lock_file(struct inod
6396  int posix_lock_file(struct file *filp, struct file_lock *fl,
6397                         struct file_lock *conflock)
6398  {
6399 -       return __posix_lock_file(filp->f_path.dentry->d_inode, fl, conflock);
6400 +       return __posix_lock_file(filp->f_path.dentry->d_inode,
6401 +               fl, conflock, filp->f_xid);
6402  }
6403  EXPORT_SYMBOL(posix_lock_file);
6404
6405 @@ -1118,7 +1149,7 @@ int locks_mandatory_area(int read_write,
6406         fl.fl_end = offset + count - 1;
6407
6408         for (;;) {
6409 -               error = __posix_lock_file(inode, &fl, NULL);
6410 +               error = __posix_lock_file(inode, &fl, NULL, filp->f_xid);
6411                 if (error != FILE_LOCK_DEFERRED)
6412                         break;
6413                 error = wait_event_interruptible(fl.fl_wait, !fl.fl_next);
6414 @@ -1431,6 +1462,7 @@ int generic_setlease(struct file *filp,
6415                 goto out;
6416
6417         locks_insert_lock(before, lease);
6418 +       vx_locks_inc(lease);
6419         return 0;
6420
6421  out:
6422 @@ -1815,6 +1847,11 @@ int fcntl_setlk(unsigned int fd, struct
6423         if (file_lock == NULL)
6424                 return -ENOLCK;
6425
6426 +       vxd_assert(filp->f_xid == vx_current_xid(),
6427 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
6428 +       file_lock->fl_xid = filp->f_xid;
6429 +       vx_locks_inc(file_lock);
6430 +
6431         /*
6432          * This might block, so we do it before checking the inode.
6433          */
6434 @@ -1933,6 +1970,11 @@ int fcntl_setlk64(unsigned int fd, struc
6435         if (file_lock == NULL)
6436                 return -ENOLCK;
6437
6438 +       vxd_assert(filp->f_xid == vx_current_xid(),
6439 +               "f_xid(%d) == current(%d)", filp->f_xid, vx_current_xid());
6440 +       file_lock->fl_xid = filp->f_xid;
6441 +       vx_locks_inc(file_lock);
6442 +
6443         /*
6444          * This might block, so we do it before checking the inode.
6445          */
6446 @@ -2198,8 +2240,11 @@ static int locks_show(struct seq_file *f
6447
6448         lock_get_status(f, fl, *((loff_t *)f->private), "");
6449
6450 -       list_for_each_entry(bfl, &fl->fl_block, fl_block)
6451 +       list_for_each_entry(bfl, &fl->fl_block, fl_block) {
6452 +               if (!vx_check(fl->fl_xid, VS_WATCH_P | VS_IDENT))
6453 +                       continue;
6454                 lock_get_status(f, bfl, *((loff_t *)f->private), " ->");
6455 +       }
6456
6457         return 0;
6458  }
6459 diff -NurpP --minimal linux-3.0/fs/namei.c linux-3.0-vs2.3.1-pre7/fs/namei.c
6460 --- linux-3.0/fs/namei.c        2011-07-22 11:18:05.000000000 +0200
6461 +++ linux-3.0-vs2.3.1-pre7/fs/namei.c   2011-07-23 17:04:13.000000000 +0200
6462 @@ -32,6 +32,14 @@
6463  #include <linux/fcntl.h>
6464  #include <linux/device_cgroup.h>
6465  #include <linux/fs_struct.h>
6466 +#include <linux/proc_fs.h>
6467 +#include <linux/vserver/inode.h>
6468 +#include <linux/vs_base.h>
6469 +#include <linux/vs_tag.h>
6470 +#include <linux/vs_cowbl.h>
6471 +#include <linux/vs_device.h>
6472 +#include <linux/vs_context.h>
6473 +#include <linux/pid_namespace.h>
6474  #include <asm/uaccess.h>
6475
6476  #include "internal.h"
6477 @@ -173,6 +181,84 @@ void putname(const char *name)
6478  EXPORT_SYMBOL(putname);
6479  #endif
6480
6481 +static inline int dx_barrier(const struct inode *inode)
6482 +{
6483 +       if (IS_BARRIER(inode) && !vx_check(0, VS_ADMIN | VS_WATCH)) {
6484 +               vxwprintk_task(1, "did hit the barrier.");
6485 +               return 1;
6486 +       }
6487 +       return 0;
6488 +}
6489 +
6490 +static int __dx_permission(const struct inode *inode, int mask)
6491 +{
6492 +       if (dx_barrier(inode))
6493 +               return -EACCES;
6494 +
6495 +       if (inode->i_sb->s_magic == DEVPTS_SUPER_MAGIC) {
6496 +               /* devpts is xid tagged */
6497 +               if (S_ISDIR(inode->i_mode) ||
6498 +                   vx_check((xid_t)inode->i_tag, VS_IDENT | VS_WATCH_P))
6499 +                       return 0;
6500 +
6501 +               /* just pretend we didn't find anything */
6502 +               return -ENOENT;
6503 +       }
6504 +       else if (inode->i_sb->s_magic == PROC_SUPER_MAGIC) {
6505 +               struct proc_dir_entry *de = PDE(inode);
6506 +
6507 +               if (de && !vx_hide_check(0, de->vx_flags))
6508 +                       goto out;
6509 +
6510 +               if ((mask & (MAY_WRITE | MAY_APPEND))) {
6511 +                       struct pid *pid;
6512 +                       struct task_struct *tsk;
6513 +
6514 +                       if (vx_check(0, VS_ADMIN | VS_WATCH_P) ||
6515 +                           vx_flags(VXF_STATE_SETUP, 0))
6516 +                               return 0;
6517 +
6518 +                       pid = PROC_I(inode)->pid;
6519 +                       if (!pid)
6520 +                               goto out;
6521 +
6522 +                       tsk = pid_task(pid, PIDTYPE_PID);
6523 +                       vxdprintk(VXD_CBIT(tag, 0), "accessing %p[#%u]",
6524 +                                 tsk, (tsk ? vx_task_xid(tsk) : 0));
6525 +                       if (tsk && vx_check(vx_task_xid(tsk), VS_IDENT | VS_WATCH_P))
6526 +                               return 0;
6527 +               }
6528 +               else {
6529 +                       /* FIXME: Should we block some entries here? */
6530 +                       return 0;
6531 +               }
6532 +       }
6533 +       else {
6534 +               if (dx_notagcheck(inode->i_sb) ||
6535 +                   dx_check(inode->i_tag, DX_HOSTID | DX_ADMIN | DX_WATCH |
6536 +                            DX_IDENT))
6537 +                       return 0;
6538 +       }
6539 +
6540 +out:
6541 +       return -EACCES;
6542 +}
6543 +
6544 +int dx_permission(const struct inode *inode, int mask)
6545 +{
6546 +       int ret = __dx_permission(inode, mask);
6547 +       if (unlikely(ret)) {
6548 +#ifndef        CONFIG_VSERVER_WARN_DEVPTS
6549 +               if (inode->i_sb->s_magic != DEVPTS_SUPER_MAGIC)
6550 +#endif
6551 +                   vxwprintk_task(1,
6552 +                       "denied [0x%x] access to inode %s:%p[#%d,%lu]",
6553 +                       mask, inode->i_sb->s_id, inode, inode->i_tag,
6554 +                       inode->i_ino);
6555 +       }
6556 +       return ret;
6557 +}
6558 +
6559  /*
6560   * This does basic POSIX ACL permission checking
6561   */
6562 @@ -283,10 +369,14 @@ int inode_permission(struct inode *inode
6563                 /*
6564                  * Nobody gets write access to an immutable file.
6565                  */
6566 -               if (IS_IMMUTABLE(inode))
6567 +               if (IS_IMMUTABLE(inode) && !IS_COW(inode))
6568                         return -EACCES;
6569         }
6570
6571 +       retval = dx_permission(inode, mask);
6572 +       if (retval)
6573 +               return retval;
6574 +
6575         if (inode->i_op->permission)
6576                 retval = inode->i_op->permission(inode, mask, 0);
6577         else
6578 @@ -582,6 +672,9 @@ static inline int exec_permission(struct
6579         int ret;
6580         struct user_namespace *ns = inode_userns(inode);
6581
6582 +       if (dx_barrier(inode))
6583 +               return -EACCES;
6584 +
6585         if (inode->i_op->permission) {
6586                 ret = inode->i_op->permission(inode, MAY_EXEC, flags);
6587         } else {
6588 @@ -1089,7 +1182,8 @@ static void follow_dotdot(struct nameida
6589
6590                 if (nd->path.dentry == nd->root.dentry &&
6591                     nd->path.mnt == nd->root.mnt) {
6592 -                       break;
6593 +                       /* for sane '/' avoid follow_mount() */
6594 +                       return;
6595                 }
6596                 if (nd->path.dentry != nd->path.mnt->mnt_root) {
6597                         /* rare case of legitimate dget_parent()... */
6598 @@ -1172,6 +1266,9 @@ static int do_lookup(struct nameidata *n
6599                                 goto unlazy;
6600                         }
6601                 }
6602 +
6603 +               /* FIXME: check dx permission */
6604 +
6605                 path->mnt = mnt;
6606                 path->dentry = dentry;
6607                 if (unlikely(!__follow_mount_rcu(nd, path, inode)))
6608 @@ -1220,6 +1317,8 @@ retry:
6609                 }
6610         }
6611
6612 +       /* FIXME: check dx permission */
6613 +
6614         path->mnt = mnt;
6615         path->dentry = dentry;
6616         err = follow_managed(path, nd->flags);
6617 @@ -1839,7 +1938,7 @@ static int may_delete(struct inode *dir,
6618         if (IS_APPEND(dir))
6619                 return -EPERM;
6620         if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
6621 -           IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
6622 +               IS_IXORUNLINK(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
6623                 return -EPERM;
6624         if (isdir) {
6625                 if (!S_ISDIR(victim->d_inode->i_mode))
6626 @@ -1966,6 +2065,15 @@ static int may_open(struct path *path, i
6627                 break;
6628         }
6629
6630 +#ifdef CONFIG_VSERVER_COWBL
6631 +       if (IS_COW(inode) &&
6632 +               ((flag & O_ACCMODE) != O_RDONLY)) {
6633 +               if (IS_COW_LINK(inode))
6634 +                       return -EMLINK;
6635 +               inode->i_flags &= ~(S_IXUNLINK|S_IMMUTABLE);
6636 +               mark_inode_dirty(inode);
6637 +       }
6638 +#endif
6639         error = inode_permission(inode, acc_mode);
6640         if (error)
6641                 return error;
6642 @@ -2859,7 +2967,7 @@ int vfs_link(struct dentry *old_dentry,
6643         /*
6644          * A link to an append-only or immutable file cannot be created.
6645          */
6646 -       if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
6647 +       if (IS_APPEND(inode) || IS_IXORUNLINK(inode))
6648                 return -EPERM;
6649         if (!dir->i_op->link)
6650                 return -EPERM;
6651 @@ -3245,6 +3353,222 @@ int vfs_follow_link(struct nameidata *nd
6652         return __vfs_follow_link(nd, link);
6653  }
6654
6655 +
6656 +#ifdef CONFIG_VSERVER_COWBL
6657 +
6658 +#include <linux/file.h>
6659 +
6660 +static inline
6661 +long do_cow_splice(struct file *in, struct file *out, size_t len)
6662 +{
6663 +       loff_t ppos = 0;
6664 +
6665 +       return do_splice_direct(in, &ppos, out, len, 0);
6666 +}
6667 +
6668 +struct dentry *cow_break_link(const char *pathname)
6669 +{
6670 +       int ret, mode, pathlen, redo = 0;
6671 +       struct nameidata old_nd, dir_nd;
6672 +       struct path old_path, new_path;
6673 +       struct dentry *dir, *res = NULL;
6674 +       struct file *old_file;
6675 +       struct file *new_file;
6676 +       char *to, *path, pad='\251';
6677 +       loff_t size;
6678 +
6679 +       vxdprintk(VXD_CBIT(misc, 1),
6680 +               "cow_break_link(" VS_Q("%s") ")", pathname);
6681 +       path = kmalloc(PATH_MAX, GFP_KERNEL);
6682 +       ret = -ENOMEM;
6683 +       if (!path)
6684 +               goto out;
6685 +
6686 +       /* old_nd will have refs to dentry and mnt */
6687 +       ret = do_path_lookup(AT_FDCWD, pathname, LOOKUP_FOLLOW, &old_nd);
6688 +       vxdprintk(VXD_CBIT(misc, 2), "do_path_lookup(old): %d", ret);
6689 +       if (ret < 0)
6690 +               goto out_free_path;
6691 +
6692 +       old_path = old_nd.path;
6693 +       mode = old_path.dentry->d_inode->i_mode;
6694 +
6695 +       to = d_path(&old_path, path, PATH_MAX-2);
6696 +       pathlen = strlen(to);
6697 +       vxdprintk(VXD_CBIT(misc, 2),
6698 +               "old path " VS_Q("%s") " [" VS_Q("%.*s") ":%d]", to,
6699 +               old_path.dentry->d_name.len, old_path.dentry->d_name.name,
6700 +               old_path.dentry->d_name.len);
6701 +
6702 +       to[pathlen + 1] = 0;
6703 +retry:
6704 +       to[pathlen] = pad--;
6705 +       ret = -EMLINK;
6706 +       if (pad <= '\240')
6707 +               goto out_rel_old;
6708 +
6709 +       vxdprintk(VXD_CBIT(misc, 1), "temp copy " VS_Q("%s"), to);
6710 +       /* dir_nd will have refs to dentry and mnt */
6711 +       ret = do_path_lookup(AT_FDCWD, to,
6712 +               LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE, &dir_nd);
6713 +       vxdprintk(VXD_CBIT(misc, 2),
6714 +               "do_path_lookup(new): %d", ret);
6715 +       if (ret < 0)
6716 +               goto retry;
6717 +
6718 +       /* this puppy downs the inode mutex */
6719 +       new_path.dentry = lookup_create(&dir_nd, 0);
6720 +       if (!new_path.dentry || IS_ERR(new_path.dentry)) {
6721 +               vxdprintk(VXD_CBIT(misc, 2),
6722 +                       "lookup_create(new): %p", new_path.dentry);
6723 +               mutex_unlock(&dir_nd.path.dentry->d_inode->i_mutex);
6724 +               path_put(&dir_nd.path);
6725 +               goto retry;
6726 +       }
6727 +       vxdprintk(VXD_CBIT(misc, 2),
6728 +               "lookup_create(new): %p [" VS_Q("%.*s") ":%d]",
6729 +               new_path.dentry,
6730 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
6731 +               new_path.dentry->d_name.len);
6732 +       dir = dir_nd.path.dentry;
6733 +
6734 +       ret = vfs_create(dir_nd.path.dentry->d_inode, new_path.dentry, mode, &dir_nd);
6735 +       vxdprintk(VXD_CBIT(misc, 2),
6736 +               "vfs_create(new): %d", ret);
6737 +       if (ret == -EEXIST) {
6738 +               mutex_unlock(&dir->d_inode->i_mutex);
6739 +               dput(new_path.dentry);
6740 +               path_put(&dir_nd.path);
6741 +               goto retry;
6742 +       }
6743 +       else if (ret < 0)
6744 +               goto out_unlock_new;
6745 +
6746 +       /* drop out early, ret passes ENOENT */
6747 +       ret = -ENOENT;
6748 +       if ((redo = d_unhashed(old_path.dentry)))
6749 +               goto out_unlock_new;
6750 +
6751 +       new_path.mnt = dir_nd.path.mnt;
6752 +       dget(old_path.dentry);
6753 +       mntget(old_path.mnt);
6754 +       /* this one cleans up the dentry/mnt in case of failure */
6755 +       old_file = dentry_open(old_path.dentry, old_path.mnt,
6756 +               O_RDONLY, current_cred());
6757 +       vxdprintk(VXD_CBIT(misc, 2),
6758 +               "dentry_open(old): %p", old_file);
6759 +       if (!old_file || IS_ERR(old_file)) {
6760 +               res = IS_ERR(old_file) ? (void *) old_file : res;
6761 +               goto out_unlock_new;
6762 +       }
6763 +
6764 +       dget(new_path.dentry);
6765 +       mntget(new_path.mnt);
6766 +       /* this one cleans up the dentry/mnt in case of failure */
6767 +       new_file = dentry_open(new_path.dentry, new_path.mnt,
6768 +               O_WRONLY, current_cred());
6769 +       vxdprintk(VXD_CBIT(misc, 2),
6770 +               "dentry_open(new): %p", new_file);
6771 +
6772 +       ret = IS_ERR(new_file) ? PTR_ERR(new_file) : -ENOENT;
6773 +       if (!new_file || IS_ERR(new_file))
6774 +               goto out_fput_old;
6775 +
6776 +       size = i_size_read(old_file->f_dentry->d_inode);
6777 +       ret = do_cow_splice(old_file, new_file, size);
6778 +       vxdprintk(VXD_CBIT(misc, 2), "do_splice_direct: %d", ret);
6779 +       if (ret < 0) {
6780 +               goto out_fput_both;
6781 +       } else if (ret < size) {
6782 +               ret = -ENOSPC;
6783 +               goto out_fput_both;
6784 +       } else {
6785 +               struct inode *old_inode = old_path.dentry->d_inode;
6786 +               struct inode *new_inode = new_path.dentry->d_inode;
6787 +               struct iattr attr = {
6788 +                       .ia_uid = old_inode->i_uid,
6789 +                       .ia_gid = old_inode->i_gid,
6790 +                       .ia_valid = ATTR_UID | ATTR_GID
6791 +                       };
6792 +
6793 +               setattr_copy(new_inode, &attr);
6794 +               mark_inode_dirty(new_inode);
6795 +       }
6796 +
6797 +       mutex_lock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
6798 +
6799 +       /* drop out late */
6800 +       ret = -ENOENT;
6801 +       if ((redo = d_unhashed(old_path.dentry)))
6802 +               goto out_unlock;
6803 +
6804 +       vxdprintk(VXD_CBIT(misc, 2),
6805 +               "vfs_rename: [" VS_Q("%*s") ":%d] -> [" VS_Q("%*s") ":%d]",
6806 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
6807 +               new_path.dentry->d_name.len,
6808 +               old_path.dentry->d_name.len, old_path.dentry->d_name.name,
6809 +               old_path.dentry->d_name.len);
6810 +       ret = vfs_rename(dir_nd.path.dentry->d_inode, new_path.dentry,
6811 +               old_nd.path.dentry->d_parent->d_inode, old_path.dentry);
6812 +       vxdprintk(VXD_CBIT(misc, 2), "vfs_rename: %d", ret);
6813 +       res = new_path.dentry;
6814 +
6815 +out_unlock:
6816 +       mutex_unlock(&old_path.dentry->d_inode->i_sb->s_vfs_rename_mutex);
6817 +
6818 +out_fput_both:
6819 +       vxdprintk(VXD_CBIT(misc, 3),
6820 +               "fput(new_file=%p[#%ld])", new_file,
6821 +               atomic_long_read(&new_file->f_count));
6822 +       fput(new_file);
6823 +
6824 +out_fput_old:
6825 +       vxdprintk(VXD_CBIT(misc, 3),
6826 +               "fput(old_file=%p[#%ld])", old_file,
6827 +               atomic_long_read(&old_file->f_count));
6828 +       fput(old_file);
6829 +
6830 +out_unlock_new:
6831 +       mutex_unlock(&dir->d_inode->i_mutex);
6832 +       if (!ret)
6833 +               goto out_redo;
6834 +
6835 +       /* error path cleanup */
6836 +       vfs_unlink(dir->d_inode, new_path.dentry);
6837 +       dput(new_path.dentry);
6838 +
6839 +out_redo:
6840 +       if (!redo)
6841 +               goto out_rel_both;
6842 +       /* lookup dentry once again */
6843 +       path_put(&old_nd.path);
6844 +       ret = do_path_lookup(AT_FDCWD, pathname, LOOKUP_FOLLOW, &old_nd);
6845 +       if (ret)
6846 +               goto out_rel_both;
6847 +
6848 +       new_path.dentry = old_nd.path.dentry;
6849 +       vxdprintk(VXD_CBIT(misc, 2),
6850 +               "do_path_lookup(redo): %p [" VS_Q("%.*s") ":%d]",
6851 +               new_path.dentry,
6852 +               new_path.dentry->d_name.len, new_path.dentry->d_name.name,
6853 +               new_path.dentry->d_name.len);
6854 +       dget(new_path.dentry);
6855 +       res = new_path.dentry;
6856 +
6857 +out_rel_both:
6858 +       path_put(&dir_nd.path);
6859 +out_rel_old:
6860 +       path_put(&old_nd.path);
6861 +out_free_path:
6862 +       kfree(path);
6863 +out:
6864 +       if (ret)
6865 +               res = ERR_PTR(ret);
6866 +       return res;
6867 +}
6868 +
6869 +#endif
6870 +
6871  /* get the link contents into pagecache */
6872  static char *page_getlink(struct dentry * dentry, struct page **ppage)
6873  {
6874 diff -NurpP --minimal linux-3.0/fs/namespace.c linux-3.0-vs2.3.1-pre7/fs/namespace.c
6875 --- linux-3.0/fs/namespace.c    2011-07-22 11:18:05.000000000 +0200
6876 +++ linux-3.0-vs2.3.1-pre7/fs/namespace.c       2011-07-23 16:43:31.000000000 +0200
6877 @@ -31,6 +31,11 @@
6878  #include <linux/idr.h>
6879  #include <linux/fs_struct.h>
6880  #include <linux/fsnotify.h>
6881 +#include <linux/vs_base.h>
6882 +#include <linux/vs_context.h>
6883 +#include <linux/vs_tag.h>
6884 +#include <linux/vserver/space.h>
6885 +#include <linux/vserver/global.h>
6886  #include <asm/uaccess.h>
6887  #include <asm/unistd.h>
6888  #include "pnode.h"
6889 @@ -679,6 +684,10 @@ vfs_kern_mount(struct file_system_type *
6890         if (!type)
6891                 return ERR_PTR(-ENODEV);
6892
6893 +       if ((type->fs_flags & FS_BINARY_MOUNTDATA) &&
6894 +               !vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT))
6895 +               return ERR_PTR(-EPERM);
6896 +
6897         mnt = alloc_vfsmnt(name);
6898         if (!mnt)
6899                 return ERR_PTR(-ENOMEM);
6900 @@ -724,6 +733,7 @@ static struct vfsmount *clone_mnt(struct
6901                 mnt->mnt_root = dget(root);
6902                 mnt->mnt_mountpoint = mnt->mnt_root;
6903                 mnt->mnt_parent = mnt;
6904 +               mnt->mnt_tag = old->mnt_tag;
6905
6906                 if (flag & CL_SLAVE) {
6907                         list_add(&mnt->mnt_slave, &old->mnt_slave_list);
6908 @@ -852,6 +862,31 @@ static inline void mangle(struct seq_fil
6909         seq_escape(m, s, " \t\n\\");
6910  }
6911
6912 +static int mnt_is_reachable(struct vfsmount *mnt)
6913 +{
6914 +       struct path root;
6915 +       struct dentry *point;
6916 +       int ret;
6917 +
6918 +       if (mnt == mnt->mnt_ns->root)
6919 +               return 1;
6920 +
6921 +       br_read_lock(vfsmount_lock);
6922 +       root = current->fs->root;
6923 +       point = root.dentry;
6924 +
6925 +       while ((mnt != mnt->mnt_parent) && (mnt != root.mnt)) {
6926 +               point = mnt->mnt_mountpoint;
6927 +               mnt = mnt->mnt_parent;
6928 +       }
6929 +
6930 +       ret = (mnt == root.mnt) && is_subdir(point, root.dentry);
6931 +
6932 +       br_read_unlock(vfsmount_lock);
6933 +
6934 +       return ret;
6935 +}
6936 +
6937  /*
6938   * Simple .show_options callback for filesystems which don't want to
6939   * implement more complex mount option showing.
6940 @@ -954,6 +989,8 @@ static int show_sb_opts(struct seq_file
6941                 { MS_SYNCHRONOUS, ",sync" },
6942                 { MS_DIRSYNC, ",dirsync" },
6943                 { MS_MANDLOCK, ",mand" },
6944 +               { MS_TAGGED, ",tag" },
6945 +               { MS_NOTAGCHECK, ",notagcheck" },
6946                 { 0, NULL }
6947         };
6948         const struct proc_fs_info *fs_infop;
6949 @@ -1000,16 +1037,26 @@ static int show_vfsmnt(struct seq_file *
6950         int err = 0;
6951         struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
6952
6953 -       if (mnt->mnt_sb->s_op->show_devname) {
6954 -               err = mnt->mnt_sb->s_op->show_devname(m, mnt);
6955 -               if (err)
6956 -                       goto out;
6957 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
6958 +               return SEQ_SKIP;
6959 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
6960 +               return SEQ_SKIP;
6961 +
6962 +       if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
6963 +               mnt == current->fs->root.mnt) {
6964 +               seq_puts(m, "/dev/root / ");
6965         } else {
6966 -               mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
6967 +               if (mnt->mnt_sb->s_op->show_devname) {
6968 +                       err = mnt->mnt_sb->s_op->show_devname(m, mnt);
6969 +                       if (err)
6970 +                               goto out;
6971 +               } else {
6972 +                       mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none");
6973 +               }
6974 +               seq_putc(m, ' ');
6975 +               seq_path(m, &mnt_path, " \t\n\\");
6976 +               seq_putc(m, ' ');
6977         }
6978 -       seq_putc(m, ' ');
6979 -       seq_path(m, &mnt_path, " \t\n\\");
6980 -       seq_putc(m, ' ');
6981         show_type(m, mnt->mnt_sb);
6982         seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw");
6983         err = show_sb_opts(m, mnt->mnt_sb);
6984 @@ -1039,6 +1086,11 @@ static int show_mountinfo(struct seq_fil
6985         struct path root = p->root;
6986         int err = 0;
6987
6988 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
6989 +               return SEQ_SKIP;
6990 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
6991 +               return SEQ_SKIP;
6992 +
6993         seq_printf(m, "%i %i %u:%u ", mnt->mnt_id, mnt->mnt_parent->mnt_id,
6994                    MAJOR(sb->s_dev), MINOR(sb->s_dev));
6995         if (sb->s_op->show_path)
6996 @@ -1107,21 +1159,31 @@ static int show_vfsstat(struct seq_file
6997         struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
6998         int err = 0;
6999
7000 -       /* device */
7001 -       if (mnt->mnt_sb->s_op->show_devname) {
7002 -               err = mnt->mnt_sb->s_op->show_devname(m, mnt);
7003 +       if (vx_flags(VXF_HIDE_MOUNT, 0))
7004 +               return SEQ_SKIP;
7005 +       if (!mnt_is_reachable(mnt) && !vx_check(0, VS_WATCH_P))
7006 +               return SEQ_SKIP;
7007 +
7008 +       if (!vx_check(0, VS_ADMIN|VS_WATCH) &&
7009 +               mnt == current->fs->root.mnt) {
7010 +               seq_puts(m, "device /dev/root mounted on / ");
7011         } else {
7012 -               if (mnt->mnt_devname) {
7013 -                       seq_puts(m, "device ");
7014 -                       mangle(m, mnt->mnt_devname);
7015 -               } else
7016 -                       seq_puts(m, "no device");
7017 -       }
7018 +               /* device */
7019 +               if (mnt->mnt_sb->s_op->show_devname) {
7020 +                       err = mnt->mnt_sb->s_op->show_devname(m, mnt);
7021 +               } else {
7022 +                       if (mnt->mnt_devname) {
7023 +                               seq_puts(m, "device ");
7024 +                               mangle(m, mnt->mnt_devname);
7025 +                       } else
7026 +                               seq_puts(m, "no device");
7027 +               }
7028
7029 -       /* mount point */
7030 -       seq_puts(m, " mounted on ");
7031 -       seq_path(m, &mnt_path, " \t\n\\");
7032 -       seq_putc(m, ' ');
7033 +               /* mount point */
7034 +               seq_puts(m, " mounted on ");
7035 +               seq_path(m, &mnt_path, " \t\n\\");
7036 +               seq_putc(m, ' ');
7037 +       }
7038
7039         /* file system type */
7040         seq_puts(m, "with fstype ");
7041 @@ -1380,7 +1442,7 @@ SYSCALL_DEFINE2(umount, char __user *, n
7042                 goto dput_and_out;
7043
7044         retval = -EPERM;
7045 -       if (!capable(CAP_SYS_ADMIN))
7046 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
7047                 goto dput_and_out;
7048
7049         retval = do_umount(path.mnt, flags);
7050 @@ -1406,7 +1468,7 @@ SYSCALL_DEFINE1(oldumount, char __user *
7051
7052  static int mount_is_safe(struct path *path)
7053  {
7054 -       if (capable(CAP_SYS_ADMIN))
7055 +       if (vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
7056                 return 0;
7057         return -EPERM;
7058  #ifdef notyet
7059 @@ -1716,7 +1778,7 @@ static int do_change_type(struct path *p
7060         int type;
7061         int err = 0;
7062
7063 -       if (!capable(CAP_SYS_ADMIN))
7064 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_NAMESPACE))
7065                 return -EPERM;
7066
7067         if (path->dentry != path->mnt->mnt_root)
7068 @@ -1732,6 +1794,7 @@ static int do_change_type(struct path *p
7069                 if (err)
7070                         goto out_unlock;
7071         }
7072 +       // mnt->mnt_flags = mnt_flags;
7073
7074         br_write_lock(vfsmount_lock);
7075         for (m = mnt; m; m = (recurse ? next_mnt(m, mnt) : NULL))
7076 @@ -1747,12 +1810,14 @@ static int do_change_type(struct path *p
7077   * do loopback mount.
7078   */
7079  static int do_loopback(struct path *path, char *old_name,
7080 -                               int recurse)
7081 +       tag_t tag, unsigned long flags, int mnt_flags)
7082  {
7083         LIST_HEAD(umount_list);
7084         struct path old_path;
7085         struct vfsmount *mnt = NULL;
7086         int err = mount_is_safe(path);
7087 +       int recurse = flags & MS_REC;
7088 +
7089         if (err)
7090                 return err;
7091         if (!old_name || !*old_name)
7092 @@ -1818,12 +1883,12 @@ static int change_mount_flags(struct vfs
7093   * on it - tough luck.
7094   */
7095  static int do_remount(struct path *path, int flags, int mnt_flags,
7096 -                     void *data)
7097 +       void *data, xid_t xid)
7098  {
7099         int err;
7100         struct super_block *sb = path->mnt->mnt_sb;
7101
7102 -       if (!capable(CAP_SYS_ADMIN))
7103 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_REMOUNT))
7104                 return -EPERM;
7105
7106         if (!check_mnt(path->mnt))
7107 @@ -1871,7 +1936,7 @@ static int do_move_mount(struct path *pa
7108         struct path old_path, parent_path;
7109         struct vfsmount *p;
7110         int err = 0;
7111 -       if (!capable(CAP_SYS_ADMIN))
7112 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
7113                 return -EPERM;
7114         if (!old_name || !*old_name)
7115                 return -EINVAL;
7116 @@ -2022,7 +2087,7 @@ static int do_new_mount(struct path *pat
7117                 return -EINVAL;
7118
7119         /* we need capabilities... */
7120 -       if (!capable(CAP_SYS_ADMIN))
7121 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT))
7122                 return -EPERM;
7123
7124         mnt = do_kern_mount(type, flags, name, data);
7125 @@ -2291,6 +2356,7 @@ long do_mount(char *dev_name, char *dir_
7126         struct path path;
7127         int retval = 0;
7128         int mnt_flags = 0;
7129 +       tag_t tag = 0;
7130
7131         /* Discard magic */
7132         if ((flags & MS_MGC_MSK) == MS_MGC_VAL)
7133 @@ -2318,6 +2384,12 @@ long do_mount(char *dev_name, char *dir_
7134         if (!(flags & MS_NOATIME))
7135                 mnt_flags |= MNT_RELATIME;
7136
7137 +       if (dx_parse_tag(data_page, &tag, 1, &mnt_flags, &flags)) {
7138 +               /* FIXME: bind and re-mounts get the tag flag? */
7139 +               if (flags & (MS_BIND|MS_REMOUNT))
7140 +                       flags |= MS_TAGID;
7141 +       }
7142 +
7143         /* Separate the per-mountpoint flags */
7144         if (flags & MS_NOSUID)
7145                 mnt_flags |= MNT_NOSUID;
7146 @@ -2334,15 +2406,17 @@ long do_mount(char *dev_name, char *dir_
7147         if (flags & MS_RDONLY)
7148                 mnt_flags |= MNT_READONLY;
7149
7150 +       if (!capable(CAP_SYS_ADMIN))
7151 +               mnt_flags |= MNT_NODEV;
7152         flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_BORN |
7153                    MS_NOATIME | MS_NODIRATIME | MS_RELATIME| MS_KERNMOUNT |
7154                    MS_STRICTATIME);
7155
7156         if (flags & MS_REMOUNT)
7157                 retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags,
7158 -                                   data_page);
7159 +                                   data_page, tag);
7160         else if (flags & MS_BIND)
7161 -               retval = do_loopback(&path, dev_name, flags & MS_REC);
7162 +               retval = do_loopback(&path, dev_name, tag, flags, mnt_flags);
7163         else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE))
7164                 retval = do_change_type(&path, flags);
7165         else if (flags & MS_MOVE)
7166 @@ -2442,6 +2516,7 @@ static struct mnt_namespace *dup_mnt_ns(
7167                 q = next_mnt(q, new_ns->root);
7168         }
7169         up_write(&namespace_sem);
7170 +       atomic_inc(&vs_global_mnt_ns);
7171
7172         if (rootmnt)
7173                 mntput(rootmnt);
7174 @@ -2581,9 +2656,10 @@ SYSCALL_DEFINE2(pivot_root, const char _
7175                 goto out3;
7176
7177         error = -EINVAL;
7178 -       if (IS_MNT_SHARED(old.mnt) ||
7179 -               IS_MNT_SHARED(new.mnt->mnt_parent) ||
7180 -               IS_MNT_SHARED(root.mnt->mnt_parent))
7181 +       if ((IS_MNT_SHARED(old.mnt) ||
7182 +               IS_MNT_SHARED(new.mnt->mnt_parent) ||
7183 +               IS_MNT_SHARED(root.mnt->mnt_parent)) &&
7184 +               !vx_flags(VXF_STATE_SETUP, 0))
7185                 goto out4;
7186         if (!check_mnt(root.mnt) || !check_mnt(new.mnt))
7187                 goto out4;
7188 @@ -2715,6 +2791,7 @@ void put_mnt_ns(struct mnt_namespace *ns
7189         br_write_unlock(vfsmount_lock);
7190         up_write(&namespace_sem);
7191         release_mounts(&umount_list);
7192 +       atomic_dec(&vs_global_mnt_ns);
7193         kfree(ns);
7194  }
7195  EXPORT_SYMBOL(put_mnt_ns);
7196 diff -NurpP --minimal linux-3.0/fs/nfs/client.c linux-3.0-vs2.3.1-pre7/fs/nfs/client.c
7197 --- linux-3.0/fs/nfs/client.c   2011-07-22 11:18:05.000000000 +0200
7198 +++ linux-3.0-vs2.3.1-pre7/fs/nfs/client.c      2011-06-10 22:11:24.000000000 +0200
7199 @@ -780,6 +780,9 @@ static int nfs_init_server_rpcclient(str
7200         if (server->flags & NFS_MOUNT_SOFT)
7201                 server->client->cl_softrtry = 1;
7202
7203 +       server->client->cl_tag = 0;
7204 +       if (server->flags & NFS_MOUNT_TAGGED)
7205 +               server->client->cl_tag = 1;
7206         return 0;
7207  }
7208
7209 @@ -951,6 +954,10 @@ static void nfs_server_set_fsinfo(struct
7210                 server->acdirmin = server->acdirmax = 0;
7211         }
7212
7213 +       /* FIXME: needs fsinfo
7214 +       if (server->flags & NFS_MOUNT_TAGGED)
7215 +               sb->s_flags |= MS_TAGGED;       */
7216 +
7217         server->maxfilesize = fsinfo->maxfilesize;
7218
7219         server->time_delta = fsinfo->time_delta;
7220 diff -NurpP --minimal linux-3.0/fs/nfs/dir.c linux-3.0-vs2.3.1-pre7/fs/nfs/dir.c
7221 --- linux-3.0/fs/nfs/dir.c      2011-07-22 11:18:05.000000000 +0200
7222 +++ linux-3.0-vs2.3.1-pre7/fs/nfs/dir.c 2011-06-10 22:11:24.000000000 +0200
7223 @@ -35,6 +35,7 @@
7224  #include <linux/sched.h>
7225  #include <linux/kmemleak.h>
7226  #include <linux/xattr.h>
7227 +#include <linux/vs_tag.h>
7228
7229  #include "delegation.h"
7230  #include "iostat.h"
7231 @@ -1298,6 +1299,7 @@ static struct dentry *nfs_lookup(struct
7232         if (IS_ERR(res))
7233                 goto out_unblock_sillyrename;
7234
7235 +       dx_propagate_tag(nd, inode);
7236  no_entry:
7237         res = d_materialise_unique(dentry, inode);
7238         if (res != NULL) {
7239 diff -NurpP --minimal linux-3.0/fs/nfs/inode.c linux-3.0-vs2.3.1-pre7/fs/nfs/inode.c
7240 --- linux-3.0/fs/nfs/inode.c    2011-07-22 11:18:05.000000000 +0200
7241 +++ linux-3.0-vs2.3.1-pre7/fs/nfs/inode.c       2011-07-01 11:35:34.000000000 +0200
7242 @@ -38,6 +38,7 @@
7243  #include <linux/nfs_xdr.h>
7244  #include <linux/slab.h>
7245  #include <linux/compat.h>
7246 +#include <linux/vs_tag.h>
7247
7248  #include <asm/system.h>
7249  #include <asm/uaccess.h>
7250 @@ -273,6 +274,8 @@ nfs_fhget(struct super_block *sb, struct
7251         if (inode->i_state & I_NEW) {
7252                 struct nfs_inode *nfsi = NFS_I(inode);
7253                 unsigned long now = jiffies;
7254 +               uid_t uid;
7255 +               gid_t gid;
7256
7257                 /* We set i_ino for the few things that still rely on it,
7258                  * such as stat(2) */
7259 @@ -321,8 +324,8 @@ nfs_fhget(struct super_block *sb, struct
7260                 nfsi->change_attr = 0;
7261                 inode->i_size = 0;
7262                 inode->i_nlink = 0;
7263 -               inode->i_uid = -2;
7264 -               inode->i_gid = -2;
7265 +               uid = -2;
7266 +               gid = -2;
7267                 inode->i_blocks = 0;
7268                 memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf));
7269
7270 @@ -359,13 +362,13 @@ nfs_fhget(struct super_block *sb, struct
7271                 else if (nfs_server_capable(inode, NFS_CAP_NLINK))
7272                         nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
7273                 if (fattr->valid & NFS_ATTR_FATTR_OWNER)
7274 -                       inode->i_uid = fattr->uid;
7275 +                       uid = fattr->uid;
7276                 else if (nfs_server_capable(inode, NFS_CAP_OWNER))
7277                         nfsi->cache_validity |= NFS_INO_INVALID_ATTR
7278                                 | NFS_INO_INVALID_ACCESS
7279                                 | NFS_INO_INVALID_ACL;
7280                 if (fattr->valid & NFS_ATTR_FATTR_GROUP)
7281 -                       inode->i_gid = fattr->gid;
7282 +                       gid = fattr->gid;
7283                 else if (nfs_server_capable(inode, NFS_CAP_OWNER_GROUP))
7284                         nfsi->cache_validity |= NFS_INO_INVALID_ATTR
7285                                 | NFS_INO_INVALID_ACCESS
7286 @@ -378,6 +381,11 @@ nfs_fhget(struct super_block *sb, struct
7287                          */
7288                         inode->i_blocks = nfs_calc_block_size(fattr->du.nfs3.used);
7289                 }
7290 +               inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
7291 +               inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
7292 +               inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
7293 +                               /* maybe fattr->xid someday */
7294 +
7295                 nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
7296                 nfsi->attrtimeo_timestamp = now;
7297                 nfsi->access_cache = RB_ROOT;
7298 @@ -494,6 +502,8 @@ void nfs_setattr_update_inode(struct ino
7299                         inode->i_uid = attr->ia_uid;
7300                 if ((attr->ia_valid & ATTR_GID) != 0)
7301                         inode->i_gid = attr->ia_gid;
7302 +               if ((attr->ia_valid & ATTR_TAG) && IS_TAGGED(inode))
7303 +                       inode->i_tag = attr->ia_tag;
7304                 NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
7305                 spin_unlock(&inode->i_lock);
7306         }
7307 @@ -941,6 +951,9 @@ static int nfs_check_inode_attributes(st
7308         struct nfs_inode *nfsi = NFS_I(inode);
7309         loff_t cur_size, new_isize;
7310         unsigned long invalid = 0;
7311 +       uid_t uid;
7312 +       gid_t gid;
7313 +       tag_t tag;
7314
7315
7316         /* Has the inode gone and changed behind our back? */
7317 @@ -964,13 +977,18 @@ static int nfs_check_inode_attributes(st
7318                         invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
7319         }
7320
7321 +       uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
7322 +       gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
7323 +       tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
7324 +
7325         /* Have any file permissions changed? */
7326         if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO))
7327                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
7328 -       if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && inode->i_uid != fattr->uid)
7329 +       if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && uid != fattr->uid)
7330                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
7331 -       if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && inode->i_gid != fattr->gid)
7332 +       if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && gid != fattr->gid)
7333                 invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
7334 +               /* maybe check for tag too? */
7335
7336         /* Has the link count changed? */
7337         if ((fattr->valid & NFS_ATTR_FATTR_NLINK) && inode->i_nlink != fattr->nlink)
7338 @@ -1205,6 +1223,9 @@ static int nfs_update_inode(struct inode
7339         unsigned long invalid = 0;
7340         unsigned long now = jiffies;
7341         unsigned long save_cache_validity;
7342 +       uid_t uid;
7343 +       gid_t gid;
7344 +       tag_t tag;
7345
7346         dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
7347                         __func__, inode->i_sb->s_id, inode->i_ino,
7348 @@ -1312,6 +1333,9 @@ static int nfs_update_inode(struct inode
7349                                 | NFS_INO_REVAL_PAGECACHE
7350                                 | NFS_INO_REVAL_FORCED);
7351
7352 +       uid = INOTAG_UID(DX_TAG(inode), fattr->uid, fattr->gid);
7353 +       gid = INOTAG_GID(DX_TAG(inode), fattr->uid, fattr->gid);
7354 +       tag = INOTAG_TAG(DX_TAG(inode), fattr->uid, fattr->gid, 0);
7355
7356         if (fattr->valid & NFS_ATTR_FATTR_ATIME)
7357                 memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
7358 @@ -1333,9 +1357,9 @@ static int nfs_update_inode(struct inode
7359                                 | NFS_INO_REVAL_FORCED);
7360
7361         if (fattr->valid & NFS_ATTR_FATTR_OWNER) {
7362 -               if (inode->i_uid != fattr->uid) {
7363 +               if (uid != fattr->uid) {
7364                         invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
7365 -                       inode->i_uid = fattr->uid;
7366 +                       uid = fattr->uid;
7367                 }
7368         } else if (server->caps & NFS_CAP_OWNER)
7369                 invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
7370 @@ -1344,9 +1368,9 @@ static int nfs_update_inode(struct inode
7371                                 | NFS_INO_REVAL_FORCED);
7372
7373         if (fattr->valid & NFS_ATTR_FATTR_GROUP) {
7374 -               if (inode->i_gid != fattr->gid) {
7375 +               if (gid != fattr->gid) {
7376                         invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
7377 -                       inode->i_gid = fattr->gid;
7378 +                       gid = fattr->gid;
7379                 }
7380         } else if (server->caps & NFS_CAP_OWNER_GROUP)
7381                 invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
7382 @@ -1354,6 +1378,10 @@ static int nfs_update_inode(struct inode
7383                                 | NFS_INO_INVALID_ACL
7384                                 | NFS_INO_REVAL_FORCED);
7385
7386 +       inode->i_uid = uid;
7387 +       inode->i_gid = gid;
7388 +       inode->i_tag = tag;
7389 +
7390         if (fattr->valid & NFS_ATTR_FATTR_NLINK) {
7391                 if (inode->i_nlink != fattr->nlink) {
7392                         invalid |= NFS_INO_INVALID_ATTR;
7393 diff -NurpP --minimal linux-3.0/fs/nfs/nfs3xdr.c linux-3.0-vs2.3.1-pre7/fs/nfs/nfs3xdr.c
7394 --- linux-3.0/fs/nfs/nfs3xdr.c  2011-03-15 18:07:32.000000000 +0100
7395 +++ linux-3.0-vs2.3.1-pre7/fs/nfs/nfs3xdr.c     2011-06-10 22:11:24.000000000 +0200
7396 @@ -20,6 +20,7 @@
7397  #include <linux/nfs3.h>
7398  #include <linux/nfs_fs.h>
7399  #include <linux/nfsacl.h>
7400 +#include <linux/vs_tag.h>
7401  #include "internal.h"
7402
7403  #define NFSDBG_FACILITY                NFSDBG_XDR
7404 @@ -562,7 +563,8 @@ static __be32 *xdr_decode_nfstime3(__be3
7405   *             set_mtime       mtime;
7406   *     };
7407   */
7408 -static void encode_sattr3(struct xdr_stream *xdr, const struct iattr *attr)
7409 +static void encode_sattr3(struct xdr_stream *xdr,
7410 +       const struct iattr *attr, int tag)
7411  {
7412         u32 nbytes;
7413         __be32 *p;
7414 @@ -594,15 +596,19 @@ static void encode_sattr3(struct xdr_str
7415         } else
7416                 *p++ = xdr_zero;
7417
7418 -       if (attr->ia_valid & ATTR_UID) {
7419 +       if (attr->ia_valid & ATTR_UID ||
7420 +               (tag && (attr->ia_valid & ATTR_TAG))) {
7421                 *p++ = xdr_one;
7422 -               *p++ = cpu_to_be32(attr->ia_uid);
7423 +               *p++ = cpu_to_be32(TAGINO_UID(tag,
7424 +                       attr->ia_uid, attr->ia_tag));
7425         } else
7426                 *p++ = xdr_zero;
7427
7428 -       if (attr->ia_valid & ATTR_GID) {
7429 +       if (attr->ia_valid & ATTR_GID ||
7430 +               (tag && (attr->ia_valid & ATTR_TAG))) {
7431                 *p++ = xdr_one;
7432 -               *p++ = cpu_to_be32(attr->ia_gid);
7433 +               *p++ = cpu_to_be32(TAGINO_GID(tag,
7434 +                       attr->ia_gid, attr->ia_tag));
7435         } else
7436                 *p++ = xdr_zero;
7437
7438 @@ -878,7 +884,7 @@ static void nfs3_xdr_enc_setattr3args(st
7439                                       const struct nfs3_sattrargs *args)
7440  {
7441         encode_nfs_fh3(xdr, args->fh);
7442 -       encode_sattr3(xdr, args->sattr);
7443 +       encode_sattr3(xdr, args->sattr, req->rq_task->tk_client->cl_tag);
7444         encode_sattrguard3(xdr, args);
7445  }
7446
7447 @@ -1028,13 +1034,13 @@ static void nfs3_xdr_enc_write3args(stru
7448   *     };
7449   */
7450  static void encode_createhow3(struct xdr_stream *xdr,
7451 -                             const struct nfs3_createargs *args)
7452 +       const struct nfs3_createargs *args, int tag)
7453  {
7454         encode_uint32(xdr, args->createmode);
7455         switch (args->createmode) {
7456         case NFS3_CREATE_UNCHECKED:
7457         case NFS3_CREATE_GUARDED:
7458 -               encode_sattr3(xdr, args->sattr);
7459 +               encode_sattr3(xdr, args->sattr, tag);
7460                 break;
7461         case NFS3_CREATE_EXCLUSIVE:
7462                 encode_createverf3(xdr, args->verifier);
7463 @@ -1049,7 +1055,7 @@ static void nfs3_xdr_enc_create3args(str
7464                                      const struct nfs3_createargs *args)
7465  {
7466         encode_diropargs3(xdr, args->fh, args->name, args->len);
7467 -       encode_createhow3(xdr, args);
7468 +       encode_createhow3(xdr, args, req->rq_task->tk_client->cl_tag);
7469  }
7470
7471  /*
7472 @@ -1065,7 +1071,7 @@ static void nfs3_xdr_enc_mkdir3args(stru
7473                                     const struct nfs3_mkdirargs *args)
7474  {
7475         encode_diropargs3(xdr, args->fh, args->name, args->len);
7476 -       encode_sattr3(xdr, args->sattr);
7477 +       encode_sattr3(xdr, args->sattr, req->rq_task->tk_client->cl_tag);
7478  }
7479
7480  /*
7481 @@ -1082,9 +1088,9 @@ static void nfs3_xdr_enc_mkdir3args(stru
7482   *     };
7483   */
7484  static void encode_symlinkdata3(struct xdr_stream *xdr,
7485 -                               const struct nfs3_symlinkargs *args)
7486 +       const struct nfs3_symlinkargs *args, int tag)
7487  {
7488 -       encode_sattr3(xdr, args->sattr);
7489 +       encode_sattr3(xdr, args->sattr, tag);
7490         encode_nfspath3(xdr, args->pages, args->pathlen);
7491  }
7492
7493 @@ -1093,7 +1099,7 @@ static void nfs3_xdr_enc_symlink3args(st
7494                                       const struct nfs3_symlinkargs *args)
7495  {
7496         encode_diropargs3(xdr, args->fromfh, args->fromname, args->fromlen);
7497 -       encode_symlinkdata3(xdr, args);
7498 +       encode_symlinkdata3(xdr, args, req->rq_task->tk_client->cl_tag);
7499  }
7500
7501  /*
7502 @@ -1121,24 +1127,24 @@ static void nfs3_xdr_enc_symlink3args(st
7503   *     };
7504   */
7505  static void encode_devicedata3(struct xdr_stream *xdr,
7506 -                              const struct nfs3_mknodargs *args)
7507 +       const struct nfs3_mknodargs *args, int tag)
7508  {
7509 -       encode_sattr3(xdr, args->sattr);
7510 +       encode_sattr3(xdr, args->sattr, tag);
7511         encode_specdata3(xdr, args->rdev);
7512  }
7513
7514  static void encode_mknoddata3(struct xdr_stream *xdr,
7515 -                             const struct nfs3_mknodargs *args)
7516 +       const struct nfs3_mknodargs *args, int tag)
7517  {
7518         encode_ftype3(xdr, args->type);
7519         switch (args->type) {
7520         case NF3CHR:
7521         case NF3BLK:
7522 -               encode_devicedata3(xdr, args);
7523 +               encode_devicedata3(xdr, args, tag);
7524                 break;
7525         case NF3SOCK:
7526         case NF3FIFO:
7527 -               encode_sattr3(xdr, args->sattr);
7528 +               encode_sattr3(xdr, args->sattr, tag);
7529                 break;
7530         case NF3REG:
7531         case NF3DIR:
7532 @@ -1153,7 +1159,7 @@ static void nfs3_xdr_enc_mknod3args(stru
7533                                     const struct nfs3_mknodargs *args)
7534  {
7535         encode_diropargs3(xdr, args->fh, args->name, args->len);
7536 -       encode_mknoddata3(xdr, args);
7537 +       encode_mknoddata3(xdr, args, req->rq_task->tk_client->cl_tag);
7538  }
7539
7540  /*
7541 diff -NurpP --minimal linux-3.0/fs/nfs/super.c linux-3.0-vs2.3.1-pre7/fs/nfs/super.c
7542 --- linux-3.0/fs/nfs/super.c    2011-07-22 11:18:05.000000000 +0200
7543 +++ linux-3.0-vs2.3.1-pre7/fs/nfs/super.c       2011-06-10 22:11:24.000000000 +0200
7544 @@ -53,6 +53,7 @@
7545  #include <linux/nfs_xdr.h>
7546  #include <linux/magic.h>
7547  #include <linux/parser.h>
7548 +#include <linux/vs_tag.h>
7549
7550  #include <asm/system.h>
7551  #include <asm/uaccess.h>
7552 @@ -87,6 +88,7 @@ enum {
7553         Opt_sharecache, Opt_nosharecache,
7554         Opt_resvport, Opt_noresvport,
7555         Opt_fscache, Opt_nofscache,
7556 +       Opt_tag, Opt_notag,
7557
7558         /* Mount options that take integer arguments */
7559         Opt_port,
7560 @@ -100,6 +102,7 @@ enum {
7561         Opt_mountvers,
7562         Opt_nfsvers,
7563         Opt_minorversion,
7564 +       Opt_tagid,
7565
7566         /* Mount options that take string arguments */
7567         Opt_sec, Opt_proto, Opt_mountproto, Opt_mounthost,
7568 @@ -180,6 +183,10 @@ static const match_table_t nfs_mount_opt
7569         { Opt_fscache_uniq, "fsc=%s" },
7570         { Opt_local_lock, "local_lock=%s" },
7571
7572 +       { Opt_tag, "tag" },
7573 +       { Opt_notag, "notag" },
7574 +       { Opt_tagid, "tagid=%u" },
7575 +
7576         { Opt_err, NULL }
7577  };
7578
7579 @@ -650,6 +657,7 @@ static void nfs_show_mount_options(struc
7580                 { NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" },
7581                 { NFS_MOUNT_UNSHARED, ",nosharecache", "" },
7582                 { NFS_MOUNT_NORESVPORT, ",noresvport", "" },
7583 +               { NFS_MOUNT_TAGGED, ",tag", "" },
7584                 { 0, NULL, NULL }
7585         };
7586         const struct proc_nfs_info *nfs_infop;
7587 @@ -1198,6 +1206,14 @@ static int nfs_parse_mount_options(char
7588                         kfree(mnt->fscache_uniq);
7589                         mnt->fscache_uniq = NULL;
7590                         break;
7591 +#ifndef CONFIG_TAGGING_NONE
7592 +               case Opt_tag:
7593 +                       mnt->flags |= NFS_MOUNT_TAGGED;
7594 +                       break;
7595 +               case Opt_notag:
7596 +                       mnt->flags &= ~NFS_MOUNT_TAGGED;
7597 +                       break;
7598 +#endif
7599
7600                 /*
7601                  * options that take numeric values
7602 @@ -1304,6 +1320,12 @@ static int nfs_parse_mount_options(char
7603                                 goto out_invalid_value;
7604                         mnt->minorversion = option;
7605                         break;
7606 +#ifdef CONFIG_PROPAGATE
7607 +               case Opt_tagid:
7608 +                       /* use args[0] */
7609 +                       nfs_data.flags |= NFS_MOUNT_TAGGED;
7610 +                       break;
7611 +#endif
7612
7613                 /*
7614                  * options that take text values
7615 diff -NurpP --minimal linux-3.0/fs/nfsd/auth.c linux-3.0-vs2.3.1-pre7/fs/nfsd/auth.c
7616 --- linux-3.0/fs/nfsd/auth.c    2010-02-25 11:52:05.000000000 +0100
7617 +++ linux-3.0-vs2.3.1-pre7/fs/nfsd/auth.c       2011-06-10 22:11:24.000000000 +0200
7618 @@ -1,6 +1,7 @@
7619  /* Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de> */
7620
7621  #include <linux/sched.h>
7622 +#include <linux/vs_tag.h>
7623  #include "nfsd.h"
7624  #include "auth.h"
7625
7626 @@ -36,6 +37,9 @@ int nfsd_setuser(struct svc_rqst *rqstp,
7627
7628         new->fsuid = rqstp->rq_cred.cr_uid;
7629         new->fsgid = rqstp->rq_cred.cr_gid;
7630 +       /* FIXME: this desperately needs a tag :)
7631 +       new->xid = (xid_t)INOTAG_TAG(DX_TAG_NFSD, cred.cr_uid, cred.cr_gid, 0);
7632 +                       */
7633
7634         rqgi = rqstp->rq_cred.cr_group_info;
7635
7636 diff -NurpP --minimal linux-3.0/fs/nfsd/nfs3xdr.c linux-3.0-vs2.3.1-pre7/fs/nfsd/nfs3xdr.c
7637 --- linux-3.0/fs/nfsd/nfs3xdr.c 2011-07-22 11:18:05.000000000 +0200
7638 +++ linux-3.0-vs2.3.1-pre7/fs/nfsd/nfs3xdr.c    2011-06-10 22:11:24.000000000 +0200
7639 @@ -7,6 +7,7 @@
7640   */
7641
7642  #include <linux/namei.h>
7643 +#include <linux/vs_tag.h>
7644  #include "xdr3.h"
7645  #include "auth.h"
7646
7647 @@ -95,6 +96,8 @@ static __be32 *
7648  decode_sattr3(__be32 *p, struct iattr *iap)
7649  {
7650         u32     tmp;
7651 +       uid_t   uid = 0;
7652 +       gid_t   gid = 0;
7653
7654         iap->ia_valid = 0;
7655
7656 @@ -104,12 +107,15 @@ decode_sattr3(__be32 *p, struct iattr *i
7657         }
7658         if (*p++) {
7659                 iap->ia_valid |= ATTR_UID;
7660 -               iap->ia_uid = ntohl(*p++);
7661 +               uid = ntohl(*p++);
7662         }
7663         if (*p++) {
7664                 iap->ia_valid |= ATTR_GID;
7665 -               iap->ia_gid = ntohl(*p++);
7666 +               gid = ntohl(*p++);
7667         }
7668 +       iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
7669 +       iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
7670 +       iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
7671         if (*p++) {
7672                 u64     newsize;
7673
7674 @@ -165,8 +171,12 @@ encode_fattr3(struct svc_rqst *rqstp, __
7675         *p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]);
7676         *p++ = htonl((u32) stat->mode);
7677         *p++ = htonl((u32) stat->nlink);
7678 -       *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
7679 -       *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
7680 +       *p++ = htonl((u32) nfsd_ruid(rqstp,
7681 +               TAGINO_UID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
7682 +               stat->uid, stat->tag)));
7683 +       *p++ = htonl((u32) nfsd_rgid(rqstp,
7684 +               TAGINO_GID(0 /* FIXME: DX_TAG(dentry->d_inode) */,
7685 +               stat->gid, stat->tag)));
7686         if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) {
7687                 p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN);
7688         } else {
7689 diff -NurpP --minimal linux-3.0/fs/nfsd/nfs4xdr.c linux-3.0-vs2.3.1-pre7/fs/nfsd/nfs4xdr.c
7690 --- linux-3.0/fs/nfsd/nfs4xdr.c 2011-07-22 11:18:05.000000000 +0200
7691 +++ linux-3.0-vs2.3.1-pre7/fs/nfsd/nfs4xdr.c    2011-06-10 22:11:24.000000000 +0200
7692 @@ -45,6 +45,7 @@
7693  #include <linux/statfs.h>
7694  #include <linux/utsname.h>
7695  #include <linux/sunrpc/svcauth_gss.h>
7696 +#include <linux/vs_tag.h>
7697
7698  #include "idmap.h"
7699  #include "acl.h"
7700 @@ -2090,14 +2091,18 @@ out_acl:
7701                 WRITE32(stat.nlink);
7702         }
7703         if (bmval1 & FATTR4_WORD1_OWNER) {
7704 -               status = nfsd4_encode_user(rqstp, stat.uid, &p, &buflen);
7705 +               status = nfsd4_encode_user(rqstp,
7706 +                       TAGINO_UID(DX_TAG(dentry->d_inode),
7707 +                       stat.uid, stat.tag), &p, &buflen);
7708                 if (status == nfserr_resource)
7709                         goto out_resource;
7710                 if (status)
7711                         goto out;
7712         }
7713         if (bmval1 & FATTR4_WORD1_OWNER_GROUP) {
7714 -               status = nfsd4_encode_group(rqstp, stat.gid, &p, &buflen);
7715 +               status = nfsd4_encode_group(rqstp,
7716 +                       TAGINO_GID(DX_TAG(dentry->d_inode),
7717 +                       stat.gid, stat.tag), &p, &buflen);
7718                 if (status == nfserr_resource)
7719                         goto out_resource;
7720                 if (status)
7721 diff -NurpP --minimal linux-3.0/fs/nfsd/nfsxdr.c linux-3.0-vs2.3.1-pre7/fs/nfsd/nfsxdr.c
7722 --- linux-3.0/fs/nfsd/nfsxdr.c  2011-05-22 16:17:53.000000000 +0200
7723 +++ linux-3.0-vs2.3.1-pre7/fs/nfsd/nfsxdr.c     2011-06-10 22:11:24.000000000 +0200
7724 @@ -6,6 +6,7 @@
7725
7726  #include "xdr.h"
7727  #include "auth.h"
7728 +#include <linux/vs_tag.h>
7729
7730  #define NFSDDBG_FACILITY               NFSDDBG_XDR
7731
7732 @@ -88,6 +89,8 @@ static __be32 *
7733  decode_sattr(__be32 *p, struct iattr *iap)
7734  {
7735         u32     tmp, tmp1;
7736 +       uid_t   uid = 0;
7737 +       gid_t   gid = 0;
7738
7739         iap->ia_valid = 0;
7740
7741 @@ -101,12 +104,15 @@ decode_sattr(__be32 *p, struct iattr *ia
7742         }
7743         if ((tmp = ntohl(*p++)) != (u32)-1) {
7744                 iap->ia_valid |= ATTR_UID;
7745 -               iap->ia_uid = tmp;
7746 +               uid = tmp;
7747         }
7748         if ((tmp = ntohl(*p++)) != (u32)-1) {
7749                 iap->ia_valid |= ATTR_GID;
7750 -               iap->ia_gid = tmp;
7751 +               gid = tmp;
7752         }
7753 +       iap->ia_uid = INOTAG_UID(DX_TAG_NFSD, uid, gid);
7754 +       iap->ia_gid = INOTAG_GID(DX_TAG_NFSD, uid, gid);
7755 +       iap->ia_tag = INOTAG_TAG(DX_TAG_NFSD, uid, gid, 0);
7756         if ((tmp = ntohl(*p++)) != (u32)-1) {
7757                 iap->ia_valid |= ATTR_SIZE;
7758                 iap->ia_size = tmp;
7759 @@ -151,8 +157,10 @@ encode_fattr(struct svc_rqst *rqstp, __b
7760         *p++ = htonl(nfs_ftypes[type >> 12]);
7761         *p++ = htonl((u32) stat->mode);
7762         *p++ = htonl((u32) stat->nlink);
7763 -       *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid));
7764 -       *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid));
7765 +       *p++ = htonl((u32) nfsd_ruid(rqstp,
7766 +               TAGINO_UID(DX_TAG(dentry->d_inode), stat->uid, stat->tag)));
7767 +       *p++ = htonl((u32) nfsd_rgid(rqstp,
7768 +               TAGINO_GID(DX_TAG(dentry->d_inode), stat->gid, stat->tag)));
7769
7770         if (S_ISLNK(type) && stat->size > NFS_MAXPATHLEN) {
7771                 *p++ = htonl(NFS_MAXPATHLEN);
7772 diff -NurpP --minimal linux-3.0/fs/ocfs2/dlmglue.c linux-3.0-vs2.3.1-pre7/fs/ocfs2/dlmglue.c
7773 --- linux-3.0/fs/ocfs2/dlmglue.c        2011-05-22 16:17:53.000000000 +0200
7774 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/dlmglue.c   2011-06-10 22:11:24.000000000 +0200
7775 @@ -2041,6 +2041,7 @@ static void __ocfs2_stuff_meta_lvb(struc
7776         lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters);
7777         lvb->lvb_iuid      = cpu_to_be32(inode->i_uid);
7778         lvb->lvb_igid      = cpu_to_be32(inode->i_gid);
7779 +       lvb->lvb_itag      = cpu_to_be16(inode->i_tag);
7780         lvb->lvb_imode     = cpu_to_be16(inode->i_mode);
7781         lvb->lvb_inlink    = cpu_to_be16(inode->i_nlink);
7782         lvb->lvb_iatime_packed  =
7783 @@ -2091,6 +2092,7 @@ static void ocfs2_refresh_inode_from_lvb
7784
7785         inode->i_uid     = be32_to_cpu(lvb->lvb_iuid);
7786         inode->i_gid     = be32_to_cpu(lvb->lvb_igid);
7787 +       inode->i_tag     = be16_to_cpu(lvb->lvb_itag);
7788         inode->i_mode    = be16_to_cpu(lvb->lvb_imode);
7789         inode->i_nlink   = be16_to_cpu(lvb->lvb_inlink);
7790         ocfs2_unpack_timespec(&inode->i_atime,
7791 diff -NurpP --minimal linux-3.0/fs/ocfs2/dlmglue.h linux-3.0-vs2.3.1-pre7/fs/ocfs2/dlmglue.h
7792 --- linux-3.0/fs/ocfs2/dlmglue.h        2010-10-21 13:07:50.000000000 +0200
7793 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/dlmglue.h   2011-06-10 22:11:24.000000000 +0200
7794 @@ -46,7 +46,8 @@ struct ocfs2_meta_lvb {
7795         __be16       lvb_inlink;
7796         __be32       lvb_iattr;
7797         __be32       lvb_igeneration;
7798 -       __be32       lvb_reserved2;
7799 +       __be16       lvb_itag;
7800 +       __be16       lvb_reserved2;
7801  };
7802
7803  #define OCFS2_QINFO_LVB_VERSION 1
7804 diff -NurpP --minimal linux-3.0/fs/ocfs2/file.c linux-3.0-vs2.3.1-pre7/fs/ocfs2/file.c
7805 --- linux-3.0/fs/ocfs2/file.c   2011-07-22 11:18:06.000000000 +0200
7806 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/file.c      2011-06-10 22:43:33.000000000 +0200
7807 @@ -1111,7 +1111,7 @@ int ocfs2_setattr(struct dentry *dentry,
7808                 attr->ia_valid &= ~ATTR_SIZE;
7809
7810  #define OCFS2_VALID_ATTRS (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME | ATTR_SIZE \
7811 -                          | ATTR_GID | ATTR_UID | ATTR_MODE)
7812 +                          | ATTR_GID | ATTR_UID | ATTR_TAG | ATTR_MODE)
7813         if (!(attr->ia_valid & OCFS2_VALID_ATTRS))
7814                 return 0;
7815
7816 diff -NurpP --minimal linux-3.0/fs/ocfs2/inode.c linux-3.0-vs2.3.1-pre7/fs/ocfs2/inode.c
7817 --- linux-3.0/fs/ocfs2/inode.c  2011-05-22 16:17:53.000000000 +0200
7818 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/inode.c     2011-06-10 22:11:24.000000000 +0200
7819 @@ -28,6 +28,7 @@
7820  #include <linux/highmem.h>
7821  #include <linux/pagemap.h>
7822  #include <linux/quotaops.h>
7823 +#include <linux/vs_tag.h>
7824
7825  #include <asm/byteorder.h>
7826
7827 @@ -78,11 +79,13 @@ void ocfs2_set_inode_flags(struct inode
7828  {
7829         unsigned int flags = OCFS2_I(inode)->ip_attr;
7830
7831 -       inode->i_flags &= ~(S_IMMUTABLE |
7832 +       inode->i_flags &= ~(S_IMMUTABLE | S_IXUNLINK |
7833                 S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
7834
7835         if (flags & OCFS2_IMMUTABLE_FL)
7836                 inode->i_flags |= S_IMMUTABLE;
7837 +       if (flags & OCFS2_IXUNLINK_FL)
7838 +               inode->i_flags |= S_IXUNLINK;
7839
7840         if (flags & OCFS2_SYNC_FL)
7841                 inode->i_flags |= S_SYNC;
7842 @@ -92,25 +95,44 @@ void ocfs2_set_inode_flags(struct inode
7843                 inode->i_flags |= S_NOATIME;
7844         if (flags & OCFS2_DIRSYNC_FL)
7845                 inode->i_flags |= S_DIRSYNC;
7846 +
7847 +       inode->i_vflags &= ~(V_BARRIER | V_COW);
7848 +
7849 +       if (flags & OCFS2_BARRIER_FL)
7850 +               inode->i_vflags |= V_BARRIER;
7851 +       if (flags & OCFS2_COW_FL)
7852 +               inode->i_vflags |= V_COW;
7853  }
7854
7855  /* Propagate flags from i_flags to OCFS2_I(inode)->ip_attr */
7856  void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi)
7857  {
7858         unsigned int flags = oi->vfs_inode.i_flags;
7859 +       unsigned int vflags = oi->vfs_inode.i_vflags;
7860 +
7861 +       oi->ip_attr &= ~(OCFS2_SYNC_FL | OCFS2_APPEND_FL |
7862 +                       OCFS2_IMMUTABLE_FL | OCFS2_IXUNLINK_FL |
7863 +                       OCFS2_NOATIME_FL | OCFS2_DIRSYNC_FL |
7864 +                       OCFS2_BARRIER_FL | OCFS2_COW_FL);
7865 +
7866 +       if (flags & S_IMMUTABLE)
7867 +               oi->ip_attr |= OCFS2_IMMUTABLE_FL;
7868 +       if (flags & S_IXUNLINK)
7869 +               oi->ip_attr |= OCFS2_IXUNLINK_FL;
7870
7871 -       oi->ip_attr &= ~(OCFS2_SYNC_FL|OCFS2_APPEND_FL|
7872 -                       OCFS2_IMMUTABLE_FL|OCFS2_NOATIME_FL|OCFS2_DIRSYNC_FL);
7873         if (flags & S_SYNC)
7874                 oi->ip_attr |= OCFS2_SYNC_FL;
7875         if (flags & S_APPEND)
7876                 oi->ip_attr |= OCFS2_APPEND_FL;
7877 -       if (flags & S_IMMUTABLE)
7878 -               oi->ip_attr |= OCFS2_IMMUTABLE_FL;
7879         if (flags & S_NOATIME)
7880                 oi->ip_attr |= OCFS2_NOATIME_FL;
7881         if (flags & S_DIRSYNC)
7882                 oi->ip_attr |= OCFS2_DIRSYNC_FL;
7883 +
7884 +       if (vflags & V_BARRIER)
7885 +               oi->ip_attr |= OCFS2_BARRIER_FL;
7886 +       if (vflags & V_COW)
7887 +               oi->ip_attr |= OCFS2_COW_FL;
7888  }
7889
7890  struct inode *ocfs2_ilookup(struct super_block *sb, u64 blkno)
7891 @@ -241,6 +263,8 @@ void ocfs2_populate_inode(struct inode *
7892         struct super_block *sb;
7893         struct ocfs2_super *osb;
7894         int use_plocks = 1;
7895 +       uid_t uid;
7896 +       gid_t gid;
7897
7898         sb = inode->i_sb;
7899         osb = OCFS2_SB(sb);
7900 @@ -269,8 +293,12 @@ void ocfs2_populate_inode(struct inode *
7901         inode->i_generation = le32_to_cpu(fe->i_generation);
7902         inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
7903         inode->i_mode = le16_to_cpu(fe->i_mode);
7904 -       inode->i_uid = le32_to_cpu(fe->i_uid);
7905 -       inode->i_gid = le32_to_cpu(fe->i_gid);
7906 +       uid = le32_to_cpu(fe->i_uid);
7907 +       gid = le32_to_cpu(fe->i_gid);
7908 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
7909 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
7910 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid,
7911 +               /* le16_to_cpu(raw_inode->i_raw_tag)i */ 0);
7912
7913         /* Fast symlinks will have i_size but no allocated clusters. */
7914         if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
7915 diff -NurpP --minimal linux-3.0/fs/ocfs2/inode.h linux-3.0-vs2.3.1-pre7/fs/ocfs2/inode.h
7916 --- linux-3.0/fs/ocfs2/inode.h  2011-01-05 21:50:26.000000000 +0100
7917 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/inode.h     2011-06-10 22:11:24.000000000 +0200
7918 @@ -151,6 +151,7 @@ struct buffer_head *ocfs2_bread(struct i
7919
7920  void ocfs2_set_inode_flags(struct inode *inode);
7921  void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi);
7922 +int ocfs2_sync_flags(struct inode *inode, int, int);
7923
7924  static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode)
7925  {
7926 diff -NurpP --minimal linux-3.0/fs/ocfs2/ioctl.c linux-3.0-vs2.3.1-pre7/fs/ocfs2/ioctl.c
7927 --- linux-3.0/fs/ocfs2/ioctl.c  2011-07-22 11:18:06.000000000 +0200
7928 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/ioctl.c     2011-06-10 22:11:24.000000000 +0200
7929 @@ -78,7 +78,41 @@ static int ocfs2_get_inode_attr(struct i
7930         return status;
7931  }
7932
7933 -static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
7934 +int ocfs2_sync_flags(struct inode *inode, int flags, int vflags)
7935 +{
7936 +       struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
7937 +       struct buffer_head *bh = NULL;
7938 +       handle_t *handle = NULL;
7939 +       int status;
7940 +
7941 +       status = ocfs2_inode_lock(inode, &bh, 1);
7942 +       if (status < 0) {
7943 +               mlog_errno(status);
7944 +               return status;
7945 +       }
7946 +       handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
7947 +       if (IS_ERR(handle)) {
7948 +               status = PTR_ERR(handle);
7949 +               mlog_errno(status);
7950 +               goto bail_unlock;
7951 +       }
7952 +
7953 +       inode->i_flags = flags;
7954 +       inode->i_vflags = vflags;
7955 +       ocfs2_get_inode_flags(OCFS2_I(inode));
7956 +
7957 +       status = ocfs2_mark_inode_dirty(handle, inode, bh);
7958 +       if (status < 0)
7959 +               mlog_errno(status);
7960 +
7961 +       ocfs2_commit_trans(osb, handle);
7962 +bail_unlock:
7963 +       ocfs2_inode_unlock(inode, 1);
7964 +       brelse(bh);
7965 +       return status;
7966 +}
7967 +
7968 +int ocfs2_set_inode_attr(struct inode *inode, unsigned flags,
7969                                 unsigned mask)
7970  {
7971         struct ocfs2_inode_info *ocfs2_inode = OCFS2_I(inode);
7972 @@ -103,6 +137,11 @@ static int ocfs2_set_inode_attr(struct i
7973         if (!S_ISDIR(inode->i_mode))
7974                 flags &= ~OCFS2_DIRSYNC_FL;
7975
7976 +       if (IS_BARRIER(inode)) {
7977 +               vxwprintk_task(1, "messing with the barrier.");
7978 +               goto bail_unlock;
7979 +       }
7980 +
7981         handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
7982         if (IS_ERR(handle)) {
7983                 status = PTR_ERR(handle);
7984 @@ -880,6 +919,7 @@ bail:
7985         return status;
7986  }
7987
7988 +
7989  long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
7990  {
7991         struct inode *inode = filp->f_path.dentry->d_inode;
7992 diff -NurpP --minimal linux-3.0/fs/ocfs2/namei.c linux-3.0-vs2.3.1-pre7/fs/ocfs2/namei.c
7993 --- linux-3.0/fs/ocfs2/namei.c  2011-05-22 16:17:53.000000000 +0200
7994 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/namei.c     2011-06-10 22:11:24.000000000 +0200
7995 @@ -41,6 +41,7 @@
7996  #include <linux/slab.h>
7997  #include <linux/highmem.h>
7998  #include <linux/quotaops.h>
7999 +#include <linux/vs_tag.h>
8000
8001  #include <cluster/masklog.h>
8002
8003 @@ -477,6 +478,7 @@ static int __ocfs2_mknod_locked(struct i
8004         struct ocfs2_dinode *fe = NULL;
8005         struct ocfs2_extent_list *fel;
8006         u16 feat;
8007 +       tag_t tag;
8008
8009         *new_fe_bh = NULL;
8010
8011 @@ -514,8 +516,11 @@ static int __ocfs2_mknod_locked(struct i
8012         fe->i_suballoc_loc = cpu_to_le64(suballoc_loc);
8013         fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
8014         fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot);
8015 -       fe->i_uid = cpu_to_le32(inode->i_uid);
8016 -       fe->i_gid = cpu_to_le32(inode->i_gid);
8017 +
8018 +       tag = dx_current_fstag(osb->sb);
8019 +       fe->i_uid = cpu_to_le32(TAGINO_UID(DX_TAG(inode), inode->i_uid, tag));
8020 +       fe->i_gid = cpu_to_le32(TAGINO_GID(DX_TAG(inode), inode->i_gid, tag));
8021 +       inode->i_tag = tag;
8022         fe->i_mode = cpu_to_le16(inode->i_mode);
8023         if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
8024                 fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
8025 diff -NurpP --minimal linux-3.0/fs/ocfs2/ocfs2.h linux-3.0-vs2.3.1-pre7/fs/ocfs2/ocfs2.h
8026 --- linux-3.0/fs/ocfs2/ocfs2.h  2011-05-22 16:17:53.000000000 +0200
8027 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/ocfs2.h     2011-06-10 22:11:24.000000000 +0200
8028 @@ -272,6 +272,7 @@ enum ocfs2_mount_options
8029                                                      writes */
8030         OCFS2_MOUNT_HB_NONE = 1 << 13, /* No heartbeat */
8031         OCFS2_MOUNT_HB_GLOBAL = 1 << 14, /* Global heartbeat */
8032 +       OCFS2_MOUNT_TAGGED = 1 << 15, /* use tagging */
8033  };
8034
8035  #define OCFS2_OSB_SOFT_RO                      0x0001
8036 diff -NurpP --minimal linux-3.0/fs/ocfs2/ocfs2_fs.h linux-3.0-vs2.3.1-pre7/fs/ocfs2/ocfs2_fs.h
8037 --- linux-3.0/fs/ocfs2/ocfs2_fs.h       2011-05-22 16:17:53.000000000 +0200
8038 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/ocfs2_fs.h  2011-06-10 22:11:24.000000000 +0200
8039 @@ -266,6 +266,11 @@
8040  #define OCFS2_TOPDIR_FL                        FS_TOPDIR_FL    /* Top of directory hierarchies*/
8041  #define OCFS2_RESERVED_FL              FS_RESERVED_FL  /* reserved for ext2 lib */
8042
8043 +#define OCFS2_IXUNLINK_FL              FS_IXUNLINK_FL  /* Immutable invert on unlink */
8044 +
8045 +#define OCFS2_BARRIER_FL               FS_BARRIER_FL   /* Barrier for chroot() */
8046 +#define OCFS2_COW_FL                   FS_COW_FL       /* Copy on Write marker */
8047 +
8048  #define OCFS2_FL_VISIBLE               FS_FL_USER_VISIBLE      /* User visible flags */
8049  #define OCFS2_FL_MODIFIABLE            FS_FL_USER_MODIFIABLE   /* User modifiable flags */
8050
8051 diff -NurpP --minimal linux-3.0/fs/ocfs2/super.c linux-3.0-vs2.3.1-pre7/fs/ocfs2/super.c
8052 --- linux-3.0/fs/ocfs2/super.c  2011-07-22 11:18:06.000000000 +0200
8053 +++ linux-3.0-vs2.3.1-pre7/fs/ocfs2/super.c     2011-06-15 02:40:14.000000000 +0200
8054 @@ -184,6 +184,7 @@ enum {
8055         Opt_coherency_full,
8056         Opt_resv_level,
8057         Opt_dir_resv_level,
8058 +       Opt_tag, Opt_notag, Opt_tagid,
8059         Opt_err,
8060  };
8061
8062 @@ -215,6 +216,9 @@ static const match_table_t tokens = {
8063         {Opt_coherency_full, "coherency=full"},
8064         {Opt_resv_level, "resv_level=%u"},
8065         {Opt_dir_resv_level, "dir_resv_level=%u"},
8066 +       {Opt_tag, "tag"},
8067 +       {Opt_notag, "notag"},
8068 +       {Opt_tagid, "tagid=%u"},
8069         {Opt_err, NULL}
8070  };
8071
8072 @@ -662,6 +666,13 @@ static int ocfs2_remount(struct super_bl
8073                 goto out;
8074         }
8075
8076 +       if ((osb->s_mount_opt & OCFS2_MOUNT_TAGGED) !=
8077 +           (parsed_options.mount_opt & OCFS2_MOUNT_TAGGED)) {
8078 +               ret = -EINVAL;
8079 +               mlog(ML_ERROR, "Cannot change tagging on remount\n");
8080 +               goto out;
8081 +       }
8082 +
8083         /* We're going to/from readonly mode. */
8084         if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY)) {
8085                 /* Disable quota accounting before remounting RO */
8086 @@ -1177,6 +1188,9 @@ static int ocfs2_fill_super(struct super
8087
8088         ocfs2_complete_mount_recovery(osb);
8089
8090 +       if (osb->s_mount_opt & OCFS2_MOUNT_TAGGED)
8091 +               sb->s_flags |= MS_TAGGED;
8092 +
8093         if (ocfs2_mount_local(osb))
8094                 snprintf(nodestr, sizeof(nodestr), "local");
8095         else
8096 @@ -1506,6 +1520,20 @@ static int ocfs2_parse_options(struct su
8097                             option < OCFS2_MAX_RESV_LEVEL)
8098                                 mopt->dir_resv_level = option;
8099                         break;
8100 +#ifndef CONFIG_TAGGING_NONE
8101 +               case Opt_tag:
8102 +                       mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
8103 +                       break;
8104 +               case Opt_notag:
8105 +                       mopt->mount_opt &= ~OCFS2_MOUNT_TAGGED;
8106 +                       break;
8107 +#endif
8108 +#ifdef CONFIG_PROPAGATE
8109 +               case Opt_tagid:
8110 +                       /* use args[0] */
8111 +                       mopt->mount_opt |= OCFS2_MOUNT_TAGGED;
8112 +                       break;
8113 +#endif
8114                 default:
8115                         mlog(ML_ERROR,
8116                              "Unrecognized mount option \"%s\" "
8117 diff -NurpP --minimal linux-3.0/fs/open.c linux-3.0-vs2.3.1-pre7/fs/open.c
8118 --- linux-3.0/fs/open.c 2011-05-22 16:17:53.000000000 +0200
8119 +++ linux-3.0-vs2.3.1-pre7/fs/open.c    2011-06-10 22:11:24.000000000 +0200
8120 @@ -30,6 +30,11 @@
8121  #include <linux/fs_struct.h>
8122  #include <linux/ima.h>
8123  #include <linux/dnotify.h>
8124 +#include <linux/vs_base.h>
8125 +#include <linux/vs_limit.h>
8126 +#include <linux/vs_tag.h>
8127 +#include <linux/vs_cowbl.h>
8128 +#include <linux/vserver/dlimit.h>
8129
8130  #include "internal.h"
8131
8132 @@ -494,6 +499,12 @@ SYSCALL_DEFINE3(fchmodat, int, dfd, cons
8133         error = user_path_at(dfd, filename, LOOKUP_FOLLOW, &path);
8134         if (error)
8135                 goto out;
8136 +
8137 +#ifdef CONFIG_VSERVER_COWBL
8138 +       error = cow_check_and_break(&path);
8139 +       if (error)
8140 +               goto dput_and_out;
8141 +#endif
8142         inode = path.dentry->d_inode;
8143
8144         error = mnt_want_write(path.mnt);
8145 @@ -531,11 +542,11 @@ static int chown_common(struct path *pat
8146         newattrs.ia_valid =  ATTR_CTIME;
8147         if (user != (uid_t) -1) {
8148                 newattrs.ia_valid |= ATTR_UID;
8149 -               newattrs.ia_uid = user;
8150 +               newattrs.ia_uid = dx_map_uid(user);
8151         }
8152         if (group != (gid_t) -1) {
8153                 newattrs.ia_valid |= ATTR_GID;
8154 -               newattrs.ia_gid = group;
8155 +               newattrs.ia_gid = dx_map_gid(group);
8156         }
8157         if (!S_ISDIR(inode->i_mode))
8158                 newattrs.ia_valid |=
8159 @@ -560,6 +571,10 @@ SYSCALL_DEFINE3(chown, const char __user
8160         error = mnt_want_write(path.mnt);
8161         if (error)
8162                 goto out_release;
8163 +#ifdef CONFIG_VSERVER_COWBL
8164 +       error = cow_check_and_break(&path);
8165 +       if (!error)
8166 +#endif
8167         error = chown_common(&path, user, group);
8168         mnt_drop_write(path.mnt);
8169  out_release:
8170 @@ -587,6 +602,10 @@ SYSCALL_DEFINE5(fchownat, int, dfd, cons
8171         error = mnt_want_write(path.mnt);
8172         if (error)
8173                 goto out_release;
8174 +#ifdef CONFIG_VSERVER_COWBL
8175 +       error = cow_check_and_break(&path);
8176 +       if (!error)
8177 +#endif
8178         error = chown_common(&path, user, group);
8179         mnt_drop_write(path.mnt);
8180  out_release:
8181 @@ -606,6 +625,10 @@ SYSCALL_DEFINE3(lchown, const char __use
8182         error = mnt_want_write(path.mnt);
8183         if (error)
8184                 goto out_release;
8185 +#ifdef CONFIG_VSERVER_COWBL
8186 +       error = cow_check_and_break(&path);
8187 +       if (!error)
8188 +#endif
8189         error = chown_common(&path, user, group);
8190         mnt_drop_write(path.mnt);
8191  out_release:
8192 @@ -857,6 +880,7 @@ static void __put_unused_fd(struct files
8193         __FD_CLR(fd, fdt->open_fds);
8194         if (fd < files->next_fd)
8195                 files->next_fd = fd;
8196 +       vx_openfd_dec(fd);
8197  }
8198
8199  void put_unused_fd(unsigned int fd)
8200 diff -NurpP --minimal linux-3.0/fs/proc/array.c linux-3.0-vs2.3.1-pre7/fs/proc/array.c
8201 --- linux-3.0/fs/proc/array.c   2011-07-22 11:18:06.000000000 +0200
8202 +++ linux-3.0-vs2.3.1-pre7/fs/proc/array.c      2011-06-10 22:11:24.000000000 +0200
8203 @@ -81,6 +81,8 @@
8204  #include <linux/pid_namespace.h>
8205  #include <linux/ptrace.h>
8206  #include <linux/tracehook.h>
8207 +#include <linux/vs_context.h>
8208 +#include <linux/vs_network.h>
8209
8210  #include <asm/pgtable.h>
8211  #include <asm/processor.h>
8212 @@ -170,6 +172,9 @@ static inline void task_state(struct seq
8213         rcu_read_lock();
8214         ppid = pid_alive(p) ?
8215                 task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
8216 +       if (unlikely(vx_current_initpid(p->pid)))
8217 +               ppid = 0;
8218 +
8219         tpid = 0;
8220         if (pid_alive(p)) {
8221                 struct task_struct *tracer = tracehook_tracer_task(p);
8222 @@ -287,7 +292,7 @@ static inline void task_sig(struct seq_f
8223  }
8224
8225  static void render_cap_t(struct seq_file *m, const char *header,
8226 -                       kernel_cap_t *a)
8227 +                       struct vx_info *vxi, kernel_cap_t *a)
8228  {
8229         unsigned __capi;
8230
8231 @@ -312,10 +317,11 @@ static inline void task_cap(struct seq_f
8232         cap_bset        = cred->cap_bset;
8233         rcu_read_unlock();
8234
8235 -       render_cap_t(m, "CapInh:\t", &cap_inheritable);
8236 -       render_cap_t(m, "CapPrm:\t", &cap_permitted);
8237 -       render_cap_t(m, "CapEff:\t", &cap_effective);
8238 -       render_cap_t(m, "CapBnd:\t", &cap_bset);
8239 +       /* FIXME: maybe move the p->vx_info masking to __task_cred() ? */
8240 +       render_cap_t(m, "CapInh:\t", p->vx_info, &cap_inheritable);
8241 +       render_cap_t(m, "CapPrm:\t", p->vx_info, &cap_permitted);
8242 +       render_cap_t(m, "CapEff:\t", p->vx_info, &cap_effective);
8243 +       render_cap_t(m, "CapBnd:\t", p->vx_info, &cap_bset);
8244  }
8245
8246  static inline void task_context_switch_counts(struct seq_file *m,
8247 @@ -337,6 +343,42 @@ static void task_cpus_allowed(struct seq
8248         seq_putc(m, '\n');
8249  }
8250
8251 +int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
8252 +                       struct pid *pid, struct task_struct *task)
8253 +{
8254 +       seq_printf(m,   "Proxy:\t%p(%c)\n"
8255 +                       "Count:\t%u\n"
8256 +                       "uts:\t%p(%c)\n"
8257 +                       "ipc:\t%p(%c)\n"
8258 +                       "mnt:\t%p(%c)\n"
8259 +                       "pid:\t%p(%c)\n"
8260 +                       "net:\t%p(%c)\n",
8261 +                       task->nsproxy,
8262 +                       (task->nsproxy == init_task.nsproxy ? 'I' : '-'),
8263 +                       atomic_read(&task->nsproxy->count),
8264 +                       task->nsproxy->uts_ns,
8265 +                       (task->nsproxy->uts_ns == init_task.nsproxy->uts_ns ? 'I' : '-'),
8266 +                       task->nsproxy->ipc_ns,
8267 +                       (task->nsproxy->ipc_ns == init_task.nsproxy->ipc_ns ? 'I' : '-'),
8268 +                       task->nsproxy->mnt_ns,
8269 +                       (task->nsproxy->mnt_ns == init_task.nsproxy->mnt_ns ? 'I' : '-'),
8270 +                       task->nsproxy->pid_ns,
8271 +                       (task->nsproxy->pid_ns == init_task.nsproxy->pid_ns ? 'I' : '-'),
8272 +                       task->nsproxy->net_ns,
8273 +                       (task->nsproxy->net_ns == init_task.nsproxy->net_ns ? 'I' : '-'));
8274 +       return 0;
8275 +}
8276 +
8277 +void task_vs_id(struct seq_file *m, struct task_struct *task)
8278 +{
8279 +       if (task_vx_flags(task, VXF_HIDE_VINFO, 0))
8280 +               return;
8281 +
8282 +       seq_printf(m, "VxID: %d\n", vx_task_xid(task));
8283 +       seq_printf(m, "NxID: %d\n", nx_task_nid(task));
8284 +}
8285 +
8286 +
8287  int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
8288                         struct pid *pid, struct task_struct *task)
8289  {
8290 @@ -353,6 +395,7 @@ int proc_pid_status(struct seq_file *m,
8291         task_cap(m, task);
8292         task_cpus_allowed(m, task);
8293         cpuset_task_status_allowed(m, task);
8294 +       task_vs_id(m, task);
8295         task_context_switch_counts(m, task);
8296         return 0;
8297  }
8298 @@ -462,6 +505,17 @@ static int do_task_stat(struct seq_file
8299         /* convert nsec -> ticks */
8300         start_time = nsec_to_clock_t(start_time);
8301
8302 +       /* fixup start time for virt uptime */
8303 +       if (vx_flags(VXF_VIRT_UPTIME, 0)) {
8304 +               unsigned long long bias =
8305 +                       current->vx_info->cvirt.bias_clock;
8306 +
8307 +               if (start_time > bias)
8308 +                       start_time -= bias;
8309 +               else
8310 +                       start_time = 0;
8311 +       }
8312 +
8313         seq_printf(m, "%d (%s) %c %d %d %d %d %d %u %lu \
8314  %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \
8315  %lu %lu %lu %lu %lu %lu %lu %lu %d %d %u %u %llu %lu %ld\n",
8316 diff -NurpP --minimal linux-3.0/fs/proc/base.c linux-3.0-vs2.3.1-pre7/fs/proc/base.c
8317 --- linux-3.0/fs/proc/base.c    2011-07-22 11:18:06.000000000 +0200
8318 +++ linux-3.0-vs2.3.1-pre7/fs/proc/base.c       2011-07-19 00:44:39.000000000 +0200
8319 @@ -83,6 +83,8 @@
8320  #include <linux/pid_namespace.h>
8321  #include <linux/fs_struct.h>
8322  #include <linux/slab.h>
8323 +#include <linux/vs_context.h>
8324 +#include <linux/vs_network.h>
8325  #ifdef CONFIG_HARDWALL
8326  #include <asm/hardwall.h>
8327  #endif
8328 @@ -1102,11 +1104,16 @@ static ssize_t oom_adjust_write(struct f
8329                 goto err_task_lock;
8330         }
8331
8332 -       if (oom_adjust < task->signal->oom_adj && !capable(CAP_SYS_RESOURCE)) {
8333 +       if (oom_adjust < task->signal->oom_adj &&
8334 +               !vx_capable(CAP_SYS_RESOURCE, VXC_OOM_ADJUST)) {
8335                 err = -EACCES;
8336                 goto err_sighand;
8337         }
8338
8339 +       /* prevent guest processes from circumventing the oom killer */
8340 +       if (vx_current_xid() && (oom_adjust == OOM_DISABLE))
8341 +               oom_adjust = OOM_ADJUST_MIN;
8342 +
8343         if (oom_adjust != task->signal->oom_adj) {
8344                 if (oom_adjust == OOM_DISABLE)
8345                         atomic_inc(&task->mm->oom_disable_count);
8346 @@ -1275,7 +1282,7 @@ static ssize_t proc_loginuid_write(struc
8347         ssize_t length;
8348         uid_t loginuid;
8349
8350 -       if (!capable(CAP_AUDIT_CONTROL))
8351 +       if (!vx_capable(CAP_AUDIT_CONTROL, VXC_AUDIT_CONTROL))
8352                 return -EPERM;
8353
8354         rcu_read_lock();
8355 @@ -1722,6 +1729,8 @@ struct inode *proc_pid_make_inode(struct
8356                 inode->i_gid = cred->egid;
8357                 rcu_read_unlock();
8358         }
8359 +       /* procfs is xid tagged */
8360 +       inode->i_tag = (tag_t)vx_task_xid(task);
8361         security_task_to_inode(task, inode);
8362
8363  out:
8364 @@ -1758,6 +1767,8 @@ int pid_getattr(struct vfsmount *mnt, st
8365
8366  /* dentry stuff */
8367
8368 +static unsigned name_to_int(struct dentry *dentry);
8369 +
8370  /*
8371   *     Exceptional case: normally we are not allowed to unhash a busy
8372   * directory. In this case, however, we can do it - no aliasing problems
8373 @@ -1786,6 +1797,12 @@ int pid_revalidate(struct dentry *dentry
8374         task = get_proc_task(inode);
8375
8376         if (task) {
8377 +               unsigned pid = name_to_int(dentry);
8378 +
8379 +               if (pid != ~0U && pid != vx_map_pid(task->pid)) {
8380 +                       put_task_struct(task);
8381 +                       goto drop;
8382 +               }
8383                 if ((inode->i_mode == (S_IFDIR|S_IRUGO|S_IXUGO)) ||
8384                     task_dumpable(task)) {
8385                         rcu_read_lock();
8386 @@ -1802,6 +1819,7 @@ int pid_revalidate(struct dentry *dentry
8387                 put_task_struct(task);
8388                 return 1;
8389         }
8390 +drop:
8391         d_drop(dentry);
8392         return 0;
8393  }
8394 @@ -2283,6 +2301,13 @@ static struct dentry *proc_pident_lookup
8395         if (!task)
8396                 goto out_no_task;
8397
8398 +       /* TODO: maybe we can come up with a generic approach? */
8399 +       if (task_vx_flags(task, VXF_HIDE_VINFO, 0) &&
8400 +               (dentry->d_name.len == 5) &&
8401 +               (!memcmp(dentry->d_name.name, "vinfo", 5) ||
8402 +               !memcmp(dentry->d_name.name, "ninfo", 5)))
8403 +               goto out;
8404 +
8405         /*
8406          * Yes, it does not scale. And it should not. Don't add
8407          * new entries into /proc/<tgid>/ without very good reasons.
8408 @@ -2668,7 +2693,7 @@ out_iput:
8409  static struct dentry *proc_base_lookup(struct inode *dir, struct dentry *dentry)
8410  {
8411         struct dentry *error;
8412 -       struct task_struct *task = get_proc_task(dir);
8413 +       struct task_struct *task = get_proc_task_real(dir);
8414         const struct pid_entry *p, *last;
8415
8416         error = ERR_PTR(-ENOENT);
8417 @@ -2765,6 +2790,9 @@ static int proc_pid_personality(struct s
8418  static const struct file_operations proc_task_operations;
8419  static const struct inode_operations proc_task_inode_operations;
8420
8421 +extern int proc_pid_vx_info(struct task_struct *, char *);
8422 +extern int proc_pid_nx_info(struct task_struct *, char *);
8423 +
8424  static const struct pid_entry tgid_base_stuff[] = {
8425         DIR("task",       S_IRUGO|S_IXUGO, proc_task_inode_operations, proc_task_operations),
8426         DIR("fd",         S_IRUSR|S_IXUSR, proc_fd_inode_operations, proc_fd_operations),
8427 @@ -2828,6 +2856,8 @@ static const struct pid_entry tgid_base_
8428  #ifdef CONFIG_CGROUPS
8429         REG("cgroup",  S_IRUGO, proc_cgroup_operations),
8430  #endif
8431 +       INF("vinfo",      S_IRUGO, proc_pid_vx_info),
8432 +       INF("ninfo",      S_IRUGO, proc_pid_nx_info),
8433         INF("oom_score",  S_IRUGO, proc_oom_score),
8434         REG("oom_adj",    S_IRUGO|S_IWUSR, proc_oom_adjust_operations),
8435         REG("oom_score_adj", S_IRUGO|S_IWUSR, proc_oom_score_adj_operations),
8436 @@ -2847,6 +2877,7 @@ static const struct pid_entry tgid_base_
8437  #ifdef CONFIG_HARDWALL
8438         INF("hardwall",   S_IRUGO, proc_pid_hardwall),
8439  #endif
8440 +       ONE("nsproxy",  S_IRUGO, proc_pid_nsproxy),
8441  };
8442
8443  static int proc_tgid_base_readdir(struct file * filp,
8444 @@ -3039,7 +3070,7 @@ retry:
8445         iter.task = NULL;
8446         pid = find_ge_pid(iter.tgid, ns);
8447         if (pid) {
8448 -               iter.tgid = pid_nr_ns(pid, ns);
8449 +               iter.tgid = pid_unmapped_nr_ns(pid, ns);
8450                 iter.task = pid_task(pid, PIDTYPE_PID);
8451                 /* What we to know is if the pid we have find is the
8452                  * pid of a thread_group_leader.  Testing for task
8453 @@ -3069,7 +3100,7 @@ static int proc_pid_fill_cache(struct fi
8454         struct tgid_iter iter)
8455  {
8456         char name[PROC_NUMBUF];
8457 -       int len = snprintf(name, sizeof(name), "%d", iter.tgid);
8458 +       int len = snprintf(name, sizeof(name), "%d", vx_map_tgid(iter.tgid));
8459         return proc_fill_cache(filp, dirent, filldir, name, len,
8460                                 proc_pid_instantiate, iter.task, NULL);
8461  }
8462 @@ -3086,7 +3117,7 @@ int proc_pid_readdir(struct file * filp,
8463                 goto out_no_task;
8464         nr = filp->f_pos - FIRST_PROCESS_ENTRY;
8465
8466 -       reaper = get_proc_task(filp->f_path.dentry->d_inode);
8467 +       reaper = get_proc_task_real(filp->f_path.dentry->d_inode);
8468         if (!reaper)
8469                 goto out_no_task;
8470
8471 @@ -3103,6 +3134,8 @@ int proc_pid_readdir(struct file * filp,
8472              iter.task;
8473              iter.tgid += 1, iter = next_tgid(ns, iter)) {
8474                 filp->f_pos = iter.tgid + TGID_OFFSET;
8475 +               if (!vx_proc_task_visible(iter.task))
8476 +                       continue;
8477                 if (proc_pid_fill_cache(filp, dirent, filldir, iter) < 0) {
8478                         put_task_struct(iter.task);
8479                         goto out;
8480 @@ -3256,6 +3289,8 @@ static struct dentry *proc_task_lookup(s
8481         tid = name_to_int(dentry);
8482         if (tid == ~0U)
8483                 goto out;
8484 +       if (vx_current_initpid(tid))
8485 +               goto out;
8486
8487         ns = dentry->d_sb->s_fs_info;
8488         rcu_read_lock();
8489 diff -NurpP --minimal linux-3.0/fs/proc/generic.c linux-3.0-vs2.3.1-pre7/fs/proc/generic.c
8490 --- linux-3.0/fs/proc/generic.c 2011-07-22 11:18:06.000000000 +0200
8491 +++ linux-3.0-vs2.3.1-pre7/fs/proc/generic.c    2011-06-10 22:11:24.000000000 +0200
8492 @@ -22,6 +22,7 @@
8493  #include <linux/bitops.h>
8494  #include <linux/spinlock.h>
8495  #include <linux/completion.h>
8496 +#include <linux/vserver/inode.h>
8497  #include <asm/uaccess.h>
8498
8499  #include "internal.h"
8500 @@ -424,11 +425,15 @@ struct dentry *proc_lookup_de(struct pro
8501         for (de = de->subdir; de ; de = de->next) {
8502                 if (de->namelen != dentry->d_name.len)
8503                         continue;
8504 +               if (!vx_hide_check(0, de->vx_flags))
8505 +                       continue;
8506                 if (!memcmp(dentry->d_name.name, de->name, de->namelen)) {
8507                         pde_get(de);
8508                         spin_unlock(&proc_subdir_lock);
8509                         error = -EINVAL;
8510                         inode = proc_get_inode(dir->i_sb, de);
8511 +                       /* generic proc entries belong to the host */
8512 +                       inode->i_tag = 0;
8513                         goto out_unlock;
8514                 }
8515         }
8516 @@ -506,6 +511,8 @@ int proc_readdir_de(struct proc_dir_entr
8517
8518                                 /* filldir passes info to user space */
8519                                 pde_get(de);
8520 +                               if (!vx_hide_check(0, de->vx_flags))
8521 +                                       goto skip;
8522                                 spin_unlock(&proc_subdir_lock);
8523                                 if (filldir(dirent, de->name, de->namelen, filp->f_pos,
8524                                             de->low_ino, de->mode >> 12) < 0) {
8525 @@ -513,6 +520,7 @@ int proc_readdir_de(struct proc_dir_entr
8526                                         goto out;
8527                                 }
8528                                 spin_lock(&proc_subdir_lock);
8529 +                       skip:
8530                                 filp->f_pos++;
8531                                 next = de->next;
8532                                 pde_put(de);
8533 @@ -627,6 +635,7 @@ static struct proc_dir_entry *__proc_cre
8534         ent->nlink = nlink;
8535         atomic_set(&ent->count, 1);
8536         ent->pde_users = 0;
8537 +       ent->vx_flags = IATTR_PROC_DEFAULT;
8538         spin_lock_init(&ent->pde_unload_lock);
8539         ent->pde_unload_completion = NULL;
8540         INIT_LIST_HEAD(&ent->pde_openers);
8541 @@ -650,7 +659,8 @@ struct proc_dir_entry *proc_symlink(cons
8542                                 kfree(ent->data);
8543                                 kfree(ent);
8544                                 ent = NULL;
8545 -                       }
8546 +                       } else
8547 +                               ent->vx_flags = IATTR_PROC_SYMLINK;
8548                 } else {
8549                         kfree(ent);
8550                         ent = NULL;
8551 diff -NurpP --minimal linux-3.0/fs/proc/inode.c linux-3.0-vs2.3.1-pre7/fs/proc/inode.c
8552 --- linux-3.0/fs/proc/inode.c   2011-07-22 11:18:06.000000000 +0200
8553 +++ linux-3.0-vs2.3.1-pre7/fs/proc/inode.c      2011-06-10 22:11:24.000000000 +0200
8554 @@ -442,6 +442,8 @@ struct inode *proc_get_inode(struct supe
8555                         inode->i_uid = de->uid;
8556                         inode->i_gid = de->gid;
8557                 }
8558 +               if (de->vx_flags)
8559 +                       PROC_I(inode)->vx_flags = de->vx_flags;
8560                 if (de->size)
8561                         inode->i_size = de->size;
8562                 if (de->nlink)
8563 diff -NurpP --minimal linux-3.0/fs/proc/internal.h linux-3.0-vs2.3.1-pre7/fs/proc/internal.h
8564 --- linux-3.0/fs/proc/internal.h        2011-07-22 11:18:06.000000000 +0200
8565 +++ linux-3.0-vs2.3.1-pre7/fs/proc/internal.h   2011-06-10 22:11:24.000000000 +0200
8566 @@ -10,6 +10,7 @@
8567   */
8568
8569  #include <linux/proc_fs.h>
8570 +#include <linux/vs_pid.h>
8571
8572  extern struct proc_dir_entry proc_root;
8573  #ifdef CONFIG_PROC_SYSCTL
8574 @@ -51,6 +52,9 @@ extern int proc_pid_status(struct seq_fi
8575                                 struct pid *pid, struct task_struct *task);
8576  extern int proc_pid_statm(struct seq_file *m, struct pid_namespace *ns,
8577                                 struct pid *pid, struct task_struct *task);
8578 +extern int proc_pid_nsproxy(struct seq_file *m, struct pid_namespace *ns,
8579 +                               struct pid *pid, struct task_struct *task);
8580 +
8581  extern loff_t mem_lseek(struct file *file, loff_t offset, int orig);
8582
8583  extern const struct file_operations proc_maps_operations;
8584 @@ -76,11 +80,16 @@ static inline struct pid *proc_pid(struc
8585         return PROC_I(inode)->pid;
8586  }
8587
8588 -static inline struct task_struct *get_proc_task(struct inode *inode)
8589 +static inline struct task_struct *get_proc_task_real(struct inode *inode)
8590  {
8591         return get_pid_task(proc_pid(inode), PIDTYPE_PID);
8592  }
8593
8594 +static inline struct task_struct *get_proc_task(struct inode *inode)
8595 +{
8596 +       return vx_get_proc_task(inode, proc_pid(inode));
8597 +}
8598 +
8599  static inline int proc_fd(struct inode *inode)
8600  {
8601         return PROC_I(inode)->fd;
8602 diff -NurpP --minimal linux-3.0/fs/proc/loadavg.c linux-3.0-vs2.3.1-pre7/fs/proc/loadavg.c
8603 --- linux-3.0/fs/proc/loadavg.c 2009-09-10 15:26:23.000000000 +0200
8604 +++ linux-3.0-vs2.3.1-pre7/fs/proc/loadavg.c    2011-06-10 22:11:24.000000000 +0200
8605 @@ -12,15 +12,27 @@
8606
8607  static int loadavg_proc_show(struct seq_file *m, void *v)
8608  {
8609 +       unsigned long running;
8610 +       unsigned int threads;
8611         unsigned long avnrun[3];
8612
8613         get_avenrun(avnrun, FIXED_1/200, 0);
8614
8615 +       if (vx_flags(VXF_VIRT_LOAD, 0)) {
8616 +               struct vx_info *vxi = current_vx_info();
8617 +
8618 +               running = atomic_read(&vxi->cvirt.nr_running);
8619 +               threads = atomic_read(&vxi->cvirt.nr_threads);
8620 +       } else {
8621 +               running = nr_running();
8622 +               threads = nr_threads;
8623 +       }
8624 +
8625         seq_printf(m, "%lu.%02lu %lu.%02lu %lu.%02lu %ld/%d %d\n",
8626                 LOAD_INT(avnrun[0]), LOAD_FRAC(avnrun[0]),
8627                 LOAD_INT(avnrun[1]), LOAD_FRAC(avnrun[1]),
8628                 LOAD_INT(avnrun[2]), LOAD_FRAC(avnrun[2]),
8629 -               nr_running(), nr_threads,
8630 +               running, threads,
8631                 task_active_pid_ns(current)->last_pid);
8632         return 0;
8633  }
8634 diff -NurpP --minimal linux-3.0/fs/proc/meminfo.c linux-3.0-vs2.3.1-pre7/fs/proc/meminfo.c
8635 --- linux-3.0/fs/proc/meminfo.c 2011-03-15 18:07:33.000000000 +0100
8636 +++ linux-3.0-vs2.3.1-pre7/fs/proc/meminfo.c    2011-06-10 22:11:24.000000000 +0200
8637 @@ -39,7 +39,8 @@ static int meminfo_proc_show(struct seq_
8638         allowed = ((totalram_pages - hugetlb_total_pages())
8639                 * sysctl_overcommit_ratio / 100) + total_swap_pages;
8640
8641 -       cached = global_page_state(NR_FILE_PAGES) -
8642 +       cached = vx_flags(VXF_VIRT_MEM, 0) ?
8643 +               vx_vsi_cached(&i) : global_page_state(NR_FILE_PAGES) -
8644                         total_swapcache_pages - i.bufferram;
8645         if (cached < 0)
8646                 cached = 0;
8647 diff -NurpP --minimal linux-3.0/fs/proc/root.c linux-3.0-vs2.3.1-pre7/fs/proc/root.c
8648 --- linux-3.0/fs/proc/root.c    2011-07-22 11:18:06.000000000 +0200
8649 +++ linux-3.0-vs2.3.1-pre7/fs/proc/root.c       2011-06-22 12:39:15.000000000 +0200
8650 @@ -18,9 +18,14 @@
8651  #include <linux/bitops.h>
8652  #include <linux/mount.h>
8653  #include <linux/pid_namespace.h>
8654 +#include <linux/vserver/inode.h>
8655
8656  #include "internal.h"
8657
8658 +struct proc_dir_entry *proc_virtual;
8659 +
8660 +extern void proc_vx_init(void);
8661 +
8662  static int proc_test_super(struct super_block *sb, void *data)
8663  {
8664         return sb->s_fs_info == data;
8665 @@ -125,6 +130,7 @@ void __init proc_root_init(void)
8666  #endif
8667         proc_mkdir("bus", NULL);
8668         proc_sys_init();
8669 +       proc_vx_init();
8670  }
8671
8672  static int proc_root_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat
8673 @@ -193,6 +199,7 @@ struct proc_dir_entry proc_root = {
8674         .proc_iops      = &proc_root_inode_operations,
8675         .proc_fops      = &proc_root_operations,
8676         .parent         = &proc_root,
8677 +       .vx_flags       = IATTR_ADMIN | IATTR_WATCH,
8678  };
8679
8680  int pid_ns_prepare_proc(struct pid_namespace *ns)
8681 diff -NurpP --minimal linux-3.0/fs/proc/uptime.c linux-3.0-vs2.3.1-pre7/fs/proc/uptime.c
8682 --- linux-3.0/fs/proc/uptime.c  2009-12-03 20:02:53.000000000 +0100
8683 +++ linux-3.0-vs2.3.1-pre7/fs/proc/uptime.c     2011-06-10 22:11:24.000000000 +0200
8684 @@ -4,22 +4,22 @@
8685  #include <linux/sched.h>
8686  #include <linux/seq_file.h>
8687  #include <linux/time.h>
8688 -#include <linux/kernel_stat.h>
8689 +#include <linux/vserver/cvirt.h>
8690  #include <asm/cputime.h>
8691
8692  static int uptime_proc_show(struct seq_file *m, void *v)
8693  {
8694         struct timespec uptime;
8695         struct timespec idle;
8696 -       int i;
8697 -       cputime_t idletime = cputime_zero;
8698 -
8699 -       for_each_possible_cpu(i)
8700 -               idletime = cputime64_add(idletime, kstat_cpu(i).cpustat.idle);
8701 +       cputime_t idletime = cputime_add(init_task.utime, init_task.stime);
8702
8703         do_posix_clock_monotonic_gettime(&uptime);
8704         monotonic_to_bootbased(&uptime);
8705         cputime_to_timespec(idletime, &idle);
8706 +
8707 +       if (vx_flags(VXF_VIRT_UPTIME, 0))
8708 +               vx_vsi_uptime(&uptime, &idle);
8709 +
8710         seq_printf(m, "%lu.%02lu %lu.%02lu\n",
8711                         (unsigned long) uptime.tv_sec,
8712                         (uptime.tv_nsec / (NSEC_PER_SEC / 100)),
8713 diff -NurpP --minimal linux-3.0/fs/quota/dquot.c linux-3.0-vs2.3.1-pre7/fs/quota/dquot.c
8714 --- linux-3.0/fs/quota/dquot.c  2011-07-22 11:18:06.000000000 +0200
8715 +++ linux-3.0-vs2.3.1-pre7/fs/quota/dquot.c     2011-06-10 22:11:24.000000000 +0200
8716 @@ -1548,6 +1548,9 @@ int __dquot_alloc_space(struct inode *in
8717         int reserve = flags & DQUOT_SPACE_RESERVE;
8718         int nofail = flags & DQUOT_SPACE_NOFAIL;
8719
8720 +       if ((ret = dl_alloc_space(inode, number)))
8721 +               return ret;
8722 +
8723         /*
8724          * First test before acquiring mutex - solves deadlocks when we
8725          * re-enter the quota code and are already holding the mutex
8726 @@ -1602,6 +1605,9 @@ int dquot_alloc_inode(const struct inode
8727         int cnt, ret = 0;
8728         char warntype[MAXQUOTAS];
8729
8730 +       if ((ret = dl_alloc_inode(inode)))
8731 +               return ret;
8732 +
8733         /* First test before acquiring mutex - solves deadlocks when we
8734           * re-enter the quota code and are already holding the mutex */
8735         if (!dquot_active(inode))
8736 @@ -1672,6 +1678,8 @@ void __dquot_free_space(struct inode *in
8737         char warntype[MAXQUOTAS];
8738         int reserve = flags & DQUOT_SPACE_RESERVE;
8739
8740 +       dl_free_space(inode, number);
8741 +
8742         /* First test before acquiring mutex - solves deadlocks when we
8743           * re-enter the quota code and are already holding the mutex */
8744         if (!dquot_active(inode)) {
8745 @@ -1710,6 +1718,8 @@ void dquot_free_inode(const struct inode
8746         unsigned int cnt;
8747         char warntype[MAXQUOTAS];
8748
8749 +       dl_free_inode(inode);
8750 +
8751         /* First test before acquiring mutex - solves deadlocks when we
8752           * re-enter the quota code and are already holding the mutex */
8753         if (!dquot_active(inode))
8754 diff -NurpP --minimal linux-3.0/fs/quota/quota.c linux-3.0-vs2.3.1-pre7/fs/quota/quota.c
8755 --- linux-3.0/fs/quota/quota.c  2011-03-15 18:07:34.000000000 +0100
8756 +++ linux-3.0-vs2.3.1-pre7/fs/quota/quota.c     2011-06-13 18:19:47.000000000 +0200
8757 @@ -8,6 +8,7 @@
8758  #include <linux/fs.h>
8759  #include <linux/namei.h>
8760  #include <linux/slab.h>
8761 +#include <linux/vs_context.h>
8762  #include <asm/current.h>
8763  #include <asm/uaccess.h>
8764  #include <linux/kernel.h>
8765 @@ -38,7 +39,7 @@ static int check_quotactl_permission(str
8766                         break;
8767                 /*FALLTHROUGH*/
8768         default:
8769 -               if (!capable(CAP_SYS_ADMIN))
8770 +               if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL))
8771                         return -EPERM;
8772         }
8773
8774 @@ -293,6 +294,46 @@ static int do_quotactl(struct super_bloc
8775         }
8776  }
8777
8778 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
8779 +
8780 +#include <linux/vroot.h>
8781 +#include <linux/major.h>
8782 +#include <linux/module.h>
8783 +#include <linux/kallsyms.h>
8784 +#include <linux/vserver/debug.h>
8785 +
8786 +static vroot_grb_func *vroot_get_real_bdev = NULL;
8787 +
8788 +static DEFINE_SPINLOCK(vroot_grb_lock);
8789 +
8790 +int register_vroot_grb(vroot_grb_func *func) {
8791 +       int ret = -EBUSY;
8792 +
8793 +       spin_lock(&vroot_grb_lock);
8794 +       if (!vroot_get_real_bdev) {
8795 +               vroot_get_real_bdev = func;
8796 +               ret = 0;
8797 +       }
8798 +       spin_unlock(&vroot_grb_lock);
8799 +       return ret;
8800 +}
8801 +EXPORT_SYMBOL(register_vroot_grb);
8802 +
8803 +int unregister_vroot_grb(vroot_grb_func *func) {
8804 +       int ret = -EINVAL;
8805 +
8806 +       spin_lock(&vroot_grb_lock);
8807 +       if (vroot_get_real_bdev) {
8808 +               vroot_get_real_bdev = NULL;
8809 +               ret = 0;
8810 +       }
8811 +       spin_unlock(&vroot_grb_lock);
8812 +       return ret;
8813 +}
8814 +EXPORT_SYMBOL(unregister_vroot_grb);
8815 +
8816 +#endif
8817 +
8818  /*
8819   * look up a superblock on which quota ops will be performed
8820   * - use the name of a block device to find the superblock thereon
8821 @@ -310,6 +351,22 @@ static struct super_block *quotactl_bloc
8822         putname(tmp);
8823         if (IS_ERR(bdev))
8824                 return ERR_CAST(bdev);
8825 +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE)
8826 +       if (bdev && bdev->bd_inode &&
8827 +                       imajor(bdev->bd_inode) == VROOT_MAJOR) {
8828 +               struct block_device *bdnew = (void *)-EINVAL;
8829 +
8830 +               if (vroot_get_real_bdev)
8831 +                       bdnew = vroot_get_real_bdev(bdev);
8832 +               else
8833 +                       vxdprintk(VXD_CBIT(misc, 0),
8834 +                                       "vroot_get_real_bdev not set");
8835 +               bdput(bdev);
8836 +               if (IS_ERR(bdnew))
8837 +                       return ERR_PTR(PTR_ERR(bdnew));
8838 +               bdev = bdnew;
8839 +       }
8840 +#endif
8841         sb = get_super(bdev);
8842         bdput(bdev);
8843         if (!sb)
8844 diff -NurpP --minimal linux-3.0/fs/reiserfs/file.c linux-3.0-vs2.3.1-pre7/fs/reiserfs/file.c
8845 --- linux-3.0/fs/reiserfs/file.c        2011-01-05 21:50:26.000000000 +0100
8846 +++ linux-3.0-vs2.3.1-pre7/fs/reiserfs/file.c   2011-06-10 22:11:24.000000000 +0200
8847 @@ -312,4 +312,5 @@ const struct inode_operations reiserfs_f
8848         .listxattr = reiserfs_listxattr,
8849         .removexattr = reiserfs_removexattr,
8850         .permission = reiserfs_permission,
8851 +       .sync_flags = reiserfs_sync_flags,
8852  };
8853 diff -NurpP --minimal linux-3.0/fs/reiserfs/inode.c linux-3.0-vs2.3.1-pre7/fs/reiserfs/inode.c
8854 --- linux-3.0/fs/reiserfs/inode.c       2011-05-22 16:17:53.000000000 +0200
8855 +++ linux-3.0-vs2.3.1-pre7/fs/reiserfs/inode.c  2011-06-10 22:11:24.000000000 +0200
8856 @@ -18,6 +18,7 @@
8857  #include <linux/writeback.h>
8858  #include <linux/quotaops.h>
8859  #include <linux/swap.h>
8860 +#include <linux/vs_tag.h>
8861
8862  int reiserfs_commit_write(struct file *f, struct page *page,
8863                           unsigned from, unsigned to);
8864 @@ -1131,6 +1132,8 @@ static void init_inode(struct inode *ino
8865         struct buffer_head *bh;
8866         struct item_head *ih;
8867         __u32 rdev;
8868 +       uid_t uid;
8869 +       gid_t gid;
8870         //int version = ITEM_VERSION_1;
8871
8872         bh = PATH_PLAST_BUFFER(path);
8873 @@ -1151,12 +1154,13 @@ static void init_inode(struct inode *ino
8874                     (struct stat_data_v1 *)B_I_PITEM(bh, ih);
8875                 unsigned long blocks;
8876
8877 +               uid = sd_v1_uid(sd);
8878 +               gid = sd_v1_gid(sd);
8879 +
8880                 set_inode_item_key_version(inode, KEY_FORMAT_3_5);
8881                 set_inode_sd_version(inode, STAT_DATA_V1);
8882                 inode->i_mode = sd_v1_mode(sd);
8883                 inode->i_nlink = sd_v1_nlink(sd);
8884 -               inode->i_uid = sd_v1_uid(sd);
8885 -               inode->i_gid = sd_v1_gid(sd);
8886                 inode->i_size = sd_v1_size(sd);
8887                 inode->i_atime.tv_sec = sd_v1_atime(sd);
8888                 inode->i_mtime.tv_sec = sd_v1_mtime(sd);
8889 @@ -1198,11 +1202,12 @@ static void init_inode(struct inode *ino
8890                 // (directories and symlinks)
8891                 struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih);
8892
8893 +               uid    = sd_v2_uid(sd);
8894 +               gid    = sd_v2_gid(sd);
8895 +
8896                 inode->i_mode = sd_v2_mode(sd);
8897                 inode->i_nlink = sd_v2_nlink(sd);
8898 -               inode->i_uid = sd_v2_uid(sd);
8899                 inode->i_size = sd_v2_size(sd);
8900 -               inode->i_gid = sd_v2_gid(sd);
8901                 inode->i_mtime.tv_sec = sd_v2_mtime(sd);
8902                 inode->i_atime.tv_sec = sd_v2_atime(sd);
8903                 inode->i_ctime.tv_sec = sd_v2_ctime(sd);
8904 @@ -1232,6 +1237,10 @@ static void init_inode(struct inode *ino
8905                 sd_attrs_to_i_attrs(sd_v2_attrs(sd), inode);
8906         }
8907
8908 +       inode->i_uid = INOTAG_UID(DX_TAG(inode), uid, gid);
8909 +       inode->i_gid = INOTAG_GID(DX_TAG(inode), uid, gid);
8910 +       inode->i_tag = INOTAG_TAG(DX_TAG(inode), uid, gid, 0);
8911 +
8912         pathrelse(path);
8913         if (S_ISREG(inode->i_mode)) {
8914                 inode->i_op = &reiserfs_file_inode_operations;
8915 @@ -1254,13 +1263,15 @@ static void init_inode(struct inode *ino
8916  static void inode2sd(void *sd, struct inode *inode, loff_t size)
8917  {
8918         struct stat_data *sd_v2 = (struct stat_data *)sd;
8919 +       uid_t uid = TAGINO_UID(DX_TAG(inode), inode->i_uid, inode->i_tag);
8920 +       gid_t gid = TAGINO_GID(DX_TAG(inode), inode->i_gid, inode->i_tag);
8921         __u16 flags;
8922
8923 +       set_sd_v2_uid(sd_v2, uid);
8924 +       set_sd_v2_gid(sd_v2, gid);
8925         set_sd_v2_mode(sd_v2, inode->i_mode);
8926         set_sd_v2_nlink(sd_v2, inode->i_nlink);
8927 -       set_sd_v2_uid(sd_v2, inode->i_uid);
8928         set_sd_v2_size(sd_v2, size);
8929 -       set_sd_v2_gid(sd_v2, inode->i_gid);
8930         set_sd_v2_mtime(sd_v2, inode->i_mtime.tv_sec);
8931         set_sd_v2_atime(sd_v2, inode->i_atime.tv_sec);
8932         set_sd_v2_ctime(sd_v2, inode->i_ctime.tv_sec);
8933 @@ -2863,14 +2874,19 @@ int reiserfs_commit_write(struct file *f
8934  void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode)
8935  {
8936         if (reiserfs_attrs(inode->i_sb)) {
8937 -               if (sd_attrs & REISERFS_SYNC_FL)
8938 -                       inode->i_flags |= S_SYNC;
8939 -               else
8940 -                       inode->i_flags &= ~S_SYNC;
8941                 if (sd_attrs & REISERFS_IMMUTABLE_FL)
8942                         inode->i_flags |= S_IMMUTABLE;
8943                 else
8944                         inode->i_flags &= ~S_IMMUTABLE;
8945 +               if (sd_attrs & REISERFS_IXUNLINK_FL)
8946 +                       inode->i_flags |= S_IXUNLINK;
8947 +               else
8948 +                       inode->i_flags &= ~S_IXUNLINK;
8949 +
8950 +               if (sd_attrs & REISERFS_SYNC_FL)
8951 +                       inode->i_flags |= S_SYNC;
8952 +               else
8953 +                       inode->i_flags &= ~S_SYNC;
8954                 if (sd_attrs & REISERFS_APPEND_FL)
8955                         inode->i_flags |= S_APPEND;
8956                 else
8957 @@ -2883,6 +2899,15 @@ void sd_attrs_to_i_attrs(__u16 sd_attrs,
8958                         REISERFS_I(inode)->i_flags |= i_nopack_mask;
8959                 else
8960                         REISERFS_I(inode)->i_flags &= ~i_nopack_mask;
8961 +
8962 +               if (sd_attrs & REISERFS_BARRIER_FL)
8963 +                       inode->i_vflags |= V_BARRIER;
8964 +               else
8965 +                       inode->i_vflags &= ~V_BARRIER;
8966 +               if (sd_attrs & REISERFS_COW_FL)
8967 +                       inode->i_vflags |= V_COW;
8968 +               else
8969 +                       inode->i_vflags &= ~V_COW;
8970         }
8971  }
8972
8973 @@ -2893,6 +2918,11 @@ void i_attrs_to_sd_attrs(struct inode *i
8974                         *sd_attrs |= REISERFS_IMMUTABLE_FL;
8975                 else
8976                         *sd_attrs &= ~REISERFS_IMMUTABLE_FL;
8977 +               if (inode->i_flags & S_IXUNLINK)
8978 +                       *sd_attrs |= REISERFS_IXUNLINK_FL;
8979 +               else
8980 +                       *sd_attrs &= ~REISERFS_IXUNLINK_FL;
8981 +
8982                 if (inode->i_flags & S_SYNC)
8983                         *sd_attrs |= REISERFS_SYNC_FL;
8984                 else
8985 @@ -2905,6 +2935,15 @@ void i_attrs_to_sd_attrs(struct inode *i
8986                         *sd_attrs |= REISERFS_NOTAIL_FL;
8987                 else
8988                         *sd_attrs &= ~REISERFS_NOTAIL_FL;
8989 +
8990 +               if (inode->i_vflags & V_BARRIER)
8991 +                       *sd_attrs |= REISERFS_BARRIER_FL;
8992 +               else
8993 +                       *sd_attrs &= ~REISERFS_BARRIER_FL;
8994 +               if (inode->i_vflags & V_COW)
8995 +                       *sd_attrs |= REISERFS_COW_FL;
8996 +               else
8997 +                       *sd_attrs &= ~REISERFS_COW_FL;
8998         }
8999  }
9000
9001 @@ -3148,7 +3187,8 @@ int reiserfs_setattr(struct dentry *dent
9002         }
9003
9004         if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
9005 -           (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
9006 +           (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) ||
9007 +           (ia_valid & ATTR_TAG && attr->ia_tag != inode->i_tag)) {
9008                 struct reiserfs_transaction_handle th;
9009                 int jbegin_count =
9010                     2 *
9011 @@ -3177,6 +3217,9 @@ int reiserfs_setattr(struct dentry *dent
9012                         inode->i_uid = attr->ia_uid;
9013                 if (attr->ia_valid & ATTR_GID)
9014                         inode->i_gid = attr->ia_gid;
9015 +                               if ((attr->ia_valid & ATTR_TAG) &&
9016 +                                       IS_TAGGED(inode))
9017 +                                       inode->i_tag = attr->ia_tag;
9018                 mark_inode_dirty(inode);
9019                 error = journal_end(&th, inode->i_sb, jbegin_count);
9020                 if (error)
9021 diff -NurpP --minimal linux-3.0/fs/reiserfs/ioctl.c linux-3.0-vs2.3.1-pre7/fs/reiserfs/ioctl.c
9022 --- linux-3.0/fs/reiserfs/ioctl.c       2011-05-22 16:17:53.000000000 +0200
9023 +++ linux-3.0-vs2.3.1-pre7/fs/reiserfs/ioctl.c  2011-06-10 22:11:24.000000000 +0200
9024 @@ -11,6 +11,21 @@
9025  #include <linux/pagemap.h>
9026  #include <linux/compat.h>
9027
9028 +
9029 +int reiserfs_sync_flags(struct inode *inode, int flags, int vflags)
9030 +{
9031 +       __u16 sd_attrs = 0;
9032 +
9033 +       inode->i_flags = flags;
9034 +       inode->i_vflags = vflags;
9035 +
9036 +       i_attrs_to_sd_attrs(inode, &sd_attrs);
9037 +       REISERFS_I(inode)->i_attrs = sd_attrs;
9038 +       inode->i_ctime = CURRENT_TIME_SEC;
9039 +       mark_inode_dirty(inode);
9040 +       return 0;
9041 +}
9042 +
9043  /*
9044   * reiserfs_ioctl - handler for ioctl for inode
9045   * supported commands:
9046 @@ -22,7 +37,7 @@
9047  long reiserfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
9048  {
9049         struct inode *inode = filp->f_path.dentry->d_inode;
9050 -       unsigned int flags;
9051 +       unsigned int flags, oldflags;
9052         int err = 0;
9053
9054         reiserfs_write_lock(inode->i_sb);
9055 @@ -47,6 +62,7 @@ long reiserfs_ioctl(struct file *filp, u
9056
9057                 flags = REISERFS_I(inode)->i_attrs;
9058                 i_attrs_to_sd_attrs(inode, (__u16 *) & flags);
9059 +               flags &= REISERFS_FL_USER_VISIBLE;
9060                 err = put_user(flags, (int __user *)arg);
9061                 break;
9062         case REISERFS_IOC_SETFLAGS:{
9063 @@ -67,6 +83,10 @@ long reiserfs_ioctl(struct file *filp, u
9064                                 err = -EFAULT;
9065                                 goto setflags_out;
9066                         }
9067 +                       if (IS_BARRIER(inode)) {
9068 +                               vxwprintk_task(1, "messing with the barrier.");
9069 +                               return -EACCES;
9070 +                       }
9071                         /*
9072                          * Is it quota file? Do not allow user to mess with it
9073                          */
9074 @@ -91,6 +111,10 @@ long reiserfs_ioctl(struct file *filp, u
9075                                         goto setflags_out;
9076                                 }
9077                         }
9078 +
9079 +                       oldflags = REISERFS_I(inode)->i_attrs;
9080 +                       flags &= REISERFS_FL_USER_MODIFIABLE;
9081 +                       flags |= oldflags & ~REISERFS_FL_USER_MODIFIABLE;
9082                         sd_attrs_to_i_attrs(flags, inode);
9083                         REISERFS_I(inode)->i_attrs = flags;
9084                         inode->i_ctime = CURRENT_TIME_SEC;
9085 diff -NurpP --minimal linux-3.0/fs/reiserfs/namei.c linux-3.0-vs2.3.1-pre7/fs/reiserfs/namei.c
9086 --- linux-3.0/fs/reiserfs/namei.c       2011-05-22 16:17:53.000000000 +0200
9087 +++ linux-3.0-vs2.3.1-pre7/fs/reiserfs/namei.c  2011-06-10 22:11:24.000000000 +0200
9088 @@ -18,6 +18,7 @@
9089  #include <linux/reiserfs_acl.h>
9090  #include <linux/reiserfs_xattr.h>
9091  #include <linux/quotaops.h>
9092 +#include <linux/vs_tag.h>
9093
9094  #define INC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) { inc_nlink(i); if (i->i_nlink >= REISERFS_LINK_MAX) i->i_nlink=1; }
9095  #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) drop_nlink(i);
9096 @@ -362,6 +363,7 @@ static struct dentry *reiserfs_lookup(st
9097         if (retval == IO_ERROR) {
9098                 return ERR_PTR(-EIO);
9099         }
9100 +               dx_propagate_tag(nd, inode);
9101
9102         return d_splice_alias(inode, dentry);
9103  }
9104 @@ -1529,6 +1531,7 @@ const struct inode_operations reiserfs_d
9105         .listxattr = reiserfs_listxattr,
9106         .removexattr = reiserfs_removexattr,
9107         .permission = reiserfs_permission,
9108 +       .sync_flags = reiserfs_sync_flags,
9109  };
9110
9111  /*
9112 diff -NurpP --minimal linux-3.0/fs/reiserfs/super.c linux-3.0-vs2.3.1-pre7/fs/reiserfs/super.c
9113 --- linux-3.0/fs/reiserfs/super.c       2011-07-22 11:18:06.000000000 +0200
9114 +++ linux-3.0-vs2.3.1-pre7/fs/reiserfs/super.c  2011-06-10 22:11:24.000000000 +0200
9115 @@ -899,6 +899,14 @@ static int reiserfs_parse_options(struct
9116                 {"user_xattr",.setmask = 1 << REISERFS_UNSUPPORTED_OPT},
9117                 {"nouser_xattr",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT},
9118  #endif
9119 +#ifndef CONFIG_TAGGING_NONE
9120 +               {"tagxid",.setmask = 1 << REISERFS_TAGGED},
9121 +               {"tag",.setmask = 1 << REISERFS_TAGGED},
9122 +               {"notag",.clrmask = 1 << REISERFS_TAGGED},
9123 +#endif
9124 +#ifdef CONFIG_PROPAGATE
9125 +               {"tag",.arg_required = 'T',.values = NULL},
9126 +#endif
9127  #ifdef CONFIG_REISERFS_FS_POSIX_ACL
9128                 {"acl",.setmask = 1 << REISERFS_POSIXACL},
9129                 {"noacl",.clrmask = 1 << REISERFS_POSIXACL},
9130 @@ -1208,6 +1216,14 @@ static int reiserfs_remount(struct super
9131         handle_quota_files(s, qf_names, &qfmt);
9132  #endif
9133
9134 +       if ((mount_options & (1 << REISERFS_TAGGED)) &&
9135 +               !(s->s_flags & MS_TAGGED)) {
9136 +               reiserfs_warning(s, "super-vs01",
9137 +                       "reiserfs: tagging not permitted on remount.");
9138 +               err = -EINVAL;
9139 +               goto out_err;
9140 +       }
9141 +
9142         handle_attrs(s);
9143
9144         /* Add options that are safe here */
9145 @@ -1690,6 +1706,10 @@ static int reiserfs_fill_super(struct su
9146                 goto error;
9147         }
9148
9149 +       /* map mount option tagxid */
9150 +       if (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_TAGGED))
9151 +               s->s_flags |= MS_TAGGED;
9152 +
9153         rs = SB_DISK_SUPER_BLOCK(s);
9154         /* Let's do basic sanity check to verify that underlying device is not
9155            smaller than the filesystem. If the check fails then abort and scream,
9156 diff -NurpP --minimal linux-3.0/fs/reiserfs/xattr.c linux-3.0-vs2.3.1-pre7/fs/reiserfs/xattr.c
9157 --- linux-3.0/fs/reiserfs/xattr.c       2011-07-22 11:18:06.000000000 +0200
9158 +++ linux-3.0-vs2.3.1-pre7/fs/reiserfs/xattr.c  2011-06-22 12:39:15.000000000 +0200
9159 @@ -40,6 +40,7 @@
9160  #include <linux/errno.h>
9161  #include <linux/gfp.h>
9162  #include <linux/fs.h>
9163 +#include <linux/mount.h>
9164  #include <linux/file.h>
9165  #include <linux/pagemap.h>
9166  #include <linux/xattr.h>
9167 diff -NurpP --minimal linux-3.0/fs/stat.c linux-3.0-vs2.3.1-pre7/fs/stat.c
9168 --- linux-3.0/fs/stat.c 2011-05-22 16:17:54.000000000 +0200
9169 +++ linux-3.0-vs2.3.1-pre7/fs/stat.c    2011-06-10 22:11:24.000000000 +0200
9170 @@ -26,6 +26,7 @@ void generic_fillattr(struct inode *inod
9171         stat->nlink = inode->i_nlink;
9172         stat->uid = inode->i_uid;
9173         stat->gid = inode->i_gid;
9174 +       stat->tag = inode->i_tag;
9175         stat->rdev = inode->i_rdev;
9176         stat->atime = inode->i_atime;
9177         stat->mtime = inode->i_mtime;
9178 diff -NurpP --minimal linux-3.0/fs/statfs.c linux-3.0-vs2.3.1-pre7/fs/statfs.c
9179 --- linux-3.0/fs/statfs.c       2011-05-22 16:17:54.000000000 +0200
9180 +++ linux-3.0-vs2.3.1-pre7/fs/statfs.c  2011-06-10 22:11:24.000000000 +0200
9181 @@ -7,6 +7,8 @@
9182  #include <linux/statfs.h>
9183  #include <linux/security.h>
9184  #include <linux/uaccess.h>
9185 +#include <linux/vs_base.h>
9186 +#include <linux/vs_dlimit.h>
9187
9188  static int flags_by_mnt(int mnt_flags)
9189  {
9190 @@ -59,6 +61,8 @@ int statfs_by_dentry(struct dentry *dent
9191         retval = dentry->d_sb->s_op->statfs(dentry, buf);
9192         if (retval == 0 && buf->f_frsize == 0)
9193                 buf->f_frsize = buf->f_bsize;
9194 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
9195 +               vx_vsi_statfs(dentry->d_sb, buf);
9196         return retval;
9197  }
9198
9199 diff -NurpP --minimal linux-3.0/fs/super.c linux-3.0-vs2.3.1-pre7/fs/super.c
9200 --- linux-3.0/fs/super.c        2011-07-22 11:18:06.000000000 +0200
9201 +++ linux-3.0-vs2.3.1-pre7/fs/super.c   2011-07-23 16:49:09.000000000 +0200
9202 @@ -32,6 +32,9 @@
9203  #include <linux/backing-dev.h>
9204  #include <linux/rculist_bl.h>
9205  #include <linux/cleancache.h>
9206 +#include <linux/devpts_fs.h>
9207 +#include <linux/proc_fs.h>
9208 +#include <linux/vs_context.h>
9209  #include "internal.h"
9210
9211
9212 @@ -943,6 +946,13 @@ mount_fs(struct file_system_type *type,
9213         WARN_ON(sb->s_bdi == &default_backing_dev_info);
9214         sb->s_flags |= MS_BORN;
9215
9216 +       error = -EPERM;
9217 +       if (!vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT) &&
9218 +               !sb->s_bdev &&
9219 +               (sb->s_magic != PROC_SUPER_MAGIC) &&
9220 +               (sb->s_magic != DEVPTS_SUPER_MAGIC))
9221 +               goto out_sb;
9222 +
9223         error = security_sb_kern_mount(sb, flags, secdata);
9224         if (error)
9225                 goto out_sb;
9226 diff -NurpP --minimal linux-3.0/fs/sysfs/mount.c linux-3.0-vs2.3.1-pre7/fs/sysfs/mount.c
9227 --- linux-3.0/fs/sysfs/mount.c  2011-07-22 11:18:06.000000000 +0200
9228 +++ linux-3.0-vs2.3.1-pre7/fs/sysfs/mount.c     2011-06-22 12:39:15.000000000 +0200
9229 @@ -47,7 +47,7 @@ static int sysfs_fill_super(struct super
9230
9231         sb->s_blocksize = PAGE_CACHE_SIZE;
9232         sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
9233 -       sb->s_magic = SYSFS_MAGIC;
9234 +       sb->s_magic = SYSFS_SUPER_MAGIC;
9235         sb->s_op = &sysfs_ops;
9236         sb->s_time_gran = 1;
9237
9238 diff -NurpP --minimal linux-3.0/fs/utimes.c linux-3.0-vs2.3.1-pre7/fs/utimes.c
9239 --- linux-3.0/fs/utimes.c       2011-05-22 16:17:54.000000000 +0200
9240 +++ linux-3.0-vs2.3.1-pre7/fs/utimes.c  2011-06-10 22:11:24.000000000 +0200
9241 @@ -8,6 +8,8 @@
9242  #include <linux/stat.h>
9243  #include <linux/utime.h>
9244  #include <linux/syscalls.h>
9245 +#include <linux/mount.h>
9246 +#include <linux/vs_cowbl.h>
9247  #include <asm/uaccess.h>
9248  #include <asm/unistd.h>
9249
9250 @@ -52,12 +54,18 @@ static int utimes_common(struct path *pa
9251  {
9252         int error;
9253         struct iattr newattrs;
9254 -       struct inode *inode = path->dentry->d_inode;
9255 +       struct inode *inode;
9256
9257         error = mnt_want_write(path->mnt);
9258         if (error)
9259                 goto out;
9260
9261 +       error = cow_check_and_break(path);
9262 +       if (error)
9263 +               goto mnt_drop_write_and_out;
9264 +
9265 +       inode = path->dentry->d_inode;
9266 +
9267         if (times && times[0].tv_nsec == UTIME_NOW &&
9268                      times[1].tv_nsec == UTIME_NOW)
9269                 times = NULL;
9270 diff -NurpP --minimal linux-3.0/fs/xattr.c linux-3.0-vs2.3.1-pre7/fs/xattr.c
9271 --- linux-3.0/fs/xattr.c        2011-07-22 11:18:09.000000000 +0200
9272 +++ linux-3.0-vs2.3.1-pre7/fs/xattr.c   2011-06-10 23:10:19.000000000 +0200
9273 @@ -18,6 +18,7 @@
9274  #include <linux/module.h>
9275  #include <linux/fsnotify.h>
9276  #include <linux/audit.h>
9277 +#include <linux/mount.h>
9278  #include <asm/uaccess.h>
9279
9280
9281 @@ -49,7 +50,7 @@ xattr_permission(struct inode *inode, co
9282          * The trusted.* namespace can only be accessed by privileged users.
9283          */
9284         if (!strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN)) {
9285 -               if (!capable(CAP_SYS_ADMIN))
9286 +               if (!vx_capable(CAP_SYS_ADMIN, VXC_FS_TRUSTED))
9287                         return (mask & MAY_WRITE) ? -EPERM : -ENODATA;
9288                 return 0;
9289         }
9290 diff -NurpP --minimal linux-3.0/fs/xfs/linux-2.6/xfs_ioctl.c linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_ioctl.c
9291 --- linux-3.0/fs/xfs/linux-2.6/xfs_ioctl.c      2011-05-22 16:17:54.000000000 +0200
9292 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_ioctl.c 2011-06-10 22:11:24.000000000 +0200
9293 @@ -28,7 +28,7 @@
9294  #include "xfs_bmap_btree.h"
9295  #include "xfs_dinode.h"
9296  #include "xfs_inode.h"
9297 -#include "xfs_ioctl.h"
9298 +// #include "xfs_ioctl.h"
9299  #include "xfs_rtalloc.h"
9300  #include "xfs_itable.h"
9301  #include "xfs_error.h"
9302 @@ -748,6 +748,10 @@ xfs_merge_ioc_xflags(
9303                 xflags |= XFS_XFLAG_IMMUTABLE;
9304         else
9305                 xflags &= ~XFS_XFLAG_IMMUTABLE;
9306 +       if (flags & FS_IXUNLINK_FL)
9307 +               xflags |= XFS_XFLAG_IXUNLINK;
9308 +       else
9309 +               xflags &= ~XFS_XFLAG_IXUNLINK;
9310         if (flags & FS_APPEND_FL)
9311                 xflags |= XFS_XFLAG_APPEND;
9312         else
9313 @@ -776,6 +780,8 @@ xfs_di2lxflags(
9314
9315         if (di_flags & XFS_DIFLAG_IMMUTABLE)
9316                 flags |= FS_IMMUTABLE_FL;
9317 +       if (di_flags & XFS_DIFLAG_IXUNLINK)
9318 +               flags |= FS_IXUNLINK_FL;
9319         if (di_flags & XFS_DIFLAG_APPEND)
9320                 flags |= FS_APPEND_FL;
9321         if (di_flags & XFS_DIFLAG_SYNC)
9322 @@ -836,6 +842,8 @@ xfs_set_diflags(
9323         di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
9324         if (xflags & XFS_XFLAG_IMMUTABLE)
9325                 di_flags |= XFS_DIFLAG_IMMUTABLE;
9326 +       if (xflags & XFS_XFLAG_IXUNLINK)
9327 +               di_flags |= XFS_DIFLAG_IXUNLINK;
9328         if (xflags & XFS_XFLAG_APPEND)
9329                 di_flags |= XFS_DIFLAG_APPEND;
9330         if (xflags & XFS_XFLAG_SYNC)
9331 @@ -878,6 +886,10 @@ xfs_diflags_to_linux(
9332                 inode->i_flags |= S_IMMUTABLE;
9333         else
9334                 inode->i_flags &= ~S_IMMUTABLE;
9335 +       if (xflags & XFS_XFLAG_IXUNLINK)
9336 +               inode->i_flags |= S_IXUNLINK;
9337 +       else
9338 +               inode->i_flags &= ~S_IXUNLINK;
9339         if (xflags & XFS_XFLAG_APPEND)
9340                 inode->i_flags |= S_APPEND;
9341         else
9342 @@ -1370,10 +1382,18 @@ xfs_file_ioctl(
9343         case XFS_IOC_FSGETXATTRA:
9344                 return xfs_ioc_fsgetxattr(ip, 1, arg);
9345         case XFS_IOC_FSSETXATTR:
9346 +               if (IS_BARRIER(inode)) {
9347 +                       vxwprintk_task(1, "messing with the barrier.");
9348 +                       return -XFS_ERROR(EACCES);
9349 +               }
9350                 return xfs_ioc_fssetxattr(ip, filp, arg);
9351         case XFS_IOC_GETXFLAGS:
9352                 return xfs_ioc_getxflags(ip, arg);
9353         case XFS_IOC_SETXFLAGS:
9354 +               if (IS_BARRIER(inode)) {
9355 +                       vxwprintk_task(1, "messing with the barrier.");
9356 +                       return -XFS_ERROR(EACCES);
9357 +               }
9358                 return xfs_ioc_setxflags(ip, filp, arg);
9359
9360         case XFS_IOC_FSSETDM: {
9361 diff -NurpP --minimal linux-3.0/fs/xfs/linux-2.6/xfs_ioctl.h linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_ioctl.h
9362 --- linux-3.0/fs/xfs/linux-2.6/xfs_ioctl.h      2010-07-07 18:31:54.000000000 +0200
9363 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_ioctl.h 2011-06-10 22:11:24.000000000 +0200
9364 @@ -70,6 +70,12 @@ xfs_handle_to_dentry(
9365         void __user             *uhandle,
9366         u32                     hlen);
9367
9368 +extern int
9369 +xfs_sync_flags(
9370 +       struct inode            *inode,
9371 +       int                     flags,
9372 +       int                     vflags);
9373 +
9374  extern long
9375  xfs_file_ioctl(
9376         struct file             *filp,
9377 diff -NurpP --minimal linux-3.0/fs/xfs/linux-2.6/xfs_iops.c linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_iops.c
9378 --- linux-3.0/fs/xfs/linux-2.6/xfs_iops.c       2011-07-22 11:18:09.000000000 +0200
9379 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_iops.c  2011-06-22 12:39:15.000000000 +0200
9380 @@ -30,6 +30,7 @@
9381  #include "xfs_bmap_btree.h"
9382  #include "xfs_dinode.h"
9383  #include "xfs_inode.h"
9384 +#include "xfs_ioctl.h"
9385  #include "xfs_bmap.h"
9386  #include "xfs_rtalloc.h"
9387  #include "xfs_error.h"
9388 @@ -48,6 +49,7 @@
9389  #include <linux/security.h>
9390  #include <linux/fiemap.h>
9391  #include <linux/slab.h>
9392 +#include <linux/vs_tag.h>
9393
9394  /*
9395   * Bring the timestamps in the XFS inode uptodate.
9396 @@ -464,6 +466,7 @@ xfs_vn_getattr(
9397         stat->nlink = ip->i_d.di_nlink;
9398         stat->uid = ip->i_d.di_uid;
9399         stat->gid = ip->i_d.di_gid;
9400 +       stat->tag = ip->i_d.di_tag;
9401         stat->ino = ip->i_ino;
9402         stat->atime = inode->i_atime;
9403         stat->mtime = inode->i_mtime;
9404 @@ -599,6 +602,7 @@ static const struct inode_operations xfs
9405         .removexattr            = generic_removexattr,
9406         .listxattr              = xfs_vn_listxattr,
9407         .fiemap                 = xfs_vn_fiemap,
9408 +       .sync_flags             = xfs_sync_flags,
9409  };
9410
9411  static const struct inode_operations xfs_dir_inode_operations = {
9412 @@ -624,6 +628,7 @@ static const struct inode_operations xfs
9413         .getxattr               = generic_getxattr,
9414         .removexattr            = generic_removexattr,
9415         .listxattr              = xfs_vn_listxattr,
9416 +       .sync_flags             = xfs_sync_flags,
9417  };
9418
9419  static const struct inode_operations xfs_dir_ci_inode_operations = {
9420 @@ -673,6 +678,10 @@ xfs_diflags_to_iflags(
9421                 inode->i_flags |= S_IMMUTABLE;
9422         else
9423                 inode->i_flags &= ~S_IMMUTABLE;
9424 +       if (ip->i_d.di_flags & XFS_DIFLAG_IXUNLINK)
9425 +               inode->i_flags |= S_IXUNLINK;
9426 +       else
9427 +               inode->i_flags &= ~S_IXUNLINK;
9428         if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
9429                 inode->i_flags |= S_APPEND;
9430         else
9431 @@ -685,6 +694,15 @@ xfs_diflags_to_iflags(
9432                 inode->i_flags |= S_NOATIME;
9433         else
9434                 inode->i_flags &= ~S_NOATIME;
9435 +
9436 +       if (ip->i_d.di_vflags & XFS_DIVFLAG_BARRIER)
9437 +               inode->i_vflags |= V_BARRIER;
9438 +       else
9439 +               inode->i_vflags &= ~V_BARRIER;
9440 +       if (ip->i_d.di_vflags & XFS_DIVFLAG_COW)
9441 +               inode->i_vflags |= V_COW;
9442 +       else
9443 +               inode->i_vflags &= ~V_COW;
9444  }
9445
9446  /*
9447 @@ -716,6 +734,7 @@ xfs_setup_inode(
9448         inode->i_nlink  = ip->i_d.di_nlink;
9449         inode->i_uid    = ip->i_d.di_uid;
9450         inode->i_gid    = ip->i_d.di_gid;
9451 +       inode->i_tag    = ip->i_d.di_tag;
9452
9453         switch (inode->i_mode & S_IFMT) {
9454         case S_IFBLK:
9455 diff -NurpP --minimal linux-3.0/fs/xfs/linux-2.6/xfs_linux.h linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_linux.h
9456 --- linux-3.0/fs/xfs/linux-2.6/xfs_linux.h      2011-07-22 11:18:09.000000000 +0200
9457 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_linux.h 2011-06-10 22:11:24.000000000 +0200
9458 @@ -115,6 +115,7 @@
9459
9460  #define current_cpu()          (raw_smp_processor_id())
9461  #define current_pid()          (current->pid)
9462 +#define current_fstag(vp)      (dx_current_fstag((vp)->i_sb))
9463  #define current_test_flags(f)  (current->flags & (f))
9464  #define current_set_flags_nested(sp, f)                \
9465                 (*(sp) = current->flags, current->flags |= (f))
9466 diff -NurpP --minimal linux-3.0/fs/xfs/linux-2.6/xfs_super.c linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_super.c
9467 --- linux-3.0/fs/xfs/linux-2.6/xfs_super.c      2011-07-22 11:18:09.000000000 +0200
9468 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/linux-2.6/xfs_super.c 2011-06-22 12:39:15.000000000 +0200
9469 @@ -114,6 +114,9 @@ mempool_t *xfs_ioend_pool;
9470  #define MNTOPT_NODELAYLOG  "nodelaylog"        /* Delayed logging disabled */
9471  #define MNTOPT_DISCARD    "discard"    /* Discard unused blocks */
9472  #define MNTOPT_NODISCARD   "nodiscard" /* Do not discard unused blocks */
9473 +#define MNTOPT_TAGXID  "tagxid"        /* context tagging for inodes */
9474 +#define MNTOPT_TAGGED  "tag"           /* context tagging for inodes */
9475 +#define MNTOPT_NOTAGTAG        "notag"         /* do not use context tagging */
9476
9477  /*
9478   * Table driven mount option parser.
9479 @@ -122,10 +125,14 @@ mempool_t *xfs_ioend_pool;
9480   * in the future, too.
9481   */
9482  enum {
9483 +       Opt_tag, Opt_notag,
9484         Opt_barrier, Opt_nobarrier, Opt_err
9485  };
9486
9487  static const match_table_t tokens = {
9488 +       {Opt_tag, "tagxid"},
9489 +       {Opt_tag, "tag"},
9490 +       {Opt_notag, "notag"},
9491         {Opt_barrier, "barrier"},
9492         {Opt_nobarrier, "nobarrier"},
9493         {Opt_err, NULL}
9494 @@ -373,6 +380,19 @@ xfs_parseargs(
9495                 } else if (!strcmp(this_char, "irixsgid")) {
9496                         xfs_warn(mp,
9497         "irixsgid is now a sysctl(2) variable, option is deprecated.");
9498 +#ifndef CONFIG_TAGGING_NONE
9499 +               } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
9500 +                       mp->m_flags |= XFS_MOUNT_TAGGED;
9501 +               } else if (!strcmp(this_char, MNTOPT_NOTAGTAG)) {
9502 +                       mp->m_flags &= ~XFS_MOUNT_TAGGED;
9503 +               } else if (!strcmp(this_char, MNTOPT_TAGXID)) {
9504 +                       mp->m_flags |= XFS_MOUNT_TAGGED;
9505 +#endif
9506 +#ifdef CONFIG_PROPAGATE
9507 +               } else if (!strcmp(this_char, MNTOPT_TAGGED)) {
9508 +                       /* use value */
9509 +                       mp->m_flags |= XFS_MOUNT_TAGGED;
9510 +#endif
9511                 } else {
9512                         xfs_warn(mp, "unknown mount option [%s].", this_char);
9513                         return EINVAL;
9514 @@ -1182,6 +1202,16 @@ xfs_fs_remount(
9515                 case Opt_nobarrier:
9516                         mp->m_flags &= ~XFS_MOUNT_BARRIER;
9517                         break;
9518 +               case Opt_tag:
9519 +                       if (!(sb->s_flags & MS_TAGGED)) {
9520 +                               printk(KERN_INFO
9521 +                                       "XFS: %s: tagging not permitted on remount.\n",
9522 +                                       sb->s_id);
9523 +                               return -EINVAL;
9524 +                       }
9525 +                       break;
9526 +               case Opt_notag:
9527 +                       break;
9528                 default:
9529                         /*
9530                          * Logically we would return an error here to prevent
9531 @@ -1397,6 +1427,9 @@ xfs_fs_fill_super(
9532         if (error)
9533                 goto out_free_sb;
9534
9535 +       if (mp->m_flags & XFS_MOUNT_TAGGED)
9536 +               sb->s_flags |= MS_TAGGED;
9537 +
9538         /*
9539          * we must configure the block size in the superblock before we run the
9540          * full mount process as the mount process can lookup and cache inodes.
9541 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_dinode.h linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_dinode.h
9542 --- linux-3.0/fs/xfs/xfs_dinode.h       2011-01-05 21:50:28.000000000 +0100
9543 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_dinode.h  2011-06-10 22:11:24.000000000 +0200
9544 @@ -51,7 +51,9 @@ typedef struct xfs_dinode {
9545         __be32          di_nlink;       /* number of links to file */
9546         __be16          di_projid_lo;   /* lower part of owner's project id */
9547         __be16          di_projid_hi;   /* higher part owner's project id */
9548 -       __u8            di_pad[6];      /* unused, zeroed space */
9549 +       __u8            di_pad[2];      /* unused, zeroed space */
9550 +       __be16          di_tag;         /* context tagging */
9551 +       __be16          di_vflags;      /* vserver specific flags */
9552         __be16          di_flushiter;   /* incremented on flush */
9553         xfs_timestamp_t di_atime;       /* time last accessed */
9554         xfs_timestamp_t di_mtime;       /* time last modified */
9555 @@ -184,6 +186,8 @@ static inline void xfs_dinode_put_rdev(s
9556  #define XFS_DIFLAG_EXTSZINHERIT_BIT 12 /* inherit inode extent size */
9557  #define XFS_DIFLAG_NODEFRAG_BIT     13 /* do not reorganize/defragment */
9558  #define XFS_DIFLAG_FILESTREAM_BIT   14  /* use filestream allocator */
9559 +#define XFS_DIFLAG_IXUNLINK_BIT     15 /* Immutable inver on unlink */
9560 +
9561  #define XFS_DIFLAG_REALTIME      (1 << XFS_DIFLAG_REALTIME_BIT)
9562  #define XFS_DIFLAG_PREALLOC      (1 << XFS_DIFLAG_PREALLOC_BIT)
9563  #define XFS_DIFLAG_NEWRTBM       (1 << XFS_DIFLAG_NEWRTBM_BIT)
9564 @@ -199,6 +203,7 @@ static inline void xfs_dinode_put_rdev(s
9565  #define XFS_DIFLAG_EXTSZINHERIT  (1 << XFS_DIFLAG_EXTSZINHERIT_BIT)
9566  #define XFS_DIFLAG_NODEFRAG      (1 << XFS_DIFLAG_NODEFRAG_BIT)
9567  #define XFS_DIFLAG_FILESTREAM    (1 << XFS_DIFLAG_FILESTREAM_BIT)
9568 +#define XFS_DIFLAG_IXUNLINK      (1 << XFS_DIFLAG_IXUNLINK_BIT)
9569
9570  #ifdef CONFIG_XFS_RT
9571  #define XFS_IS_REALTIME_INODE(ip) ((ip)->i_d.di_flags & XFS_DIFLAG_REALTIME)
9572 @@ -211,6 +216,10 @@ static inline void xfs_dinode_put_rdev(s
9573          XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND | XFS_DIFLAG_SYNC | \
9574          XFS_DIFLAG_NOATIME | XFS_DIFLAG_NODUMP | XFS_DIFLAG_RTINHERIT | \
9575          XFS_DIFLAG_PROJINHERIT | XFS_DIFLAG_NOSYMLINKS | XFS_DIFLAG_EXTSIZE | \
9576 -        XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM)
9577 +        XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_NODEFRAG | XFS_DIFLAG_FILESTREAM | \
9578 +        XFS_DIFLAG_IXUNLINK)
9579 +
9580 +#define XFS_DIVFLAG_BARRIER    0x01
9581 +#define XFS_DIVFLAG_COW                0x02
9582
9583  #endif /* __XFS_DINODE_H__ */
9584 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_fs.h linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_fs.h
9585 --- linux-3.0/fs/xfs/xfs_fs.h   2011-01-05 21:50:28.000000000 +0100
9586 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_fs.h      2011-06-10 22:11:24.000000000 +0200
9587 @@ -67,6 +67,9 @@ struct fsxattr {
9588  #define XFS_XFLAG_EXTSZINHERIT 0x00001000      /* inherit inode extent size */
9589  #define XFS_XFLAG_NODEFRAG     0x00002000      /* do not defragment */
9590  #define XFS_XFLAG_FILESTREAM   0x00004000      /* use filestream allocator */
9591 +#define XFS_XFLAG_IXUNLINK     0x00008000      /* immutable invert on unlink */
9592 +#define XFS_XFLAG_BARRIER      0x10000000      /* chroot() barrier */
9593 +#define XFS_XFLAG_COW          0x20000000      /* copy on write mark */
9594  #define XFS_XFLAG_HASATTR      0x80000000      /* no DIFLAG for this   */
9595
9596  /*
9597 @@ -297,7 +300,8 @@ typedef struct xfs_bstat {
9598  #define        bs_projid       bs_projid_lo    /* (previously just bs_projid)  */
9599         __u16           bs_forkoff;     /* inode fork offset in bytes   */
9600         __u16           bs_projid_hi;   /* higher part of project id    */
9601 -       unsigned char   bs_pad[10];     /* pad space, unused            */
9602 +       unsigned char   bs_pad[8];      /* pad space, unused            */
9603 +       __u16           bs_tag;         /* context tagging              */
9604         __u32           bs_dmevmask;    /* DMIG event mask              */
9605         __u16           bs_dmstate;     /* DMIG state info              */
9606         __u16           bs_aextents;    /* attribute number of extents  */
9607 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_ialloc.c linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_ialloc.c
9608 --- linux-3.0/fs/xfs/xfs_ialloc.c       2011-05-22 16:17:54.000000000 +0200
9609 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_ialloc.c  2011-06-10 22:11:24.000000000 +0200
9610 @@ -37,7 +37,6 @@
9611  #include "xfs_error.h"
9612  #include "xfs_bmap.h"
9613
9614 -
9615  /*
9616   * Allocation group level functions.
9617   */
9618 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_inode.c linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_inode.c
9619 --- linux-3.0/fs/xfs/xfs_inode.c        2011-07-22 11:18:10.000000000 +0200
9620 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_inode.c   2011-06-10 22:11:24.000000000 +0200
9621 @@ -243,6 +243,7 @@ xfs_inotobp(
9622         return 0;
9623  }
9624
9625 +#include <linux/vs_tag.h>
9626
9627  /*
9628   * This routine is called to map an inode to the buffer containing
9629 @@ -641,15 +642,25 @@ xfs_iformat_btree(
9630  STATIC void
9631  xfs_dinode_from_disk(
9632         xfs_icdinode_t          *to,
9633 -       xfs_dinode_t            *from)
9634 +       xfs_dinode_t            *from,
9635 +       int tagged)
9636  {
9637 +       uint32_t uid, gid, tag;
9638 +
9639         to->di_magic = be16_to_cpu(from->di_magic);
9640         to->di_mode = be16_to_cpu(from->di_mode);
9641         to->di_version = from ->di_version;
9642         to->di_format = from->di_format;
9643         to->di_onlink = be16_to_cpu(from->di_onlink);
9644 -       to->di_uid = be32_to_cpu(from->di_uid);
9645 -       to->di_gid = be32_to_cpu(from->di_gid);
9646 +
9647 +       uid = be32_to_cpu(from->di_uid);
9648 +       gid = be32_to_cpu(from->di_gid);
9649 +       tag = be16_to_cpu(from->di_tag);
9650 +
9651 +       to->di_uid = INOTAG_UID(tagged, uid, gid);
9652 +       to->di_gid = INOTAG_GID(tagged, uid, gid);
9653 +       to->di_tag = INOTAG_TAG(tagged, uid, gid, tag);
9654 +
9655         to->di_nlink = be32_to_cpu(from->di_nlink);
9656         to->di_projid_lo = be16_to_cpu(from->di_projid_lo);
9657         to->di_projid_hi = be16_to_cpu(from->di_projid_hi);
9658 @@ -671,21 +682,26 @@ xfs_dinode_from_disk(
9659         to->di_dmevmask = be32_to_cpu(from->di_dmevmask);
9660         to->di_dmstate  = be16_to_cpu(from->di_dmstate);
9661         to->di_flags    = be16_to_cpu(from->di_flags);
9662 +       to->di_vflags   = be16_to_cpu(from->di_vflags);
9663         to->di_gen      = be32_to_cpu(from->di_gen);
9664  }
9665
9666  void
9667  xfs_dinode_to_disk(
9668         xfs_dinode_t            *to,
9669 -       xfs_icdinode_t          *from)
9670 +       xfs_icdinode_t          *from,
9671 +       int tagged)
9672  {
9673         to->di_magic = cpu_to_be16(from->di_magic);
9674         to->di_mode = cpu_to_be16(from->di_mode);
9675         to->di_version = from ->di_version;
9676         to->di_format = from->di_format;
9677         to->di_onlink = cpu_to_be16(from->di_onlink);
9678 -       to->di_uid = cpu_to_be32(from->di_uid);
9679 -       to->di_gid = cpu_to_be32(from->di_gid);
9680 +
9681 +       to->di_uid = cpu_to_be32(TAGINO_UID(tagged, from->di_uid, from->di_tag));
9682 +       to->di_gid = cpu_to_be32(TAGINO_GID(tagged, from->di_gid, from->di_tag));
9683 +       to->di_tag = cpu_to_be16(TAGINO_TAG(tagged, from->di_tag));
9684 +
9685         to->di_nlink = cpu_to_be32(from->di_nlink);
9686         to->di_projid_lo = cpu_to_be16(from->di_projid_lo);
9687         to->di_projid_hi = cpu_to_be16(from->di_projid_hi);
9688 @@ -707,12 +723,14 @@ xfs_dinode_to_disk(
9689         to->di_dmevmask = cpu_to_be32(from->di_dmevmask);
9690         to->di_dmstate = cpu_to_be16(from->di_dmstate);
9691         to->di_flags = cpu_to_be16(from->di_flags);
9692 +       to->di_vflags = cpu_to_be16(from->di_vflags);
9693         to->di_gen = cpu_to_be32(from->di_gen);
9694  }
9695
9696  STATIC uint
9697  _xfs_dic2xflags(
9698 -       __uint16_t              di_flags)
9699 +       __uint16_t              di_flags,
9700 +       __uint16_t              di_vflags)
9701  {
9702         uint                    flags = 0;
9703
9704 @@ -723,6 +741,8 @@ _xfs_dic2xflags(
9705                         flags |= XFS_XFLAG_PREALLOC;
9706                 if (di_flags & XFS_DIFLAG_IMMUTABLE)
9707                         flags |= XFS_XFLAG_IMMUTABLE;
9708 +               if (di_flags & XFS_DIFLAG_IXUNLINK)
9709 +                       flags |= XFS_XFLAG_IXUNLINK;
9710                 if (di_flags & XFS_DIFLAG_APPEND)
9711                         flags |= XFS_XFLAG_APPEND;
9712                 if (di_flags & XFS_DIFLAG_SYNC)
9713 @@ -747,6 +767,10 @@ _xfs_dic2xflags(
9714                         flags |= XFS_XFLAG_FILESTREAM;
9715         }
9716
9717 +       if (di_vflags & XFS_DIVFLAG_BARRIER)
9718 +               flags |= FS_BARRIER_FL;
9719 +       if (di_vflags & XFS_DIVFLAG_COW)
9720 +               flags |= FS_COW_FL;
9721         return flags;
9722  }
9723
9724 @@ -756,7 +780,7 @@ xfs_ip2xflags(
9725  {
9726         xfs_icdinode_t          *dic = &ip->i_d;
9727
9728 -       return _xfs_dic2xflags(dic->di_flags) |
9729 +       return _xfs_dic2xflags(dic->di_flags, dic->di_vflags) |
9730                                 (XFS_IFORK_Q(ip) ? XFS_XFLAG_HASATTR : 0);
9731  }
9732
9733 @@ -764,7 +788,8 @@ uint
9734  xfs_dic2xflags(
9735         xfs_dinode_t            *dip)
9736  {
9737 -       return _xfs_dic2xflags(be16_to_cpu(dip->di_flags)) |
9738 +       return _xfs_dic2xflags(be16_to_cpu(dip->di_flags),
9739 +                               be16_to_cpu(dip->di_vflags)) |
9740                                 (XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
9741  }
9742
9743 @@ -797,7 +822,6 @@ xfs_iread(
9744         if (error)
9745                 return error;
9746         dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
9747 -
9748         /*
9749          * If we got something that isn't an inode it means someone
9750          * (nfs or dmi) has a stale handle.
9751 @@ -820,7 +844,8 @@ xfs_iread(
9752          * Otherwise, just get the truly permanent information.
9753          */
9754         if (dip->di_mode) {
9755 -               xfs_dinode_from_disk(&ip->i_d, dip);
9756 +               xfs_dinode_from_disk(&ip->i_d, dip,
9757 +                       mp->m_flags & XFS_MOUNT_TAGGED);
9758                 error = xfs_iformat(ip, dip);
9759                 if (error)  {
9760  #ifdef DEBUG
9761 @@ -1015,6 +1040,7 @@ xfs_ialloc(
9762         ASSERT(ip->i_d.di_nlink == nlink);
9763         ip->i_d.di_uid = current_fsuid();
9764         ip->i_d.di_gid = current_fsgid();
9765 +       ip->i_d.di_tag = current_fstag(&ip->i_vnode);
9766         xfs_set_projid(ip, prid);
9767         memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
9768
9769 @@ -1075,6 +1101,7 @@ xfs_ialloc(
9770         ip->i_d.di_dmevmask = 0;
9771         ip->i_d.di_dmstate = 0;
9772         ip->i_d.di_flags = 0;
9773 +       ip->i_d.di_vflags = 0;
9774         flags = XFS_ILOG_CORE;
9775         switch (mode & S_IFMT) {
9776         case S_IFIFO:
9777 @@ -2108,6 +2135,7 @@ xfs_ifree(
9778         }
9779         ip->i_d.di_mode = 0;            /* mark incore inode as free */
9780         ip->i_d.di_flags = 0;
9781 +       ip->i_d.di_vflags = 0;
9782         ip->i_d.di_dmevmask = 0;
9783         ip->i_d.di_forkoff = 0;         /* mark the attr fork not in use */
9784         ip->i_df.if_ext_max =
9785 @@ -2987,7 +3015,8 @@ xfs_iflush_int(
9786          * because if the inode is dirty at all the core must
9787          * be.
9788          */
9789 -       xfs_dinode_to_disk(dip, &ip->i_d);
9790 +       xfs_dinode_to_disk(dip, &ip->i_d,
9791 +               mp->m_flags & XFS_MOUNT_TAGGED);
9792
9793         /* Wrap, we never let the log put out DI_MAX_FLUSH */
9794         if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
9795 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_inode.h linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_inode.h
9796 --- linux-3.0/fs/xfs/xfs_inode.h        2011-07-22 11:18:10.000000000 +0200
9797 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_inode.h   2011-07-01 11:35:35.000000000 +0200
9798 @@ -135,7 +135,9 @@ typedef struct xfs_icdinode {
9799         __uint32_t      di_nlink;       /* number of links to file */
9800         __uint16_t      di_projid_lo;   /* lower part of owner's project id */
9801         __uint16_t      di_projid_hi;   /* higher part of owner's project id */
9802 -       __uint8_t       di_pad[6];      /* unused, zeroed space */
9803 +       __uint8_t       di_pad[2];      /* unused, zeroed space */
9804 +       __uint16_t      di_tag;         /* context tagging */
9805 +       __uint16_t      di_vflags;      /* vserver specific flags */
9806         __uint16_t      di_flushiter;   /* incremented on flush */
9807         xfs_ictimestamp_t di_atime;     /* time last accessed */
9808         xfs_ictimestamp_t di_mtime;     /* time last modified */
9809 @@ -546,7 +548,7 @@ int         xfs_itobp(struct xfs_mount *, struc
9810  int            xfs_iread(struct xfs_mount *, struct xfs_trans *,
9811                           struct xfs_inode *, uint);
9812  void           xfs_dinode_to_disk(struct xfs_dinode *,
9813 -                                  struct xfs_icdinode *);
9814 +                                  struct xfs_icdinode *, int);
9815  void           xfs_idestroy_fork(struct xfs_inode *, int);
9816  void           xfs_idata_realloc(struct xfs_inode *, int, int);
9817  void           xfs_iroot_realloc(struct xfs_inode *, int, int);
9818 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_itable.c linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_itable.c
9819 --- linux-3.0/fs/xfs/xfs_itable.c       2011-05-22 16:17:54.000000000 +0200
9820 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_itable.c  2011-06-10 22:11:24.000000000 +0200
9821 @@ -98,6 +98,7 @@ xfs_bulkstat_one_int(
9822         buf->bs_mode = dic->di_mode;
9823         buf->bs_uid = dic->di_uid;
9824         buf->bs_gid = dic->di_gid;
9825 +       buf->bs_tag = dic->di_tag;
9826         buf->bs_size = dic->di_size;
9827
9828         /*
9829 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_log_recover.c linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_log_recover.c
9830 --- linux-3.0/fs/xfs/xfs_log_recover.c  2011-07-22 11:18:10.000000000 +0200
9831 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_log_recover.c     2011-06-10 22:11:24.000000000 +0200
9832 @@ -2343,7 +2343,8 @@ xlog_recover_inode_pass2(
9833         }
9834
9835         /* The core is in in-core format */
9836 -       xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr);
9837 +       xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr,
9838 +               mp->m_flags & XFS_MOUNT_TAGGED);
9839
9840         /* the rest is in on-disk format */
9841         if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
9842 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_mount.h linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_mount.h
9843 --- linux-3.0/fs/xfs/xfs_mount.h        2011-07-22 11:18:10.000000000 +0200
9844 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_mount.h   2011-06-10 22:11:24.000000000 +0200
9845 @@ -249,6 +249,7 @@ typedef struct xfs_mount {
9846                                                    allocator */
9847  #define XFS_MOUNT_NOATTR2      (1ULL << 25)    /* disable use of attr2 format */
9848
9849 +#define XFS_MOUNT_TAGGED       (1ULL << 31)    /* context tagging */
9850
9851  /*
9852   * Default minimum read and write sizes.
9853 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_vnodeops.c linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_vnodeops.c
9854 --- linux-3.0/fs/xfs/xfs_vnodeops.c     2011-07-22 11:18:10.000000000 +0200
9855 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_vnodeops.c        2011-07-01 11:35:35.000000000 +0200
9856 @@ -50,6 +50,78 @@
9857  #include "xfs_vnodeops.h"
9858  #include "xfs_trace.h"
9859
9860 +
9861 +STATIC void
9862 +xfs_get_inode_flags(
9863 +       xfs_inode_t     *ip)
9864 +{
9865 +       struct inode    *inode = VFS_I(ip);
9866 +       unsigned int    flags = inode->i_flags;
9867 +       unsigned int    vflags = inode->i_vflags;
9868 +
9869 +       if (flags & S_IMMUTABLE)
9870 +               ip->i_d.di_flags |= XFS_DIFLAG_IMMUTABLE;
9871 +       else
9872 +               ip->i_d.di_flags &= ~XFS_DIFLAG_IMMUTABLE;
9873 +       if (flags & S_IXUNLINK)
9874 +               ip->i_d.di_flags |= XFS_DIFLAG_IXUNLINK;
9875 +       else
9876 +               ip->i_d.di_flags &= ~XFS_DIFLAG_IXUNLINK;
9877 +
9878 +       if (vflags & V_BARRIER)
9879 +               ip->i_d.di_vflags |= XFS_DIVFLAG_BARRIER;
9880 +       else
9881 +               ip->i_d.di_vflags &= ~XFS_DIVFLAG_BARRIER;
9882 +       if (vflags & V_COW)
9883 +               ip->i_d.di_vflags |= XFS_DIVFLAG_COW;
9884 +       else
9885 +               ip->i_d.di_vflags &= ~XFS_DIVFLAG_COW;
9886 +}
9887 +
9888 +int
9889 +xfs_sync_flags(
9890 +       struct inode            *inode,
9891 +       int                     flags,
9892 +       int                     vflags)
9893 +{
9894 +       struct xfs_inode        *ip = XFS_I(inode);
9895 +       struct xfs_mount        *mp = ip->i_mount;
9896 +       struct xfs_trans        *tp;
9897 +       unsigned int            lock_flags = 0;
9898 +       int                     code;
9899 +
9900 +       tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
9901 +       code = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
9902 +       if (code)
9903 +               goto error_out;
9904 +
9905 +       xfs_ilock(ip, XFS_ILOCK_EXCL);
9906 +
9907 +       xfs_trans_ijoin(tp, ip);
9908 +
9909 +       inode->i_flags = flags;
9910 +       inode->i_vflags = vflags;
9911 +       xfs_get_inode_flags(ip);
9912 +
9913 +       xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
9914 +       xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
9915 +
9916 +       XFS_STATS_INC(xs_ig_attrchg);
9917 +
9918 +       if (mp->m_flags & XFS_MOUNT_WSYNC)
9919 +               xfs_trans_set_sync(tp);
9920 +       code = xfs_trans_commit(tp, 0);
9921 +       xfs_iunlock(ip, XFS_ILOCK_EXCL);
9922 +       return code;
9923 +
9924 +error_out:
9925 +       xfs_trans_cancel(tp, 0);
9926 +       if (lock_flags)
9927 +               xfs_iunlock(ip, XFS_ILOCK_EXCL);
9928 +       return code;
9929 +}
9930 +
9931 +
9932  int
9933  xfs_setattr(
9934         struct xfs_inode        *ip,
9935 @@ -65,6 +137,7 @@ xfs_setattr(
9936         uint                    commit_flags=0;
9937         uid_t                   uid=0, iuid=0;
9938         gid_t                   gid=0, igid=0;
9939 +       tag_t                   tag=0, itag=0;
9940         struct xfs_dquot        *udqp, *gdqp, *olddquot1, *olddquot2;
9941         int                     need_iolock = 1;
9942
9943 @@ -147,7 +220,7 @@ xfs_setattr(
9944         /*
9945          * Change file ownership.  Must be the owner or privileged.
9946          */
9947 -       if (mask & (ATTR_UID|ATTR_GID)) {
9948 +       if (mask & (ATTR_UID|ATTR_GID|ATTR_TAG)) {
9949                 /*
9950                  * These IDs could have changed since we last looked at them.
9951                  * But, we're assured that if the ownership did change
9952 @@ -156,8 +229,10 @@ xfs_setattr(
9953                  */
9954                 iuid = ip->i_d.di_uid;
9955                 igid = ip->i_d.di_gid;
9956 +               itag = ip->i_d.di_tag;
9957                 gid = (mask & ATTR_GID) ? iattr->ia_gid : igid;
9958                 uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;
9959 +               tag = (mask & ATTR_TAG) ? iattr->ia_tag : itag;
9960
9961                 /*
9962                  * Do a quota reservation only if uid/gid is actually
9963 @@ -165,7 +240,8 @@ xfs_setattr(
9964                  */
9965                 if (XFS_IS_QUOTA_RUNNING(mp) &&
9966                     ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
9967 -                    (XFS_IS_GQUOTA_ON(mp) && igid != gid))) {
9968 +                    (XFS_IS_GQUOTA_ON(mp) && igid != gid) ||
9969 +                    (XFS_IS_GQUOTA_ON(mp) && itag != tag))) {
9970                         ASSERT(tp);
9971                         code = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
9972                                                 capable(CAP_FOWNER) ?
9973 @@ -329,7 +405,7 @@ xfs_setattr(
9974         /*
9975          * Change file ownership.  Must be the owner or privileged.
9976          */
9977 -       if (mask & (ATTR_UID|ATTR_GID)) {
9978 +       if (mask & (ATTR_UID|ATTR_GID|ATTR_TAG)) {
9979                 /*
9980                  * CAP_FSETID overrides the following restrictions:
9981                  *
9982 @@ -345,6 +421,10 @@ xfs_setattr(
9983                  * Change the ownerships and register quota modifications
9984                  * in the transaction.
9985                  */
9986 +               if (itag != tag) {
9987 +                       ip->i_d.di_tag = tag;
9988 +                       inode->i_tag = tag;
9989 +               }
9990                 if (iuid != uid) {
9991                         if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_UQUOTA_ON(mp)) {
9992                                 ASSERT(mask & ATTR_UID);
9993 diff -NurpP --minimal linux-3.0/fs/xfs/xfs_vnodeops.h linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_vnodeops.h
9994 --- linux-3.0/fs/xfs/xfs_vnodeops.h     2011-05-22 16:17:54.000000000 +0200
9995 +++ linux-3.0-vs2.3.1-pre7/fs/xfs/xfs_vnodeops.h        2011-06-10 22:11:24.000000000 +0200
9996 @@ -13,6 +13,7 @@ struct xfs_inode;
9997  struct xfs_iomap;
9998
9999
10000 +int xfs_sync_xflags(struct xfs_inode *ip);
10001  int xfs_setattr(struct xfs_inode *ip, struct iattr *vap, int flags);
10002  #define        XFS_ATTR_DMI            0x01    /* invocation from a DMI function */
10003  #define        XFS_ATTR_NONBLOCK       0x02    /* return EAGAIN if operation would block */
10004 diff -NurpP --minimal linux-3.0/include/asm-generic/tlb.h linux-3.0-vs2.3.1-pre7/include/asm-generic/tlb.h
10005 --- linux-3.0/include/asm-generic/tlb.h 2011-07-22 11:18:10.000000000 +0200
10006 +++ linux-3.0-vs2.3.1-pre7/include/asm-generic/tlb.h    2011-06-10 22:11:24.000000000 +0200
10007 @@ -16,6 +16,7 @@
10008  #define _ASM_GENERIC__TLB_H
10009
10010  #include <linux/swap.h>
10011 +#include <linux/vs_memory.h>
10012  #include <asm/pgalloc.h>
10013  #include <asm/tlbflush.h>
10014
10015 diff -NurpP --minimal linux-3.0/include/linux/Kbuild linux-3.0-vs2.3.1-pre7/include/linux/Kbuild
10016 --- linux-3.0/include/linux/Kbuild      2011-07-22 11:18:10.000000000 +0200
10017 +++ linux-3.0-vs2.3.1-pre7/include/linux/Kbuild 2011-06-10 22:11:24.000000000 +0200
10018 @@ -17,6 +17,7 @@ header-y += netfilter_bridge/
10019  header-y += netfilter_ipv4/
10020  header-y += netfilter_ipv6/
10021  header-y += usb/
10022 +header-y += vserver/
10023  header-y += wimax/
10024
10025  objhdr-y += version.h
10026 diff -NurpP --minimal linux-3.0/include/linux/capability.h linux-3.0-vs2.3.1-pre7/include/linux/capability.h
10027 --- linux-3.0/include/linux/capability.h        2011-07-22 11:18:10.000000000 +0200
10028 +++ linux-3.0-vs2.3.1-pre7/include/linux/capability.h   2011-06-10 22:11:24.000000000 +0200
10029 @@ -279,6 +279,7 @@ struct cpu_vfs_cap_data {
10030     arbitrary SCSI commands */
10031  /* Allow setting encryption key on loopback filesystem */
10032  /* Allow setting zone reclaim policy */
10033 +/* Allow the selection of a security context */
10034
10035  #define CAP_SYS_ADMIN        21
10036
10037 @@ -362,7 +363,12 @@ struct cpu_vfs_cap_data {
10038
10039  #define CAP_LAST_CAP         CAP_WAKE_ALARM
10040
10041 -#define cap_valid(x) ((x) >= 0 && (x) <= CAP_LAST_CAP)
10042 +/* Allow context manipulations */
10043 +/* Allow changing context info on files */
10044 +
10045 +#define CAP_CONTEXT         63
10046 +
10047 +#define cap_valid(x) ((x) >= 0 && ((x) <= CAP_LAST_CAP || (x) == CAP_CONTEXT))
10048
10049  /*
10050   * Bit location of each capability (used by user-space library and kernel)
10051 diff -NurpP --minimal linux-3.0/include/linux/cred.h linux-3.0-vs2.3.1-pre7/include/linux/cred.h
10052 --- linux-3.0/include/linux/cred.h      2011-07-22 11:18:10.000000000 +0200
10053 +++ linux-3.0-vs2.3.1-pre7/include/linux/cred.h 2011-06-10 22:11:24.000000000 +0200
10054 @@ -156,6 +156,7 @@ extern void exit_creds(struct task_struc
10055  extern int copy_creds(struct task_struct *, unsigned long);
10056  extern const struct cred *get_task_cred(struct task_struct *);
10057  extern struct cred *cred_alloc_blank(void);
10058 +extern struct cred *__prepare_creds(const struct cred *);
10059  extern struct cred *prepare_creds(void);
10060  extern struct cred *prepare_exec_creds(void);
10061  extern int commit_creds(struct cred *);
10062 @@ -209,6 +210,31 @@ static inline void validate_process_cred
10063  }
10064  #endif
10065
10066 +static inline void set_cred_subscribers(struct cred *cred, int n)
10067 +{
10068 +#ifdef CONFIG_DEBUG_CREDENTIALS
10069 +       atomic_set(&cred->subscribers, n);
10070 +#endif
10071 +}
10072 +
10073 +static inline int read_cred_subscribers(const struct cred *cred)
10074 +{
10075 +#ifdef CONFIG_DEBUG_CREDENTIALS
10076 +       return atomic_read(&cred->subscribers);
10077 +#else
10078 +       return 0;
10079 +#endif
10080 +}
10081 +
10082 +static inline void alter_cred_subscribers(const struct cred *_cred, int n)
10083 +{
10084 +#ifdef CONFIG_DEBUG_CREDENTIALS
10085 +       struct cred *cred = (struct cred *) _cred;
10086 +
10087 +       atomic_add(n, &cred->subscribers);
10088 +#endif
10089 +}
10090 +
10091  /**
10092   * get_new_cred - Get a reference on a new set of credentials
10093   * @cred: The new credentials to reference
10094 diff -NurpP --minimal linux-3.0/include/linux/devpts_fs.h linux-3.0-vs2.3.1-pre7/include/linux/devpts_fs.h
10095 --- linux-3.0/include/linux/devpts_fs.h 2008-12-25 00:26:37.000000000 +0100
10096 +++ linux-3.0-vs2.3.1-pre7/include/linux/devpts_fs.h    2011-06-10 22:11:24.000000000 +0200
10097 @@ -45,5 +45,4 @@ static inline void devpts_pty_kill(struc
10098
10099  #endif
10100
10101 -
10102  #endif /* _LINUX_DEVPTS_FS_H */
10103 diff -NurpP --minimal linux-3.0/include/linux/ext2_fs.h linux-3.0-vs2.3.1-pre7/include/linux/ext2_fs.h
10104 --- linux-3.0/include/linux/ext2_fs.h   2010-02-25 11:52:07.000000000 +0100
10105 +++ linux-3.0-vs2.3.1-pre7/include/linux/ext2_fs.h      2011-06-10 22:11:24.000000000 +0200
10106 @@ -189,8 +189,12 @@ struct ext2_group_desc
10107  #define EXT2_NOTAIL_FL                 FS_NOTAIL_FL    /* file tail should not be merged */
10108  #define EXT2_DIRSYNC_FL                        FS_DIRSYNC_FL   /* dirsync behaviour (directories only) */
10109  #define EXT2_TOPDIR_FL                 FS_TOPDIR_FL    /* Top of directory hierarchies*/
10110 +#define EXT2_IXUNLINK_FL               FS_IXUNLINK_FL  /* Immutable invert on unlink */
10111  #define EXT2_RESERVED_FL               FS_RESERVED_FL  /* reserved for ext2 lib */
10112
10113 +#define EXT2_BARRIER_FL                        FS_BARRIER_FL   /* Barrier for chroot() */
10114 +#define EXT2_COW_FL                    FS_COW_FL       /* Copy on Write marker */
10115 +
10116  #define EXT2_FL_USER_VISIBLE           FS_FL_USER_VISIBLE      /* User visible flags */
10117  #define EXT2_FL_USER_MODIFIABLE                FS_FL_USER_MODIFIABLE   /* User modifiable flags */
10118
10119 @@ -274,7 +278,8 @@ struct ext2_inode {
10120                         __u16   i_pad1;
10121                         __le16  l_i_uid_high;   /* these 2 fields    */
10122                         __le16  l_i_gid_high;   /* were reserved2[0] */
10123 -                       __u32   l_i_reserved2;
10124 +                       __le16  l_i_tag;        /* Context Tag */
10125 +                       __u16   l_i_reserved2;
10126                 } linux2;
10127                 struct {
10128                         __u8    h_i_frag;       /* Fragment number */
10129 @@ -303,6 +308,7 @@ struct ext2_inode {
10130  #define i_gid_low      i_gid
10131  #define i_uid_high     osd2.linux2.l_i_uid_high
10132  #define i_gid_high     osd2.linux2.l_i_gid_high
10133 +#define i_raw_tag      osd2.linux2.l_i_tag
10134  #define i_reserved2    osd2.linux2.l_i_reserved2
10135  #endif
10136
10137 @@ -347,6 +353,7 @@ struct ext2_inode {
10138  #define EXT2_MOUNT_USRQUOTA            0x020000  /* user quota */
10139  #define EXT2_MOUNT_GRPQUOTA            0x040000  /* group quota */
10140  #define EXT2_MOUNT_RESERVATION         0x080000  /* Preallocation */
10141 +#define EXT2_MOUNT_TAGGED              (1<<24)   /* Enable Context Tags */
10142
10143
10144  #define clear_opt(o, opt)              o &= ~EXT2_MOUNT_##opt
10145 diff -NurpP --minimal linux-3.0/include/linux/ext3_fs.h linux-3.0-vs2.3.1-pre7/include/linux/ext3_fs.h
10146 --- linux-3.0/include/linux/ext3_fs.h   2011-07-22 11:18:10.000000000 +0200
10147 +++ linux-3.0-vs2.3.1-pre7/include/linux/ext3_fs.h      2011-06-10 22:11:24.000000000 +0200
10148 @@ -173,10 +173,14 @@ struct ext3_group_desc
10149  #define EXT3_NOTAIL_FL                 0x00008000 /* file tail should not be merged */
10150  #define EXT3_DIRSYNC_FL                        0x00010000 /* dirsync behaviour (directories only) */
10151  #define EXT3_TOPDIR_FL                 0x00020000 /* Top of directory hierarchies*/
10152 +#define EXT3_IXUNLINK_FL               0x08000000 /* Immutable invert on unlink */
10153  #define EXT3_RESERVED_FL               0x80000000 /* reserved for ext3 lib */
10154
10155 -#define EXT3_FL_USER_VISIBLE           0x0003DFFF /* User visible flags */
10156 -#define EXT3_FL_USER_MODIFIABLE                0x000380FF /* User modifiable flags */
10157 +#define EXT3_BARRIER_FL                        0x04000000 /* Barrier for chroot() */
10158 +#define EXT3_COW_FL                    0x20000000 /* Copy on Write marker */
10159 +
10160 +#define EXT3_FL_USER_VISIBLE           0x0103DFFF /* User visible flags */
10161 +#define EXT3_FL_USER_MODIFIABLE                0x010380FF /* User modifiable flags */
10162
10163  /* Flags that should be inherited by new inodes from their parent. */
10164  #define EXT3_FL_INHERITED (EXT3_SECRM_FL | EXT3_UNRM_FL | EXT3_COMPR_FL |\
10165 @@ -312,7 +316,8 @@ struct ext3_inode {
10166                         __u16   i_pad1;
10167                         __le16  l_i_uid_high;   /* these 2 fields    */
10168                         __le16  l_i_gid_high;   /* were reserved2[0] */
10169 -                       __u32   l_i_reserved2;
10170 +                       __le16  l_i_tag;        /* Context Tag */
10171 +                       __u16   l_i_reserved2;
10172                 } linux2;
10173                 struct {
10174                         __u8    h_i_frag;       /* Fragment number */
10175 @@ -343,6 +348,7 @@ struct ext3_inode {
10176  #define i_gid_low      i_gid
10177  #define i_uid_high     osd2.linux2.l_i_uid_high
10178  #define i_gid_high     osd2.linux2.l_i_gid_high
10179 +#define i_raw_tag      osd2.linux2.l_i_tag
10180  #define i_reserved2    osd2.linux2.l_i_reserved2
10181
10182  #elif defined(__GNU__)
10183 @@ -405,6 +411,7 @@ struct ext3_inode {
10184  #define EXT3_MOUNT_GRPQUOTA            0x200000 /* "old" group quota */
10185  #define EXT3_MOUNT_DATA_ERR_ABORT      0x400000 /* Abort on file data write
10186                                                   * error in ordered mode */
10187 +#define EXT3_MOUNT_TAGGED              (1<<24) /* Enable Context Tags */
10188
10189  /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */
10190  #ifndef _LINUX_EXT2_FS_H
10191 @@ -919,6 +926,7 @@ extern void ext3_get_inode_flags(struct
10192  extern void ext3_set_aops(struct inode *inode);
10193  extern int ext3_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
10194                        u64 start, u64 len);
10195 +extern int ext3_sync_flags(struct inode *, int, int);
10196
10197  /* ioctl.c */
10198  extern long ext3_ioctl(struct file *, unsigned int, unsigned long);
10199 diff -NurpP --minimal linux-3.0/include/linux/fs.h linux-3.0-vs2.3.1-pre7/include/linux/fs.h
10200 --- linux-3.0/include/linux/fs.h        2011-07-22 11:18:10.000000000 +0200
10201 +++ linux-3.0-vs2.3.1-pre7/include/linux/fs.h   2011-07-01 11:35:35.000000000 +0200
10202 @@ -208,6 +208,9 @@ struct inodes_stat_t {
10203  #define MS_KERNMOUNT   (1<<22) /* this is a kern_mount call */
10204  #define MS_I_VERSION   (1<<23) /* Update inode I_version field */
10205  #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
10206 +#define MS_TAGGED      (1<<25) /* use generic inode tagging */
10207 +#define MS_TAGID       (1<<26) /* use specific tag for this mount */
10208 +#define MS_NOTAGCHECK  (1<<27) /* don't check tags */
10209  #define MS_NOSEC       (1<<28)
10210  #define MS_BORN                (1<<29)
10211  #define MS_ACTIVE      (1<<30)
10212 @@ -239,6 +242,14 @@ struct inodes_stat_t {
10213  #define S_IMA          1024    /* Inode has an associated IMA struct */
10214  #define S_AUTOMOUNT    2048    /* Automount/referral quasi-directory */
10215  #define S_NOSEC                4096    /* no suid or xattr security attributes */
10216 +#define S_IXUNLINK     8192    /* Immutable Invert on unlink */
10217 +
10218 +/* Linux-VServer related Inode flags */
10219 +
10220 +#define V_VALID                1
10221 +#define V_XATTR                2
10222 +#define V_BARRIER      4       /* Barrier for chroot() */
10223 +#define V_COW          8       /* Copy on Write */
10224
10225  /*
10226   * Note that nosuid etc flags are inode-specific: setting some file-system
10227 @@ -261,12 +272,15 @@ struct inodes_stat_t {
10228  #define IS_DIRSYNC(inode)      (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
10229                                         ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
10230  #define IS_MANDLOCK(inode)     __IS_FLG(inode, MS_MANDLOCK)
10231 -#define IS_NOATIME(inode)   __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
10232 -#define IS_I_VERSION(inode)   __IS_FLG(inode, MS_I_VERSION)
10233 +#define IS_NOATIME(inode)      __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
10234 +#define IS_I_VERSION(inode)    __IS_FLG(inode, MS_I_VERSION)
10235 +#define IS_TAGGED(inode)       __IS_FLG(inode, MS_TAGGED)
10236
10237  #define IS_NOQUOTA(inode)      ((inode)->i_flags & S_NOQUOTA)
10238  #define IS_APPEND(inode)       ((inode)->i_flags & S_APPEND)
10239  #define IS_IMMUTABLE(inode)    ((inode)->i_flags & S_IMMUTABLE)
10240 +#define IS_IXUNLINK(inode)     ((inode)->i_flags & S_IXUNLINK)
10241 +#define IS_IXORUNLINK(inode)   ((IS_IXUNLINK(inode) ? S_IMMUTABLE : 0) ^ IS_IMMUTABLE(inode))
10242  #define IS_POSIXACL(inode)     __IS_FLG(inode, MS_POSIXACL)
10243
10244  #define IS_DEADDIR(inode)      ((inode)->i_flags & S_DEAD)
10245 @@ -277,6 +291,16 @@ struct inodes_stat_t {
10246  #define IS_AUTOMOUNT(inode)    ((inode)->i_flags & S_AUTOMOUNT)
10247  #define IS_NOSEC(inode)                ((inode)->i_flags & S_NOSEC)
10248
10249 +#define IS_BARRIER(inode)      (S_ISDIR((inode)->i_mode) && ((inode)->i_vflags & V_BARRIER))
10250 +
10251 +#ifdef CONFIG_VSERVER_COWBL
10252 +#  define IS_COW(inode)                (IS_IXUNLINK(inode) && IS_IMMUTABLE(inode))
10253 +#  define IS_COW_LINK(inode)   (S_ISREG((inode)->i_mode) && ((inode)->i_nlink > 1))
10254 +#else
10255 +#  define IS_COW(inode)                (0)
10256 +#  define IS_COW_LINK(inode)   (0)
10257 +#endif
10258 +
10259  /* the read-only stuff doesn't really belong here, but any other place is
10260     probably as bad and I don't want to create yet another include file. */
10261
10262 @@ -362,11 +386,14 @@ struct inodes_stat_t {
10263  #define FS_EXTENT_FL                   0x00080000 /* Extents */
10264  #define FS_DIRECTIO_FL                 0x00100000 /* Use direct i/o */
10265  #define FS_NOCOW_FL                    0x00800000 /* Do not cow file */
10266 +#define FS_IXUNLINK_FL                 0x08000000 /* Immutable invert on unlink */
10267  #define FS_RESERVED_FL                 0x80000000 /* reserved for ext2 lib */
10268
10269 -#define FS_FL_USER_VISIBLE             0x0003DFFF /* User visible flags */
10270 -#define FS_FL_USER_MODIFIABLE          0x000380FF /* User modifiable flags */
10271 -
10272 +#define FS_BARRIER_FL                  0x04000000 /* Barrier for chroot() */
10273 +#define FS_COW_FL                      0x20000000 /* Copy on Write marker */
10274 +
10275 +#define FS_FL_USER_VISIBLE             0x0103DFFF /* User visible flags */
10276 +#define FS_FL_USER_MODIFIABLE          0x010380FF /* User modifiable flags */
10277
10278  #define SYNC_FILE_RANGE_WAIT_BEFORE    1
10279  #define SYNC_FILE_RANGE_WRITE          2
10280 @@ -447,6 +474,7 @@ typedef void (dio_iodone_t)(struct kiocb
10281  #define ATTR_KILL_PRIV (1 << 14)
10282  #define ATTR_OPEN      (1 << 15) /* Truncating from open(O_TRUNC) */
10283  #define ATTR_TIMES_SET (1 << 16)
10284 +#define ATTR_TAG       (1 << 17)
10285
10286  /*
10287   * This is the Inode Attributes structure, used for notify_change().  It
10288 @@ -462,6 +490,7 @@ struct iattr {
10289         umode_t         ia_mode;
10290         uid_t           ia_uid;
10291         gid_t           ia_gid;
10292 +       tag_t           ia_tag;
10293         loff_t          ia_size;
10294         struct timespec ia_atime;
10295         struct timespec ia_mtime;
10296 @@ -475,6 +504,9 @@ struct iattr {
10297         struct file     *ia_file;
10298  };
10299
10300 +#define ATTR_FLAG_BARRIER      512     /* Barrier for chroot() */
10301 +#define ATTR_FLAG_IXUNLINK     1024    /* Immutable invert on unlink */
10302 +
10303  /*
10304   * Includes for diskquotas.
10305   */
10306 @@ -740,11 +772,13 @@ struct inode {
10307         umode_t                 i_mode;
10308         uid_t                   i_uid;
10309         gid_t                   i_gid;
10310 +       tag_t                   i_tag;
10311         const struct inode_operations   *i_op;
10312         struct super_block      *i_sb;
10313
10314         spinlock_t              i_lock; /* i_blocks, i_bytes, maybe i_size */
10315 -       unsigned int            i_flags;
10316 +       unsigned short          i_flags;
10317 +       unsigned short          i_vflags;
10318         unsigned long           i_state;
10319  #ifdef CONFIG_SECURITY
10320         void                    *i_security;
10321 @@ -766,6 +800,7 @@ struct inode {
10322         atomic_t                i_count;
10323         unsigned int            i_nlink;
10324         dev_t                   i_rdev;
10325 +       dev_t                   i_mdev;
10326         unsigned int            i_blkbits;
10327         u64                     i_version;
10328         loff_t                  i_size;
10329 @@ -890,12 +925,12 @@ static inline void i_size_write(struct i
10330
10331  static inline unsigned iminor(const struct inode *inode)
10332  {
10333 -       return MINOR(inode->i_rdev);
10334 +       return MINOR(inode->i_mdev);
10335  }
10336
10337  static inline unsigned imajor(const struct inode *inode)
10338  {
10339 -       return MAJOR(inode->i_rdev);
10340 +       return MAJOR(inode->i_mdev);
10341  }
10342
10343  extern struct block_device *I_BDEV(struct inode *inode);
10344 @@ -957,6 +992,7 @@ struct file {
10345         loff_t                  f_pos;
10346         struct fown_struct      f_owner;
10347         const struct cred       *f_cred;
10348 +       xid_t                   f_xid;
10349         struct file_ra_state    f_ra;
10350
10351         u64                     f_version;
10352 @@ -1101,6 +1137,7 @@ struct file_lock {
10353         struct file *fl_file;
10354         loff_t fl_start;
10355         loff_t fl_end;
10356 +       xid_t fl_xid;
10357
10358         struct fasync_struct *  fl_fasync; /* for lease break notifications */
10359         unsigned long fl_break_time;    /* for nonblocking lease breaks */
10360 @@ -1600,6 +1637,7 @@ struct inode_operations {
10361         ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
10362         ssize_t (*listxattr) (struct dentry *, char *, size_t);
10363         int (*removexattr) (struct dentry *, const char *);
10364 +       int (*sync_flags) (struct inode *, int, int);
10365         void (*truncate_range)(struct inode *, loff_t, loff_t);
10366         int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
10367                       u64 len);
10368 @@ -1618,6 +1656,7 @@ extern ssize_t vfs_readv(struct file *,
10369                 unsigned long, loff_t *);
10370  extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
10371                 unsigned long, loff_t *);
10372 +ssize_t vfs_sendfile(struct file *, struct file *, loff_t *, size_t, loff_t);
10373
10374  struct super_operations {
10375         struct inode *(*alloc_inode)(struct super_block *sb);
10376 @@ -2437,6 +2476,7 @@ extern int dcache_dir_open(struct inode
10377  extern int dcache_dir_close(struct inode *, struct file *);
10378  extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
10379  extern int dcache_readdir(struct file *, void *, filldir_t);
10380 +extern int dcache_readdir_filter(struct file *, void *, filldir_t, int (*)(struct dentry *));
10381  extern int simple_setattr(struct dentry *, struct iattr *);
10382  extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
10383  extern int simple_statfs(struct dentry *, struct kstatfs *);
10384 diff -NurpP --minimal linux-3.0/include/linux/gfs2_ondisk.h linux-3.0-vs2.3.1-pre7/include/linux/gfs2_ondisk.h
10385 --- linux-3.0/include/linux/gfs2_ondisk.h       2010-07-07 18:31:55.000000000 +0200
10386 +++ linux-3.0-vs2.3.1-pre7/include/linux/gfs2_ondisk.h  2011-06-10 22:11:24.000000000 +0200
10387 @@ -211,6 +211,9 @@ enum {
10388         gfs2fl_NoAtime          = 7,
10389         gfs2fl_Sync             = 8,
10390         gfs2fl_System           = 9,
10391 +       gfs2fl_IXUnlink         = 16,
10392 +       gfs2fl_Barrier          = 17,
10393 +       gfs2fl_Cow              = 18,
10394         gfs2fl_TruncInProg      = 29,
10395         gfs2fl_InheritDirectio  = 30,
10396         gfs2fl_InheritJdata     = 31,
10397 @@ -227,6 +230,9 @@ enum {
10398  #define GFS2_DIF_NOATIME               0x00000080
10399  #define GFS2_DIF_SYNC                  0x00000100
10400  #define GFS2_DIF_SYSTEM                        0x00000200 /* New in gfs2 */
10401 +#define GFS2_DIF_IXUNLINK              0x00010000
10402 +#define GFS2_DIF_BARRIER               0x00020000
10403 +#define GFS2_DIF_COW                   0x00040000
10404  #define GFS2_DIF_TRUNC_IN_PROG         0x20000000 /* New in gfs2 */
10405  #define GFS2_DIF_INHERIT_DIRECTIO      0x40000000
10406  #define GFS2_DIF_INHERIT_JDATA         0x80000000
10407 diff -NurpP --minimal linux-3.0/include/linux/if_tun.h linux-3.0-vs2.3.1-pre7/include/linux/if_tun.h
10408 --- linux-3.0/include/linux/if_tun.h    2010-08-02 16:52:54.000000000 +0200
10409 +++ linux-3.0-vs2.3.1-pre7/include/linux/if_tun.h       2011-06-10 22:11:24.000000000 +0200
10410 @@ -53,6 +53,7 @@
10411  #define TUNDETACHFILTER _IOW('T', 214, struct sock_fprog)
10412  #define TUNGETVNETHDRSZ _IOR('T', 215, int)
10413  #define TUNSETVNETHDRSZ _IOW('T', 216, int)
10414 +#define TUNSETNID     _IOW('T', 217, int)
10415
10416  /* TUNSETIFF ifr flags */
10417  #define IFF_TUN                0x0001
10418 diff -NurpP --minimal linux-3.0/include/linux/init_task.h linux-3.0-vs2.3.1-pre7/include/linux/init_task.h
10419 --- linux-3.0/include/linux/init_task.h 2011-07-22 11:18:10.000000000 +0200
10420 +++ linux-3.0-vs2.3.1-pre7/include/linux/init_task.h    2011-06-10 22:11:24.000000000 +0200
10421 @@ -193,6 +193,10 @@ extern struct cred init_cred;
10422         INIT_FTRACE_GRAPH                                               \
10423         INIT_TRACE_RECURSION                                            \
10424         INIT_TASK_RCU_PREEMPT(tsk)                                      \
10425 +       .xid            = 0,                                            \
10426 +       .vx_info        = NULL,                                         \
10427 +       .nid            = 0,                                            \
10428 +       .nx_info        = NULL,                                         \
10429  }
10430
10431
10432 diff -NurpP --minimal linux-3.0/include/linux/ipc.h linux-3.0-vs2.3.1-pre7/include/linux/ipc.h
10433 --- linux-3.0/include/linux/ipc.h       2009-12-03 20:02:55.000000000 +0100
10434 +++ linux-3.0-vs2.3.1-pre7/include/linux/ipc.h  2011-06-10 22:11:24.000000000 +0200
10435 @@ -91,6 +91,7 @@ struct kern_ipc_perm
10436         key_t           key;
10437         uid_t           uid;
10438         gid_t           gid;
10439 +       xid_t           xid;
10440         uid_t           cuid;
10441         gid_t           cgid;
10442         mode_t          mode;
10443 diff -NurpP --minimal linux-3.0/include/linux/ipc_namespace.h linux-3.0-vs2.3.1-pre7/include/linux/ipc_namespace.h
10444 --- linux-3.0/include/linux/ipc_namespace.h     2011-05-22 16:17:55.000000000 +0200
10445 +++ linux-3.0-vs2.3.1-pre7/include/linux/ipc_namespace.h        2011-06-13 14:09:44.000000000 +0200
10446 @@ -94,7 +94,8 @@ static inline int mq_init_ns(struct ipc_
10447
10448  #if defined(CONFIG_IPC_NS)
10449  extern struct ipc_namespace *copy_ipcs(unsigned long flags,
10450 -                                      struct task_struct *tsk);
10451 +                                      struct ipc_namespace *old_ns,
10452 +                                      struct user_namespace *user_ns);
10453  static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns)
10454  {
10455         if (ns)
10456 @@ -105,12 +106,13 @@ static inline struct ipc_namespace *get_
10457  extern void put_ipc_ns(struct ipc_namespace *ns);
10458  #else
10459  static inline struct ipc_namespace *copy_ipcs(unsigned long flags,
10460 -                                             struct task_struct *tsk)
10461 +                                             struct ipc_namespace *old_ns,
10462 +                                             struct user_namespace *user_ns)
10463  {
10464         if (flags & CLONE_NEWIPC)
10465                 return ERR_PTR(-EINVAL);
10466
10467 -       return tsk->nsproxy->ipc_ns;
10468 +       return old_ns;
10469  }
10470
10471  static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns)
10472 diff -NurpP --minimal linux-3.0/include/linux/loop.h linux-3.0-vs2.3.1-pre7/include/linux/loop.h
10473 --- linux-3.0/include/linux/loop.h      2009-09-10 15:26:25.000000000 +0200
10474 +++ linux-3.0-vs2.3.1-pre7/include/linux/loop.h 2011-06-10 22:11:24.000000000 +0200
10475 @@ -45,6 +45,7 @@ struct loop_device {
10476         struct loop_func_table *lo_encryption;
10477         __u32           lo_init[2];
10478         uid_t           lo_key_owner;   /* Who set the key */
10479 +       xid_t           lo_xid;
10480         int             (*ioctl)(struct loop_device *, int cmd,
10481                                  unsigned long arg);
10482
10483 diff -NurpP --minimal linux-3.0/include/linux/magic.h linux-3.0-vs2.3.1-pre7/include/linux/magic.h
10484 --- linux-3.0/include/linux/magic.h     2011-05-22 16:17:55.000000000 +0200
10485 +++ linux-3.0-vs2.3.1-pre7/include/linux/magic.h        2011-06-10 22:11:24.000000000 +0200
10486 @@ -3,7 +3,7 @@
10487
10488  #define ADFS_SUPER_MAGIC       0xadf5
10489  #define AFFS_SUPER_MAGIC       0xadff
10490 -#define AFS_SUPER_MAGIC                0x5346414F
10491 +#define AFS_SUPER_MAGIC                0x5346414F
10492  #define AUTOFS_SUPER_MAGIC     0x0187
10493  #define CODA_SUPER_MAGIC       0x73757245
10494  #define CRAMFS_MAGIC           0x28cd3d45      /* some random number */
10495 @@ -41,6 +41,7 @@
10496  #define NFS_SUPER_MAGIC                0x6969
10497  #define OPENPROM_SUPER_MAGIC   0x9fa1
10498  #define PROC_SUPER_MAGIC       0x9fa0
10499 +#define DEVPTS_SUPER_MAGIC     0x1cd1
10500  #define QNX4_SUPER_MAGIC       0x002f          /* qnx4 fs detection */
10501
10502  #define REISERFS_SUPER_MAGIC   0x52654973      /* used by gcc */
10503 diff -NurpP --minimal linux-3.0/include/linux/major.h linux-3.0-vs2.3.1-pre7/include/linux/major.h
10504 --- linux-3.0/include/linux/major.h     2009-09-10 15:26:25.000000000 +0200
10505 +++ linux-3.0-vs2.3.1-pre7/include/linux/major.h        2011-06-10 22:11:24.000000000 +0200
10506 @@ -15,6 +15,7 @@
10507  #define HD_MAJOR               IDE0_MAJOR
10508  #define PTY_SLAVE_MAJOR                3
10509  #define TTY_MAJOR              4
10510 +#define VROOT_MAJOR            4
10511  #define TTYAUX_MAJOR           5
10512  #define LP_MAJOR               6
10513  #define VCS_MAJOR              7
10514 diff -NurpP --minimal linux-3.0/include/linux/memcontrol.h linux-3.0-vs2.3.1-pre7/include/linux/memcontrol.h
10515 --- linux-3.0/include/linux/memcontrol.h        2011-07-22 11:18:10.000000000 +0200
10516 +++ linux-3.0-vs2.3.1-pre7/include/linux/memcontrol.h   2011-06-22 12:39:15.000000000 +0200
10517 @@ -86,6 +86,13 @@ extern struct mem_cgroup *try_get_mem_cg
10518  extern struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
10519  extern struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm);
10520
10521 +extern u64 mem_cgroup_res_read_u64(struct mem_cgroup *mem, int member);
10522 +extern u64 mem_cgroup_memsw_read_u64(struct mem_cgroup *mem, int member);
10523 +
10524 +extern s64 mem_cgroup_stat_read_cache(struct mem_cgroup *mem);
10525 +extern s64 mem_cgroup_stat_read_anon(struct mem_cgroup *mem);
10526 +extern s64 mem_cgroup_stat_read_mapped(struct mem_cgroup *mem);
10527 +
10528  static inline
10529  int mm_match_cgroup(const struct mm_struct *mm, const struct mem_cgroup *cgroup)
10530  {
10531 diff -NurpP --minimal linux-3.0/include/linux/mm_types.h linux-3.0-vs2.3.1-pre7/include/linux/mm_types.h
10532 --- linux-3.0/include/linux/mm_types.h  2011-07-22 11:18:11.000000000 +0200
10533 +++ linux-3.0-vs2.3.1-pre7/include/linux/mm_types.h     2011-06-10 22:11:24.000000000 +0200
10534 @@ -268,6 +268,7 @@ struct mm_struct {
10535
10536         /* Architecture-specific MM context */
10537         mm_context_t context;
10538 +       struct vx_info *mm_vx_info;
10539
10540         /* Swap token stuff */
10541         /*
10542 diff -NurpP --minimal linux-3.0/include/linux/mmzone.h linux-3.0-vs2.3.1-pre7/include/linux/mmzone.h
10543 --- linux-3.0/include/linux/mmzone.h    2011-07-22 11:18:11.000000000 +0200
10544 +++ linux-3.0-vs2.3.1-pre7/include/linux/mmzone.h       2011-07-01 11:35:35.000000000 +0200
10545 @@ -654,6 +654,13 @@ typedef struct pglist_data {
10546         __pgdat->node_start_pfn + __pgdat->node_spanned_pages;\
10547  })
10548
10549 +#define node_start_pfn(nid)    (NODE_DATA(nid)->node_start_pfn)
10550 +
10551 +#define node_end_pfn(nid) ({\
10552 +       pg_data_t *__pgdat = NODE_DATA(nid);\
10553 +       __pgdat->node_start_pfn + __pgdat->node_spanned_pages;\
10554 +})
10555 +
10556  #include <linux/memory_hotplug.h>
10557
10558  extern struct mutex zonelists_mutex;
10559 diff -NurpP --minimal linux-3.0/include/linux/mount.h linux-3.0-vs2.3.1-pre7/include/linux/mount.h
10560 --- linux-3.0/include/linux/mount.h     2011-03-15 18:07:39.000000000 +0100
10561 +++ linux-3.0-vs2.3.1-pre7/include/linux/mount.h        2011-06-10 22:11:24.000000000 +0200
10562 @@ -52,6 +52,9 @@ struct mnt_pcp {
10563         int mnt_writers;
10564  };
10565
10566 +#define MNT_TAGID      0x10000
10567 +#define MNT_NOTAG      0x20000
10568 +
10569  struct vfsmount {
10570         struct list_head mnt_hash;
10571         struct vfsmount *mnt_parent;    /* fs we are mounted on */
10572 @@ -86,6 +89,7 @@ struct vfsmount {
10573         int mnt_expiry_mark;            /* true if marked for expiry */
10574         int mnt_pinned;
10575         int mnt_ghosts;
10576 +       tag_t mnt_tag;                  /* tagging used for vfsmount */
10577  };
10578
10579  struct file; /* forward dec */
10580 diff -NurpP --minimal linux-3.0/include/linux/net.h linux-3.0-vs2.3.1-pre7/include/linux/net.h
10581 --- linux-3.0/include/linux/net.h       2011-07-22 11:18:11.000000000 +0200
10582 +++ linux-3.0-vs2.3.1-pre7/include/linux/net.h  2011-06-10 22:11:24.000000000 +0200
10583 @@ -72,6 +72,7 @@ struct net;
10584  #define SOCK_NOSPACE           2
10585  #define SOCK_PASSCRED          3
10586  #define SOCK_PASSSEC           4
10587 +#define SOCK_USER_SOCKET       5
10588
10589  #ifndef ARCH_HAS_SOCKET_TYPES
10590  /**
10591 diff -NurpP --minimal linux-3.0/include/linux/nfs_mount.h linux-3.0-vs2.3.1-pre7/include/linux/nfs_mount.h
10592 --- linux-3.0/include/linux/nfs_mount.h 2011-01-05 21:50:31.000000000 +0100
10593 +++ linux-3.0-vs2.3.1-pre7/include/linux/nfs_mount.h    2011-06-10 22:11:24.000000000 +0200
10594 @@ -63,7 +63,8 @@ struct nfs_mount_data {
10595  #define NFS_MOUNT_SECFLAVOUR   0x2000  /* 5 */
10596  #define NFS_MOUNT_NORDIRPLUS   0x4000  /* 5 */
10597  #define NFS_MOUNT_UNSHARED     0x8000  /* 5 */
10598 -#define NFS_MOUNT_FLAGMASK     0xFFFF
10599 +#define NFS_MOUNT_TAGGED       0x10000 /* context tagging */
10600 +#define NFS_MOUNT_FLAGMASK     0x1FFFF
10601
10602  /* The following are for internal use only */
10603  #define NFS_MOUNT_LOOKUP_CACHE_NONEG   0x10000
10604 diff -NurpP --minimal linux-3.0/include/linux/nsproxy.h linux-3.0-vs2.3.1-pre7/include/linux/nsproxy.h
10605 --- linux-3.0/include/linux/nsproxy.h   2011-07-22 11:18:11.000000000 +0200
10606 +++ linux-3.0-vs2.3.1-pre7/include/linux/nsproxy.h      2011-06-10 22:11:24.000000000 +0200
10607 @@ -3,6 +3,7 @@
10608
10609  #include <linux/spinlock.h>
10610  #include <linux/sched.h>
10611 +#include <linux/vserver/debug.h>
10612
10613  struct mnt_namespace;
10614  struct uts_namespace;
10615 @@ -63,22 +64,33 @@ static inline struct nsproxy *task_nspro
10616  }
10617
10618  int copy_namespaces(unsigned long flags, struct task_struct *tsk);
10619 +struct nsproxy *copy_nsproxy(struct nsproxy *orig);
10620  void exit_task_namespaces(struct task_struct *tsk);
10621  void switch_task_namespaces(struct task_struct *tsk, struct nsproxy *new);
10622  void free_nsproxy(struct nsproxy *ns);
10623  int unshare_nsproxy_namespaces(unsigned long, struct nsproxy **,
10624         struct fs_struct *);
10625
10626 -static inline void put_nsproxy(struct nsproxy *ns)
10627 +#define        get_nsproxy(n)  __get_nsproxy(n, __FILE__, __LINE__)
10628 +
10629 +static inline void __get_nsproxy(struct nsproxy *ns,
10630 +       const char *_file, int _line)
10631  {
10632 -       if (atomic_dec_and_test(&ns->count)) {
10633 -               free_nsproxy(ns);
10634 -       }
10635 +       vxlprintk(VXD_CBIT(space, 0), "get_nsproxy(%p[%u])",
10636 +               ns, atomic_read(&ns->count), _file, _line);
10637 +       atomic_inc(&ns->count);
10638  }
10639
10640 -static inline void get_nsproxy(struct nsproxy *ns)
10641 +#define        put_nsproxy(n)  __put_nsproxy(n, __FILE__, __LINE__)
10642 +
10643 +static inline void __put_nsproxy(struct nsproxy *ns,
10644 +       const char *_file, int _line)
10645  {
10646 -       atomic_inc(&ns->count);
10647 +       vxlprintk(VXD_CBIT(space, 0), "put_nsproxy(%p[%u])",
10648 +               ns, atomic_read(&ns->count), _file, _line);
10649 +       if (atomic_dec_and_test(&ns->count)) {
10650 +               free_nsproxy(ns);
10651 +       }
10652  }
10653
10654  #endif
10655 diff -NurpP --minimal linux-3.0/include/linux/pid.h linux-3.0-vs2.3.1-pre7/include/linux/pid.h
10656 --- linux-3.0/include/linux/pid.h       2011-07-22 11:18:11.000000000 +0200
10657 +++ linux-3.0-vs2.3.1-pre7/include/linux/pid.h  2011-06-10 22:11:24.000000000 +0200
10658 @@ -8,7 +8,8 @@ enum pid_type
10659         PIDTYPE_PID,
10660         PIDTYPE_PGID,
10661         PIDTYPE_SID,
10662 -       PIDTYPE_MAX
10663 +       PIDTYPE_MAX,
10664 +       PIDTYPE_REALPID
10665  };
10666
10667  /*
10668 @@ -171,6 +172,7 @@ static inline pid_t pid_nr(struct pid *p
10669  }
10670
10671  pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns);
10672 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns);
10673  pid_t pid_vnr(struct pid *pid);
10674
10675  #define do_each_pid_task(pid, type, task)                              \
10676 diff -NurpP --minimal linux-3.0/include/linux/proc_fs.h linux-3.0-vs2.3.1-pre7/include/linux/proc_fs.h
10677 --- linux-3.0/include/linux/proc_fs.h   2011-07-22 11:18:11.000000000 +0200
10678 +++ linux-3.0-vs2.3.1-pre7/include/linux/proc_fs.h      2011-06-10 22:11:24.000000000 +0200
10679 @@ -56,6 +56,7 @@ struct proc_dir_entry {
10680         nlink_t nlink;
10681         uid_t uid;
10682         gid_t gid;
10683 +       int vx_flags;
10684         loff_t size;
10685         const struct inode_operations *proc_iops;
10686         /*
10687 @@ -252,12 +253,18 @@ extern const struct proc_ns_operations n
10688  extern const struct proc_ns_operations utsns_operations;
10689  extern const struct proc_ns_operations ipcns_operations;
10690
10691 +struct vx_info;
10692 +struct nx_info;
10693 +
10694  union proc_op {
10695         int (*proc_get_link)(struct inode *, struct path *);
10696         int (*proc_read)(struct task_struct *task, char *page);
10697         int (*proc_show)(struct seq_file *m,
10698                 struct pid_namespace *ns, struct pid *pid,
10699                 struct task_struct *task);
10700 +       int (*proc_vs_read)(char *page);
10701 +       int (*proc_vxi_read)(struct vx_info *vxi, char *page);
10702 +       int (*proc_nxi_read)(struct nx_info *nxi, char *page);
10703  };
10704
10705  struct ctl_table_header;
10706 @@ -265,6 +272,7 @@ struct ctl_table;
10707
10708  struct proc_inode {
10709         struct pid *pid;
10710 +       int vx_flags;
10711         int fd;
10712         union proc_op op;
10713         struct proc_dir_entry *pde;
10714 diff -NurpP --minimal linux-3.0/include/linux/quotaops.h linux-3.0-vs2.3.1-pre7/include/linux/quotaops.h
10715 --- linux-3.0/include/linux/quotaops.h  2011-05-22 16:17:57.000000000 +0200
10716 +++ linux-3.0-vs2.3.1-pre7/include/linux/quotaops.h     2011-06-10 22:11:24.000000000 +0200
10717 @@ -8,6 +8,7 @@
10718  #define _LINUX_QUOTAOPS_
10719
10720  #include <linux/fs.h>
10721 +#include <linux/vs_dlimit.h>
10722
10723  #define DQUOT_SPACE_WARN       0x1
10724  #define DQUOT_SPACE_RESERVE    0x2
10725 @@ -204,11 +205,12 @@ static inline void dquot_drop(struct ino
10726
10727  static inline int dquot_alloc_inode(const struct inode *inode)
10728  {
10729 -       return 0;
10730 +       return dl_alloc_inode(inode);
10731  }
10732
10733  static inline void dquot_free_inode(const struct inode *inode)
10734  {
10735 +       dl_free_inode(inode);
10736  }
10737
10738  static inline int dquot_transfer(struct inode *inode, struct iattr *iattr)
10739 @@ -219,6 +221,10 @@ static inline int dquot_transfer(struct
10740  static inline int __dquot_alloc_space(struct inode *inode, qsize_t number,
10741                 int flags)
10742  {
10743 +       int ret = 0;
10744 +
10745 +       if ((ret = dl_alloc_space(inode, number)))
10746 +               return ret;
10747         if (!(flags & DQUOT_SPACE_RESERVE))
10748                 inode_add_bytes(inode, number);
10749         return 0;
10750 @@ -229,6 +235,7 @@ static inline void __dquot_free_space(st
10751  {
10752         if (!(flags & DQUOT_SPACE_RESERVE))
10753                 inode_sub_bytes(inode, number);
10754 +       dl_free_space(inode, number);
10755  }
10756
10757  static inline int dquot_claim_space_nodirty(struct inode *inode, qsize_t number)
10758 diff -NurpP --minimal linux-3.0/include/linux/reboot.h linux-3.0-vs2.3.1-pre7/include/linux/reboot.h
10759 --- linux-3.0/include/linux/reboot.h    2010-07-07 18:31:56.000000000 +0200
10760 +++ linux-3.0-vs2.3.1-pre7/include/linux/reboot.h       2011-06-10 22:11:24.000000000 +0200
10761 @@ -33,6 +33,7 @@
10762  #define        LINUX_REBOOT_CMD_RESTART2       0xA1B2C3D4
10763  #define        LINUX_REBOOT_CMD_SW_SUSPEND     0xD000FCE2
10764  #define        LINUX_REBOOT_CMD_KEXEC          0x45584543
10765 +#define        LINUX_REBOOT_CMD_OOM            0xDEADBEEF
10766
10767
10768  #ifdef __KERNEL__
10769 diff -NurpP --minimal linux-3.0/include/linux/reiserfs_fs.h linux-3.0-vs2.3.1-pre7/include/linux/reiserfs_fs.h
10770 --- linux-3.0/include/linux/reiserfs_fs.h       2011-05-22 16:17:58.000000000 +0200
10771 +++ linux-3.0-vs2.3.1-pre7/include/linux/reiserfs_fs.h  2011-06-10 22:11:24.000000000 +0200
10772 @@ -976,6 +976,11 @@ struct stat_data_v1 {
10773  #define REISERFS_COMPR_FL     FS_COMPR_FL
10774  #define REISERFS_NOTAIL_FL    FS_NOTAIL_FL
10775
10776 +/* unfortunately reiserfs sdattr is only 16 bit */
10777 +#define REISERFS_IXUNLINK_FL  (FS_IXUNLINK_FL >> 16)
10778 +#define REISERFS_BARRIER_FL   (FS_BARRIER_FL >> 16)
10779 +#define REISERFS_COW_FL       (FS_COW_FL >> 16)
10780 +
10781  /* persistent flags that file inherits from the parent directory */
10782  #define REISERFS_INHERIT_MASK ( REISERFS_IMMUTABLE_FL |        \
10783                                 REISERFS_SYNC_FL |      \
10784 @@ -985,6 +990,9 @@ struct stat_data_v1 {
10785                                 REISERFS_COMPR_FL |     \
10786                                 REISERFS_NOTAIL_FL )
10787
10788 +#define REISERFS_FL_USER_VISIBLE       0x80FF
10789 +#define REISERFS_FL_USER_MODIFIABLE    0x80FF
10790 +
10791  /* Stat Data on disk (reiserfs version of UFS disk inode minus the
10792     address blocks) */
10793  struct stat_data {
10794 @@ -2073,6 +2081,7 @@ static inline void reiserfs_update_sd(st
10795  void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode);
10796  void i_attrs_to_sd_attrs(struct inode *inode, __u16 * sd_attrs);
10797  int reiserfs_setattr(struct dentry *dentry, struct iattr *attr);
10798 +int reiserfs_sync_flags(struct inode *inode, int, int);
10799
10800  int __reiserfs_write_begin(struct page *page, unsigned from, unsigned len);
10801
10802 diff -NurpP --minimal linux-3.0/include/linux/reiserfs_fs_sb.h linux-3.0-vs2.3.1-pre7/include/linux/reiserfs_fs_sb.h
10803 --- linux-3.0/include/linux/reiserfs_fs_sb.h    2010-02-25 11:52:07.000000000 +0100
10804 +++ linux-3.0-vs2.3.1-pre7/include/linux/reiserfs_fs_sb.h       2011-06-10 22:11:24.000000000 +0200
10805 @@ -476,6 +476,7 @@ enum reiserfs_mount_options {
10806         REISERFS_EXPOSE_PRIVROOT,
10807         REISERFS_BARRIER_NONE,
10808         REISERFS_BARRIER_FLUSH,
10809 +       REISERFS_TAGGED,
10810
10811         /* Actions on error */
10812         REISERFS_ERROR_PANIC,
10813 diff -NurpP --minimal linux-3.0/include/linux/sched.h linux-3.0-vs2.3.1-pre7/include/linux/sched.h
10814 --- linux-3.0/include/linux/sched.h     2011-07-22 11:18:11.000000000 +0200
10815 +++ linux-3.0-vs2.3.1-pre7/include/linux/sched.h        2011-07-22 11:20:39.000000000 +0200
10816 @@ -1406,6 +1406,14 @@ struct task_struct {
10817  #endif
10818         seccomp_t seccomp;
10819
10820 +/* vserver context data */
10821 +       struct vx_info *vx_info;
10822 +       struct nx_info *nx_info;
10823 +
10824 +       xid_t xid;
10825 +       nid_t nid;
10826 +       tag_t tag;
10827 +
10828  /* Thread group tracking */
10829         u32 parent_exec_id;
10830         u32 self_exec_id;
10831 @@ -1649,6 +1657,11 @@ struct pid_namespace;
10832  pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
10833                         struct pid_namespace *ns);
10834
10835 +#include <linux/vserver/base.h>
10836 +#include <linux/vserver/context.h>
10837 +#include <linux/vserver/debug.h>
10838 +#include <linux/vserver/pid.h>
10839 +
10840  static inline pid_t task_pid_nr(struct task_struct *tsk)
10841  {
10842         return tsk->pid;
10843 @@ -1662,7 +1675,8 @@ static inline pid_t task_pid_nr_ns(struc
10844
10845  static inline pid_t task_pid_vnr(struct task_struct *tsk)
10846  {
10847 -       return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
10848 +       // return __task_pid_nr_ns(tsk, PIDTYPE_PID, NULL);
10849 +       return vx_map_pid(__task_pid_nr_ns(tsk, PIDTYPE_PID, NULL));
10850  }
10851
10852
10853 @@ -1675,7 +1689,7 @@ pid_t task_tgid_nr_ns(struct task_struct
10854
10855  static inline pid_t task_tgid_vnr(struct task_struct *tsk)
10856  {
10857 -       return pid_vnr(task_tgid(tsk));
10858 +       return vx_map_tgid(pid_vnr(task_tgid(tsk)));
10859  }
10860
10861
10862 diff -NurpP --minimal linux-3.0/include/linux/shmem_fs.h linux-3.0-vs2.3.1-pre7/include/linux/shmem_fs.h
10863 --- linux-3.0/include/linux/shmem_fs.h  2011-07-22 11:18:11.000000000 +0200
10864 +++ linux-3.0-vs2.3.1-pre7/include/linux/shmem_fs.h     2011-07-01 11:35:35.000000000 +0200
10865 @@ -12,6 +12,9 @@
10866
10867  #define SHMEM_SYMLINK_INLINE_LEN (SHMEM_NR_DIRECT * sizeof(swp_entry_t))
10868
10869 +#define TMPFS_SUPER_MAGIC      0x01021994
10870 +
10871 +
10872  struct shmem_inode_info {
10873         spinlock_t              lock;
10874         unsigned long           flags;
10875 diff -NurpP --minimal linux-3.0/include/linux/stat.h linux-3.0-vs2.3.1-pre7/include/linux/stat.h
10876 --- linux-3.0/include/linux/stat.h      2008-12-25 00:26:37.000000000 +0100
10877 +++ linux-3.0-vs2.3.1-pre7/include/linux/stat.h 2011-06-10 22:11:24.000000000 +0200
10878 @@ -66,6 +66,7 @@ struct kstat {
10879         unsigned int    nlink;
10880         uid_t           uid;
10881         gid_t           gid;
10882 +       tag_t           tag;
10883         dev_t           rdev;
10884         loff_t          size;
10885         struct timespec  atime;
10886 diff -NurpP --minimal linux-3.0/include/linux/sunrpc/auth.h linux-3.0-vs2.3.1-pre7/include/linux/sunrpc/auth.h
10887 --- linux-3.0/include/linux/sunrpc/auth.h       2011-03-15 18:07:39.000000000 +0100
10888 +++ linux-3.0-vs2.3.1-pre7/include/linux/sunrpc/auth.h  2011-06-10 22:11:24.000000000 +0200
10889 @@ -25,6 +25,7 @@
10890  struct auth_cred {
10891         uid_t   uid;
10892         gid_t   gid;
10893 +       tag_t   tag;
10894         struct group_info *group_info;
10895         unsigned char machine_cred : 1;
10896  };
10897 diff -NurpP --minimal linux-3.0/include/linux/sunrpc/clnt.h linux-3.0-vs2.3.1-pre7/include/linux/sunrpc/clnt.h
10898 --- linux-3.0/include/linux/sunrpc/clnt.h       2011-05-22 16:17:58.000000000 +0200
10899 +++ linux-3.0-vs2.3.1-pre7/include/linux/sunrpc/clnt.h  2011-06-10 22:11:24.000000000 +0200
10900 @@ -49,7 +49,8 @@ struct rpc_clnt {
10901         unsigned int            cl_softrtry : 1,/* soft timeouts */
10902                                 cl_discrtry : 1,/* disconnect before retry */
10903                                 cl_autobind : 1,/* use getport() */
10904 -                               cl_chatty   : 1;/* be verbose */
10905 +                               cl_chatty   : 1,/* be verbose */
10906 +                               cl_tag      : 1;/* context tagging */
10907
10908         struct rpc_rtt *        cl_rtt;         /* RTO estimator data */
10909         const struct rpc_timeout *cl_timeout;   /* Timeout strategy */
10910 diff -NurpP --minimal linux-3.0/include/linux/syscalls.h linux-3.0-vs2.3.1-pre7/include/linux/syscalls.h
10911 --- linux-3.0/include/linux/syscalls.h  2011-07-22 11:18:11.000000000 +0200
10912 +++ linux-3.0-vs2.3.1-pre7/include/linux/syscalls.h     2011-06-10 22:11:24.000000000 +0200
10913 @@ -483,6 +483,8 @@ asmlinkage long sys_symlink(const char _
10914  asmlinkage long sys_unlink(const char __user *pathname);
10915  asmlinkage long sys_rename(const char __user *oldname,
10916                                 const char __user *newname);
10917 +asmlinkage long sys_copyfile(const char __user *from, const char __user *to,
10918 +                               umode_t mode);
10919  asmlinkage long sys_chmod(const char __user *filename, mode_t mode);
10920  asmlinkage long sys_fchmod(unsigned int fd, mode_t mode);
10921
10922 diff -NurpP --minimal linux-3.0/include/linux/sysctl.h linux-3.0-vs2.3.1-pre7/include/linux/sysctl.h
10923 --- linux-3.0/include/linux/sysctl.h    2011-03-15 18:07:40.000000000 +0100
10924 +++ linux-3.0-vs2.3.1-pre7/include/linux/sysctl.h       2011-06-10 22:11:24.000000000 +0200
10925 @@ -60,6 +60,7 @@ enum
10926         CTL_ABI=9,              /* Binary emulation */
10927         CTL_CPU=10,             /* CPU stuff (speed scaling, etc) */
10928         CTL_ARLAN=254,          /* arlan wireless driver */
10929 +       CTL_VSERVER=4242,       /* Linux-VServer debug */
10930         CTL_S390DBF=5677,       /* s390 debug */
10931         CTL_SUNRPC=7249,        /* sunrpc debug */
10932         CTL_PM=9899,            /* frv power management */
10933 @@ -94,6 +95,7 @@ enum
10934
10935         KERN_PANIC=15,          /* int: panic timeout */
10936         KERN_REALROOTDEV=16,    /* real root device to mount after initrd */
10937 +       KERN_VSHELPER=17,       /* string: path to vshelper policy agent */
10938
10939         KERN_SPARC_REBOOT=21,   /* reboot command on Sparc */
10940         KERN_CTLALTDEL=22,      /* int: allow ctl-alt-del to reboot */
10941 diff -NurpP --minimal linux-3.0/include/linux/sysfs.h linux-3.0-vs2.3.1-pre7/include/linux/sysfs.h
10942 --- linux-3.0/include/linux/sysfs.h     2011-07-22 11:18:11.000000000 +0200
10943 +++ linux-3.0-vs2.3.1-pre7/include/linux/sysfs.h        2011-06-22 12:39:15.000000000 +0200
10944 @@ -19,6 +19,8 @@
10945  #include <linux/kobject_ns.h>
10946  #include <asm/atomic.h>
10947
10948 +#define SYSFS_SUPER_MAGIC      0x62656572
10949 +
10950  struct kobject;
10951  struct module;
10952  enum kobj_ns_type;
10953 diff -NurpP --minimal linux-3.0/include/linux/time.h linux-3.0-vs2.3.1-pre7/include/linux/time.h
10954 --- linux-3.0/include/linux/time.h      2011-07-22 11:18:11.000000000 +0200
10955 +++ linux-3.0-vs2.3.1-pre7/include/linux/time.h 2011-06-10 22:11:24.000000000 +0200
10956 @@ -256,6 +256,9 @@ static __always_inline void timespec_add
10957         a->tv_sec += __iter_div_u64_rem(a->tv_nsec + ns, NSEC_PER_SEC, &ns);
10958         a->tv_nsec = ns;
10959  }
10960 +
10961 +#include <linux/vs_time.h>
10962 +
10963  #endif /* __KERNEL__ */
10964
10965  #define NFDBITS                        __NFDBITS
10966 diff -NurpP --minimal linux-3.0/include/linux/types.h linux-3.0-vs2.3.1-pre7/include/linux/types.h
10967 --- linux-3.0/include/linux/types.h     2011-05-22 16:17:58.000000000 +0200
10968 +++ linux-3.0-vs2.3.1-pre7/include/linux/types.h        2011-06-10 22:11:24.000000000 +0200
10969 @@ -40,6 +40,9 @@ typedef __kernel_uid32_t      uid_t;
10970  typedef __kernel_gid32_t       gid_t;
10971  typedef __kernel_uid16_t        uid16_t;
10972  typedef __kernel_gid16_t        gid16_t;
10973 +typedef unsigned int           xid_t;
10974 +typedef unsigned int           nid_t;
10975 +typedef unsigned int           tag_t;
10976
10977  typedef unsigned long          uintptr_t;
10978
10979 diff -NurpP --minimal linux-3.0/include/linux/utsname.h linux-3.0-vs2.3.1-pre7/include/linux/utsname.h
10980 --- linux-3.0/include/linux/utsname.h   2011-05-22 16:17:58.000000000 +0200
10981 +++ linux-3.0-vs2.3.1-pre7/include/linux/utsname.h      2011-06-13 14:36:48.000000000 +0200
10982 @@ -54,7 +54,8 @@ static inline void get_uts_ns(struct uts
10983  }
10984
10985  extern struct uts_namespace *copy_utsname(unsigned long flags,
10986 -                                         struct task_struct *tsk);
10987 +                                         struct uts_namespace *old_ns,
10988 +                                         struct user_namespace *user_ns);
10989  extern void free_uts_ns(struct kref *kref);
10990
10991  static inline void put_uts_ns(struct uts_namespace *ns)
10992 @@ -71,12 +72,13 @@ static inline void put_uts_ns(struct uts
10993  }
10994
10995  static inline struct uts_namespace *copy_utsname(unsigned long flags,
10996 -                                                struct task_struct *tsk)
10997 +                                                struct uts_namespace *old_ns,
10998 +                                                struct user_namespace *user_ns)
10999  {
11000         if (flags & CLONE_NEWUTS)
11001                 return ERR_PTR(-EINVAL);
11002
11003 -       return tsk->nsproxy->uts_ns;
11004 +       return old_ns;
11005  }
11006  #endif
11007
11008 diff -NurpP --minimal linux-3.0/include/linux/vroot.h linux-3.0-vs2.3.1-pre7/include/linux/vroot.h
11009 --- linux-3.0/include/linux/vroot.h     1970-01-01 01:00:00.000000000 +0100
11010 +++ linux-3.0-vs2.3.1-pre7/include/linux/vroot.h        2011-06-10 22:11:24.000000000 +0200
11011 @@ -0,0 +1,51 @@
11012 +
11013 +/*
11014 + * include/linux/vroot.h
11015 + *
11016 + * written by Herbert Pötzl, 9/11/2002
11017 + * ported to 2.6 by Herbert Pötzl, 30/12/2004
11018 + *
11019 + * Copyright (C) 2002-2007 by Herbert Pötzl.
11020 + * Redistribution of this file is permitted under the
11021 + * GNU General Public License.
11022 + */
11023 +
11024 +#ifndef _LINUX_VROOT_H
11025 +#define _LINUX_VROOT_H
11026 +
11027 +
11028 +#ifdef __KERNEL__
11029 +
11030 +/* Possible states of device */
11031 +enum {
11032 +       Vr_unbound,
11033 +       Vr_bound,
11034 +};
11035 +
11036 +struct vroot_device {
11037 +       int             vr_number;
11038 +       int             vr_refcnt;
11039 +
11040 +       struct semaphore        vr_ctl_mutex;
11041 +       struct block_device    *vr_device;
11042 +       int                     vr_state;
11043 +};
11044 +
11045 +
11046 +typedef struct block_device *(vroot_grb_func)(struct block_device *);
11047 +
11048 +extern int register_vroot_grb(vroot_grb_func *);
11049 +extern int unregister_vroot_grb(vroot_grb_func *);
11050 +
11051 +#endif /* __KERNEL__ */
11052 +
11053 +#define MAX_VROOT_DEFAULT      8
11054 +
11055 +/*
11056 + * IOCTL commands --- we will commandeer 0x56 ('V')
11057 + */
11058 +
11059 +#define VROOT_SET_DEV          0x5600
11060 +#define VROOT_CLR_DEV          0x5601
11061 +
11062 +#endif /* _LINUX_VROOT_H */
11063 diff -NurpP --minimal linux-3.0/include/linux/vs_base.h linux-3.0-vs2.3.1-pre7/include/linux/vs_base.h
11064 --- linux-3.0/include/linux/vs_base.h   1970-01-01 01:00:00.000000000 +0100
11065 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_base.h      2011-06-10 22:11:24.000000000 +0200
11066 @@ -0,0 +1,10 @@
11067 +#ifndef _VS_BASE_H
11068 +#define _VS_BASE_H
11069 +
11070 +#include "vserver/base.h"
11071 +#include "vserver/check.h"
11072 +#include "vserver/debug.h"
11073 +
11074 +#else
11075 +#warning duplicate inclusion
11076 +#endif
11077 diff -NurpP --minimal linux-3.0/include/linux/vs_context.h linux-3.0-vs2.3.1-pre7/include/linux/vs_context.h
11078 --- linux-3.0/include/linux/vs_context.h        1970-01-01 01:00:00.000000000 +0100
11079 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_context.h   2011-06-10 22:11:24.000000000 +0200
11080 @@ -0,0 +1,242 @@
11081 +#ifndef _VS_CONTEXT_H
11082 +#define _VS_CONTEXT_H
11083 +
11084 +#include "vserver/base.h"
11085 +#include "vserver/check.h"
11086 +#include "vserver/context.h"
11087 +#include "vserver/history.h"
11088 +#include "vserver/debug.h"
11089 +
11090 +#include <linux/sched.h>
11091 +
11092 +
11093 +#define get_vx_info(i) __get_vx_info(i, __FILE__, __LINE__, __HERE__)
11094 +
11095 +static inline struct vx_info *__get_vx_info(struct vx_info *vxi,
11096 +       const char *_file, int _line, void *_here)
11097 +{
11098 +       if (!vxi)
11099 +               return NULL;
11100 +
11101 +       vxlprintk(VXD_CBIT(xid, 2), "get_vx_info(%p[#%d.%d])",
11102 +               vxi, vxi ? vxi->vx_id : 0,
11103 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11104 +               _file, _line);
11105 +       __vxh_get_vx_info(vxi, _here);
11106 +
11107 +       atomic_inc(&vxi->vx_usecnt);
11108 +       return vxi;
11109 +}
11110 +
11111 +
11112 +extern void free_vx_info(struct vx_info *);
11113 +
11114 +#define put_vx_info(i) __put_vx_info(i, __FILE__, __LINE__, __HERE__)
11115 +
11116 +static inline void __put_vx_info(struct vx_info *vxi,
11117 +       const char *_file, int _line, void *_here)
11118 +{
11119 +       if (!vxi)
11120 +               return;
11121 +
11122 +       vxlprintk(VXD_CBIT(xid, 2), "put_vx_info(%p[#%d.%d])",
11123 +               vxi, vxi ? vxi->vx_id : 0,
11124 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11125 +               _file, _line);
11126 +       __vxh_put_vx_info(vxi, _here);
11127 +
11128 +       if (atomic_dec_and_test(&vxi->vx_usecnt))
11129 +               free_vx_info(vxi);
11130 +}
11131 +
11132 +
11133 +#define init_vx_info(p, i) \
11134 +       __init_vx_info(p, i, __FILE__, __LINE__, __HERE__)
11135 +
11136 +static inline void __init_vx_info(struct vx_info **vxp, struct vx_info *vxi,
11137 +       const char *_file, int _line, void *_here)
11138 +{
11139 +       if (vxi) {
11140 +               vxlprintk(VXD_CBIT(xid, 3),
11141 +                       "init_vx_info(%p[#%d.%d])",
11142 +                       vxi, vxi ? vxi->vx_id : 0,
11143 +                       vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11144 +                       _file, _line);
11145 +               __vxh_init_vx_info(vxi, vxp, _here);
11146 +
11147 +               atomic_inc(&vxi->vx_usecnt);
11148 +       }
11149 +       *vxp = vxi;
11150 +}
11151 +
11152 +
11153 +#define set_vx_info(p, i) \
11154 +       __set_vx_info(p, i, __FILE__, __LINE__, __HERE__)
11155 +
11156 +static inline void __set_vx_info(struct vx_info **vxp, struct vx_info *vxi,
11157 +       const char *_file, int _line, void *_here)
11158 +{
11159 +       struct vx_info *vxo;
11160 +
11161 +       if (!vxi)
11162 +               return;
11163 +
11164 +       vxlprintk(VXD_CBIT(xid, 3), "set_vx_info(%p[#%d.%d])",
11165 +               vxi, vxi ? vxi->vx_id : 0,
11166 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11167 +               _file, _line);
11168 +       __vxh_set_vx_info(vxi, vxp, _here);
11169 +
11170 +       atomic_inc(&vxi->vx_usecnt);
11171 +       vxo = xchg(vxp, vxi);
11172 +       BUG_ON(vxo);
11173 +}
11174 +
11175 +
11176 +#define clr_vx_info(p) __clr_vx_info(p, __FILE__, __LINE__, __HERE__)
11177 +
11178 +static inline void __clr_vx_info(struct vx_info **vxp,
11179 +       const char *_file, int _line, void *_here)
11180 +{
11181 +       struct vx_info *vxo;
11182 +
11183 +       vxo = xchg(vxp, NULL);
11184 +       if (!vxo)
11185 +               return;
11186 +
11187 +       vxlprintk(VXD_CBIT(xid, 3), "clr_vx_info(%p[#%d.%d])",
11188 +               vxo, vxo ? vxo->vx_id : 0,
11189 +               vxo ? atomic_read(&vxo->vx_usecnt) : 0,
11190 +               _file, _line);
11191 +       __vxh_clr_vx_info(vxo, vxp, _here);
11192 +
11193 +       if (atomic_dec_and_test(&vxo->vx_usecnt))
11194 +               free_vx_info(vxo);
11195 +}
11196 +
11197 +
11198 +#define claim_vx_info(v, p) \
11199 +       __claim_vx_info(v, p, __FILE__, __LINE__, __HERE__)
11200 +
11201 +static inline void __claim_vx_info(struct vx_info *vxi,
11202 +       struct task_struct *task,
11203 +       const char *_file, int _line, void *_here)
11204 +{
11205 +       vxlprintk(VXD_CBIT(xid, 3), "claim_vx_info(%p[#%d.%d.%d]) %p",
11206 +               vxi, vxi ? vxi->vx_id : 0,
11207 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11208 +               vxi ? atomic_read(&vxi->vx_tasks) : 0,
11209 +               task, _file, _line);
11210 +       __vxh_claim_vx_info(vxi, task, _here);
11211 +
11212 +       atomic_inc(&vxi->vx_tasks);
11213 +}
11214 +
11215 +
11216 +extern void unhash_vx_info(struct vx_info *);
11217 +
11218 +#define release_vx_info(v, p) \
11219 +       __release_vx_info(v, p, __FILE__, __LINE__, __HERE__)
11220 +
11221 +static inline void __release_vx_info(struct vx_info *vxi,
11222 +       struct task_struct *task,
11223 +       const char *_file, int _line, void *_here)
11224 +{
11225 +       vxlprintk(VXD_CBIT(xid, 3), "release_vx_info(%p[#%d.%d.%d]) %p",
11226 +               vxi, vxi ? vxi->vx_id : 0,
11227 +               vxi ? atomic_read(&vxi->vx_usecnt) : 0,
11228 +               vxi ? atomic_read(&vxi->vx_tasks) : 0,
11229 +               task, _file, _line);
11230 +       __vxh_release_vx_info(vxi, task, _here);
11231 +
11232 +       might_sleep();
11233 +
11234 +       if (atomic_dec_and_test(&vxi->vx_tasks))
11235 +               unhash_vx_info(vxi);
11236 +}
11237 +
11238 +
11239 +#define task_get_vx_info(p) \
11240 +       __task_get_vx_info(p, __FILE__, __LINE__, __HERE__)
11241 +
11242 +static inline struct vx_info *__task_get_vx_info(struct task_struct *p,
11243 +       const char *_file, int _line, void *_here)
11244 +{
11245 +       struct vx_info *vxi;
11246 +
11247 +       task_lock(p);
11248 +       vxlprintk(VXD_CBIT(xid, 5), "task_get_vx_info(%p)",
11249 +               p, _file, _line);
11250 +       vxi = __get_vx_info(p->vx_info, _file, _line, _here);
11251 +       task_unlock(p);
11252 +       return vxi;
11253 +}
11254 +
11255 +
11256 +static inline void __wakeup_vx_info(struct vx_info *vxi)
11257 +{
11258 +       if (waitqueue_active(&vxi->vx_wait))
11259 +               wake_up_interruptible(&vxi->vx_wait);
11260 +}
11261 +
11262 +
11263 +#define enter_vx_info(v, s) __enter_vx_info(v, s, __FILE__, __LINE__)
11264 +
11265 +static inline void __enter_vx_info(struct vx_info *vxi,
11266 +       struct vx_info_save *vxis, const char *_file, int _line)
11267 +{
11268 +       vxlprintk(VXD_CBIT(xid, 5), "enter_vx_info(%p[#%d],%p) %p[#%d,%p]",
11269 +               vxi, vxi ? vxi->vx_id : 0, vxis, current,
11270 +               current->xid, current->vx_info, _file, _line);
11271 +       vxis->vxi = xchg(&current->vx_info, vxi);
11272 +       vxis->xid = current->xid;
11273 +       current->xid = vxi ? vxi->vx_id : 0;
11274 +}
11275 +
11276 +#define leave_vx_info(s) __leave_vx_info(s, __FILE__, __LINE__)
11277 +
11278 +static inline void __leave_vx_info(struct vx_info_save *vxis,
11279 +       const char *_file, int _line)
11280 +{
11281 +       vxlprintk(VXD_CBIT(xid, 5), "leave_vx_info(%p[#%d,%p]) %p[#%d,%p]",
11282 +               vxis, vxis->xid, vxis->vxi, current,
11283 +               current->xid, current->vx_info, _file, _line);
11284 +       (void)xchg(&current->vx_info, vxis->vxi);
11285 +       current->xid = vxis->xid;
11286 +}
11287 +
11288 +
11289 +static inline void __enter_vx_admin(struct vx_info_save *vxis)
11290 +{
11291 +       vxis->vxi = xchg(&current->vx_info, NULL);
11292 +       vxis->xid = xchg(&current->xid, (xid_t)0);
11293 +}
11294 +
11295 +static inline void __leave_vx_admin(struct vx_info_save *vxis)
11296 +{
11297 +       (void)xchg(&current->xid, vxis->xid);
11298 +       (void)xchg(&current->vx_info, vxis->vxi);
11299 +}
11300 +
11301 +#define task_is_init(p) \
11302 +       __task_is_init(p, __FILE__, __LINE__, __HERE__)
11303 +
11304 +static inline int __task_is_init(struct task_struct *p,
11305 +       const char *_file, int _line, void *_here)
11306 +{
11307 +       int is_init = is_global_init(p);
11308 +
11309 +       task_lock(p);
11310 +       if (p->vx_info)
11311 +               is_init = p->vx_info->vx_initpid == p->pid;
11312 +       task_unlock(p);
11313 +       return is_init;
11314 +}
11315 +
11316 +extern void exit_vx_info(struct task_struct *, int);
11317 +extern void exit_vx_info_early(struct task_struct *, int);
11318 +
11319 +
11320 +#else
11321 +#warning duplicate inclusion
11322 +#endif
11323 diff -NurpP --minimal linux-3.0/include/linux/vs_cowbl.h linux-3.0-vs2.3.1-pre7/include/linux/vs_cowbl.h
11324 --- linux-3.0/include/linux/vs_cowbl.h  1970-01-01 01:00:00.000000000 +0100
11325 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_cowbl.h     2011-06-10 22:11:24.000000000 +0200
11326 @@ -0,0 +1,48 @@
11327 +#ifndef _VS_COWBL_H
11328 +#define _VS_COWBL_H
11329 +
11330 +#include <linux/fs.h>
11331 +#include <linux/dcache.h>
11332 +#include <linux/namei.h>
11333 +#include <linux/slab.h>
11334 +
11335 +extern struct dentry *cow_break_link(const char *pathname);
11336 +
11337 +static inline int cow_check_and_break(struct path *path)
11338 +{
11339 +       struct inode *inode = path->dentry->d_inode;
11340 +       int error = 0;
11341 +
11342 +       /* do we need this check? */
11343 +       if (IS_RDONLY(inode))
11344 +               return -EROFS;
11345 +
11346 +       if (IS_COW(inode)) {
11347 +               if (IS_COW_LINK(inode)) {
11348 +                       struct dentry *new_dentry, *old_dentry = path->dentry;
11349 +                       char *pp, *buf;
11350 +
11351 +                       buf = kmalloc(PATH_MAX, GFP_KERNEL);
11352 +                       if (!buf) {
11353 +                               return -ENOMEM;
11354 +                       }
11355 +                       pp = d_path(path, buf, PATH_MAX);
11356 +                       new_dentry = cow_break_link(pp);
11357 +                       kfree(buf);
11358 +                       if (!IS_ERR(new_dentry)) {
11359 +                               path->dentry = new_dentry;
11360 +                               dput(old_dentry);
11361 +                       } else
11362 +                               error = PTR_ERR(new_dentry);
11363 +               } else {
11364 +                       inode->i_flags &= ~(S_IXUNLINK | S_IMMUTABLE);
11365 +                       inode->i_ctime = CURRENT_TIME;
11366 +                       mark_inode_dirty(inode);
11367 +               }
11368 +       }
11369 +       return error;
11370 +}
11371 +
11372 +#else
11373 +#warning duplicate inclusion
11374 +#endif
11375 diff -NurpP --minimal linux-3.0/include/linux/vs_cvirt.h linux-3.0-vs2.3.1-pre7/include/linux/vs_cvirt.h
11376 --- linux-3.0/include/linux/vs_cvirt.h  1970-01-01 01:00:00.000000000 +0100
11377 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_cvirt.h     2011-06-10 22:11:24.000000000 +0200
11378 @@ -0,0 +1,50 @@
11379 +#ifndef _VS_CVIRT_H
11380 +#define _VS_CVIRT_H
11381 +
11382 +#include "vserver/cvirt.h"
11383 +#include "vserver/context.h"
11384 +#include "vserver/base.h"
11385 +#include "vserver/check.h"
11386 +#include "vserver/debug.h"
11387 +
11388 +
11389 +static inline void vx_activate_task(struct task_struct *p)
11390 +{
11391 +       struct vx_info *vxi;
11392 +
11393 +       if ((vxi = p->vx_info)) {
11394 +               vx_update_load(vxi);
11395 +               atomic_inc(&vxi->cvirt.nr_running);
11396 +       }
11397 +}
11398 +
11399 +static inline void vx_deactivate_task(struct task_struct *p)
11400 +{
11401 +       struct vx_info *vxi;
11402 +
11403 +       if ((vxi = p->vx_info)) {
11404 +               vx_update_load(vxi);
11405 +               atomic_dec(&vxi->cvirt.nr_running);
11406 +       }
11407 +}
11408 +
11409 +static inline void vx_uninterruptible_inc(struct task_struct *p)
11410 +{
11411 +       struct vx_info *vxi;
11412 +
11413 +       if ((vxi = p->vx_info))
11414 +               atomic_inc(&vxi->cvirt.nr_uninterruptible);
11415 +}
11416 +
11417 +static inline void vx_uninterruptible_dec(struct task_struct *p)
11418 +{
11419 +       struct vx_info *vxi;
11420 +
11421 +       if ((vxi = p->vx_info))
11422 +               atomic_dec(&vxi->cvirt.nr_uninterruptible);
11423 +}
11424 +
11425 +
11426 +#else
11427 +#warning duplicate inclusion
11428 +#endif
11429 diff -NurpP --minimal linux-3.0/include/linux/vs_device.h linux-3.0-vs2.3.1-pre7/include/linux/vs_device.h
11430 --- linux-3.0/include/linux/vs_device.h 1970-01-01 01:00:00.000000000 +0100
11431 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_device.h    2011-06-10 22:11:24.000000000 +0200
11432 @@ -0,0 +1,45 @@
11433 +#ifndef _VS_DEVICE_H
11434 +#define _VS_DEVICE_H
11435 +
11436 +#include "vserver/base.h"
11437 +#include "vserver/device.h"
11438 +#include "vserver/debug.h"
11439 +
11440 +
11441 +#ifdef CONFIG_VSERVER_DEVICE
11442 +
11443 +int vs_map_device(struct vx_info *, dev_t, dev_t *, umode_t);
11444 +
11445 +#define vs_device_perm(v, d, m, p) \
11446 +       ((vs_map_device(current_vx_info(), d, NULL, m) & (p)) == (p))
11447 +
11448 +#else
11449 +
11450 +static inline
11451 +int vs_map_device(struct vx_info *vxi,
11452 +       dev_t device, dev_t *target, umode_t mode)
11453 +{
11454 +       if (target)
11455 +               *target = device;
11456 +       return ~0;
11457 +}
11458 +
11459 +#define vs_device_perm(v, d, m, p) ((p) == (p))
11460 +
11461 +#endif
11462 +
11463 +
11464 +#define vs_map_chrdev(d, t, p) \
11465 +       ((vs_map_device(current_vx_info(), d, t, S_IFCHR) & (p)) == (p))
11466 +#define vs_map_blkdev(d, t, p) \
11467 +       ((vs_map_device(current_vx_info(), d, t, S_IFBLK) & (p)) == (p))
11468 +
11469 +#define vs_chrdev_perm(d, p) \
11470 +       vs_device_perm(current_vx_info(), d, S_IFCHR, p)
11471 +#define vs_blkdev_perm(d, p) \
11472 +       vs_device_perm(current_vx_info(), d, S_IFBLK, p)
11473 +
11474 +
11475 +#else
11476 +#warning duplicate inclusion
11477 +#endif
11478 diff -NurpP --minimal linux-3.0/include/linux/vs_dlimit.h linux-3.0-vs2.3.1-pre7/include/linux/vs_dlimit.h
11479 --- linux-3.0/include/linux/vs_dlimit.h 1970-01-01 01:00:00.000000000 +0100
11480 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_dlimit.h    2011-06-10 22:11:24.000000000 +0200
11481 @@ -0,0 +1,215 @@
11482 +#ifndef _VS_DLIMIT_H
11483 +#define _VS_DLIMIT_H
11484 +
11485 +#include <linux/fs.h>
11486 +
11487 +#include "vserver/dlimit.h"
11488 +#include "vserver/base.h"
11489 +#include "vserver/debug.h"
11490 +
11491 +
11492 +#define get_dl_info(i) __get_dl_info(i, __FILE__, __LINE__)
11493 +
11494 +static inline struct dl_info *__get_dl_info(struct dl_info *dli,
11495 +       const char *_file, int _line)
11496 +{
11497 +       if (!dli)
11498 +               return NULL;
11499 +       vxlprintk(VXD_CBIT(dlim, 4), "get_dl_info(%p[#%d.%d])",
11500 +               dli, dli ? dli->dl_tag : 0,
11501 +               dli ? atomic_read(&dli->dl_usecnt) : 0,
11502 +               _file, _line);
11503 +       atomic_inc(&dli->dl_usecnt);
11504 +       return dli;
11505 +}
11506 +
11507 +
11508 +#define free_dl_info(i) \
11509 +       call_rcu(&(i)->dl_rcu, rcu_free_dl_info)
11510 +
11511 +#define put_dl_info(i) __put_dl_info(i, __FILE__, __LINE__)
11512 +
11513 +static inline void __put_dl_info(struct dl_info *dli,
11514 +       const char *_file, int _line)
11515 +{
11516 +       if (!dli)
11517 +               return;
11518 +       vxlprintk(VXD_CBIT(dlim, 4), "put_dl_info(%p[#%d.%d])",
11519 +               dli, dli ? dli->dl_tag : 0,
11520 +               dli ? atomic_read(&dli->dl_usecnt) : 0,
11521 +               _file, _line);
11522 +       if (atomic_dec_and_test(&dli->dl_usecnt))
11523 +               free_dl_info(dli);
11524 +}
11525 +
11526 +
11527 +#define __dlimit_char(d)       ((d) ? '*' : ' ')
11528 +
11529 +static inline int __dl_alloc_space(struct super_block *sb,
11530 +       tag_t tag, dlsize_t nr, const char *file, int line)
11531 +{
11532 +       struct dl_info *dli = NULL;
11533 +       int ret = 0;
11534 +
11535 +       if (nr == 0)
11536 +               goto out;
11537 +       dli = locate_dl_info(sb, tag);
11538 +       if (!dli)
11539 +               goto out;
11540 +
11541 +       spin_lock(&dli->dl_lock);
11542 +       ret = (dli->dl_space_used + nr > dli->dl_space_total);
11543 +       if (!ret)
11544 +               dli->dl_space_used += nr;
11545 +       spin_unlock(&dli->dl_lock);
11546 +       put_dl_info(dli);
11547 +out:
11548 +       vxlprintk(VXD_CBIT(dlim, 1),
11549 +               "ALLOC (%p,#%d)%c %lld bytes (%d)",
11550 +               sb, tag, __dlimit_char(dli), (long long)nr,
11551 +               ret, file, line);
11552 +       return ret ? -ENOSPC : 0;
11553 +}
11554 +
11555 +static inline void __dl_free_space(struct super_block *sb,
11556 +       tag_t tag, dlsize_t nr, const char *_file, int _line)
11557 +{
11558 +       struct dl_info *dli = NULL;
11559 +
11560 +       if (nr == 0)
11561 +               goto out;
11562 +       dli = locate_dl_info(sb, tag);
11563 +       if (!dli)
11564 +               goto out;
11565 +
11566 +       spin_lock(&dli->dl_lock);
11567 +       if (dli->dl_space_used > nr)
11568 +               dli->dl_space_used -= nr;
11569 +       else
11570 +               dli->dl_space_used = 0;
11571 +       spin_unlock(&dli->dl_lock);
11572 +       put_dl_info(dli);
11573 +out:
11574 +       vxlprintk(VXD_CBIT(dlim, 1),
11575 +               "FREE  (%p,#%d)%c %lld bytes",
11576 +               sb, tag, __dlimit_char(dli), (long long)nr,
11577 +               _file, _line);
11578 +}
11579 +
11580 +static inline int __dl_alloc_inode(struct super_block *sb,
11581 +       tag_t tag, const char *_file, int _line)
11582 +{
11583 +       struct dl_info *dli;
11584 +       int ret = 0;
11585 +
11586 +       dli = locate_dl_info(sb, tag);
11587 +       if (!dli)
11588 +               goto out;
11589 +
11590 +       spin_lock(&dli->dl_lock);
11591 +       dli->dl_inodes_used++;
11592 +       ret = (dli->dl_inodes_used > dli->dl_inodes_total);
11593 +       spin_unlock(&dli->dl_lock);
11594 +       put_dl_info(dli);
11595 +out:
11596 +       vxlprintk(VXD_CBIT(dlim, 0),
11597 +               "ALLOC (%p,#%d)%c inode (%d)",
11598 +               sb, tag, __dlimit_char(dli), ret, _file, _line);
11599 +       return ret ? -ENOSPC : 0;
11600 +}
11601 +
11602 +static inline void __dl_free_inode(struct super_block *sb,
11603 +       tag_t tag, const char *_file, int _line)
11604 +{
11605 +       struct dl_info *dli;
11606 +
11607 +       dli = locate_dl_info(sb, tag);
11608 +       if (!dli)
11609 +               goto out;
11610 +
11611 +       spin_lock(&dli->dl_lock);
11612 +       if (dli->dl_inodes_used > 1)
11613 +               dli->dl_inodes_used--;
11614 +       else
11615 +               dli->dl_inodes_used = 0;
11616 +       spin_unlock(&dli->dl_lock);
11617 +       put_dl_info(dli);
11618 +out:
11619 +       vxlprintk(VXD_CBIT(dlim, 0),
11620 +               "FREE  (%p,#%d)%c inode",
11621 +               sb, tag, __dlimit_char(dli), _file, _line);
11622 +}
11623 +
11624 +static inline void __dl_adjust_block(struct super_block *sb, tag_t tag,
11625 +       unsigned long long *free_blocks, unsigned long long *root_blocks,
11626 +       const char *_file, int _line)
11627 +{
11628 +       struct dl_info *dli;
11629 +       uint64_t broot, bfree;
11630 +
11631 +       dli = locate_dl_info(sb, tag);
11632 +       if (!dli)
11633 +               return;
11634 +
11635 +       spin_lock(&dli->dl_lock);
11636 +       broot = (dli->dl_space_total -
11637 +               (dli->dl_space_total >> 10) * dli->dl_nrlmult)
11638 +               >> sb->s_blocksize_bits;
11639 +       bfree = (dli->dl_space_total - dli->dl_space_used)
11640 +                       >> sb->s_blocksize_bits;
11641 +       spin_unlock(&dli->dl_lock);
11642 +
11643 +       vxlprintk(VXD_CBIT(dlim, 2),
11644 +               "ADJUST: %lld,%lld on %lld,%lld [mult=%d]",
11645 +               (long long)bfree, (long long)broot,
11646 +               *free_blocks, *root_blocks, dli->dl_nrlmult,
11647 +               _file, _line);
11648 +       if (free_blocks) {
11649 +               if (*free_blocks > bfree)
11650 +                       *free_blocks = bfree;
11651 +       }
11652 +       if (root_blocks) {
11653 +               if (*root_blocks > broot)
11654 +                       *root_blocks = broot;
11655 +       }
11656 +       put_dl_info(dli);
11657 +}
11658 +
11659 +#define dl_prealloc_space(in, bytes) \
11660 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11661 +               __FILE__, __LINE__ )
11662 +
11663 +#define dl_alloc_space(in, bytes) \
11664 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11665 +               __FILE__, __LINE__ )
11666 +
11667 +#define dl_reserve_space(in, bytes) \
11668 +       __dl_alloc_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11669 +               __FILE__, __LINE__ )
11670 +
11671 +#define dl_claim_space(in, bytes) (0)
11672 +
11673 +#define dl_release_space(in, bytes) \
11674 +       __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11675 +               __FILE__, __LINE__ )
11676 +
11677 +#define dl_free_space(in, bytes) \
11678 +       __dl_free_space((in)->i_sb, (in)->i_tag, (dlsize_t)(bytes), \
11679 +               __FILE__, __LINE__ )
11680 +
11681 +
11682 +
11683 +#define dl_alloc_inode(in) \
11684 +       __dl_alloc_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
11685 +
11686 +#define dl_free_inode(in) \
11687 +       __dl_free_inode((in)->i_sb, (in)->i_tag, __FILE__, __LINE__ )
11688 +
11689 +
11690 +#define dl_adjust_block(sb, tag, fb, rb) \
11691 +       __dl_adjust_block(sb, tag, fb, rb, __FILE__, __LINE__ )
11692 +
11693 +
11694 +#else
11695 +#warning duplicate inclusion
11696 +#endif
11697 diff -NurpP --minimal linux-3.0/include/linux/vs_inet.h linux-3.0-vs2.3.1-pre7/include/linux/vs_inet.h
11698 --- linux-3.0/include/linux/vs_inet.h   1970-01-01 01:00:00.000000000 +0100
11699 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_inet.h      2011-06-10 22:11:24.000000000 +0200
11700 @@ -0,0 +1,351 @@
11701 +#ifndef _VS_INET_H
11702 +#define _VS_INET_H
11703 +
11704 +#include "vserver/base.h"
11705 +#include "vserver/network.h"
11706 +#include "vserver/debug.h"
11707 +
11708 +#define IPI_LOOPBACK   htonl(INADDR_LOOPBACK)
11709 +
11710 +#define NXAV4(a)       NIPQUAD((a)->ip[0]), NIPQUAD((a)->ip[1]), \
11711 +                       NIPQUAD((a)->mask), (a)->type
11712 +#define NXAV4_FMT      "[" NIPQUAD_FMT "-" NIPQUAD_FMT "/" NIPQUAD_FMT ":%04x]"
11713 +
11714 +#define NIPQUAD(addr) \
11715 +       ((unsigned char *)&addr)[0], \
11716 +       ((unsigned char *)&addr)[1], \
11717 +       ((unsigned char *)&addr)[2], \
11718 +       ((unsigned char *)&addr)[3]
11719 +
11720 +#define NIPQUAD_FMT "%u.%u.%u.%u"
11721 +
11722 +
11723 +static inline
11724 +int v4_addr_match(struct nx_addr_v4 *nxa, __be32 addr, uint16_t tmask)
11725 +{
11726 +       __be32 ip = nxa->ip[0].s_addr;
11727 +       __be32 mask = nxa->mask.s_addr;
11728 +       __be32 bcast = ip | ~mask;
11729 +       int ret = 0;
11730 +
11731 +       switch (nxa->type & tmask) {
11732 +       case NXA_TYPE_MASK:
11733 +               ret = (ip == (addr & mask));
11734 +               break;
11735 +       case NXA_TYPE_ADDR:
11736 +               ret = 3;
11737 +               if (addr == ip)
11738 +                       break;
11739 +               /* fall through to broadcast */
11740 +       case NXA_MOD_BCAST:
11741 +               ret = ((tmask & NXA_MOD_BCAST) && (addr == bcast));
11742 +               break;
11743 +       case NXA_TYPE_RANGE:
11744 +               ret = ((nxa->ip[0].s_addr <= addr) &&
11745 +                       (nxa->ip[1].s_addr > addr));
11746 +               break;
11747 +       case NXA_TYPE_ANY:
11748 +               ret = 2;
11749 +               break;
11750 +       }
11751 +
11752 +       vxdprintk(VXD_CBIT(net, 0),
11753 +               "v4_addr_match(%p" NXAV4_FMT "," NIPQUAD_FMT ",%04x) = %d",
11754 +               nxa, NXAV4(nxa), NIPQUAD(addr), tmask, ret);
11755 +       return ret;
11756 +}
11757 +
11758 +static inline
11759 +int v4_addr_in_nx_info(struct nx_info *nxi, __be32 addr, uint16_t tmask)
11760 +{
11761 +       struct nx_addr_v4 *nxa;
11762 +       int ret = 1;
11763 +
11764 +       if (!nxi)
11765 +               goto out;
11766 +
11767 +       ret = 2;
11768 +       /* allow 127.0.0.1 when remapping lback */
11769 +       if ((tmask & NXA_LOOPBACK) &&
11770 +               (addr == IPI_LOOPBACK) &&
11771 +               nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
11772 +               goto out;
11773 +       ret = 3;
11774 +       /* check for lback address */
11775 +       if ((tmask & NXA_MOD_LBACK) &&
11776 +               (nxi->v4_lback.s_addr == addr))
11777 +               goto out;
11778 +       ret = 4;
11779 +       /* check for broadcast address */
11780 +       if ((tmask & NXA_MOD_BCAST) &&
11781 +               (nxi->v4_bcast.s_addr == addr))
11782 +               goto out;
11783 +       ret = 5;
11784 +       /* check for v4 addresses */
11785 +       for (nxa = &nxi->v4; nxa; nxa = nxa->next)
11786 +               if (v4_addr_match(nxa, addr, tmask))
11787 +                       goto out;
11788 +       ret = 0;
11789 +out:
11790 +       vxdprintk(VXD_CBIT(net, 0),
11791 +               "v4_addr_in_nx_info(%p[#%u]," NIPQUAD_FMT ",%04x) = %d",
11792 +               nxi, nxi ? nxi->nx_id : 0, NIPQUAD(addr), tmask, ret);
11793 +       return ret;
11794 +}
11795 +
11796 +static inline
11797 +int v4_nx_addr_match(struct nx_addr_v4 *nxa, struct nx_addr_v4 *addr, uint16_t mask)
11798 +{
11799 +       /* FIXME: needs full range checks */
11800 +       return v4_addr_match(nxa, addr->ip[0].s_addr, mask);
11801 +}
11802 +
11803 +static inline
11804 +int v4_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v4 *nxa, uint16_t mask)
11805 +{
11806 +       struct nx_addr_v4 *ptr;
11807 +
11808 +       for (ptr = &nxi->v4; ptr; ptr = ptr->next)
11809 +               if (v4_nx_addr_match(ptr, nxa, mask))
11810 +                       return 1;
11811 +       return 0;
11812 +}
11813 +
11814 +#include <net/inet_sock.h>
11815 +
11816 +/*
11817 + *     Check if a given address matches for a socket
11818 + *
11819 + *     nxi:            the socket's nx_info if any
11820 + *     addr:           to be verified address
11821 + */
11822 +static inline
11823 +int v4_sock_addr_match (
11824 +       struct nx_info *nxi,
11825 +       struct inet_sock *inet,
11826 +       __be32 addr)
11827 +{
11828 +       __be32 saddr = inet->inet_rcv_saddr;
11829 +       __be32 bcast = nxi ? nxi->v4_bcast.s_addr : INADDR_BROADCAST;
11830 +
11831 +       if (addr && (saddr == addr || bcast == addr))
11832 +               return 1;
11833 +       if (!saddr)
11834 +               return v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND);
11835 +       return 0;
11836 +}
11837 +
11838 +
11839 +/* inet related checks and helpers */
11840 +
11841 +
11842 +struct in_ifaddr;
11843 +struct net_device;
11844 +struct sock;
11845 +
11846 +#ifdef CONFIG_INET
11847 +
11848 +#include <linux/netdevice.h>
11849 +#include <linux/inetdevice.h>
11850 +#include <net/inet_sock.h>
11851 +#include <net/inet_timewait_sock.h>
11852 +
11853 +
11854 +int dev_in_nx_info(struct net_device *, struct nx_info *);
11855 +int v4_dev_in_nx_info(struct net_device *, struct nx_info *);
11856 +int nx_v4_addr_conflict(struct nx_info *, struct nx_info *);
11857 +
11858 +
11859 +/*
11860 + *     check if address is covered by socket
11861 + *
11862 + *     sk:     the socket to check against
11863 + *     addr:   the address in question (must be != 0)
11864 + */
11865 +
11866 +static inline
11867 +int __v4_addr_match_socket(const struct sock *sk, struct nx_addr_v4 *nxa)
11868 +{
11869 +       struct nx_info *nxi = sk->sk_nx_info;
11870 +       __be32 saddr = sk_rcv_saddr(sk);
11871 +
11872 +       vxdprintk(VXD_CBIT(net, 5),
11873 +               "__v4_addr_in_socket(%p," NXAV4_FMT ") %p:" NIPQUAD_FMT " %p;%lx",
11874 +               sk, NXAV4(nxa), nxi, NIPQUAD(saddr), sk->sk_socket,
11875 +               (sk->sk_socket?sk->sk_socket->flags:0));
11876 +
11877 +       if (saddr) {            /* direct address match */
11878 +               return v4_addr_match(nxa, saddr, -1);
11879 +       } else if (nxi) {       /* match against nx_info */
11880 +               return v4_nx_addr_in_nx_info(nxi, nxa, -1);
11881 +       } else {                /* unrestricted any socket */
11882 +               return 1;
11883 +       }
11884 +}
11885 +
11886 +
11887 +
11888 +static inline
11889 +int nx_dev_visible(struct nx_info *nxi, struct net_device *dev)
11890 +{
11891 +       vxdprintk(VXD_CBIT(net, 1),
11892 +               "nx_dev_visible(%p[#%u],%p " VS_Q("%s") ") %d",
11893 +               nxi, nxi ? nxi->nx_id : 0, dev, dev->name,
11894 +               nxi ? dev_in_nx_info(dev, nxi) : 0);
11895 +
11896 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
11897 +               return 1;
11898 +       if (dev_in_nx_info(dev, nxi))
11899 +               return 1;
11900 +       return 0;
11901 +}
11902 +
11903 +
11904 +static inline
11905 +int v4_ifa_in_nx_info(struct in_ifaddr *ifa, struct nx_info *nxi)
11906 +{
11907 +       if (!nxi)
11908 +               return 1;
11909 +       if (!ifa)
11910 +               return 0;
11911 +       return v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW);
11912 +}
11913 +
11914 +static inline
11915 +int nx_v4_ifa_visible(struct nx_info *nxi, struct in_ifaddr *ifa)
11916 +{
11917 +       vxdprintk(VXD_CBIT(net, 1), "nx_v4_ifa_visible(%p[#%u],%p) %d",
11918 +               nxi, nxi ? nxi->nx_id : 0, ifa,
11919 +               nxi ? v4_ifa_in_nx_info(ifa, nxi) : 0);
11920 +
11921 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
11922 +               return 1;
11923 +       if (v4_ifa_in_nx_info(ifa, nxi))
11924 +               return 1;
11925 +       return 0;
11926 +}
11927 +
11928 +
11929 +struct nx_v4_sock_addr {
11930 +       __be32 saddr;   /* Address used for validation */
11931 +       __be32 baddr;   /* Address used for socket bind */
11932 +};
11933 +
11934 +static inline
11935 +int v4_map_sock_addr(struct inet_sock *inet, struct sockaddr_in *addr,
11936 +       struct nx_v4_sock_addr *nsa)
11937 +{
11938 +       struct sock *sk = &inet->sk;
11939 +       struct nx_info *nxi = sk->sk_nx_info;
11940 +       __be32 saddr = addr->sin_addr.s_addr;
11941 +       __be32 baddr = saddr;
11942 +
11943 +       vxdprintk(VXD_CBIT(net, 3),
11944 +               "inet_bind(%p)* %p,%p;%lx " NIPQUAD_FMT,
11945 +               sk, sk->sk_nx_info, sk->sk_socket,
11946 +               (sk->sk_socket ? sk->sk_socket->flags : 0),
11947 +               NIPQUAD(saddr));
11948 +
11949 +       if (nxi) {
11950 +               if (saddr == INADDR_ANY) {
11951 +                       if (nx_info_flags(nxi, NXF_SINGLE_IP, 0))
11952 +                               baddr = nxi->v4.ip[0].s_addr;
11953 +               } else if (saddr == IPI_LOOPBACK) {
11954 +                       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
11955 +                               baddr = nxi->v4_lback.s_addr;
11956 +               } else {        /* normal address bind */
11957 +                       if (!v4_addr_in_nx_info(nxi, saddr, NXA_MASK_BIND))
11958 +                               return -EADDRNOTAVAIL;
11959 +               }
11960 +       }
11961 +
11962 +       vxdprintk(VXD_CBIT(net, 3),
11963 +               "inet_bind(%p) " NIPQUAD_FMT ", " NIPQUAD_FMT,
11964 +               sk, NIPQUAD(saddr), NIPQUAD(baddr));
11965 +
11966 +       nsa->saddr = saddr;
11967 +       nsa->baddr = baddr;
11968 +       return 0;
11969 +}
11970 +
11971 +static inline
11972 +void v4_set_sock_addr(struct inet_sock *inet, struct nx_v4_sock_addr *nsa)
11973 +{
11974 +       inet->inet_saddr = nsa->baddr;
11975 +       inet->inet_rcv_saddr = nsa->baddr;
11976 +}
11977 +
11978 +
11979 +/*
11980 + *      helper to simplify inet_lookup_listener
11981 + *
11982 + *      nxi:   the socket's nx_info if any
11983 + *      addr:  to be verified address
11984 + *      saddr: socket address
11985 + */
11986 +static inline int v4_inet_addr_match (
11987 +       struct nx_info *nxi,
11988 +       __be32 addr,
11989 +       __be32 saddr)
11990 +{
11991 +       if (addr && (saddr == addr))
11992 +               return 1;
11993 +       if (!saddr)
11994 +               return nxi ? v4_addr_in_nx_info(nxi, addr, NXA_MASK_BIND) : 1;
11995 +       return 0;
11996 +}
11997 +
11998 +static inline __be32 nx_map_sock_lback(struct nx_info *nxi, __be32 addr)
11999 +{
12000 +       if (nx_info_flags(nxi, NXF_HIDE_LBACK, 0) &&
12001 +               (addr == nxi->v4_lback.s_addr))
12002 +               return IPI_LOOPBACK;
12003 +       return addr;
12004 +}
12005 +
12006 +static inline
12007 +int nx_info_has_v4(struct nx_info *nxi)
12008 +{
12009 +       if (!nxi)
12010 +               return 1;
12011 +       if (NX_IPV4(nxi))
12012 +               return 1;
12013 +       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0))
12014 +               return 1;
12015 +       return 0;
12016 +}
12017 +
12018 +#else /* CONFIG_INET */
12019 +
12020 +static inline
12021 +int nx_dev_visible(struct nx_info *n, struct net_device *d)
12022 +{
12023 +       return 1;
12024 +}
12025 +
12026 +static inline
12027 +int nx_v4_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
12028 +{
12029 +       return 1;
12030 +}
12031 +
12032 +static inline
12033 +int v4_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
12034 +{
12035 +       return 1;
12036 +}
12037 +
12038 +static inline
12039 +int nx_info_has_v4(struct nx_info *nxi)
12040 +{
12041 +       return 0;
12042 +}
12043 +
12044 +#endif /* CONFIG_INET */
12045 +
12046 +#define current_nx_info_has_v4() \
12047 +       nx_info_has_v4(current_nx_info())
12048 +
12049 +#else
12050 +// #warning duplicate inclusion
12051 +#endif
12052 diff -NurpP --minimal linux-3.0/include/linux/vs_inet6.h linux-3.0-vs2.3.1-pre7/include/linux/vs_inet6.h
12053 --- linux-3.0/include/linux/vs_inet6.h  1970-01-01 01:00:00.000000000 +0100
12054 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_inet6.h     2011-06-10 22:11:24.000000000 +0200
12055 @@ -0,0 +1,246 @@
12056 +#ifndef _VS_INET6_H
12057 +#define _VS_INET6_H
12058 +
12059 +#include "vserver/base.h"
12060 +#include "vserver/network.h"
12061 +#include "vserver/debug.h"
12062 +
12063 +#include <net/ipv6.h>
12064 +
12065 +#define NXAV6(a)       &(a)->ip, &(a)->mask, (a)->prefix, (a)->type
12066 +#define NXAV6_FMT      "[%pI6/%pI6/%d:%04x]"
12067 +
12068 +
12069 +#ifdef CONFIG_IPV6
12070 +
12071 +static inline
12072 +int v6_addr_match(struct nx_addr_v6 *nxa,
12073 +       const struct in6_addr *addr, uint16_t mask)
12074 +{
12075 +       int ret = 0;
12076 +
12077 +       switch (nxa->type & mask) {
12078 +       case NXA_TYPE_MASK:
12079 +               ret = ipv6_masked_addr_cmp(&nxa->ip, &nxa->mask, addr);
12080 +               break;
12081 +       case NXA_TYPE_ADDR:
12082 +               ret = ipv6_addr_equal(&nxa->ip, addr);
12083 +               break;
12084 +       case NXA_TYPE_ANY:
12085 +               ret = 1;
12086 +               break;
12087 +       }
12088 +       vxdprintk(VXD_CBIT(net, 0),
12089 +               "v6_addr_match(%p" NXAV6_FMT ",%pI6,%04x) = %d",
12090 +               nxa, NXAV6(nxa), addr, mask, ret);
12091 +       return ret;
12092 +}
12093 +
12094 +static inline
12095 +int v6_addr_in_nx_info(struct nx_info *nxi,
12096 +       const struct in6_addr *addr, uint16_t mask)
12097 +{
12098 +       struct nx_addr_v6 *nxa;
12099 +       int ret = 1;
12100 +
12101 +       if (!nxi)
12102 +               goto out;
12103 +       for (nxa = &nxi->v6; nxa; nxa = nxa->next)
12104 +               if (v6_addr_match(nxa, addr, mask))
12105 +                       goto out;
12106 +       ret = 0;
12107 +out:
12108 +       vxdprintk(VXD_CBIT(net, 0),
12109 +               "v6_addr_in_nx_info(%p[#%u],%pI6,%04x) = %d",
12110 +               nxi, nxi ? nxi->nx_id : 0, addr, mask, ret);
12111 +       return ret;
12112 +}
12113 +
12114 +static inline
12115 +int v6_nx_addr_match(struct nx_addr_v6 *nxa, struct nx_addr_v6 *addr, uint16_t mask)
12116 +{
12117 +       /* FIXME: needs full range checks */
12118 +       return v6_addr_match(nxa, &addr->ip, mask);
12119 +}
12120 +
12121 +static inline
12122 +int v6_nx_addr_in_nx_info(struct nx_info *nxi, struct nx_addr_v6 *nxa, uint16_t mask)
12123 +{
12124 +       struct nx_addr_v6 *ptr;
12125 +
12126 +       for (ptr = &nxi->v6; ptr; ptr = ptr->next)
12127 +               if (v6_nx_addr_match(ptr, nxa, mask))
12128 +                       return 1;
12129 +       return 0;
12130 +}
12131 +
12132 +
12133 +/*
12134 + *     Check if a given address matches for a socket
12135 + *
12136 + *     nxi:            the socket's nx_info if any
12137 + *     addr:           to be verified address
12138 + */
12139 +static inline
12140 +int v6_sock_addr_match (
12141 +       struct nx_info *nxi,
12142 +       struct inet_sock *inet,
12143 +       struct in6_addr *addr)
12144 +{
12145 +       struct sock *sk = &inet->sk;
12146 +       struct in6_addr *saddr = inet6_rcv_saddr(sk);
12147 +
12148 +       if (!ipv6_addr_any(addr) &&
12149 +               ipv6_addr_equal(saddr, addr))
12150 +               return 1;
12151 +       if (ipv6_addr_any(saddr))
12152 +               return v6_addr_in_nx_info(nxi, addr, -1);
12153 +       return 0;
12154 +}
12155 +
12156 +/*
12157 + *     check if address is covered by socket
12158 + *
12159 + *     sk:     the socket to check against
12160 + *     addr:   the address in question (must be != 0)
12161 + */
12162 +
12163 +static inline
12164 +int __v6_addr_match_socket(const struct sock *sk, struct nx_addr_v6 *nxa)
12165 +{
12166 +       struct nx_info *nxi = sk->sk_nx_info;
12167 +       struct in6_addr *saddr = inet6_rcv_saddr(sk);
12168 +
12169 +       vxdprintk(VXD_CBIT(net, 5),
12170 +               "__v6_addr_in_socket(%p," NXAV6_FMT ") %p:%pI6 %p;%lx",
12171 +               sk, NXAV6(nxa), nxi, saddr, sk->sk_socket,
12172 +               (sk->sk_socket?sk->sk_socket->flags:0));
12173 +
12174 +       if (!ipv6_addr_any(saddr)) {    /* direct address match */
12175 +               return v6_addr_match(nxa, saddr, -1);
12176 +       } else if (nxi) {               /* match against nx_info */
12177 +               return v6_nx_addr_in_nx_info(nxi, nxa, -1);
12178 +       } else {                        /* unrestricted any socket */
12179 +               return 1;
12180 +       }
12181 +}
12182 +
12183 +
12184 +/* inet related checks and helpers */
12185 +
12186 +
12187 +struct in_ifaddr;
12188 +struct net_device;
12189 +struct sock;
12190 +
12191 +
12192 +#include <linux/netdevice.h>
12193 +#include <linux/inetdevice.h>
12194 +#include <net/inet_timewait_sock.h>
12195 +
12196 +
12197 +int dev_in_nx_info(struct net_device *, struct nx_info *);
12198 +int v6_dev_in_nx_info(struct net_device *, struct nx_info *);
12199 +int nx_v6_addr_conflict(struct nx_info *, struct nx_info *);
12200 +
12201 +
12202 +
12203 +static inline
12204 +int v6_ifa_in_nx_info(struct inet6_ifaddr *ifa, struct nx_info *nxi)
12205 +{
12206 +       if (!nxi)
12207 +               return 1;
12208 +       if (!ifa)
12209 +               return 0;
12210 +       return v6_addr_in_nx_info(nxi, &ifa->addr, -1);
12211 +}
12212 +
12213 +static inline
12214 +int nx_v6_ifa_visible(struct nx_info *nxi, struct inet6_ifaddr *ifa)
12215 +{
12216 +       vxdprintk(VXD_CBIT(net, 1), "nx_v6_ifa_visible(%p[#%u],%p) %d",
12217 +               nxi, nxi ? nxi->nx_id : 0, ifa,
12218 +               nxi ? v6_ifa_in_nx_info(ifa, nxi) : 0);
12219 +
12220 +       if (!nx_info_flags(nxi, NXF_HIDE_NETIF, 0))
12221 +               return 1;
12222 +       if (v6_ifa_in_nx_info(ifa, nxi))
12223 +               return 1;
12224 +       return 0;
12225 +}
12226 +
12227 +
12228 +struct nx_v6_sock_addr {
12229 +       struct in6_addr saddr;  /* Address used for validation */
12230 +       struct in6_addr baddr;  /* Address used for socket bind */
12231 +};
12232 +
12233 +static inline
12234 +int v6_map_sock_addr(struct inet_sock *inet, struct sockaddr_in6 *addr,
12235 +       struct nx_v6_sock_addr *nsa)
12236 +{
12237 +       // struct sock *sk = &inet->sk;
12238 +       // struct nx_info *nxi = sk->sk_nx_info;
12239 +       struct in6_addr saddr = addr->sin6_addr;
12240 +       struct in6_addr baddr = saddr;
12241 +
12242 +       nsa->saddr = saddr;
12243 +       nsa->baddr = baddr;
12244 +       return 0;
12245 +}
12246 +
12247 +static inline
12248 +void v6_set_sock_addr(struct inet_sock *inet, struct nx_v6_sock_addr *nsa)
12249 +{
12250 +       // struct sock *sk = &inet->sk;
12251 +       // struct in6_addr *saddr = inet6_rcv_saddr(sk);
12252 +
12253 +       // *saddr = nsa->baddr;
12254 +       // inet->inet_saddr = nsa->baddr;
12255 +}
12256 +
12257 +static inline
12258 +int nx_info_has_v6(struct nx_info *nxi)
12259 +{
12260 +       if (!nxi)
12261 +               return 1;
12262 +       if (NX_IPV6(nxi))
12263 +               return 1;
12264 +       return 0;
12265 +}
12266 +
12267 +#else /* CONFIG_IPV6 */
12268 +
12269 +static inline
12270 +int nx_v6_dev_visible(struct nx_info *n, struct net_device *d)
12271 +{
12272 +       return 1;
12273 +}
12274 +
12275 +
12276 +static inline
12277 +int nx_v6_addr_conflict(struct nx_info *n, uint32_t a, const struct sock *s)
12278 +{
12279 +       return 1;
12280 +}
12281 +
12282 +static inline
12283 +int v6_ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n)
12284 +{
12285 +       return 1;
12286 +}
12287 +
12288 +static inline
12289 +int nx_info_has_v6(struct nx_info *nxi)
12290 +{
12291 +       return 0;
12292 +}
12293 +
12294 +#endif /* CONFIG_IPV6 */
12295 +
12296 +#define current_nx_info_has_v6() \
12297 +       nx_info_has_v6(current_nx_info())
12298 +
12299 +#else
12300 +#warning duplicate inclusion
12301 +#endif
12302 diff -NurpP --minimal linux-3.0/include/linux/vs_limit.h linux-3.0-vs2.3.1-pre7/include/linux/vs_limit.h
12303 --- linux-3.0/include/linux/vs_limit.h  1970-01-01 01:00:00.000000000 +0100
12304 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_limit.h     2011-06-10 22:11:24.000000000 +0200
12305 @@ -0,0 +1,140 @@
12306 +#ifndef _VS_LIMIT_H
12307 +#define _VS_LIMIT_H
12308 +
12309 +#include "vserver/limit.h"
12310 +#include "vserver/base.h"
12311 +#include "vserver/context.h"
12312 +#include "vserver/debug.h"
12313 +#include "vserver/context.h"
12314 +#include "vserver/limit_int.h"
12315 +
12316 +
12317 +#define vx_acc_cres(v, d, p, r) \
12318 +       __vx_acc_cres(v, r, d, p, __FILE__, __LINE__)
12319 +
12320 +#define vx_acc_cres_cond(x, d, p, r) \
12321 +       __vx_acc_cres(((x) == vx_current_xid()) ? current_vx_info() : 0, \
12322 +       r, d, p, __FILE__, __LINE__)
12323 +
12324 +
12325 +#define vx_add_cres(v, a, p, r) \
12326 +       __vx_add_cres(v, r, a, p, __FILE__, __LINE__)
12327 +#define vx_sub_cres(v, a, p, r)                vx_add_cres(v, -(a), p, r)
12328 +
12329 +#define vx_add_cres_cond(x, a, p, r) \
12330 +       __vx_add_cres(((x) == vx_current_xid()) ? current_vx_info() : 0, \
12331 +       r, a, p, __FILE__, __LINE__)
12332 +#define vx_sub_cres_cond(x, a, p, r)   vx_add_cres_cond(x, -(a), p, r)
12333 +
12334 +
12335 +/* process and file limits */
12336 +
12337 +#define vx_nproc_inc(p) \
12338 +       vx_acc_cres((p)->vx_info, 1, p, RLIMIT_NPROC)
12339 +
12340 +#define vx_nproc_dec(p) \
12341 +       vx_acc_cres((p)->vx_info,-1, p, RLIMIT_NPROC)
12342 +
12343 +#define vx_files_inc(f) \
12344 +       vx_acc_cres_cond((f)->f_xid, 1, f, RLIMIT_NOFILE)
12345 +
12346 +#define vx_files_dec(f) \
12347 +       vx_acc_cres_cond((f)->f_xid,-1, f, RLIMIT_NOFILE)
12348 +
12349 +#define vx_locks_inc(l) \
12350 +       vx_acc_cres_cond((l)->fl_xid, 1, l, RLIMIT_LOCKS)
12351 +
12352 +#define vx_locks_dec(l) \
12353 +       vx_acc_cres_cond((l)->fl_xid,-1, l, RLIMIT_LOCKS)
12354 +
12355 +#define vx_openfd_inc(f) \
12356 +       vx_acc_cres(current_vx_info(), 1, (void *)(long)(f), VLIMIT_OPENFD)
12357 +
12358 +#define vx_openfd_dec(f) \
12359 +       vx_acc_cres(current_vx_info(),-1, (void *)(long)(f), VLIMIT_OPENFD)
12360 +
12361 +
12362 +#define vx_cres_avail(v, n, r) \
12363 +       __vx_cres_avail(v, r, n, __FILE__, __LINE__)
12364 +
12365 +
12366 +#define vx_nproc_avail(n) \
12367 +       vx_cres_avail(current_vx_info(), n, RLIMIT_NPROC)
12368 +
12369 +#define vx_files_avail(n) \
12370 +       vx_cres_avail(current_vx_info(), n, RLIMIT_NOFILE)
12371 +
12372 +#define vx_locks_avail(n) \
12373 +       vx_cres_avail(current_vx_info(), n, RLIMIT_LOCKS)
12374 +
12375 +#define vx_openfd_avail(n) \
12376 +       vx_cres_avail(current_vx_info(), n, VLIMIT_OPENFD)
12377 +
12378 +
12379 +/* dentry limits */
12380 +
12381 +#define vx_dentry_inc(d) do {                                          \
12382 +       if ((d)->d_count == 1)                                          \
12383 +               vx_acc_cres(current_vx_info(), 1, d, VLIMIT_DENTRY);    \
12384 +       } while (0)
12385 +
12386 +#define vx_dentry_dec(d) do {                                          \
12387 +       if ((d)->d_count == 0)                                          \
12388 +               vx_acc_cres(current_vx_info(),-1, d, VLIMIT_DENTRY);    \
12389 +       } while (0)
12390 +
12391 +#define vx_dentry_avail(n) \
12392 +       vx_cres_avail(current_vx_info(), n, VLIMIT_DENTRY)
12393 +
12394 +
12395 +/* socket limits */
12396 +
12397 +#define vx_sock_inc(s) \
12398 +       vx_acc_cres((s)->sk_vx_info, 1, s, VLIMIT_NSOCK)
12399 +
12400 +#define vx_sock_dec(s) \
12401 +       vx_acc_cres((s)->sk_vx_info,-1, s, VLIMIT_NSOCK)
12402 +
12403 +#define vx_sock_avail(n) \
12404 +       vx_cres_avail(current_vx_info(), n, VLIMIT_NSOCK)
12405 +
12406 +
12407 +/* ipc resource limits */
12408 +
12409 +#define vx_ipcmsg_add(v, u, a) \
12410 +       vx_add_cres(v, a, u, RLIMIT_MSGQUEUE)
12411 +
12412 +#define vx_ipcmsg_sub(v, u, a) \
12413 +       vx_sub_cres(v, a, u, RLIMIT_MSGQUEUE)
12414 +
12415 +#define vx_ipcmsg_avail(v, a) \
12416 +       vx_cres_avail(v, a, RLIMIT_MSGQUEUE)
12417 +
12418 +
12419 +#define vx_ipcshm_add(v, k, a) \
12420 +       vx_add_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
12421 +
12422 +#define vx_ipcshm_sub(v, k, a) \
12423 +       vx_sub_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM)
12424 +
12425 +#define vx_ipcshm_avail(v, a) \
12426 +       vx_cres_avail(v, a, VLIMIT_SHMEM)
12427 +
12428 +
12429 +#define vx_semary_inc(a) \
12430 +       vx_acc_cres(current_vx_info(), 1, a, VLIMIT_SEMARY)
12431 +
12432 +#define vx_semary_dec(a) \
12433 +       vx_acc_cres(current_vx_info(), -1, a, VLIMIT_SEMARY)
12434 +
12435 +
12436 +#define vx_nsems_add(a,n) \
12437 +       vx_add_cres(current_vx_info(), n, a, VLIMIT_NSEMS)
12438 +
12439 +#define vx_nsems_sub(a,n) \
12440 +       vx_sub_cres(current_vx_info(), n, a, VLIMIT_NSEMS)
12441 +
12442 +
12443 +#else
12444 +#warning duplicate inclusion
12445 +#endif
12446 diff -NurpP --minimal linux-3.0/include/linux/vs_memory.h linux-3.0-vs2.3.1-pre7/include/linux/vs_memory.h
12447 --- linux-3.0/include/linux/vs_memory.h 1970-01-01 01:00:00.000000000 +0100
12448 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_memory.h    2011-06-10 22:11:24.000000000 +0200
12449 @@ -0,0 +1,58 @@
12450 +#ifndef _VS_MEMORY_H
12451 +#define _VS_MEMORY_H
12452 +
12453 +#include "vserver/limit.h"
12454 +#include "vserver/base.h"
12455 +#include "vserver/context.h"
12456 +#include "vserver/debug.h"
12457 +#include "vserver/context.h"
12458 +#include "vserver/limit_int.h"
12459 +
12460 +enum {
12461 +       VXPT_UNKNOWN = 0,
12462 +       VXPT_ANON,
12463 +       VXPT_NONE,
12464 +       VXPT_FILE,
12465 +       VXPT_SWAP,
12466 +       VXPT_WRITE
12467 +};
12468 +
12469 +#if 0
12470 +#define        vx_page_fault(mm, vma, type, ret)
12471 +#else
12472 +
12473 +static inline
12474 +void __vx_page_fault(struct mm_struct *mm,
12475 +       struct vm_area_struct *vma, int type, int ret)
12476 +{
12477 +       struct vx_info *vxi = mm->mm_vx_info;
12478 +       int what;
12479 +/*
12480 +       static char *page_type[6] =
12481 +               { "UNKNOWN", "ANON", "NONE", "FILE", "SWAP", "WRITE" };
12482 +       static char *page_what[4] =
12483 +               { "FAULT_OOM", "FAULT_SIGBUS", "FAULT_MINOR", "FAULT_MAJOR" };
12484 +*/
12485 +
12486 +       if (!vxi)
12487 +               return;
12488 +
12489 +       what = (ret & 0x3);
12490 +
12491 +/*     printk("[%d] page[%d][%d] %2x %s %s\n", vxi->vx_id,
12492 +               type, what, ret, page_type[type], page_what[what]);
12493 +*/
12494 +       if (ret & VM_FAULT_WRITE)
12495 +               what |= 0x4;
12496 +       atomic_inc(&vxi->cacct.page[type][what]);
12497 +}
12498 +
12499 +#define        vx_page_fault(mm, vma, type, ret)       __vx_page_fault(mm, vma, type, ret)
12500 +#endif
12501 +
12502 +
12503 +extern unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm);
12504 +
12505 +#else
12506 +#warning duplicate inclusion
12507 +#endif
12508 diff -NurpP --minimal linux-3.0/include/linux/vs_network.h linux-3.0-vs2.3.1-pre7/include/linux/vs_network.h
12509 --- linux-3.0/include/linux/vs_network.h        1970-01-01 01:00:00.000000000 +0100
12510 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_network.h   2011-06-10 22:11:24.000000000 +0200
12511 @@ -0,0 +1,169 @@
12512 +#ifndef _NX_VS_NETWORK_H
12513 +#define _NX_VS_NETWORK_H
12514 +
12515 +#include "vserver/context.h"
12516 +#include "vserver/network.h"
12517 +#include "vserver/base.h"
12518 +#include "vserver/check.h"
12519 +#include "vserver/debug.h"
12520 +
12521 +#include <linux/sched.h>
12522 +
12523 +
12524 +#define get_nx_info(i) __get_nx_info(i, __FILE__, __LINE__)
12525 +
12526 +static inline struct nx_info *__get_nx_info(struct nx_info *nxi,
12527 +       const char *_file, int _line)
12528 +{
12529 +       if (!nxi)
12530 +               return NULL;
12531 +
12532 +       vxlprintk(VXD_CBIT(nid, 2), "get_nx_info(%p[#%d.%d])",
12533 +               nxi, nxi ? nxi->nx_id : 0,
12534 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
12535 +               _file, _line);
12536 +
12537 +       atomic_inc(&nxi->nx_usecnt);
12538 +       return nxi;
12539 +}
12540 +
12541 +
12542 +extern void free_nx_info(struct nx_info *);
12543 +
12544 +#define put_nx_info(i) __put_nx_info(i, __FILE__, __LINE__)
12545 +
12546 +static inline void __put_nx_info(struct nx_info *nxi, const char *_file, int _line)
12547 +{
12548 +       if (!nxi)
12549 +               return;
12550 +
12551 +       vxlprintk(VXD_CBIT(nid, 2), "put_nx_info(%p[#%d.%d])",
12552 +               nxi, nxi ? nxi->nx_id : 0,
12553 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
12554 +               _file, _line);
12555 +
12556 +       if (atomic_dec_and_test(&nxi->nx_usecnt))
12557 +               free_nx_info(nxi);
12558 +}
12559 +
12560 +
12561 +#define init_nx_info(p, i) __init_nx_info(p, i, __FILE__, __LINE__)
12562 +
12563 +static inline void __init_nx_info(struct nx_info **nxp, struct nx_info *nxi,
12564 +               const char *_file, int _line)
12565 +{
12566 +       if (nxi) {
12567 +               vxlprintk(VXD_CBIT(nid, 3),
12568 +                       "init_nx_info(%p[#%d.%d])",
12569 +                       nxi, nxi ? nxi->nx_id : 0,
12570 +                       nxi ? atomic_read(&nxi->nx_usecnt) : 0,
12571 +                       _file, _line);
12572 +
12573 +               atomic_inc(&nxi->nx_usecnt);
12574 +       }
12575 +       *nxp = nxi;
12576 +}
12577 +
12578 +
12579 +#define set_nx_info(p, i) __set_nx_info(p, i, __FILE__, __LINE__)
12580 +
12581 +static inline void __set_nx_info(struct nx_info **nxp, struct nx_info *nxi,
12582 +       const char *_file, int _line)
12583 +{
12584 +       struct nx_info *nxo;
12585 +
12586 +       if (!nxi)
12587 +               return;
12588 +
12589 +       vxlprintk(VXD_CBIT(nid, 3), "set_nx_info(%p[#%d.%d])",
12590 +               nxi, nxi ? nxi->nx_id : 0,
12591 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
12592 +               _file, _line);
12593 +
12594 +       atomic_inc(&nxi->nx_usecnt);
12595 +       nxo = xchg(nxp, nxi);
12596 +       BUG_ON(nxo);
12597 +}
12598 +
12599 +#define clr_nx_info(p) __clr_nx_info(p, __FILE__, __LINE__)
12600 +
12601 +static inline void __clr_nx_info(struct nx_info **nxp,
12602 +       const char *_file, int _line)
12603 +{
12604 +       struct nx_info *nxo;
12605 +
12606 +       nxo = xchg(nxp, NULL);
12607 +       if (!nxo)
12608 +               return;
12609 +
12610 +       vxlprintk(VXD_CBIT(nid, 3), "clr_nx_info(%p[#%d.%d])",
12611 +               nxo, nxo ? nxo->nx_id : 0,
12612 +               nxo ? atomic_read(&nxo->nx_usecnt) : 0,
12613 +               _file, _line);
12614 +
12615 +       if (atomic_dec_and_test(&nxo->nx_usecnt))
12616 +               free_nx_info(nxo);
12617 +}
12618 +
12619 +
12620 +#define claim_nx_info(v, p) __claim_nx_info(v, p, __FILE__, __LINE__)
12621 +
12622 +static inline void __claim_nx_info(struct nx_info *nxi,
12623 +       struct task_struct *task, const char *_file, int _line)
12624 +{
12625 +       vxlprintk(VXD_CBIT(nid, 3), "claim_nx_info(%p[#%d.%d.%d]) %p",
12626 +               nxi, nxi ? nxi->nx_id : 0,
12627 +               nxi?atomic_read(&nxi->nx_usecnt):0,
12628 +               nxi?atomic_read(&nxi->nx_tasks):0,
12629 +               task, _file, _line);
12630 +
12631 +       atomic_inc(&nxi->nx_tasks);
12632 +}
12633 +
12634 +
12635 +extern void unhash_nx_info(struct nx_info *);
12636 +
12637 +#define release_nx_info(v, p) __release_nx_info(v, p, __FILE__, __LINE__)
12638 +
12639 +static inline void __release_nx_info(struct nx_info *nxi,
12640 +       struct task_struct *task, const char *_file, int _line)
12641 +{
12642 +       vxlprintk(VXD_CBIT(nid, 3), "release_nx_info(%p[#%d.%d.%d]) %p",
12643 +               nxi, nxi ? nxi->nx_id : 0,
12644 +               nxi ? atomic_read(&nxi->nx_usecnt) : 0,
12645 +               nxi ? atomic_read(&nxi->nx_tasks) : 0,
12646 +               task, _file, _line);
12647 +
12648 +       might_sleep();
12649 +
12650 +       if (atomic_dec_and_test(&nxi->nx_tasks))
12651 +               unhash_nx_info(nxi);
12652 +}
12653 +
12654 +
12655 +#define task_get_nx_info(i)    __task_get_nx_info(i, __FILE__, __LINE__)
12656 +
12657 +static __inline__ struct nx_info *__task_get_nx_info(struct task_struct *p,
12658 +       const char *_file, int _line)
12659 +{
12660 +       struct nx_info *nxi;
12661 +
12662 +       task_lock(p);
12663 +       vxlprintk(VXD_CBIT(nid, 5), "task_get_nx_info(%p)",
12664 +               p, _file, _line);
12665 +       nxi = __get_nx_info(p->nx_info, _file, _line);
12666 +       task_unlock(p);
12667 +       return nxi;
12668 +}
12669 +
12670 +
12671 +static inline void exit_nx_info(struct task_struct *p)
12672 +{
12673 +       if (p->nx_info)
12674 +               release_nx_info(p->nx_info, p);
12675 +}
12676 +
12677 +
12678 +#else
12679 +#warning duplicate inclusion
12680 +#endif
12681 diff -NurpP --minimal linux-3.0/include/linux/vs_pid.h linux-3.0-vs2.3.1-pre7/include/linux/vs_pid.h
12682 --- linux-3.0/include/linux/vs_pid.h    1970-01-01 01:00:00.000000000 +0100
12683 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_pid.h       2011-06-10 22:11:24.000000000 +0200
12684 @@ -0,0 +1,50 @@
12685 +#ifndef _VS_PID_H
12686 +#define _VS_PID_H
12687 +
12688 +#include "vserver/base.h"
12689 +#include "vserver/check.h"
12690 +#include "vserver/context.h"
12691 +#include "vserver/debug.h"
12692 +#include "vserver/pid.h"
12693 +#include <linux/pid_namespace.h>
12694 +
12695 +
12696 +#define VXF_FAKE_INIT  (VXF_INFO_INIT | VXF_STATE_INIT)
12697 +
12698 +static inline
12699 +int vx_proc_task_visible(struct task_struct *task)
12700 +{
12701 +       if ((task->pid == 1) &&
12702 +               !vx_flags(VXF_FAKE_INIT, VXF_FAKE_INIT))
12703 +               /* show a blend through init */
12704 +               goto visible;
12705 +       if (vx_check(vx_task_xid(task), VS_WATCH | VS_IDENT))
12706 +               goto visible;
12707 +       return 0;
12708 +visible:
12709 +       return 1;
12710 +}
12711 +
12712 +#define find_task_by_real_pid(pid) find_task_by_pid_ns(pid, &init_pid_ns)
12713 +
12714 +
12715 +static inline
12716 +struct task_struct *vx_get_proc_task(struct inode *inode, struct pid *pid)
12717 +{
12718 +       struct task_struct *task = get_pid_task(pid, PIDTYPE_PID);
12719 +
12720 +       if (task && !vx_proc_task_visible(task)) {
12721 +               vxdprintk(VXD_CBIT(misc, 6),
12722 +                       "dropping task (get) %p[#%u,%u] for %p[#%u,%u]",
12723 +                       task, task->xid, task->pid,
12724 +                       current, current->xid, current->pid);
12725 +               put_task_struct(task);
12726 +               task = NULL;
12727 +       }
12728 +       return task;
12729 +}
12730 +
12731 +
12732 +#else
12733 +#warning duplicate inclusion
12734 +#endif
12735 diff -NurpP --minimal linux-3.0/include/linux/vs_sched.h linux-3.0-vs2.3.1-pre7/include/linux/vs_sched.h
12736 --- linux-3.0/include/linux/vs_sched.h  1970-01-01 01:00:00.000000000 +0100
12737 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_sched.h     2011-06-10 22:11:24.000000000 +0200
12738 @@ -0,0 +1,40 @@
12739 +#ifndef _VS_SCHED_H
12740 +#define _VS_SCHED_H
12741 +
12742 +#include "vserver/base.h"
12743 +#include "vserver/context.h"
12744 +#include "vserver/sched.h"
12745 +
12746 +
12747 +#define MAX_PRIO_BIAS           20
12748 +#define MIN_PRIO_BIAS          -20
12749 +
12750 +static inline
12751 +int vx_adjust_prio(struct task_struct *p, int prio, int max_user)
12752 +{
12753 +       struct vx_info *vxi = p->vx_info;
12754 +
12755 +       if (vxi)
12756 +               prio += vx_cpu(vxi, sched_pc).prio_bias;
12757 +       return prio;
12758 +}
12759 +
12760 +static inline void vx_account_user(struct vx_info *vxi,
12761 +       cputime_t cputime, int nice)
12762 +{
12763 +       if (!vxi)
12764 +               return;
12765 +       vx_cpu(vxi, sched_pc).user_ticks += cputime;
12766 +}
12767 +
12768 +static inline void vx_account_system(struct vx_info *vxi,
12769 +       cputime_t cputime, int idle)
12770 +{
12771 +       if (!vxi)
12772 +               return;
12773 +       vx_cpu(vxi, sched_pc).sys_ticks += cputime;
12774 +}
12775 +
12776 +#else
12777 +#warning duplicate inclusion
12778 +#endif
12779 diff -NurpP --minimal linux-3.0/include/linux/vs_socket.h linux-3.0-vs2.3.1-pre7/include/linux/vs_socket.h
12780 --- linux-3.0/include/linux/vs_socket.h 1970-01-01 01:00:00.000000000 +0100
12781 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_socket.h    2011-06-10 22:11:24.000000000 +0200
12782 @@ -0,0 +1,67 @@
12783 +#ifndef _VS_SOCKET_H
12784 +#define _VS_SOCKET_H
12785 +
12786 +#include "vserver/debug.h"
12787 +#include "vserver/base.h"
12788 +#include "vserver/cacct.h"
12789 +#include "vserver/context.h"
12790 +#include "vserver/tag.h"
12791 +
12792 +
12793 +/* socket accounting */
12794 +
12795 +#include <linux/socket.h>
12796 +
12797 +static inline int vx_sock_type(int family)
12798 +{
12799 +       switch (family) {
12800 +       case PF_UNSPEC:
12801 +               return VXA_SOCK_UNSPEC;
12802 +       case PF_UNIX:
12803 +               return VXA_SOCK_UNIX;
12804 +       case PF_INET:
12805 +               return VXA_SOCK_INET;
12806 +       case PF_INET6:
12807 +               return VXA_SOCK_INET6;
12808 +       case PF_PACKET:
12809 +               return VXA_SOCK_PACKET;
12810 +       default:
12811 +               return VXA_SOCK_OTHER;
12812 +       }
12813 +}
12814 +
12815 +#define vx_acc_sock(v, f, p, s) \
12816 +       __vx_acc_sock(v, f, p, s, __FILE__, __LINE__)
12817 +
12818 +static inline void __vx_acc_sock(struct vx_info *vxi,
12819 +       int family, int pos, int size, char *file, int line)
12820 +{
12821 +       if (vxi) {
12822 +               int type = vx_sock_type(family);
12823 +
12824 +               atomic_long_inc(&vxi->cacct.sock[type][pos].count);
12825 +               atomic_long_add(size, &vxi->cacct.sock[type][pos].total);
12826 +       }
12827 +}
12828 +
12829 +#define vx_sock_recv(sk, s) \
12830 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 0, s)
12831 +#define vx_sock_send(sk, s) \
12832 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 1, s)
12833 +#define vx_sock_fail(sk, s) \
12834 +       vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 2, s)
12835 +
12836 +
12837 +#define sock_vx_init(s) do {           \
12838 +       (s)->sk_xid = 0;                \
12839 +       (s)->sk_vx_info = NULL;         \
12840 +       } while (0)
12841 +
12842 +#define sock_nx_init(s) do {           \
12843 +       (s)->sk_nid = 0;                \
12844 +       (s)->sk_nx_info = NULL;         \
12845 +       } while (0)
12846 +
12847 +#else
12848 +#warning duplicate inclusion
12849 +#endif
12850 diff -NurpP --minimal linux-3.0/include/linux/vs_tag.h linux-3.0-vs2.3.1-pre7/include/linux/vs_tag.h
12851 --- linux-3.0/include/linux/vs_tag.h    1970-01-01 01:00:00.000000000 +0100
12852 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_tag.h       2011-06-10 22:11:24.000000000 +0200
12853 @@ -0,0 +1,47 @@
12854 +#ifndef _VS_TAG_H
12855 +#define _VS_TAG_H
12856 +
12857 +#include <linux/vserver/tag.h>
12858 +
12859 +/* check conditions */
12860 +
12861 +#define DX_ADMIN       0x0001
12862 +#define DX_WATCH       0x0002
12863 +#define DX_HOSTID      0x0008
12864 +
12865 +#define DX_IDENT       0x0010
12866 +
12867 +#define DX_ARG_MASK    0x0010
12868 +
12869 +
12870 +#define dx_task_tag(t) ((t)->tag)
12871 +
12872 +#define dx_current_tag() dx_task_tag(current)
12873 +
12874 +#define dx_check(c, m) __dx_check(dx_current_tag(), c, m)
12875 +
12876 +#define dx_weak_check(c, m)    ((m) ? dx_check(c, m) : 1)
12877 +
12878 +
12879 +/*
12880 + * check current context for ADMIN/WATCH and
12881 + * optionally against supplied argument
12882 + */
12883 +static inline int __dx_check(tag_t cid, tag_t id, unsigned int mode)
12884 +{
12885 +       if (mode & DX_ARG_MASK) {
12886 +               if ((mode & DX_IDENT) && (id == cid))
12887 +                       return 1;
12888 +       }
12889 +       return (((mode & DX_ADMIN) && (cid == 0)) ||
12890 +               ((mode & DX_WATCH) && (cid == 1)) ||
12891 +               ((mode & DX_HOSTID) && (id == 0)));
12892 +}
12893 +
12894 +struct inode;
12895 +int dx_permission(const struct inode *inode, int mask);
12896 +
12897 +
12898 +#else
12899 +#warning duplicate inclusion
12900 +#endif
12901 diff -NurpP --minimal linux-3.0/include/linux/vs_time.h linux-3.0-vs2.3.1-pre7/include/linux/vs_time.h
12902 --- linux-3.0/include/linux/vs_time.h   1970-01-01 01:00:00.000000000 +0100
12903 +++ linux-3.0-vs2.3.1-pre7/include/linux/vs_time.h      2011-06-13 14:57:45.000000000 +0200
12904 @@ -0,0 +1,19 @@
12905 +#ifndef _VS_TIME_H
12906 +#define _VS_TIME_H
12907 +
12908 +
12909 +/* time faking stuff */
12910 +
12911 +#ifdef CONFIG_VSERVER_VTIME
12912 +
12913 +extern void vx_adjust_timespec(struct timespec *ts);
12914 +extern int vx_settimeofday(const struct timespec *ts);
12915 +
12916 +#else
12917 +#define        vx_adjust_timespec(t)   do { } while (0)
12918 +#define        vx_settimeofday(t)      do_settimeofday(t)
12919 +#endif
12920 +
12921 +#else
12922 +#warning duplicate inclusion
12923 +#endif
12924 diff -NurpP --minimal linux-3.0/include/linux/vserver/Kbuild linux-3.0-vs2.3.1-pre7/include/linux/vserver/Kbuild
12925 --- linux-3.0/include/linux/vserver/Kbuild      1970-01-01 01:00:00.000000000 +0100
12926 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/Kbuild 2011-06-10 22:11:24.000000000 +0200
12927 @@ -0,0 +1,8 @@
12928 +
12929 +unifdef-y += context_cmd.h network_cmd.h space_cmd.h \
12930 +       cacct_cmd.h cvirt_cmd.h limit_cmd.h dlimit_cmd.h \
12931 +       inode_cmd.h tag_cmd.h sched_cmd.h signal_cmd.h \
12932 +       debug_cmd.h device_cmd.h
12933 +
12934 +unifdef-y += switch.h network.h monitor.h inode.h device.h
12935 +
12936 diff -NurpP --minimal linux-3.0/include/linux/vserver/base.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/base.h
12937 --- linux-3.0/include/linux/vserver/base.h      1970-01-01 01:00:00.000000000 +0100
12938 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/base.h 2011-06-13 14:09:44.000000000 +0200
12939 @@ -0,0 +1,173 @@
12940 +#ifndef _VX_BASE_H
12941 +#define _VX_BASE_H
12942 +
12943 +
12944 +/* context state changes */
12945 +
12946 +enum {
12947 +       VSC_STARTUP = 1,
12948 +       VSC_SHUTDOWN,
12949 +
12950 +       VSC_NETUP,
12951 +       VSC_NETDOWN,
12952 +};
12953 +
12954 +
12955 +
12956 +#define vx_task_xid(t) ((t)->xid)
12957 +
12958 +#define vx_current_xid() vx_task_xid(current)
12959 +
12960 +#define current_vx_info() (current->vx_info)
12961 +
12962 +
12963 +#define nx_task_nid(t) ((t)->nid)
12964 +
12965 +#define nx_current_nid() nx_task_nid(current)
12966 +
12967 +#define current_nx_info() (current->nx_info)
12968 +
12969 +
12970 +/* generic flag merging */
12971 +
12972 +#define vs_check_flags(v, m, f)        (((v) & (m)) ^ (f))
12973 +
12974 +#define vs_mask_flags(v, f, m) (((v) & ~(m)) | ((f) & (m)))
12975 +
12976 +#define vs_mask_mask(v, f, m)  (((v) & ~(m)) | ((v) & (f) & (m)))
12977 +
12978 +#define vs_check_bit(v, n)     ((v) & (1LL << (n)))
12979 +
12980 +
12981 +/* context flags */
12982 +
12983 +#define __vx_flags(v)  ((v) ? (v)->vx_flags : 0)
12984 +
12985 +#define vx_current_flags()     __vx_flags(current_vx_info())
12986 +
12987 +#define vx_info_flags(v, m, f) \
12988 +       vs_check_flags(__vx_flags(v), m, f)
12989 +
12990 +#define task_vx_flags(t, m, f) \
12991 +       ((t) && vx_info_flags((t)->vx_info, m, f))
12992 +
12993 +#define vx_flags(m, f) vx_info_flags(current_vx_info(), m, f)
12994 +
12995 +
12996 +/* context caps */
12997 +
12998 +#define __vx_ccaps(v)  ((v) ? (v)->vx_ccaps : 0)
12999 +
13000 +#define vx_current_ccaps()     __vx_ccaps(current_vx_info())
13001 +
13002 +#define vx_info_ccaps(v, c)    (__vx_ccaps(v) & (c))
13003 +
13004 +#define vx_ccaps(c)    vx_info_ccaps(current_vx_info(), (c))
13005 +
13006 +
13007 +
13008 +/* network flags */
13009 +
13010 +#define __nx_flags(n)  ((n) ? (n)->nx_flags : 0)
13011 +
13012 +#define nx_current_flags()     __nx_flags(current_nx_info())
13013 +
13014 +#define nx_info_flags(n, m, f) \
13015 +       vs_check_flags(__nx_flags(n), m, f)
13016 +
13017 +#define task_nx_flags(t, m, f) \
13018 +       ((t) && nx_info_flags((t)->nx_info, m, f))
13019 +
13020 +#define nx_flags(m, f) nx_info_flags(current_nx_info(), m, f)
13021 +
13022 +
13023 +/* network caps */
13024 +
13025 +#define __nx_ncaps(n)  ((n) ? (n)->nx_ncaps : 0)
13026 +
13027 +#define nx_current_ncaps()     __nx_ncaps(current_nx_info())
13028 +
13029 +#define nx_info_ncaps(n, c)    (__nx_ncaps(n) & (c))
13030 +
13031 +#define nx_ncaps(c)    nx_info_ncaps(current_nx_info(), c)
13032 +
13033 +
13034 +/* context mask capabilities */
13035 +
13036 +#define __vx_mcaps(v)  ((v) ? (v)->vx_ccaps >> 32UL : ~0 )
13037 +
13038 +#define vx_info_mcaps(v, c)    (__vx_mcaps(v) & (c))
13039 +
13040 +#define vx_mcaps(c)    vx_info_mcaps(current_vx_info(), c)
13041 +
13042 +
13043 +/* context bcap mask */
13044 +
13045 +#define __vx_bcaps(v)          ((v)->vx_bcaps)
13046 +
13047 +#define vx_current_bcaps()     __vx_bcaps(current_vx_info())
13048 +
13049 +
13050 +/* mask given bcaps */
13051 +
13052 +#define vx_info_mbcaps(v, c)   ((v) ? cap_intersect(__vx_bcaps(v), c) : c)
13053 +
13054 +#define vx_mbcaps(c)           vx_info_mbcaps(current_vx_info(), c)
13055 +
13056 +
13057 +/* masked cap_bset */
13058 +
13059 +#define vx_info_cap_bset(v)    vx_info_mbcaps(v, current->cap_bset)
13060 +
13061 +#define vx_current_cap_bset()  vx_info_cap_bset(current_vx_info())
13062 +
13063 +#if 0
13064 +#define vx_info_mbcap(v, b) \
13065 +       (!vx_info_flags(v, VXF_STATE_SETUP, 0) ? \
13066 +       vx_info_bcaps(v, b) : (b))
13067 +
13068 +#define task_vx_mbcap(t, b) \
13069 +       vx_info_mbcap((t)->vx_info, (t)->b)
13070 +
13071 +#define vx_mbcap(b)    task_vx_mbcap(current, b)
13072 +#endif
13073 +
13074 +#define vx_cap_raised(v, c, f) cap_raised(vx_info_mbcaps(v, c), f)
13075 +
13076 +#define vx_capable(b, c) (capable(b) || \
13077 +       (cap_raised(current_cap(), b) && vx_ccaps(c)))
13078 +
13079 +#define vx_ns_capable(n, b, c) (ns_capable(n, b) || \
13080 +       (cap_raised(current_cap(), b) && vx_ccaps(c)))
13081 +
13082 +#define nx_capable(b, c) (capable(b) || \
13083 +       (cap_raised(current_cap(), b) && nx_ncaps(c)))
13084 +
13085 +#define vx_task_initpid(t, n) \
13086 +       ((t)->vx_info && \
13087 +       ((t)->vx_info->vx_initpid == (n)))
13088 +
13089 +#define vx_current_initpid(n)  vx_task_initpid(current, n)
13090 +
13091 +
13092 +/* context unshare mask */
13093 +
13094 +#define __vx_umask(v)          ((v)->vx_umask)
13095 +
13096 +#define vx_current_umask()     __vx_umask(current_vx_info())
13097 +
13098 +#define vx_can_unshare(b, f) (capable(b) || \
13099 +       (cap_raised(current_cap(), b) && \
13100 +       !((f) & ~vx_current_umask())))
13101 +
13102 +
13103 +#define __vx_state(v)  ((v) ? ((v)->vx_state) : 0)
13104 +
13105 +#define vx_info_state(v, m)    (__vx_state(v) & (m))
13106 +
13107 +
13108 +#define __nx_state(n)  ((n) ? ((n)->nx_state) : 0)
13109 +
13110 +#define nx_info_state(n, m)    (__nx_state(n) & (m))
13111 +
13112 +#endif
13113 diff -NurpP --minimal linux-3.0/include/linux/vserver/cacct.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct.h
13114 --- linux-3.0/include/linux/vserver/cacct.h     1970-01-01 01:00:00.000000000 +0100
13115 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct.h        2011-06-10 22:11:24.000000000 +0200
13116 @@ -0,0 +1,15 @@
13117 +#ifndef _VX_CACCT_H
13118 +#define _VX_CACCT_H
13119 +
13120 +
13121 +enum sock_acc_field {
13122 +       VXA_SOCK_UNSPEC = 0,
13123 +       VXA_SOCK_UNIX,
13124 +       VXA_SOCK_INET,
13125 +       VXA_SOCK_INET6,
13126 +       VXA_SOCK_PACKET,
13127 +       VXA_SOCK_OTHER,
13128 +       VXA_SOCK_SIZE   /* array size */
13129 +};
13130 +
13131 +#endif /* _VX_CACCT_H */
13132 diff -NurpP --minimal linux-3.0/include/linux/vserver/cacct_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct_cmd.h
13133 --- linux-3.0/include/linux/vserver/cacct_cmd.h 1970-01-01 01:00:00.000000000 +0100
13134 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct_cmd.h    2011-06-10 22:11:24.000000000 +0200
13135 @@ -0,0 +1,23 @@
13136 +#ifndef _VX_CACCT_CMD_H
13137 +#define _VX_CACCT_CMD_H
13138 +
13139 +
13140 +/* virtual host info name commands */
13141 +
13142 +#define VCMD_sock_stat         VC_CMD(VSTAT, 5, 0)
13143 +
13144 +struct vcmd_sock_stat_v0 {
13145 +       uint32_t field;
13146 +       uint32_t count[3];
13147 +       uint64_t total[3];
13148 +};
13149 +
13150 +
13151 +#ifdef __KERNEL__
13152 +
13153 +#include <linux/compiler.h>
13154 +
13155 +extern int vc_sock_stat(struct vx_info *, void __user *);
13156 +
13157 +#endif /* __KERNEL__ */
13158 +#endif /* _VX_CACCT_CMD_H */
13159 diff -NurpP --minimal linux-3.0/include/linux/vserver/cacct_def.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct_def.h
13160 --- linux-3.0/include/linux/vserver/cacct_def.h 1970-01-01 01:00:00.000000000 +0100
13161 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct_def.h    2011-06-10 22:11:24.000000000 +0200
13162 @@ -0,0 +1,43 @@
13163 +#ifndef _VX_CACCT_DEF_H
13164 +#define _VX_CACCT_DEF_H
13165 +
13166 +#include <asm/atomic.h>
13167 +#include <linux/vserver/cacct.h>
13168 +
13169 +
13170 +struct _vx_sock_acc {
13171 +       atomic_long_t count;
13172 +       atomic_long_t total;
13173 +};
13174 +
13175 +/* context sub struct */
13176 +
13177 +struct _vx_cacct {
13178 +       struct _vx_sock_acc sock[VXA_SOCK_SIZE][3];
13179 +       atomic_t slab[8];
13180 +       atomic_t page[6][8];
13181 +};
13182 +
13183 +#ifdef CONFIG_VSERVER_DEBUG
13184 +
13185 +static inline void __dump_vx_cacct(struct _vx_cacct *cacct)
13186 +{
13187 +       int i, j;
13188 +
13189 +       printk("\t_vx_cacct:");
13190 +       for (i = 0; i < 6; i++) {
13191 +               struct _vx_sock_acc *ptr = cacct->sock[i];
13192 +
13193 +               printk("\t [%d] =", i);
13194 +               for (j = 0; j < 3; j++) {
13195 +                       printk(" [%d] = %8lu, %8lu", j,
13196 +                               atomic_long_read(&ptr[j].count),
13197 +                               atomic_long_read(&ptr[j].total));
13198 +               }
13199 +               printk("\n");
13200 +       }
13201 +}
13202 +
13203 +#endif
13204 +
13205 +#endif /* _VX_CACCT_DEF_H */
13206 diff -NurpP --minimal linux-3.0/include/linux/vserver/cacct_int.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct_int.h
13207 --- linux-3.0/include/linux/vserver/cacct_int.h 1970-01-01 01:00:00.000000000 +0100
13208 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cacct_int.h    2011-06-10 22:11:24.000000000 +0200
13209 @@ -0,0 +1,21 @@
13210 +#ifndef _VX_CACCT_INT_H
13211 +#define _VX_CACCT_INT_H
13212 +
13213 +
13214 +#ifdef __KERNEL__
13215 +
13216 +static inline
13217 +unsigned long vx_sock_count(struct _vx_cacct *cacct, int type, int pos)
13218 +{
13219 +       return atomic_long_read(&cacct->sock[type][pos].count);
13220 +}
13221 +
13222 +
13223 +static inline
13224 +unsigned long vx_sock_total(struct _vx_cacct *cacct, int type, int pos)
13225 +{
13226 +       return atomic_long_read(&cacct->sock[type][pos].total);
13227 +}
13228 +
13229 +#endif /* __KERNEL__ */
13230 +#endif /* _VX_CACCT_INT_H */
13231 diff -NurpP --minimal linux-3.0/include/linux/vserver/check.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/check.h
13232 --- linux-3.0/include/linux/vserver/check.h     1970-01-01 01:00:00.000000000 +0100
13233 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/check.h        2011-06-10 22:11:24.000000000 +0200
13234 @@ -0,0 +1,89 @@
13235 +#ifndef _VS_CHECK_H
13236 +#define _VS_CHECK_H
13237 +
13238 +
13239 +#define MAX_S_CONTEXT  65535   /* Arbitrary limit */
13240 +
13241 +#ifdef CONFIG_VSERVER_DYNAMIC_IDS
13242 +#define MIN_D_CONTEXT  49152   /* dynamic contexts start here */
13243 +#else
13244 +#define MIN_D_CONTEXT  65536
13245 +#endif
13246 +
13247 +/* check conditions */
13248 +
13249 +#define VS_ADMIN       0x0001
13250 +#define VS_WATCH       0x0002
13251 +#define VS_HIDE                0x0004
13252 +#define VS_HOSTID      0x0008
13253 +
13254 +#define VS_IDENT       0x0010
13255 +#define VS_EQUIV       0x0020
13256 +#define VS_PARENT      0x0040
13257 +#define VS_CHILD       0x0080
13258 +
13259 +#define VS_ARG_MASK    0x00F0
13260 +
13261 +#define VS_DYNAMIC     0x0100
13262 +#define VS_STATIC      0x0200
13263 +
13264 +#define VS_ATR_MASK    0x0F00
13265 +
13266 +#ifdef CONFIG_VSERVER_PRIVACY
13267 +#define VS_ADMIN_P     (0)
13268 +#define VS_WATCH_P     (0)
13269 +#else
13270 +#define VS_ADMIN_P     VS_ADMIN
13271 +#define VS_WATCH_P     VS_WATCH
13272 +#endif
13273 +
13274 +#define VS_HARDIRQ     0x1000
13275 +#define VS_SOFTIRQ     0x2000
13276 +#define VS_IRQ         0x4000
13277 +
13278 +#define VS_IRQ_MASK    0xF000
13279 +
13280 +#include <linux/hardirq.h>
13281 +
13282 +/*
13283 + * check current context for ADMIN/WATCH and
13284 + * optionally against supplied argument
13285 + */
13286 +static inline int __vs_check(int cid, int id, unsigned int mode)
13287 +{
13288 +       if (mode & VS_ARG_MASK) {
13289 +               if ((mode & VS_IDENT) && (id == cid))
13290 +                       return 1;
13291 +       }
13292 +       if (mode & VS_ATR_MASK) {
13293 +               if ((mode & VS_DYNAMIC) &&
13294 +                       (id >= MIN_D_CONTEXT) &&
13295 +                       (id <= MAX_S_CONTEXT))
13296 +                       return 1;
13297 +               if ((mode & VS_STATIC) &&
13298 +                       (id > 1) && (id < MIN_D_CONTEXT))
13299 +                       return 1;
13300 +       }
13301 +       if (mode & VS_IRQ_MASK) {
13302 +               if ((mode & VS_IRQ) && unlikely(in_interrupt()))
13303 +                       return 1;
13304 +               if ((mode & VS_HARDIRQ) && unlikely(in_irq()))
13305 +                       return 1;
13306 +               if ((mode & VS_SOFTIRQ) && unlikely(in_softirq()))
13307 +                       return 1;
13308 +       }
13309 +       return (((mode & VS_ADMIN) && (cid == 0)) ||
13310 +               ((mode & VS_WATCH) && (cid == 1)) ||
13311 +               ((mode & VS_HOSTID) && (id == 0)));
13312 +}
13313 +
13314 +#define vx_check(c, m) __vs_check(vx_current_xid(), c, (m) | VS_IRQ)
13315 +
13316 +#define vx_weak_check(c, m)    ((m) ? vx_check(c, m) : 1)
13317 +
13318 +
13319 +#define nx_check(c, m) __vs_check(nx_current_nid(), c, m)
13320 +
13321 +#define nx_weak_check(c, m)    ((m) ? nx_check(c, m) : 1)
13322 +
13323 +#endif
13324 diff -NurpP --minimal linux-3.0/include/linux/vserver/context.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/context.h
13325 --- linux-3.0/include/linux/vserver/context.h   1970-01-01 01:00:00.000000000 +0100
13326 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/context.h      2011-06-10 22:11:24.000000000 +0200
13327 @@ -0,0 +1,189 @@
13328 +#ifndef _VX_CONTEXT_H
13329 +#define _VX_CONTEXT_H
13330 +
13331 +#include <linux/types.h>
13332 +#include <linux/capability.h>
13333 +
13334 +
13335 +/* context flags */
13336 +
13337 +#define VXF_INFO_SCHED         0x00000002
13338 +#define VXF_INFO_NPROC         0x00000004
13339 +#define VXF_INFO_PRIVATE       0x00000008
13340 +
13341 +#define VXF_INFO_INIT          0x00000010
13342 +#define VXF_INFO_HIDE          0x00000020
13343 +#define VXF_INFO_ULIMIT                0x00000040
13344 +#define VXF_INFO_NSPACE                0x00000080
13345 +
13346 +#define VXF_SCHED_HARD         0x00000100
13347 +#define VXF_SCHED_PRIO         0x00000200
13348 +#define VXF_SCHED_PAUSE                0x00000400
13349 +
13350 +#define VXF_VIRT_MEM           0x00010000
13351 +#define VXF_VIRT_UPTIME                0x00020000
13352 +#define VXF_VIRT_CPU           0x00040000
13353 +#define VXF_VIRT_LOAD          0x00080000
13354 +#define VXF_VIRT_TIME          0x00100000
13355 +
13356 +#define VXF_HIDE_MOUNT         0x01000000
13357 +/* was VXF_HIDE_NETIF          0x02000000 */
13358 +#define VXF_HIDE_VINFO         0x04000000
13359 +
13360 +#define VXF_STATE_SETUP                (1ULL << 32)
13361 +#define VXF_STATE_INIT         (1ULL << 33)
13362 +#define VXF_STATE_ADMIN                (1ULL << 34)
13363 +
13364 +#define VXF_SC_HELPER          (1ULL << 36)
13365 +#define VXF_REBOOT_KILL                (1ULL << 37)
13366 +#define VXF_PERSISTENT         (1ULL << 38)
13367 +
13368 +#define VXF_FORK_RSS           (1ULL << 48)
13369 +#define VXF_PROLIFIC           (1ULL << 49)
13370 +
13371 +#define VXF_IGNEG_NICE         (1ULL << 52)
13372 +
13373 +#define VXF_ONE_TIME           (0x0007ULL << 32)
13374 +
13375 +#define VXF_INIT_SET           (VXF_STATE_SETUP | VXF_STATE_INIT | VXF_STATE_ADMIN)
13376 +
13377 +
13378 +/* context migration */
13379 +
13380 +#define VXM_SET_INIT           0x00000001
13381 +#define VXM_SET_REAPER         0x00000002
13382 +
13383 +/* context caps */
13384 +
13385 +#define VXC_CAP_MASK           0x00000000
13386 +
13387 +#define VXC_SET_UTSNAME                0x00000001
13388 +#define VXC_SET_RLIMIT         0x00000002
13389 +#define VXC_FS_SECURITY                0x00000004
13390 +#define VXC_FS_TRUSTED         0x00000008
13391 +#define VXC_TIOCSTI            0x00000010
13392 +
13393 +/* was VXC_RAW_ICMP            0x00000100 */
13394 +#define VXC_SYSLOG             0x00001000
13395 +#define VXC_OOM_ADJUST         0x00002000
13396 +#define VXC_AUDIT_CONTROL      0x00004000
13397 +
13398 +#define VXC_SECURE_MOUNT       0x00010000
13399 +#define VXC_SECURE_REMOUNT     0x00020000
13400 +#define VXC_BINARY_MOUNT       0x00040000
13401 +
13402 +#define VXC_QUOTA_CTL          0x00100000
13403 +#define VXC_ADMIN_MAPPER       0x00200000
13404 +#define VXC_ADMIN_CLOOP                0x00400000
13405 +
13406 +#define VXC_KTHREAD            0x01000000
13407 +#define VXC_NAMESPACE          0x02000000
13408 +
13409 +
13410 +#ifdef __KERNEL__
13411 +
13412 +#include <linux/list.h>
13413 +#include <linux/spinlock.h>
13414 +#include <linux/rcupdate.h>
13415 +
13416 +#include "limit_def.h"
13417 +#include "sched_def.h"
13418 +#include "cvirt_def.h"
13419 +#include "cacct_def.h"
13420 +#include "device_def.h"
13421 +
13422 +#define VX_SPACES      2
13423 +
13424 +struct _vx_info_pc {
13425 +       struct _vx_sched_pc sched_pc;
13426 +       struct _vx_cvirt_pc cvirt_pc;
13427 +};
13428 +
13429 +struct _vx_space {
13430 +       unsigned long vx_nsmask;                /* assignment mask */
13431 +       struct nsproxy *vx_nsproxy;             /* private namespaces */
13432 +       struct fs_struct *vx_fs;                /* private namespace fs */
13433 +       const struct cred *vx_cred;             /* task credentials */
13434 +};
13435 +
13436 +struct vx_info {
13437 +       struct hlist_node vx_hlist;             /* linked list of contexts */
13438 +       xid_t vx_id;                            /* context id */
13439 +       atomic_t vx_usecnt;                     /* usage count */
13440 +       atomic_t vx_tasks;                      /* tasks count */
13441 +       struct vx_info *vx_parent;              /* parent context */
13442 +       int vx_state;                           /* context state */
13443 +
13444 +       struct _vx_space space[VX_SPACES];      /* namespace store */
13445 +
13446 +       uint64_t vx_flags;                      /* context flags */
13447 +       uint64_t vx_ccaps;                      /* context caps (vserver) */
13448 +       uint64_t vx_umask;                      /* unshare mask (guest) */
13449 +       kernel_cap_t vx_bcaps;                  /* bounding caps (system) */
13450 +
13451 +       struct task_struct *vx_reaper;          /* guest reaper process */
13452 +       pid_t vx_initpid;                       /* PID of guest init */
13453 +       int64_t vx_badness_bias;                /* OOM points bias */
13454 +
13455 +       struct _vx_limit limit;                 /* vserver limits */
13456 +       struct _vx_sched sched;                 /* vserver scheduler */
13457 +       struct _vx_cvirt cvirt;                 /* virtual/bias stuff */
13458 +       struct _vx_cacct cacct;                 /* context accounting */
13459 +
13460 +       struct _vx_device dmap;                 /* default device map targets */
13461 +
13462 +#ifndef CONFIG_SMP
13463 +       struct _vx_info_pc info_pc;             /* per cpu data */
13464 +#else
13465 +       struct _vx_info_pc *ptr_pc;             /* per cpu array */
13466 +#endif
13467 +
13468 +       wait_queue_head_t vx_wait;              /* context exit waitqueue */
13469 +       int reboot_cmd;                         /* last sys_reboot() cmd */
13470 +       int exit_code;                          /* last process exit code */
13471 +
13472 +       char vx_name[65];                       /* vserver name */
13473 +};
13474 +
13475 +#ifndef CONFIG_SMP
13476 +#define        vx_ptr_pc(vxi)          (&(vxi)->info_pc)
13477 +#define        vx_per_cpu(vxi, v, id)  vx_ptr_pc(vxi)->v
13478 +#else
13479 +#define        vx_ptr_pc(vxi)          ((vxi)->ptr_pc)
13480 +#define        vx_per_cpu(vxi, v, id)  per_cpu_ptr(vx_ptr_pc(vxi), id)->v
13481 +#endif
13482 +
13483 +#define        vx_cpu(vxi, v)          vx_per_cpu(vxi, v, smp_processor_id())
13484 +
13485 +
13486 +struct vx_info_save {
13487 +       struct vx_info *vxi;
13488 +       xid_t xid;
13489 +};
13490 +
13491 +
13492 +/* status flags */
13493 +
13494 +#define VXS_HASHED     0x0001
13495 +#define VXS_PAUSED     0x0010
13496 +#define VXS_SHUTDOWN   0x0100
13497 +#define VXS_HELPER     0x1000
13498 +#define VXS_RELEASED   0x8000
13499 +
13500 +
13501 +extern void claim_vx_info(struct vx_info *, struct task_struct *);
13502 +extern void release_vx_info(struct vx_info *, struct task_struct *);
13503 +
13504 +extern struct vx_info *lookup_vx_info(int);
13505 +extern struct vx_info *lookup_or_create_vx_info(int);
13506 +
13507 +extern int get_xid_list(int, unsigned int *, int);
13508 +extern int xid_is_hashed(xid_t);
13509 +
13510 +extern int vx_migrate_task(struct task_struct *, struct vx_info *, int);
13511 +
13512 +extern long vs_state_change(struct vx_info *, unsigned int);
13513 +
13514 +
13515 +#endif /* __KERNEL__ */
13516 +#endif /* _VX_CONTEXT_H */
13517 diff -NurpP --minimal linux-3.0/include/linux/vserver/context_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/context_cmd.h
13518 --- linux-3.0/include/linux/vserver/context_cmd.h       1970-01-01 01:00:00.000000000 +0100
13519 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/context_cmd.h  2011-06-10 22:11:24.000000000 +0200
13520 @@ -0,0 +1,145 @@
13521 +#ifndef _VX_CONTEXT_CMD_H
13522 +#define _VX_CONTEXT_CMD_H
13523 +
13524 +
13525 +/* vinfo commands */
13526 +
13527 +#define VCMD_task_xid          VC_CMD(VINFO, 1, 0)
13528 +
13529 +#ifdef __KERNEL__
13530 +extern int vc_task_xid(uint32_t);
13531 +
13532 +#endif /* __KERNEL__ */
13533 +
13534 +#define VCMD_vx_info           VC_CMD(VINFO, 5, 0)
13535 +
13536 +struct vcmd_vx_info_v0 {
13537 +       uint32_t xid;
13538 +       uint32_t initpid;
13539 +       /* more to come */
13540 +};
13541 +
13542 +#ifdef __KERNEL__
13543 +extern int vc_vx_info(struct vx_info *, void __user *);
13544 +
13545 +#endif /* __KERNEL__ */
13546 +
13547 +#define VCMD_ctx_stat          VC_CMD(VSTAT, 0, 0)
13548 +
13549 +struct vcmd_ctx_stat_v0 {
13550 +       uint32_t usecnt;
13551 +       uint32_t tasks;
13552 +       /* more to come */
13553 +};
13554 +
13555 +#ifdef __KERNEL__
13556 +extern int vc_ctx_stat(struct vx_info *, void __user *);
13557 +
13558 +#endif /* __KERNEL__ */
13559 +
13560 +/* context commands */
13561 +
13562 +#define VCMD_ctx_create_v0     VC_CMD(VPROC, 1, 0)
13563 +#define VCMD_ctx_create                VC_CMD(VPROC, 1, 1)
13564 +
13565 +struct vcmd_ctx_create {
13566 +       uint64_t flagword;
13567 +};
13568 +
13569 +#define VCMD_ctx_migrate_v0    VC_CMD(PROCMIG, 1, 0)
13570 +#define VCMD_ctx_migrate       VC_CMD(PROCMIG, 1, 1)
13571 +
13572 +struct vcmd_ctx_migrate {
13573 +       uint64_t flagword;
13574 +};
13575 +
13576 +#ifdef __KERNEL__
13577 +extern int vc_ctx_create(uint32_t, void __user *);
13578 +extern int vc_ctx_migrate(struct vx_info *, void __user *);
13579 +
13580 +#endif /* __KERNEL__ */
13581 +
13582 +
13583 +/* flag commands */
13584 +
13585 +#define VCMD_get_cflags                VC_CMD(FLAGS, 1, 0)
13586 +#define VCMD_set_cflags                VC_CMD(FLAGS, 2, 0)
13587 +
13588 +struct vcmd_ctx_flags_v0 {
13589 +       uint64_t flagword;
13590 +       uint64_t mask;
13591 +};
13592 +
13593 +#ifdef __KERNEL__
13594 +extern int vc_get_cflags(struct vx_info *, void __user *);
13595 +extern int vc_set_cflags(struct vx_info *, void __user *);
13596 +
13597 +#endif /* __KERNEL__ */
13598 +
13599 +
13600 +/* context caps commands */
13601 +
13602 +#define VCMD_get_ccaps         VC_CMD(FLAGS, 3, 1)
13603 +#define VCMD_set_ccaps         VC_CMD(FLAGS, 4, 1)
13604 +
13605 +struct vcmd_ctx_caps_v1 {
13606 +       uint64_t ccaps;
13607 +       uint64_t cmask;
13608 +};
13609 +
13610 +#ifdef __KERNEL__
13611 +extern int vc_get_ccaps(struct vx_info *, void __user *);
13612 +extern int vc_set_ccaps(struct vx_info *, void __user *);
13613 +
13614 +#endif /* __KERNEL__ */
13615 +
13616 +
13617 +/* bcaps commands */
13618 +
13619 +#define VCMD_get_bcaps         VC_CMD(FLAGS, 9, 0)
13620 +#define VCMD_set_bcaps         VC_CMD(FLAGS, 10, 0)
13621 +
13622 +struct vcmd_bcaps {
13623 +       uint64_t bcaps;
13624 +       uint64_t bmask;
13625 +};
13626 +
13627 +#ifdef __KERNEL__
13628 +extern int vc_get_bcaps(struct vx_info *, void __user *);
13629 +extern int vc_set_bcaps(struct vx_info *, void __user *);
13630 +
13631 +#endif /* __KERNEL__ */
13632 +
13633 +
13634 +/* umask commands */
13635 +
13636 +#define VCMD_get_umask         VC_CMD(FLAGS, 13, 0)
13637 +#define VCMD_set_umask         VC_CMD(FLAGS, 14, 0)
13638 +
13639 +struct vcmd_umask {
13640 +       uint64_t umask;
13641 +       uint64_t mask;
13642 +};
13643 +
13644 +#ifdef __KERNEL__
13645 +extern int vc_get_umask(struct vx_info *, void __user *);
13646 +extern int vc_set_umask(struct vx_info *, void __user *);
13647 +
13648 +#endif /* __KERNEL__ */
13649 +
13650 +
13651 +/* OOM badness */
13652 +
13653 +#define VCMD_get_badness       VC_CMD(MEMCTRL, 5, 0)
13654 +#define VCMD_set_badness       VC_CMD(MEMCTRL, 6, 0)
13655 +
13656 +struct vcmd_badness_v0 {
13657 +       int64_t bias;
13658 +};
13659 +
13660 +#ifdef __KERNEL__
13661 +extern int vc_get_badness(struct vx_info *, void __user *);
13662 +extern int vc_set_badness(struct vx_info *, void __user *);
13663 +
13664 +#endif /* __KERNEL__ */
13665 +#endif /* _VX_CONTEXT_CMD_H */
13666 diff -NurpP --minimal linux-3.0/include/linux/vserver/cvirt.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cvirt.h
13667 --- linux-3.0/include/linux/vserver/cvirt.h     1970-01-01 01:00:00.000000000 +0100
13668 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cvirt.h        2011-06-10 22:11:24.000000000 +0200
13669 @@ -0,0 +1,20 @@
13670 +#ifndef _VX_CVIRT_H
13671 +#define _VX_CVIRT_H
13672 +
13673 +
13674 +#ifdef __KERNEL__
13675 +
13676 +struct timespec;
13677 +
13678 +void vx_vsi_uptime(struct timespec *, struct timespec *);
13679 +
13680 +
13681 +struct vx_info;
13682 +
13683 +void vx_update_load(struct vx_info *);
13684 +
13685 +
13686 +int vx_do_syslog(int, char __user *, int);
13687 +
13688 +#endif /* __KERNEL__ */
13689 +#endif /* _VX_CVIRT_H */
13690 diff -NurpP --minimal linux-3.0/include/linux/vserver/cvirt_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cvirt_cmd.h
13691 --- linux-3.0/include/linux/vserver/cvirt_cmd.h 1970-01-01 01:00:00.000000000 +0100
13692 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cvirt_cmd.h    2011-06-10 22:11:24.000000000 +0200
13693 @@ -0,0 +1,53 @@
13694 +#ifndef _VX_CVIRT_CMD_H
13695 +#define _VX_CVIRT_CMD_H
13696 +
13697 +
13698 +/* virtual host info name commands */
13699 +
13700 +#define VCMD_set_vhi_name      VC_CMD(VHOST, 1, 0)
13701 +#define VCMD_get_vhi_name      VC_CMD(VHOST, 2, 0)
13702 +
13703 +struct vcmd_vhi_name_v0 {
13704 +       uint32_t field;
13705 +       char name[65];
13706 +};
13707 +
13708 +
13709 +enum vhi_name_field {
13710 +       VHIN_CONTEXT = 0,
13711 +       VHIN_SYSNAME,
13712 +       VHIN_NODENAME,
13713 +       VHIN_RELEASE,
13714 +       VHIN_VERSION,
13715 +       VHIN_MACHINE,
13716 +       VHIN_DOMAINNAME,
13717 +};
13718 +
13719 +
13720 +#ifdef __KERNEL__
13721 +
13722 +#include <linux/compiler.h>
13723 +
13724 +extern int vc_set_vhi_name(struct vx_info *, void __user *);
13725 +extern int vc_get_vhi_name(struct vx_info *, void __user *);
13726 +
13727 +#endif /* __KERNEL__ */
13728 +
13729 +#define VCMD_virt_stat         VC_CMD(VSTAT, 3, 0)
13730 +
13731 +struct vcmd_virt_stat_v0 {
13732 +       uint64_t offset;
13733 +       uint64_t uptime;
13734 +       uint32_t nr_threads;
13735 +       uint32_t nr_running;
13736 +       uint32_t nr_uninterruptible;
13737 +       uint32_t nr_onhold;
13738 +       uint32_t nr_forks;
13739 +       uint32_t load[3];
13740 +};
13741 +
13742 +#ifdef __KERNEL__
13743 +extern int vc_virt_stat(struct vx_info *, void __user *);
13744 +
13745 +#endif /* __KERNEL__ */
13746 +#endif /* _VX_CVIRT_CMD_H */
13747 diff -NurpP --minimal linux-3.0/include/linux/vserver/cvirt_def.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/cvirt_def.h
13748 --- linux-3.0/include/linux/vserver/cvirt_def.h 1970-01-01 01:00:00.000000000 +0100
13749 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/cvirt_def.h    2011-06-10 22:11:24.000000000 +0200
13750 @@ -0,0 +1,80 @@
13751 +#ifndef _VX_CVIRT_DEF_H
13752 +#define _VX_CVIRT_DEF_H
13753 +
13754 +#include <linux/jiffies.h>
13755 +#include <linux/spinlock.h>
13756 +#include <linux/wait.h>
13757 +#include <linux/time.h>
13758 +#include <asm/atomic.h>
13759 +
13760 +
13761 +struct _vx_usage_stat {
13762 +       uint64_t user;
13763 +       uint64_t nice;
13764 +       uint64_t system;
13765 +       uint64_t softirq;
13766 +       uint64_t irq;
13767 +       uint64_t idle;
13768 +       uint64_t iowait;
13769 +};
13770 +
13771 +struct _vx_syslog {
13772 +       wait_queue_head_t log_wait;
13773 +       spinlock_t logbuf_lock;         /* lock for the log buffer */
13774 +
13775 +       unsigned long log_start;        /* next char to be read by syslog() */
13776 +       unsigned long con_start;        /* next char to be sent to consoles */
13777 +       unsigned long log_end;  /* most-recently-written-char + 1 */
13778 +       unsigned long logged_chars;     /* #chars since last read+clear operation */
13779 +
13780 +       char log_buf[1024];
13781 +};
13782 +
13783 +
13784 +/* context sub struct */
13785 +
13786 +struct _vx_cvirt {
13787 +       atomic_t nr_threads;            /* number of current threads */
13788 +       atomic_t nr_running;            /* number of running threads */
13789 +       atomic_t nr_uninterruptible;    /* number of uninterruptible threads */
13790 +
13791 +       atomic_t nr_onhold;             /* processes on hold */
13792 +       uint32_t onhold_last;           /* jiffies when put on hold */
13793 +
13794 +       struct timespec bias_ts;        /* time offset to the host */
13795 +       struct timespec bias_idle;
13796 +       struct timespec bias_uptime;    /* context creation point */
13797 +       uint64_t bias_clock;            /* offset in clock_t */
13798 +
13799 +       spinlock_t load_lock;           /* lock for the load averages */
13800 +       atomic_t load_updates;          /* nr of load updates done so far */
13801 +       uint32_t load_last;             /* last time load was calculated */
13802 +       uint32_t load[3];               /* load averages 1,5,15 */
13803 +
13804 +       atomic_t total_forks;           /* number of forks so far */
13805 +
13806 +       struct _vx_syslog syslog;
13807 +};
13808 +
13809 +struct _vx_cvirt_pc {
13810 +       struct _vx_usage_stat cpustat;
13811 +};
13812 +
13813 +
13814 +#ifdef CONFIG_VSERVER_DEBUG
13815 +
13816 +static inline void __dump_vx_cvirt(struct _vx_cvirt *cvirt)
13817 +{
13818 +       printk("\t_vx_cvirt:\n");
13819 +       printk("\t threads: %4d, %4d, %4d, %4d\n",
13820 +               atomic_read(&cvirt->nr_threads),
13821 +               atomic_read(&cvirt->nr_running),
13822 +               atomic_read(&cvirt->nr_uninterruptible),
13823 +               atomic_read(&cvirt->nr_onhold));
13824 +       /* add rest here */
13825 +       printk("\t total_forks = %d\n", atomic_read(&cvirt->total_forks));
13826 +}
13827 +
13828 +#endif
13829 +
13830 +#endif /* _VX_CVIRT_DEF_H */
13831 diff -NurpP --minimal linux-3.0/include/linux/vserver/debug.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/debug.h
13832 --- linux-3.0/include/linux/vserver/debug.h     1970-01-01 01:00:00.000000000 +0100
13833 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/debug.h        2011-06-10 22:11:24.000000000 +0200
13834 @@ -0,0 +1,140 @@
13835 +#ifndef _VX_DEBUG_H
13836 +#define _VX_DEBUG_H
13837 +
13838 +
13839 +#define VXD_CBIT(n, m) (vx_debug_ ## n & (1 << (m)))
13840 +#define VXD_CMIN(n, m) (vx_debug_ ## n > (m))
13841 +#define VXD_MASK(n, m) (vx_debug_ ## n & (m))
13842 +
13843 +#define VXD_DEV(d)     (d), (d)->bd_inode->i_ino,              \
13844 +                       imajor((d)->bd_inode), iminor((d)->bd_inode)
13845 +#define VXF_DEV                "%p[%lu,%d:%d]"
13846 +
13847 +#if    defined(CONFIG_QUOTES_UTF8)
13848 +#define        VS_Q_LQM        "\xc2\xbb"
13849 +#define        VS_Q_RQM        "\xc2\xab"
13850 +#elif  defined(CONFIG_QUOTES_ASCII)
13851 +#define        VS_Q_LQM        "\x27"
13852 +#define        VS_Q_RQM        "\x27"
13853 +#else
13854 +#define        VS_Q_LQM        "\xbb"
13855 +#define        VS_Q_RQM        "\xab"
13856 +#endif
13857 +
13858 +#define        VS_Q(f)         VS_Q_LQM f VS_Q_RQM
13859 +
13860 +
13861 +#define vxd_path(p)                                            \
13862 +       ({ static char _buffer[PATH_MAX];                       \
13863 +          d_path(p, _buffer, sizeof(_buffer)); })
13864 +
13865 +#define vxd_cond_path(n)                                       \
13866 +       ((n) ? vxd_path(&(n)->path) : "<null>" )
13867 +
13868 +
13869 +#ifdef CONFIG_VSERVER_DEBUG
13870 +
13871 +extern unsigned int vx_debug_switch;
13872 +extern unsigned int vx_debug_xid;
13873 +extern unsigned int vx_debug_nid;
13874 +extern unsigned int vx_debug_tag;
13875 +extern unsigned int vx_debug_net;
13876 +extern unsigned int vx_debug_limit;
13877 +extern unsigned int vx_debug_cres;
13878 +extern unsigned int vx_debug_dlim;
13879 +extern unsigned int vx_debug_quota;
13880 +extern unsigned int vx_debug_cvirt;
13881 +extern unsigned int vx_debug_space;
13882 +extern unsigned int vx_debug_misc;
13883 +
13884 +
13885 +#define VX_LOGLEVEL    "vxD: "
13886 +#define VX_PROC_FMT    "%p: "
13887 +#define VX_PROCESS     current
13888 +
13889 +#define vxdprintk(c, f, x...)                                  \
13890 +       do {                                                    \
13891 +               if (c)                                          \
13892 +                       printk(VX_LOGLEVEL VX_PROC_FMT f "\n",  \
13893 +                               VX_PROCESS , ##x);              \
13894 +       } while (0)
13895 +
13896 +#define vxlprintk(c, f, x...)                                  \
13897 +       do {                                                    \
13898 +               if (c)                                          \
13899 +                       printk(VX_LOGLEVEL f " @%s:%d\n", x);   \
13900 +       } while (0)
13901 +
13902 +#define vxfprintk(c, f, x...)                                  \
13903 +       do {                                                    \
13904 +               if (c)                                          \
13905 +                       printk(VX_LOGLEVEL f " %s@%s:%d\n", x); \
13906 +       } while (0)
13907 +
13908 +
13909 +struct vx_info;
13910 +
13911 +void dump_vx_info(struct vx_info *, int);
13912 +void dump_vx_info_inactive(int);
13913 +
13914 +#else  /* CONFIG_VSERVER_DEBUG */
13915 +
13916 +#define vx_debug_switch 0
13917 +#define vx_debug_xid   0
13918 +#define vx_debug_nid   0
13919 +#define vx_debug_tag   0
13920 +#define vx_debug_net   0
13921 +#define vx_debug_limit 0
13922 +#define vx_debug_cres  0
13923 +#define vx_debug_dlim  0
13924 +#define vx_debug_cvirt 0
13925 +
13926 +#define vxdprintk(x...) do { } while (0)
13927 +#define vxlprintk(x...) do { } while (0)
13928 +#define vxfprintk(x...) do { } while (0)
13929 +
13930 +#endif /* CONFIG_VSERVER_DEBUG */
13931 +
13932 +
13933 +#ifdef CONFIG_VSERVER_WARN
13934 +
13935 +#define VX_WARNLEVEL   KERN_WARNING "vxW: "
13936 +#define VX_WARN_TASK   "[" VS_Q("%s") ",%u:#%u|%u|%u] "
13937 +#define VX_WARN_XID    "[xid #%u] "
13938 +#define VX_WARN_NID    "[nid #%u] "
13939 +#define VX_WARN_TAG    "[tag #%u] "
13940 +
13941 +#define vxwprintk(c, f, x...)                                  \
13942 +       do {                                                    \
13943 +               if (c)                                          \
13944 +                       printk(VX_WARNLEVEL f "\n", ##x);       \
13945 +       } while (0)
13946 +
13947 +#else  /* CONFIG_VSERVER_WARN */
13948 +
13949 +#define vxwprintk(x...) do { } while (0)
13950 +
13951 +#endif /* CONFIG_VSERVER_WARN */
13952 +
13953 +#define vxwprintk_task(c, f, x...)                             \
13954 +       vxwprintk(c, VX_WARN_TASK f,                            \
13955 +               current->comm, current->pid,                    \
13956 +               current->xid, current->nid, current->tag, ##x)
13957 +#define vxwprintk_xid(c, f, x...)                              \
13958 +       vxwprintk(c, VX_WARN_XID f, current->xid, x)
13959 +#define vxwprintk_nid(c, f, x...)                              \
13960 +       vxwprintk(c, VX_WARN_NID f, current->nid, x)
13961 +#define vxwprintk_tag(c, f, x...)                              \
13962 +       vxwprintk(c, VX_WARN_TAG f, current->tag, x)
13963 +
13964 +#ifdef CONFIG_VSERVER_DEBUG
13965 +#define vxd_assert_lock(l)     assert_spin_locked(l)
13966 +#define vxd_assert(c, f, x...) vxlprintk(!(c), \
13967 +       "assertion [" f "] failed.", ##x, __FILE__, __LINE__)
13968 +#else
13969 +#define vxd_assert_lock(l)     do { } while (0)
13970 +#define vxd_assert(c, f, x...) do { } while (0)
13971 +#endif
13972 +
13973 +
13974 +#endif /* _VX_DEBUG_H */
13975 diff -NurpP --minimal linux-3.0/include/linux/vserver/debug_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/debug_cmd.h
13976 --- linux-3.0/include/linux/vserver/debug_cmd.h 1970-01-01 01:00:00.000000000 +0100
13977 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/debug_cmd.h    2011-06-10 22:11:24.000000000 +0200
13978 @@ -0,0 +1,58 @@
13979 +#ifndef _VX_DEBUG_CMD_H
13980 +#define _VX_DEBUG_CMD_H
13981 +
13982 +
13983 +/* debug commands */
13984 +
13985 +#define VCMD_dump_history      VC_CMD(DEBUG, 1, 0)
13986 +
13987 +#define VCMD_read_history      VC_CMD(DEBUG, 5, 0)
13988 +#define VCMD_read_monitor      VC_CMD(DEBUG, 6, 0)
13989 +
13990 +struct  vcmd_read_history_v0 {
13991 +       uint32_t index;
13992 +       uint32_t count;
13993 +       char __user *data;
13994 +};
13995 +
13996 +struct  vcmd_read_monitor_v0 {
13997 +       uint32_t index;
13998 +       uint32_t count;
13999 +       char __user *data;
14000 +};
14001 +
14002 +
14003 +#ifdef __KERNEL__
14004 +
14005 +#ifdef CONFIG_COMPAT
14006 +
14007 +#include <asm/compat.h>
14008 +
14009 +struct vcmd_read_history_v0_x32 {
14010 +       uint32_t index;
14011 +       uint32_t count;
14012 +       compat_uptr_t data_ptr;
14013 +};
14014 +
14015 +struct vcmd_read_monitor_v0_x32 {
14016 +       uint32_t index;
14017 +       uint32_t count;
14018 +       compat_uptr_t data_ptr;
14019 +};
14020 +
14021 +#endif  /* CONFIG_COMPAT */
14022 +
14023 +extern int vc_dump_history(uint32_t);
14024 +
14025 +extern int vc_read_history(uint32_t, void __user *);
14026 +extern int vc_read_monitor(uint32_t, void __user *);
14027 +
14028 +#ifdef CONFIG_COMPAT
14029 +
14030 +extern int vc_read_history_x32(uint32_t, void __user *);
14031 +extern int vc_read_monitor_x32(uint32_t, void __user *);
14032 +
14033 +#endif  /* CONFIG_COMPAT */
14034 +
14035 +#endif /* __KERNEL__ */
14036 +#endif /* _VX_DEBUG_CMD_H */
14037 diff -NurpP --minimal linux-3.0/include/linux/vserver/device.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/device.h
14038 --- linux-3.0/include/linux/vserver/device.h    1970-01-01 01:00:00.000000000 +0100
14039 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/device.h       2011-06-10 22:11:24.000000000 +0200
14040 @@ -0,0 +1,15 @@
14041 +#ifndef _VX_DEVICE_H
14042 +#define _VX_DEVICE_H
14043 +
14044 +
14045 +#define DATTR_CREATE   0x00000001
14046 +#define DATTR_OPEN     0x00000002
14047 +
14048 +#define DATTR_REMAP    0x00000010
14049 +
14050 +#define DATTR_MASK     0x00000013
14051 +
14052 +
14053 +#else  /* _VX_DEVICE_H */
14054 +#warning duplicate inclusion
14055 +#endif /* _VX_DEVICE_H */
14056 diff -NurpP --minimal linux-3.0/include/linux/vserver/device_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/device_cmd.h
14057 --- linux-3.0/include/linux/vserver/device_cmd.h        1970-01-01 01:00:00.000000000 +0100
14058 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/device_cmd.h   2011-06-10 22:11:24.000000000 +0200
14059 @@ -0,0 +1,44 @@
14060 +#ifndef _VX_DEVICE_CMD_H
14061 +#define _VX_DEVICE_CMD_H
14062 +
14063 +
14064 +/*  device vserver commands */
14065 +
14066 +#define VCMD_set_mapping       VC_CMD(DEVICE, 1, 0)
14067 +#define VCMD_unset_mapping     VC_CMD(DEVICE, 2, 0)
14068 +
14069 +struct vcmd_set_mapping_v0 {
14070 +       const char __user *device;
14071 +       const char __user *target;
14072 +       uint32_t flags;
14073 +};
14074 +
14075 +
14076 +#ifdef __KERNEL__
14077 +
14078 +#ifdef CONFIG_COMPAT
14079 +
14080 +#include <asm/compat.h>
14081 +
14082 +struct vcmd_set_mapping_v0_x32 {
14083 +       compat_uptr_t device_ptr;
14084 +       compat_uptr_t target_ptr;
14085 +       uint32_t flags;
14086 +};
14087 +
14088 +#endif /* CONFIG_COMPAT */
14089 +
14090 +#include <linux/compiler.h>
14091 +
14092 +extern int vc_set_mapping(struct vx_info *, void __user *);
14093 +extern int vc_unset_mapping(struct vx_info *, void __user *);
14094 +
14095 +#ifdef CONFIG_COMPAT
14096 +
14097 +extern int vc_set_mapping_x32(struct vx_info *, void __user *);
14098 +extern int vc_unset_mapping_x32(struct vx_info *, void __user *);
14099 +
14100 +#endif /* CONFIG_COMPAT */
14101 +
14102 +#endif /* __KERNEL__ */
14103 +#endif /* _VX_DEVICE_CMD_H */
14104 diff -NurpP --minimal linux-3.0/include/linux/vserver/device_def.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/device_def.h
14105 --- linux-3.0/include/linux/vserver/device_def.h        1970-01-01 01:00:00.000000000 +0100
14106 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/device_def.h   2011-06-10 22:11:24.000000000 +0200
14107 @@ -0,0 +1,17 @@
14108 +#ifndef _VX_DEVICE_DEF_H
14109 +#define _VX_DEVICE_DEF_H
14110 +
14111 +#include <linux/types.h>
14112 +
14113 +struct vx_dmap_target {
14114 +       dev_t target;
14115 +       uint32_t flags;
14116 +};
14117 +
14118 +struct _vx_device {
14119 +#ifdef CONFIG_VSERVER_DEVICE
14120 +       struct vx_dmap_target targets[2];
14121 +#endif
14122 +};
14123 +
14124 +#endif /* _VX_DEVICE_DEF_H */
14125 diff -NurpP --minimal linux-3.0/include/linux/vserver/dlimit.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/dlimit.h
14126 --- linux-3.0/include/linux/vserver/dlimit.h    1970-01-01 01:00:00.000000000 +0100
14127 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/dlimit.h       2011-06-10 22:11:24.000000000 +0200
14128 @@ -0,0 +1,54 @@
14129 +#ifndef _VX_DLIMIT_H
14130 +#define _VX_DLIMIT_H
14131 +
14132 +#include "switch.h"
14133 +
14134 +
14135 +#ifdef __KERNEL__
14136 +
14137 +/*      keep in sync with CDLIM_INFINITY       */
14138 +
14139 +#define DLIM_INFINITY          (~0ULL)
14140 +
14141 +#include <linux/spinlock.h>
14142 +#include <linux/rcupdate.h>
14143 +
14144 +struct super_block;
14145 +
14146 +struct dl_info {
14147 +       struct hlist_node dl_hlist;             /* linked list of contexts */
14148 +       struct rcu_head dl_rcu;                 /* the rcu head */
14149 +       tag_t dl_tag;                           /* context tag */
14150 +       atomic_t dl_usecnt;                     /* usage count */
14151 +       atomic_t dl_refcnt;                     /* reference count */
14152 +
14153 +       struct super_block *dl_sb;              /* associated superblock */
14154 +
14155 +       spinlock_t dl_lock;                     /* protect the values */
14156 +
14157 +       unsigned long long dl_space_used;       /* used space in bytes */
14158 +       unsigned long long dl_space_total;      /* maximum space in bytes */
14159 +       unsigned long dl_inodes_used;           /* used inodes */
14160 +       unsigned long dl_inodes_total;          /* maximum inodes */
14161 +
14162 +       unsigned int dl_nrlmult;                /* non root limit mult */
14163 +};
14164 +
14165 +struct rcu_head;
14166 +
14167 +extern void rcu_free_dl_info(struct rcu_head *);
14168 +extern void unhash_dl_info(struct dl_info *);
14169 +
14170 +extern struct dl_info *locate_dl_info(struct super_block *, tag_t);
14171 +
14172 +
14173 +struct kstatfs;
14174 +
14175 +extern void vx_vsi_statfs(struct super_block *, struct kstatfs *);
14176 +
14177 +typedef uint64_t dlsize_t;
14178 +
14179 +#endif /* __KERNEL__ */
14180 +#else  /* _VX_DLIMIT_H */
14181 +#warning duplicate inclusion
14182 +#endif /* _VX_DLIMIT_H */
14183 diff -NurpP --minimal linux-3.0/include/linux/vserver/dlimit_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/dlimit_cmd.h
14184 --- linux-3.0/include/linux/vserver/dlimit_cmd.h        1970-01-01 01:00:00.000000000 +0100
14185 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/dlimit_cmd.h   2011-06-10 22:11:24.000000000 +0200
14186 @@ -0,0 +1,109 @@
14187 +#ifndef _VX_DLIMIT_CMD_H
14188 +#define _VX_DLIMIT_CMD_H
14189 +
14190 +
14191 +/*  dlimit vserver commands */
14192 +
14193 +#define VCMD_add_dlimit                VC_CMD(DLIMIT, 1, 0)
14194 +#define VCMD_rem_dlimit                VC_CMD(DLIMIT, 2, 0)
14195 +
14196 +#define VCMD_set_dlimit                VC_CMD(DLIMIT, 5, 0)
14197 +#define VCMD_get_dlimit                VC_CMD(DLIMIT, 6, 0)
14198 +
14199 +struct vcmd_ctx_dlimit_base_v0 {
14200 +       const char __user *name;
14201 +       uint32_t flags;
14202 +};
14203 +
14204 +struct vcmd_ctx_dlimit_v0 {
14205 +       const char __user *name;
14206 +       uint32_t space_used;                    /* used space in kbytes */
14207 +       uint32_t space_total;                   /* maximum space in kbytes */
14208 +       uint32_t inodes_used;                   /* used inodes */
14209 +       uint32_t inodes_total;                  /* maximum inodes */
14210 +       uint32_t reserved;                      /* reserved for root in % */
14211 +       uint32_t flags;
14212 +};
14213 +
14214 +#define CDLIM_UNSET            ((uint32_t)0UL)
14215 +#define CDLIM_INFINITY         ((uint32_t)~0UL)
14216 +#define CDLIM_KEEP             ((uint32_t)~1UL)
14217 +
14218 +#define DLIME_UNIT     0
14219 +#define DLIME_KILO     1
14220 +#define DLIME_MEGA     2
14221 +#define DLIME_GIGA     3
14222 +
14223 +#define DLIMF_SHIFT    0x10
14224 +
14225 +#define DLIMS_USED     0
14226 +#define DLIMS_TOTAL    2
14227 +
14228 +static inline
14229 +uint64_t dlimit_space_32to64(uint32_t val, uint32_t flags, int shift)
14230 +{
14231 +       int exp = (flags & DLIMF_SHIFT) ?
14232 +               (flags >> shift) & DLIME_GIGA : DLIME_KILO;
14233 +       return ((uint64_t)val) << (10 * exp);
14234 +}
14235 +
14236 +static inline
14237 +uint32_t dlimit_space_64to32(uint64_t val, uint32_t *flags, int shift)
14238 +{
14239 +       int exp = 0;
14240 +
14241 +       if (*flags & DLIMF_SHIFT) {
14242 +               while (val > (1LL << 32) && (exp < 3)) {
14243 +                       val >>= 10;
14244 +                       exp++;
14245 +               }
14246 +               *flags &= ~(DLIME_GIGA << shift);
14247 +               *flags |= exp << shift;
14248 +       } else
14249 +               val >>= 10;
14250 +       return val;
14251 +}
14252 +
14253 +#ifdef __KERNEL__
14254 +
14255 +#ifdef CONFIG_COMPAT
14256 +
14257 +#include <asm/compat.h>
14258 +
14259 +struct vcmd_ctx_dlimit_base_v0_x32 {
14260 +       compat_uptr_t name_ptr;
14261 +       uint32_t flags;
14262 +};
14263 +
14264 +struct vcmd_ctx_dlimit_v0_x32 {
14265 +       compat_uptr_t name_ptr;
14266 +       uint32_t space_used;                    /* used space in kbytes */
14267 +       uint32_t space_total;                   /* maximum space in kbytes */
14268 +       uint32_t inodes_used;                   /* used inodes */
14269 +       uint32_t inodes_total;                  /* maximum inodes */
14270 +       uint32_t reserved;                      /* reserved for root in % */
14271 +       uint32_t flags;
14272 +};
14273 +
14274 +#endif /* CONFIG_COMPAT */
14275 +
14276 +#include <linux/compiler.h>
14277 +
14278 +extern int vc_add_dlimit(uint32_t, void __user *);
14279 +extern int vc_rem_dlimit(uint32_t, void __user *);
14280 +
14281 +extern int vc_set_dlimit(uint32_t, void __user *);
14282 +extern int vc_get_dlimit(uint32_t, void __user *);
14283 +
14284 +#ifdef CONFIG_COMPAT
14285 +
14286 +extern int vc_add_dlimit_x32(uint32_t, void __user *);
14287 +extern int vc_rem_dlimit_x32(uint32_t, void __user *);
14288 +
14289 +extern int vc_set_dlimit_x32(uint32_t, void __user *);
14290 +extern int vc_get_dlimit_x32(uint32_t, void __user *);
14291 +
14292 +#endif /* CONFIG_COMPAT */
14293 +
14294 +#endif /* __KERNEL__ */
14295 +#endif /* _VX_DLIMIT_CMD_H */
14296 diff -NurpP --minimal linux-3.0/include/linux/vserver/global.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/global.h
14297 --- linux-3.0/include/linux/vserver/global.h    1970-01-01 01:00:00.000000000 +0100
14298 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/global.h       2011-06-10 22:11:24.000000000 +0200
14299 @@ -0,0 +1,19 @@
14300 +#ifndef _VX_GLOBAL_H
14301 +#define _VX_GLOBAL_H
14302 +
14303 +
14304 +extern atomic_t vx_global_ctotal;
14305 +extern atomic_t vx_global_cactive;
14306 +
14307 +extern atomic_t nx_global_ctotal;
14308 +extern atomic_t nx_global_cactive;
14309 +
14310 +extern atomic_t vs_global_nsproxy;
14311 +extern atomic_t vs_global_fs;
14312 +extern atomic_t vs_global_mnt_ns;
14313 +extern atomic_t vs_global_uts_ns;
14314 +extern atomic_t vs_global_user_ns;
14315 +extern atomic_t vs_global_pid_ns;
14316 +
14317 +
14318 +#endif /* _VX_GLOBAL_H */
14319 diff -NurpP --minimal linux-3.0/include/linux/vserver/history.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/history.h
14320 --- linux-3.0/include/linux/vserver/history.h   1970-01-01 01:00:00.000000000 +0100
14321 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/history.h      2011-06-10 22:11:24.000000000 +0200
14322 @@ -0,0 +1,197 @@
14323 +#ifndef _VX_HISTORY_H
14324 +#define _VX_HISTORY_H
14325 +
14326 +
14327 +enum {
14328 +       VXH_UNUSED = 0,
14329 +       VXH_THROW_OOPS = 1,
14330 +
14331 +       VXH_GET_VX_INFO,
14332 +       VXH_PUT_VX_INFO,
14333 +       VXH_INIT_VX_INFO,
14334 +       VXH_SET_VX_INFO,
14335 +       VXH_CLR_VX_INFO,
14336 +       VXH_CLAIM_VX_INFO,
14337 +       VXH_RELEASE_VX_INFO,
14338 +       VXH_ALLOC_VX_INFO,
14339 +       VXH_DEALLOC_VX_INFO,
14340 +       VXH_HASH_VX_INFO,
14341 +       VXH_UNHASH_VX_INFO,
14342 +       VXH_LOC_VX_INFO,
14343 +       VXH_LOOKUP_VX_INFO,
14344 +       VXH_CREATE_VX_INFO,
14345 +};
14346 +
14347 +struct _vxhe_vxi {
14348 +       struct vx_info *ptr;
14349 +       unsigned xid;
14350 +       unsigned usecnt;
14351 +       unsigned tasks;
14352 +};
14353 +
14354 +struct _vxhe_set_clr {
14355 +       void *data;
14356 +};
14357 +
14358 +struct _vxhe_loc_lookup {
14359 +       unsigned arg;
14360 +};
14361 +
14362 +struct _vx_hist_entry {
14363 +       void *loc;
14364 +       unsigned short seq;
14365 +       unsigned short type;
14366 +       struct _vxhe_vxi vxi;
14367 +       union {
14368 +               struct _vxhe_set_clr sc;
14369 +               struct _vxhe_loc_lookup ll;
14370 +       };
14371 +};
14372 +
14373 +#ifdef CONFIG_VSERVER_HISTORY
14374 +
14375 +extern unsigned volatile int vxh_active;
14376 +
14377 +struct _vx_hist_entry *vxh_advance(void *loc);
14378 +
14379 +
14380 +static inline
14381 +void   __vxh_copy_vxi(struct _vx_hist_entry *entry, struct vx_info *vxi)
14382 +{
14383 +       entry->vxi.ptr = vxi;
14384 +       if (vxi) {
14385 +               entry->vxi.usecnt = atomic_read(&vxi->vx_usecnt);
14386 +               entry->vxi.tasks = atomic_read(&vxi->vx_tasks);
14387 +               entry->vxi.xid = vxi->vx_id;
14388 +       }
14389 +}
14390 +
14391 +
14392 +#define        __HERE__ current_text_addr()
14393 +
14394 +#define __VXH_BODY(__type, __data, __here)     \
14395 +       struct _vx_hist_entry *entry;           \
14396 +                                               \
14397 +       preempt_disable();                      \
14398 +       entry = vxh_advance(__here);            \
14399 +       __data;                                 \
14400 +       entry->type = __type;                   \
14401 +       preempt_enable();
14402 +
14403 +
14404 +       /* pass vxi only */
14405 +
14406 +#define __VXH_SMPL                             \
14407 +       __vxh_copy_vxi(entry, vxi)
14408 +
14409 +static inline
14410 +void   __vxh_smpl(struct vx_info *vxi, int __type, void *__here)
14411 +{
14412 +       __VXH_BODY(__type, __VXH_SMPL, __here)
14413 +}
14414 +
14415 +       /* pass vxi and data (void *) */
14416 +
14417 +#define __VXH_DATA                             \
14418 +       __vxh_copy_vxi(entry, vxi);             \
14419 +       entry->sc.data = data
14420 +
14421 +static inline
14422 +void   __vxh_data(struct vx_info *vxi, void *data,
14423 +                       int __type, void *__here)
14424 +{
14425 +       __VXH_BODY(__type, __VXH_DATA, __here)
14426 +}
14427 +
14428 +       /* pass vxi and arg (long) */
14429 +
14430 +#define __VXH_LONG                             \
14431 +       __vxh_copy_vxi(entry, vxi);             \
14432 +       entry->ll.arg = arg
14433 +
14434 +static inline
14435 +void   __vxh_long(struct vx_info *vxi, long arg,
14436 +                       int __type, void *__here)
14437 +{
14438 +       __VXH_BODY(__type, __VXH_LONG, __here)
14439 +}
14440 +
14441 +
14442 +static inline
14443 +void   __vxh_throw_oops(void *__here)
14444 +{
14445 +       __VXH_BODY(VXH_THROW_OOPS, {}, __here);
14446 +       /* prevent further acquisition */
14447 +       vxh_active = 0;
14448 +}
14449 +
14450 +
14451 +#define vxh_throw_oops()       __vxh_throw_oops(__HERE__);
14452 +
14453 +#define __vxh_get_vx_info(v, h)        __vxh_smpl(v, VXH_GET_VX_INFO, h);
14454 +#define __vxh_put_vx_info(v, h)        __vxh_smpl(v, VXH_PUT_VX_INFO, h);
14455 +
14456 +#define __vxh_init_vx_info(v, d, h) \
14457 +       __vxh_data(v, d, VXH_INIT_VX_INFO, h);
14458 +#define __vxh_set_vx_info(v, d, h) \
14459 +       __vxh_data(v, d, VXH_SET_VX_INFO, h);
14460 +#define __vxh_clr_vx_info(v, d, h) \
14461 +       __vxh_data(v, d, VXH_CLR_VX_INFO, h);
14462 +
14463 +#define __vxh_claim_vx_info(v, d, h) \
14464 +       __vxh_data(v, d, VXH_CLAIM_VX_INFO, h);
14465 +#define __vxh_release_vx_info(v, d, h) \
14466 +       __vxh_data(v, d, VXH_RELEASE_VX_INFO, h);
14467 +
14468 +#define vxh_alloc_vx_info(v) \
14469 +       __vxh_smpl(v, VXH_ALLOC_VX_INFO, __HERE__);
14470 +#define vxh_dealloc_vx_info(v) \
14471 +       __vxh_smpl(v, VXH_DEALLOC_VX_INFO, __HERE__);
14472 +
14473 +#define vxh_hash_vx_info(v) \
14474 +       __vxh_smpl(v, VXH_HASH_VX_INFO, __HERE__);
14475 +#define vxh_unhash_vx_info(v) \
14476 +       __vxh_smpl(v, VXH_UNHASH_VX_INFO, __HERE__);
14477 +
14478 +#define vxh_loc_vx_info(v, l) \
14479 +       __vxh_long(v, l, VXH_LOC_VX_INFO, __HERE__);
14480 +#define vxh_lookup_vx_info(v, l) \
14481 +       __vxh_long(v, l, VXH_LOOKUP_VX_INFO, __HERE__);
14482 +#define vxh_create_vx_info(v, l) \
14483 +       __vxh_long(v, l, VXH_CREATE_VX_INFO, __HERE__);
14484 +
14485 +extern void vxh_dump_history(void);
14486 +
14487 +
14488 +#else  /* CONFIG_VSERVER_HISTORY */
14489 +
14490 +#define        __HERE__        0
14491 +
14492 +#define vxh_throw_oops()               do { } while (0)
14493 +
14494 +#define __vxh_get_vx_info(v, h)                do { } while (0)
14495 +#define __vxh_put_vx_info(v, h)                do { } while (0)
14496 +
14497 +#define __vxh_init_vx_info(v, d, h)    do { } while (0)
14498 +#define __vxh_set_vx_info(v, d, h)     do { } while (0)
14499 +#define __vxh_clr_vx_info(v, d, h)     do { } while (0)
14500 +
14501 +#define __vxh_claim_vx_info(v, d, h)   do { } while (0)
14502 +#define __vxh_release_vx_info(v, d, h) do { } while (0)
14503 +
14504 +#define vxh_alloc_vx_info(v)           do { } while (0)
14505 +#define vxh_dealloc_vx_info(v)         do { } while (0)
14506 +
14507 +#define vxh_hash_vx_info(v)            do { } while (0)
14508 +#define vxh_unhash_vx_info(v)          do { } while (0)
14509 +
14510 +#define vxh_loc_vx_info(v, l)          do { } while (0)
14511 +#define vxh_lookup_vx_info(v, l)       do { } while (0)
14512 +#define vxh_create_vx_info(v, l)       do { } while (0)
14513 +
14514 +#define vxh_dump_history()             do { } while (0)
14515 +
14516 +
14517 +#endif /* CONFIG_VSERVER_HISTORY */
14518 +
14519 +#endif /* _VX_HISTORY_H */
14520 diff -NurpP --minimal linux-3.0/include/linux/vserver/inode.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/inode.h
14521 --- linux-3.0/include/linux/vserver/inode.h     1970-01-01 01:00:00.000000000 +0100
14522 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/inode.h        2011-06-10 22:11:24.000000000 +0200
14523 @@ -0,0 +1,39 @@
14524 +#ifndef _VX_INODE_H
14525 +#define _VX_INODE_H
14526 +
14527 +
14528 +#define IATTR_TAG      0x01000000
14529 +
14530 +#define IATTR_ADMIN    0x00000001
14531 +#define IATTR_WATCH    0x00000002
14532 +#define IATTR_HIDE     0x00000004
14533 +#define IATTR_FLAGS    0x00000007
14534 +
14535 +#define IATTR_BARRIER  0x00010000
14536 +#define IATTR_IXUNLINK 0x00020000
14537 +#define IATTR_IMMUTABLE 0x00040000
14538 +#define IATTR_COW      0x00080000
14539 +
14540 +#ifdef __KERNEL__
14541 +
14542 +
14543 +#ifdef CONFIG_VSERVER_PROC_SECURE
14544 +#define IATTR_PROC_DEFAULT     ( IATTR_ADMIN | IATTR_HIDE )
14545 +#define IATTR_PROC_SYMLINK     ( IATTR_ADMIN )
14546 +#else
14547 +#define IATTR_PROC_DEFAULT     ( IATTR_ADMIN )
14548 +#define IATTR_PROC_SYMLINK     ( IATTR_ADMIN )
14549 +#endif
14550 +
14551 +#define vx_hide_check(c, m)    (((m) & IATTR_HIDE) ? vx_check(c, m) : 1)
14552 +
14553 +#endif /* __KERNEL__ */
14554 +
14555 +/* inode ioctls */
14556 +
14557 +#define FIOC_GETXFLG   _IOR('x', 5, long)
14558 +#define FIOC_SETXFLG   _IOW('x', 6, long)
14559 +
14560 +#else  /* _VX_INODE_H */
14561 +#warning duplicate inclusion
14562 +#endif /* _VX_INODE_H */
14563 diff -NurpP --minimal linux-3.0/include/linux/vserver/inode_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/inode_cmd.h
14564 --- linux-3.0/include/linux/vserver/inode_cmd.h 1970-01-01 01:00:00.000000000 +0100
14565 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/inode_cmd.h    2011-06-10 22:11:24.000000000 +0200
14566 @@ -0,0 +1,59 @@
14567 +#ifndef _VX_INODE_CMD_H
14568 +#define _VX_INODE_CMD_H
14569 +
14570 +
14571 +/*  inode vserver commands */
14572 +
14573 +#define VCMD_get_iattr         VC_CMD(INODE, 1, 1)
14574 +#define VCMD_set_iattr         VC_CMD(INODE, 2, 1)
14575 +
14576 +#define VCMD_fget_iattr                VC_CMD(INODE, 3, 0)
14577 +#define VCMD_fset_iattr                VC_CMD(INODE, 4, 0)
14578 +
14579 +struct vcmd_ctx_iattr_v1 {
14580 +       const char __user *name;
14581 +       uint32_t tag;
14582 +       uint32_t flags;
14583 +       uint32_t mask;
14584 +};
14585 +
14586 +struct vcmd_ctx_fiattr_v0 {
14587 +       uint32_t tag;
14588 +       uint32_t flags;
14589 +       uint32_t mask;
14590 +};
14591 +
14592 +
14593 +#ifdef __KERNEL__
14594 +
14595 +
14596 +#ifdef CONFIG_COMPAT
14597 +
14598 +#include <asm/compat.h>
14599 +
14600 +struct vcmd_ctx_iattr_v1_x32 {
14601 +       compat_uptr_t name_ptr;
14602 +       uint32_t tag;
14603 +       uint32_t flags;
14604 +       uint32_t mask;
14605 +};
14606 +
14607 +#endif /* CONFIG_COMPAT */
14608 +
14609 +#include <linux/compiler.h>
14610 +
14611 +extern int vc_get_iattr(void __user *);
14612 +extern int vc_set_iattr(void __user *);
14613 +
14614 +extern int vc_fget_iattr(uint32_t, void __user *);
14615 +extern int vc_fset_iattr(uint32_t, void __user *);
14616 +
14617 +#ifdef CONFIG_COMPAT
14618 +
14619 +extern int vc_get_iattr_x32(void __user *);
14620 +extern int vc_set_iattr_x32(void __user *);
14621 +
14622 +#endif /* CONFIG_COMPAT */
14623 +
14624 +#endif /* __KERNEL__ */
14625 +#endif /* _VX_INODE_CMD_H */
14626 diff -NurpP --minimal linux-3.0/include/linux/vserver/limit.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit.h
14627 --- linux-3.0/include/linux/vserver/limit.h     1970-01-01 01:00:00.000000000 +0100
14628 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit.h        2011-06-10 22:11:24.000000000 +0200
14629 @@ -0,0 +1,71 @@
14630 +#ifndef _VX_LIMIT_H
14631 +#define _VX_LIMIT_H
14632 +
14633 +#define VLIMIT_NSOCK   16
14634 +#define VLIMIT_OPENFD  17
14635 +#define VLIMIT_ANON    18
14636 +#define VLIMIT_SHMEM   19
14637 +#define VLIMIT_SEMARY  20
14638 +#define VLIMIT_NSEMS   21
14639 +#define VLIMIT_DENTRY  22
14640 +#define VLIMIT_MAPPED  23
14641 +
14642 +
14643 +#ifdef __KERNEL__
14644 +
14645 +#define        VLIM_NOCHECK    ((1L << VLIMIT_DENTRY) | (1L << RLIMIT_RSS))
14646 +
14647 +/*     keep in sync with CRLIM_INFINITY */
14648 +
14649 +#define        VLIM_INFINITY   (~0ULL)
14650 +
14651 +#include <asm/atomic.h>
14652 +#include <asm/resource.h>
14653 +
14654 +#ifndef RLIM_INFINITY
14655 +#warning RLIM_INFINITY is undefined
14656 +#endif
14657 +
14658 +#define __rlim_val(l, r, v)    ((l)->res[r].v)
14659 +
14660 +#define __rlim_soft(l, r)      __rlim_val(l, r, soft)
14661 +#define __rlim_hard(l, r)      __rlim_val(l, r, hard)
14662 +
14663 +#define __rlim_rcur(l, r)      __rlim_val(l, r, rcur)
14664 +#define __rlim_rmin(l, r)      __rlim_val(l, r, rmin)
14665 +#define __rlim_rmax(l, r)      __rlim_val(l, r, rmax)
14666 +
14667 +#define __rlim_lhit(l, r)      __rlim_val(l, r, lhit)
14668 +#define __rlim_hit(l, r)       atomic_inc(&__rlim_lhit(l, r))
14669 +
14670 +typedef atomic_long_t rlim_atomic_t;
14671 +typedef unsigned long rlim_t;
14672 +
14673 +#define __rlim_get(l, r)       atomic_long_read(&__rlim_rcur(l, r))
14674 +#define __rlim_set(l, r, v)    atomic_long_set(&__rlim_rcur(l, r), v)
14675 +#define __rlim_inc(l, r)       atomic_long_inc(&__rlim_rcur(l, r))
14676 +#define __rlim_dec(l, r)       atomic_long_dec(&__rlim_rcur(l, r))
14677 +#define __rlim_add(l, r, v)    atomic_long_add(v, &__rlim_rcur(l, r))
14678 +#define __rlim_sub(l, r, v)    atomic_long_sub(v, &__rlim_rcur(l, r))
14679 +
14680 +
14681 +#if    (RLIM_INFINITY == VLIM_INFINITY)
14682 +#define        VX_VLIM(r) ((long long)(long)(r))
14683 +#define        VX_RLIM(v) ((rlim_t)(v))
14684 +#else
14685 +#define        VX_VLIM(r) (((r) == RLIM_INFINITY) \
14686 +               ? VLIM_INFINITY : (long long)(r))
14687 +#define        VX_RLIM(v) (((v) == VLIM_INFINITY) \
14688 +               ? RLIM_INFINITY : (rlim_t)(v))
14689 +#endif
14690 +
14691 +struct sysinfo;
14692 +
14693 +void vx_vsi_meminfo(struct sysinfo *);
14694 +void vx_vsi_swapinfo(struct sysinfo *);
14695 +long vx_vsi_cached(struct sysinfo *);
14696 +
14697 +#define NUM_LIMITS     24
14698 +
14699 +#endif /* __KERNEL__ */
14700 +#endif /* _VX_LIMIT_H */
14701 diff -NurpP --minimal linux-3.0/include/linux/vserver/limit_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit_cmd.h
14702 --- linux-3.0/include/linux/vserver/limit_cmd.h 1970-01-01 01:00:00.000000000 +0100
14703 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit_cmd.h    2011-06-10 22:11:24.000000000 +0200
14704 @@ -0,0 +1,71 @@
14705 +#ifndef _VX_LIMIT_CMD_H
14706 +#define _VX_LIMIT_CMD_H
14707 +
14708 +
14709 +/*  rlimit vserver commands */
14710 +
14711 +#define VCMD_get_rlimit                VC_CMD(RLIMIT, 1, 0)
14712 +#define VCMD_set_rlimit                VC_CMD(RLIMIT, 2, 0)
14713 +#define VCMD_get_rlimit_mask   VC_CMD(RLIMIT, 3, 0)
14714 +#define VCMD_reset_hits                VC_CMD(RLIMIT, 7, 0)
14715 +#define VCMD_reset_minmax      VC_CMD(RLIMIT, 9, 0)
14716 +
14717 +struct vcmd_ctx_rlimit_v0 {
14718 +       uint32_t id;
14719 +       uint64_t minimum;
14720 +       uint64_t softlimit;
14721 +       uint64_t maximum;
14722 +};
14723 +
14724 +struct vcmd_ctx_rlimit_mask_v0 {
14725 +       uint32_t minimum;
14726 +       uint32_t softlimit;
14727 +       uint32_t maximum;
14728 +};
14729 +
14730 +#define VCMD_rlimit_stat       VC_CMD(VSTAT, 1, 0)
14731 +
14732 +struct vcmd_rlimit_stat_v0 {
14733 +       uint32_t id;
14734 +       uint32_t hits;
14735 +       uint64_t value;
14736 +       uint64_t minimum;
14737 +       uint64_t maximum;
14738 +};
14739 +
14740 +#define CRLIM_UNSET            (0ULL)
14741 +#define CRLIM_INFINITY         (~0ULL)
14742 +#define CRLIM_KEEP             (~1ULL)
14743 +
14744 +#ifdef __KERNEL__
14745 +
14746 +#ifdef CONFIG_IA32_EMULATION
14747 +
14748 +struct vcmd_ctx_rlimit_v0_x32 {
14749 +       uint32_t id;
14750 +       uint64_t minimum;
14751 +       uint64_t softlimit;
14752 +       uint64_t maximum;
14753 +} __attribute__ ((packed));
14754 +
14755 +#endif /* CONFIG_IA32_EMULATION */
14756 +
14757 +#include <linux/compiler.h>
14758 +
14759 +extern int vc_get_rlimit_mask(uint32_t, void __user *);
14760 +extern int vc_get_rlimit(struct vx_info *, void __user *);
14761 +extern int vc_set_rlimit(struct vx_info *, void __user *);
14762 +extern int vc_reset_hits(struct vx_info *, void __user *);
14763 +extern int vc_reset_minmax(struct vx_info *, void __user *);
14764 +
14765 +extern int vc_rlimit_stat(struct vx_info *, void __user *);
14766 +
14767 +#ifdef CONFIG_IA32_EMULATION
14768 +
14769 +extern int vc_get_rlimit_x32(struct vx_info *, void __user *);
14770 +extern int vc_set_rlimit_x32(struct vx_info *, void __user *);
14771 +
14772 +#endif /* CONFIG_IA32_EMULATION */
14773 +
14774 +#endif /* __KERNEL__ */
14775 +#endif /* _VX_LIMIT_CMD_H */
14776 diff -NurpP --minimal linux-3.0/include/linux/vserver/limit_def.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit_def.h
14777 --- linux-3.0/include/linux/vserver/limit_def.h 1970-01-01 01:00:00.000000000 +0100
14778 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit_def.h    2011-06-10 22:11:24.000000000 +0200
14779 @@ -0,0 +1,47 @@
14780 +#ifndef _VX_LIMIT_DEF_H
14781 +#define _VX_LIMIT_DEF_H
14782 +
14783 +#include <asm/atomic.h>
14784 +#include <asm/resource.h>
14785 +
14786 +#include "limit.h"
14787 +
14788 +
14789 +struct _vx_res_limit {
14790 +       rlim_t soft;            /* Context soft limit */
14791 +       rlim_t hard;            /* Context hard limit */
14792 +
14793 +       rlim_atomic_t rcur;     /* Current value */
14794 +       rlim_t rmin;            /* Context minimum */
14795 +       rlim_t rmax;            /* Context maximum */
14796 +
14797 +       atomic_t lhit;          /* Limit hits */
14798 +};
14799 +
14800 +/* context sub struct */
14801 +
14802 +struct _vx_limit {
14803 +       struct _vx_res_limit res[NUM_LIMITS];
14804 +};
14805 +
14806 +#ifdef CONFIG_VSERVER_DEBUG
14807 +
14808 +static inline void __dump_vx_limit(struct _vx_limit *limit)
14809 +{
14810 +       int i;
14811 +
14812 +       printk("\t_vx_limit:");
14813 +       for (i = 0; i < NUM_LIMITS; i++) {
14814 +               printk("\t [%2d] = %8lu %8lu/%8lu, %8ld/%8ld, %8d\n",
14815 +                       i, (unsigned long)__rlim_get(limit, i),
14816 +                       (unsigned long)__rlim_rmin(limit, i),
14817 +                       (unsigned long)__rlim_rmax(limit, i),
14818 +                       (long)__rlim_soft(limit, i),
14819 +                       (long)__rlim_hard(limit, i),
14820 +                       atomic_read(&__rlim_lhit(limit, i)));
14821 +       }
14822 +}
14823 +
14824 +#endif
14825 +
14826 +#endif /* _VX_LIMIT_DEF_H */
14827 diff -NurpP --minimal linux-3.0/include/linux/vserver/limit_int.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit_int.h
14828 --- linux-3.0/include/linux/vserver/limit_int.h 1970-01-01 01:00:00.000000000 +0100
14829 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/limit_int.h    2011-06-10 22:11:24.000000000 +0200
14830 @@ -0,0 +1,198 @@
14831 +#ifndef _VX_LIMIT_INT_H
14832 +#define _VX_LIMIT_INT_H
14833 +
14834 +#include "context.h"
14835 +
14836 +#ifdef __KERNEL__
14837 +
14838 +#define VXD_RCRES_COND(r)      VXD_CBIT(cres, r)
14839 +#define VXD_RLIMIT_COND(r)     VXD_CBIT(limit, r)
14840 +
14841 +extern const char *vlimit_name[NUM_LIMITS];
14842 +
14843 +static inline void __vx_acc_cres(struct vx_info *vxi,
14844 +       int res, int dir, void *_data, char *_file, int _line)
14845 +{
14846 +       if (VXD_RCRES_COND(res))
14847 +               vxlprintk(1, "vx_acc_cres[%5d,%s,%2d]: %5ld%s (%p)",
14848 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
14849 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
14850 +                       (dir > 0) ? "++" : "--", _data, _file, _line);
14851 +       if (!vxi)
14852 +               return;
14853 +
14854 +       if (dir > 0)
14855 +               __rlim_inc(&vxi->limit, res);
14856 +       else
14857 +               __rlim_dec(&vxi->limit, res);
14858 +}
14859 +
14860 +static inline void __vx_add_cres(struct vx_info *vxi,
14861 +       int res, int amount, void *_data, char *_file, int _line)
14862 +{
14863 +       if (VXD_RCRES_COND(res))
14864 +               vxlprintk(1, "vx_add_cres[%5d,%s,%2d]: %5ld += %5d (%p)",
14865 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
14866 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
14867 +                       amount, _data, _file, _line);
14868 +       if (amount == 0)
14869 +               return;
14870 +       if (!vxi)
14871 +               return;
14872 +       __rlim_add(&vxi->limit, res, amount);
14873 +}
14874 +
14875 +static inline
14876 +int __vx_cres_adjust_max(struct _vx_limit *limit, int res, rlim_t value)
14877 +{
14878 +       int cond = (value > __rlim_rmax(limit, res));
14879 +
14880 +       if (cond)
14881 +               __rlim_rmax(limit, res) = value;
14882 +       return cond;
14883 +}
14884 +
14885 +static inline
14886 +int __vx_cres_adjust_min(struct _vx_limit *limit, int res, rlim_t value)
14887 +{
14888 +       int cond = (value < __rlim_rmin(limit, res));
14889 +
14890 +       if (cond)
14891 +               __rlim_rmin(limit, res) = value;
14892 +       return cond;
14893 +}
14894 +
14895 +static inline
14896 +void __vx_cres_fixup(struct _vx_limit *limit, int res, rlim_t value)
14897 +{
14898 +       if (!__vx_cres_adjust_max(limit, res, value))
14899 +               __vx_cres_adjust_min(limit, res, value);
14900 +}
14901 +
14902 +
14903 +/*     return values:
14904 +        +1 ... no limit hit
14905 +        -1 ... over soft limit
14906 +         0 ... over hard limit         */
14907 +
14908 +static inline int __vx_cres_avail(struct vx_info *vxi,
14909 +       int res, int num, char *_file, int _line)
14910 +{
14911 +       struct _vx_limit *limit;
14912 +       rlim_t value;
14913 +
14914 +       if (VXD_RLIMIT_COND(res))
14915 +               vxlprintk(1, "vx_cres_avail[%5d,%s,%2d]: %5ld/%5ld > %5ld + %5d",
14916 +                       (vxi ? vxi->vx_id : -1), vlimit_name[res], res,
14917 +                       (vxi ? (long)__rlim_soft(&vxi->limit, res) : -1),
14918 +                       (vxi ? (long)__rlim_hard(&vxi->limit, res) : -1),
14919 +                       (vxi ? (long)__rlim_get(&vxi->limit, res) : 0),
14920 +                       num, _file, _line);
14921 +       if (!vxi)
14922 +               return 1;
14923 +
14924 +       limit = &vxi->limit;
14925 +       value = __rlim_get(limit, res);
14926 +
14927 +       if (!__vx_cres_adjust_max(limit, res, value))
14928 +               __vx_cres_adjust_min(limit, res, value);
14929 +
14930 +       if (num == 0)
14931 +               return 1;
14932 +
14933 +       if (__rlim_soft(limit, res) == RLIM_INFINITY)
14934 +               return -1;
14935 +       if (value + num <= __rlim_soft(limit, res))
14936 +               return -1;
14937 +
14938 +       if (__rlim_hard(limit, res) == RLIM_INFINITY)
14939 +               return 1;
14940 +       if (value + num <= __rlim_hard(limit, res))
14941 +               return 1;
14942 +
14943 +       __rlim_hit(limit, res);
14944 +       return 0;
14945 +}
14946 +
14947 +
14948 +static const int VLA_RSS[] = { RLIMIT_RSS, VLIMIT_ANON, VLIMIT_MAPPED, 0 };
14949 +
14950 +static inline
14951 +rlim_t __vx_cres_array_sum(struct _vx_limit *limit, const int *array)
14952 +{
14953 +       rlim_t value, sum = 0;
14954 +       int res;
14955 +
14956 +       while ((res = *array++)) {
14957 +               value = __rlim_get(limit, res);
14958 +               __vx_cres_fixup(limit, res, value);
14959 +               sum += value;
14960 +       }
14961 +       return sum;
14962 +}
14963 +
14964 +static inline
14965 +rlim_t __vx_cres_array_fixup(struct _vx_limit *limit, const int *array)
14966 +{
14967 +       rlim_t value = __vx_cres_array_sum(limit, array + 1);
14968 +       int res = *array;
14969 +
14970 +       if (value == __rlim_get(limit, res))
14971 +               return value;
14972 +
14973 +       __rlim_set(limit, res, value);
14974 +       /* now adjust min/max */
14975 +       if (!__vx_cres_adjust_max(limit, res, value))
14976 +               __vx_cres_adjust_min(limit, res, value);
14977 +
14978 +       return value;
14979 +}
14980 +
14981 +static inline int __vx_cres_array_avail(struct vx_info *vxi,
14982 +       const int *array, int num, char *_file, int _line)
14983 +{
14984 +       struct _vx_limit *limit;
14985 +       rlim_t value = 0;
14986 +       int res;
14987 +
14988 +       if (num == 0)
14989 +               return 1;
14990 +       if (!vxi)
14991 +               return 1;
14992 +
14993 +       limit = &vxi->limit;
14994 +       res = *array;
14995 +       value = __vx_cres_array_sum(limit, array + 1);
14996 +
14997 +       __rlim_set(limit, res, value);
14998 +       __vx_cres_fixup(limit, res, value);
14999 +
15000 +       return __vx_cres_avail(vxi, res, num, _file, _line);
15001 +}
15002 +
15003 +
15004 +static inline void vx_limit_fixup(struct _vx_limit *limit, int id)
15005 +{
15006 +       rlim_t value;
15007 +       int res;
15008 +
15009 +       /* complex resources first */
15010 +       if ((id < 0) || (id == RLIMIT_RSS))
15011 +               __vx_cres_array_fixup(limit, VLA_RSS);
15012 +
15013 +       for (res = 0; res < NUM_LIMITS; res++) {
15014 +               if ((id > 0) && (res != id))
15015 +                       continue;
15016 +
15017 +               value = __rlim_get(limit, res);
15018 +               __vx_cres_fixup(limit, res, value);
15019 +
15020 +               /* not supposed to happen, maybe warn? */
15021 +               if (__rlim_rmax(limit, res) > __rlim_hard(limit, res))
15022 +                       __rlim_rmax(limit, res) = __rlim_hard(limit, res);
15023 +       }
15024 +}
15025 +
15026 +
15027 +#endif /* __KERNEL__ */
15028 +#endif /* _VX_LIMIT_INT_H */
15029 diff -NurpP --minimal linux-3.0/include/linux/vserver/monitor.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/monitor.h
15030 --- linux-3.0/include/linux/vserver/monitor.h   1970-01-01 01:00:00.000000000 +0100
15031 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/monitor.h      2011-06-10 22:11:24.000000000 +0200
15032 @@ -0,0 +1,96 @@
15033 +#ifndef _VX_MONITOR_H
15034 +#define _VX_MONITOR_H
15035 +
15036 +#include <linux/types.h>
15037 +
15038 +enum {
15039 +       VXM_UNUSED = 0,
15040 +
15041 +       VXM_SYNC = 0x10,
15042 +
15043 +       VXM_UPDATE = 0x20,
15044 +       VXM_UPDATE_1,
15045 +       VXM_UPDATE_2,
15046 +
15047 +       VXM_RQINFO_1 = 0x24,
15048 +       VXM_RQINFO_2,
15049 +
15050 +       VXM_ACTIVATE = 0x40,
15051 +       VXM_DEACTIVATE,
15052 +       VXM_IDLE,
15053 +
15054 +       VXM_HOLD = 0x44,
15055 +       VXM_UNHOLD,
15056 +
15057 +       VXM_MIGRATE = 0x48,
15058 +       VXM_RESCHED,
15059 +
15060 +       /* all other bits are flags */
15061 +       VXM_SCHED = 0x80,
15062 +};
15063 +
15064 +struct _vxm_update_1 {
15065 +       uint32_t tokens_max;
15066 +       uint32_t fill_rate;
15067 +       uint32_t interval;
15068 +};
15069 +
15070 +struct _vxm_update_2 {
15071 +       uint32_t tokens_min;
15072 +       uint32_t fill_rate;
15073 +       uint32_t interval;
15074 +};
15075 +
15076 +struct _vxm_rqinfo_1 {
15077 +       uint16_t running;
15078 +       uint16_t onhold;
15079 +       uint16_t iowait;
15080 +       uint16_t uintr;
15081 +       uint32_t idle_tokens;
15082 +};
15083 +
15084 +struct _vxm_rqinfo_2 {
15085 +       uint32_t norm_time;
15086 +       uint32_t idle_time;
15087 +       uint32_t idle_skip;
15088 +};
15089 +
15090 +struct _vxm_sched {
15091 +       uint32_t tokens;
15092 +       uint32_t norm_time;
15093 +       uint32_t idle_time;
15094 +};
15095 +
15096 +struct _vxm_task {
15097 +       uint16_t pid;
15098 +       uint16_t state;
15099 +};
15100 +
15101 +struct _vxm_event {
15102 +       uint32_t jif;
15103 +       union {
15104 +               uint32_t seq;
15105 +               uint32_t sec;
15106 +       };
15107 +       union {
15108 +               uint32_t tokens;
15109 +               uint32_t nsec;
15110 +               struct _vxm_task tsk;
15111 +       };
15112 +};
15113 +
15114 +struct _vx_mon_entry {
15115 +       uint16_t type;
15116 +       uint16_t xid;
15117 +       union {
15118 +               struct _vxm_event ev;
15119 +               struct _vxm_sched sd;
15120 +               struct _vxm_update_1 u1;
15121 +               struct _vxm_update_2 u2;
15122 +               struct _vxm_rqinfo_1 q1;
15123 +               struct _vxm_rqinfo_2 q2;
15124 +       };
15125 +};
15126 +
15127 +
15128 +#endif /* _VX_MONITOR_H */
15129 diff -NurpP --minimal linux-3.0/include/linux/vserver/network.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/network.h
15130 --- linux-3.0/include/linux/vserver/network.h   1970-01-01 01:00:00.000000000 +0100
15131 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/network.h      2011-06-10 22:11:24.000000000 +0200
15132 @@ -0,0 +1,146 @@
15133 +#ifndef _VX_NETWORK_H
15134 +#define _VX_NETWORK_H
15135 +
15136 +#include <linux/types.h>
15137 +
15138 +
15139 +#define MAX_N_CONTEXT  65535   /* Arbitrary limit */
15140 +
15141 +
15142 +/* network flags */
15143 +
15144 +#define NXF_INFO_PRIVATE       0x00000008
15145 +
15146 +#define NXF_SINGLE_IP          0x00000100
15147 +#define NXF_LBACK_REMAP                0x00000200
15148 +#define NXF_LBACK_ALLOW                0x00000400
15149 +
15150 +#define NXF_HIDE_NETIF         0x02000000
15151 +#define NXF_HIDE_LBACK         0x04000000
15152 +
15153 +#define NXF_STATE_SETUP                (1ULL << 32)
15154 +#define NXF_STATE_ADMIN                (1ULL << 34)
15155 +
15156 +#define NXF_SC_HELPER          (1ULL << 36)
15157 +#define NXF_PERSISTENT         (1ULL << 38)
15158 +
15159 +#define NXF_ONE_TIME           (0x0005ULL << 32)
15160 +
15161 +
15162 +#define        NXF_INIT_SET            (__nxf_init_set())
15163 +
15164 +static inline uint64_t __nxf_init_set(void) {
15165 +       return    NXF_STATE_ADMIN
15166 +#ifdef CONFIG_VSERVER_AUTO_LBACK
15167 +               | NXF_LBACK_REMAP
15168 +               | NXF_HIDE_LBACK
15169 +#endif
15170 +#ifdef CONFIG_VSERVER_AUTO_SINGLE
15171 +               | NXF_SINGLE_IP
15172 +#endif
15173 +               | NXF_HIDE_NETIF;
15174 +}
15175 +
15176 +
15177 +/* network caps */
15178 +
15179 +#define NXC_TUN_CREATE         0x00000001
15180 +
15181 +#define NXC_RAW_ICMP           0x00000100
15182 +
15183 +
15184 +/* address types */
15185 +
15186 +#define NXA_TYPE_IPV4          0x0001
15187 +#define NXA_TYPE_IPV6          0x0002
15188 +
15189 +#define NXA_TYPE_NONE          0x0000
15190 +#define NXA_TYPE_ANY           0x00FF
15191 +
15192 +#define NXA_TYPE_ADDR          0x0010
15193 +#define NXA_TYPE_MASK          0x0020
15194 +#define NXA_TYPE_RANGE         0x0040
15195 +
15196 +#define NXA_MASK_ALL           (NXA_TYPE_ADDR | NXA_TYPE_MASK | NXA_TYPE_RANGE)
15197 +
15198 +#define NXA_MOD_BCAST          0x0100
15199 +#define NXA_MOD_LBACK          0x0200
15200 +
15201 +#define NXA_LOOPBACK           0x1000
15202 +
15203 +#define NXA_MASK_BIND          (NXA_MASK_ALL | NXA_MOD_BCAST | NXA_MOD_LBACK)
15204 +#define NXA_MASK_SHOW          (NXA_MASK_ALL | NXA_LOOPBACK)
15205 +
15206 +#ifdef __KERNEL__
15207 +
15208 +#include <linux/list.h>
15209 +#include <linux/spinlock.h>
15210 +#include <linux/rcupdate.h>
15211 +#include <linux/in.h>
15212 +#include <linux/in6.h>
15213 +#include <asm/atomic.h>
15214 +
15215 +struct nx_addr_v4 {
15216 +       struct nx_addr_v4 *next;
15217 +       struct in_addr ip[2];
15218 +       struct in_addr mask;
15219 +       uint16_t type;
15220 +       uint16_t flags;
15221 +};
15222 +
15223 +struct nx_addr_v6 {
15224 +       struct nx_addr_v6 *next;
15225 +       struct in6_addr ip;
15226 +       struct in6_addr mask;
15227 +       uint32_t prefix;
15228 +       uint16_t type;
15229 +       uint16_t flags;
15230 +};
15231 +
15232 +struct nx_info {
15233 +       struct hlist_node nx_hlist;     /* linked list of nxinfos */
15234 +       nid_t nx_id;                    /* vnet id */
15235 +       atomic_t nx_usecnt;             /* usage count */
15236 +       atomic_t nx_tasks;              /* tasks count */
15237 +       int nx_state;                   /* context state */
15238 +
15239 +       uint64_t nx_flags;              /* network flag word */
15240 +       uint64_t nx_ncaps;              /* network capabilities */
15241 +
15242 +       struct in_addr v4_lback;        /* Loopback address */
15243 +       struct in_addr v4_bcast;        /* Broadcast address */
15244 +       struct nx_addr_v4 v4;           /* First/Single ipv4 address */
15245 +#ifdef CONFIG_IPV6
15246 +       struct nx_addr_v6 v6;           /* First/Single ipv6 address */
15247 +#endif
15248 +       char nx_name[65];               /* network context name */
15249 +};
15250 +
15251 +
15252 +/* status flags */
15253 +
15254 +#define NXS_HASHED      0x0001
15255 +#define NXS_SHUTDOWN    0x0100
15256 +#define NXS_RELEASED    0x8000
15257 +
15258 +extern struct nx_info *lookup_nx_info(int);
15259 +
15260 +extern int get_nid_list(int, unsigned int *, int);
15261 +extern int nid_is_hashed(nid_t);
15262 +
15263 +extern int nx_migrate_task(struct task_struct *, struct nx_info *);
15264 +
15265 +extern long vs_net_change(struct nx_info *, unsigned int);
15266 +
15267 +struct sock;
15268 +
15269 +
15270 +#define NX_IPV4(n)     ((n)->v4.type != NXA_TYPE_NONE)
15271 +#ifdef  CONFIG_IPV6
15272 +#define NX_IPV6(n)     ((n)->v6.type != NXA_TYPE_NONE)
15273 +#else
15274 +#define NX_IPV6(n)     (0)
15275 +#endif
15276 +
15277 +#endif /* __KERNEL__ */
15278 +#endif /* _VX_NETWORK_H */
15279 diff -NurpP --minimal linux-3.0/include/linux/vserver/network_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/network_cmd.h
15280 --- linux-3.0/include/linux/vserver/network_cmd.h       1970-01-01 01:00:00.000000000 +0100
15281 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/network_cmd.h  2011-06-10 22:11:24.000000000 +0200
15282 @@ -0,0 +1,164 @@
15283 +#ifndef _VX_NETWORK_CMD_H
15284 +#define _VX_NETWORK_CMD_H
15285 +
15286 +
15287 +/* vinfo commands */
15288 +
15289 +#define VCMD_task_nid          VC_CMD(VINFO, 2, 0)
15290 +
15291 +#ifdef __KERNEL__
15292 +extern int vc_task_nid(uint32_t);
15293 +
15294 +#endif /* __KERNEL__ */
15295 +
15296 +#define VCMD_nx_info           VC_CMD(VINFO, 6, 0)
15297 +
15298 +struct vcmd_nx_info_v0 {
15299 +       uint32_t nid;
15300 +       /* more to come */
15301 +};
15302 +
15303 +#ifdef __KERNEL__
15304 +extern int vc_nx_info(struct nx_info *, void __user *);
15305 +
15306 +#endif /* __KERNEL__ */
15307 +
15308 +#include <linux/in.h>
15309 +#include <linux/in6.h>
15310 +
15311 +#define VCMD_net_create_v0     VC_CMD(VNET, 1, 0)
15312 +#define VCMD_net_create                VC_CMD(VNET, 1, 1)
15313 +
15314 +struct  vcmd_net_create {
15315 +       uint64_t flagword;
15316 +};
15317 +
15318 +#define VCMD_net_migrate       VC_CMD(NETMIG, 1, 0)
15319 +
15320 +#define VCMD_net_add           VC_CMD(NETALT, 1, 0)
15321 +#define VCMD_net_remove                VC_CMD(NETALT, 2, 0)
15322 +
15323 +struct vcmd_net_addr_v0 {
15324 +       uint16_t type;
15325 +       uint16_t count;
15326 +       struct in_addr ip[4];
15327 +       struct in_addr mask[4];
15328 +};
15329 +
15330 +#define VCMD_net_add_ipv4_v1   VC_CMD(NETALT, 1, 1)
15331 +#define VCMD_net_rem_ipv4_v1   VC_CMD(NETALT, 2, 1)
15332 +
15333 +struct vcmd_net_addr_ipv4_v1 {
15334 +       uint16_t type;
15335 +       uint16_t flags;
15336 +       struct in_addr ip;
15337 +       struct in_addr mask;
15338 +};
15339 +
15340 +#define VCMD_net_add_ipv4      VC_CMD(NETALT, 1, 2)
15341 +#define VCMD_net_rem_ipv4      VC_CMD(NETALT, 2, 2)
15342 +
15343 +struct vcmd_net_addr_ipv4_v2 {
15344 +       uint16_t type;
15345 +       uint16_t flags;
15346 +       struct in_addr ip;
15347 +       struct in_addr ip2;
15348 +       struct in_addr mask;
15349 +};
15350 +
15351 +#define VCMD_net_add_ipv6      VC_CMD(NETALT, 3, 1)
15352 +#define VCMD_net_remove_ipv6   VC_CMD(NETALT, 4, 1)
15353 +
15354 +struct vcmd_net_addr_ipv6_v1 {
15355 +       uint16_t type;
15356 +       uint16_t flags;
15357 +       uint32_t prefix;
15358 +       struct in6_addr ip;
15359 +       struct in6_addr mask;
15360 +};
15361 +
15362 +#define VCMD_add_match_ipv4    VC_CMD(NETALT, 5, 0)
15363 +#define VCMD_get_match_ipv4    VC_CMD(NETALT, 6, 0)
15364 +
15365 +struct vcmd_match_ipv4_v0 {
15366 +       uint16_t type;
15367 +       uint16_t flags;
15368 +       uint16_t parent;
15369 +       uint16_t prefix;
15370 +       struct in_addr ip;
15371 +       struct in_addr ip2;
15372 +       struct in_addr mask;
15373 +};
15374 +
15375 +#define VCMD_add_match_ipv6    VC_CMD(NETALT, 7, 0)
15376 +#define VCMD_get_match_ipv6    VC_CMD(NETALT, 8, 0)
15377 +
15378 +struct vcmd_match_ipv6_v0 {
15379 +       uint16_t type;
15380 +       uint16_t flags;
15381 +       uint16_t parent;
15382 +       uint16_t prefix;
15383 +       struct in6_addr ip;
15384 +       struct in6_addr ip2;
15385 +       struct in6_addr mask;
15386 +};
15387 +
15388 +
15389 +#ifdef __KERNEL__
15390 +extern int vc_net_create(uint32_t, void __user *);
15391 +extern int vc_net_migrate(struct nx_info *, void __user *);
15392 +
15393 +extern int vc_net_add(struct nx_info *, void __user *);
15394 +extern int vc_net_remove(struct nx_info *, void __user *);
15395 +
15396 +extern int vc_net_add_ipv4_v1(struct nx_info *, void __user *);
15397 +extern int vc_net_add_ipv4(struct nx_info *, void __user *);
15398 +
15399 +extern int vc_net_rem_ipv4_v1(struct nx_info *, void __user *);
15400 +extern int vc_net_rem_ipv4(struct nx_info *, void __user *);
15401 +
15402 +extern int vc_net_add_ipv6(struct nx_info *, void __user *);
15403 +extern int vc_net_remove_ipv6(struct nx_info *, void __user *);
15404 +
15405 +extern int vc_add_match_ipv4(struct nx_info *, void __user *);
15406 +extern int vc_get_match_ipv4(struct nx_info *, void __user *);
15407 +
15408 +extern int vc_add_match_ipv6(struct nx_info *, void __user *);
15409 +extern int vc_get_match_ipv6(struct nx_info *, void __user *);
15410 +
15411 +#endif /* __KERNEL__ */
15412 +
15413 +
15414 +/* flag commands */
15415 +
15416 +#define VCMD_get_nflags                VC_CMD(FLAGS, 5, 0)
15417 +#define VCMD_set_nflags                VC_CMD(FLAGS, 6, 0)
15418 +
15419 +struct vcmd_net_flags_v0 {
15420 +       uint64_t flagword;
15421 +       uint64_t mask;
15422 +};
15423 +
15424 +#ifdef __KERNEL__
15425 +extern int vc_get_nflags(struct nx_info *, void __user *);
15426 +extern int vc_set_nflags(struct nx_info *, void __user *);
15427 +
15428 +#endif /* __KERNEL__ */
15429 +
15430 +
15431 +/* network caps commands */
15432 +
15433 +#define VCMD_get_ncaps         VC_CMD(FLAGS, 7, 0)
15434 +#define VCMD_set_ncaps         VC_CMD(FLAGS, 8, 0)
15435 +
15436 +struct vcmd_net_caps_v0 {
15437 +       uint64_t ncaps;
15438 +       uint64_t cmask;
15439 +};
15440 +
15441 +#ifdef __KERNEL__
15442 +extern int vc_get_ncaps(struct nx_info *, void __user *);
15443 +extern int vc_set_ncaps(struct nx_info *, void __user *);
15444 +
15445 +#endif /* __KERNEL__ */
15446 +#endif /* _VX_CONTEXT_CMD_H */
15447 diff -NurpP --minimal linux-3.0/include/linux/vserver/percpu.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/percpu.h
15448 --- linux-3.0/include/linux/vserver/percpu.h    1970-01-01 01:00:00.000000000 +0100
15449 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/percpu.h       2011-06-10 22:11:24.000000000 +0200
15450 @@ -0,0 +1,14 @@
15451 +#ifndef _VX_PERCPU_H
15452 +#define _VX_PERCPU_H
15453 +
15454 +#include "cvirt_def.h"
15455 +#include "sched_def.h"
15456 +
15457 +struct _vx_percpu {
15458 +       struct _vx_cvirt_pc cvirt;
15459 +       struct _vx_sched_pc sched;
15460 +};
15461 +
15462 +#define        PERCPU_PERCTX   (sizeof(struct _vx_percpu))
15463 +
15464 +#endif /* _VX_PERCPU_H */
15465 diff -NurpP --minimal linux-3.0/include/linux/vserver/pid.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/pid.h
15466 --- linux-3.0/include/linux/vserver/pid.h       1970-01-01 01:00:00.000000000 +0100
15467 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/pid.h  2011-06-10 22:11:24.000000000 +0200
15468 @@ -0,0 +1,51 @@
15469 +#ifndef _VSERVER_PID_H
15470 +#define _VSERVER_PID_H
15471 +
15472 +/* pid faking stuff */
15473 +
15474 +#define vx_info_map_pid(v, p) \
15475 +       __vx_info_map_pid((v), (p), __func__, __FILE__, __LINE__)
15476 +#define vx_info_map_tgid(v,p)  vx_info_map_pid(v,p)
15477 +#define vx_map_pid(p) vx_info_map_pid(current_vx_info(), p)
15478 +#define vx_map_tgid(p) vx_map_pid(p)
15479 +
15480 +static inline int __vx_info_map_pid(struct vx_info *vxi, int pid,
15481 +       const char *func, const char *file, int line)
15482 +{
15483 +       if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
15484 +               vxfprintk(VXD_CBIT(cvirt, 2),
15485 +                       "vx_map_tgid: %p/%llx: %d -> %d",
15486 +                       vxi, (long long)vxi->vx_flags, pid,
15487 +                       (pid && pid == vxi->vx_initpid) ? 1 : pid,
15488 +                       func, file, line);
15489 +               if (pid == 0)
15490 +                       return 0;
15491 +               if (pid == vxi->vx_initpid)
15492 +                       return 1;
15493 +       }
15494 +       return pid;
15495 +}
15496 +
15497 +#define vx_info_rmap_pid(v, p) \
15498 +       __vx_info_rmap_pid((v), (p), __func__, __FILE__, __LINE__)
15499 +#define vx_rmap_pid(p) vx_info_rmap_pid(current_vx_info(), p)
15500 +#define vx_rmap_tgid(p) vx_rmap_pid(p)
15501 +
15502 +static inline int __vx_info_rmap_pid(struct vx_info *vxi, int pid,
15503 +       const char *func, const char *file, int line)
15504 +{
15505 +       if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) {
15506 +               vxfprintk(VXD_CBIT(cvirt, 2),
15507 +                       "vx_rmap_tgid: %p/%llx: %d -> %d",
15508 +                       vxi, (long long)vxi->vx_flags, pid,
15509 +                       (pid == 1) ? vxi->vx_initpid : pid,
15510 +                       func, file, line);
15511 +               if ((pid == 1) && vxi->vx_initpid)
15512 +                       return vxi->vx_initpid;
15513 +               if (pid == vxi->vx_initpid)
15514 +                       return ~0U;
15515 +       }
15516 +       return pid;
15517 +}
15518 +
15519 +#endif
15520 diff -NurpP --minimal linux-3.0/include/linux/vserver/sched.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/sched.h
15521 --- linux-3.0/include/linux/vserver/sched.h     1970-01-01 01:00:00.000000000 +0100
15522 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/sched.h        2011-06-10 22:11:24.000000000 +0200
15523 @@ -0,0 +1,23 @@
15524 +#ifndef _VX_SCHED_H
15525 +#define _VX_SCHED_H
15526 +
15527 +
15528 +#ifdef __KERNEL__
15529 +
15530 +struct timespec;
15531 +
15532 +void vx_vsi_uptime(struct timespec *, struct timespec *);
15533 +
15534 +
15535 +struct vx_info;
15536 +
15537 +void vx_update_load(struct vx_info *);
15538 +
15539 +
15540 +void vx_update_sched_param(struct _vx_sched *sched,
15541 +       struct _vx_sched_pc *sched_pc);
15542 +
15543 +#endif /* __KERNEL__ */
15544 +#else  /* _VX_SCHED_H */
15545 +#warning duplicate inclusion
15546 +#endif /* _VX_SCHED_H */
15547 diff -NurpP --minimal linux-3.0/include/linux/vserver/sched_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/sched_cmd.h
15548 --- linux-3.0/include/linux/vserver/sched_cmd.h 1970-01-01 01:00:00.000000000 +0100
15549 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/sched_cmd.h    2011-06-10 22:11:24.000000000 +0200
15550 @@ -0,0 +1,21 @@
15551 +#ifndef _VX_SCHED_CMD_H
15552 +#define _VX_SCHED_CMD_H
15553 +
15554 +
15555 +struct vcmd_prio_bias {
15556 +       int32_t cpu_id;
15557 +       int32_t prio_bias;
15558 +};
15559 +
15560 +#define VCMD_set_prio_bias     VC_CMD(SCHED, 4, 0)
15561 +#define VCMD_get_prio_bias     VC_CMD(SCHED, 5, 0)
15562 +
15563 +#ifdef __KERNEL__
15564 +
15565 +#include <linux/compiler.h>
15566 +
15567 +extern int vc_set_prio_bias(struct vx_info *, void __user *);
15568 +extern int vc_get_prio_bias(struct vx_info *, void __user *);
15569 +
15570 +#endif /* __KERNEL__ */
15571 +#endif /* _VX_SCHED_CMD_H */
15572 diff -NurpP --minimal linux-3.0/include/linux/vserver/sched_def.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/sched_def.h
15573 --- linux-3.0/include/linux/vserver/sched_def.h 1970-01-01 01:00:00.000000000 +0100
15574 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/sched_def.h    2011-06-10 22:11:24.000000000 +0200
15575 @@ -0,0 +1,38 @@
15576 +#ifndef _VX_SCHED_DEF_H
15577 +#define _VX_SCHED_DEF_H
15578 +
15579 +#include <linux/spinlock.h>
15580 +#include <linux/jiffies.h>
15581 +#include <linux/cpumask.h>
15582 +#include <asm/atomic.h>
15583 +#include <asm/param.h>
15584 +
15585 +
15586 +/* context sub struct */
15587 +
15588 +struct _vx_sched {
15589 +       int prio_bias;                  /* bias offset for priority */
15590 +
15591 +       cpumask_t update;               /* CPUs which should update */
15592 +};
15593 +
15594 +struct _vx_sched_pc {
15595 +       int prio_bias;                  /* bias offset for priority */
15596 +
15597 +       uint64_t user_ticks;            /* token tick events */
15598 +       uint64_t sys_ticks;             /* token tick events */
15599 +       uint64_t hold_ticks;            /* token ticks paused */
15600 +};
15601 +
15602 +
15603 +#ifdef CONFIG_VSERVER_DEBUG
15604 +
15605 +static inline void __dump_vx_sched(struct _vx_sched *sched)
15606 +{
15607 +       printk("\t_vx_sched:\n");
15608 +       printk("\t priority = %4d\n", sched->prio_bias);
15609 +}
15610 +
15611 +#endif
15612 +
15613 +#endif /* _VX_SCHED_DEF_H */
15614 diff -NurpP --minimal linux-3.0/include/linux/vserver/signal.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/signal.h
15615 --- linux-3.0/include/linux/vserver/signal.h    1970-01-01 01:00:00.000000000 +0100
15616 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/signal.h       2011-06-10 22:11:24.000000000 +0200
15617 @@ -0,0 +1,14 @@
15618 +#ifndef _VX_SIGNAL_H
15619 +#define _VX_SIGNAL_H
15620 +
15621 +
15622 +#ifdef __KERNEL__
15623 +
15624 +struct vx_info;
15625 +
15626 +int vx_info_kill(struct vx_info *, int, int);
15627 +
15628 +#endif /* __KERNEL__ */
15629 +#else  /* _VX_SIGNAL_H */
15630 +#warning duplicate inclusion
15631 +#endif /* _VX_SIGNAL_H */
15632 diff -NurpP --minimal linux-3.0/include/linux/vserver/signal_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/signal_cmd.h
15633 --- linux-3.0/include/linux/vserver/signal_cmd.h        1970-01-01 01:00:00.000000000 +0100
15634 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/signal_cmd.h   2011-06-10 22:11:24.000000000 +0200
15635 @@ -0,0 +1,43 @@
15636 +#ifndef _VX_SIGNAL_CMD_H
15637 +#define _VX_SIGNAL_CMD_H
15638 +
15639 +
15640 +/*  signalling vserver commands */
15641 +
15642 +#define VCMD_ctx_kill          VC_CMD(PROCTRL, 1, 0)
15643 +#define VCMD_wait_exit         VC_CMD(EVENT, 99, 0)
15644 +
15645 +struct vcmd_ctx_kill_v0 {
15646 +       int32_t pid;
15647 +       int32_t sig;
15648 +};
15649 +
15650 +struct vcmd_wait_exit_v0 {
15651 +       int32_t reboot_cmd;
15652 +       int32_t exit_code;
15653 +};
15654 +
15655 +#ifdef __KERNEL__
15656 +
15657 +extern int vc_ctx_kill(struct vx_info *, void __user *);
15658 +extern int vc_wait_exit(struct vx_info *, void __user *);
15659 +
15660 +#endif /* __KERNEL__ */
15661 +
15662 +/*  process alteration commands */
15663 +
15664 +#define VCMD_get_pflags                VC_CMD(PROCALT, 5, 0)
15665 +#define VCMD_set_pflags                VC_CMD(PROCALT, 6, 0)
15666 +
15667 +struct vcmd_pflags_v0 {
15668 +       uint32_t flagword;
15669 +       uint32_t mask;
15670 +};
15671 +
15672 +#ifdef __KERNEL__
15673 +
15674 +extern int vc_get_pflags(uint32_t pid, void __user *);
15675 +extern int vc_set_pflags(uint32_t pid, void __user *);
15676 +
15677 +#endif /* __KERNEL__ */
15678 +#endif /* _VX_SIGNAL_CMD_H */
15679 diff -NurpP --minimal linux-3.0/include/linux/vserver/space.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/space.h
15680 --- linux-3.0/include/linux/vserver/space.h     1970-01-01 01:00:00.000000000 +0100
15681 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/space.h        2011-06-10 22:11:24.000000000 +0200
15682 @@ -0,0 +1,12 @@
15683 +#ifndef _VX_SPACE_H
15684 +#define _VX_SPACE_H
15685 +
15686 +#include <linux/types.h>
15687 +
15688 +struct vx_info;
15689 +
15690 +int vx_set_space(struct vx_info *vxi, unsigned long mask, unsigned index);
15691 +
15692 +#else  /* _VX_SPACE_H */
15693 +#warning duplicate inclusion
15694 +#endif /* _VX_SPACE_H */
15695 diff -NurpP --minimal linux-3.0/include/linux/vserver/space_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/space_cmd.h
15696 --- linux-3.0/include/linux/vserver/space_cmd.h 1970-01-01 01:00:00.000000000 +0100
15697 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/space_cmd.h    2011-06-10 22:11:24.000000000 +0200
15698 @@ -0,0 +1,38 @@
15699 +#ifndef _VX_SPACE_CMD_H
15700 +#define _VX_SPACE_CMD_H
15701 +
15702 +
15703 +#define VCMD_enter_space_v0    VC_CMD(PROCALT, 1, 0)
15704 +#define VCMD_enter_space_v1    VC_CMD(PROCALT, 1, 1)
15705 +#define VCMD_enter_space       VC_CMD(PROCALT, 1, 2)
15706 +
15707 +#define VCMD_set_space_v0      VC_CMD(PROCALT, 3, 0)
15708 +#define VCMD_set_space_v1      VC_CMD(PROCALT, 3, 1)
15709 +#define VCMD_set_space         VC_CMD(PROCALT, 3, 2)
15710 +
15711 +#define VCMD_get_space_mask_v0 VC_CMD(PROCALT, 4, 0)
15712 +
15713 +#define VCMD_get_space_mask    VC_CMD(VSPACE, 0, 1)
15714 +#define VCMD_get_space_default VC_CMD(VSPACE, 1, 0)
15715 +
15716 +
15717 +struct vcmd_space_mask_v1 {
15718 +       uint64_t mask;
15719 +};
15720 +
15721 +struct vcmd_space_mask_v2 {
15722 +       uint64_t mask;
15723 +       uint32_t index;
15724 +};
15725 +
15726 +
15727 +#ifdef __KERNEL__
15728 +
15729 +extern int vc_enter_space_v1(struct vx_info *, void __user *);
15730 +extern int vc_set_space_v1(struct vx_info *, void __user *);
15731 +extern int vc_enter_space(struct vx_info *, void __user *);
15732 +extern int vc_set_space(struct vx_info *, void __user *);
15733 +extern int vc_get_space_mask(void __user *, int);
15734 +
15735 +#endif /* __KERNEL__ */
15736 +#endif /* _VX_SPACE_CMD_H */
15737 diff -NurpP --minimal linux-3.0/include/linux/vserver/switch.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/switch.h
15738 --- linux-3.0/include/linux/vserver/switch.h    1970-01-01 01:00:00.000000000 +0100
15739 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/switch.h       2011-06-10 22:11:24.000000000 +0200
15740 @@ -0,0 +1,98 @@
15741 +#ifndef _VX_SWITCH_H
15742 +#define _VX_SWITCH_H
15743 +
15744 +#include <linux/types.h>
15745 +
15746 +
15747 +#define VC_CATEGORY(c)         (((c) >> 24) & 0x3F)
15748 +#define VC_COMMAND(c)          (((c) >> 16) & 0xFF)
15749 +#define VC_VERSION(c)          ((c) & 0xFFF)
15750 +
15751 +#define VC_CMD(c, i, v)                ((((VC_CAT_ ## c) & 0x3F) << 24) \
15752 +                               | (((i) & 0xFF) << 16) | ((v) & 0xFFF))
15753 +
15754 +/*
15755 +
15756 +  Syscall Matrix V2.8
15757 +
15758 +        |VERSION|CREATE |MODIFY |MIGRATE|CONTROL|EXPERIM| |SPECIAL|SPECIAL|
15759 +        |STATS  |DESTROY|ALTER  |CHANGE |LIMIT  |TEST   | |       |       |
15760 +        |INFO   |SETUP  |       |MOVE   |       |       | |       |       |
15761 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15762 +  SYSTEM |VERSION|VSETUP |VHOST  |       |       |       | |DEVICE |       |
15763 +  HOST   |     00|     01|     02|     03|     04|     05| |     06|     07|
15764 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15765 +  CPU    |       |VPROC  |PROCALT|PROCMIG|PROCTRL|       | |SCHED. |       |
15766 +  PROCESS|     08|     09|     10|     11|     12|     13| |     14|     15|
15767 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15768 +  MEMORY |       |       |       |       |MEMCTRL|       | |SWAP   |       |
15769 +        |     16|     17|     18|     19|     20|     21| |     22|     23|
15770 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15771 +  NETWORK|       |VNET   |NETALT |NETMIG |NETCTL |       | |SERIAL |       |
15772 +        |     24|     25|     26|     27|     28|     29| |     30|     31|
15773 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15774 +  DISK   |       |       |       |TAGMIG |DLIMIT |       | |INODE  |       |
15775 +  VFS    |     32|     33|     34|     35|     36|     37| |     38|     39|
15776 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15777 +  OTHER  |VSTAT  |       |       |       |       |       | |VINFO  |       |
15778 +        |     40|     41|     42|     43|     44|     45| |     46|     47|
15779 +  =======+=======+=======+=======+=======+=======+=======+ +=======+=======+
15780 +  SPECIAL|EVENT  |       |       |       |FLAGS  |       | |VSPACE |       |
15781 +        |     48|     49|     50|     51|     52|     53| |     54|     55|
15782 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15783 +  SPECIAL|DEBUG  |       |       |       |RLIMIT |SYSCALL| |       |COMPAT |
15784 +        |     56|     57|     58|     59|     60|TEST 61| |     62|     63|
15785 +  -------+-------+-------+-------+-------+-------+-------+ +-------+-------+
15786 +
15787 +*/
15788 +
15789 +#define VC_CAT_VERSION         0
15790 +
15791 +#define VC_CAT_VSETUP          1
15792 +#define VC_CAT_VHOST           2
15793 +
15794 +#define VC_CAT_DEVICE          6
15795 +
15796 +#define VC_CAT_VPROC           9
15797 +#define VC_CAT_PROCALT         10
15798 +#define VC_CAT_PROCMIG         11
15799 +#define VC_CAT_PROCTRL         12
15800 +
15801 +#define VC_CAT_SCHED           14
15802 +#define VC_CAT_MEMCTRL         20
15803 +
15804 +#define VC_CAT_VNET            25
15805 +#define VC_CAT_NETALT          26
15806 +#define VC_CAT_NETMIG          27
15807 +#define VC_CAT_NETCTRL         28
15808 +
15809 +#define VC_CAT_TAGMIG          35
15810 +#define VC_CAT_DLIMIT          36
15811 +#define VC_CAT_INODE           38
15812 +
15813 +#define VC_CAT_VSTAT           40
15814 +#define VC_CAT_VINFO           46
15815 +#define VC_CAT_EVENT           48
15816 +
15817 +#define VC_CAT_FLAGS           52
15818 +#define VC_CAT_VSPACE          54
15819 +#define VC_CAT_DEBUG           56
15820 +#define VC_CAT_RLIMIT          60
15821 +
15822 +#define VC_CAT_SYSTEST         61
15823 +#define VC_CAT_COMPAT          63
15824 +
15825 +/*  query version */
15826 +
15827 +#define VCMD_get_version       VC_CMD(VERSION, 0, 0)
15828 +#define VCMD_get_vci           VC_CMD(VERSION, 1, 0)
15829 +
15830 +
15831 +#ifdef __KERNEL__
15832 +
15833 +#include <linux/errno.h>
15834 +
15835 +#endif /* __KERNEL__ */
15836 +
15837 +#endif /* _VX_SWITCH_H */
15838 +
15839 diff -NurpP --minimal linux-3.0/include/linux/vserver/tag.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/tag.h
15840 --- linux-3.0/include/linux/vserver/tag.h       1970-01-01 01:00:00.000000000 +0100
15841 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/tag.h  2011-06-10 22:11:24.000000000 +0200
15842 @@ -0,0 +1,143 @@
15843 +#ifndef _DX_TAG_H
15844 +#define _DX_TAG_H
15845 +
15846 +#include <linux/types.h>
15847 +
15848 +
15849 +#define DX_TAG(in)     (IS_TAGGED(in))
15850 +
15851 +
15852 +#ifdef CONFIG_TAG_NFSD
15853 +#define DX_TAG_NFSD    1
15854 +#else
15855 +#define DX_TAG_NFSD    0
15856 +#endif
15857 +
15858 +
15859 +#ifdef CONFIG_TAGGING_NONE
15860 +
15861 +#define MAX_UID                0xFFFFFFFF
15862 +#define MAX_GID                0xFFFFFFFF
15863 +
15864 +#define INOTAG_TAG(cond, uid, gid, tag)        (0)
15865 +
15866 +#define TAGINO_UID(cond, uid, tag)     (uid)
15867 +#define TAGINO_GID(cond, gid, tag)     (gid)
15868 +
15869 +#endif
15870 +
15871 +
15872 +#ifdef CONFIG_TAGGING_GID16
15873 +
15874 +#define MAX_UID                0xFFFFFFFF
15875 +#define MAX_GID                0x0000FFFF
15876 +
15877 +#define INOTAG_TAG(cond, uid, gid, tag)        \
15878 +       ((cond) ? (((gid) >> 16) & 0xFFFF) : 0)
15879 +
15880 +#define TAGINO_UID(cond, uid, tag)     (uid)
15881 +#define TAGINO_GID(cond, gid, tag)     \
15882 +       ((cond) ? (((gid) & 0xFFFF) | ((tag) << 16)) : (gid))
15883 +
15884 +#endif
15885 +
15886 +
15887 +#ifdef CONFIG_TAGGING_ID24
15888 +
15889 +#define MAX_UID                0x00FFFFFF
15890 +#define MAX_GID                0x00FFFFFF
15891 +
15892 +#define INOTAG_TAG(cond, uid, gid, tag)        \
15893 +       ((cond) ? ((((uid) >> 16) & 0xFF00) | (((gid) >> 24) & 0xFF)) : 0)
15894 +
15895 +#define TAGINO_UID(cond, uid, tag)     \
15896 +       ((cond) ? (((uid) & 0xFFFFFF) | (((tag) & 0xFF00) << 16)) : (uid))
15897 +#define TAGINO_GID(cond, gid, tag)     \
15898 +       ((cond) ? (((gid) & 0xFFFFFF) | (((tag) & 0x00FF) << 24)) : (gid))
15899 +
15900 +#endif
15901 +
15902 +
15903 +#ifdef CONFIG_TAGGING_UID16
15904 +
15905 +#define MAX_UID                0x0000FFFF
15906 +#define MAX_GID                0xFFFFFFFF
15907 +
15908 +#define INOTAG_TAG(cond, uid, gid, tag)        \
15909 +       ((cond) ? (((uid) >> 16) & 0xFFFF) : 0)
15910 +
15911 +#define TAGINO_UID(cond, uid, tag)     \
15912 +       ((cond) ? (((uid) & 0xFFFF) | ((tag) << 16)) : (uid))
15913 +#define TAGINO_GID(cond, gid, tag)     (gid)
15914 +
15915 +#endif
15916 +
15917 +
15918 +#ifdef CONFIG_TAGGING_INTERN
15919 +
15920 +#define MAX_UID                0xFFFFFFFF
15921 +#define MAX_GID                0xFFFFFFFF
15922 +
15923 +#define INOTAG_TAG(cond, uid, gid, tag)        \
15924 +       ((cond) ? (tag) : 0)
15925 +
15926 +#define TAGINO_UID(cond, uid, tag)     (uid)
15927 +#define TAGINO_GID(cond, gid, tag)     (gid)
15928 +
15929 +#endif
15930 +
15931 +
15932 +#ifndef CONFIG_TAGGING_NONE
15933 +#define dx_current_fstag(sb)   \
15934 +       ((sb)->s_flags & MS_TAGGED ? dx_current_tag() : 0)
15935 +#else
15936 +#define dx_current_fstag(sb)   (0)
15937 +#endif
15938 +
15939 +#ifndef CONFIG_TAGGING_INTERN
15940 +#define TAGINO_TAG(cond, tag)  (0)
15941 +#else
15942 +#define TAGINO_TAG(cond, tag)  ((cond) ? (tag) : 0)
15943 +#endif
15944 +
15945 +#define INOTAG_UID(cond, uid, gid)     \
15946 +       ((cond) ? ((uid) & MAX_UID) : (uid))
15947 +#define INOTAG_GID(cond, uid, gid)     \
15948 +       ((cond) ? ((gid) & MAX_GID) : (gid))
15949 +
15950 +
15951 +static inline uid_t dx_map_uid(uid_t uid)
15952 +{
15953 +       if ((uid > MAX_UID) && (uid != -1))
15954 +               uid = -2;
15955 +       return (uid & MAX_UID);
15956 +}
15957 +
15958 +static inline gid_t dx_map_gid(gid_t gid)
15959 +{
15960 +       if ((gid > MAX_GID) && (gid != -1))
15961 +               gid = -2;
15962 +       return (gid & MAX_GID);
15963 +}
15964 +
15965 +struct peer_tag {
15966 +       int32_t xid;
15967 +       int32_t nid;
15968 +};
15969 +
15970 +#define dx_notagcheck(sb) ((sb) && ((sb)->s_flags & MS_NOTAGCHECK))
15971 +
15972 +int dx_parse_tag(char *string, tag_t *tag, int remove, int *mnt_flags,
15973 +                unsigned long *flags);
15974 +
15975 +#ifdef CONFIG_PROPAGATE
15976 +
15977 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode);
15978 +
15979 +#define dx_propagate_tag(n, i) __dx_propagate_tag(n, i)
15980 +
15981 +#else
15982 +#define dx_propagate_tag(n, i) do { } while (0)
15983 +#endif
15984 +
15985 +#endif /* _DX_TAG_H */
15986 diff -NurpP --minimal linux-3.0/include/linux/vserver/tag_cmd.h linux-3.0-vs2.3.1-pre7/include/linux/vserver/tag_cmd.h
15987 --- linux-3.0/include/linux/vserver/tag_cmd.h   1970-01-01 01:00:00.000000000 +0100
15988 +++ linux-3.0-vs2.3.1-pre7/include/linux/vserver/tag_cmd.h      2011-06-10 22:11:24.000000000 +0200
15989 @@ -0,0 +1,22 @@
15990 +#ifndef _VX_TAG_CMD_H
15991 +#define _VX_TAG_CMD_H
15992 +
15993 +
15994 +/* vinfo commands */
15995 +
15996 +#define VCMD_task_tag          VC_CMD(VINFO, 3, 0)
15997 +
15998 +#ifdef __KERNEL__
15999 +extern int vc_task_tag(uint32_t);
16000 +
16001 +#endif /* __KERNEL__ */
16002 +
16003 +/* context commands */
16004 +
16005 +#define VCMD_tag_migrate       VC_CMD(TAGMIG, 1, 0)
16006 +
16007 +#ifdef __KERNEL__
16008 +extern int vc_tag_migrate(uint32_t);
16009 +
16010 +#endif /* __KERNEL__ */
16011 +#endif /* _VX_TAG_CMD_H */
16012 diff -NurpP --minimal linux-3.0/include/net/addrconf.h linux-3.0-vs2.3.1-pre7/include/net/addrconf.h
16013 --- linux-3.0/include/net/addrconf.h    2011-07-22 11:18:11.000000000 +0200
16014 +++ linux-3.0-vs2.3.1-pre7/include/net/addrconf.h       2011-06-10 22:11:24.000000000 +0200
16015 @@ -80,7 +80,8 @@ extern int                    ipv6_dev_get_saddr(struct n
16016                                                struct net_device *dev,
16017                                                const struct in6_addr *daddr,
16018                                                unsigned int srcprefs,
16019 -                                              struct in6_addr *saddr);
16020 +                                              struct in6_addr *saddr,
16021 +                                              struct nx_info *nxi);
16022  extern int                     ipv6_get_lladdr(struct net_device *dev,
16023                                                 struct in6_addr *addr,
16024                                                 unsigned char banned_flags);
16025 diff -NurpP --minimal linux-3.0/include/net/af_unix.h linux-3.0-vs2.3.1-pre7/include/net/af_unix.h
16026 --- linux-3.0/include/net/af_unix.h     2011-07-22 11:18:11.000000000 +0200
16027 +++ linux-3.0-vs2.3.1-pre7/include/net/af_unix.h        2011-06-10 22:11:24.000000000 +0200
16028 @@ -4,6 +4,7 @@
16029  #include <linux/socket.h>
16030  #include <linux/un.h>
16031  #include <linux/mutex.h>
16032 +#include <linux/vs_base.h>
16033  #include <net/sock.h>
16034
16035  extern void unix_inflight(struct file *fp);
16036 diff -NurpP --minimal linux-3.0/include/net/inet_timewait_sock.h linux-3.0-vs2.3.1-pre7/include/net/inet_timewait_sock.h
16037 --- linux-3.0/include/net/inet_timewait_sock.h  2011-03-15 18:07:40.000000000 +0100
16038 +++ linux-3.0-vs2.3.1-pre7/include/net/inet_timewait_sock.h     2011-06-10 22:11:24.000000000 +0200
16039 @@ -113,6 +113,10 @@ struct inet_timewait_sock {
16040  #define tw_net                 __tw_common.skc_net
16041  #define tw_daddr               __tw_common.skc_daddr
16042  #define tw_rcv_saddr           __tw_common.skc_rcv_saddr
16043 +#define tw_xid                 __tw_common.skc_xid
16044 +#define tw_vx_info             __tw_common.skc_vx_info
16045 +#define tw_nid                 __tw_common.skc_nid
16046 +#define tw_nx_info             __tw_common.skc_nx_info
16047         int                     tw_timeout;
16048         volatile unsigned char  tw_substate;
16049         unsigned char           tw_rcv_wscale;
16050 diff -NurpP --minimal linux-3.0/include/net/ip6_route.h linux-3.0-vs2.3.1-pre7/include/net/ip6_route.h
16051 --- linux-3.0/include/net/ip6_route.h   2011-07-22 11:18:11.000000000 +0200
16052 +++ linux-3.0-vs2.3.1-pre7/include/net/ip6_route.h      2011-06-16 14:16:51.000000000 +0200
16053 @@ -86,7 +86,8 @@ extern int                    ip6_route_get_saddr(struct
16054                                                     struct rt6_info *rt,
16055                                                     const struct in6_addr *daddr,
16056                                                     unsigned int prefs,
16057 -                                                   struct in6_addr *saddr);
16058 +                                                   struct in6_addr *saddr,
16059 +                                                   struct nx_info *nxi);
16060
16061  extern struct rt6_info         *rt6_lookup(struct net *net,
16062                                             const struct in6_addr *daddr,
16063 diff -NurpP --minimal linux-3.0/include/net/route.h linux-3.0-vs2.3.1-pre7/include/net/route.h
16064 --- linux-3.0/include/net/route.h       2011-07-22 11:18:11.000000000 +0200
16065 +++ linux-3.0-vs2.3.1-pre7/include/net/route.h  2011-06-13 18:20:44.000000000 +0200
16066 @@ -202,6 +202,9 @@ static inline void ip_rt_put(struct rtab
16067                 dst_release(&rt->dst);
16068  }
16069
16070 +#include <linux/vs_base.h>
16071 +#include <linux/vs_inet.h>
16072 +
16073  #define IPTOS_RT_MASK  (IPTOS_TOS_MASK & ~3)
16074
16075  extern const __u8 ip_tos2prio[16];
16076 @@ -253,6 +256,9 @@ static inline void ip_route_connect_init
16077                            protocol, flow_flags, dst, src, dport, sport);
16078  }
16079
16080 +extern struct rtable *ip_v4_find_src(struct net *net, struct nx_info *,
16081 +       struct flowi4 *);
16082 +
16083  static inline struct rtable *ip_route_connect(struct flowi4 *fl4,
16084                                               __be32 dst, __be32 src, u32 tos,
16085                                               int oif, u8 protocol,
16086 @@ -261,11 +267,24 @@ static inline struct rtable *ip_route_co
16087  {
16088         struct net *net = sock_net(sk);
16089         struct rtable *rt;
16090 +       struct nx_info *nx_info = current_nx_info();
16091
16092         ip_route_connect_init(fl4, dst, src, tos, oif, protocol,
16093                               sport, dport, sk, can_sleep);
16094
16095 -       if (!dst || !src) {
16096 +       if (sk)
16097 +               nx_info = sk->sk_nx_info;
16098 +
16099 +       vxdprintk(VXD_CBIT(net, 4),
16100 +               "ip_route_connect(%p) %p,%p;%lx",
16101 +               sk, nx_info, sk->sk_socket,
16102 +               (sk->sk_socket?sk->sk_socket->flags:0));
16103 +
16104 +       rt = ip_v4_find_src(net, nx_info, fl4);
16105 +       if (IS_ERR(rt))
16106 +               return rt;
16107 +
16108 +       if (!fl4->daddr || !fl4->saddr) {
16109                 rt = __ip_route_output_key(net, fl4);
16110                 if (IS_ERR(rt))
16111                         return rt;
16112 diff -NurpP --minimal linux-3.0/include/net/sock.h linux-3.0-vs2.3.1-pre7/include/net/sock.h
16113 --- linux-3.0/include/net/sock.h        2011-07-22 11:18:12.000000000 +0200
16114 +++ linux-3.0-vs2.3.1-pre7/include/net/sock.h   2011-07-01 11:35:35.000000000 +0200
16115 @@ -149,6 +149,10 @@ struct sock_common {
16116  #ifdef CONFIG_NET_NS
16117         struct net              *skc_net;
16118  #endif
16119 +       xid_t                   skc_xid;
16120 +       struct vx_info          *skc_vx_info;
16121 +       nid_t                   skc_nid;
16122 +       struct nx_info          *skc_nx_info;
16123         /*
16124          * fields between dontcopy_begin/dontcopy_end
16125          * are not copied in sock_copy()
16126 @@ -256,6 +260,10 @@ struct sock {
16127  #define sk_bind_node           __sk_common.skc_bind_node
16128  #define sk_prot                        __sk_common.skc_prot
16129  #define sk_net                 __sk_common.skc_net
16130 +#define sk_xid                 __sk_common.skc_xid
16131 +#define sk_vx_info             __sk_common.skc_vx_info
16132 +#define sk_nid                 __sk_common.skc_nid
16133 +#define sk_nx_info             __sk_common.skc_nx_info
16134         socket_lock_t           sk_lock;
16135         struct sk_buff_head     sk_receive_queue;
16136         /*
16137 diff -NurpP --minimal linux-3.0/init/Kconfig linux-3.0-vs2.3.1-pre7/init/Kconfig
16138 --- linux-3.0/init/Kconfig      2011-07-22 11:18:12.000000000 +0200
16139 +++ linux-3.0-vs2.3.1-pre7/init/Kconfig 2011-07-20 02:16:43.000000000 +0200
16140 @@ -574,6 +574,7 @@ config HAVE_UNSTABLE_SCHED_CLOCK
16141  menuconfig CGROUPS
16142         boolean "Control Group support"
16143         depends on EVENTFD
16144 +       default y
16145         help
16146           This option adds support for grouping sets of processes together, for
16147           use with process control subsystems such as Cpusets, CFS, memory
16148 @@ -790,7 +791,7 @@ config IPC_NS
16149  config USER_NS
16150         bool "User namespace (EXPERIMENTAL)"
16151         depends on EXPERIMENTAL
16152 -       default y
16153 +       default n
16154         help
16155           This allows containers, i.e. vservers, to use user namespaces
16156           to provide different user info for different servers.
16157 diff -NurpP --minimal linux-3.0/init/main.c linux-3.0-vs2.3.1-pre7/init/main.c
16158 --- linux-3.0/init/main.c       2011-07-22 11:18:12.000000000 +0200
16159 +++ linux-3.0-vs2.3.1-pre7/init/main.c  2011-06-22 12:39:15.000000000 +0200
16160 @@ -68,6 +68,7 @@
16161  #include <linux/shmem_fs.h>
16162  #include <linux/slab.h>
16163  #include <linux/perf_event.h>
16164 +#include <linux/vserver/percpu.h>
16165
16166  #include <asm/io.h>
16167  #include <asm/bugs.h>
16168 diff -NurpP --minimal linux-3.0/ipc/mqueue.c linux-3.0-vs2.3.1-pre7/ipc/mqueue.c
16169 --- linux-3.0/ipc/mqueue.c      2011-03-15 18:07:41.000000000 +0100
16170 +++ linux-3.0-vs2.3.1-pre7/ipc/mqueue.c 2011-06-10 22:11:24.000000000 +0200
16171 @@ -33,6 +33,8 @@
16172  #include <linux/pid.h>
16173  #include <linux/ipc_namespace.h>
16174  #include <linux/slab.h>
16175 +#include <linux/vs_context.h>
16176 +#include <linux/vs_limit.h>
16177
16178  #include <net/sock.h>
16179  #include "util.h"
16180 @@ -66,6 +68,7 @@ struct mqueue_inode_info {
16181         struct sigevent notify;
16182         struct pid* notify_owner;
16183         struct user_struct *user;       /* user who created, for accounting */
16184 +       struct vx_info *vxi;
16185         struct sock *notify_sock;
16186         struct sk_buff *notify_cookie;
16187
16188 @@ -126,6 +129,7 @@ static struct inode *mqueue_get_inode(st
16189                 if (S_ISREG(mode)) {
16190                         struct mqueue_inode_info *info;
16191                         struct task_struct *p = current;
16192 +                       struct vx_info *vxi = p->vx_info;
16193                         unsigned long mq_bytes, mq_msg_tblsz;
16194
16195                         inode->i_fop = &mqueue_file_operations;
16196 @@ -139,6 +143,7 @@ static struct inode *mqueue_get_inode(st
16197                         info->notify_owner = NULL;
16198                         info->qsize = 0;
16199                         info->user = NULL;      /* set when all is ok */
16200 +                       info->vxi = NULL;
16201                         memset(&info->attr, 0, sizeof(info->attr));
16202                         info->attr.mq_maxmsg = ipc_ns->mq_msg_max;
16203                         info->attr.mq_msgsize = ipc_ns->mq_msgsize_max;
16204 @@ -157,16 +162,19 @@ static struct inode *mqueue_get_inode(st
16205                         spin_lock(&mq_lock);
16206                         if (u->mq_bytes + mq_bytes < u->mq_bytes ||
16207                             u->mq_bytes + mq_bytes >
16208 -                           task_rlimit(p, RLIMIT_MSGQUEUE)) {
16209 +                           task_rlimit(p, RLIMIT_MSGQUEUE) ||
16210 +                           !vx_ipcmsg_avail(vxi, mq_bytes)) {
16211                                 spin_unlock(&mq_lock);
16212                                 /* mqueue_evict_inode() releases info->messages */
16213                                 goto out_inode;
16214                         }
16215                         u->mq_bytes += mq_bytes;
16216 +                       vx_ipcmsg_add(vxi, u, mq_bytes);
16217                         spin_unlock(&mq_lock);
16218
16219                         /* all is ok */
16220                         info->user = get_uid(u);
16221 +                       info->vxi = get_vx_info(vxi);
16222                 } else if (S_ISDIR(mode)) {
16223                         inc_nlink(inode);
16224                         /* Some things misbehave if size == 0 on a directory */
16225 @@ -275,8 +283,11 @@ static void mqueue_evict_inode(struct in
16226             + info->attr.mq_msgsize);
16227         user = info->user;
16228         if (user) {
16229 +               struct vx_info *vxi = info->vxi;
16230 +
16231                 spin_lock(&mq_lock);
16232                 user->mq_bytes -= mq_bytes;
16233 +               vx_ipcmsg_sub(vxi, user, mq_bytes);
16234                 /*
16235                  * get_ns_from_inode() ensures that the
16236                  * (ipc_ns = sb->s_fs_info) is either a valid ipc_ns
16237 @@ -286,6 +297,7 @@ static void mqueue_evict_inode(struct in
16238                 if (ipc_ns)
16239                         ipc_ns->mq_queues_count--;
16240                 spin_unlock(&mq_lock);
16241 +               put_vx_info(vxi);
16242                 free_uid(user);
16243         }
16244         if (ipc_ns)
16245 diff -NurpP --minimal linux-3.0/ipc/msg.c linux-3.0-vs2.3.1-pre7/ipc/msg.c
16246 --- linux-3.0/ipc/msg.c 2011-05-22 16:17:59.000000000 +0200
16247 +++ linux-3.0-vs2.3.1-pre7/ipc/msg.c    2011-06-10 22:11:24.000000000 +0200
16248 @@ -37,6 +37,7 @@
16249  #include <linux/rwsem.h>
16250  #include <linux/nsproxy.h>
16251  #include <linux/ipc_namespace.h>
16252 +#include <linux/vs_base.h>
16253
16254  #include <asm/current.h>
16255  #include <asm/uaccess.h>
16256 @@ -190,6 +191,7 @@ static int newque(struct ipc_namespace *
16257
16258         msq->q_perm.mode = msgflg & S_IRWXUGO;
16259         msq->q_perm.key = key;
16260 +       msq->q_perm.xid = vx_current_xid();
16261
16262         msq->q_perm.security = NULL;
16263         retval = security_msg_queue_alloc(msq);
16264 diff -NurpP --minimal linux-3.0/ipc/namespace.c linux-3.0-vs2.3.1-pre7/ipc/namespace.c
16265 --- linux-3.0/ipc/namespace.c   2011-07-22 11:18:12.000000000 +0200
16266 +++ linux-3.0-vs2.3.1-pre7/ipc/namespace.c      2011-06-13 14:09:44.000000000 +0200
16267 @@ -13,11 +13,12 @@
16268  #include <linux/mount.h>
16269  #include <linux/user_namespace.h>
16270  #include <linux/proc_fs.h>
16271 +#include <linux/vs_base.h>
16272 +#include <linux/vserver/global.h>
16273
16274  #include "util.h"
16275
16276 -static struct ipc_namespace *create_ipc_ns(struct task_struct *tsk,
16277 -                                          struct ipc_namespace *old_ns)
16278 +static struct ipc_namespace *create_ipc_ns(struct user_namespace *user_ns)
16279  {
16280         struct ipc_namespace *ns;
16281         int err;
16282 @@ -46,19 +47,18 @@ static struct ipc_namespace *create_ipc_
16283         ipcns_notify(IPCNS_CREATED);
16284         register_ipcns_notifier(ns);
16285
16286 -       ns->user_ns = get_user_ns(task_cred_xxx(tsk, user)->user_ns);
16287 +       ns->user_ns = get_user_ns(user_ns);
16288
16289         return ns;
16290  }
16291
16292  struct ipc_namespace *copy_ipcs(unsigned long flags,
16293 -                               struct task_struct *tsk)
16294 +                               struct ipc_namespace *old_ns,
16295 +                               struct user_namespace *user_ns)
16296  {
16297 -       struct ipc_namespace *ns = tsk->nsproxy->ipc_ns;
16298 -
16299         if (!(flags & CLONE_NEWIPC))
16300 -               return get_ipc_ns(ns);
16301 -       return create_ipc_ns(tsk, ns);
16302 +               return get_ipc_ns(old_ns);
16303 +       return create_ipc_ns(user_ns);
16304  }
16305
16306  /*
16307 diff -NurpP --minimal linux-3.0/ipc/sem.c linux-3.0-vs2.3.1-pre7/ipc/sem.c
16308 --- linux-3.0/ipc/sem.c 2011-05-22 16:17:59.000000000 +0200
16309 +++ linux-3.0-vs2.3.1-pre7/ipc/sem.c    2011-06-10 22:11:24.000000000 +0200
16310 @@ -86,6 +86,8 @@
16311  #include <linux/rwsem.h>
16312  #include <linux/nsproxy.h>
16313  #include <linux/ipc_namespace.h>
16314 +#include <linux/vs_base.h>
16315 +#include <linux/vs_limit.h>
16316
16317  #include <asm/uaccess.h>
16318  #include "util.h"
16319 @@ -260,6 +262,7 @@ static int newary(struct ipc_namespace *
16320
16321         sma->sem_perm.mode = (semflg & S_IRWXUGO);
16322         sma->sem_perm.key = key;
16323 +       sma->sem_perm.xid = vx_current_xid();
16324
16325         sma->sem_perm.security = NULL;
16326         retval = security_sem_alloc(sma);
16327 @@ -275,6 +278,9 @@ static int newary(struct ipc_namespace *
16328                 return id;
16329         }
16330         ns->used_sems += nsems;
16331 +       /* FIXME: obsoleted? */
16332 +       vx_semary_inc(sma);
16333 +       vx_nsems_add(sma, nsems);
16334
16335         sma->sem_base = (struct sem *) &sma[1];
16336
16337 @@ -730,6 +736,9 @@ static void freeary(struct ipc_namespace
16338
16339         wake_up_sem_queue_do(&tasks);
16340         ns->used_sems -= sma->sem_nsems;
16341 +       /* FIXME: obsoleted? */
16342 +       vx_nsems_sub(sma, sma->sem_nsems);
16343 +       vx_semary_dec(sma);
16344         security_sem_free(sma);
16345         ipc_rcu_putref(sma);
16346  }
16347 diff -NurpP --minimal linux-3.0/ipc/shm.c linux-3.0-vs2.3.1-pre7/ipc/shm.c
16348 --- linux-3.0/ipc/shm.c 2011-07-22 11:18:12.000000000 +0200
16349 +++ linux-3.0-vs2.3.1-pre7/ipc/shm.c    2011-06-10 22:11:24.000000000 +0200
16350 @@ -39,6 +39,8 @@
16351  #include <linux/nsproxy.h>
16352  #include <linux/mount.h>
16353  #include <linux/ipc_namespace.h>
16354 +#include <linux/vs_context.h>
16355 +#include <linux/vs_limit.h>
16356
16357  #include <asm/uaccess.h>
16358
16359 @@ -173,7 +175,12 @@ static void shm_open(struct vm_area_stru
16360   */
16361  static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
16362  {
16363 -       ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
16364 +       struct vx_info *vxi = lookup_vx_info(shp->shm_perm.xid);
16365 +       int numpages = (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
16366 +
16367 +       vx_ipcshm_sub(vxi, shp, numpages);
16368 +       ns->shm_tot -= numpages;
16369 +
16370         shm_rmid(ns, shp);
16371         shm_unlock(shp);
16372         if (!is_file_hugepages(shp->shm_file))
16373 @@ -183,6 +190,7 @@ static void shm_destroy(struct ipc_names
16374                                                 shp->mlock_user);
16375         fput (shp->shm_file);
16376         security_shm_free(shp);
16377 +       put_vx_info(vxi);
16378         ipc_rcu_putref(shp);
16379  }
16380
16381 @@ -355,11 +363,15 @@ static int newseg(struct ipc_namespace *
16382         if (ns->shm_tot + numpages > ns->shm_ctlall)
16383                 return -ENOSPC;
16384
16385 +       if (!vx_ipcshm_avail(current_vx_info(), numpages))
16386 +               return -ENOSPC;
16387 +
16388         shp = ipc_rcu_alloc(sizeof(*shp));
16389         if (!shp)
16390                 return -ENOMEM;
16391
16392         shp->shm_perm.key = key;
16393 +       shp->shm_perm.xid = vx_current_xid();
16394         shp->shm_perm.mode = (shmflg & S_IRWXUGO);
16395         shp->mlock_user = NULL;
16396
16397 @@ -413,6 +425,7 @@ static int newseg(struct ipc_namespace *
16398         ns->shm_tot += numpages;
16399         error = shp->shm_perm.id;
16400         shm_unlock(shp);
16401 +       vx_ipcshm_add(current_vx_info(), key, numpages);
16402         return error;
16403
16404  no_id:
16405 diff -NurpP --minimal linux-3.0/kernel/Makefile linux-3.0-vs2.3.1-pre7/kernel/Makefile
16406 --- linux-3.0/kernel/Makefile   2011-07-22 11:18:12.000000000 +0200
16407 +++ linux-3.0-vs2.3.1-pre7/kernel/Makefile      2011-06-10 22:11:24.000000000 +0200
16408 @@ -24,6 +24,7 @@ CFLAGS_REMOVE_sched_clock.o = -pg
16409  CFLAGS_REMOVE_irq_work.o = -pg
16410  endif
16411
16412 +obj-y += vserver/
16413  obj-$(CONFIG_FREEZER) += freezer.o
16414  obj-$(CONFIG_PROFILING) += profile.o
16415  obj-$(CONFIG_SYSCTL_SYSCALL_CHECK) += sysctl_check.o
16416 diff -NurpP --minimal linux-3.0/kernel/capability.c linux-3.0-vs2.3.1-pre7/kernel/capability.c
16417 --- linux-3.0/kernel/capability.c       2011-07-22 11:18:12.000000000 +0200
16418 +++ linux-3.0-vs2.3.1-pre7/kernel/capability.c  2011-07-19 19:27:53.000000000 +0200
16419 @@ -15,6 +15,7 @@
16420  #include <linux/syscalls.h>
16421  #include <linux/pid_namespace.h>
16422  #include <linux/user_namespace.h>
16423 +#include <linux/vs_context.h>
16424  #include <asm/uaccess.h>
16425
16426  /*
16427 @@ -116,6 +117,7 @@ static int cap_validate_magic(cap_user_h
16428         return 0;
16429  }
16430
16431 +
16432  /*
16433   * The only thing that can change the capabilities of the current
16434   * process is the current process. As such, we can't be in this code
16435 @@ -340,6 +342,8 @@ bool has_capability_noaudit(struct task_
16436         return (ret == 0);
16437  }
16438
16439 +#include <linux/vserver/base.h>
16440 +
16441  /**
16442   * capable - Determine if the current task has a superior capability in effect
16443   * @cap: The capability to be tested for
16444 @@ -374,6 +378,9 @@ bool ns_capable(struct user_namespace *n
16445                 BUG();
16446         }
16447
16448 +       if (vs_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap))
16449 +               return true;
16450 +
16451         if (security_capable(ns, current_cred(), cap) == 0) {
16452                 current->flags |= PF_SUPERPRIV;
16453                 return true;
16454 diff -NurpP --minimal linux-3.0/kernel/compat.c linux-3.0-vs2.3.1-pre7/kernel/compat.c
16455 --- linux-3.0/kernel/compat.c   2011-07-22 11:18:12.000000000 +0200
16456 +++ linux-3.0-vs2.3.1-pre7/kernel/compat.c      2011-06-10 22:11:24.000000000 +0200
16457 @@ -970,7 +970,7 @@ asmlinkage long compat_sys_stime(compat_
16458         if (err)
16459                 return err;
16460
16461 -       do_settimeofday(&tv);
16462 +       vx_settimeofday(&tv);
16463         return 0;
16464  }
16465
16466 diff -NurpP --minimal linux-3.0/kernel/cred.c linux-3.0-vs2.3.1-pre7/kernel/cred.c
16467 --- linux-3.0/kernel/cred.c     2011-07-22 11:18:12.000000000 +0200
16468 +++ linux-3.0-vs2.3.1-pre7/kernel/cred.c        2011-06-10 22:11:24.000000000 +0200
16469 @@ -61,31 +61,6 @@ struct cred init_cred = {
16470  #endif
16471  };
16472
16473 -static inline void set_cred_subscribers(struct cred *cred, int n)
16474 -{
16475 -#ifdef CONFIG_DEBUG_CREDENTIALS
16476 -       atomic_set(&cred->subscribers, n);
16477 -#endif
16478 -}
16479 -
16480 -static inline int read_cred_subscribers(const struct cred *cred)
16481 -{
16482 -#ifdef CONFIG_DEBUG_CREDENTIALS
16483 -       return atomic_read(&cred->subscribers);
16484 -#else
16485 -       return 0;
16486 -#endif
16487 -}
16488 -
16489 -static inline void alter_cred_subscribers(const struct cred *_cred, int n)
16490 -{
16491 -#ifdef CONFIG_DEBUG_CREDENTIALS
16492 -       struct cred *cred = (struct cred *) _cred;
16493 -
16494 -       atomic_add(n, &cred->subscribers);
16495 -#endif
16496 -}
16497 -
16498  /*
16499   * Dispose of the shared task group credentials
16500   */
16501 @@ -281,21 +256,16 @@ error:
16502   *
16503   * Call commit_creds() or abort_creds() to clean up.
16504   */
16505 -struct cred *prepare_creds(void)
16506 +struct cred *__prepare_creds(const struct cred *old)
16507  {
16508 -       struct task_struct *task = current;
16509 -       const struct cred *old;
16510         struct cred *new;
16511
16512 -       validate_process_creds();
16513 -
16514         new = kmem_cache_alloc(cred_jar, GFP_KERNEL);
16515         if (!new)
16516                 return NULL;
16517
16518         kdebug("prepare_creds() alloc %p", new);
16519
16520 -       old = task->cred;
16521         memcpy(new, old, sizeof(struct cred));
16522
16523         atomic_set(&new->usage, 1);
16524 @@ -322,6 +292,13 @@ error:
16525         abort_creds(new);
16526         return NULL;
16527  }
16528 +
16529 +struct cred *prepare_creds(void)
16530 +{
16531 +       validate_process_creds();
16532 +
16533 +       return __prepare_creds(current->cred);
16534 +}
16535  EXPORT_SYMBOL(prepare_creds);
16536
16537  /*
16538 diff -NurpP --minimal linux-3.0/kernel/exit.c linux-3.0-vs2.3.1-pre7/kernel/exit.c
16539 --- linux-3.0/kernel/exit.c     2011-07-22 11:18:12.000000000 +0200
16540 +++ linux-3.0-vs2.3.1-pre7/kernel/exit.c        2011-06-22 12:39:15.000000000 +0200
16541 @@ -48,6 +48,10 @@
16542  #include <linux/fs_struct.h>
16543  #include <linux/init_task.h>
16544  #include <linux/perf_event.h>
16545 +#include <linux/vs_limit.h>
16546 +#include <linux/vs_context.h>
16547 +#include <linux/vs_network.h>
16548 +#include <linux/vs_pid.h>
16549  #include <trace/events/sched.h>
16550  #include <linux/hw_breakpoint.h>
16551  #include <linux/oom.h>
16552 @@ -494,9 +498,11 @@ static void close_files(struct files_str
16553                                         filp_close(file, files);
16554                                         cond_resched();
16555                                 }
16556 +                               vx_openfd_dec(i);
16557                         }
16558                         i++;
16559                         set >>= 1;
16560 +                       cond_resched();
16561                 }
16562         }
16563  }
16564 @@ -1047,11 +1053,16 @@ NORET_TYPE void do_exit(long code)
16565
16566         validate_creds_for_do_exit(tsk);
16567
16568 +       /* needs to stay after exit_notify() */
16569 +       exit_vx_info(tsk, code);
16570 +       exit_nx_info(tsk);
16571 +
16572         preempt_disable();
16573         exit_rcu();
16574         /* causes final put_task_struct in finish_task_switch(). */
16575         tsk->state = TASK_DEAD;
16576         schedule();
16577 +       printk("bad task: %p [%lx]\n", current, current->state);
16578         BUG();
16579         /* Avoid "noreturn function does return".  */
16580         for (;;)
16581 diff -NurpP --minimal linux-3.0/kernel/fork.c linux-3.0-vs2.3.1-pre7/kernel/fork.c
16582 --- linux-3.0/kernel/fork.c     2011-07-22 11:18:12.000000000 +0200
16583 +++ linux-3.0-vs2.3.1-pre7/kernel/fork.c        2011-06-10 22:21:02.000000000 +0200
16584 @@ -67,6 +67,10 @@
16585  #include <linux/user-return-notifier.h>
16586  #include <linux/oom.h>
16587  #include <linux/khugepaged.h>
16588 +#include <linux/vs_context.h>
16589 +#include <linux/vs_network.h>
16590 +#include <linux/vs_limit.h>
16591 +#include <linux/vs_memory.h>
16592
16593  #include <asm/pgtable.h>
16594  #include <asm/pgalloc.h>
16595 @@ -167,6 +171,8 @@ void free_task(struct task_struct *tsk)
16596         account_kernel_stack(tsk->stack, -1);
16597         free_thread_info(tsk->stack);
16598         rt_mutex_debug_task_free(tsk);
16599 +       clr_vx_info(&tsk->vx_info);
16600 +       clr_nx_info(&tsk->nx_info);
16601         ftrace_graph_exit_task(tsk);
16602         free_task_struct(tsk);
16603  }
16604 @@ -505,6 +511,7 @@ static struct mm_struct * mm_init(struct
16605         if (likely(!mm_alloc_pgd(mm))) {
16606                 mm->def_flags = 0;
16607                 mmu_notifier_mm_init(mm);
16608 +               set_vx_info(&mm->mm_vx_info, p->vx_info);
16609                 return mm;
16610         }
16611
16612 @@ -542,6 +549,7 @@ void __mmdrop(struct mm_struct *mm)
16613  #ifdef CONFIG_TRANSPARENT_HUGEPAGE
16614         VM_BUG_ON(mm->pmd_huge_pte);
16615  #endif
16616 +       clr_vx_info(&mm->mm_vx_info);
16617         free_mm(mm);
16618  }
16619  EXPORT_SYMBOL_GPL(__mmdrop);
16620 @@ -729,6 +737,7 @@ struct mm_struct *dup_mm(struct task_str
16621                 goto fail_nomem;
16622
16623         memcpy(mm, oldmm, sizeof(*mm));
16624 +       mm->mm_vx_info = NULL;
16625         mm_init_cpumask(mm);
16626
16627         /* Initializing for Swap token stuff */
16628 @@ -772,6 +781,7 @@ fail_nocontext:
16629          * If init_new_context() failed, we cannot use mmput() to free the mm
16630          * because it calls destroy_context()
16631          */
16632 +       clr_vx_info(&mm->mm_vx_info);
16633         mm_free_pgd(mm);
16634         free_mm(mm);
16635         return NULL;
16636 @@ -1057,6 +1067,8 @@ static struct task_struct *copy_process(
16637         int retval;
16638         struct task_struct *p;
16639         int cgroup_callbacks_done = 0;
16640 +       struct vx_info *vxi;
16641 +       struct nx_info *nxi;
16642
16643         if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
16644                 return ERR_PTR(-EINVAL);
16645 @@ -1103,7 +1115,12 @@ static struct task_struct *copy_process(
16646         DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
16647         DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
16648  #endif
16649 +       init_vx_info(&p->vx_info, current_vx_info());
16650 +       init_nx_info(&p->nx_info, current_nx_info());
16651 +
16652         retval = -EAGAIN;
16653 +       if (!vx_nproc_avail(1))
16654 +               goto bad_fork_free;
16655         if (atomic_read(&p->real_cred->user->processes) >=
16656                         task_rlimit(p, RLIMIT_NPROC)) {
16657                 if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
16658 @@ -1360,6 +1377,18 @@ static struct task_struct *copy_process(
16659
16660         total_forks++;
16661         spin_unlock(&current->sighand->siglock);
16662 +
16663 +       /* p is copy of current */
16664 +       vxi = p->vx_info;
16665 +       if (vxi) {
16666 +               claim_vx_info(vxi, p);
16667 +               atomic_inc(&vxi->cvirt.nr_threads);
16668 +               atomic_inc(&vxi->cvirt.total_forks);
16669 +               vx_nproc_inc(p);
16670 +       }
16671 +       nxi = p->nx_info;
16672 +       if (nxi)
16673 +               claim_nx_info(nxi, p);
16674         write_unlock_irq(&tasklist_lock);
16675         proc_fork_connector(p);
16676         cgroup_post_fork(p);
16677 diff -NurpP --minimal linux-3.0/kernel/kthread.c linux-3.0-vs2.3.1-pre7/kernel/kthread.c
16678 --- linux-3.0/kernel/kthread.c  2011-07-22 11:18:12.000000000 +0200
16679 +++ linux-3.0-vs2.3.1-pre7/kernel/kthread.c     2011-06-10 22:11:24.000000000 +0200
16680 @@ -16,6 +16,7 @@
16681  #include <linux/mutex.h>
16682  #include <linux/slab.h>
16683  #include <linux/freezer.h>
16684 +#include <linux/vs_pid.h>
16685  #include <trace/events/sched.h>
16686
16687  static DEFINE_SPINLOCK(kthread_create_lock);
16688 diff -NurpP --minimal linux-3.0/kernel/nsproxy.c linux-3.0-vs2.3.1-pre7/kernel/nsproxy.c
16689 --- linux-3.0/kernel/nsproxy.c  2011-07-22 11:18:12.000000000 +0200
16690 +++ linux-3.0-vs2.3.1-pre7/kernel/nsproxy.c     2011-06-13 14:09:44.000000000 +0200
16691 @@ -20,6 +20,8 @@
16692  #include <linux/mnt_namespace.h>
16693  #include <linux/utsname.h>
16694  #include <linux/pid_namespace.h>
16695 +#include <linux/vserver/global.h>
16696 +#include <linux/vserver/debug.h>
16697  #include <net/net_namespace.h>
16698  #include <linux/ipc_namespace.h>
16699  #include <linux/proc_fs.h>
16700 @@ -46,8 +48,11 @@ static inline struct nsproxy *create_nsp
16701         struct nsproxy *nsproxy;
16702
16703         nsproxy = kmem_cache_alloc(nsproxy_cachep, GFP_KERNEL);
16704 -       if (nsproxy)
16705 +       if (nsproxy) {
16706                 atomic_set(&nsproxy->count, 1);
16707 +               atomic_inc(&vs_global_nsproxy);
16708 +       }
16709 +       vxdprintk(VXD_CBIT(space, 2), "create_nsproxy = %p[1]", nsproxy);
16710         return nsproxy;
16711  }
16712
16713 @@ -56,8 +61,11 @@ static inline struct nsproxy *create_nsp
16714   * Return the newly created nsproxy.  Do not attach this to the task,
16715   * leave it to the caller to do proper locking and attach it to task.
16716   */
16717 -static struct nsproxy *create_new_namespaces(unsigned long flags,
16718 -                       struct task_struct *tsk, struct fs_struct *new_fs)
16719 +static struct nsproxy *unshare_namespaces(unsigned long flags,
16720 +                       struct nsproxy *orig,
16721 +                       struct fs_struct *new_fs,
16722 +                       struct user_namespace *new_user,
16723 +                       struct pid_namespace *new_pid)
16724  {
16725         struct nsproxy *new_nsp;
16726         int err;
16727 @@ -66,31 +74,31 @@ static struct nsproxy *create_new_namesp
16728         if (!new_nsp)
16729                 return ERR_PTR(-ENOMEM);
16730
16731 -       new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs);
16732 +       new_nsp->mnt_ns = copy_mnt_ns(flags, orig->mnt_ns, new_fs);
16733         if (IS_ERR(new_nsp->mnt_ns)) {
16734                 err = PTR_ERR(new_nsp->mnt_ns);
16735                 goto out_ns;
16736         }
16737
16738 -       new_nsp->uts_ns = copy_utsname(flags, tsk);
16739 +       new_nsp->uts_ns = copy_utsname(flags, orig->uts_ns, new_user);
16740         if (IS_ERR(new_nsp->uts_ns)) {
16741                 err = PTR_ERR(new_nsp->uts_ns);
16742                 goto out_uts;
16743         }
16744
16745 -       new_nsp->ipc_ns = copy_ipcs(flags, tsk);
16746 +       new_nsp->ipc_ns = copy_ipcs(flags, orig->ipc_ns, new_user);
16747         if (IS_ERR(new_nsp->ipc_ns)) {
16748                 err = PTR_ERR(new_nsp->ipc_ns);
16749                 goto out_ipc;
16750         }
16751
16752 -       new_nsp->pid_ns = copy_pid_ns(flags, task_active_pid_ns(tsk));
16753 +       new_nsp->pid_ns = copy_pid_ns(flags, new_pid);
16754         if (IS_ERR(new_nsp->pid_ns)) {
16755                 err = PTR_ERR(new_nsp->pid_ns);
16756                 goto out_pid;
16757         }
16758
16759 -       new_nsp->net_ns = copy_net_ns(flags, tsk->nsproxy->net_ns);
16760 +       new_nsp->net_ns = copy_net_ns(flags, orig->net_ns);
16761         if (IS_ERR(new_nsp->net_ns)) {
16762                 err = PTR_ERR(new_nsp->net_ns);
16763                 goto out_net;
16764 @@ -115,6 +123,40 @@ out_ns:
16765         return ERR_PTR(err);
16766  }
16767
16768 +static struct nsproxy *create_new_namespaces(unsigned long flags,
16769 +                       struct task_struct *tsk, struct fs_struct *new_fs)
16770 +{
16771 +       return unshare_namespaces(flags, tsk->nsproxy,
16772 +               new_fs, task_cred_xxx(tsk, user)->user_ns,
16773 +               task_active_pid_ns(tsk));
16774 +}
16775 +
16776 +/*
16777 + * copies the nsproxy, setting refcount to 1, and grabbing a
16778 + * reference to all contained namespaces.
16779 + */
16780 +struct nsproxy *copy_nsproxy(struct nsproxy *orig)
16781 +{
16782 +       struct nsproxy *ns = create_nsproxy();
16783 +
16784 +       if (ns) {
16785 +               memcpy(ns, orig, sizeof(struct nsproxy));
16786 +               atomic_set(&ns->count, 1);
16787 +
16788 +               if (ns->mnt_ns)
16789 +                       get_mnt_ns(ns->mnt_ns);
16790 +               if (ns->uts_ns)
16791 +                       get_uts_ns(ns->uts_ns);
16792 +               if (ns->ipc_ns)
16793 +                       get_ipc_ns(ns->ipc_ns);
16794 +               if (ns->pid_ns)
16795 +                       get_pid_ns(ns->pid_ns);
16796 +               if (ns->net_ns)
16797 +                       get_net(ns->net_ns);
16798 +       }
16799 +       return ns;
16800 +}
16801 +
16802  /*
16803   * called from clone.  This now handles copy for nsproxy and all
16804   * namespaces therein.
16805 @@ -122,9 +164,12 @@ out_ns:
16806  int copy_namespaces(unsigned long flags, struct task_struct *tsk)
16807  {
16808         struct nsproxy *old_ns = tsk->nsproxy;
16809 -       struct nsproxy *new_ns;
16810 +       struct nsproxy *new_ns = NULL;
16811         int err = 0;
16812
16813 +       vxdprintk(VXD_CBIT(space, 7), "copy_namespaces(0x%08lx,%p[%p])",
16814 +               flags, tsk, old_ns);
16815 +
16816         if (!old_ns)
16817                 return 0;
16818
16819 @@ -134,7 +179,7 @@ int copy_namespaces(unsigned long flags,
16820                                 CLONE_NEWPID | CLONE_NEWNET)))
16821                 return 0;
16822
16823 -       if (!capable(CAP_SYS_ADMIN)) {
16824 +       if (!vx_can_unshare(CAP_SYS_ADMIN, flags)) {
16825                 err = -EPERM;
16826                 goto out;
16827         }
16828 @@ -161,6 +206,9 @@ int copy_namespaces(unsigned long flags,
16829
16830  out:
16831         put_nsproxy(old_ns);
16832 +       vxdprintk(VXD_CBIT(space, 3),
16833 +               "copy_namespaces(0x%08lx,%p[%p]) = %d [%p]",
16834 +               flags, tsk, old_ns, err, new_ns);
16835         return err;
16836  }
16837
16838 @@ -174,7 +222,9 @@ void free_nsproxy(struct nsproxy *ns)
16839                 put_ipc_ns(ns->ipc_ns);
16840         if (ns->pid_ns)
16841                 put_pid_ns(ns->pid_ns);
16842 -       put_net(ns->net_ns);
16843 +       if (ns->net_ns)
16844 +               put_net(ns->net_ns);
16845 +       atomic_dec(&vs_global_nsproxy);
16846         kmem_cache_free(nsproxy_cachep, ns);
16847  }
16848
16849 @@ -187,11 +237,15 @@ int unshare_nsproxy_namespaces(unsigned
16850  {
16851         int err = 0;
16852
16853 +       vxdprintk(VXD_CBIT(space, 4),
16854 +               "unshare_nsproxy_namespaces(0x%08lx,[%p])",
16855 +               unshare_flags, current->nsproxy);
16856 +
16857         if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC |
16858                                CLONE_NEWNET)))
16859                 return 0;
16860
16861 -       if (!capable(CAP_SYS_ADMIN))
16862 +       if (!vx_can_unshare(CAP_SYS_ADMIN, unshare_flags))
16863                 return -EPERM;
16864
16865         *new_nsp = create_new_namespaces(unshare_flags, current,
16866 diff -NurpP --minimal linux-3.0/kernel/pid.c linux-3.0-vs2.3.1-pre7/kernel/pid.c
16867 --- linux-3.0/kernel/pid.c      2011-05-22 16:17:59.000000000 +0200
16868 +++ linux-3.0-vs2.3.1-pre7/kernel/pid.c 2011-06-10 22:11:24.000000000 +0200
16869 @@ -36,6 +36,7 @@
16870  #include <linux/pid_namespace.h>
16871  #include <linux/init_task.h>
16872  #include <linux/syscalls.h>
16873 +#include <linux/vs_pid.h>
16874
16875  #define pid_hashfn(nr, ns)     \
16876         hash_long((unsigned long)nr + (unsigned long)ns, pidhash_shift)
16877 @@ -342,7 +343,7 @@ EXPORT_SYMBOL_GPL(find_pid_ns);
16878
16879  struct pid *find_vpid(int nr)
16880  {
16881 -       return find_pid_ns(nr, current->nsproxy->pid_ns);
16882 +       return find_pid_ns(vx_rmap_pid(nr), current->nsproxy->pid_ns);
16883  }
16884  EXPORT_SYMBOL_GPL(find_vpid);
16885
16886 @@ -402,6 +403,9 @@ void transfer_pid(struct task_struct *ol
16887  struct task_struct *pid_task(struct pid *pid, enum pid_type type)
16888  {
16889         struct task_struct *result = NULL;
16890 +
16891 +       if (type == PIDTYPE_REALPID)
16892 +               type = PIDTYPE_PID;
16893         if (pid) {
16894                 struct hlist_node *first;
16895                 first = rcu_dereference_check(hlist_first_rcu(&pid->tasks[type]),
16896 @@ -420,7 +424,7 @@ EXPORT_SYMBOL(pid_task);
16897  struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns)
16898  {
16899         rcu_lockdep_assert(rcu_read_lock_held());
16900 -       return pid_task(find_pid_ns(nr, ns), PIDTYPE_PID);
16901 +       return pid_task(find_pid_ns(vx_rmap_pid(nr), ns), PIDTYPE_PID);
16902  }
16903
16904  struct task_struct *find_task_by_vpid(pid_t vnr)
16905 @@ -464,7 +468,7 @@ struct pid *find_get_pid(pid_t nr)
16906  }
16907  EXPORT_SYMBOL_GPL(find_get_pid);
16908
16909 -pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
16910 +pid_t pid_unmapped_nr_ns(struct pid *pid, struct pid_namespace *ns)
16911  {
16912         struct upid *upid;
16913         pid_t nr = 0;
16914 @@ -477,6 +481,11 @@ pid_t pid_nr_ns(struct pid *pid, struct
16915         return nr;
16916  }
16917
16918 +pid_t pid_nr_ns(struct pid *pid, struct pid_namespace *ns)
16919 +{
16920 +       return vx_map_pid(pid_unmapped_nr_ns(pid, ns));
16921 +}
16922 +
16923  pid_t pid_vnr(struct pid *pid)
16924  {
16925         return pid_nr_ns(pid, current->nsproxy->pid_ns);
16926 diff -NurpP --minimal linux-3.0/kernel/pid_namespace.c linux-3.0-vs2.3.1-pre7/kernel/pid_namespace.c
16927 --- linux-3.0/kernel/pid_namespace.c    2011-05-22 16:17:59.000000000 +0200
16928 +++ linux-3.0-vs2.3.1-pre7/kernel/pid_namespace.c       2011-06-10 22:17:45.000000000 +0200
16929 @@ -15,6 +15,7 @@
16930  #include <linux/acct.h>
16931  #include <linux/slab.h>
16932  #include <linux/proc_fs.h>
16933 +#include <linux/vserver/global.h>
16934
16935  #define BITS_PER_PAGE          (PAGE_SIZE*8)
16936
16937 @@ -88,6 +89,7 @@ static struct pid_namespace *create_pid_
16938                 goto out_free_map;
16939
16940         kref_init(&ns->kref);
16941 +       atomic_inc(&vs_global_pid_ns);
16942         ns->level = level;
16943         ns->parent = get_pid_ns(parent_pid_ns);
16944
16945 @@ -119,6 +121,7 @@ static void destroy_pid_namespace(struct
16946
16947         for (i = 0; i < PIDMAP_ENTRIES; i++)
16948                 kfree(ns->pidmap[i].page);
16949 +       atomic_dec(&vs_global_pid_ns);
16950         kmem_cache_free(pid_ns_cachep, ns);
16951  }
16952
16953 diff -NurpP --minimal linux-3.0/kernel/posix-timers.c linux-3.0-vs2.3.1-pre7/kernel/posix-timers.c
16954 --- linux-3.0/kernel/posix-timers.c     2011-07-22 11:18:12.000000000 +0200
16955 +++ linux-3.0-vs2.3.1-pre7/kernel/posix-timers.c        2011-06-10 22:11:24.000000000 +0200
16956 @@ -47,6 +47,7 @@
16957  #include <linux/wait.h>
16958  #include <linux/workqueue.h>
16959  #include <linux/module.h>
16960 +#include <linux/vs_context.h>
16961
16962  /*
16963   * Management arrays for POSIX timers.  Timers are kept in slab memory
16964 @@ -340,6 +341,7 @@ int posix_timer_event(struct k_itimer *t
16965  {
16966         struct task_struct *task;
16967         int shared, ret = -1;
16968 +
16969         /*
16970          * FIXME: if ->sigq is queued we can race with
16971          * dequeue_signal()->do_schedule_next_timer().
16972 @@ -356,10 +358,18 @@ int posix_timer_event(struct k_itimer *t
16973         rcu_read_lock();
16974         task = pid_task(timr->it_pid, PIDTYPE_PID);
16975         if (task) {
16976 +               struct vx_info_save vxis;
16977 +               struct vx_info *vxi;
16978 +
16979 +               vxi = get_vx_info(task->vx_info);
16980 +               enter_vx_info(vxi, &vxis);
16981                 shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID);
16982                 ret = send_sigqueue(timr->sigq, task, shared);
16983 +               leave_vx_info(&vxis);
16984 +               put_vx_info(vxi);
16985         }
16986         rcu_read_unlock();
16987 +
16988         /* If we failed to send the signal the timer stops. */
16989         return ret > 0;
16990  }
16991 diff -NurpP --minimal linux-3.0/kernel/printk.c linux-3.0-vs2.3.1-pre7/kernel/printk.c
16992 --- linux-3.0/kernel/printk.c   2011-07-22 11:18:12.000000000 +0200
16993 +++ linux-3.0-vs2.3.1-pre7/kernel/printk.c      2011-06-10 22:11:24.000000000 +0200
16994 @@ -41,6 +41,7 @@
16995  #include <linux/cpu.h>
16996  #include <linux/notifier.h>
16997  #include <linux/rculist.h>
16998 +#include <linux/vs_cvirt.h>
16999
17000  #include <asm/uaccess.h>
17001
17002 @@ -314,7 +315,7 @@ static int check_syslog_permissions(int
17003                 return 0;
17004
17005         if (syslog_action_restricted(type)) {
17006 -               if (capable(CAP_SYSLOG))
17007 +               if (vx_capable(CAP_SYSLOG, VXC_SYSLOG))
17008                         return 0;
17009                 /* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */
17010                 if (capable(CAP_SYS_ADMIN)) {
17011 @@ -342,12 +343,9 @@ int do_syslog(int type, char __user *buf
17012         if (error)
17013                 return error;
17014
17015 -       switch (type) {
17016 -       case SYSLOG_ACTION_CLOSE:       /* Close log */
17017 -               break;
17018 -       case SYSLOG_ACTION_OPEN:        /* Open log */
17019 -               break;
17020 -       case SYSLOG_ACTION_READ:        /* Read from log */
17021 +       if ((type == SYSLOG_ACTION_READ) ||
17022 +           (type == SYSLOG_ACTION_READ_ALL) ||
17023 +           (type == SYSLOG_ACTION_READ_CLEAR)) {
17024                 error = -EINVAL;
17025                 if (!buf || len < 0)
17026                         goto out;
17027 @@ -358,6 +356,16 @@ int do_syslog(int type, char __user *buf
17028                         error = -EFAULT;
17029                         goto out;
17030                 }
17031 +       }
17032 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
17033 +               return vx_do_syslog(type, buf, len);
17034 +
17035 +       switch (type) {
17036 +       case SYSLOG_ACTION_CLOSE:       /* Close log */
17037 +               break;
17038 +       case SYSLOG_ACTION_OPEN:        /* Open log */
17039 +               break;
17040 +       case SYSLOG_ACTION_READ:        /* Read from log */
17041                 error = wait_event_interruptible(log_wait,
17042                                                         (log_start - log_end));
17043                 if (error)
17044 @@ -384,16 +392,6 @@ int do_syslog(int type, char __user *buf
17045                 /* FALL THRU */
17046         /* Read last kernel messages */
17047         case SYSLOG_ACTION_READ_ALL:
17048 -               error = -EINVAL;
17049 -               if (!buf || len < 0)
17050 -                       goto out;
17051 -               error = 0;
17052 -               if (!len)
17053 -                       goto out;
17054 -               if (!access_ok(VERIFY_WRITE, buf, len)) {
17055 -                       error = -EFAULT;
17056 -                       goto out;
17057 -               }
17058                 count = len;
17059                 if (count > log_buf_len)
17060                         count = log_buf_len;
17061 diff -NurpP --minimal linux-3.0/kernel/ptrace.c linux-3.0-vs2.3.1-pre7/kernel/ptrace.c
17062 --- linux-3.0/kernel/ptrace.c   2011-07-22 11:18:12.000000000 +0200
17063 +++ linux-3.0-vs2.3.1-pre7/kernel/ptrace.c      2011-06-10 22:11:24.000000000 +0200
17064 @@ -22,6 +22,7 @@
17065  #include <linux/syscalls.h>
17066  #include <linux/uaccess.h>
17067  #include <linux/regset.h>
17068 +#include <linux/vs_context.h>
17069  #include <linux/hw_breakpoint.h>
17070
17071
17072 @@ -169,6 +170,11 @@ ok:
17073                 dumpable = get_dumpable(task->mm);
17074         if (!dumpable && !task_ns_capable(task, CAP_SYS_PTRACE))
17075                 return -EPERM;
17076 +       if (!vx_check(task->xid, VS_ADMIN_P|VS_WATCH_P|VS_IDENT))
17077 +               return -EPERM;
17078 +       if (!vx_check(task->xid, VS_IDENT) &&
17079 +               !task_vx_flags(task, VXF_STATE_ADMIN, 0))
17080 +               return -EACCES;
17081
17082         return security_ptrace_access_check(task, mode);
17083  }
17084 diff -NurpP --minimal linux-3.0/kernel/sched.c linux-3.0-vs2.3.1-pre7/kernel/sched.c
17085 --- linux-3.0/kernel/sched.c    2011-07-22 11:18:12.000000000 +0200
17086 +++ linux-3.0-vs2.3.1-pre7/kernel/sched.c       2011-07-22 11:20:39.000000000 +0200
17087 @@ -71,6 +71,8 @@
17088  #include <linux/ctype.h>
17089  #include <linux/ftrace.h>
17090  #include <linux/slab.h>
17091 +#include <linux/vs_sched.h>
17092 +#include <linux/vs_cvirt.h>
17093
17094  #include <asm/tlb.h>
17095  #include <asm/irq_regs.h>
17096 @@ -3461,9 +3463,17 @@ static void calc_global_nohz(unsigned lo
17097   */
17098  void get_avenrun(unsigned long *loads, unsigned long offset, int shift)
17099  {
17100 -       loads[0] = (avenrun[0] + offset) << shift;
17101 -       loads[1] = (avenrun[1] + offset) << shift;
17102 -       loads[2] = (avenrun[2] + offset) << shift;
17103 +       if (vx_flags(VXF_VIRT_LOAD, 0)) {
17104 +               struct vx_info *vxi = current_vx_info();
17105 +
17106 +               loads[0] = (vxi->cvirt.load[0] + offset) << shift;
17107 +               loads[1] = (vxi->cvirt.load[1] + offset) << shift;
17108 +               loads[2] = (vxi->cvirt.load[2] + offset) << shift;
17109 +       } else {
17110 +               loads[0] = (avenrun[0] + offset) << shift;
17111 +               loads[1] = (avenrun[1] + offset) << shift;
17112 +               loads[2] = (avenrun[2] + offset) << shift;
17113 +       }
17114  }
17115
17116  /*
17117 @@ -3746,16 +3756,19 @@ void account_user_time(struct task_struc
17118                        cputime_t cputime_scaled)
17119  {
17120         struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
17121 +       struct vx_info *vxi = p->vx_info;  /* p is _always_ current */
17122         cputime64_t tmp;
17123 +       int nice = (TASK_NICE(p) > 0);
17124
17125         /* Add user time to process. */
17126         p->utime = cputime_add(p->utime, cputime);
17127         p->utimescaled = cputime_add(p->utimescaled, cputime_scaled);
17128 +       vx_account_user(vxi, cputime, nice);
17129         account_group_user_time(p, cputime);
17130
17131         /* Add user time to cpustat. */
17132         tmp = cputime_to_cputime64(cputime);
17133 -       if (TASK_NICE(p) > 0)
17134 +       if (nice)
17135                 cpustat->nice = cputime64_add(cpustat->nice, tmp);
17136         else
17137                 cpustat->user = cputime64_add(cpustat->user, tmp);
17138 @@ -3807,10 +3820,12 @@ void __account_system_time(struct task_s
17139                         cputime_t cputime_scaled, cputime64_t *target_cputime64)
17140  {
17141         cputime64_t tmp = cputime_to_cputime64(cputime);
17142 +       struct vx_info *vxi = p->vx_info;  /* p is _always_ current */
17143
17144         /* Add system time to process. */
17145         p->stime = cputime_add(p->stime, cputime);
17146         p->stimescaled = cputime_add(p->stimescaled, cputime_scaled);
17147 +       vx_account_system(vxi, cputime, 0 /* do we have idle time? */);
17148         account_group_system_time(p, cputime);
17149
17150         /* Add system time to cpustat. */
17151 @@ -4968,7 +4983,7 @@ SYSCALL_DEFINE1(nice, int, increment)
17152                 nice = 19;
17153
17154         if (increment < 0 && !can_nice(current, nice))
17155 -               return -EPERM;
17156 +               return vx_flags(VXF_IGNEG_NICE, 0) ? 0 : -EPERM;
17157
17158         retval = security_task_setnice(current, nice);
17159         if (retval)
17160 diff -NurpP --minimal linux-3.0/kernel/sched_fair.c linux-3.0-vs2.3.1-pre7/kernel/sched_fair.c
17161 --- linux-3.0/kernel/sched_fair.c       2011-07-22 11:18:12.000000000 +0200
17162 +++ linux-3.0-vs2.3.1-pre7/kernel/sched_fair.c  2011-07-22 11:20:39.000000000 +0200
17163 @@ -998,6 +998,8 @@ enqueue_entity(struct cfs_rq *cfs_rq, st
17164                 __enqueue_entity(cfs_rq, se);
17165         se->on_rq = 1;
17166
17167 +       if (entity_is_task(se))
17168 +               vx_activate_task(task_of(se));
17169         if (cfs_rq->nr_running == 1)
17170                 list_add_leaf_cfs_rq(cfs_rq);
17171  }
17172 @@ -1074,6 +1076,8 @@ dequeue_entity(struct cfs_rq *cfs_rq, st
17173         if (se != cfs_rq->curr)
17174                 __dequeue_entity(cfs_rq, se);
17175         se->on_rq = 0;
17176 +       if (entity_is_task(se))
17177 +               vx_deactivate_task(task_of(se));
17178         update_cfs_load(cfs_rq, 0);
17179         account_entity_dequeue(cfs_rq, se);
17180
17181 diff -NurpP --minimal linux-3.0/kernel/signal.c linux-3.0-vs2.3.1-pre7/kernel/signal.c
17182 --- linux-3.0/kernel/signal.c   2011-07-22 11:18:12.000000000 +0200
17183 +++ linux-3.0-vs2.3.1-pre7/kernel/signal.c      2011-07-22 11:20:39.000000000 +0200
17184 @@ -28,6 +28,8 @@
17185  #include <linux/freezer.h>
17186  #include <linux/pid_namespace.h>
17187  #include <linux/nsproxy.h>
17188 +#include <linux/vs_context.h>
17189 +#include <linux/vs_pid.h>
17190  #define CREATE_TRACE_POINTS
17191  #include <trace/events/signal.h>
17192
17193 @@ -744,9 +746,18 @@ static int check_kill_permission(int sig
17194         struct pid *sid;
17195         int error;
17196
17197 +       vxdprintk(VXD_CBIT(misc, 7),
17198 +               "check_kill_permission(%d,%p,%p[#%u,%u])",
17199 +               sig, info, t, vx_task_xid(t), t->pid);
17200 +
17201         if (!valid_signal(sig))
17202                 return -EINVAL;
17203
17204 +/*     FIXME: needed? if so, why?
17205 +       if ((info != SEND_SIG_NOINFO) &&
17206 +               (is_si_special(info) || !si_fromuser(info)))
17207 +               goto skip;      */
17208 +
17209         if (!si_fromuser(info))
17210                 return 0;
17211
17212 @@ -770,6 +781,20 @@ static int check_kill_permission(int sig
17213                 }
17214         }
17215
17216 +       error = -EPERM;
17217 +       if (t->pid == 1 && current->xid)
17218 +               return error;
17219 +
17220 +       error = -ESRCH;
17221 +       /* FIXME: we shouldn't return ESRCH ever, to avoid
17222 +                 loops, maybe ENOENT or EACCES? */
17223 +       if (!vx_check(vx_task_xid(t), VS_WATCH_P | VS_IDENT)) {
17224 +               vxdprintk(current->xid || VXD_CBIT(misc, 7),
17225 +                       "signal %d[%p] xid mismatch %p[#%u,%u] xid=#%u",
17226 +                       sig, info, t, vx_task_xid(t), t->pid, current->xid);
17227 +               return error;
17228 +       }
17229 +/* skip: */
17230         return security_task_kill(t, info, sig, 0);
17231  }
17232
17233 @@ -1246,7 +1271,7 @@ int kill_pid_info(int sig, struct siginf
17234         rcu_read_lock();
17235  retry:
17236         p = pid_task(pid, PIDTYPE_PID);
17237 -       if (p) {
17238 +       if (p && vx_check(vx_task_xid(p), VS_IDENT)) {
17239                 error = group_send_sig_info(sig, info, p);
17240                 if (unlikely(error == -ESRCH))
17241                         /*
17242 @@ -1285,7 +1310,7 @@ int kill_pid_info_as_uid(int sig, struct
17243
17244         rcu_read_lock();
17245         p = pid_task(pid, PIDTYPE_PID);
17246 -       if (!p) {
17247 +       if (!p || !vx_check(vx_task_xid(p), VS_IDENT)) {
17248                 ret = -ESRCH;
17249                 goto out_unlock;
17250         }
17251 @@ -1340,8 +1365,10 @@ static int kill_something_info(int sig,
17252                 struct task_struct * p;
17253
17254                 for_each_process(p) {
17255 -                       if (task_pid_vnr(p) > 1 &&
17256 -                                       !same_thread_group(p, current)) {
17257 +                       if (vx_check(vx_task_xid(p), VS_ADMIN|VS_IDENT) &&
17258 +                               task_pid_vnr(p) > 1 &&
17259 +                               !same_thread_group(p, current) &&
17260 +                               !vx_current_initpid(p->pid)) {
17261                                 int err = group_send_sig_info(sig, info, p);
17262                                 ++count;
17263                                 if (err != -EPERM)
17264 @@ -2140,6 +2167,11 @@ relock:
17265                                 !sig_kernel_only(signr))
17266                         continue;
17267
17268 +               /* virtual init is protected against user signals */
17269 +               if ((info->si_code == SI_USER) &&
17270 +                       vx_current_initpid(current->pid))
17271 +                       continue;
17272 +
17273                 if (sig_kernel_stop(signr)) {
17274                         /*
17275                          * The default action is to stop all threads in
17276 diff -NurpP --minimal linux-3.0/kernel/softirq.c linux-3.0-vs2.3.1-pre7/kernel/softirq.c
17277 --- linux-3.0/kernel/softirq.c  2011-07-22 11:18:12.000000000 +0200
17278 +++ linux-3.0-vs2.3.1-pre7/kernel/softirq.c     2011-07-22 11:20:39.000000000 +0200
17279 @@ -24,6 +24,7 @@
17280  #include <linux/ftrace.h>
17281  #include <linux/smp.h>
17282  #include <linux/tick.h>
17283 +#include <linux/vs_context.h>
17284
17285  #define CREATE_TRACE_POINTS
17286  #include <trace/events/irq.h>
17287 diff -NurpP --minimal linux-3.0/kernel/sys.c linux-3.0-vs2.3.1-pre7/kernel/sys.c
17288 --- linux-3.0/kernel/sys.c      2011-07-22 11:18:12.000000000 +0200
17289 +++ linux-3.0-vs2.3.1-pre7/kernel/sys.c 2011-07-23 15:55:50.000000000 +0200
17290 @@ -43,6 +43,7 @@
17291  #include <linux/syscalls.h>
17292  #include <linux/kprobes.h>
17293  #include <linux/user_namespace.h>
17294 +#include <linux/vs_pid.h>
17295
17296  #include <linux/kmsg_dump.h>
17297
17298 @@ -151,7 +152,10 @@ static int set_one_prio(struct task_stru
17299                 goto out;
17300         }
17301         if (niceval < task_nice(p) && !can_nice(p, niceval)) {
17302 -               error = -EACCES;
17303 +               if (vx_flags(VXF_IGNEG_NICE, 0))
17304 +                       error = 0;
17305 +               else
17306 +                       error = -EACCES;
17307                 goto out;
17308         }
17309         no_nice = security_task_setnice(p, niceval);
17310 @@ -201,6 +205,8 @@ SYSCALL_DEFINE3(setpriority, int, which,
17311                         else
17312                                 pgrp = task_pgrp(current);
17313                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
17314 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
17315 +                                       continue;
17316                                 error = set_one_prio(p, niceval, error);
17317                         } while_each_pid_thread(pgrp, PIDTYPE_PGID, p);
17318                         break;
17319 @@ -264,6 +270,8 @@ SYSCALL_DEFINE2(getpriority, int, which,
17320                         else
17321                                 pgrp = task_pgrp(current);
17322                         do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
17323 +                               if (!vx_check(p->xid, VS_ADMIN_P | VS_IDENT))
17324 +                                       continue;
17325                                 niceval = 20 - task_nice(p);
17326                                 if (niceval > retval)
17327                                         retval = niceval;
17328 @@ -383,6 +391,8 @@ EXPORT_SYMBOL_GPL(kernel_power_off);
17329
17330  static DEFINE_MUTEX(reboot_mutex);
17331
17332 +long vs_reboot(unsigned int, void __user *);
17333 +
17334  /*
17335   * Reboot system call: for obvious reasons only root may call it,
17336   * and even root needs to set up some magic numbers in the registers
17337 @@ -415,6 +425,9 @@ SYSCALL_DEFINE4(reboot, int, magic1, int
17338         if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off)
17339                 cmd = LINUX_REBOOT_CMD_HALT;
17340
17341 +       if (!vx_check(0, VS_ADMIN|VS_WATCH))
17342 +               return vs_reboot(cmd, arg);
17343 +
17344         mutex_lock(&reboot_mutex);
17345         switch (cmd) {
17346         case LINUX_REBOOT_CMD_RESTART:
17347 @@ -1197,7 +1210,8 @@ SYSCALL_DEFINE2(sethostname, char __user
17348         int errno;
17349         char tmp[__NEW_UTS_LEN];
17350
17351 -       if (!ns_capable(current->nsproxy->uts_ns->user_ns, CAP_SYS_ADMIN))
17352 +       if (!vx_ns_capable(current->nsproxy->uts_ns->user_ns,
17353 +               CAP_SYS_ADMIN, VXC_SET_UTSNAME))
17354                 return -EPERM;
17355
17356         if (len < 0 || len > __NEW_UTS_LEN)
17357 @@ -1247,7 +1261,8 @@ SYSCALL_DEFINE2(setdomainname, char __us
17358         int errno;
17359         char tmp[__NEW_UTS_LEN];
17360
17361 -       if (!ns_capable(current->nsproxy->uts_ns->user_ns, CAP_SYS_ADMIN))
17362 +       if (!vx_ns_capable(current->nsproxy->uts_ns->user_ns,
17363 +               CAP_SYS_ADMIN, VXC_SET_UTSNAME))
17364                 return -EPERM;
17365         if (len < 0 || len > __NEW_UTS_LEN)
17366                 return -EINVAL;
17367 @@ -1365,7 +1380,7 @@ int do_prlimit(struct task_struct *tsk,
17368                 /* Keep the capable check against init_user_ns until
17369                    cgroups can contain all limits */
17370                 if (new_rlim->rlim_max > rlim->rlim_max &&
17371 -                               !capable(CAP_SYS_RESOURCE))
17372 +                       !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT))
17373                         retval = -EPERM;
17374                 if (!retval)
17375                         retval = security_task_setrlimit(tsk->group_leader,
17376 @@ -1419,7 +1434,8 @@ static int check_prlimit_permission(stru
17377              cred->gid == tcred->sgid &&
17378              cred->gid == tcred->gid))
17379                 return 0;
17380 -       if (ns_capable(tcred->user->user_ns, CAP_SYS_RESOURCE))
17381 +       if (vx_ns_capable(tcred->user->user_ns,
17382 +               CAP_SYS_RESOURCE, VXC_SET_RLIMIT))
17383                 return 0;
17384
17385         return -EPERM;
17386 diff -NurpP --minimal linux-3.0/kernel/sysctl.c linux-3.0-vs2.3.1-pre7/kernel/sysctl.c
17387 --- linux-3.0/kernel/sysctl.c   2011-07-22 11:18:12.000000000 +0200
17388 +++ linux-3.0-vs2.3.1-pre7/kernel/sysctl.c      2011-06-15 02:40:14.000000000 +0200
17389 @@ -75,6 +75,7 @@
17390  #if defined(CONFIG_PROVE_LOCKING) || defined(CONFIG_LOCK_STAT)
17391  #include <linux/lockdep.h>
17392  #endif
17393 +extern char vshelper_path[];
17394  #ifdef CONFIG_CHR_DEV_SG
17395  #include <scsi/sg.h>
17396  #endif
17397 @@ -568,6 +569,13 @@ static struct ctl_table kern_table[] = {
17398                 .proc_handler   = proc_dostring,
17399         },
17400  #endif
17401 +       {
17402 +               .procname       = "vshelper",
17403 +               .data           = &vshelper_path,
17404 +               .maxlen         = 256,
17405 +               .mode           = 0644,
17406 +               .proc_handler   = &proc_dostring,
17407 +       },
17408  #ifdef CONFIG_CHR_DEV_SG
17409         {
17410                 .procname       = "sg-big-buff",
17411 diff -NurpP --minimal linux-3.0/kernel/sysctl_binary.c linux-3.0-vs2.3.1-pre7/kernel/sysctl_binary.c
17412 --- linux-3.0/kernel/sysctl_binary.c    2011-05-22 16:17:59.000000000 +0200
17413 +++ linux-3.0-vs2.3.1-pre7/kernel/sysctl_binary.c       2011-06-10 22:11:24.000000000 +0200
17414 @@ -73,6 +73,7 @@ static const struct bin_table bin_kern_t
17415
17416         { CTL_INT,      KERN_PANIC,                     "panic" },
17417         { CTL_INT,      KERN_REALROOTDEV,               "real-root-dev" },
17418 +       { CTL_STR,      KERN_VSHELPER,                  "vshelper" },
17419
17420         { CTL_STR,      KERN_SPARC_REBOOT,              "reboot-cmd" },
17421         { CTL_INT,      KERN_CTLALTDEL,                 "ctrl-alt-del" },
17422 diff -NurpP --minimal linux-3.0/kernel/time/timekeeping.c linux-3.0-vs2.3.1-pre7/kernel/time/timekeeping.c
17423 --- linux-3.0/kernel/time/timekeeping.c 2011-07-22 11:18:12.000000000 +0200
17424 +++ linux-3.0-vs2.3.1-pre7/kernel/time/timekeeping.c    2011-06-10 22:11:24.000000000 +0200
17425 @@ -233,6 +233,7 @@ void getnstimeofday(struct timespec *ts)
17426         } while (read_seqretry(&xtime_lock, seq));
17427
17428         timespec_add_ns(ts, nsecs);
17429 +       vx_adjust_timespec(ts);
17430  }
17431
17432  EXPORT_SYMBOL(getnstimeofday);
17433 diff -NurpP --minimal linux-3.0/kernel/time.c linux-3.0-vs2.3.1-pre7/kernel/time.c
17434 --- linux-3.0/kernel/time.c     2011-05-22 16:17:59.000000000 +0200
17435 +++ linux-3.0-vs2.3.1-pre7/kernel/time.c        2011-06-10 22:11:24.000000000 +0200
17436 @@ -92,7 +92,7 @@ SYSCALL_DEFINE1(stime, time_t __user *,
17437         if (err)
17438                 return err;
17439
17440 -       do_settimeofday(&tv);
17441 +       vx_settimeofday(&tv);
17442         return 0;
17443  }
17444
17445 @@ -177,7 +177,7 @@ int do_sys_settimeofday(const struct tim
17446                 /* SMP safe, again the code in arch/foo/time.c should
17447                  * globally block out interrupts when it runs.
17448                  */
17449 -               return do_settimeofday(tv);
17450 +               return vx_settimeofday(tv);
17451         }
17452         return 0;
17453  }
17454 diff -NurpP --minimal linux-3.0/kernel/timer.c linux-3.0-vs2.3.1-pre7/kernel/timer.c
17455 --- linux-3.0/kernel/timer.c    2011-07-22 11:18:12.000000000 +0200
17456 +++ linux-3.0-vs2.3.1-pre7/kernel/timer.c       2011-06-15 02:40:14.000000000 +0200
17457 @@ -40,6 +40,10 @@
17458  #include <linux/irq_work.h>
17459  #include <linux/sched.h>
17460  #include <linux/slab.h>
17461 +#include <linux/vs_base.h>
17462 +#include <linux/vs_cvirt.h>
17463 +#include <linux/vs_pid.h>
17464 +#include <linux/vserver/sched.h>
17465
17466  #include <asm/uaccess.h>
17467  #include <asm/unistd.h>
17468 @@ -1336,12 +1340,6 @@ SYSCALL_DEFINE1(alarm, unsigned int, sec
17469
17470  #endif
17471
17472 -#ifndef __alpha__
17473 -
17474 -/*
17475 - * The Alpha uses getxpid, getxuid, and getxgid instead.  Maybe this
17476 - * should be moved into arch/i386 instead?
17477 - */
17478
17479  /**
17480   * sys_getpid - return the thread group id of the current process
17481 @@ -1370,10 +1368,23 @@ SYSCALL_DEFINE0(getppid)
17482         rcu_read_lock();
17483         pid = task_tgid_vnr(current->real_parent);
17484         rcu_read_unlock();
17485 +       return vx_map_pid(pid);
17486 +}
17487
17488 -       return pid;
17489 +#ifdef __alpha__
17490 +
17491 +/*
17492 + * The Alpha uses getxpid, getxuid, and getxgid instead.
17493 + */
17494 +
17495 +asmlinkage long do_getxpid(long *ppid)
17496 +{
17497 +       *ppid = sys_getppid();
17498 +       return sys_getpid();
17499  }
17500
17501 +#else /* _alpha_ */
17502 +
17503  SYSCALL_DEFINE0(getuid)
17504  {
17505         /* Only we change this so SMP safe */
17506 diff -NurpP --minimal linux-3.0/kernel/user_namespace.c linux-3.0-vs2.3.1-pre7/kernel/user_namespace.c
17507 --- linux-3.0/kernel/user_namespace.c   2011-03-15 18:07:42.000000000 +0100
17508 +++ linux-3.0-vs2.3.1-pre7/kernel/user_namespace.c      2011-06-10 22:11:24.000000000 +0200
17509 @@ -11,6 +11,7 @@
17510  #include <linux/user_namespace.h>
17511  #include <linux/highuid.h>
17512  #include <linux/cred.h>
17513 +#include <linux/vserver/global.h>
17514
17515  static struct kmem_cache *user_ns_cachep __read_mostly;
17516
17517 @@ -33,6 +34,7 @@ int create_user_ns(struct cred *new)
17518                 return -ENOMEM;
17519
17520         kref_init(&ns->kref);
17521 +       atomic_inc(&vs_global_user_ns);
17522
17523         for (n = 0; n < UIDHASH_SZ; ++n)
17524                 INIT_HLIST_HEAD(ns->uidhash_table + n);
17525 @@ -81,6 +83,8 @@ void free_user_ns(struct kref *kref)
17526         struct user_namespace *ns =
17527                 container_of(kref, struct user_namespace, kref);
17528
17529 +       /* FIXME: maybe move into destroyer? */
17530 +       atomic_dec(&vs_global_user_ns);
17531         INIT_WORK(&ns->destroyer, free_user_ns_work);
17532         schedule_work(&ns->destroyer);
17533  }
17534 diff -NurpP --minimal linux-3.0/kernel/utsname.c linux-3.0-vs2.3.1-pre7/kernel/utsname.c
17535 --- linux-3.0/kernel/utsname.c  2011-07-22 11:18:12.000000000 +0200
17536 +++ linux-3.0-vs2.3.1-pre7/kernel/utsname.c     2011-06-13 14:09:44.000000000 +0200
17537 @@ -16,14 +16,17 @@
17538  #include <linux/slab.h>
17539  #include <linux/user_namespace.h>
17540  #include <linux/proc_fs.h>
17541 +#include <linux/vserver/global.h>
17542
17543  static struct uts_namespace *create_uts_ns(void)
17544  {
17545         struct uts_namespace *uts_ns;
17546
17547         uts_ns = kmalloc(sizeof(struct uts_namespace), GFP_KERNEL);
17548 -       if (uts_ns)
17549 +       if (uts_ns) {
17550                 kref_init(&uts_ns->kref);
17551 +               atomic_inc(&vs_global_uts_ns);
17552 +       }
17553         return uts_ns;
17554  }
17555
17556 @@ -32,8 +35,8 @@ static struct uts_namespace *create_uts_
17557   * @old_ns: namespace to clone
17558   * Return NULL on error (failure to kmalloc), new ns otherwise
17559   */
17560 -static struct uts_namespace *clone_uts_ns(struct task_struct *tsk,
17561 -                                         struct uts_namespace *old_ns)
17562 +static struct uts_namespace *clone_uts_ns(struct uts_namespace *old_ns,
17563 +                                         struct user_namespace *old_user)
17564  {
17565         struct uts_namespace *ns;
17566
17567 @@ -43,7 +46,7 @@ static struct uts_namespace *clone_uts_n
17568
17569         down_read(&uts_sem);
17570         memcpy(&ns->name, &old_ns->name, sizeof(ns->name));
17571 -       ns->user_ns = get_user_ns(task_cred_xxx(tsk, user)->user_ns);
17572 +       ns->user_ns = get_user_ns(old_user);
17573         up_read(&uts_sem);
17574         return ns;
17575  }
17576 @@ -55,9 +58,9 @@ static struct uts_namespace *clone_uts_n
17577   * versa.
17578   */
17579  struct uts_namespace *copy_utsname(unsigned long flags,
17580 -                                  struct task_struct *tsk)
17581 +                                  struct uts_namespace *old_ns,
17582 +                                  struct user_namespace *user_ns)
17583  {
17584 -       struct uts_namespace *old_ns = tsk->nsproxy->uts_ns;
17585         struct uts_namespace *new_ns;
17586
17587         BUG_ON(!old_ns);
17588 @@ -66,7 +69,7 @@ struct uts_namespace *copy_utsname(unsig
17589         if (!(flags & CLONE_NEWUTS))
17590                 return old_ns;
17591
17592 -       new_ns = clone_uts_ns(tsk, old_ns);
17593 +       new_ns = clone_uts_ns(old_ns, user_ns);
17594
17595         put_uts_ns(old_ns);
17596         return new_ns;
17597 @@ -78,6 +81,7 @@ void free_uts_ns(struct kref *kref)
17598
17599         ns = container_of(kref, struct uts_namespace, kref);
17600         put_user_ns(ns->user_ns);
17601 +       atomic_dec(&vs_global_uts_ns);
17602         kfree(ns);
17603  }
17604
17605 diff -NurpP --minimal linux-3.0/kernel/vserver/Kconfig linux-3.0-vs2.3.1-pre7/kernel/vserver/Kconfig
17606 --- linux-3.0/kernel/vserver/Kconfig    1970-01-01 01:00:00.000000000 +0100
17607 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/Kconfig       2011-07-20 02:15:23.000000000 +0200
17608 @@ -0,0 +1,228 @@
17609 +#
17610 +# Linux VServer configuration
17611 +#
17612 +
17613 +menu "Linux VServer"
17614 +
17615 +config VSERVER_AUTO_LBACK
17616 +       bool    "Automatically Assign Loopback IP"
17617 +       default y
17618 +       help
17619 +         Automatically assign a guest specific loopback
17620 +         IP and add it to the kernel network stack on
17621 +         startup.
17622 +
17623 +config VSERVER_AUTO_SINGLE
17624 +       bool    "Automatic Single IP Special Casing"
17625 +       depends on EXPERIMENTAL
17626 +       default y
17627 +       help
17628 +         This allows network contexts with a single IP to
17629 +         automatically remap 0.0.0.0 bindings to that IP,
17630 +         avoiding further network checks and improving
17631 +         performance.
17632 +
17633 +         (note: such guests do not allow to change the ip
17634 +          on the fly and do not show loopback addresses)
17635 +
17636 +config VSERVER_COWBL
17637 +       bool    "Enable COW Immutable Link Breaking"
17638 +       default y
17639 +       help
17640 +         This enables the COW (Copy-On-Write) link break code.
17641 +         It allows you to treat unified files like normal files
17642 +         when writing to them (which will implicitely break the
17643 +         link and create a copy of the unified file)
17644 +
17645 +config VSERVER_VTIME
17646 +       bool    "Enable Virtualized Guest Time"
17647 +       depends on EXPERIMENTAL
17648 +       default n
17649 +       help
17650 +         This enables per guest time offsets to allow for
17651 +         adjusting the system clock individually per guest.
17652 +         this adds some overhead to the time functions and
17653 +         therefore should not be enabled without good reason.
17654 +
17655 +config VSERVER_DEVICE
17656 +       bool    "Enable Guest Device Mapping"
17657 +       depends on EXPERIMENTAL
17658 +       default n
17659 +       help
17660 +         This enables generic device remapping.
17661 +
17662 +config VSERVER_PROC_SECURE
17663 +       bool    "Enable Proc Security"
17664 +       depends on PROC_FS
17665 +       default y
17666 +       help
17667 +         This configures ProcFS security to initially hide
17668 +         non-process entries for all contexts except the main and
17669 +         spectator context (i.e. for all guests), which is a secure
17670 +         default.
17671 +
17672 +         (note: on 1.2x the entries were visible by default)
17673 +
17674 +choice
17675 +       prompt  "Persistent Inode Tagging"
17676 +       default TAGGING_ID24
17677 +       help
17678 +         This adds persistent context information to filesystems
17679 +         mounted with the tagxid option. Tagging is a requirement
17680 +         for per-context disk limits and per-context quota.
17681 +
17682 +
17683 +config TAGGING_NONE
17684 +       bool    "Disabled"
17685 +       help
17686 +         do not store per-context information in inodes.
17687 +
17688 +config TAGGING_UID16
17689 +       bool    "UID16/GID32"
17690 +       help
17691 +         reduces UID to 16 bit, but leaves GID at 32 bit.
17692 +
17693 +config TAGGING_GID16
17694 +       bool    "UID32/GID16"
17695 +       help
17696 +         reduces GID to 16 bit, but leaves UID at 32 bit.
17697 +
17698 +config TAGGING_ID24
17699 +       bool    "UID24/GID24"
17700 +       help
17701 +         uses the upper 8bit from UID and GID for XID tagging
17702 +         which leaves 24bit for UID/GID each, which should be
17703 +         more than sufficient for normal use.
17704 +
17705 +config TAGGING_INTERN
17706 +       bool    "UID32/GID32"
17707 +       help
17708 +         this uses otherwise reserved inode fields in the on
17709 +         disk representation, which limits the use to a few
17710 +         filesystems (currently ext2 and ext3)
17711 +
17712 +endchoice
17713 +
17714 +config TAG_NFSD
17715 +       bool    "Tag NFSD User Auth and Files"
17716 +       default n
17717 +       help
17718 +         Enable this if you do want the in-kernel NFS
17719 +         Server to use the tagging specified above.
17720 +         (will require patched clients too)
17721 +
17722 +config VSERVER_PRIVACY
17723 +       bool    "Honor Privacy Aspects of Guests"
17724 +       default n
17725 +       help
17726 +         When enabled, most context checks will disallow
17727 +         access to structures assigned to a specific context,
17728 +         like ptys or loop devices.
17729 +
17730 +config VSERVER_CONTEXTS
17731 +       int     "Maximum number of Contexts (1-65533)"  if EMBEDDED
17732 +       range 1 65533
17733 +       default "768"   if 64BIT
17734 +       default "256"
17735 +       help
17736 +         This setting will optimize certain data structures
17737 +         and memory allocations according to the expected
17738 +         maximum.
17739 +
17740 +         note: this is not a strict upper limit.
17741 +
17742 +config VSERVER_WARN
17743 +       bool    "VServer Warnings"
17744 +       default y
17745 +       help
17746 +         This enables various runtime warnings, which will
17747 +         notify about potential manipulation attempts or
17748 +         resource shortage. It is generally considered to
17749 +         be a good idea to have that enabled.
17750 +
17751 +config VSERVER_WARN_DEVPTS
17752 +       bool    "VServer DevPTS Warnings"
17753 +       depends on VSERVER_WARN
17754 +       default y
17755 +       help
17756 +         This enables DevPTS related warnings, issued when a
17757 +         process inside a context tries to lookup or access
17758 +         a dynamic pts from the host or a different context.
17759 +
17760 +config VSERVER_DEBUG
17761 +       bool    "VServer Debugging Code"
17762 +       default n
17763 +       help
17764 +         Set this to yes if you want to be able to activate
17765 +         debugging output at runtime. It adds a very small
17766 +         overhead to all vserver related functions and
17767 +         increases the kernel size by about 20k.
17768 +
17769 +config VSERVER_HISTORY
17770 +       bool    "VServer History Tracing"
17771 +       depends on VSERVER_DEBUG
17772 +       default n
17773 +       help
17774 +         Set this to yes if you want to record the history of
17775 +         linux-vserver activities, so they can be replayed in
17776 +         the event of a kernel panic or oops.
17777 +
17778 +config VSERVER_HISTORY_SIZE
17779 +       int     "Per-CPU History Size (32-65536)"
17780 +       depends on VSERVER_HISTORY
17781 +       range 32 65536
17782 +       default 64
17783 +       help
17784 +         This allows you to specify the number of entries in
17785 +         the per-CPU history buffer.
17786 +
17787 +config VSERVER_LEGACY_MEM
17788 +       bool    "Legacy Memory Limits"
17789 +       default n
17790 +       help
17791 +         This provides fake memory limits to keep
17792 +         older tools happy in the face of memory
17793 +         cgroups
17794 +
17795 +choice
17796 +       prompt  "Quotes used in debug and warn messages"
17797 +       default QUOTES_ISO8859
17798 +
17799 +config QUOTES_ISO8859
17800 +       bool    "Extended ASCII (ISO 8859) angle quotes"
17801 +       help
17802 +         This uses the extended ASCII characters \xbb
17803 +         and \xab for quoting file and process names.
17804 +
17805 +config QUOTES_UTF8
17806 +       bool    "UTF-8 angle quotes"
17807 +       help
17808 +         This uses the the UTF-8 sequences for angle
17809 +         quotes to quote file and process names.
17810 +
17811 +config QUOTES_ASCII
17812 +       bool    "ASCII single quotes"
17813 +       help
17814 +         This uses the ASCII single quote character
17815 +         (\x27) to quote file and process names.
17816 +
17817 +endchoice
17818 +
17819 +endmenu
17820 +
17821 +
17822 +config VSERVER
17823 +       bool
17824 +       default y
17825 +       select NAMESPACES
17826 +       select UTS_NS
17827 +       select IPC_NS
17828 +#      select USER_NS
17829 +       select SYSVIPC
17830 +
17831 +config VSERVER_SECURITY
17832 +       bool
17833 +       depends on SECURITY
17834 +       default y
17835 +       select SECURITY_CAPABILITIES
17836 +
17837 diff -NurpP --minimal linux-3.0/kernel/vserver/Makefile linux-3.0-vs2.3.1-pre7/kernel/vserver/Makefile
17838 --- linux-3.0/kernel/vserver/Makefile   1970-01-01 01:00:00.000000000 +0100
17839 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/Makefile      2011-06-10 22:11:24.000000000 +0200
17840 @@ -0,0 +1,18 @@
17841 +#
17842 +# Makefile for the Linux vserver routines.
17843 +#
17844 +
17845 +
17846 +obj-y          += vserver.o
17847 +
17848 +vserver-y      := switch.o context.o space.o sched.o network.o inode.o \
17849 +                  limit.o cvirt.o cacct.o signal.o helper.o init.o \
17850 +                  dlimit.o tag.o
17851 +
17852 +vserver-$(CONFIG_INET) += inet.o
17853 +vserver-$(CONFIG_PROC_FS) += proc.o
17854 +vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o debug.o
17855 +vserver-$(CONFIG_VSERVER_HISTORY) += history.o
17856 +vserver-$(CONFIG_VSERVER_MONITOR) += monitor.o
17857 +vserver-$(CONFIG_VSERVER_DEVICE) += device.o
17858 +
17859 diff -NurpP --minimal linux-3.0/kernel/vserver/cacct.c linux-3.0-vs2.3.1-pre7/kernel/vserver/cacct.c
17860 --- linux-3.0/kernel/vserver/cacct.c    1970-01-01 01:00:00.000000000 +0100
17861 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/cacct.c       2011-06-10 22:11:24.000000000 +0200
17862 @@ -0,0 +1,42 @@
17863 +/*
17864 + *  linux/kernel/vserver/cacct.c
17865 + *
17866 + *  Virtual Server: Context Accounting
17867 + *
17868 + *  Copyright (C) 2006-2007 Herbert Pötzl
17869 + *
17870 + *  V0.01  added accounting stats
17871 + *
17872 + */
17873 +
17874 +#include <linux/types.h>
17875 +#include <linux/vs_context.h>
17876 +#include <linux/vserver/cacct_cmd.h>
17877 +#include <linux/vserver/cacct_int.h>
17878 +
17879 +#include <asm/errno.h>
17880 +#include <asm/uaccess.h>
17881 +
17882 +
17883 +int vc_sock_stat(struct vx_info *vxi, void __user *data)
17884 +{
17885 +       struct vcmd_sock_stat_v0 vc_data;
17886 +       int j, field;
17887 +
17888 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
17889 +               return -EFAULT;
17890 +
17891 +       field = vc_data.field;
17892 +       if ((field < 0) || (field >= VXA_SOCK_SIZE))
17893 +               return -EINVAL;
17894 +
17895 +       for (j = 0; j < 3; j++) {
17896 +               vc_data.count[j] = vx_sock_count(&vxi->cacct, field, j);
17897 +               vc_data.total[j] = vx_sock_total(&vxi->cacct, field, j);
17898 +       }
17899 +
17900 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
17901 +               return -EFAULT;
17902 +       return 0;
17903 +}
17904 +
17905 diff -NurpP --minimal linux-3.0/kernel/vserver/cacct_init.h linux-3.0-vs2.3.1-pre7/kernel/vserver/cacct_init.h
17906 --- linux-3.0/kernel/vserver/cacct_init.h       1970-01-01 01:00:00.000000000 +0100
17907 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/cacct_init.h  2011-06-10 22:11:24.000000000 +0200
17908 @@ -0,0 +1,25 @@
17909 +
17910 +
17911 +static inline void vx_info_init_cacct(struct _vx_cacct *cacct)
17912 +{
17913 +       int i, j;
17914 +
17915 +
17916 +       for (i = 0; i < VXA_SOCK_SIZE; i++) {
17917 +               for (j = 0; j < 3; j++) {
17918 +                       atomic_long_set(&cacct->sock[i][j].count, 0);
17919 +                       atomic_long_set(&cacct->sock[i][j].total, 0);
17920 +               }
17921 +       }
17922 +       for (i = 0; i < 8; i++)
17923 +               atomic_set(&cacct->slab[i], 0);
17924 +       for (i = 0; i < 5; i++)
17925 +               for (j = 0; j < 4; j++)
17926 +                       atomic_set(&cacct->page[i][j], 0);
17927 +}
17928 +
17929 +static inline void vx_info_exit_cacct(struct _vx_cacct *cacct)
17930 +{
17931 +       return;
17932 +}
17933 +
17934 diff -NurpP --minimal linux-3.0/kernel/vserver/cacct_proc.h linux-3.0-vs2.3.1-pre7/kernel/vserver/cacct_proc.h
17935 --- linux-3.0/kernel/vserver/cacct_proc.h       1970-01-01 01:00:00.000000000 +0100
17936 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/cacct_proc.h  2011-06-10 22:11:24.000000000 +0200
17937 @@ -0,0 +1,53 @@
17938 +#ifndef _VX_CACCT_PROC_H
17939 +#define _VX_CACCT_PROC_H
17940 +
17941 +#include <linux/vserver/cacct_int.h>
17942 +
17943 +
17944 +#define VX_SOCKA_TOP   \
17945 +       "Type\t    recv #/bytes\t\t   send #/bytes\t\t    fail #/bytes\n"
17946 +
17947 +static inline int vx_info_proc_cacct(struct _vx_cacct *cacct, char *buffer)
17948 +{
17949 +       int i, j, length = 0;
17950 +       static char *type[VXA_SOCK_SIZE] = {
17951 +               "UNSPEC", "UNIX", "INET", "INET6", "PACKET", "OTHER"
17952 +       };
17953 +
17954 +       length += sprintf(buffer + length, VX_SOCKA_TOP);
17955 +       for (i = 0; i < VXA_SOCK_SIZE; i++) {
17956 +               length += sprintf(buffer + length, "%s:", type[i]);
17957 +               for (j = 0; j < 3; j++) {
17958 +                       length += sprintf(buffer + length,
17959 +                               "\t%10lu/%-10lu",
17960 +                               vx_sock_count(cacct, i, j),
17961 +                               vx_sock_total(cacct, i, j));
17962 +               }
17963 +               buffer[length++] = '\n';
17964 +       }
17965 +
17966 +       length += sprintf(buffer + length, "\n");
17967 +       length += sprintf(buffer + length,
17968 +               "slab:\t %8u %8u %8u %8u\n",
17969 +               atomic_read(&cacct->slab[1]),
17970 +               atomic_read(&cacct->slab[4]),
17971 +               atomic_read(&cacct->slab[0]),
17972 +               atomic_read(&cacct->slab[2]));
17973 +
17974 +       length += sprintf(buffer + length, "\n");
17975 +       for (i = 0; i < 5; i++) {
17976 +               length += sprintf(buffer + length,
17977 +                       "page[%d]: %8u %8u %8u %8u\t %8u %8u %8u %8u\n", i,
17978 +                       atomic_read(&cacct->page[i][0]),
17979 +                       atomic_read(&cacct->page[i][1]),
17980 +                       atomic_read(&cacct->page[i][2]),
17981 +                       atomic_read(&cacct->page[i][3]),
17982 +                       atomic_read(&cacct->page[i][4]),
17983 +                       atomic_read(&cacct->page[i][5]),
17984 +                       atomic_read(&cacct->page[i][6]),
17985 +                       atomic_read(&cacct->page[i][7]));
17986 +       }
17987 +       return length;
17988 +}
17989 +
17990 +#endif /* _VX_CACCT_PROC_H */
17991 diff -NurpP --minimal linux-3.0/kernel/vserver/context.c linux-3.0-vs2.3.1-pre7/kernel/vserver/context.c
17992 --- linux-3.0/kernel/vserver/context.c  1970-01-01 01:00:00.000000000 +0100
17993 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/context.c     2011-07-19 19:22:49.000000000 +0200
17994 @@ -0,0 +1,1080 @@
17995 +/*
17996 + *  linux/kernel/vserver/context.c
17997 + *
17998 + *  Virtual Server: Context Support
17999 + *
18000 + *  Copyright (C) 2003-2010  Herbert Pötzl
18001 + *
18002 + *  V0.01  context helper
18003 + *  V0.02  vx_ctx_kill syscall command
18004 + *  V0.03  replaced context_info calls
18005 + *  V0.04  redesign of struct (de)alloc
18006 + *  V0.05  rlimit basic implementation
18007 + *  V0.06  task_xid and info commands
18008 + *  V0.07  context flags and caps
18009 + *  V0.08  switch to RCU based hash
18010 + *  V0.09  revert to non RCU for now
18011 + *  V0.10  and back to working RCU hash
18012 + *  V0.11  and back to locking again
18013 + *  V0.12  referenced context store
18014 + *  V0.13  separate per cpu data
18015 + *  V0.14  changed vcmds to vxi arg
18016 + *  V0.15  added context stat
18017 + *  V0.16  have __create claim() the vxi
18018 + *  V0.17  removed older and legacy stuff
18019 + *  V0.18  added user credentials
18020 + *
18021 + */
18022 +
18023 +#include <linux/slab.h>
18024 +#include <linux/types.h>
18025 +#include <linux/security.h>
18026 +#include <linux/pid_namespace.h>
18027 +#include <linux/capability.h>
18028 +
18029 +#include <linux/vserver/context.h>
18030 +#include <linux/vserver/network.h>
18031 +#include <linux/vserver/debug.h>
18032 +#include <linux/vserver/limit.h>
18033 +#include <linux/vserver/limit_int.h>
18034 +#include <linux/vserver/space.h>
18035 +#include <linux/init_task.h>
18036 +#include <linux/fs_struct.h>
18037 +#include <linux/cred.h>
18038 +
18039 +#include <linux/vs_context.h>
18040 +#include <linux/vs_limit.h>
18041 +#include <linux/vs_pid.h>
18042 +#include <linux/vserver/context_cmd.h>
18043 +
18044 +#include "cvirt_init.h"
18045 +#include "cacct_init.h"
18046 +#include "limit_init.h"
18047 +#include "sched_init.h"
18048 +
18049 +
18050 +atomic_t vx_global_ctotal      = ATOMIC_INIT(0);
18051 +atomic_t vx_global_cactive     = ATOMIC_INIT(0);
18052 +
18053 +
18054 +/*     now inactive context structures */
18055 +
18056 +static struct hlist_head vx_info_inactive = HLIST_HEAD_INIT;
18057 +
18058 +static DEFINE_SPINLOCK(vx_info_inactive_lock);
18059 +
18060 +
18061 +/*     __alloc_vx_info()
18062 +
18063 +       * allocate an initialized vx_info struct
18064 +       * doesn't make it visible (hash)                        */
18065 +
18066 +static struct vx_info *__alloc_vx_info(xid_t xid)
18067 +{
18068 +       struct vx_info *new = NULL;
18069 +       int cpu, index;
18070 +
18071 +       vxdprintk(VXD_CBIT(xid, 0), "alloc_vx_info(%d)*", xid);
18072 +
18073 +       /* would this benefit from a slab cache? */
18074 +       new = kmalloc(sizeof(struct vx_info), GFP_KERNEL);
18075 +       if (!new)
18076 +               return 0;
18077 +
18078 +       memset(new, 0, sizeof(struct vx_info));
18079 +#ifdef CONFIG_SMP
18080 +       new->ptr_pc = alloc_percpu(struct _vx_info_pc);
18081 +       if (!new->ptr_pc)
18082 +               goto error;
18083 +#endif
18084 +       new->vx_id = xid;
18085 +       INIT_HLIST_NODE(&new->vx_hlist);
18086 +       atomic_set(&new->vx_usecnt, 0);
18087 +       atomic_set(&new->vx_tasks, 0);
18088 +       new->vx_parent = NULL;
18089 +       new->vx_state = 0;
18090 +       init_waitqueue_head(&new->vx_wait);
18091 +
18092 +       /* prepare reaper */
18093 +       get_task_struct(init_pid_ns.child_reaper);
18094 +       new->vx_reaper = init_pid_ns.child_reaper;
18095 +       new->vx_badness_bias = 0;
18096 +
18097 +       /* rest of init goes here */
18098 +       vx_info_init_limit(&new->limit);
18099 +       vx_info_init_sched(&new->sched);
18100 +       vx_info_init_cvirt(&new->cvirt);
18101 +       vx_info_init_cacct(&new->cacct);
18102 +
18103 +       /* per cpu data structures */
18104 +       for_each_possible_cpu(cpu) {
18105 +               vx_info_init_sched_pc(
18106 +                       &vx_per_cpu(new, sched_pc, cpu), cpu);
18107 +               vx_info_init_cvirt_pc(
18108 +                       &vx_per_cpu(new, cvirt_pc, cpu), cpu);
18109 +       }
18110 +
18111 +       new->vx_flags = VXF_INIT_SET;
18112 +       new->vx_bcaps = CAP_FULL_SET;   // maybe ~CAP_SETPCAP
18113 +       new->vx_ccaps = 0;
18114 +       new->vx_umask = 0;
18115 +
18116 +       new->reboot_cmd = 0;
18117 +       new->exit_code = 0;
18118 +
18119 +       // preconfig spaces
18120 +       for (index = 0; index < VX_SPACES; index++) {
18121 +               struct _vx_space *space = &new->space[index];
18122 +
18123 +               // filesystem
18124 +               spin_lock(&init_fs.lock);
18125 +               init_fs.users++;
18126 +               spin_unlock(&init_fs.lock);
18127 +               space->vx_fs = &init_fs;
18128 +
18129 +               /* FIXME: do we want defaults? */
18130 +               // space->vx_real_cred = 0;
18131 +               // space->vx_cred = 0;
18132 +       }
18133 +
18134 +
18135 +       vxdprintk(VXD_CBIT(xid, 0),
18136 +               "alloc_vx_info(%d) = %p", xid, new);
18137 +       vxh_alloc_vx_info(new);
18138 +       atomic_inc(&vx_global_ctotal);
18139 +       return new;
18140 +#ifdef CONFIG_SMP
18141 +error:
18142 +       kfree(new);
18143 +       return 0;
18144 +#endif
18145 +}
18146 +
18147 +/*     __dealloc_vx_info()
18148 +
18149 +       * final disposal of vx_info                             */
18150 +
18151 +static void __dealloc_vx_info(struct vx_info *vxi)
18152 +{
18153 +#ifdef CONFIG_VSERVER_WARN
18154 +       struct vx_info_save vxis;
18155 +       int cpu;
18156 +#endif
18157 +       vxdprintk(VXD_CBIT(xid, 0),
18158 +               "dealloc_vx_info(%p)", vxi);
18159 +       vxh_dealloc_vx_info(vxi);
18160 +
18161 +#ifdef CONFIG_VSERVER_WARN
18162 +       enter_vx_info(vxi, &vxis);
18163 +       vx_info_exit_limit(&vxi->limit);
18164 +       vx_info_exit_sched(&vxi->sched);
18165 +       vx_info_exit_cvirt(&vxi->cvirt);
18166 +       vx_info_exit_cacct(&vxi->cacct);
18167 +
18168 +       for_each_possible_cpu(cpu) {
18169 +               vx_info_exit_sched_pc(
18170 +                       &vx_per_cpu(vxi, sched_pc, cpu), cpu);
18171 +               vx_info_exit_cvirt_pc(
18172 +                       &vx_per_cpu(vxi, cvirt_pc, cpu), cpu);
18173 +       }
18174 +       leave_vx_info(&vxis);
18175 +#endif
18176 +
18177 +       vxi->vx_id = -1;
18178 +       vxi->vx_state |= VXS_RELEASED;
18179 +
18180 +#ifdef CONFIG_SMP
18181 +       free_percpu(vxi->ptr_pc);
18182 +#endif
18183 +       kfree(vxi);
18184 +       atomic_dec(&vx_global_ctotal);
18185 +}
18186 +
18187 +static void __shutdown_vx_info(struct vx_info *vxi)
18188 +{
18189 +       struct nsproxy *nsproxy;
18190 +       struct fs_struct *fs;
18191 +       struct cred *cred;
18192 +       int index, kill;
18193 +
18194 +       might_sleep();
18195 +
18196 +       vxi->vx_state |= VXS_SHUTDOWN;
18197 +       vs_state_change(vxi, VSC_SHUTDOWN);
18198 +
18199 +       for (index = 0; index < VX_SPACES; index++) {
18200 +               struct _vx_space *space = &vxi->space[index];
18201 +
18202 +               nsproxy = xchg(&space->vx_nsproxy, NULL);
18203 +               if (nsproxy)
18204 +                       put_nsproxy(nsproxy);
18205 +
18206 +               fs = xchg(&space->vx_fs, NULL);
18207 +               spin_lock(&fs->lock);
18208 +               kill = !--fs->users;
18209 +               spin_unlock(&fs->lock);
18210 +               if (kill)
18211 +                       free_fs_struct(fs);
18212 +
18213 +               cred = (struct cred *)xchg(&space->vx_cred, NULL);
18214 +               if (cred)
18215 +                       abort_creds(cred);
18216 +       }
18217 +}
18218 +
18219 +/* exported stuff */
18220 +
18221 +void free_vx_info(struct vx_info *vxi)
18222 +{
18223 +       unsigned long flags;
18224 +       unsigned index;
18225 +
18226 +       /* check for reference counts first */
18227 +       BUG_ON(atomic_read(&vxi->vx_usecnt));
18228 +       BUG_ON(atomic_read(&vxi->vx_tasks));
18229 +
18230 +       /* context must not be hashed */
18231 +       BUG_ON(vx_info_state(vxi, VXS_HASHED));
18232 +
18233 +       /* context shutdown is mandatory */
18234 +       BUG_ON(!vx_info_state(vxi, VXS_SHUTDOWN));
18235 +
18236 +       /* spaces check */
18237 +       for (index = 0; index < VX_SPACES; index++) {
18238 +               struct _vx_space *space = &vxi->space[index];
18239 +
18240 +               BUG_ON(space->vx_nsproxy);
18241 +               BUG_ON(space->vx_fs);
18242 +               // BUG_ON(space->vx_real_cred);
18243 +               // BUG_ON(space->vx_cred);
18244 +       }
18245 +
18246 +       spin_lock_irqsave(&vx_info_inactive_lock, flags);
18247 +       hlist_del(&vxi->vx_hlist);
18248 +       spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
18249 +
18250 +       __dealloc_vx_info(vxi);
18251 +}
18252 +
18253 +
18254 +/*     hash table for vx_info hash */
18255 +
18256 +#define VX_HASH_SIZE   13
18257 +
18258 +static struct hlist_head vx_info_hash[VX_HASH_SIZE] =
18259 +       { [0 ... VX_HASH_SIZE-1] = HLIST_HEAD_INIT };
18260 +
18261 +static DEFINE_SPINLOCK(vx_info_hash_lock);
18262 +
18263 +
18264 +static inline unsigned int __hashval(xid_t xid)
18265 +{
18266 +       return (xid % VX_HASH_SIZE);
18267 +}
18268 +
18269 +
18270 +
18271 +/*     __hash_vx_info()
18272 +
18273 +       * add the vxi to the global hash table
18274 +       * requires the hash_lock to be held                     */
18275 +
18276 +static inline void __hash_vx_info(struct vx_info *vxi)
18277 +{
18278 +       struct hlist_head *head;
18279 +
18280 +       vxd_assert_lock(&vx_info_hash_lock);
18281 +       vxdprintk(VXD_CBIT(xid, 4),
18282 +               "__hash_vx_info: %p[#%d]", vxi, vxi->vx_id);
18283 +       vxh_hash_vx_info(vxi);
18284 +
18285 +       /* context must not be hashed */
18286 +       BUG_ON(vx_info_state(vxi, VXS_HASHED));
18287 +
18288 +       vxi->vx_state |= VXS_HASHED;
18289 +       head = &vx_info_hash[__hashval(vxi->vx_id)];
18290 +       hlist_add_head(&vxi->vx_hlist, head);
18291 +       atomic_inc(&vx_global_cactive);
18292 +}
18293 +
18294 +/*     __unhash_vx_info()
18295 +
18296 +       * remove the vxi from the global hash table
18297 +       * requires the hash_lock to be held                     */
18298 +
18299 +static inline void __unhash_vx_info(struct vx_info *vxi)
18300 +{
18301 +       unsigned long flags;
18302 +
18303 +       vxd_assert_lock(&vx_info_hash_lock);
18304 +       vxdprintk(VXD_CBIT(xid, 4),
18305 +               "__unhash_vx_info: %p[#%d.%d.%d]", vxi, vxi->vx_id,
18306 +               atomic_read(&vxi->vx_usecnt), atomic_read(&vxi->vx_tasks));
18307 +       vxh_unhash_vx_info(vxi);
18308 +
18309 +       /* context must be hashed */
18310 +       BUG_ON(!vx_info_state(vxi, VXS_HASHED));
18311 +       /* but without tasks */
18312 +       BUG_ON(atomic_read(&vxi->vx_tasks));
18313 +
18314 +       vxi->vx_state &= ~VXS_HASHED;
18315 +       hlist_del_init(&vxi->vx_hlist);
18316 +       spin_lock_irqsave(&vx_info_inactive_lock, flags);
18317 +       hlist_add_head(&vxi->vx_hlist, &vx_info_inactive);
18318 +       spin_unlock_irqrestore(&vx_info_inactive_lock, flags);
18319 +       atomic_dec(&vx_global_cactive);
18320 +}
18321 +
18322 +
18323 +/*     __lookup_vx_info()
18324 +
18325 +       * requires the hash_lock to be held
18326 +       * doesn't increment the vx_refcnt                       */
18327 +
18328 +static inline struct vx_info *__lookup_vx_info(xid_t xid)
18329 +{
18330 +       struct hlist_head *head = &vx_info_hash[__hashval(xid)];
18331 +       struct hlist_node *pos;
18332 +       struct vx_info *vxi;
18333 +
18334 +       vxd_assert_lock(&vx_info_hash_lock);
18335 +       hlist_for_each(pos, head) {
18336 +               vxi = hlist_entry(pos, struct vx_info, vx_hlist);
18337 +
18338 +               if (vxi->vx_id == xid)
18339 +                       goto found;
18340 +       }
18341 +       vxi = NULL;
18342 +found:
18343 +       vxdprintk(VXD_CBIT(xid, 0),
18344 +               "__lookup_vx_info(#%u): %p[#%u]",
18345 +               xid, vxi, vxi ? vxi->vx_id : 0);
18346 +       vxh_lookup_vx_info(vxi, xid);
18347 +       return vxi;
18348 +}
18349 +
18350 +
18351 +/*     __create_vx_info()
18352 +
18353 +       * create the requested context
18354 +       * get(), claim() and hash it                            */
18355 +
18356 +static struct vx_info *__create_vx_info(int id)
18357 +{
18358 +       struct vx_info *new, *vxi = NULL;
18359 +
18360 +       vxdprintk(VXD_CBIT(xid, 1), "create_vx_info(%d)*", id);
18361 +
18362 +       if (!(new = __alloc_vx_info(id)))
18363 +               return ERR_PTR(-ENOMEM);
18364 +
18365 +       /* required to make dynamic xids unique */
18366 +       spin_lock(&vx_info_hash_lock);
18367 +
18368 +       /* static context requested */
18369 +       if ((vxi = __lookup_vx_info(id))) {
18370 +               vxdprintk(VXD_CBIT(xid, 0),
18371 +                       "create_vx_info(%d) = %p (already there)", id, vxi);
18372 +               if (vx_info_flags(vxi, VXF_STATE_SETUP, 0))
18373 +                       vxi = ERR_PTR(-EBUSY);
18374 +               else
18375 +                       vxi = ERR_PTR(-EEXIST);
18376 +               goto out_unlock;
18377 +       }
18378 +       /* new context */
18379 +       vxdprintk(VXD_CBIT(xid, 0),
18380 +               "create_vx_info(%d) = %p (new)", id, new);
18381 +       claim_vx_info(new, NULL);
18382 +       __hash_vx_info(get_vx_info(new));
18383 +       vxi = new, new = NULL;
18384 +
18385 +out_unlock:
18386 +       spin_unlock(&vx_info_hash_lock);
18387 +       vxh_create_vx_info(IS_ERR(vxi) ? NULL : vxi, id);
18388 +       if (new)
18389 +               __dealloc_vx_info(new);
18390 +       return vxi;
18391 +}
18392 +
18393 +
18394 +/*     exported stuff                                          */
18395 +
18396 +
18397 +void unhash_vx_info(struct vx_info *vxi)
18398 +{
18399 +       spin_lock(&vx_info_hash_lock);
18400 +       __unhash_vx_info(vxi);
18401 +       spin_unlock(&vx_info_hash_lock);
18402 +       __shutdown_vx_info(vxi);
18403 +       __wakeup_vx_info(vxi);
18404 +}
18405 +
18406 +
18407 +/*     lookup_vx_info()
18408 +
18409 +       * search for a vx_info and get() it
18410 +       * negative id means current                             */
18411 +
18412 +struct vx_info *lookup_vx_info(int id)
18413 +{
18414 +       struct vx_info *vxi = NULL;
18415 +
18416 +       if (id < 0) {
18417 +               vxi = get_vx_info(current_vx_info());
18418 +       } else if (id > 1) {
18419 +               spin_lock(&vx_info_hash_lock);
18420 +               vxi = get_vx_info(__lookup_vx_info(id));
18421 +               spin_unlock(&vx_info_hash_lock);
18422 +       }
18423 +       return vxi;
18424 +}
18425 +
18426 +/*     xid_is_hashed()
18427 +
18428 +       * verify that xid is still hashed                       */
18429 +
18430 +int xid_is_hashed(xid_t xid)
18431 +{
18432 +       int hashed;
18433 +
18434 +       spin_lock(&vx_info_hash_lock);
18435 +       hashed = (__lookup_vx_info(xid) != NULL);
18436 +       spin_unlock(&vx_info_hash_lock);
18437 +       return hashed;
18438 +}
18439 +
18440 +#ifdef CONFIG_PROC_FS
18441 +
18442 +/*     get_xid_list()
18443 +
18444 +       * get a subset of hashed xids for proc
18445 +       * assumes size is at least one                          */
18446 +
18447 +int get_xid_list(int index, unsigned int *xids, int size)
18448 +{
18449 +       int hindex, nr_xids = 0;
18450 +
18451 +       /* only show current and children */
18452 +       if (!vx_check(0, VS_ADMIN | VS_WATCH)) {
18453 +               if (index > 0)
18454 +                       return 0;
18455 +               xids[nr_xids] = vx_current_xid();
18456 +               return 1;
18457 +       }
18458 +
18459 +       for (hindex = 0; hindex < VX_HASH_SIZE; hindex++) {
18460 +               struct hlist_head *head = &vx_info_hash[hindex];
18461 +               struct hlist_node *pos;
18462 +
18463 +               spin_lock(&vx_info_hash_lock);
18464 +               hlist_for_each(pos, head) {
18465 +                       struct vx_info *vxi;
18466 +
18467 +                       if (--index > 0)
18468 +                               continue;
18469 +
18470 +                       vxi = hlist_entry(pos, struct vx_info, vx_hlist);
18471 +                       xids[nr_xids] = vxi->vx_id;
18472 +                       if (++nr_xids >= size) {
18473 +                               spin_unlock(&vx_info_hash_lock);
18474 +                               goto out;
18475 +                       }
18476 +               }
18477 +               /* keep the lock time short */
18478 +               spin_unlock(&vx_info_hash_lock);
18479 +       }
18480 +out:
18481 +       return nr_xids;
18482 +}
18483 +#endif
18484 +
18485 +#ifdef CONFIG_VSERVER_DEBUG
18486 +
18487 +void   dump_vx_info_inactive(int level)
18488 +{
18489 +       struct hlist_node *entry, *next;
18490 +
18491 +       hlist_for_each_safe(entry, next, &vx_info_inactive) {
18492 +               struct vx_info *vxi =
18493 +                       list_entry(entry, struct vx_info, vx_hlist);
18494 +
18495 +               dump_vx_info(vxi, level);
18496 +       }
18497 +}
18498 +
18499 +#endif
18500 +
18501 +#if 0
18502 +int vx_migrate_user(struct task_struct *p, struct vx_info *vxi)
18503 +{
18504 +       struct user_struct *new_user, *old_user;
18505 +
18506 +       if (!p || !vxi)
18507 +               BUG();
18508 +
18509 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
18510 +               return -EACCES;
18511 +
18512 +       new_user = alloc_uid(vxi->vx_id, p->uid);
18513 +       if (!new_user)
18514 +               return -ENOMEM;
18515 +
18516 +       old_user = p->user;
18517 +       if (new_user != old_user) {
18518 +               atomic_inc(&new_user->processes);
18519 +               atomic_dec(&old_user->processes);
18520 +               p->user = new_user;
18521 +       }
18522 +       free_uid(old_user);
18523 +       return 0;
18524 +}
18525 +#endif
18526 +
18527 +#if 0
18528 +void vx_mask_cap_bset(struct vx_info *vxi, struct task_struct *p)
18529 +{
18530 +       // p->cap_effective &= vxi->vx_cap_bset;
18531 +       p->cap_effective =
18532 +               cap_intersect(p->cap_effective, vxi->cap_bset);
18533 +       // p->cap_inheritable &= vxi->vx_cap_bset;
18534 +       p->cap_inheritable =
18535 +               cap_intersect(p->cap_inheritable, vxi->cap_bset);
18536 +       // p->cap_permitted &= vxi->vx_cap_bset;
18537 +       p->cap_permitted =
18538 +               cap_intersect(p->cap_permitted, vxi->cap_bset);
18539 +}
18540 +#endif
18541 +
18542 +
18543 +#include <linux/file.h>
18544 +#include <linux/fdtable.h>
18545 +
18546 +static int vx_openfd_task(struct task_struct *tsk)
18547 +{
18548 +       struct files_struct *files = tsk->files;
18549 +       struct fdtable *fdt;
18550 +       const unsigned long *bptr;
18551 +       int count, total;
18552 +
18553 +       /* no rcu_read_lock() because of spin_lock() */
18554 +       spin_lock(&files->file_lock);
18555 +       fdt = files_fdtable(files);
18556 +       bptr = fdt->open_fds->fds_bits;
18557 +       count = fdt->max_fds / (sizeof(unsigned long) * 8);
18558 +       for (total = 0; count > 0; count--) {
18559 +               if (*bptr)
18560 +                       total += hweight_long(*bptr);
18561 +               bptr++;
18562 +       }
18563 +       spin_unlock(&files->file_lock);
18564 +       return total;
18565 +}
18566 +
18567 +
18568 +/*     for *space compatibility */
18569 +
18570 +asmlinkage long sys_unshare(unsigned long);
18571 +
18572 +/*
18573 + *     migrate task to new context
18574 + *     gets vxi, puts old_vxi on change
18575 + *     optionally unshares namespaces (hack)
18576 + */
18577 +
18578 +int vx_migrate_task(struct task_struct *p, struct vx_info *vxi, int unshare)
18579 +{
18580 +       struct vx_info *old_vxi;
18581 +       int ret = 0;
18582 +
18583 +       if (!p || !vxi)
18584 +               BUG();
18585 +
18586 +       vxdprintk(VXD_CBIT(xid, 5),
18587 +               "vx_migrate_task(%p,%p[#%d.%d])", p, vxi,
18588 +               vxi->vx_id, atomic_read(&vxi->vx_usecnt));
18589 +
18590 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0) &&
18591 +               !vx_info_flags(vxi, VXF_STATE_SETUP, 0))
18592 +               return -EACCES;
18593 +
18594 +       if (vx_info_state(vxi, VXS_SHUTDOWN))
18595 +               return -EFAULT;
18596 +
18597 +       old_vxi = task_get_vx_info(p);
18598 +       if (old_vxi == vxi)
18599 +               goto out;
18600 +
18601 +//     if (!(ret = vx_migrate_user(p, vxi))) {
18602 +       {
18603 +               int openfd;
18604 +
18605 +               task_lock(p);
18606 +               openfd = vx_openfd_task(p);
18607 +
18608 +               if (old_vxi) {
18609 +                       atomic_dec(&old_vxi->cvirt.nr_threads);
18610 +                       atomic_dec(&old_vxi->cvirt.nr_running);
18611 +                       __rlim_dec(&old_vxi->limit, RLIMIT_NPROC);
18612 +                       /* FIXME: what about the struct files here? */
18613 +                       __rlim_sub(&old_vxi->limit, VLIMIT_OPENFD, openfd);
18614 +                       /* account for the executable */
18615 +                       __rlim_dec(&old_vxi->limit, VLIMIT_DENTRY);
18616 +               }
18617 +               atomic_inc(&vxi->cvirt.nr_threads);
18618 +               atomic_inc(&vxi->cvirt.nr_running);
18619 +               __rlim_inc(&vxi->limit, RLIMIT_NPROC);
18620 +               /* FIXME: what about the struct files here? */
18621 +               __rlim_add(&vxi->limit, VLIMIT_OPENFD, openfd);
18622 +               /* account for the executable */
18623 +               __rlim_inc(&vxi->limit, VLIMIT_DENTRY);
18624 +
18625 +               if (old_vxi) {
18626 +                       release_vx_info(old_vxi, p);
18627 +                       clr_vx_info(&p->vx_info);
18628 +               }
18629 +               claim_vx_info(vxi, p);
18630 +               set_vx_info(&p->vx_info, vxi);
18631 +               p->xid = vxi->vx_id;
18632 +
18633 +               vxdprintk(VXD_CBIT(xid, 5),
18634 +                       "moved task %p into vxi:%p[#%d]",
18635 +                       p, vxi, vxi->vx_id);
18636 +
18637 +               // vx_mask_cap_bset(vxi, p);
18638 +               task_unlock(p);
18639 +
18640 +               /* hack for *spaces to provide compatibility */
18641 +               if (unshare) {
18642 +                       struct nsproxy *old_nsp, *new_nsp;
18643 +
18644 +                       ret = unshare_nsproxy_namespaces(
18645 +                               CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER,
18646 +                               &new_nsp, NULL);
18647 +                       if (ret)
18648 +                               goto out;
18649 +
18650 +                       old_nsp = xchg(&p->nsproxy, new_nsp);
18651 +                       vx_set_space(vxi,
18652 +                               CLONE_NEWUTS | CLONE_NEWIPC | CLONE_NEWUSER, 0);
18653 +                       put_nsproxy(old_nsp);
18654 +               }
18655 +       }
18656 +out:
18657 +       put_vx_info(old_vxi);
18658 +       return ret;
18659 +}
18660 +
18661 +int vx_set_reaper(struct vx_info *vxi, struct task_struct *p)
18662 +{
18663 +       struct task_struct *old_reaper;
18664 +
18665 +       if (!vxi)
18666 +               return -EINVAL;
18667 +
18668 +       vxdprintk(VXD_CBIT(xid, 6),
18669 +               "vx_set_reaper(%p[#%d],%p[#%d,%d])",
18670 +               vxi, vxi->vx_id, p, p->xid, p->pid);
18671 +
18672 +       old_reaper = vxi->vx_reaper;
18673 +       if (old_reaper == p)
18674 +               return 0;
18675 +
18676 +       /* set new child reaper */
18677 +       get_task_struct(p);
18678 +       vxi->vx_reaper = p;
18679 +       put_task_struct(old_reaper);
18680 +       return 0;
18681 +}
18682 +
18683 +int vx_set_init(struct vx_info *vxi, struct task_struct *p)
18684 +{
18685 +       if (!vxi)
18686 +               return -EINVAL;
18687 +
18688 +       vxdprintk(VXD_CBIT(xid, 6),
18689 +               "vx_set_init(%p[#%d],%p[#%d,%d,%d])",
18690 +               vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
18691 +
18692 +       vxi->vx_flags &= ~VXF_STATE_INIT;
18693 +       // vxi->vx_initpid = p->tgid;
18694 +       vxi->vx_initpid = p->pid;
18695 +       return 0;
18696 +}
18697 +
18698 +void vx_exit_init(struct vx_info *vxi, struct task_struct *p, int code)
18699 +{
18700 +       vxdprintk(VXD_CBIT(xid, 6),
18701 +               "vx_exit_init(%p[#%d],%p[#%d,%d,%d])",
18702 +               vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid);
18703 +
18704 +       vxi->exit_code = code;
18705 +       vxi->vx_initpid = 0;
18706 +}
18707 +
18708 +
18709 +void vx_set_persistent(struct vx_info *vxi)
18710 +{
18711 +       vxdprintk(VXD_CBIT(xid, 6),
18712 +               "vx_set_persistent(%p[#%d])", vxi, vxi->vx_id);
18713 +
18714 +       get_vx_info(vxi);
18715 +       claim_vx_info(vxi, NULL);
18716 +}
18717 +
18718 +void vx_clear_persistent(struct vx_info *vxi)
18719 +{
18720 +       vxdprintk(VXD_CBIT(xid, 6),
18721 +               "vx_clear_persistent(%p[#%d])", vxi, vxi->vx_id);
18722 +
18723 +       release_vx_info(vxi, NULL);
18724 +       put_vx_info(vxi);
18725 +}
18726 +
18727 +void vx_update_persistent(struct vx_info *vxi)
18728 +{
18729 +       if (vx_info_flags(vxi, VXF_PERSISTENT, 0))
18730 +               vx_set_persistent(vxi);
18731 +       else
18732 +               vx_clear_persistent(vxi);
18733 +}
18734 +
18735 +
18736 +/*     task must be current or locked          */
18737 +
18738 +void   exit_vx_info(struct task_struct *p, int code)
18739 +{
18740 +       struct vx_info *vxi = p->vx_info;
18741 +
18742 +       if (vxi) {
18743 +               atomic_dec(&vxi->cvirt.nr_threads);
18744 +               vx_nproc_dec(p);
18745 +
18746 +               vxi->exit_code = code;
18747 +               release_vx_info(vxi, p);
18748 +       }
18749 +}
18750 +
18751 +void   exit_vx_info_early(struct task_struct *p, int code)
18752 +{
18753 +       struct vx_info *vxi = p->vx_info;
18754 +
18755 +       if (vxi) {
18756 +               if (vxi->vx_initpid == p->pid)
18757 +                       vx_exit_init(vxi, p, code);
18758 +               if (vxi->vx_reaper == p)
18759 +                       vx_set_reaper(vxi, init_pid_ns.child_reaper);
18760 +       }
18761 +}
18762 +
18763 +
18764 +/* vserver syscall commands below here */
18765 +
18766 +/* taks xid and vx_info functions */
18767 +
18768 +#include <asm/uaccess.h>
18769 +
18770 +
18771 +int vc_task_xid(uint32_t id)
18772 +{
18773 +       xid_t xid;
18774 +
18775 +       if (id) {
18776 +               struct task_struct *tsk;
18777 +
18778 +               rcu_read_lock();
18779 +               tsk = find_task_by_real_pid(id);
18780 +               xid = (tsk) ? tsk->xid : -ESRCH;
18781 +               rcu_read_unlock();
18782 +       } else
18783 +               xid = vx_current_xid();
18784 +       return xid;
18785 +}
18786 +
18787 +
18788 +int vc_vx_info(struct vx_info *vxi, void __user *data)
18789 +{
18790 +       struct vcmd_vx_info_v0 vc_data;
18791 +
18792 +       vc_data.xid = vxi->vx_id;
18793 +       vc_data.initpid = vxi->vx_initpid;
18794 +
18795 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18796 +               return -EFAULT;
18797 +       return 0;
18798 +}
18799 +
18800 +
18801 +int vc_ctx_stat(struct vx_info *vxi, void __user *data)
18802 +{
18803 +       struct vcmd_ctx_stat_v0 vc_data;
18804 +
18805 +       vc_data.usecnt = atomic_read(&vxi->vx_usecnt);
18806 +       vc_data.tasks = atomic_read(&vxi->vx_tasks);
18807 +
18808 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18809 +               return -EFAULT;
18810 +       return 0;
18811 +}
18812 +
18813 +
18814 +/* context functions */
18815 +
18816 +int vc_ctx_create(uint32_t xid, void __user *data)
18817 +{
18818 +       struct vcmd_ctx_create vc_data = { .flagword = VXF_INIT_SET };
18819 +       struct vx_info *new_vxi;
18820 +       int ret;
18821 +
18822 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
18823 +               return -EFAULT;
18824 +
18825 +       if ((xid > MAX_S_CONTEXT) || (xid < 2))
18826 +               return -EINVAL;
18827 +
18828 +       new_vxi = __create_vx_info(xid);
18829 +       if (IS_ERR(new_vxi))
18830 +               return PTR_ERR(new_vxi);
18831 +
18832 +       /* initial flags */
18833 +       new_vxi->vx_flags = vc_data.flagword;
18834 +
18835 +       ret = -ENOEXEC;
18836 +       if (vs_state_change(new_vxi, VSC_STARTUP))
18837 +               goto out;
18838 +
18839 +       ret = vx_migrate_task(current, new_vxi, (!data));
18840 +       if (ret)
18841 +               goto out;
18842 +
18843 +       /* return context id on success */
18844 +       ret = new_vxi->vx_id;
18845 +
18846 +       /* get a reference for persistent contexts */
18847 +       if ((vc_data.flagword & VXF_PERSISTENT))
18848 +               vx_set_persistent(new_vxi);
18849 +out:
18850 +       release_vx_info(new_vxi, NULL);
18851 +       put_vx_info(new_vxi);
18852 +       return ret;
18853 +}
18854 +
18855 +
18856 +int vc_ctx_migrate(struct vx_info *vxi, void __user *data)
18857 +{
18858 +       struct vcmd_ctx_migrate vc_data = { .flagword = 0 };
18859 +       int ret;
18860 +
18861 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
18862 +               return -EFAULT;
18863 +
18864 +       ret = vx_migrate_task(current, vxi, 0);
18865 +       if (ret)
18866 +               return ret;
18867 +       if (vc_data.flagword & VXM_SET_INIT)
18868 +               ret = vx_set_init(vxi, current);
18869 +       if (ret)
18870 +               return ret;
18871 +       if (vc_data.flagword & VXM_SET_REAPER)
18872 +               ret = vx_set_reaper(vxi, current);
18873 +       return ret;
18874 +}
18875 +
18876 +
18877 +int vc_get_cflags(struct vx_info *vxi, void __user *data)
18878 +{
18879 +       struct vcmd_ctx_flags_v0 vc_data;
18880 +
18881 +       vc_data.flagword = vxi->vx_flags;
18882 +
18883 +       /* special STATE flag handling */
18884 +       vc_data.mask = vs_mask_flags(~0ULL, vxi->vx_flags, VXF_ONE_TIME);
18885 +
18886 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18887 +               return -EFAULT;
18888 +       return 0;
18889 +}
18890 +
18891 +int vc_set_cflags(struct vx_info *vxi, void __user *data)
18892 +{
18893 +       struct vcmd_ctx_flags_v0 vc_data;
18894 +       uint64_t mask, trigger;
18895 +
18896 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18897 +               return -EFAULT;
18898 +
18899 +       /* special STATE flag handling */
18900 +       mask = vs_mask_mask(vc_data.mask, vxi->vx_flags, VXF_ONE_TIME);
18901 +       trigger = (mask & vxi->vx_flags) ^ (mask & vc_data.flagword);
18902 +
18903 +       if (vxi == current_vx_info()) {
18904 +               /* if (trigger & VXF_STATE_SETUP)
18905 +                       vx_mask_cap_bset(vxi, current); */
18906 +               if (trigger & VXF_STATE_INIT) {
18907 +                       int ret;
18908 +
18909 +                       ret = vx_set_init(vxi, current);
18910 +                       if (ret)
18911 +                               return ret;
18912 +                       ret = vx_set_reaper(vxi, current);
18913 +                       if (ret)
18914 +                               return ret;
18915 +               }
18916 +       }
18917 +
18918 +       vxi->vx_flags = vs_mask_flags(vxi->vx_flags,
18919 +               vc_data.flagword, mask);
18920 +       if (trigger & VXF_PERSISTENT)
18921 +               vx_update_persistent(vxi);
18922 +
18923 +       return 0;
18924 +}
18925 +
18926 +
18927 +static inline uint64_t caps_from_cap_t(kernel_cap_t c)
18928 +{
18929 +       uint64_t v = c.cap[0] | ((uint64_t)c.cap[1] << 32);
18930 +
18931 +       // printk("caps_from_cap_t(%08x:%08x) = %016llx\n", c.cap[1], c.cap[0], v);
18932 +       return v;
18933 +}
18934 +
18935 +static inline kernel_cap_t cap_t_from_caps(uint64_t v)
18936 +{
18937 +       kernel_cap_t c = __cap_empty_set;
18938 +
18939 +       c.cap[0] = v & 0xFFFFFFFF;
18940 +       c.cap[1] = (v >> 32) & 0xFFFFFFFF;
18941 +
18942 +       // printk("cap_t_from_caps(%016llx) = %08x:%08x\n", v, c.cap[1], c.cap[0]);
18943 +       return c;
18944 +}
18945 +
18946 +
18947 +static int do_get_caps(struct vx_info *vxi, uint64_t *bcaps, uint64_t *ccaps)
18948 +{
18949 +       if (bcaps)
18950 +               *bcaps = caps_from_cap_t(vxi->vx_bcaps);
18951 +       if (ccaps)
18952 +               *ccaps = vxi->vx_ccaps;
18953 +
18954 +       return 0;
18955 +}
18956 +
18957 +int vc_get_ccaps(struct vx_info *vxi, void __user *data)
18958 +{
18959 +       struct vcmd_ctx_caps_v1 vc_data;
18960 +       int ret;
18961 +
18962 +       ret = do_get_caps(vxi, NULL, &vc_data.ccaps);
18963 +       if (ret)
18964 +               return ret;
18965 +       vc_data.cmask = ~0ULL;
18966 +
18967 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
18968 +               return -EFAULT;
18969 +       return 0;
18970 +}
18971 +
18972 +static int do_set_caps(struct vx_info *vxi,
18973 +       uint64_t bcaps, uint64_t bmask, uint64_t ccaps, uint64_t cmask)
18974 +{
18975 +       uint64_t bcold = caps_from_cap_t(vxi->vx_bcaps);
18976 +
18977 +#if 0
18978 +       printk("do_set_caps(%16llx, %16llx, %16llx, %16llx)\n",
18979 +               bcaps, bmask, ccaps, cmask);
18980 +#endif
18981 +       vxi->vx_bcaps = cap_t_from_caps(
18982 +               vs_mask_flags(bcold, bcaps, bmask));
18983 +       vxi->vx_ccaps = vs_mask_flags(vxi->vx_ccaps, ccaps, cmask);
18984 +
18985 +       return 0;
18986 +}
18987 +
18988 +int vc_set_ccaps(struct vx_info *vxi, void __user *data)
18989 +{
18990 +       struct vcmd_ctx_caps_v1 vc_data;
18991 +
18992 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
18993 +               return -EFAULT;
18994 +
18995 +       return do_set_caps(vxi, 0, 0, vc_data.ccaps, vc_data.cmask);
18996 +}
18997 +
18998 +int vc_get_bcaps(struct vx_info *vxi, void __user *data)
18999 +{
19000 +       struct vcmd_bcaps vc_data;
19001 +       int ret;
19002 +
19003 +       ret = do_get_caps(vxi, &vc_data.bcaps, NULL);
19004 +       if (ret)
19005 +               return ret;
19006 +       vc_data.bmask = ~0ULL;
19007 +
19008 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19009 +               return -EFAULT;
19010 +       return 0;
19011 +}
19012 +
19013 +int vc_set_bcaps(struct vx_info *vxi, void __user *data)
19014 +{
19015 +       struct vcmd_bcaps vc_data;
19016 +
19017 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19018 +               return -EFAULT;
19019 +
19020 +       return do_set_caps(vxi, vc_data.bcaps, vc_data.bmask, 0, 0);
19021 +}
19022 +
19023 +
19024 +int vc_get_umask(struct vx_info *vxi, void __user *data)
19025 +{
19026 +       struct vcmd_umask vc_data;
19027 +
19028 +       vc_data.umask = vxi->vx_umask;
19029 +       vc_data.mask = ~0ULL;
19030 +
19031 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19032 +               return -EFAULT;
19033 +       return 0;
19034 +}
19035 +
19036 +int vc_set_umask(struct vx_info *vxi, void __user *data)
19037 +{
19038 +       struct vcmd_umask vc_data;
19039 +
19040 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19041 +               return -EFAULT;
19042 +
19043 +       vxi->vx_umask = vs_mask_flags(vxi->vx_umask,
19044 +               vc_data.umask, vc_data.mask);
19045 +       return 0;
19046 +}
19047 +
19048 +
19049 +int vc_get_badness(struct vx_info *vxi, void __user *data)
19050 +{
19051 +       struct vcmd_badness_v0 vc_data;
19052 +
19053 +       vc_data.bias = vxi->vx_badness_bias;
19054 +
19055 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19056 +               return -EFAULT;
19057 +       return 0;
19058 +}
19059 +
19060 +int vc_set_badness(struct vx_info *vxi, void __user *data)
19061 +{
19062 +       struct vcmd_badness_v0 vc_data;
19063 +
19064 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19065 +               return -EFAULT;
19066 +
19067 +       vxi->vx_badness_bias = vc_data.bias;
19068 +       return 0;
19069 +}
19070 +
19071 +#include <linux/module.h>
19072 +
19073 +EXPORT_SYMBOL_GPL(free_vx_info);
19074 +
19075 diff -NurpP --minimal linux-3.0/kernel/vserver/cvirt.c linux-3.0-vs2.3.1-pre7/kernel/vserver/cvirt.c
19076 --- linux-3.0/kernel/vserver/cvirt.c    1970-01-01 01:00:00.000000000 +0100
19077 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/cvirt.c       2011-06-13 14:57:26.000000000 +0200
19078 @@ -0,0 +1,303 @@
19079 +/*
19080 + *  linux/kernel/vserver/cvirt.c
19081 + *
19082 + *  Virtual Server: Context Virtualization
19083 + *
19084 + *  Copyright (C) 2004-2007  Herbert Pötzl
19085 + *
19086 + *  V0.01  broken out from limit.c
19087 + *  V0.02  added utsname stuff
19088 + *  V0.03  changed vcmds to vxi arg
19089 + *
19090 + */
19091 +
19092 +#include <linux/types.h>
19093 +#include <linux/utsname.h>
19094 +#include <linux/vs_cvirt.h>
19095 +#include <linux/vserver/switch.h>
19096 +#include <linux/vserver/cvirt_cmd.h>
19097 +
19098 +#include <asm/uaccess.h>
19099 +
19100 +
19101 +void vx_vsi_uptime(struct timespec *uptime, struct timespec *idle)
19102 +{
19103 +       struct vx_info *vxi = current_vx_info();
19104 +
19105 +       set_normalized_timespec(uptime,
19106 +               uptime->tv_sec - vxi->cvirt.bias_uptime.tv_sec,
19107 +               uptime->tv_nsec - vxi->cvirt.bias_uptime.tv_nsec);
19108 +       if (!idle)
19109 +               return;
19110 +       set_normalized_timespec(idle,
19111 +               idle->tv_sec - vxi->cvirt.bias_idle.tv_sec,
19112 +               idle->tv_nsec - vxi->cvirt.bias_idle.tv_nsec);
19113 +       return;
19114 +}
19115 +
19116 +uint64_t vx_idle_jiffies(void)
19117 +{
19118 +       return init_task.utime + init_task.stime;
19119 +}
19120 +
19121 +
19122 +
19123 +static inline uint32_t __update_loadavg(uint32_t load,
19124 +       int wsize, int delta, int n)
19125 +{
19126 +       unsigned long long calc, prev;
19127 +
19128 +       /* just set it to n */
19129 +       if (unlikely(delta >= wsize))
19130 +               return (n << FSHIFT);
19131 +
19132 +       calc = delta * n;
19133 +       calc <<= FSHIFT;
19134 +       prev = (wsize - delta);
19135 +       prev *= load;
19136 +       calc += prev;
19137 +       do_div(calc, wsize);
19138 +       return calc;
19139 +}
19140 +
19141 +
19142 +void vx_update_load(struct vx_info *vxi)
19143 +{
19144 +       uint32_t now, last, delta;
19145 +       unsigned int nr_running, nr_uninterruptible;
19146 +       unsigned int total;
19147 +       unsigned long flags;
19148 +
19149 +       spin_lock_irqsave(&vxi->cvirt.load_lock, flags);
19150 +
19151 +       now = jiffies;
19152 +       last = vxi->cvirt.load_last;
19153 +       delta = now - last;
19154 +
19155 +       if (delta < 5*HZ)
19156 +               goto out;
19157 +
19158 +       nr_running = atomic_read(&vxi->cvirt.nr_running);
19159 +       nr_uninterruptible = atomic_read(&vxi->cvirt.nr_uninterruptible);
19160 +       total = nr_running + nr_uninterruptible;
19161 +
19162 +       vxi->cvirt.load[0] = __update_loadavg(vxi->cvirt.load[0],
19163 +               60*HZ, delta, total);
19164 +       vxi->cvirt.load[1] = __update_loadavg(vxi->cvirt.load[1],
19165 +               5*60*HZ, delta, total);
19166 +       vxi->cvirt.load[2] = __update_loadavg(vxi->cvirt.load[2],
19167 +               15*60*HZ, delta, total);
19168 +
19169 +       vxi->cvirt.load_last = now;
19170 +out:
19171 +       atomic_inc(&vxi->cvirt.load_updates);
19172 +       spin_unlock_irqrestore(&vxi->cvirt.load_lock, flags);
19173 +}
19174 +
19175 +
19176 +/*
19177 + * Commands to do_syslog:
19178 + *
19179 + *      0 -- Close the log.  Currently a NOP.
19180 + *      1 -- Open the log. Currently a NOP.
19181 + *      2 -- Read from the log.
19182 + *      3 -- Read all messages remaining in the ring buffer.
19183 + *      4 -- Read and clear all messages remaining in the ring buffer
19184 + *      5 -- Clear ring buffer.
19185 + *      6 -- Disable printk's to console
19186 + *      7 -- Enable printk's to console
19187 + *      8 -- Set level of messages printed to console
19188 + *      9 -- Return number of unread characters in the log buffer
19189 + *     10 -- Return size of the log buffer
19190 + */
19191 +int vx_do_syslog(int type, char __user *buf, int len)
19192 +{
19193 +       int error = 0;
19194 +       int do_clear = 0;
19195 +       struct vx_info *vxi = current_vx_info();
19196 +       struct _vx_syslog *log;
19197 +
19198 +       if (!vxi)
19199 +               return -EINVAL;
19200 +       log = &vxi->cvirt.syslog;
19201 +
19202 +       switch (type) {
19203 +       case 0:         /* Close log */
19204 +       case 1:         /* Open log */
19205 +               break;
19206 +       case 2:         /* Read from log */
19207 +               error = wait_event_interruptible(log->log_wait,
19208 +                       (log->log_start - log->log_end));
19209 +               if (error)
19210 +                       break;
19211 +               spin_lock_irq(&log->logbuf_lock);
19212 +               spin_unlock_irq(&log->logbuf_lock);
19213 +               break;
19214 +       case 4:         /* Read/clear last kernel messages */
19215 +               do_clear = 1;
19216 +               /* fall through */
19217 +       case 3:         /* Read last kernel messages */
19218 +               return 0;
19219 +
19220 +       case 5:         /* Clear ring buffer */
19221 +               return 0;
19222 +
19223 +       case 6:         /* Disable logging to console */
19224 +       case 7:         /* Enable logging to console */
19225 +       case 8:         /* Set level of messages printed to console */
19226 +               break;
19227 +
19228 +       case 9:         /* Number of chars in the log buffer */
19229 +               return 0;
19230 +       case 10:        /* Size of the log buffer */
19231 +               return 0;
19232 +       default:
19233 +               error = -EINVAL;
19234 +               break;
19235 +       }
19236 +       return error;
19237 +}
19238 +
19239 +
19240 +/* virtual host info names */
19241 +
19242 +static char *vx_vhi_name(struct vx_info *vxi, int id)
19243 +{
19244 +       struct nsproxy *nsproxy;
19245 +       struct uts_namespace *uts;
19246 +
19247 +       if (id == VHIN_CONTEXT)
19248 +               return vxi->vx_name;
19249 +
19250 +       nsproxy = vxi->space[0].vx_nsproxy;
19251 +       if (!nsproxy)
19252 +               return NULL;
19253 +
19254 +       uts = nsproxy->uts_ns;
19255 +       if (!uts)
19256 +               return NULL;
19257 +
19258 +       switch (id) {
19259 +       case VHIN_SYSNAME:
19260 +               return uts->name.sysname;
19261 +       case VHIN_NODENAME:
19262 +               return uts->name.nodename;
19263 +       case VHIN_RELEASE:
19264 +               return uts->name.release;
19265 +       case VHIN_VERSION:
19266 +               return uts->name.version;
19267 +       case VHIN_MACHINE:
19268 +               return uts->name.machine;
19269 +       case VHIN_DOMAINNAME:
19270 +               return uts->name.domainname;
19271 +       default:
19272 +               return NULL;
19273 +       }
19274 +       return NULL;
19275 +}
19276 +
19277 +int vc_set_vhi_name(struct vx_info *vxi, void __user *data)
19278 +{
19279 +       struct vcmd_vhi_name_v0 vc_data;
19280 +       char *name;
19281 +
19282 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19283 +               return -EFAULT;
19284 +
19285 +       name = vx_vhi_name(vxi, vc_data.field);
19286 +       if (!name)
19287 +               return -EINVAL;
19288 +
19289 +       memcpy(name, vc_data.name, 65);
19290 +       return 0;
19291 +}
19292 +
19293 +int vc_get_vhi_name(struct vx_info *vxi, void __user *data)
19294 +{
19295 +       struct vcmd_vhi_name_v0 vc_data;
19296 +       char *name;
19297 +
19298 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
19299 +               return -EFAULT;
19300 +
19301 +       name = vx_vhi_name(vxi, vc_data.field);
19302 +       if (!name)
19303 +               return -EINVAL;
19304 +
19305 +       memcpy(vc_data.name, name, 65);
19306 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19307 +               return -EFAULT;
19308 +       return 0;
19309 +}
19310 +
19311 +
19312 +int vc_virt_stat(struct vx_info *vxi, void __user *data)
19313 +{
19314 +       struct vcmd_virt_stat_v0 vc_data;
19315 +       struct _vx_cvirt *cvirt = &vxi->cvirt;
19316 +       struct timespec uptime;
19317 +
19318 +       do_posix_clock_monotonic_gettime(&uptime);
19319 +       set_normalized_timespec(&uptime,
19320 +               uptime.tv_sec - cvirt->bias_uptime.tv_sec,
19321 +               uptime.tv_nsec - cvirt->bias_uptime.tv_nsec);
19322 +
19323 +       vc_data.offset = timespec_to_ns(&cvirt->bias_ts);
19324 +       vc_data.uptime = timespec_to_ns(&uptime);
19325 +       vc_data.nr_threads = atomic_read(&cvirt->nr_threads);
19326 +       vc_data.nr_running = atomic_read(&cvirt->nr_running);
19327 +       vc_data.nr_uninterruptible = atomic_read(&cvirt->nr_uninterruptible);
19328 +       vc_data.nr_onhold = atomic_read(&cvirt->nr_onhold);
19329 +       vc_data.nr_forks = atomic_read(&cvirt->total_forks);
19330 +       vc_data.load[0] = cvirt->load[0];
19331 +       vc_data.load[1] = cvirt->load[1];
19332 +       vc_data.load[2] = cvirt->load[2];
19333 +
19334 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
19335 +               return -EFAULT;
19336 +       return 0;
19337 +}
19338 +
19339 +
19340 +#ifdef CONFIG_VSERVER_VTIME
19341 +
19342 +/* virtualized time base */
19343 +
19344 +void vx_adjust_timespec(struct timespec *ts)
19345 +{
19346 +       struct vx_info *vxi;
19347 +
19348 +       if (!vx_flags(VXF_VIRT_TIME, 0))
19349 +               return;
19350 +
19351 +       vxi = current_vx_info();
19352 +       ts->tv_sec += vxi->cvirt.bias_ts.tv_sec;
19353 +       ts->tv_nsec += vxi->cvirt.bias_ts.tv_nsec;
19354 +
19355 +       if (ts->tv_nsec >= NSEC_PER_SEC) {
19356 +               ts->tv_sec++;
19357 +               ts->tv_nsec -= NSEC_PER_SEC;
19358 +       } else if (ts->tv_nsec < 0) {
19359 +               ts->tv_sec--;
19360 +               ts->tv_nsec += NSEC_PER_SEC;
19361 +       }
19362 +}
19363 +
19364 +int vx_settimeofday(const struct timespec *ts)
19365 +{
19366 +       struct timespec ats, delta;
19367 +       struct vx_info *vxi;
19368 +
19369 +       if (!vx_flags(VXF_VIRT_TIME, 0))
19370 +               return do_settimeofday(ts);
19371 +
19372 +       getnstimeofday(&ats);
19373 +       delta = timespec_sub(*ts, ats);
19374 +
19375 +       vxi = current_vx_info();
19376 +       vxi->cvirt.bias_ts = timespec_add(vxi->cvirt.bias_ts, delta);
19377 +       return 0;
19378 +}
19379 +
19380 +#endif
19381 +
19382 diff -NurpP --minimal linux-3.0/kernel/vserver/cvirt_init.h linux-3.0-vs2.3.1-pre7/kernel/vserver/cvirt_init.h
19383 --- linux-3.0/kernel/vserver/cvirt_init.h       1970-01-01 01:00:00.000000000 +0100
19384 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/cvirt_init.h  2011-06-10 22:11:24.000000000 +0200
19385 @@ -0,0 +1,70 @@
19386 +
19387 +
19388 +extern uint64_t vx_idle_jiffies(void);
19389 +
19390 +static inline void vx_info_init_cvirt(struct _vx_cvirt *cvirt)
19391 +{
19392 +       uint64_t idle_jiffies = vx_idle_jiffies();
19393 +       uint64_t nsuptime;
19394 +
19395 +       do_posix_clock_monotonic_gettime(&cvirt->bias_uptime);
19396 +       nsuptime = (unsigned long long)cvirt->bias_uptime.tv_sec
19397 +               * NSEC_PER_SEC + cvirt->bias_uptime.tv_nsec;
19398 +       cvirt->bias_clock = nsec_to_clock_t(nsuptime);
19399 +       cvirt->bias_ts.tv_sec = 0;
19400 +       cvirt->bias_ts.tv_nsec = 0;
19401 +
19402 +       jiffies_to_timespec(idle_jiffies, &cvirt->bias_idle);
19403 +       atomic_set(&cvirt->nr_threads, 0);
19404 +       atomic_set(&cvirt->nr_running, 0);
19405 +       atomic_set(&cvirt->nr_uninterruptible, 0);
19406 +       atomic_set(&cvirt->nr_onhold, 0);
19407 +
19408 +       spin_lock_init(&cvirt->load_lock);
19409 +       cvirt->load_last = jiffies;
19410 +       atomic_set(&cvirt->load_updates, 0);
19411 +       cvirt->load[0] = 0;
19412 +       cvirt->load[1] = 0;
19413 +       cvirt->load[2] = 0;
19414 +       atomic_set(&cvirt->total_forks, 0);
19415 +
19416 +       spin_lock_init(&cvirt->syslog.logbuf_lock);
19417 +       init_waitqueue_head(&cvirt->syslog.log_wait);
19418 +       cvirt->syslog.log_start = 0;
19419 +       cvirt->syslog.log_end = 0;
19420 +       cvirt->syslog.con_start = 0;
19421 +       cvirt->syslog.logged_chars = 0;
19422 +}
19423 +
19424 +static inline
19425 +void vx_info_init_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
19426 +{
19427 +       // cvirt_pc->cpustat = { 0 };
19428 +}
19429 +
19430 +static inline void vx_info_exit_cvirt(struct _vx_cvirt *cvirt)
19431 +{
19432 +#ifdef CONFIG_VSERVER_WARN
19433 +       int value;
19434 +#endif
19435 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_threads)),
19436 +               "!!! cvirt: %p[nr_threads] = %d on exit.",
19437 +               cvirt, value);
19438 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_running)),
19439 +               "!!! cvirt: %p[nr_running] = %d on exit.",
19440 +               cvirt, value);
19441 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_uninterruptible)),
19442 +               "!!! cvirt: %p[nr_uninterruptible] = %d on exit.",
19443 +               cvirt, value);
19444 +       vxwprintk_xid((value = atomic_read(&cvirt->nr_onhold)),
19445 +               "!!! cvirt: %p[nr_onhold] = %d on exit.",
19446 +               cvirt, value);
19447 +       return;
19448 +}
19449 +
19450 +static inline
19451 +void vx_info_exit_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc, int cpu)
19452 +{
19453 +       return;
19454 +}
19455 +
19456 diff -NurpP --minimal linux-3.0/kernel/vserver/cvirt_proc.h linux-3.0-vs2.3.1-pre7/kernel/vserver/cvirt_proc.h
19457 --- linux-3.0/kernel/vserver/cvirt_proc.h       1970-01-01 01:00:00.000000000 +0100
19458 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/cvirt_proc.h  2011-06-10 22:11:24.000000000 +0200
19459 @@ -0,0 +1,135 @@
19460 +#ifndef _VX_CVIRT_PROC_H
19461 +#define _VX_CVIRT_PROC_H
19462 +
19463 +#include <linux/nsproxy.h>
19464 +#include <linux/mnt_namespace.h>
19465 +#include <linux/ipc_namespace.h>
19466 +#include <linux/utsname.h>
19467 +#include <linux/ipc.h>
19468 +
19469 +
19470 +static inline
19471 +int vx_info_proc_nsproxy(struct nsproxy *nsproxy, char *buffer)
19472 +{
19473 +       struct mnt_namespace *ns;
19474 +       struct uts_namespace *uts;
19475 +       struct ipc_namespace *ipc;
19476 +       struct path path;
19477 +       char *pstr, *root;
19478 +       int length = 0;
19479 +
19480 +       if (!nsproxy)
19481 +               goto out;
19482 +
19483 +       length += sprintf(buffer + length,
19484 +               "NSProxy:\t%p [%p,%p,%p]\n",
19485 +               nsproxy, nsproxy->mnt_ns,
19486 +               nsproxy->uts_ns, nsproxy->ipc_ns);
19487 +
19488 +       ns = nsproxy->mnt_ns;
19489 +       if (!ns)
19490 +               goto skip_ns;
19491 +
19492 +       pstr = kmalloc(PATH_MAX, GFP_KERNEL);
19493 +       if (!pstr)
19494 +               goto skip_ns;
19495 +
19496 +       path.mnt = ns->root;
19497 +       path.dentry = ns->root->mnt_root;
19498 +       root = d_path(&path, pstr, PATH_MAX - 2);
19499 +       length += sprintf(buffer + length,
19500 +               "Namespace:\t%p [#%u]\n"
19501 +               "RootPath:\t%s\n",
19502 +               ns, atomic_read(&ns->count),
19503 +               root);
19504 +       kfree(pstr);
19505 +skip_ns:
19506 +
19507 +       uts = nsproxy->uts_ns;
19508 +       if (!uts)
19509 +               goto skip_uts;
19510 +
19511 +       length += sprintf(buffer + length,
19512 +               "SysName:\t%.*s\n"
19513 +               "NodeName:\t%.*s\n"
19514 +               "Release:\t%.*s\n"
19515 +               "Version:\t%.*s\n"
19516 +               "Machine:\t%.*s\n"
19517 +               "DomainName:\t%.*s\n",
19518 +               __NEW_UTS_LEN, uts->name.sysname,
19519 +               __NEW_UTS_LEN, uts->name.nodename,
19520 +               __NEW_UTS_LEN, uts->name.release,
19521 +               __NEW_UTS_LEN, uts->name.version,
19522 +               __NEW_UTS_LEN, uts->name.machine,
19523 +               __NEW_UTS_LEN, uts->name.domainname);
19524 +skip_uts:
19525 +
19526 +       ipc = nsproxy->ipc_ns;
19527 +       if (!ipc)
19528 +               goto skip_ipc;
19529 +
19530 +       length += sprintf(buffer + length,
19531 +               "SEMS:\t\t%d %d %d %d  %d\n"
19532 +               "MSG:\t\t%d %d %d\n"
19533 +               "SHM:\t\t%lu %lu  %d %d\n",
19534 +               ipc->sem_ctls[0], ipc->sem_ctls[1],
19535 +               ipc->sem_ctls[2], ipc->sem_ctls[3],
19536 +               ipc->used_sems,
19537 +               ipc->msg_ctlmax, ipc->msg_ctlmnb, ipc->msg_ctlmni,
19538 +               (unsigned long)ipc->shm_ctlmax,
19539 +               (unsigned long)ipc->shm_ctlall,
19540 +               ipc->shm_ctlmni, ipc->shm_tot);
19541 +skip_ipc:
19542 +out:
19543 +       return length;
19544 +}
19545 +
19546 +
19547 +#include <linux/sched.h>
19548 +
19549 +#define LOAD_INT(x) ((x) >> FSHIFT)
19550 +#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1 - 1)) * 100)
19551 +
19552 +static inline
19553 +int vx_info_proc_cvirt(struct _vx_cvirt *cvirt, char *buffer)
19554 +{
19555 +       int length = 0;
19556 +       int a, b, c;
19557 +
19558 +       length += sprintf(buffer + length,
19559 +               "BiasUptime:\t%lu.%02lu\n",
19560 +               (unsigned long)cvirt->bias_uptime.tv_sec,
19561 +               (cvirt->bias_uptime.tv_nsec / (NSEC_PER_SEC / 100)));
19562 +
19563 +       a = cvirt->load[0] + (FIXED_1 / 200);
19564 +       b = cvirt->load[1] + (FIXED_1 / 200);
19565 +       c = cvirt->load[2] + (FIXED_1 / 200);
19566 +       length += sprintf(buffer + length,
19567 +               "nr_threads:\t%d\n"
19568 +               "nr_running:\t%d\n"
19569 +               "nr_unintr:\t%d\n"
19570 +               "nr_onhold:\t%d\n"
19571 +               "load_updates:\t%d\n"
19572 +               "loadavg:\t%d.%02d %d.%02d %d.%02d\n"
19573 +               "total_forks:\t%d\n",
19574 +               atomic_read(&cvirt->nr_threads),
19575 +               atomic_read(&cvirt->nr_running),
19576 +               atomic_read(&cvirt->nr_uninterruptible),
19577 +               atomic_read(&cvirt->nr_onhold),
19578 +               atomic_read(&cvirt->load_updates),
19579 +               LOAD_INT(a), LOAD_FRAC(a),
19580 +               LOAD_INT(b), LOAD_FRAC(b),
19581 +               LOAD_INT(c), LOAD_FRAC(c),
19582 +               atomic_read(&cvirt->total_forks));
19583 +       return length;
19584 +}
19585 +
19586 +static inline
19587 +int vx_info_proc_cvirt_pc(struct _vx_cvirt_pc *cvirt_pc,
19588 +       char *buffer, int cpu)
19589 +{
19590 +       int length = 0;
19591 +       return length;
19592 +}
19593 +
19594 +#endif /* _VX_CVIRT_PROC_H */
19595 diff -NurpP --minimal linux-3.0/kernel/vserver/debug.c linux-3.0-vs2.3.1-pre7/kernel/vserver/debug.c
19596 --- linux-3.0/kernel/vserver/debug.c    1970-01-01 01:00:00.000000000 +0100
19597 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/debug.c       2011-06-10 22:11:24.000000000 +0200
19598 @@ -0,0 +1,32 @@
19599 +/*
19600 + *  kernel/vserver/debug.c
19601 + *
19602 + *  Copyright (C) 2005-2007 Herbert Pötzl
19603 + *
19604 + *  V0.01  vx_info dump support
19605 + *
19606 + */
19607 +
19608 +#include <linux/module.h>
19609 +
19610 +#include <linux/vserver/context.h>
19611 +
19612 +
19613 +void   dump_vx_info(struct vx_info *vxi, int level)
19614 +{
19615 +       printk("vx_info %p[#%d, %d.%d, %4x]\n", vxi, vxi->vx_id,
19616 +               atomic_read(&vxi->vx_usecnt),
19617 +               atomic_read(&vxi->vx_tasks),
19618 +               vxi->vx_state);
19619 +       if (level > 0) {
19620 +               __dump_vx_limit(&vxi->limit);
19621 +               __dump_vx_sched(&vxi->sched);
19622 +               __dump_vx_cvirt(&vxi->cvirt);
19623 +               __dump_vx_cacct(&vxi->cacct);
19624 +       }
19625 +       printk("---\n");
19626 +}
19627 +
19628 +
19629 +EXPORT_SYMBOL_GPL(dump_vx_info);
19630 +
19631 diff -NurpP --minimal linux-3.0/kernel/vserver/device.c linux-3.0-vs2.3.1-pre7/kernel/vserver/device.c
19632 --- linux-3.0/kernel/vserver/device.c   1970-01-01 01:00:00.000000000 +0100
19633 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/device.c      2011-06-10 23:20:56.000000000 +0200
19634 @@ -0,0 +1,443 @@
19635 +/*
19636 + *  linux/kernel/vserver/device.c
19637 + *
19638 + *  Linux-VServer: Device Support
19639 + *
19640 + *  Copyright (C) 2006  Herbert Pötzl
19641 + *  Copyright (C) 2007  Daniel Hokka Zakrisson
19642 + *
19643 + *  V0.01  device mapping basics
19644 + *  V0.02  added defaults
19645 + *
19646 + */
19647 +
19648 +#include <linux/slab.h>
19649 +#include <linux/rcupdate.h>
19650 +#include <linux/fs.h>
19651 +#include <linux/namei.h>
19652 +#include <linux/hash.h>
19653 +
19654 +#include <asm/errno.h>
19655 +#include <asm/uaccess.h>
19656 +#include <linux/vserver/base.h>
19657 +#include <linux/vserver/debug.h>
19658 +#include <linux/vserver/context.h>
19659 +#include <linux/vserver/device.h>
19660 +#include <linux/vserver/device_cmd.h>
19661 +
19662 +
19663 +#define DMAP_HASH_BITS 4
19664 +
19665 +
19666 +struct vs_mapping {
19667 +       union {
19668 +               struct hlist_node hlist;
19669 +               struct list_head list;
19670 +       } u;
19671 +#define dm_hlist       u.hlist
19672 +#define dm_list                u.list
19673 +       xid_t xid;
19674 +       dev_t device;
19675 +       struct vx_dmap_target target;
19676 +};
19677 +
19678 +
19679 +static struct hlist_head dmap_main_hash[1 << DMAP_HASH_BITS];
19680 +
19681 +static DEFINE_SPINLOCK(dmap_main_hash_lock);
19682 +
19683 +static struct vx_dmap_target dmap_defaults[2] = {
19684 +       { .flags = DATTR_OPEN },
19685 +       { .flags = DATTR_OPEN },
19686 +};
19687 +
19688 +
19689 +struct kmem_cache *dmap_cachep __read_mostly;
19690 +
19691 +int __init dmap_cache_init(void)
19692 +{
19693 +       dmap_cachep = kmem_cache_create("dmap_cache",
19694 +               sizeof(struct vs_mapping), 0,
19695 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
19696 +       return 0;
19697 +}
19698 +
19699 +__initcall(dmap_cache_init);
19700 +
19701 +
19702 +static inline unsigned int __hashval(dev_t dev, int bits)
19703 +{
19704 +       return hash_long((unsigned long)dev, bits);
19705 +}
19706 +
19707 +
19708 +/*     __hash_mapping()
19709 + *     add the mapping to the hash table
19710 + */
19711 +static inline void __hash_mapping(struct vx_info *vxi, struct vs_mapping *vdm)
19712 +{
19713 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
19714 +       struct hlist_head *head, *hash = dmap_main_hash;
19715 +       int device = vdm->device;
19716 +
19717 +       spin_lock(hash_lock);
19718 +       vxdprintk(VXD_CBIT(misc, 8), "__hash_mapping: %p[#%d] %08x:%08x",
19719 +               vxi, vxi ? vxi->vx_id : 0, device, vdm->target.target);
19720 +
19721 +       head = &hash[__hashval(device, DMAP_HASH_BITS)];
19722 +       hlist_add_head(&vdm->dm_hlist, head);
19723 +       spin_unlock(hash_lock);
19724 +}
19725 +
19726 +
19727 +static inline int __mode_to_default(umode_t mode)
19728 +{
19729 +       switch (mode) {
19730 +       case S_IFBLK:
19731 +               return 0;
19732 +       case S_IFCHR:
19733 +               return 1;
19734 +       default:
19735 +               BUG();
19736 +       }
19737 +}
19738 +
19739 +
19740 +/*     __set_default()
19741 + *     set a default
19742 + */
19743 +static inline void __set_default(struct vx_info *vxi, umode_t mode,
19744 +       struct vx_dmap_target *vdmt)
19745 +{
19746 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
19747 +       spin_lock(hash_lock);
19748 +
19749 +       if (vxi)
19750 +               vxi->dmap.targets[__mode_to_default(mode)] = *vdmt;
19751 +       else
19752 +               dmap_defaults[__mode_to_default(mode)] = *vdmt;
19753 +
19754 +
19755 +       spin_unlock(hash_lock);
19756 +
19757 +       vxdprintk(VXD_CBIT(misc, 8), "__set_default: %p[#%u] %08x %04x",
19758 +                 vxi, vxi ? vxi->vx_id : 0, vdmt->target, vdmt->flags);
19759 +}
19760 +
19761 +
19762 +/*     __remove_default()
19763 + *     remove a default
19764 + */
19765 +static inline int __remove_default(struct vx_info *vxi, umode_t mode)
19766 +{
19767 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
19768 +       spin_lock(hash_lock);
19769 +
19770 +       if (vxi)
19771 +               vxi->dmap.targets[__mode_to_default(mode)].flags = 0;
19772 +       else    /* remove == reset */
19773 +               dmap_defaults[__mode_to_default(mode)].flags = DATTR_OPEN | mode;
19774 +
19775 +       spin_unlock(hash_lock);
19776 +       return 0;
19777 +}
19778 +
19779 +
19780 +/*     __find_mapping()
19781 + *     find a mapping in the hash table
19782 + *
19783 + *     caller must hold hash_lock
19784 + */
19785 +static inline int __find_mapping(xid_t xid, dev_t device, umode_t mode,
19786 +       struct vs_mapping **local, struct vs_mapping **global)
19787 +{
19788 +       struct hlist_head *hash = dmap_main_hash;
19789 +       struct hlist_head *head = &hash[__hashval(device, DMAP_HASH_BITS)];
19790 +       struct hlist_node *pos;
19791 +       struct vs_mapping *vdm;
19792 +
19793 +       *local = NULL;
19794 +       if (global)
19795 +               *global = NULL;
19796 +
19797 +       hlist_for_each(pos, head) {
19798 +               vdm = hlist_entry(pos, struct vs_mapping, dm_hlist);
19799 +
19800 +               if ((vdm->device == device) &&
19801 +                       !((vdm->target.flags ^ mode) & S_IFMT)) {
19802 +                       if (vdm->xid == xid) {
19803 +                               *local = vdm;
19804 +                               return 1;
19805 +                       } else if (global && vdm->xid == 0)
19806 +                               *global = vdm;
19807 +               }
19808 +       }
19809 +
19810 +       if (global && *global)
19811 +               return 0;
19812 +       else
19813 +               return -ENOENT;
19814 +}
19815 +
19816 +
19817 +/*     __lookup_mapping()
19818 + *     find a mapping and store the result in target and flags
19819 + */
19820 +static inline int __lookup_mapping(struct vx_info *vxi,
19821 +       dev_t device, dev_t *target, int *flags, umode_t mode)
19822 +{
19823 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
19824 +       struct vs_mapping *vdm, *global;
19825 +       struct vx_dmap_target *vdmt;
19826 +       int ret = 0;
19827 +       xid_t xid = vxi->vx_id;
19828 +       int index;
19829 +
19830 +       spin_lock(hash_lock);
19831 +       if (__find_mapping(xid, device, mode, &vdm, &global) > 0) {
19832 +               ret = 1;
19833 +               vdmt = &vdm->target;
19834 +               goto found;
19835 +       }
19836 +
19837 +       index = __mode_to_default(mode);
19838 +       if (vxi && vxi->dmap.targets[index].flags) {
19839 +               ret = 2;
19840 +               vdmt = &vxi->dmap.targets[index];
19841 +       } else if (global) {
19842 +               ret = 3;
19843 +               vdmt = &global->target;
19844 +               goto found;
19845 +       } else {
19846 +               ret = 4;
19847 +               vdmt = &dmap_defaults[index];
19848 +       }
19849 +
19850 +found:
19851 +       if (target && (vdmt->flags & DATTR_REMAP))
19852 +               *target = vdmt->target;
19853 +       else if (target)
19854 +               *target = device;
19855 +       if (flags)
19856 +               *flags = vdmt->flags;
19857 +
19858 +       spin_unlock(hash_lock);
19859 +
19860 +       return ret;
19861 +}
19862 +
19863 +
19864 +/*     __remove_mapping()
19865 + *     remove a mapping from the hash table
19866 + */
19867 +static inline int __remove_mapping(struct vx_info *vxi, dev_t device,
19868 +       umode_t mode)
19869 +{
19870 +       spinlock_t *hash_lock = &dmap_main_hash_lock;
19871 +       struct vs_mapping *vdm = NULL;
19872 +       int ret = 0;
19873 +
19874 +       spin_lock(hash_lock);
19875 +
19876 +       ret = __find_mapping((vxi ? vxi->vx_id : 0), device, mode, &vdm,
19877 +               NULL);
19878 +       vxdprintk(VXD_CBIT(misc, 8), "__remove_mapping: %p[#%d] %08x %04x",
19879 +               vxi, vxi ? vxi->vx_id : 0, device, mode);
19880 +       if (ret < 0)
19881 +               goto out;
19882 +       hlist_del(&vdm->dm_hlist);
19883 +
19884 +out:
19885 +       spin_unlock(hash_lock);
19886 +       if (vdm)
19887 +               kmem_cache_free(dmap_cachep, vdm);
19888 +       return ret;
19889 +}
19890 +
19891 +
19892 +
19893 +int vs_map_device(struct vx_info *vxi,
19894 +       dev_t device, dev_t *target, umode_t mode)
19895 +{
19896 +       int ret, flags = DATTR_MASK;
19897 +
19898 +       if (!vxi) {
19899 +               if (target)
19900 +                       *target = device;
19901 +               goto out;
19902 +       }
19903 +       ret = __lookup_mapping(vxi, device, target, &flags, mode);
19904 +       vxdprintk(VXD_CBIT(misc, 8), "vs_map_device: %08x target: %08x flags: %04x mode: %04x mapped=%d",
19905 +               device, target ? *target : 0, flags, mode, ret);
19906 +out:
19907 +       return (flags & DATTR_MASK);
19908 +}
19909 +
19910 +
19911 +
19912 +static int do_set_mapping(struct vx_info *vxi,
19913 +       dev_t device, dev_t target, int flags, umode_t mode)
19914 +{
19915 +       if (device) {
19916 +               struct vs_mapping *new;
19917 +
19918 +               new = kmem_cache_alloc(dmap_cachep, GFP_KERNEL);
19919 +               if (!new)
19920 +                       return -ENOMEM;
19921 +
19922 +               INIT_HLIST_NODE(&new->dm_hlist);
19923 +               new->device = device;
19924 +               new->target.target = target;
19925 +               new->target.flags = flags | mode;
19926 +               new->xid = (vxi ? vxi->vx_id : 0);
19927 +
19928 +               vxdprintk(VXD_CBIT(misc, 8), "do_set_mapping: %08x target: %08x flags: %04x", device, target, flags);
19929 +               __hash_mapping(vxi, new);
19930 +       } else {
19931 +               struct vx_dmap_target new = {
19932 +                       .target = target,
19933 +                       .flags = flags | mode,
19934 +               };
19935 +               __set_default(vxi, mode, &new);
19936 +       }
19937 +       return 0;
19938 +}
19939 +
19940 +
19941 +static int do_unset_mapping(struct vx_info *vxi,
19942 +       dev_t device, dev_t target, int flags, umode_t mode)
19943 +{
19944 +       int ret = -EINVAL;
19945 +
19946 +       if (device) {
19947 +               ret = __remove_mapping(vxi, device, mode);
19948 +               if (ret < 0)
19949 +                       goto out;
19950 +       } else {
19951 +               ret = __remove_default(vxi, mode);
19952 +               if (ret < 0)
19953 +                       goto out;
19954 +       }
19955 +
19956 +out:
19957 +       return ret;
19958 +}
19959 +
19960 +
19961 +static inline int __user_device(const char __user *name, dev_t *dev,
19962 +       umode_t *mode)
19963 +{
19964 +       struct nameidata nd;
19965 +       int ret;
19966 +
19967 +       if (!name) {
19968 +               *dev = 0;
19969 +               return 0;
19970 +       }
19971 +       ret = user_lpath(name, &nd.path);
19972 +       if (ret)
19973 +               return ret;
19974 +       if (nd.path.dentry->d_inode) {
19975 +               *dev = nd.path.dentry->d_inode->i_rdev;
19976 +               *mode = nd.path.dentry->d_inode->i_mode;
19977 +       }
19978 +       path_put(&nd.path);
19979 +       return 0;
19980 +}
19981 +
19982 +static inline int __mapping_mode(dev_t device, dev_t target,
19983 +       umode_t device_mode, umode_t target_mode, umode_t *mode)
19984 +{
19985 +       if (device)
19986 +               *mode = device_mode & S_IFMT;
19987 +       else if (target)
19988 +               *mode = target_mode & S_IFMT;
19989 +       else
19990 +               return -EINVAL;
19991 +
19992 +       /* if both given, device and target mode have to match */
19993 +       if (device && target &&
19994 +               ((device_mode ^ target_mode) & S_IFMT))
19995 +               return -EINVAL;
19996 +       return 0;
19997 +}
19998 +
19999 +
20000 +static inline int do_mapping(struct vx_info *vxi, const char __user *device_path,
20001 +       const char __user *target_path, int flags, int set)
20002 +{
20003 +       dev_t device = ~0, target = ~0;
20004 +       umode_t device_mode = 0, target_mode = 0, mode;
20005 +       int ret;
20006 +
20007 +       ret = __user_device(device_path, &device, &device_mode);
20008 +       if (ret)
20009 +               return ret;
20010 +       ret = __user_device(target_path, &target, &target_mode);
20011 +       if (ret)
20012 +               return ret;
20013 +
20014 +       ret = __mapping_mode(device, target,
20015 +               device_mode, target_mode, &mode);
20016 +       if (ret)
20017 +               return ret;
20018 +
20019 +       if (set)
20020 +               return do_set_mapping(vxi, device, target,
20021 +                       flags, mode);
20022 +       else
20023 +               return do_unset_mapping(vxi, device, target,
20024 +                       flags, mode);
20025 +}
20026 +
20027 +
20028 +int vc_set_mapping(struct vx_info *vxi, void __user *data)
20029 +{
20030 +       struct vcmd_set_mapping_v0 vc_data;
20031 +
20032 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20033 +               return -EFAULT;
20034 +
20035 +       return do_mapping(vxi, vc_data.device, vc_data.target,
20036 +               vc_data.flags, 1);
20037 +}
20038 +
20039 +int vc_unset_mapping(struct vx_info *vxi, void __user *data)
20040 +{
20041 +       struct vcmd_set_mapping_v0 vc_data;
20042 +
20043 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20044 +               return -EFAULT;
20045 +
20046 +       return do_mapping(vxi, vc_data.device, vc_data.target,
20047 +               vc_data.flags, 0);
20048 +}
20049 +
20050 +
20051 +#ifdef CONFIG_COMPAT
20052 +
20053 +int vc_set_mapping_x32(struct vx_info *vxi, void __user *data)
20054 +{
20055 +       struct vcmd_set_mapping_v0_x32 vc_data;
20056 +
20057 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20058 +               return -EFAULT;
20059 +
20060 +       return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
20061 +               compat_ptr(vc_data.target_ptr), vc_data.flags, 1);
20062 +}
20063 +
20064 +int vc_unset_mapping_x32(struct vx_info *vxi, void __user *data)
20065 +{
20066 +       struct vcmd_set_mapping_v0_x32 vc_data;
20067 +
20068 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20069 +               return -EFAULT;
20070 +
20071 +       return do_mapping(vxi, compat_ptr(vc_data.device_ptr),
20072 +               compat_ptr(vc_data.target_ptr), vc_data.flags, 0);
20073 +}
20074 +
20075 +#endif /* CONFIG_COMPAT */
20076 +
20077 +
20078 diff -NurpP --minimal linux-3.0/kernel/vserver/dlimit.c linux-3.0-vs2.3.1-pre7/kernel/vserver/dlimit.c
20079 --- linux-3.0/kernel/vserver/dlimit.c   1970-01-01 01:00:00.000000000 +0100
20080 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/dlimit.c      2011-06-10 23:20:56.000000000 +0200
20081 @@ -0,0 +1,531 @@
20082 +/*
20083 + *  linux/kernel/vserver/dlimit.c
20084 + *
20085 + *  Virtual Server: Context Disk Limits
20086 + *
20087 + *  Copyright (C) 2004-2009  Herbert Pötzl
20088 + *
20089 + *  V0.01  initial version
20090 + *  V0.02  compat32 splitup
20091 + *  V0.03  extended interface
20092 + *
20093 + */
20094 +
20095 +#include <linux/statfs.h>
20096 +#include <linux/sched.h>
20097 +#include <linux/namei.h>
20098 +#include <linux/vs_tag.h>
20099 +#include <linux/vs_dlimit.h>
20100 +#include <linux/vserver/dlimit_cmd.h>
20101 +#include <linux/slab.h>
20102 +// #include <linux/gfp.h>
20103 +
20104 +#include <asm/uaccess.h>
20105 +
20106 +/*     __alloc_dl_info()
20107 +
20108 +       * allocate an initialized dl_info struct
20109 +       * doesn't make it visible (hash)                        */
20110 +
20111 +static struct dl_info *__alloc_dl_info(struct super_block *sb, tag_t tag)
20112 +{
20113 +       struct dl_info *new = NULL;
20114 +
20115 +       vxdprintk(VXD_CBIT(dlim, 5),
20116 +               "alloc_dl_info(%p,%d)*", sb, tag);
20117 +
20118 +       /* would this benefit from a slab cache? */
20119 +       new = kmalloc(sizeof(struct dl_info), GFP_KERNEL);
20120 +       if (!new)
20121 +               return 0;
20122 +
20123 +       memset(new, 0, sizeof(struct dl_info));
20124 +       new->dl_tag = tag;
20125 +       new->dl_sb = sb;
20126 +       // INIT_RCU_HEAD(&new->dl_rcu);
20127 +       INIT_HLIST_NODE(&new->dl_hlist);
20128 +       spin_lock_init(&new->dl_lock);
20129 +       atomic_set(&new->dl_refcnt, 0);
20130 +       atomic_set(&new->dl_usecnt, 0);
20131 +
20132 +       /* rest of init goes here */
20133 +
20134 +       vxdprintk(VXD_CBIT(dlim, 4),
20135 +               "alloc_dl_info(%p,%d) = %p", sb, tag, new);
20136 +       return new;
20137 +}
20138 +
20139 +/*     __dealloc_dl_info()
20140 +
20141 +       * final disposal of dl_info                             */
20142 +
20143 +static void __dealloc_dl_info(struct dl_info *dli)
20144 +{
20145 +       vxdprintk(VXD_CBIT(dlim, 4),
20146 +               "dealloc_dl_info(%p)", dli);
20147 +
20148 +       dli->dl_hlist.next = LIST_POISON1;
20149 +       dli->dl_tag = -1;
20150 +       dli->dl_sb = 0;
20151 +
20152 +       BUG_ON(atomic_read(&dli->dl_usecnt));
20153 +       BUG_ON(atomic_read(&dli->dl_refcnt));
20154 +
20155 +       kfree(dli);
20156 +}
20157 +
20158 +
20159 +/*     hash table for dl_info hash */
20160 +
20161 +#define DL_HASH_SIZE   13
20162 +
20163 +struct hlist_head dl_info_hash[DL_HASH_SIZE];
20164 +
20165 +static DEFINE_SPINLOCK(dl_info_hash_lock);
20166 +
20167 +
20168 +static inline unsigned int __hashval(struct super_block *sb, tag_t tag)
20169 +{
20170 +       return ((tag ^ (unsigned long)sb) % DL_HASH_SIZE);
20171 +}
20172 +
20173 +
20174 +
20175 +/*     __hash_dl_info()
20176 +
20177 +       * add the dli to the global hash table
20178 +       * requires the hash_lock to be held                     */
20179 +
20180 +static inline void __hash_dl_info(struct dl_info *dli)
20181 +{
20182 +       struct hlist_head *head;
20183 +
20184 +       vxdprintk(VXD_CBIT(dlim, 6),
20185 +               "__hash_dl_info: %p[#%d]", dli, dli->dl_tag);
20186 +       get_dl_info(dli);
20187 +       head = &dl_info_hash[__hashval(dli->dl_sb, dli->dl_tag)];
20188 +       hlist_add_head_rcu(&dli->dl_hlist, head);
20189 +}
20190 +
20191 +/*     __unhash_dl_info()
20192 +
20193 +       * remove the dli from the global hash table
20194 +       * requires the hash_lock to be held                     */
20195 +
20196 +static inline void __unhash_dl_info(struct dl_info *dli)
20197 +{
20198 +       vxdprintk(VXD_CBIT(dlim, 6),
20199 +               "__unhash_dl_info: %p[#%d]", dli, dli->dl_tag);
20200 +       hlist_del_rcu(&dli->dl_hlist);
20201 +       put_dl_info(dli);
20202 +}
20203 +
20204 +
20205 +/*     __lookup_dl_info()
20206 +
20207 +       * requires the rcu_read_lock()
20208 +       * doesn't increment the dl_refcnt                       */
20209 +
20210 +static inline struct dl_info *__lookup_dl_info(struct super_block *sb, tag_t tag)
20211 +{
20212 +       struct hlist_head *head = &dl_info_hash[__hashval(sb, tag)];
20213 +       struct hlist_node *pos;
20214 +       struct dl_info *dli;
20215 +
20216 +       hlist_for_each_entry_rcu(dli, pos, head, dl_hlist) {
20217 +
20218 +               if (dli->dl_tag == tag && dli->dl_sb == sb) {
20219 +                       return dli;
20220 +               }
20221 +       }
20222 +       return NULL;
20223 +}
20224 +
20225 +
20226 +struct dl_info *locate_dl_info(struct super_block *sb, tag_t tag)
20227 +{
20228 +       struct dl_info *dli;
20229 +
20230 +       rcu_read_lock();
20231 +       dli = get_dl_info(__lookup_dl_info(sb, tag));
20232 +       vxdprintk(VXD_CBIT(dlim, 7),
20233 +               "locate_dl_info(%p,#%d) = %p", sb, tag, dli);
20234 +       rcu_read_unlock();
20235 +       return dli;
20236 +}
20237 +
20238 +void rcu_free_dl_info(struct rcu_head *head)
20239 +{
20240 +       struct dl_info *dli = container_of(head, struct dl_info, dl_rcu);
20241 +       int usecnt, refcnt;
20242 +
20243 +       BUG_ON(!dli || !head);
20244 +
20245 +       usecnt = atomic_read(&dli->dl_usecnt);
20246 +       BUG_ON(usecnt < 0);
20247 +
20248 +       refcnt = atomic_read(&dli->dl_refcnt);
20249 +       BUG_ON(refcnt < 0);
20250 +
20251 +       vxdprintk(VXD_CBIT(dlim, 3),
20252 +               "rcu_free_dl_info(%p)", dli);
20253 +       if (!usecnt)
20254 +               __dealloc_dl_info(dli);
20255 +       else
20256 +               printk("!!! rcu didn't free\n");
20257 +}
20258 +
20259 +
20260 +
20261 +
20262 +static int do_addrem_dlimit(uint32_t id, const char __user *name,
20263 +       uint32_t flags, int add)
20264 +{
20265 +       struct path path;
20266 +       int ret;
20267 +
20268 +       ret = user_lpath(name, &path);
20269 +       if (!ret) {
20270 +               struct super_block *sb;
20271 +               struct dl_info *dli;
20272 +
20273 +               ret = -EINVAL;
20274 +               if (!path.dentry->d_inode)
20275 +                       goto out_release;
20276 +               if (!(sb = path.dentry->d_inode->i_sb))
20277 +                       goto out_release;
20278 +
20279 +               if (add) {
20280 +                       dli = __alloc_dl_info(sb, id);
20281 +                       spin_lock(&dl_info_hash_lock);
20282 +
20283 +                       ret = -EEXIST;
20284 +                       if (__lookup_dl_info(sb, id))
20285 +                               goto out_unlock;
20286 +                       __hash_dl_info(dli);
20287 +                       dli = NULL;
20288 +               } else {
20289 +                       spin_lock(&dl_info_hash_lock);
20290 +                       dli = __lookup_dl_info(sb, id);
20291 +
20292 +                       ret = -ESRCH;
20293 +                       if (!dli)
20294 +                               goto out_unlock;
20295 +                       __unhash_dl_info(dli);
20296 +               }
20297 +               ret = 0;
20298 +       out_unlock:
20299 +               spin_unlock(&dl_info_hash_lock);
20300 +               if (add && dli)
20301 +                       __dealloc_dl_info(dli);
20302 +       out_release:
20303 +               path_put(&path);
20304 +       }
20305 +       return ret;
20306 +}
20307 +
20308 +int vc_add_dlimit(uint32_t id, void __user *data)
20309 +{
20310 +       struct vcmd_ctx_dlimit_base_v0 vc_data;
20311 +
20312 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20313 +               return -EFAULT;
20314 +
20315 +       return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 1);
20316 +}
20317 +
20318 +int vc_rem_dlimit(uint32_t id, void __user *data)
20319 +{
20320 +       struct vcmd_ctx_dlimit_base_v0 vc_data;
20321 +
20322 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20323 +               return -EFAULT;
20324 +
20325 +       return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 0);
20326 +}
20327 +
20328 +#ifdef CONFIG_COMPAT
20329 +
20330 +int vc_add_dlimit_x32(uint32_t id, void __user *data)
20331 +{
20332 +       struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
20333 +
20334 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20335 +               return -EFAULT;
20336 +
20337 +       return do_addrem_dlimit(id,
20338 +               compat_ptr(vc_data.name_ptr), vc_data.flags, 1);
20339 +}
20340 +
20341 +int vc_rem_dlimit_x32(uint32_t id, void __user *data)
20342 +{
20343 +       struct vcmd_ctx_dlimit_base_v0_x32 vc_data;
20344 +
20345 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20346 +               return -EFAULT;
20347 +
20348 +       return do_addrem_dlimit(id,
20349 +               compat_ptr(vc_data.name_ptr), vc_data.flags, 0);
20350 +}
20351 +
20352 +#endif /* CONFIG_COMPAT */
20353 +
20354 +
20355 +static inline
20356 +int do_set_dlimit(uint32_t id, const char __user *name,
20357 +       uint32_t space_used, uint32_t space_total,
20358 +       uint32_t inodes_used, uint32_t inodes_total,
20359 +       uint32_t reserved, uint32_t flags)
20360 +{
20361 +       struct path path;
20362 +       int ret;
20363 +
20364 +       ret = user_lpath(name, &path);
20365 +       if (!ret) {
20366 +               struct super_block *sb;
20367 +               struct dl_info *dli;
20368 +
20369 +               ret = -EINVAL;
20370 +               if (!path.dentry->d_inode)
20371 +                       goto out_release;
20372 +               if (!(sb = path.dentry->d_inode->i_sb))
20373 +                       goto out_release;
20374 +
20375 +               /* sanity checks */
20376 +               if ((reserved != CDLIM_KEEP &&
20377 +                       reserved > 100) ||
20378 +                       (inodes_used != CDLIM_KEEP &&
20379 +                       inodes_used > inodes_total) ||
20380 +                       (space_used != CDLIM_KEEP &&
20381 +                       space_used > space_total))
20382 +                       goto out_release;
20383 +
20384 +               ret = -ESRCH;
20385 +               dli = locate_dl_info(sb, id);
20386 +               if (!dli)
20387 +                       goto out_release;
20388 +
20389 +               spin_lock(&dli->dl_lock);
20390 +
20391 +               if (inodes_used != CDLIM_KEEP)
20392 +                       dli->dl_inodes_used = inodes_used;
20393 +               if (inodes_total != CDLIM_KEEP)
20394 +                       dli->dl_inodes_total = inodes_total;
20395 +               if (space_used != CDLIM_KEEP)
20396 +                       dli->dl_space_used = dlimit_space_32to64(
20397 +                               space_used, flags, DLIMS_USED);
20398 +
20399 +               if (space_total == CDLIM_INFINITY)
20400 +                       dli->dl_space_total = DLIM_INFINITY;
20401 +               else if (space_total != CDLIM_KEEP)
20402 +                       dli->dl_space_total = dlimit_space_32to64(
20403 +                               space_total, flags, DLIMS_TOTAL);
20404 +
20405 +               if (reserved != CDLIM_KEEP)
20406 +                       dli->dl_nrlmult = (1 << 10) * (100 - reserved) / 100;
20407 +
20408 +               spin_unlock(&dli->dl_lock);
20409 +
20410 +               put_dl_info(dli);
20411 +               ret = 0;
20412 +
20413 +       out_release:
20414 +               path_put(&path);
20415 +       }
20416 +       return ret;
20417 +}
20418 +
20419 +int vc_set_dlimit(uint32_t id, void __user *data)
20420 +{
20421 +       struct vcmd_ctx_dlimit_v0 vc_data;
20422 +
20423 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20424 +               return -EFAULT;
20425 +
20426 +       return do_set_dlimit(id, vc_data.name,
20427 +               vc_data.space_used, vc_data.space_total,
20428 +               vc_data.inodes_used, vc_data.inodes_total,
20429 +               vc_data.reserved, vc_data.flags);
20430 +}
20431 +
20432 +#ifdef CONFIG_COMPAT
20433 +
20434 +int vc_set_dlimit_x32(uint32_t id, void __user *data)
20435 +{
20436 +       struct vcmd_ctx_dlimit_v0_x32 vc_data;
20437 +
20438 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20439 +               return -EFAULT;
20440 +
20441 +       return do_set_dlimit(id, compat_ptr(vc_data.name_ptr),
20442 +               vc_data.space_used, vc_data.space_total,
20443 +               vc_data.inodes_used, vc_data.inodes_total,
20444 +               vc_data.reserved, vc_data.flags);
20445 +}
20446 +
20447 +#endif /* CONFIG_COMPAT */
20448 +
20449 +
20450 +static inline
20451 +int do_get_dlimit(uint32_t id, const char __user *name,
20452 +       uint32_t *space_used, uint32_t *space_total,
20453 +       uint32_t *inodes_used, uint32_t *inodes_total,
20454 +       uint32_t *reserved, uint32_t *flags)
20455 +{
20456 +       struct path path;
20457 +       int ret;
20458 +
20459 +       ret = user_lpath(name, &path);
20460 +       if (!ret) {
20461 +               struct super_block *sb;
20462 +               struct dl_info *dli;
20463 +
20464 +               ret = -EINVAL;
20465 +               if (!path.dentry->d_inode)
20466 +                       goto out_release;
20467 +               if (!(sb = path.dentry->d_inode->i_sb))
20468 +                       goto out_release;
20469 +
20470 +               ret = -ESRCH;
20471 +               dli = locate_dl_info(sb, id);
20472 +               if (!dli)
20473 +                       goto out_release;
20474 +
20475 +               spin_lock(&dli->dl_lock);
20476 +               *inodes_used = dli->dl_inodes_used;
20477 +               *inodes_total = dli->dl_inodes_total;
20478 +
20479 +               *space_used = dlimit_space_64to32(
20480 +                       dli->dl_space_used, flags, DLIMS_USED);
20481 +
20482 +               if (dli->dl_space_total == DLIM_INFINITY)
20483 +                       *space_total = CDLIM_INFINITY;
20484 +               else
20485 +                       *space_total = dlimit_space_64to32(
20486 +                               dli->dl_space_total, flags, DLIMS_TOTAL);
20487 +
20488 +               *reserved = 100 - ((dli->dl_nrlmult * 100 + 512) >> 10);
20489 +               spin_unlock(&dli->dl_lock);
20490 +
20491 +               put_dl_info(dli);
20492 +               ret = -EFAULT;
20493 +
20494 +               ret = 0;
20495 +       out_release:
20496 +               path_put(&path);
20497 +       }
20498 +       return ret;
20499 +}
20500 +
20501 +
20502 +int vc_get_dlimit(uint32_t id, void __user *data)
20503 +{
20504 +       struct vcmd_ctx_dlimit_v0 vc_data;
20505 +       int ret;
20506 +
20507 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20508 +               return -EFAULT;
20509 +
20510 +       ret = do_get_dlimit(id, vc_data.name,
20511 +               &vc_data.space_used, &vc_data.space_total,
20512 +               &vc_data.inodes_used, &vc_data.inodes_total,
20513 +               &vc_data.reserved, &vc_data.flags);
20514 +       if (ret)
20515 +               return ret;
20516 +
20517 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20518 +               return -EFAULT;
20519 +       return 0;
20520 +}
20521 +
20522 +#ifdef CONFIG_COMPAT
20523 +
20524 +int vc_get_dlimit_x32(uint32_t id, void __user *data)
20525 +{
20526 +       struct vcmd_ctx_dlimit_v0_x32 vc_data;
20527 +       int ret;
20528 +
20529 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
20530 +               return -EFAULT;
20531 +
20532 +       ret = do_get_dlimit(id, compat_ptr(vc_data.name_ptr),
20533 +               &vc_data.space_used, &vc_data.space_total,
20534 +               &vc_data.inodes_used, &vc_data.inodes_total,
20535 +               &vc_data.reserved, &vc_data.flags);
20536 +       if (ret)
20537 +               return ret;
20538 +
20539 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
20540 +               return -EFAULT;
20541 +       return 0;
20542 +}
20543 +
20544 +#endif /* CONFIG_COMPAT */
20545 +
20546 +
20547 +void vx_vsi_statfs(struct super_block *sb, struct kstatfs *buf)
20548 +{
20549 +       struct dl_info *dli;
20550 +       __u64 blimit, bfree, bavail;
20551 +       __u32 ifree;
20552 +
20553 +       dli = locate_dl_info(sb, dx_current_tag());
20554 +       if (!dli)
20555 +               return;
20556 +
20557 +       spin_lock(&dli->dl_lock);
20558 +       if (dli->dl_inodes_total == (unsigned long)DLIM_INFINITY)
20559 +               goto no_ilim;
20560 +
20561 +       /* reduce max inodes available to limit */
20562 +       if (buf->f_files > dli->dl_inodes_total)
20563 +               buf->f_files = dli->dl_inodes_total;
20564 +
20565 +       ifree = dli->dl_inodes_total - dli->dl_inodes_used;
20566 +       /* reduce free inodes to min */
20567 +       if (ifree < buf->f_ffree)
20568 +               buf->f_ffree = ifree;
20569 +
20570 +no_ilim:
20571 +       if (dli->dl_space_total == DLIM_INFINITY)
20572 +               goto no_blim;
20573 +
20574 +       blimit = dli->dl_space_total >> sb->s_blocksize_bits;
20575 +
20576 +       if (dli->dl_space_total < dli->dl_space_used)
20577 +               bfree = 0;
20578 +       else
20579 +               bfree = (dli->dl_space_total - dli->dl_space_used)
20580 +                       >> sb->s_blocksize_bits;
20581 +
20582 +       bavail = ((dli->dl_space_total >> 10) * dli->dl_nrlmult);
20583 +       if (bavail < dli->dl_space_used)
20584 +               bavail = 0;
20585 +       else
20586 +               bavail = (bavail - dli->dl_space_used)
20587 +                       >> sb->s_blocksize_bits;
20588 +
20589 +       /* reduce max space available to limit */
20590 +       if (buf->f_blocks > blimit)
20591 +               buf->f_blocks = blimit;
20592 +
20593 +       /* reduce free space to min */
20594 +       if (bfree < buf->f_bfree)
20595 +               buf->f_bfree = bfree;
20596 +
20597 +       /* reduce avail space to min */
20598 +       if (bavail < buf->f_bavail)
20599 +               buf->f_bavail = bavail;
20600 +
20601 +no_blim:
20602 +       spin_unlock(&dli->dl_lock);
20603 +       put_dl_info(dli);
20604 +
20605 +       return;
20606 +}
20607 +
20608 +#include <linux/module.h>
20609 +
20610 +EXPORT_SYMBOL_GPL(locate_dl_info);
20611 +EXPORT_SYMBOL_GPL(rcu_free_dl_info);
20612 +
20613 diff -NurpP --minimal linux-3.0/kernel/vserver/helper.c linux-3.0-vs2.3.1-pre7/kernel/vserver/helper.c
20614 --- linux-3.0/kernel/vserver/helper.c   1970-01-01 01:00:00.000000000 +0100
20615 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/helper.c      2011-06-10 22:11:24.000000000 +0200
20616 @@ -0,0 +1,223 @@
20617 +/*
20618 + *  linux/kernel/vserver/helper.c
20619 + *
20620 + *  Virtual Context Support
20621 + *
20622 + *  Copyright (C) 2004-2007  Herbert Pötzl
20623 + *
20624 + *  V0.01  basic helper
20625 + *
20626 + */
20627 +
20628 +#include <linux/kmod.h>
20629 +#include <linux/reboot.h>
20630 +#include <linux/vs_context.h>
20631 +#include <linux/vs_network.h>
20632 +#include <linux/vserver/signal.h>
20633 +
20634 +
20635 +char vshelper_path[255] = "/sbin/vshelper";
20636 +
20637 +
20638 +static int do_vshelper(char *name, char *argv[], char *envp[], int sync)
20639 +{
20640 +       int ret;
20641 +
20642 +       if ((ret = call_usermodehelper(name, argv, envp, sync))) {
20643 +               printk( KERN_WARNING
20644 +                       "%s: (%s %s) returned %s with %d\n",
20645 +                       name, argv[1], argv[2],
20646 +                       sync ? "sync" : "async", ret);
20647 +       }
20648 +       vxdprintk(VXD_CBIT(switch, 4),
20649 +               "%s: (%s %s) returned %s with %d",
20650 +               name, argv[1], argv[2], sync ? "sync" : "async", ret);
20651 +       return ret;
20652 +}
20653 +
20654 +/*
20655 + *      vshelper path is set via /proc/sys
20656 + *      invoked by vserver sys_reboot(), with
20657 + *      the following arguments
20658 + *
20659 + *      argv [0] = vshelper_path;
20660 + *      argv [1] = action: "restart", "halt", "poweroff", ...
20661 + *      argv [2] = context identifier
20662 + *
20663 + *      envp [*] = type-specific parameters
20664 + */
20665 +
20666 +long vs_reboot_helper(struct vx_info *vxi, int cmd, void __user *arg)
20667 +{
20668 +       char id_buf[8], cmd_buf[16];
20669 +       char uid_buf[16], pid_buf[16];
20670 +       int ret;
20671 +
20672 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
20673 +       char *envp[] = {"HOME=/", "TERM=linux",
20674 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin",
20675 +                       uid_buf, pid_buf, cmd_buf, 0};
20676 +
20677 +       if (vx_info_state(vxi, VXS_HELPER))
20678 +               return -EAGAIN;
20679 +       vxi->vx_state |= VXS_HELPER;
20680 +
20681 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
20682 +
20683 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
20684 +       snprintf(uid_buf, sizeof(uid_buf)-1, "VS_UID=%d", current_uid());
20685 +       snprintf(pid_buf, sizeof(pid_buf)-1, "VS_PID=%d", current->pid);
20686 +
20687 +       switch (cmd) {
20688 +       case LINUX_REBOOT_CMD_RESTART:
20689 +               argv[1] = "restart";
20690 +               break;
20691 +
20692 +       case LINUX_REBOOT_CMD_HALT:
20693 +               argv[1] = "halt";
20694 +               break;
20695 +
20696 +       case LINUX_REBOOT_CMD_POWER_OFF:
20697 +               argv[1] = "poweroff";
20698 +               break;
20699 +
20700 +       case LINUX_REBOOT_CMD_SW_SUSPEND:
20701 +               argv[1] = "swsusp";
20702 +               break;
20703 +
20704 +       case LINUX_REBOOT_CMD_OOM:
20705 +               argv[1] = "oom";
20706 +               break;
20707 +
20708 +       default:
20709 +               vxi->vx_state &= ~VXS_HELPER;
20710 +               return 0;
20711 +       }
20712 +
20713 +       ret = do_vshelper(vshelper_path, argv, envp, 0);
20714 +       vxi->vx_state &= ~VXS_HELPER;
20715 +       __wakeup_vx_info(vxi);
20716 +       return (ret) ? -EPERM : 0;
20717 +}
20718 +
20719 +
20720 +long vs_reboot(unsigned int cmd, void __user *arg)
20721 +{
20722 +       struct vx_info *vxi = current_vx_info();
20723 +       long ret = 0;
20724 +
20725 +       vxdprintk(VXD_CBIT(misc, 5),
20726 +               "vs_reboot(%p[#%d],%u)",
20727 +               vxi, vxi ? vxi->vx_id : 0, cmd);
20728 +
20729 +       ret = vs_reboot_helper(vxi, cmd, arg);
20730 +       if (ret)
20731 +               return ret;
20732 +
20733 +       vxi->reboot_cmd = cmd;
20734 +       if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
20735 +               switch (cmd) {
20736 +               case LINUX_REBOOT_CMD_RESTART:
20737 +               case LINUX_REBOOT_CMD_HALT:
20738 +               case LINUX_REBOOT_CMD_POWER_OFF:
20739 +                       vx_info_kill(vxi, 0, SIGKILL);
20740 +                       vx_info_kill(vxi, 1, SIGKILL);
20741 +               default:
20742 +                       break;
20743 +               }
20744 +       }
20745 +       return 0;
20746 +}
20747 +
20748 +long vs_oom_action(unsigned int cmd)
20749 +{
20750 +       struct vx_info *vxi = current_vx_info();
20751 +       long ret = 0;
20752 +
20753 +       vxdprintk(VXD_CBIT(misc, 5),
20754 +               "vs_oom_action(%p[#%d],%u)",
20755 +               vxi, vxi ? vxi->vx_id : 0, cmd);
20756 +
20757 +       ret = vs_reboot_helper(vxi, cmd, NULL);
20758 +       if (ret)
20759 +               return ret;
20760 +
20761 +       vxi->reboot_cmd = cmd;
20762 +       if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) {
20763 +               vx_info_kill(vxi, 0, SIGKILL);
20764 +               vx_info_kill(vxi, 1, SIGKILL);
20765 +       }
20766 +       return 0;
20767 +}
20768 +
20769 +/*
20770 + *      argv [0] = vshelper_path;
20771 + *      argv [1] = action: "startup", "shutdown"
20772 + *      argv [2] = context identifier
20773 + *
20774 + *      envp [*] = type-specific parameters
20775 + */
20776 +
20777 +long vs_state_change(struct vx_info *vxi, unsigned int cmd)
20778 +{
20779 +       char id_buf[8], cmd_buf[16];
20780 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
20781 +       char *envp[] = {"HOME=/", "TERM=linux",
20782 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
20783 +
20784 +       if (!vx_info_flags(vxi, VXF_SC_HELPER, 0))
20785 +               return 0;
20786 +
20787 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id);
20788 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
20789 +
20790 +       switch (cmd) {
20791 +       case VSC_STARTUP:
20792 +               argv[1] = "startup";
20793 +               break;
20794 +       case VSC_SHUTDOWN:
20795 +               argv[1] = "shutdown";
20796 +               break;
20797 +       default:
20798 +               return 0;
20799 +       }
20800 +
20801 +       return do_vshelper(vshelper_path, argv, envp, 1);
20802 +}
20803 +
20804 +
20805 +/*
20806 + *      argv [0] = vshelper_path;
20807 + *      argv [1] = action: "netup", "netdown"
20808 + *      argv [2] = context identifier
20809 + *
20810 + *      envp [*] = type-specific parameters
20811 + */
20812 +
20813 +long vs_net_change(struct nx_info *nxi, unsigned int cmd)
20814 +{
20815 +       char id_buf[8], cmd_buf[16];
20816 +       char *argv[] = {vshelper_path, NULL, id_buf, 0};
20817 +       char *envp[] = {"HOME=/", "TERM=linux",
20818 +                       "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0};
20819 +
20820 +       if (!nx_info_flags(nxi, NXF_SC_HELPER, 0))
20821 +               return 0;
20822 +
20823 +       snprintf(id_buf, sizeof(id_buf)-1, "%d", nxi->nx_id);
20824 +       snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd);
20825 +
20826 +       switch (cmd) {
20827 +       case VSC_NETUP:
20828 +               argv[1] = "netup";
20829 +               break;
20830 +       case VSC_NETDOWN:
20831 +               argv[1] = "netdown";
20832 +               break;
20833 +       default:
20834 +               return 0;
20835 +       }
20836 +
20837 +       return do_vshelper(vshelper_path, argv, envp, 1);
20838 +}
20839 +
20840 diff -NurpP --minimal linux-3.0/kernel/vserver/history.c linux-3.0-vs2.3.1-pre7/kernel/vserver/history.c
20841 --- linux-3.0/kernel/vserver/history.c  1970-01-01 01:00:00.000000000 +0100
20842 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/history.c     2011-06-10 22:11:24.000000000 +0200
20843 @@ -0,0 +1,258 @@
20844 +/*
20845 + *  kernel/vserver/history.c
20846 + *
20847 + *  Virtual Context History Backtrace
20848 + *
20849 + *  Copyright (C) 2004-2007  Herbert Pötzl
20850 + *
20851 + *  V0.01  basic structure
20852 + *  V0.02  hash/unhash and trace
20853 + *  V0.03  preemption fixes
20854 + *
20855 + */
20856 +
20857 +#include <linux/module.h>
20858 +#include <asm/uaccess.h>
20859 +
20860 +#include <linux/vserver/context.h>
20861 +#include <linux/vserver/debug.h>
20862 +#include <linux/vserver/debug_cmd.h>
20863 +#include <linux/vserver/history.h>
20864 +
20865 +
20866 +#ifdef CONFIG_VSERVER_HISTORY
20867 +#define VXH_SIZE       CONFIG_VSERVER_HISTORY_SIZE
20868 +#else
20869 +#define VXH_SIZE       64
20870 +#endif
20871 +
20872 +struct _vx_history {
20873 +       unsigned int counter;
20874 +
20875 +       struct _vx_hist_entry entry[VXH_SIZE + 1];
20876 +};
20877 +
20878 +
20879 +DEFINE_PER_CPU(struct _vx_history, vx_history_buffer);
20880 +
20881 +unsigned volatile int vxh_active = 1;
20882 +
20883 +static atomic_t sequence = ATOMIC_INIT(0);
20884 +
20885 +
20886 +/*     vxh_advance()
20887 +
20888 +       * requires disabled preemption                          */
20889 +
20890 +struct _vx_hist_entry *vxh_advance(void *loc)
20891 +{
20892 +       unsigned int cpu = smp_processor_id();
20893 +       struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
20894 +       struct _vx_hist_entry *entry;
20895 +       unsigned int index;
20896 +
20897 +       index = vxh_active ? (hist->counter++ % VXH_SIZE) : VXH_SIZE;
20898 +       entry = &hist->entry[index];
20899 +
20900 +       entry->seq = atomic_inc_return(&sequence);
20901 +       entry->loc = loc;
20902 +       return entry;
20903 +}
20904 +
20905 +EXPORT_SYMBOL_GPL(vxh_advance);
20906 +
20907 +
20908 +#define VXH_LOC_FMTS   "(#%04x,*%d):%p"
20909 +
20910 +#define VXH_LOC_ARGS(e)        (e)->seq, cpu, (e)->loc
20911 +
20912 +
20913 +#define VXH_VXI_FMTS   "%p[#%d,%d.%d]"
20914 +
20915 +#define VXH_VXI_ARGS(e)        (e)->vxi.ptr,                           \
20916 +                       (e)->vxi.ptr ? (e)->vxi.xid : 0,        \
20917 +                       (e)->vxi.ptr ? (e)->vxi.usecnt : 0,     \
20918 +                       (e)->vxi.ptr ? (e)->vxi.tasks : 0
20919 +
20920 +void   vxh_dump_entry(struct _vx_hist_entry *e, unsigned cpu)
20921 +{
20922 +       switch (e->type) {
20923 +       case VXH_THROW_OOPS:
20924 +               printk( VXH_LOC_FMTS " oops \n", VXH_LOC_ARGS(e));
20925 +               break;
20926 +
20927 +       case VXH_GET_VX_INFO:
20928 +       case VXH_PUT_VX_INFO:
20929 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
20930 +                       VXH_LOC_ARGS(e),
20931 +                       (e->type == VXH_GET_VX_INFO) ? "get" : "put",
20932 +                       VXH_VXI_ARGS(e));
20933 +               break;
20934 +
20935 +       case VXH_INIT_VX_INFO:
20936 +       case VXH_SET_VX_INFO:
20937 +       case VXH_CLR_VX_INFO:
20938 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
20939 +                       VXH_LOC_ARGS(e),
20940 +                       (e->type == VXH_INIT_VX_INFO) ? "init" :
20941 +                       ((e->type == VXH_SET_VX_INFO) ? "set" : "clr"),
20942 +                       VXH_VXI_ARGS(e), e->sc.data);
20943 +               break;
20944 +
20945 +       case VXH_CLAIM_VX_INFO:
20946 +       case VXH_RELEASE_VX_INFO:
20947 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n",
20948 +                       VXH_LOC_ARGS(e),
20949 +                       (e->type == VXH_CLAIM_VX_INFO) ? "claim" : "release",
20950 +                       VXH_VXI_ARGS(e), e->sc.data);
20951 +               break;
20952 +
20953 +       case VXH_ALLOC_VX_INFO:
20954 +       case VXH_DEALLOC_VX_INFO:
20955 +               printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n",
20956 +                       VXH_LOC_ARGS(e),
20957 +                       (e->type == VXH_ALLOC_VX_INFO) ? "alloc" : "dealloc",
20958 +                       VXH_VXI_ARGS(e));
20959 +               break;
20960 +
20961 +       case VXH_HASH_VX_INFO:
20962 +       case VXH_UNHASH_VX_INFO:
20963 +               printk( VXH_LOC_FMTS " __%s_vx_info " VXH_VXI_FMTS "\n",
20964 +                       VXH_LOC_ARGS(e),
20965 +                       (e->type == VXH_HASH_VX_INFO) ? "hash" : "unhash",
20966 +                       VXH_VXI_ARGS(e));
20967 +               break;
20968 +
20969 +       case VXH_LOC_VX_INFO:
20970 +       case VXH_LOOKUP_VX_INFO:
20971 +       case VXH_CREATE_VX_INFO:
20972 +               printk( VXH_LOC_FMTS " __%s_vx_info [#%d] -> " VXH_VXI_FMTS "\n",
20973 +                       VXH_LOC_ARGS(e),
20974 +                       (e->type == VXH_CREATE_VX_INFO) ? "create" :
20975 +                       ((e->type == VXH_LOC_VX_INFO) ? "loc" : "lookup"),
20976 +                       e->ll.arg, VXH_VXI_ARGS(e));
20977 +               break;
20978 +       }
20979 +}
20980 +
20981 +static void __vxh_dump_history(void)
20982 +{
20983 +       unsigned int i, cpu;
20984 +
20985 +       printk("History:\tSEQ: %8x\tNR_CPUS: %d\n",
20986 +               atomic_read(&sequence), NR_CPUS);
20987 +
20988 +       for (i = 0; i < VXH_SIZE; i++) {
20989 +               for_each_online_cpu(cpu) {
20990 +                       struct _vx_history *hist =
20991 +                               &per_cpu(vx_history_buffer, cpu);
20992 +                       unsigned int index = (hist->counter - i) % VXH_SIZE;
20993 +                       struct _vx_hist_entry *entry = &hist->entry[index];
20994 +
20995 +                       vxh_dump_entry(entry, cpu);
20996 +               }
20997 +       }
20998 +}
20999 +
21000 +void   vxh_dump_history(void)
21001 +{
21002 +       vxh_active = 0;
21003 +#ifdef CONFIG_SMP
21004 +       local_irq_enable();
21005 +       smp_send_stop();
21006 +       local_irq_disable();
21007 +#endif
21008 +       __vxh_dump_history();
21009 +}
21010 +
21011 +
21012 +/* vserver syscall commands below here */
21013 +
21014 +
21015 +int vc_dump_history(uint32_t id)
21016 +{
21017 +       vxh_active = 0;
21018 +       __vxh_dump_history();
21019 +       vxh_active = 1;
21020 +
21021 +       return 0;
21022 +}
21023 +
21024 +
21025 +int do_read_history(struct __user _vx_hist_entry *data,
21026 +       int cpu, uint32_t *index, uint32_t *count)
21027 +{
21028 +       int pos, ret = 0;
21029 +       struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu);
21030 +       int end = hist->counter;
21031 +       int start = end - VXH_SIZE + 2;
21032 +       int idx = *index;
21033 +
21034 +       /* special case: get current pos */
21035 +       if (!*count) {
21036 +               *index = end;
21037 +               return 0;
21038 +       }
21039 +
21040 +       /* have we lost some data? */
21041 +       if (idx < start)
21042 +               idx = start;
21043 +
21044 +       for (pos = 0; (pos < *count) && (idx < end); pos++, idx++) {
21045 +               struct _vx_hist_entry *entry =
21046 +                       &hist->entry[idx % VXH_SIZE];
21047 +
21048 +               /* send entry to userspace */
21049 +               ret = copy_to_user(&data[pos], entry, sizeof(*entry));
21050 +               if (ret)
21051 +                       break;
21052 +       }
21053 +       /* save new index and count */
21054 +       *index = idx;
21055 +       *count = pos;
21056 +       return ret ? ret : (*index < end);
21057 +}
21058 +
21059 +int vc_read_history(uint32_t id, void __user *data)
21060 +{
21061 +       struct vcmd_read_history_v0 vc_data;
21062 +       int ret;
21063 +
21064 +       if (id >= NR_CPUS)
21065 +               return -EINVAL;
21066 +
21067 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21068 +               return -EFAULT;
21069 +
21070 +       ret = do_read_history((struct __user _vx_hist_entry *)vc_data.data,
21071 +               id, &vc_data.index, &vc_data.count);
21072 +
21073 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21074 +               return -EFAULT;
21075 +       return ret;
21076 +}
21077 +
21078 +#ifdef CONFIG_COMPAT
21079 +
21080 +int vc_read_history_x32(uint32_t id, void __user *data)
21081 +{
21082 +       struct vcmd_read_history_v0_x32 vc_data;
21083 +       int ret;
21084 +
21085 +       if (id >= NR_CPUS)
21086 +               return -EINVAL;
21087 +
21088 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21089 +               return -EFAULT;
21090 +
21091 +       ret = do_read_history((struct __user _vx_hist_entry *)
21092 +               compat_ptr(vc_data.data_ptr),
21093 +               id, &vc_data.index, &vc_data.count);
21094 +
21095 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21096 +               return -EFAULT;
21097 +       return ret;
21098 +}
21099 +
21100 +#endif /* CONFIG_COMPAT */
21101 +
21102 diff -NurpP --minimal linux-3.0/kernel/vserver/inet.c linux-3.0-vs2.3.1-pre7/kernel/vserver/inet.c
21103 --- linux-3.0/kernel/vserver/inet.c     1970-01-01 01:00:00.000000000 +0100
21104 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/inet.c        2011-06-13 18:20:44.000000000 +0200
21105 @@ -0,0 +1,225 @@
21106 +
21107 +#include <linux/in.h>
21108 +#include <linux/inetdevice.h>
21109 +#include <linux/vs_inet.h>
21110 +#include <linux/vs_inet6.h>
21111 +#include <linux/vserver/debug.h>
21112 +#include <net/route.h>
21113 +#include <net/addrconf.h>
21114 +
21115 +
21116 +int nx_v4_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
21117 +{
21118 +       int ret = 0;
21119 +
21120 +       if (!nxi1 || !nxi2 || nxi1 == nxi2)
21121 +               ret = 1;
21122 +       else {
21123 +               struct nx_addr_v4 *ptr;
21124 +
21125 +               for (ptr = &nxi1->v4; ptr; ptr = ptr->next) {
21126 +                       if (v4_nx_addr_in_nx_info(nxi2, ptr, -1)) {
21127 +                               ret = 1;
21128 +                               break;
21129 +                       }
21130 +               }
21131 +       }
21132 +
21133 +       vxdprintk(VXD_CBIT(net, 2),
21134 +               "nx_v4_addr_conflict(%p,%p): %d",
21135 +               nxi1, nxi2, ret);
21136 +
21137 +       return ret;
21138 +}
21139 +
21140 +
21141 +#ifdef CONFIG_IPV6
21142 +
21143 +int nx_v6_addr_conflict(struct nx_info *nxi1, struct nx_info *nxi2)
21144 +{
21145 +       int ret = 0;
21146 +
21147 +       if (!nxi1 || !nxi2 || nxi1 == nxi2)
21148 +               ret = 1;
21149 +       else {
21150 +               struct nx_addr_v6 *ptr;
21151 +
21152 +               for (ptr = &nxi1->v6; ptr; ptr = ptr->next) {
21153 +                       if (v6_nx_addr_in_nx_info(nxi2, ptr, -1)) {
21154 +                               ret = 1;
21155 +                               break;
21156 +                       }
21157 +               }
21158 +       }
21159 +
21160 +       vxdprintk(VXD_CBIT(net, 2),
21161 +               "nx_v6_addr_conflict(%p,%p): %d",
21162 +               nxi1, nxi2, ret);
21163 +
21164 +       return ret;
21165 +}
21166 +
21167 +#endif
21168 +
21169 +int v4_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
21170 +{
21171 +       struct in_device *in_dev;
21172 +       struct in_ifaddr **ifap;
21173 +       struct in_ifaddr *ifa;
21174 +       int ret = 0;
21175 +
21176 +       if (!dev)
21177 +               goto out;
21178 +       in_dev = in_dev_get(dev);
21179 +       if (!in_dev)
21180 +               goto out;
21181 +
21182 +       for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
21183 +               ifap = &ifa->ifa_next) {
21184 +               if (v4_addr_in_nx_info(nxi, ifa->ifa_local, NXA_MASK_SHOW)) {
21185 +                       ret = 1;
21186 +                       break;
21187 +               }
21188 +       }
21189 +       in_dev_put(in_dev);
21190 +out:
21191 +       return ret;
21192 +}
21193 +
21194 +
21195 +#ifdef CONFIG_IPV6
21196 +
21197 +int v6_dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
21198 +{
21199 +       struct inet6_dev *in_dev;
21200 +       struct inet6_ifaddr *ifa;
21201 +       int ret = 0;
21202 +
21203 +       if (!dev)
21204 +               goto out;
21205 +       in_dev = in6_dev_get(dev);
21206 +       if (!in_dev)
21207 +               goto out;
21208 +
21209 +       // for (ifap = &in_dev->addr_list; (ifa = *ifap) != NULL;
21210 +       list_for_each_entry(ifa, &in_dev->addr_list, if_list) {
21211 +               if (v6_addr_in_nx_info(nxi, &ifa->addr, -1)) {
21212 +                       ret = 1;
21213 +                       break;
21214 +               }
21215 +       }
21216 +       in6_dev_put(in_dev);
21217 +out:
21218 +       return ret;
21219 +}
21220 +
21221 +#endif
21222 +
21223 +int dev_in_nx_info(struct net_device *dev, struct nx_info *nxi)
21224 +{
21225 +       int ret = 1;
21226 +
21227 +       if (!nxi)
21228 +               goto out;
21229 +       if (nxi->v4.type && v4_dev_in_nx_info(dev, nxi))
21230 +               goto out;
21231 +#ifdef CONFIG_IPV6
21232 +       ret = 2;
21233 +       if (nxi->v6.type && v6_dev_in_nx_info(dev, nxi))
21234 +               goto out;
21235 +#endif
21236 +       ret = 0;
21237 +out:
21238 +       vxdprintk(VXD_CBIT(net, 3),
21239 +               "dev_in_nx_info(%p,%p[#%d]) = %d",
21240 +               dev, nxi, nxi ? nxi->nx_id : 0, ret);
21241 +       return ret;
21242 +}
21243 +
21244 +struct rtable *ip_v4_find_src(struct net *net, struct nx_info *nxi,
21245 +       struct flowi4 *fl4)
21246 +{
21247 +       struct rtable *rt;
21248 +
21249 +       if (!nxi)
21250 +               return NULL;
21251 +
21252 +       /* FIXME: handle lback only case */
21253 +       if (!NX_IPV4(nxi))
21254 +               return ERR_PTR(EPERM);
21255 +
21256 +       vxdprintk(VXD_CBIT(net, 4),
21257 +               "ip_v4_find_src(%p[#%u]) " NIPQUAD_FMT " -> " NIPQUAD_FMT,
21258 +               nxi, nxi ? nxi->nx_id : 0,
21259 +               NIPQUAD(fl4->saddr), NIPQUAD(fl4->daddr));
21260 +
21261 +       /* single IP is unconditional */
21262 +       if (nx_info_flags(nxi, NXF_SINGLE_IP, 0) &&
21263 +               (fl4->saddr == INADDR_ANY))
21264 +               fl4->saddr = nxi->v4.ip[0].s_addr;
21265 +
21266 +       if (fl4->saddr == INADDR_ANY) {
21267 +               struct nx_addr_v4 *ptr;
21268 +               __be32 found = 0;
21269 +
21270 +               rt = __ip_route_output_key(net, fl4);
21271 +               if (!IS_ERR(rt)) {
21272 +                       found = fl4->saddr;
21273 +                       ip_rt_put(rt);
21274 +                       vxdprintk(VXD_CBIT(net, 4),
21275 +                               "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
21276 +                               nxi, nxi ? nxi->nx_id : 0, fl4->flowi4_oif, NIPQUAD(found));
21277 +                       if (v4_addr_in_nx_info(nxi, found, NXA_MASK_BIND))
21278 +                               goto found;
21279 +               }
21280 +
21281 +               for (ptr = &nxi->v4; ptr; ptr = ptr->next) {
21282 +                       __be32 primary = ptr->ip[0].s_addr;
21283 +                       __be32 mask = ptr->mask.s_addr;
21284 +                       __be32 neta = primary & mask;
21285 +
21286 +                       vxdprintk(VXD_CBIT(net, 4), "ip_v4_find_src(%p[#%u]) chk: "
21287 +                               NIPQUAD_FMT "/" NIPQUAD_FMT "/" NIPQUAD_FMT,
21288 +                               nxi, nxi ? nxi->nx_id : 0, NIPQUAD(primary),
21289 +                               NIPQUAD(mask), NIPQUAD(neta));
21290 +                       if ((found & mask) != neta)
21291 +                               continue;
21292 +
21293 +                       fl4->saddr = primary;
21294 +                       rt = __ip_route_output_key(net, fl4);
21295 +                       vxdprintk(VXD_CBIT(net, 4),
21296 +                               "ip_v4_find_src(%p[#%u]) rok[%u]: " NIPQUAD_FMT,
21297 +                               nxi, nxi ? nxi->nx_id : 0, fl4->flowi4_oif, NIPQUAD(primary));
21298 +                       if (!IS_ERR(rt)) {
21299 +                               found = fl4->saddr;
21300 +                               ip_rt_put(rt);
21301 +                               if (found == primary)
21302 +                                       goto found;
21303 +                       }
21304 +               }
21305 +               /* still no source ip? */
21306 +               found = ipv4_is_loopback(fl4->daddr)
21307 +                       ? IPI_LOOPBACK : nxi->v4.ip[0].s_addr;
21308 +       found:
21309 +               /* assign src ip to flow */
21310 +               fl4->saddr = found;
21311 +
21312 +       } else {
21313 +               if (!v4_addr_in_nx_info(nxi, fl4->saddr, NXA_MASK_BIND))
21314 +                       return ERR_PTR(EPERM);
21315 +       }
21316 +
21317 +       if (nx_info_flags(nxi, NXF_LBACK_REMAP, 0)) {
21318 +               if (ipv4_is_loopback(fl4->daddr))
21319 +                       fl4->daddr = nxi->v4_lback.s_addr;
21320 +               if (ipv4_is_loopback(fl4->saddr))
21321 +                       fl4->saddr = nxi->v4_lback.s_addr;
21322 +       } else if (ipv4_is_loopback(fl4->daddr) &&
21323 +               !nx_info_flags(nxi, NXF_LBACK_ALLOW, 0))
21324 +               return ERR_PTR(EPERM);
21325 +
21326 +       return NULL;
21327 +}
21328 +
21329 +EXPORT_SYMBOL_GPL(ip_v4_find_src);
21330 +
21331 diff -NurpP --minimal linux-3.0/kernel/vserver/init.c linux-3.0-vs2.3.1-pre7/kernel/vserver/init.c
21332 --- linux-3.0/kernel/vserver/init.c     1970-01-01 01:00:00.000000000 +0100
21333 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/init.c        2011-06-10 22:11:24.000000000 +0200
21334 @@ -0,0 +1,45 @@
21335 +/*
21336 + *  linux/kernel/init.c
21337 + *
21338 + *  Virtual Server Init
21339 + *
21340 + *  Copyright (C) 2004-2007  Herbert Pötzl
21341 + *
21342 + *  V0.01  basic structure
21343 + *
21344 + */
21345 +
21346 +#include <linux/init.h>
21347 +
21348 +int    vserver_register_sysctl(void);
21349 +void   vserver_unregister_sysctl(void);
21350 +
21351 +
21352 +static int __init init_vserver(void)
21353 +{
21354 +       int ret = 0;
21355 +
21356 +#ifdef CONFIG_VSERVER_DEBUG
21357 +       vserver_register_sysctl();
21358 +#endif
21359 +       return ret;
21360 +}
21361 +
21362 +
21363 +static void __exit exit_vserver(void)
21364 +{
21365 +
21366 +#ifdef CONFIG_VSERVER_DEBUG
21367 +       vserver_unregister_sysctl();
21368 +#endif
21369 +       return;
21370 +}
21371 +
21372 +/* FIXME: GFP_ZONETYPES gone
21373 +long vx_slab[GFP_ZONETYPES]; */
21374 +long vx_area;
21375 +
21376 +
21377 +module_init(init_vserver);
21378 +module_exit(exit_vserver);
21379 +
21380 diff -NurpP --minimal linux-3.0/kernel/vserver/inode.c linux-3.0-vs2.3.1-pre7/kernel/vserver/inode.c
21381 --- linux-3.0/kernel/vserver/inode.c    1970-01-01 01:00:00.000000000 +0100
21382 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/inode.c       2011-06-10 22:11:24.000000000 +0200
21383 @@ -0,0 +1,437 @@
21384 +/*
21385 + *  linux/kernel/vserver/inode.c
21386 + *
21387 + *  Virtual Server: File System Support
21388 + *
21389 + *  Copyright (C) 2004-2007  Herbert Pötzl
21390 + *
21391 + *  V0.01  separated from vcontext V0.05
21392 + *  V0.02  moved to tag (instead of xid)
21393 + *
21394 + */
21395 +
21396 +#include <linux/tty.h>
21397 +#include <linux/proc_fs.h>
21398 +#include <linux/devpts_fs.h>
21399 +#include <linux/fs.h>
21400 +#include <linux/file.h>
21401 +#include <linux/mount.h>
21402 +#include <linux/parser.h>
21403 +#include <linux/namei.h>
21404 +#include <linux/vserver/inode.h>
21405 +#include <linux/vserver/inode_cmd.h>
21406 +#include <linux/vs_base.h>
21407 +#include <linux/vs_tag.h>
21408 +
21409 +#include <asm/uaccess.h>
21410 +
21411 +
21412 +static int __vc_get_iattr(struct inode *in, uint32_t *tag, uint32_t *flags, uint32_t *mask)
21413 +{
21414 +       struct proc_dir_entry *entry;
21415 +
21416 +       if (!in || !in->i_sb)
21417 +               return -ESRCH;
21418 +
21419 +       *flags = IATTR_TAG
21420 +               | (IS_IMMUTABLE(in) ? IATTR_IMMUTABLE : 0)
21421 +               | (IS_IXUNLINK(in) ? IATTR_IXUNLINK : 0)
21422 +               | (IS_BARRIER(in) ? IATTR_BARRIER : 0)
21423 +               | (IS_COW(in) ? IATTR_COW : 0);
21424 +       *mask = IATTR_IXUNLINK | IATTR_IMMUTABLE | IATTR_COW;
21425 +
21426 +       if (S_ISDIR(in->i_mode))
21427 +               *mask |= IATTR_BARRIER;
21428 +
21429 +       if (IS_TAGGED(in)) {
21430 +               *tag = in->i_tag;
21431 +               *mask |= IATTR_TAG;
21432 +       }
21433 +
21434 +       switch (in->i_sb->s_magic) {
21435 +       case PROC_SUPER_MAGIC:
21436 +               entry = PROC_I(in)->pde;
21437 +
21438 +               /* check for specific inodes? */
21439 +               if (entry)
21440 +                       *mask |= IATTR_FLAGS;
21441 +               if (entry)
21442 +                       *flags |= (entry->vx_flags & IATTR_FLAGS);
21443 +               else
21444 +                       *flags |= (PROC_I(in)->vx_flags & IATTR_FLAGS);
21445 +               break;
21446 +
21447 +       case DEVPTS_SUPER_MAGIC:
21448 +               *tag = in->i_tag;
21449 +               *mask |= IATTR_TAG;
21450 +               break;
21451 +
21452 +       default:
21453 +               break;
21454 +       }
21455 +       return 0;
21456 +}
21457 +
21458 +int vc_get_iattr(void __user *data)
21459 +{
21460 +       struct path path;
21461 +       struct vcmd_ctx_iattr_v1 vc_data = { .tag = -1 };
21462 +       int ret;
21463 +
21464 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21465 +               return -EFAULT;
21466 +
21467 +       ret = user_lpath(vc_data.name, &path);
21468 +       if (!ret) {
21469 +               ret = __vc_get_iattr(path.dentry->d_inode,
21470 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
21471 +               path_put(&path);
21472 +       }
21473 +       if (ret)
21474 +               return ret;
21475 +
21476 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21477 +               ret = -EFAULT;
21478 +       return ret;
21479 +}
21480 +
21481 +#ifdef CONFIG_COMPAT
21482 +
21483 +int vc_get_iattr_x32(void __user *data)
21484 +{
21485 +       struct path path;
21486 +       struct vcmd_ctx_iattr_v1_x32 vc_data = { .tag = -1 };
21487 +       int ret;
21488 +
21489 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21490 +               return -EFAULT;
21491 +
21492 +       ret = user_lpath(compat_ptr(vc_data.name_ptr), &path);
21493 +       if (!ret) {
21494 +               ret = __vc_get_iattr(path.dentry->d_inode,
21495 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
21496 +               path_put(&path);
21497 +       }
21498 +       if (ret)
21499 +               return ret;
21500 +
21501 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21502 +               ret = -EFAULT;
21503 +       return ret;
21504 +}
21505 +
21506 +#endif /* CONFIG_COMPAT */
21507 +
21508 +
21509 +int vc_fget_iattr(uint32_t fd, void __user *data)
21510 +{
21511 +       struct file *filp;
21512 +       struct vcmd_ctx_fiattr_v0 vc_data = { .tag = -1 };
21513 +       int ret;
21514 +
21515 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21516 +               return -EFAULT;
21517 +
21518 +       filp = fget(fd);
21519 +       if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
21520 +               return -EBADF;
21521 +
21522 +       ret = __vc_get_iattr(filp->f_dentry->d_inode,
21523 +               &vc_data.tag, &vc_data.flags, &vc_data.mask);
21524 +
21525 +       fput(filp);
21526 +
21527 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21528 +               ret = -EFAULT;
21529 +       return ret;
21530 +}
21531 +
21532 +
21533 +static int __vc_set_iattr(struct dentry *de, uint32_t *tag, uint32_t *flags, uint32_t *mask)
21534 +{
21535 +       struct inode *in = de->d_inode;
21536 +       int error = 0, is_proc = 0, has_tag = 0;
21537 +       struct iattr attr = { 0 };
21538 +
21539 +       if (!in || !in->i_sb)
21540 +               return -ESRCH;
21541 +
21542 +       is_proc = (in->i_sb->s_magic == PROC_SUPER_MAGIC);
21543 +       if ((*mask & IATTR_FLAGS) && !is_proc)
21544 +               return -EINVAL;
21545 +
21546 +       has_tag = IS_TAGGED(in) ||
21547 +               (in->i_sb->s_magic == DEVPTS_SUPER_MAGIC);
21548 +       if ((*mask & IATTR_TAG) && !has_tag)
21549 +               return -EINVAL;
21550 +
21551 +       mutex_lock(&in->i_mutex);
21552 +       if (*mask & IATTR_TAG) {
21553 +               attr.ia_tag = *tag;
21554 +               attr.ia_valid |= ATTR_TAG;
21555 +       }
21556 +
21557 +       if (*mask & IATTR_FLAGS) {
21558 +               struct proc_dir_entry *entry = PROC_I(in)->pde;
21559 +               unsigned int iflags = PROC_I(in)->vx_flags;
21560 +
21561 +               iflags = (iflags & ~(*mask & IATTR_FLAGS))
21562 +                       | (*flags & IATTR_FLAGS);
21563 +               PROC_I(in)->vx_flags = iflags;
21564 +               if (entry)
21565 +                       entry->vx_flags = iflags;
21566 +       }
21567 +
21568 +       if (*mask & (IATTR_IMMUTABLE | IATTR_IXUNLINK |
21569 +               IATTR_BARRIER | IATTR_COW)) {
21570 +               int iflags = in->i_flags;
21571 +               int vflags = in->i_vflags;
21572 +
21573 +               if (*mask & IATTR_IMMUTABLE) {
21574 +                       if (*flags & IATTR_IMMUTABLE)
21575 +                               iflags |= S_IMMUTABLE;
21576 +                       else
21577 +                               iflags &= ~S_IMMUTABLE;
21578 +               }
21579 +               if (*mask & IATTR_IXUNLINK) {
21580 +                       if (*flags & IATTR_IXUNLINK)
21581 +                               iflags |= S_IXUNLINK;
21582 +                       else
21583 +                               iflags &= ~S_IXUNLINK;
21584 +               }
21585 +               if (S_ISDIR(in->i_mode) && (*mask & IATTR_BARRIER)) {
21586 +                       if (*flags & IATTR_BARRIER)
21587 +                               vflags |= V_BARRIER;
21588 +                       else
21589 +                               vflags &= ~V_BARRIER;
21590 +               }
21591 +               if (S_ISREG(in->i_mode) && (*mask & IATTR_COW)) {
21592 +                       if (*flags & IATTR_COW)
21593 +                               vflags |= V_COW;
21594 +                       else
21595 +                               vflags &= ~V_COW;
21596 +               }
21597 +               if (in->i_op && in->i_op->sync_flags) {
21598 +                       error = in->i_op->sync_flags(in, iflags, vflags);
21599 +                       if (error)
21600 +                               goto out;
21601 +               }
21602 +       }
21603 +
21604 +       if (attr.ia_valid) {
21605 +               if (in->i_op && in->i_op->setattr)
21606 +                       error = in->i_op->setattr(de, &attr);
21607 +               else {
21608 +                       error = inode_change_ok(in, &attr);
21609 +                       if (!error) {
21610 +                               setattr_copy(in, &attr);
21611 +                               mark_inode_dirty(in);
21612 +                       }
21613 +               }
21614 +       }
21615 +
21616 +out:
21617 +       mutex_unlock(&in->i_mutex);
21618 +       return error;
21619 +}
21620 +
21621 +int vc_set_iattr(void __user *data)
21622 +{
21623 +       struct path path;
21624 +       struct vcmd_ctx_iattr_v1 vc_data;
21625 +       int ret;
21626 +
21627 +       if (!capable(CAP_LINUX_IMMUTABLE))
21628 +               return -EPERM;
21629 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21630 +               return -EFAULT;
21631 +
21632 +       ret = user_lpath(vc_data.name, &path);
21633 +       if (!ret) {
21634 +               ret = __vc_set_iattr(path.dentry,
21635 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
21636 +               path_put(&path);
21637 +       }
21638 +
21639 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21640 +               ret = -EFAULT;
21641 +       return ret;
21642 +}
21643 +
21644 +#ifdef CONFIG_COMPAT
21645 +
21646 +int vc_set_iattr_x32(void __user *data)
21647 +{
21648 +       struct path path;
21649 +       struct vcmd_ctx_iattr_v1_x32 vc_data;
21650 +       int ret;
21651 +
21652 +       if (!capable(CAP_LINUX_IMMUTABLE))
21653 +               return -EPERM;
21654 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21655 +               return -EFAULT;
21656 +
21657 +       ret = user_lpath(compat_ptr(vc_data.name_ptr), &path);
21658 +       if (!ret) {
21659 +               ret = __vc_set_iattr(path.dentry,
21660 +                       &vc_data.tag, &vc_data.flags, &vc_data.mask);
21661 +               path_put(&path);
21662 +       }
21663 +
21664 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21665 +               ret = -EFAULT;
21666 +       return ret;
21667 +}
21668 +
21669 +#endif /* CONFIG_COMPAT */
21670 +
21671 +int vc_fset_iattr(uint32_t fd, void __user *data)
21672 +{
21673 +       struct file *filp;
21674 +       struct vcmd_ctx_fiattr_v0 vc_data;
21675 +       int ret;
21676 +
21677 +       if (!capable(CAP_LINUX_IMMUTABLE))
21678 +               return -EPERM;
21679 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21680 +               return -EFAULT;
21681 +
21682 +       filp = fget(fd);
21683 +       if (!filp || !filp->f_dentry || !filp->f_dentry->d_inode)
21684 +               return -EBADF;
21685 +
21686 +       ret = __vc_set_iattr(filp->f_dentry, &vc_data.tag,
21687 +               &vc_data.flags, &vc_data.mask);
21688 +
21689 +       fput(filp);
21690 +
21691 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21692 +               return -EFAULT;
21693 +       return ret;
21694 +}
21695 +
21696 +
21697 +enum { Opt_notagcheck, Opt_tag, Opt_notag, Opt_tagid, Opt_err };
21698 +
21699 +static match_table_t tokens = {
21700 +       {Opt_notagcheck, "notagcheck"},
21701 +#ifdef CONFIG_PROPAGATE
21702 +       {Opt_notag, "notag"},
21703 +       {Opt_tag, "tag"},
21704 +       {Opt_tagid, "tagid=%u"},
21705 +#endif
21706 +       {Opt_err, NULL}
21707 +};
21708 +
21709 +
21710 +static void __dx_parse_remove(char *string, char *opt)
21711 +{
21712 +       char *p = strstr(string, opt);
21713 +       char *q = p;
21714 +
21715 +       if (p) {
21716 +               while (*q != '\0' && *q != ',')
21717 +                       q++;
21718 +               while (*q)
21719 +                       *p++ = *q++;
21720 +               while (*p)
21721 +                       *p++ = '\0';
21722 +       }
21723 +}
21724 +
21725 +int dx_parse_tag(char *string, tag_t *tag, int remove, int *mnt_flags,
21726 +                unsigned long *flags)
21727 +{
21728 +       int set = 0;
21729 +       substring_t args[MAX_OPT_ARGS];
21730 +       int token;
21731 +       char *s, *p, *opts;
21732 +#if defined(CONFIG_PROPAGATE) || defined(CONFIG_VSERVER_WARN)
21733 +       int option = 0;
21734 +#endif
21735 +
21736 +       if (!string)
21737 +               return 0;
21738 +       s = kstrdup(string, GFP_KERNEL | GFP_ATOMIC);
21739 +       if (!s)
21740 +               return 0;
21741 +
21742 +       opts = s;
21743 +       while ((p = strsep(&opts, ",")) != NULL) {
21744 +               token = match_token(p, tokens, args);
21745 +
21746 +               switch (token) {
21747 +#ifdef CONFIG_PROPAGATE
21748 +               case Opt_tag:
21749 +                       if (tag)
21750 +                               *tag = 0;
21751 +                       if (remove)
21752 +                               __dx_parse_remove(s, "tag");
21753 +                       *mnt_flags |= MNT_TAGID;
21754 +                       set |= MNT_TAGID;
21755 +                       break;
21756 +               case Opt_notag:
21757 +                       if (remove)
21758 +                               __dx_parse_remove(s, "notag");
21759 +                       *mnt_flags |= MNT_NOTAG;
21760 +                       set |= MNT_NOTAG;
21761 +                       break;
21762 +               case Opt_tagid:
21763 +                       if (tag && !match_int(args, &option))
21764 +                               *tag = option;
21765 +                       if (remove)
21766 +                               __dx_parse_remove(s, "tagid");
21767 +                       *mnt_flags |= MNT_TAGID;
21768 +                       set |= MNT_TAGID;
21769 +                       break;
21770 +#endif
21771 +               case Opt_notagcheck:
21772 +                       if (remove)
21773 +                               __dx_parse_remove(s, "notagcheck");
21774 +                       *flags |= MS_NOTAGCHECK;
21775 +                       set |= MS_NOTAGCHECK;
21776 +                       break;
21777 +               }
21778 +               vxdprintk(VXD_CBIT(tag, 7),
21779 +                       "dx_parse_tag(" VS_Q("%s") "): %d:#%d",
21780 +                       p, token, option);
21781 +       }
21782 +       if (set)
21783 +               strcpy(string, s);
21784 +       kfree(s);
21785 +       return set;
21786 +}
21787 +
21788 +#ifdef CONFIG_PROPAGATE
21789 +
21790 +void __dx_propagate_tag(struct nameidata *nd, struct inode *inode)
21791 +{
21792 +       tag_t new_tag = 0;
21793 +       struct vfsmount *mnt;
21794 +       int propagate;
21795 +
21796 +       if (!nd)
21797 +               return;
21798 +       mnt = nd->path.mnt;
21799 +       if (!mnt)
21800 +               return;
21801 +
21802 +       propagate = (mnt->mnt_flags & MNT_TAGID);
21803 +       if (propagate)
21804 +               new_tag = mnt->mnt_tag;
21805 +
21806 +       vxdprintk(VXD_CBIT(tag, 7),
21807 +               "dx_propagate_tag(%p[#%lu.%d]): %d,%d",
21808 +               inode, inode->i_ino, inode->i_tag,
21809 +               new_tag, (propagate) ? 1 : 0);
21810 +
21811 +       if (propagate)
21812 +               inode->i_tag = new_tag;
21813 +}
21814 +
21815 +#include <linux/module.h>
21816 +
21817 +EXPORT_SYMBOL_GPL(__dx_propagate_tag);
21818 +
21819 +#endif /* CONFIG_PROPAGATE */
21820 +
21821 diff -NurpP --minimal linux-3.0/kernel/vserver/limit.c linux-3.0-vs2.3.1-pre7/kernel/vserver/limit.c
21822 --- linux-3.0/kernel/vserver/limit.c    1970-01-01 01:00:00.000000000 +0100
21823 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/limit.c       2011-06-10 22:11:24.000000000 +0200
21824 @@ -0,0 +1,354 @@
21825 +/*
21826 + *  linux/kernel/vserver/limit.c
21827 + *
21828 + *  Virtual Server: Context Limits
21829 + *
21830 + *  Copyright (C) 2004-2010  Herbert Pötzl
21831 + *
21832 + *  V0.01  broken out from vcontext V0.05
21833 + *  V0.02  changed vcmds to vxi arg
21834 + *  V0.03  added memory cgroup support
21835 + *
21836 + */
21837 +
21838 +#include <linux/sched.h>
21839 +#include <linux/module.h>
21840 +#include <linux/memcontrol.h>
21841 +#include <linux/res_counter.h>
21842 +#include <linux/vs_limit.h>
21843 +#include <linux/vserver/limit.h>
21844 +#include <linux/vserver/limit_cmd.h>
21845 +
21846 +#include <asm/uaccess.h>
21847 +
21848 +
21849 +const char *vlimit_name[NUM_LIMITS] = {
21850 +#ifdef CONFIG_VSERVER_LEGACY_MEM
21851 +       [RLIMIT_RSS]            = "RSS",
21852 +       [RLIMIT_AS]             = "VM",
21853 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
21854 +       [RLIMIT_CPU]            = "CPU",
21855 +       [RLIMIT_NPROC]          = "NPROC",
21856 +       [RLIMIT_NOFILE]         = "NOFILE",
21857 +       [RLIMIT_LOCKS]          = "LOCKS",
21858 +       [RLIMIT_SIGPENDING]     = "SIGP",
21859 +       [RLIMIT_MSGQUEUE]       = "MSGQ",
21860 +
21861 +       [VLIMIT_NSOCK]          = "NSOCK",
21862 +       [VLIMIT_OPENFD]         = "OPENFD",
21863 +       [VLIMIT_SHMEM]          = "SHMEM",
21864 +       [VLIMIT_DENTRY]         = "DENTRY",
21865 +};
21866 +
21867 +EXPORT_SYMBOL_GPL(vlimit_name);
21868 +
21869 +#define MASK_ENTRY(x)  (1 << (x))
21870 +
21871 +const struct vcmd_ctx_rlimit_mask_v0 vlimit_mask = {
21872 +               /* minimum */
21873 +       0
21874 +       ,       /* softlimit */
21875 +#ifdef CONFIG_VSERVER_LEGACY_MEM
21876 +       MASK_ENTRY( RLIMIT_RSS          ) |
21877 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
21878 +       0
21879 +       ,       /* maximum */
21880 +#ifdef CONFIG_VSERVER_LEGACY_MEM
21881 +       MASK_ENTRY( RLIMIT_RSS          ) |
21882 +       MASK_ENTRY( RLIMIT_AS           ) |
21883 +#endif /* CONFIG_VSERVER_LEGACY_MEM */
21884 +       MASK_ENTRY( RLIMIT_NPROC        ) |
21885 +       MASK_ENTRY( RLIMIT_NOFILE       ) |
21886 +       MASK_ENTRY( RLIMIT_LOCKS        ) |
21887 +       MASK_ENTRY( RLIMIT_MSGQUEUE     ) |
21888 +
21889 +       MASK_ENTRY( VLIMIT_NSOCK        ) |
21890 +       MASK_ENTRY( VLIMIT_OPENFD       ) |
21891 +       MASK_ENTRY( VLIMIT_SHMEM        ) |
21892 +       MASK_ENTRY( VLIMIT_DENTRY       ) |
21893 +       0
21894 +};
21895 +               /* accounting only */
21896 +uint32_t account_mask =
21897 +       MASK_ENTRY( VLIMIT_SEMARY       ) |
21898 +       MASK_ENTRY( VLIMIT_NSEMS        ) |
21899 +       MASK_ENTRY( VLIMIT_MAPPED       ) |
21900 +       0;
21901 +
21902 +
21903 +static int is_valid_vlimit(int id)
21904 +{
21905 +       uint32_t mask = vlimit_mask.minimum |
21906 +               vlimit_mask.softlimit | vlimit_mask.maximum;
21907 +       return mask & (1 << id);
21908 +}
21909 +
21910 +static int is_accounted_vlimit(int id)
21911 +{
21912 +       if (is_valid_vlimit(id))
21913 +               return 1;
21914 +       return account_mask & (1 << id);
21915 +}
21916 +
21917 +
21918 +static inline uint64_t vc_get_soft(struct vx_info *vxi, int id)
21919 +{
21920 +       rlim_t limit = __rlim_soft(&vxi->limit, id);
21921 +       return VX_VLIM(limit);
21922 +}
21923 +
21924 +static inline uint64_t vc_get_hard(struct vx_info *vxi, int id)
21925 +{
21926 +       rlim_t limit = __rlim_hard(&vxi->limit, id);
21927 +       return VX_VLIM(limit);
21928 +}
21929 +
21930 +static int do_get_rlimit(struct vx_info *vxi, uint32_t id,
21931 +       uint64_t *minimum, uint64_t *softlimit, uint64_t *maximum)
21932 +{
21933 +       if (!is_valid_vlimit(id))
21934 +               return -EINVAL;
21935 +
21936 +       if (minimum)
21937 +               *minimum = CRLIM_UNSET;
21938 +       if (softlimit)
21939 +               *softlimit = vc_get_soft(vxi, id);
21940 +       if (maximum)
21941 +               *maximum = vc_get_hard(vxi, id);
21942 +       return 0;
21943 +}
21944 +
21945 +int vc_get_rlimit(struct vx_info *vxi, void __user *data)
21946 +{
21947 +       struct vcmd_ctx_rlimit_v0 vc_data;
21948 +       int ret;
21949 +
21950 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21951 +               return -EFAULT;
21952 +
21953 +       ret = do_get_rlimit(vxi, vc_data.id,
21954 +               &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
21955 +       if (ret)
21956 +               return ret;
21957 +
21958 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
21959 +               return -EFAULT;
21960 +       return 0;
21961 +}
21962 +
21963 +static int do_set_rlimit(struct vx_info *vxi, uint32_t id,
21964 +       uint64_t minimum, uint64_t softlimit, uint64_t maximum)
21965 +{
21966 +       if (!is_valid_vlimit(id))
21967 +               return -EINVAL;
21968 +
21969 +       if (maximum != CRLIM_KEEP)
21970 +               __rlim_hard(&vxi->limit, id) = VX_RLIM(maximum);
21971 +       if (softlimit != CRLIM_KEEP)
21972 +               __rlim_soft(&vxi->limit, id) = VX_RLIM(softlimit);
21973 +
21974 +       /* clamp soft limit */
21975 +       if (__rlim_soft(&vxi->limit, id) > __rlim_hard(&vxi->limit, id))
21976 +               __rlim_soft(&vxi->limit, id) = __rlim_hard(&vxi->limit, id);
21977 +
21978 +       return 0;
21979 +}
21980 +
21981 +int vc_set_rlimit(struct vx_info *vxi, void __user *data)
21982 +{
21983 +       struct vcmd_ctx_rlimit_v0 vc_data;
21984 +
21985 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21986 +               return -EFAULT;
21987 +
21988 +       return do_set_rlimit(vxi, vc_data.id,
21989 +               vc_data.minimum, vc_data.softlimit, vc_data.maximum);
21990 +}
21991 +
21992 +#ifdef CONFIG_IA32_EMULATION
21993 +
21994 +int vc_set_rlimit_x32(struct vx_info *vxi, void __user *data)
21995 +{
21996 +       struct vcmd_ctx_rlimit_v0_x32 vc_data;
21997 +
21998 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
21999 +               return -EFAULT;
22000 +
22001 +       return do_set_rlimit(vxi, vc_data.id,
22002 +               vc_data.minimum, vc_data.softlimit, vc_data.maximum);
22003 +}
22004 +
22005 +int vc_get_rlimit_x32(struct vx_info *vxi, void __user *data)
22006 +{
22007 +       struct vcmd_ctx_rlimit_v0_x32 vc_data;
22008 +       int ret;
22009 +
22010 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22011 +               return -EFAULT;
22012 +
22013 +       ret = do_get_rlimit(vxi, vc_data.id,
22014 +               &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum);
22015 +       if (ret)
22016 +               return ret;
22017 +
22018 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22019 +               return -EFAULT;
22020 +       return 0;
22021 +}
22022 +
22023 +#endif /* CONFIG_IA32_EMULATION */
22024 +
22025 +
22026 +int vc_get_rlimit_mask(uint32_t id, void __user *data)
22027 +{
22028 +       if (copy_to_user(data, &vlimit_mask, sizeof(vlimit_mask)))
22029 +               return -EFAULT;
22030 +       return 0;
22031 +}
22032 +
22033 +
22034 +static inline void vx_reset_hits(struct _vx_limit *limit)
22035 +{
22036 +       int lim;
22037 +
22038 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22039 +               atomic_set(&__rlim_lhit(limit, lim), 0);
22040 +       }
22041 +}
22042 +
22043 +int vc_reset_hits(struct vx_info *vxi, void __user *data)
22044 +{
22045 +       vx_reset_hits(&vxi->limit);
22046 +       return 0;
22047 +}
22048 +
22049 +static inline void vx_reset_minmax(struct _vx_limit *limit)
22050 +{
22051 +       rlim_t value;
22052 +       int lim;
22053 +
22054 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22055 +               value = __rlim_get(limit, lim);
22056 +               __rlim_rmax(limit, lim) = value;
22057 +               __rlim_rmin(limit, lim) = value;
22058 +       }
22059 +}
22060 +
22061 +int vc_reset_minmax(struct vx_info *vxi, void __user *data)
22062 +{
22063 +       vx_reset_minmax(&vxi->limit);
22064 +       return 0;
22065 +}
22066 +
22067 +
22068 +int vc_rlimit_stat(struct vx_info *vxi, void __user *data)
22069 +{
22070 +       struct vcmd_rlimit_stat_v0 vc_data;
22071 +       struct _vx_limit *limit = &vxi->limit;
22072 +       int id;
22073 +
22074 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
22075 +               return -EFAULT;
22076 +
22077 +       id = vc_data.id;
22078 +       if (!is_accounted_vlimit(id))
22079 +               return -EINVAL;
22080 +
22081 +       vx_limit_fixup(limit, id);
22082 +       vc_data.hits = atomic_read(&__rlim_lhit(limit, id));
22083 +       vc_data.value = __rlim_get(limit, id);
22084 +       vc_data.minimum = __rlim_rmin(limit, id);
22085 +       vc_data.maximum = __rlim_rmax(limit, id);
22086 +
22087 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22088 +               return -EFAULT;
22089 +       return 0;
22090 +}
22091 +
22092 +
22093 +void vx_vsi_meminfo(struct sysinfo *val)
22094 +{
22095 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
22096 +       struct mem_cgroup *mcg = mem_cgroup_from_task(current);
22097 +       u64 res_limit, res_usage;
22098 +
22099 +       if (!mcg)
22100 +               return;
22101 +
22102 +       res_limit = mem_cgroup_res_read_u64(mcg, RES_LIMIT);
22103 +       res_usage = mem_cgroup_res_read_u64(mcg, RES_USAGE);
22104 +
22105 +       if (res_limit != RESOURCE_MAX)
22106 +               val->totalram = (res_limit >> PAGE_SHIFT);
22107 +       val->freeram = val->totalram - (res_usage >> PAGE_SHIFT);
22108 +       val->bufferram = 0;
22109 +       val->totalhigh = 0;
22110 +       val->freehigh = 0;
22111 +#endif /* CONFIG_CGROUP_MEM_RES_CTLR */
22112 +       return;
22113 +}
22114 +
22115 +void vx_vsi_swapinfo(struct sysinfo *val)
22116 +{
22117 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
22118 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
22119 +       struct mem_cgroup *mcg = mem_cgroup_from_task(current);
22120 +       u64 res_limit, res_usage, memsw_limit, memsw_usage;
22121 +       s64 swap_limit, swap_usage;
22122 +
22123 +       if (!mcg)
22124 +               return;
22125 +
22126 +       res_limit = mem_cgroup_res_read_u64(mcg, RES_LIMIT);
22127 +       res_usage = mem_cgroup_res_read_u64(mcg, RES_USAGE);
22128 +       memsw_limit = mem_cgroup_memsw_read_u64(mcg, RES_LIMIT);
22129 +       memsw_usage = mem_cgroup_memsw_read_u64(mcg, RES_USAGE);
22130 +
22131 +       if (res_limit == RESOURCE_MAX)
22132 +               return;
22133 +
22134 +       swap_limit = memsw_limit - res_limit;
22135 +       if (memsw_limit != RESOURCE_MAX)
22136 +               val->totalswap = swap_limit >> PAGE_SHIFT;
22137 +
22138 +       swap_usage = memsw_usage - res_usage;
22139 +       val->freeswap = (swap_usage < swap_limit) ?
22140 +               val->totalswap - (swap_usage >> PAGE_SHIFT) : 0;
22141 +#else  /* !CONFIG_CGROUP_MEM_RES_CTLR_SWAP */
22142 +       val->totalswap = 0;
22143 +       val->freeswap = 0;
22144 +#endif /* !CONFIG_CGROUP_MEM_RES_CTLR_SWAP */
22145 +#endif /* CONFIG_CGROUP_MEM_RES_CTLR */
22146 +       return;
22147 +}
22148 +
22149 +long vx_vsi_cached(struct sysinfo *val)
22150 +{
22151 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
22152 +       struct mem_cgroup *mcg = mem_cgroup_from_task(current);
22153 +
22154 +       return mem_cgroup_stat_read_cache(mcg);
22155 +#else
22156 +       return 0;
22157 +#endif
22158 +}
22159 +
22160 +
22161 +unsigned long vx_badness(struct task_struct *task, struct mm_struct *mm)
22162 +{
22163 +       struct vx_info *vxi = mm->mm_vx_info;
22164 +       unsigned long points;
22165 +       rlim_t v, w;
22166 +
22167 +       if (!vxi)
22168 +               return 0;
22169 +
22170 +       points = vxi->vx_badness_bias;
22171 +
22172 +       v = __vx_cres_array_fixup(&vxi->limit, VLA_RSS);
22173 +       w = __rlim_soft(&vxi->limit, RLIMIT_RSS);
22174 +       points += (v > w) ? (v - w) : 0;
22175 +
22176 +       return points;
22177 +}
22178 +
22179 diff -NurpP --minimal linux-3.0/kernel/vserver/limit_init.h linux-3.0-vs2.3.1-pre7/kernel/vserver/limit_init.h
22180 --- linux-3.0/kernel/vserver/limit_init.h       1970-01-01 01:00:00.000000000 +0100
22181 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/limit_init.h  2011-06-10 22:11:24.000000000 +0200
22182 @@ -0,0 +1,31 @@
22183 +
22184 +
22185 +static inline void vx_info_init_limit(struct _vx_limit *limit)
22186 +{
22187 +       int lim;
22188 +
22189 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22190 +               __rlim_soft(limit, lim) = RLIM_INFINITY;
22191 +               __rlim_hard(limit, lim) = RLIM_INFINITY;
22192 +               __rlim_set(limit, lim, 0);
22193 +               atomic_set(&__rlim_lhit(limit, lim), 0);
22194 +               __rlim_rmin(limit, lim) = 0;
22195 +               __rlim_rmax(limit, lim) = 0;
22196 +       }
22197 +}
22198 +
22199 +static inline void vx_info_exit_limit(struct _vx_limit *limit)
22200 +{
22201 +       rlim_t value;
22202 +       int lim;
22203 +
22204 +       for (lim = 0; lim < NUM_LIMITS; lim++) {
22205 +               if ((1 << lim) & VLIM_NOCHECK)
22206 +                       continue;
22207 +               value = __rlim_get(limit, lim);
22208 +               vxwprintk_xid(value,
22209 +                       "!!! limit: %p[%s,%d] = %ld on exit.",
22210 +                       limit, vlimit_name[lim], lim, (long)value);
22211 +       }
22212 +}
22213 +
22214 diff -NurpP --minimal linux-3.0/kernel/vserver/limit_proc.h linux-3.0-vs2.3.1-pre7/kernel/vserver/limit_proc.h
22215 --- linux-3.0/kernel/vserver/limit_proc.h       1970-01-01 01:00:00.000000000 +0100
22216 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/limit_proc.h  2011-06-10 22:11:24.000000000 +0200
22217 @@ -0,0 +1,57 @@
22218 +#ifndef _VX_LIMIT_PROC_H
22219 +#define _VX_LIMIT_PROC_H
22220 +
22221 +#include <linux/vserver/limit_int.h>
22222 +
22223 +
22224 +#define VX_LIMIT_FMT   ":\t%8ld\t%8ld/%8ld\t%8lld/%8lld\t%6d\n"
22225 +#define VX_LIMIT_TOP   \
22226 +       "Limit\t current\t     min/max\t\t    soft/hard\t\thits\n"
22227 +
22228 +#define VX_LIMIT_ARG(r)                                \
22229 +       (unsigned long)__rlim_get(limit, r),    \
22230 +       (unsigned long)__rlim_rmin(limit, r),   \
22231 +       (unsigned long)__rlim_rmax(limit, r),   \
22232 +       VX_VLIM(__rlim_soft(limit, r)),         \
22233 +       VX_VLIM(__rlim_hard(limit, r)),         \
22234 +       atomic_read(&__rlim_lhit(limit, r))
22235 +
22236 +static inline int vx_info_proc_limit(struct _vx_limit *limit, char *buffer)
22237 +{
22238 +       vx_limit_fixup(limit, -1);
22239 +       return sprintf(buffer, VX_LIMIT_TOP
22240 +               "PROC"  VX_LIMIT_FMT
22241 +               "VM"    VX_LIMIT_FMT
22242 +               "VML"   VX_LIMIT_FMT
22243 +               "RSS"   VX_LIMIT_FMT
22244 +               "ANON"  VX_LIMIT_FMT
22245 +               "RMAP"  VX_LIMIT_FMT
22246 +               "FILES" VX_LIMIT_FMT
22247 +               "OFD"   VX_LIMIT_FMT
22248 +               "LOCKS" VX_LIMIT_FMT
22249 +               "SOCK"  VX_LIMIT_FMT
22250 +               "MSGQ"  VX_LIMIT_FMT
22251 +               "SHM"   VX_LIMIT_FMT
22252 +               "SEMA"  VX_LIMIT_FMT
22253 +               "SEMS"  VX_LIMIT_FMT
22254 +               "DENT"  VX_LIMIT_FMT,
22255 +               VX_LIMIT_ARG(RLIMIT_NPROC),
22256 +               VX_LIMIT_ARG(RLIMIT_AS),
22257 +               VX_LIMIT_ARG(RLIMIT_MEMLOCK),
22258 +               VX_LIMIT_ARG(RLIMIT_RSS),
22259 +               VX_LIMIT_ARG(VLIMIT_ANON),
22260 +               VX_LIMIT_ARG(VLIMIT_MAPPED),
22261 +               VX_LIMIT_ARG(RLIMIT_NOFILE),
22262 +               VX_LIMIT_ARG(VLIMIT_OPENFD),
22263 +               VX_LIMIT_ARG(RLIMIT_LOCKS),
22264 +               VX_LIMIT_ARG(VLIMIT_NSOCK),
22265 +               VX_LIMIT_ARG(RLIMIT_MSGQUEUE),
22266 +               VX_LIMIT_ARG(VLIMIT_SHMEM),
22267 +               VX_LIMIT_ARG(VLIMIT_SEMARY),
22268 +               VX_LIMIT_ARG(VLIMIT_NSEMS),
22269 +               VX_LIMIT_ARG(VLIMIT_DENTRY));
22270 +}
22271 +
22272 +#endif /* _VX_LIMIT_PROC_H */
22273 +
22274 +
22275 diff -NurpP --minimal linux-3.0/kernel/vserver/network.c linux-3.0-vs2.3.1-pre7/kernel/vserver/network.c
22276 --- linux-3.0/kernel/vserver/network.c  1970-01-01 01:00:00.000000000 +0100
22277 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/network.c     2011-06-10 23:20:56.000000000 +0200
22278 @@ -0,0 +1,912 @@
22279 +/*
22280 + *  linux/kernel/vserver/network.c
22281 + *
22282 + *  Virtual Server: Network Support
22283 + *
22284 + *  Copyright (C) 2003-2007  Herbert Pötzl
22285 + *
22286 + *  V0.01  broken out from vcontext V0.05
22287 + *  V0.02  cleaned up implementation
22288 + *  V0.03  added equiv nx commands
22289 + *  V0.04  switch to RCU based hash
22290 + *  V0.05  and back to locking again
22291 + *  V0.06  changed vcmds to nxi arg
22292 + *  V0.07  have __create claim() the nxi
22293 + *
22294 + */
22295 +
22296 +#include <linux/err.h>
22297 +#include <linux/slab.h>
22298 +#include <linux/rcupdate.h>
22299 +
22300 +#include <linux/vs_network.h>
22301 +#include <linux/vs_pid.h>
22302 +#include <linux/vserver/network_cmd.h>
22303 +
22304 +
22305 +atomic_t nx_global_ctotal      = ATOMIC_INIT(0);
22306 +atomic_t nx_global_cactive     = ATOMIC_INIT(0);
22307 +
22308 +static struct kmem_cache *nx_addr_v4_cachep = NULL;
22309 +static struct kmem_cache *nx_addr_v6_cachep = NULL;
22310 +
22311 +
22312 +static int __init init_network(void)
22313 +{
22314 +       nx_addr_v4_cachep = kmem_cache_create("nx_v4_addr_cache",
22315 +               sizeof(struct nx_addr_v4), 0,
22316 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
22317 +       nx_addr_v6_cachep = kmem_cache_create("nx_v6_addr_cache",
22318 +               sizeof(struct nx_addr_v6), 0,
22319 +               SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
22320 +       return 0;
22321 +}
22322 +
22323 +
22324 +/*     __alloc_nx_addr_v4()                                    */
22325 +
22326 +static inline struct nx_addr_v4 *__alloc_nx_addr_v4(void)
22327 +{
22328 +       struct nx_addr_v4 *nxa = kmem_cache_alloc(
22329 +               nx_addr_v4_cachep, GFP_KERNEL);
22330 +
22331 +       if (!IS_ERR(nxa))
22332 +               memset(nxa, 0, sizeof(*nxa));
22333 +       return nxa;
22334 +}
22335 +
22336 +/*     __dealloc_nx_addr_v4()                                  */
22337 +
22338 +static inline void __dealloc_nx_addr_v4(struct nx_addr_v4 *nxa)
22339 +{
22340 +       kmem_cache_free(nx_addr_v4_cachep, nxa);
22341 +}
22342 +
22343 +/*     __dealloc_nx_addr_v4_all()                              */
22344 +
22345 +static inline void __dealloc_nx_addr_v4_all(struct nx_addr_v4 *nxa)
22346 +{
22347 +       while (nxa) {
22348 +               struct nx_addr_v4 *next = nxa->next;
22349 +
22350 +               __dealloc_nx_addr_v4(nxa);
22351 +               nxa = next;
22352 +       }
22353 +}
22354 +
22355 +
22356 +#ifdef CONFIG_IPV6
22357 +
22358 +/*     __alloc_nx_addr_v6()                                    */
22359 +
22360 +static inline struct nx_addr_v6 *__alloc_nx_addr_v6(void)
22361 +{
22362 +       struct nx_addr_v6 *nxa = kmem_cache_alloc(
22363 +               nx_addr_v6_cachep, GFP_KERNEL);
22364 +
22365 +       if (!IS_ERR(nxa))
22366 +               memset(nxa, 0, sizeof(*nxa));
22367 +       return nxa;
22368 +}
22369 +
22370 +/*     __dealloc_nx_addr_v6()                                  */
22371 +
22372 +static inline void __dealloc_nx_addr_v6(struct nx_addr_v6 *nxa)
22373 +{
22374 +       kmem_cache_free(nx_addr_v6_cachep, nxa);
22375 +}
22376 +
22377 +/*     __dealloc_nx_addr_v6_all()                              */
22378 +
22379 +static inline void __dealloc_nx_addr_v6_all(struct nx_addr_v6 *nxa)
22380 +{
22381 +       while (nxa) {
22382 +               struct nx_addr_v6 *next = nxa->next;
22383 +
22384 +               __dealloc_nx_addr_v6(nxa);
22385 +               nxa = next;
22386 +       }
22387 +}
22388 +
22389 +#endif /* CONFIG_IPV6 */
22390 +
22391 +/*     __alloc_nx_info()
22392 +
22393 +       * allocate an initialized nx_info struct
22394 +       * doesn't make it visible (hash)                        */
22395 +
22396 +static struct nx_info *__alloc_nx_info(nid_t nid)
22397 +{
22398 +       struct nx_info *new = NULL;
22399 +
22400 +       vxdprintk(VXD_CBIT(nid, 1), "alloc_nx_info(%d)*", nid);
22401 +
22402 +       /* would this benefit from a slab cache? */
22403 +       new = kmalloc(sizeof(struct nx_info), GFP_KERNEL);
22404 +       if (!new)
22405 +               return 0;
22406 +
22407 +       memset(new, 0, sizeof(struct nx_info));
22408 +       new->nx_id = nid;
22409 +       INIT_HLIST_NODE(&new->nx_hlist);
22410 +       atomic_set(&new->nx_usecnt, 0);
22411 +       atomic_set(&new->nx_tasks, 0);
22412 +       new->nx_state = 0;
22413 +
22414 +       new->nx_flags = NXF_INIT_SET;
22415 +
22416 +       /* rest of init goes here */
22417 +
22418 +       new->v4_lback.s_addr = htonl(INADDR_LOOPBACK);
22419 +       new->v4_bcast.s_addr = htonl(INADDR_BROADCAST);
22420 +
22421 +       vxdprintk(VXD_CBIT(nid, 0),
22422 +               "alloc_nx_info(%d) = %p", nid, new);
22423 +       atomic_inc(&nx_global_ctotal);
22424 +       return new;
22425 +}
22426 +
22427 +/*     __dealloc_nx_info()
22428 +
22429 +       * final disposal of nx_info                             */
22430 +
22431 +static void __dealloc_nx_info(struct nx_info *nxi)
22432 +{
22433 +       vxdprintk(VXD_CBIT(nid, 0),
22434 +               "dealloc_nx_info(%p)", nxi);
22435 +
22436 +       nxi->nx_hlist.next = LIST_POISON1;
22437 +       nxi->nx_id = -1;
22438 +
22439 +       BUG_ON(atomic_read(&nxi->nx_usecnt));
22440 +       BUG_ON(atomic_read(&nxi->nx_tasks));
22441 +
22442 +       __dealloc_nx_addr_v4_all(nxi->v4.next);
22443 +
22444 +       nxi->nx_state |= NXS_RELEASED;
22445 +       kfree(nxi);
22446 +       atomic_dec(&nx_global_ctotal);
22447 +}
22448 +
22449 +static void __shutdown_nx_info(struct nx_info *nxi)
22450 +{
22451 +       nxi->nx_state |= NXS_SHUTDOWN;
22452 +       vs_net_change(nxi, VSC_NETDOWN);
22453 +}
22454 +
22455 +/*     exported stuff                                          */
22456 +
22457 +void free_nx_info(struct nx_info *nxi)
22458 +{
22459 +       /* context shutdown is mandatory */
22460 +       BUG_ON(nxi->nx_state != NXS_SHUTDOWN);
22461 +
22462 +       /* context must not be hashed */
22463 +       BUG_ON(nxi->nx_state & NXS_HASHED);
22464 +
22465 +       BUG_ON(atomic_read(&nxi->nx_usecnt));
22466 +       BUG_ON(atomic_read(&nxi->nx_tasks));
22467 +
22468 +       __dealloc_nx_info(nxi);
22469 +}
22470 +
22471 +
22472 +void __nx_set_lback(struct nx_info *nxi)
22473 +{
22474 +       int nid = nxi->nx_id;
22475 +       __be32 lback = htonl(INADDR_LOOPBACK ^ ((nid & 0xFFFF) << 8));
22476 +
22477 +       nxi->v4_lback.s_addr = lback;
22478 +}
22479 +
22480 +extern int __nx_inet_add_lback(__be32 addr);
22481 +extern int __nx_inet_del_lback(__be32 addr);
22482 +
22483 +
22484 +/*     hash table for nx_info hash */
22485 +
22486 +#define NX_HASH_SIZE   13
22487 +
22488 +struct hlist_head nx_info_hash[NX_HASH_SIZE];
22489 +
22490 +static DEFINE_SPINLOCK(nx_info_hash_lock);
22491 +
22492 +
22493 +static inline unsigned int __hashval(nid_t nid)
22494 +{
22495 +       return (nid % NX_HASH_SIZE);
22496 +}
22497 +
22498 +
22499 +
22500 +/*     __hash_nx_info()
22501 +
22502 +       * add the nxi to the global hash table
22503 +       * requires the hash_lock to be held                     */
22504 +
22505 +static inline void __hash_nx_info(struct nx_info *nxi)
22506 +{
22507 +       struct hlist_head *head;
22508 +
22509 +       vxd_assert_lock(&nx_info_hash_lock);
22510 +       vxdprintk(VXD_CBIT(nid, 4),
22511 +               "__hash_nx_info: %p[#%d]", nxi, nxi->nx_id);
22512 +
22513 +       /* context must not be hashed */
22514 +       BUG_ON(nx_info_state(nxi, NXS_HASHED));
22515 +
22516 +       nxi->nx_state |= NXS_HASHED;
22517 +       head = &nx_info_hash[__hashval(nxi->nx_id)];
22518 +       hlist_add_head(&nxi->nx_hlist, head);
22519 +       atomic_inc(&nx_global_cactive);
22520 +}
22521 +
22522 +/*     __unhash_nx_info()
22523 +
22524 +       * remove the nxi from the global hash table
22525 +       * requires the hash_lock to be held                     */
22526 +
22527 +static inline void __unhash_nx_info(struct nx_info *nxi)
22528 +{
22529 +       vxd_assert_lock(&nx_info_hash_lock);
22530 +       vxdprintk(VXD_CBIT(nid, 4),
22531 +               "__unhash_nx_info: %p[#%d.%d.%d]", nxi, nxi->nx_id,
22532 +               atomic_read(&nxi->nx_usecnt), atomic_read(&nxi->nx_tasks));
22533 +
22534 +       /* context must be hashed */
22535 +       BUG_ON(!nx_info_state(nxi, NXS_HASHED));
22536 +       /* but without tasks */
22537 +       BUG_ON(atomic_read(&nxi->nx_tasks));
22538 +
22539 +       nxi->nx_state &= ~NXS_HASHED;
22540 +       hlist_del(&nxi->nx_hlist);
22541 +       atomic_dec(&nx_global_cactive);
22542 +}
22543 +
22544 +
22545 +/*     __lookup_nx_info()
22546 +
22547 +       * requires the hash_lock to be held
22548 +       * doesn't increment the nx_refcnt                       */
22549 +
22550 +static inline struct nx_info *__lookup_nx_info(nid_t nid)
22551 +{
22552 +       struct hlist_head *head = &nx_info_hash[__hashval(nid)];
22553 +       struct hlist_node *pos;
22554 +       struct nx_info *nxi;
22555 +
22556 +       vxd_assert_lock(&nx_info_hash_lock);
22557 +       hlist_for_each(pos, head) {
22558 +               nxi = hlist_entry(pos, struct nx_info, nx_hlist);
22559 +
22560 +               if (nxi->nx_id == nid)
22561 +                       goto found;
22562 +       }
22563 +       nxi = NULL;
22564 +found:
22565 +       vxdprintk(VXD_CBIT(nid, 0),
22566 +               "__lookup_nx_info(#%u): %p[#%u]",
22567 +               nid, nxi, nxi ? nxi->nx_id : 0);
22568 +       return nxi;
22569 +}
22570 +
22571 +
22572 +/*     __create_nx_info()
22573 +
22574 +       * create the requested context
22575 +       * get(), claim() and hash it                            */
22576 +
22577 +static struct nx_info *__create_nx_info(int id)
22578 +{
22579 +       struct nx_info *new, *nxi = NULL;
22580 +
22581 +       vxdprintk(VXD_CBIT(nid, 1), "create_nx_info(%d)*", id);
22582 +
22583 +       if (!(new = __alloc_nx_info(id)))
22584 +               return ERR_PTR(-ENOMEM);
22585 +
22586 +       /* required to make dynamic xids unique */
22587 +       spin_lock(&nx_info_hash_lock);
22588 +
22589 +       /* static context requested */
22590 +       if ((nxi = __lookup_nx_info(id))) {
22591 +               vxdprintk(VXD_CBIT(nid, 0),
22592 +                       "create_nx_info(%d) = %p (already there)", id, nxi);
22593 +               if (nx_info_flags(nxi, NXF_STATE_SETUP, 0))
22594 +                       nxi = ERR_PTR(-EBUSY);
22595 +               else
22596 +                       nxi = ERR_PTR(-EEXIST);
22597 +               goto out_unlock;
22598 +       }
22599 +       /* new context */
22600 +       vxdprintk(VXD_CBIT(nid, 0),
22601 +               "create_nx_info(%d) = %p (new)", id, new);
22602 +       claim_nx_info(new, NULL);
22603 +       __nx_set_lback(new);
22604 +       __hash_nx_info(get_nx_info(new));
22605 +       nxi = new, new = NULL;
22606 +
22607 +out_unlock:
22608 +       spin_unlock(&nx_info_hash_lock);
22609 +       if (new)
22610 +               __dealloc_nx_info(new);
22611 +       return nxi;
22612 +}
22613 +
22614 +
22615 +
22616 +/*     exported stuff                                          */
22617 +
22618 +
22619 +void unhash_nx_info(struct nx_info *nxi)
22620 +{
22621 +       __shutdown_nx_info(nxi);
22622 +       spin_lock(&nx_info_hash_lock);
22623 +       __unhash_nx_info(nxi);
22624 +       spin_unlock(&nx_info_hash_lock);
22625 +}
22626 +
22627 +/*     lookup_nx_info()
22628 +
22629 +       * search for a nx_info and get() it
22630 +       * negative id means current                             */
22631 +
22632 +struct nx_info *lookup_nx_info(int id)
22633 +{
22634 +       struct nx_info *nxi = NULL;
22635 +
22636 +       if (id < 0) {
22637 +               nxi = get_nx_info(current_nx_info());
22638 +       } else if (id > 1) {
22639 +               spin_lock(&nx_info_hash_lock);
22640 +               nxi = get_nx_info(__lookup_nx_info(id));
22641 +               spin_unlock(&nx_info_hash_lock);
22642 +       }
22643 +       return nxi;
22644 +}
22645 +
22646 +/*     nid_is_hashed()
22647 +
22648 +       * verify that nid is still hashed                       */
22649 +
22650 +int nid_is_hashed(nid_t nid)
22651 +{
22652 +       int hashed;
22653 +
22654 +       spin_lock(&nx_info_hash_lock);
22655 +       hashed = (__lookup_nx_info(nid) != NULL);
22656 +       spin_unlock(&nx_info_hash_lock);
22657 +       return hashed;
22658 +}
22659 +
22660 +
22661 +#ifdef CONFIG_PROC_FS
22662 +
22663 +/*     get_nid_list()
22664 +
22665 +       * get a subset of hashed nids for proc
22666 +       * assumes size is at least one                          */
22667 +
22668 +int get_nid_list(int index, unsigned int *nids, int size)
22669 +{
22670 +       int hindex, nr_nids = 0;
22671 +
22672 +       /* only show current and children */
22673 +       if (!nx_check(0, VS_ADMIN | VS_WATCH)) {
22674 +               if (index > 0)
22675 +                       return 0;
22676 +               nids[nr_nids] = nx_current_nid();
22677 +               return 1;
22678 +       }
22679 +
22680 +       for (hindex = 0; hindex < NX_HASH_SIZE; hindex++) {
22681 +               struct hlist_head *head = &nx_info_hash[hindex];
22682 +               struct hlist_node *pos;
22683 +
22684 +               spin_lock(&nx_info_hash_lock);
22685 +               hlist_for_each(pos, head) {
22686 +                       struct nx_info *nxi;
22687 +
22688 +                       if (--index > 0)
22689 +                               continue;
22690 +
22691 +                       nxi = hlist_entry(pos, struct nx_info, nx_hlist);
22692 +                       nids[nr_nids] = nxi->nx_id;
22693 +                       if (++nr_nids >= size) {
22694 +                               spin_unlock(&nx_info_hash_lock);
22695 +                               goto out;
22696 +                       }
22697 +               }
22698 +               /* keep the lock time short */
22699 +               spin_unlock(&nx_info_hash_lock);
22700 +       }
22701 +out:
22702 +       return nr_nids;
22703 +}
22704 +#endif
22705 +
22706 +
22707 +/*
22708 + *     migrate task to new network
22709 + *     gets nxi, puts old_nxi on change
22710 + */
22711 +
22712 +int nx_migrate_task(struct task_struct *p, struct nx_info *nxi)
22713 +{
22714 +       struct nx_info *old_nxi;
22715 +       int ret = 0;
22716 +
22717 +       if (!p || !nxi)
22718 +               BUG();
22719 +
22720 +       vxdprintk(VXD_CBIT(nid, 5),
22721 +               "nx_migrate_task(%p,%p[#%d.%d.%d])",
22722 +               p, nxi, nxi->nx_id,
22723 +               atomic_read(&nxi->nx_usecnt),
22724 +               atomic_read(&nxi->nx_tasks));
22725 +
22726 +       if (nx_info_flags(nxi, NXF_INFO_PRIVATE, 0) &&
22727 +               !nx_info_flags(nxi, NXF_STATE_SETUP, 0))
22728 +               return -EACCES;
22729 +
22730 +       if (nx_info_state(nxi, NXS_SHUTDOWN))
22731 +               return -EFAULT;
22732 +
22733 +       /* maybe disallow this completely? */
22734 +       old_nxi = task_get_nx_info(p);
22735 +       if (old_nxi == nxi)
22736 +               goto out;
22737 +
22738 +       task_lock(p);
22739 +       if (old_nxi)
22740 +               clr_nx_info(&p->nx_info);
22741 +       claim_nx_info(nxi, p);
22742 +       set_nx_info(&p->nx_info, nxi);
22743 +       p->nid = nxi->nx_id;
22744 +       task_unlock(p);
22745 +
22746 +       vxdprintk(VXD_CBIT(nid, 5),
22747 +               "moved task %p into nxi:%p[#%d]",
22748 +               p, nxi, nxi->nx_id);
22749 +
22750 +       if (old_nxi)
22751 +               release_nx_info(old_nxi, p);
22752 +       ret = 0;
22753 +out:
22754 +       put_nx_info(old_nxi);
22755 +       return ret;
22756 +}
22757 +
22758 +
22759 +void nx_set_persistent(struct nx_info *nxi)
22760 +{
22761 +       vxdprintk(VXD_CBIT(nid, 6),
22762 +               "nx_set_persistent(%p[#%d])", nxi, nxi->nx_id);
22763 +
22764 +       get_nx_info(nxi);
22765 +       claim_nx_info(nxi, NULL);
22766 +}
22767 +
22768 +void nx_clear_persistent(struct nx_info *nxi)
22769 +{
22770 +       vxdprintk(VXD_CBIT(nid, 6),
22771 +               "nx_clear_persistent(%p[#%d])", nxi, nxi->nx_id);
22772 +
22773 +       release_nx_info(nxi, NULL);
22774 +       put_nx_info(nxi);
22775 +}
22776 +
22777 +void nx_update_persistent(struct nx_info *nxi)
22778 +{
22779 +       if (nx_info_flags(nxi, NXF_PERSISTENT, 0))
22780 +               nx_set_persistent(nxi);
22781 +       else
22782 +               nx_clear_persistent(nxi);
22783 +}
22784 +
22785 +/* vserver syscall commands below here */
22786 +
22787 +/* taks nid and nx_info functions */
22788 +
22789 +#include <asm/uaccess.h>
22790 +
22791 +
22792 +int vc_task_nid(uint32_t id)
22793 +{
22794 +       nid_t nid;
22795 +
22796 +       if (id) {
22797 +               struct task_struct *tsk;
22798 +
22799 +               rcu_read_lock();
22800 +               tsk = find_task_by_real_pid(id);
22801 +               nid = (tsk) ? tsk->nid : -ESRCH;
22802 +               rcu_read_unlock();
22803 +       } else
22804 +               nid = nx_current_nid();
22805 +       return nid;
22806 +}
22807 +
22808 +
22809 +int vc_nx_info(struct nx_info *nxi, void __user *data)
22810 +{
22811 +       struct vcmd_nx_info_v0 vc_data;
22812 +
22813 +       vc_data.nid = nxi->nx_id;
22814 +
22815 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
22816 +               return -EFAULT;
22817 +       return 0;
22818 +}
22819 +
22820 +
22821 +/* network functions */
22822 +
22823 +int vc_net_create(uint32_t nid, void __user *data)
22824 +{
22825 +       struct vcmd_net_create vc_data = { .flagword = NXF_INIT_SET };
22826 +       struct nx_info *new_nxi;
22827 +       int ret;
22828 +
22829 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
22830 +               return -EFAULT;
22831 +
22832 +       if ((nid > MAX_S_CONTEXT) || (nid < 2))
22833 +               return -EINVAL;
22834 +
22835 +       new_nxi = __create_nx_info(nid);
22836 +       if (IS_ERR(new_nxi))
22837 +               return PTR_ERR(new_nxi);
22838 +
22839 +       /* initial flags */
22840 +       new_nxi->nx_flags = vc_data.flagword;
22841 +
22842 +       ret = -ENOEXEC;
22843 +       if (vs_net_change(new_nxi, VSC_NETUP))
22844 +               goto out;
22845 +
22846 +       ret = nx_migrate_task(current, new_nxi);
22847 +       if (ret)
22848 +               goto out;
22849 +
22850 +       /* return context id on success */
22851 +       ret = new_nxi->nx_id;
22852 +
22853 +       /* get a reference for persistent contexts */
22854 +       if ((vc_data.flagword & NXF_PERSISTENT))
22855 +               nx_set_persistent(new_nxi);
22856 +out:
22857 +       release_nx_info(new_nxi, NULL);
22858 +       put_nx_info(new_nxi);
22859 +       return ret;
22860 +}
22861 +
22862 +
22863 +int vc_net_migrate(struct nx_info *nxi, void __user *data)
22864 +{
22865 +       return nx_migrate_task(current, nxi);
22866 +}
22867 +
22868 +
22869 +
22870 +int do_add_v4_addr(struct nx_info *nxi, __be32 ip, __be32 ip2, __be32 mask,
22871 +       uint16_t type, uint16_t flags)
22872 +{
22873 +       struct nx_addr_v4 *nxa = &nxi->v4;
22874 +
22875 +       if (NX_IPV4(nxi)) {
22876 +               /* locate last entry */
22877 +               for (; nxa->next; nxa = nxa->next);
22878 +               nxa->next = __alloc_nx_addr_v4();
22879 +               nxa = nxa->next;
22880 +
22881 +               if (IS_ERR(nxa))
22882 +                       return PTR_ERR(nxa);
22883 +       }
22884 +
22885 +       if (nxi->v4.next)
22886 +               /* remove single ip for ip list */
22887 +               nxi->nx_flags &= ~NXF_SINGLE_IP;
22888 +
22889 +       nxa->ip[0].s_addr = ip;
22890 +       nxa->ip[1].s_addr = ip2;
22891 +       nxa->mask.s_addr = mask;
22892 +       nxa->type = type;
22893 +       nxa->flags = flags;
22894 +       return 0;
22895 +}
22896 +
22897 +int do_remove_v4_addr(struct nx_info *nxi, __be32 ip, __be32 ip2, __be32 mask,
22898 +       uint16_t type, uint16_t flags)
22899 +{
22900 +       struct nx_addr_v4 *nxa = &nxi->v4;
22901 +
22902 +       switch (type) {
22903 +/*     case NXA_TYPE_ADDR:
22904 +               break;          */
22905 +
22906 +       case NXA_TYPE_ANY:
22907 +               __dealloc_nx_addr_v4_all(xchg(&nxa->next, NULL));
22908 +               memset(nxa, 0, sizeof(*nxa));
22909 +               break;
22910 +
22911 +       default:
22912 +               return -EINVAL;
22913 +       }
22914 +       return 0;
22915 +}
22916 +
22917 +
22918 +int vc_net_add(struct nx_info *nxi, void __user *data)
22919 +{
22920 +       struct vcmd_net_addr_v0 vc_data;
22921 +       int index, ret = 0;
22922 +
22923 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
22924 +               return -EFAULT;
22925 +
22926 +       switch (vc_data.type) {
22927 +       case NXA_TYPE_IPV4:
22928 +               if ((vc_data.count < 1) || (vc_data.count > 4))
22929 +                       return -EINVAL;
22930 +
22931 +               index = 0;
22932 +               while (index < vc_data.count) {
22933 +                       ret = do_add_v4_addr(nxi, vc_data.ip[index].s_addr, 0,
22934 +                               vc_data.mask[index].s_addr, NXA_TYPE_ADDR, 0);
22935 +                       if (ret)
22936 +                               return ret;
22937 +                       index++;
22938 +               }
22939 +               ret = index;
22940 +               break;
22941 +
22942 +       case NXA_TYPE_IPV4|NXA_MOD_BCAST:
22943 +               nxi->v4_bcast = vc_data.ip[0];
22944 +               ret = 1;
22945 +               break;
22946 +
22947 +       case NXA_TYPE_IPV4|NXA_MOD_LBACK:
22948 +               nxi->v4_lback = vc_data.ip[0];
22949 +               ret = 1;
22950 +               break;
22951 +
22952 +       default:
22953 +               ret = -EINVAL;
22954 +               break;
22955 +       }
22956 +       return ret;
22957 +}
22958 +
22959 +int vc_net_remove(struct nx_info *nxi, void __user *data)
22960 +{
22961 +       struct vcmd_net_addr_v0 vc_data;
22962 +
22963 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
22964 +               return -EFAULT;
22965 +
22966 +       switch (vc_data.type) {
22967 +       case NXA_TYPE_ANY:
22968 +               __dealloc_nx_addr_v4_all(xchg(&nxi->v4.next, NULL));
22969 +               memset(&nxi->v4, 0, sizeof(nxi->v4));
22970 +               break;
22971 +
22972 +       default:
22973 +               return -EINVAL;
22974 +       }
22975 +       return 0;
22976 +}
22977 +
22978 +
22979 +int vc_net_add_ipv4_v1(struct nx_info *nxi, void __user *data)
22980 +{
22981 +       struct vcmd_net_addr_ipv4_v1 vc_data;
22982 +
22983 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
22984 +               return -EFAULT;
22985 +
22986 +       switch (vc_data.type) {
22987 +       case NXA_TYPE_ADDR:
22988 +       case NXA_TYPE_MASK:
22989 +               return do_add_v4_addr(nxi, vc_data.ip.s_addr, 0,
22990 +                       vc_data.mask.s_addr, vc_data.type, vc_data.flags);
22991 +
22992 +       case NXA_TYPE_ADDR | NXA_MOD_BCAST:
22993 +               nxi->v4_bcast = vc_data.ip;
22994 +               break;
22995 +
22996 +       case NXA_TYPE_ADDR | NXA_MOD_LBACK:
22997 +               nxi->v4_lback = vc_data.ip;
22998 +               break;
22999 +
23000 +       default:
23001 +               return -EINVAL;
23002 +       }
23003 +       return 0;
23004 +}
23005 +
23006 +int vc_net_add_ipv4(struct nx_info *nxi, void __user *data)
23007 +{
23008 +       struct vcmd_net_addr_ipv4_v2 vc_data;
23009 +
23010 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23011 +               return -EFAULT;
23012 +
23013 +       switch (vc_data.type) {
23014 +       case NXA_TYPE_ADDR:
23015 +       case NXA_TYPE_MASK:
23016 +       case NXA_TYPE_RANGE:
23017 +               return do_add_v4_addr(nxi, vc_data.ip.s_addr, vc_data.ip2.s_addr,
23018 +                       vc_data.mask.s_addr, vc_data.type, vc_data.flags);
23019 +
23020 +       case NXA_TYPE_ADDR | NXA_MOD_BCAST:
23021 +               nxi->v4_bcast = vc_data.ip;
23022 +               break;
23023 +
23024 +       case NXA_TYPE_ADDR | NXA_MOD_LBACK:
23025 +               nxi->v4_lback = vc_data.ip;
23026 +               break;
23027 +
23028 +       default:
23029 +               return -EINVAL;
23030 +       }
23031 +       return 0;
23032 +}
23033 +
23034 +int vc_net_rem_ipv4_v1(struct nx_info *nxi, void __user *data)
23035 +{
23036 +       struct vcmd_net_addr_ipv4_v1 vc_data;
23037 +
23038 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23039 +               return -EFAULT;
23040 +
23041 +       return do_remove_v4_addr(nxi, vc_data.ip.s_addr, 0,
23042 +               vc_data.mask.s_addr, vc_data.type, vc_data.flags);
23043 +}
23044 +
23045 +int vc_net_rem_ipv4(struct nx_info *nxi, void __user *data)
23046 +{
23047 +       struct vcmd_net_addr_ipv4_v2 vc_data;
23048 +
23049 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23050 +               return -EFAULT;
23051 +
23052 +       return do_remove_v4_addr(nxi, vc_data.ip.s_addr, vc_data.ip2.s_addr,
23053 +               vc_data.mask.s_addr, vc_data.type, vc_data.flags);
23054 +}
23055 +
23056 +#ifdef CONFIG_IPV6
23057 +
23058 +int do_add_v6_addr(struct nx_info *nxi,
23059 +       struct in6_addr *ip, struct in6_addr *mask,
23060 +       uint32_t prefix, uint16_t type, uint16_t flags)
23061 +{
23062 +       struct nx_addr_v6 *nxa = &nxi->v6;
23063 +
23064 +       if (NX_IPV6(nxi)) {
23065 +               /* locate last entry */
23066 +               for (; nxa->next; nxa = nxa->next);
23067 +               nxa->next = __alloc_nx_addr_v6();
23068 +               nxa = nxa->next;
23069 +
23070 +               if (IS_ERR(nxa))
23071 +                       return PTR_ERR(nxa);
23072 +       }
23073 +
23074 +       nxa->ip = *ip;
23075 +       nxa->mask = *mask;
23076 +       nxa->prefix = prefix;
23077 +       nxa->type = type;
23078 +       nxa->flags = flags;
23079 +       return 0;
23080 +}
23081 +
23082 +
23083 +int vc_net_add_ipv6(struct nx_info *nxi, void __user *data)
23084 +{
23085 +       struct vcmd_net_addr_ipv6_v1 vc_data;
23086 +
23087 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23088 +               return -EFAULT;
23089 +
23090 +       switch (vc_data.type) {
23091 +       case NXA_TYPE_ADDR:
23092 +               memset(&vc_data.mask, ~0, sizeof(vc_data.mask));
23093 +               /* fallthrough */
23094 +       case NXA_TYPE_MASK:
23095 +               return do_add_v6_addr(nxi, &vc_data.ip, &vc_data.mask,
23096 +                       vc_data.prefix, vc_data.type, vc_data.flags);
23097 +       default:
23098 +               return -EINVAL;
23099 +       }
23100 +       return 0;
23101 +}
23102 +
23103 +int vc_net_remove_ipv6(struct nx_info *nxi, void __user *data)
23104 +{
23105 +       struct vcmd_net_addr_ipv6_v1 vc_data;
23106 +
23107 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
23108 +               return -EFAULT;
23109 +
23110 +       switch (vc_data.type) {
23111 +       case NXA_TYPE_ANY:
23112 +               __dealloc_nx_addr_v6_all(xchg(&nxi->v6.next, NULL));
23113 +               memset(&nxi->v6, 0, sizeof(nxi->v6));
23114 +               break;
23115 +
23116 +       default:
23117 +               return -EINVAL;
23118 +       }
23119 +       return 0;
23120 +}
23121 +
23122 +#endif /* CONFIG_IPV6 */
23123 +
23124 +
23125 +int vc_get_nflags(struct nx_info *nxi, void __user *data)
23126 +{
23127 +       struct vcmd_net_flags_v0 vc_data;
23128 +
23129 +       vc_data.flagword = nxi->nx_flags;
23130 +
23131 +       /* special STATE flag handling */
23132 +       vc_data.mask = vs_mask_flags(~0ULL, nxi->nx_flags, NXF_ONE_TIME);
23133 +
23134 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23135 +               return -EFAULT;
23136 +       return 0;
23137 +}
23138 +
23139 +int vc_set_nflags(struct nx_info *nxi, void __user *data)
23140 +{
23141 +       struct vcmd_net_flags_v0 vc_data;
23142 +       uint64_t mask, trigger;
23143 +
23144 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23145 +               return -EFAULT;
23146 +
23147 +       /* special STATE flag handling */
23148 +       mask = vs_mask_mask(vc_data.mask, nxi->nx_flags, NXF_ONE_TIME);
23149 +       trigger = (mask & nxi->nx_flags) ^ (mask & vc_data.flagword);
23150 +
23151 +       nxi->nx_flags = vs_mask_flags(nxi->nx_flags,
23152 +               vc_data.flagword, mask);
23153 +       if (trigger & NXF_PERSISTENT)
23154 +               nx_update_persistent(nxi);
23155 +
23156 +       return 0;
23157 +}
23158 +
23159 +int vc_get_ncaps(struct nx_info *nxi, void __user *data)
23160 +{
23161 +       struct vcmd_net_caps_v0 vc_data;
23162 +
23163 +       vc_data.ncaps = nxi->nx_ncaps;
23164 +       vc_data.cmask = ~0ULL;
23165 +
23166 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
23167 +               return -EFAULT;
23168 +       return 0;
23169 +}
23170 +
23171 +int vc_set_ncaps(struct nx_info *nxi, void __user *data)
23172 +{
23173 +       struct vcmd_net_caps_v0 vc_data;
23174 +
23175 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
23176 +               return -EFAULT;
23177 +
23178 +       nxi->nx_ncaps = vs_mask_flags(nxi->nx_ncaps,
23179 +               vc_data.ncaps, vc_data.cmask);
23180 +       return 0;
23181 +}
23182 +
23183 +
23184 +#include <linux/module.h>
23185 +
23186 +module_init(init_network);
23187 +
23188 +EXPORT_SYMBOL_GPL(free_nx_info);
23189 +EXPORT_SYMBOL_GPL(unhash_nx_info);
23190 +
23191 diff -NurpP --minimal linux-3.0/kernel/vserver/proc.c linux-3.0-vs2.3.1-pre7/kernel/vserver/proc.c
23192 --- linux-3.0/kernel/vserver/proc.c     1970-01-01 01:00:00.000000000 +0100
23193 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/proc.c        2011-06-10 22:11:24.000000000 +0200
23194 @@ -0,0 +1,1100 @@
23195 +/*
23196 + *  linux/kernel/vserver/proc.c
23197 + *
23198 + *  Virtual Context Support
23199 + *
23200 + *  Copyright (C) 2003-2007  Herbert Pötzl
23201 + *
23202 + *  V0.01  basic structure
23203 + *  V0.02  adaptation vs1.3.0
23204 + *  V0.03  proc permissions
23205 + *  V0.04  locking/generic
23206 + *  V0.05  next generation procfs
23207 + *  V0.06  inode validation
23208 + *  V0.07  generic rewrite vid
23209 + *  V0.08  remove inode type
23210 + *
23211 + */
23212 +
23213 +#include <linux/proc_fs.h>
23214 +#include <linux/fs_struct.h>
23215 +#include <linux/mount.h>
23216 +#include <asm/unistd.h>
23217 +
23218 +#include <linux/vs_context.h>
23219 +#include <linux/vs_network.h>
23220 +#include <linux/vs_cvirt.h>
23221 +
23222 +#include <linux/in.h>
23223 +#include <linux/inetdevice.h>
23224 +#include <linux/vs_inet.h>
23225 +#include <linux/vs_inet6.h>
23226 +
23227 +#include <linux/vserver/global.h>
23228 +
23229 +#include "cvirt_proc.h"
23230 +#include "cacct_proc.h"
23231 +#include "limit_proc.h"
23232 +#include "sched_proc.h"
23233 +#include "vci_config.h"
23234 +
23235 +
23236 +static inline char *print_cap_t(char *buffer, kernel_cap_t *c)
23237 +{
23238 +       unsigned __capi;
23239 +
23240 +       CAP_FOR_EACH_U32(__capi) {
23241 +               buffer += sprintf(buffer, "%08x",
23242 +                       c->cap[(_KERNEL_CAPABILITY_U32S-1) - __capi]);
23243 +       }
23244 +       return buffer;
23245 +}
23246 +
23247 +
23248 +static struct proc_dir_entry *proc_virtual;
23249 +
23250 +static struct proc_dir_entry *proc_virtnet;
23251 +
23252 +
23253 +/* first the actual feeds */
23254 +
23255 +
23256 +static int proc_vci(char *buffer)
23257 +{
23258 +       return sprintf(buffer,
23259 +               "VCIVersion:\t%04x:%04x\n"
23260 +               "VCISyscall:\t%d\n"
23261 +               "VCIKernel:\t%08x\n",
23262 +               VCI_VERSION >> 16,
23263 +               VCI_VERSION & 0xFFFF,
23264 +               __NR_vserver,
23265 +               vci_kernel_config());
23266 +}
23267 +
23268 +static int proc_virtual_info(char *buffer)
23269 +{
23270 +       return proc_vci(buffer);
23271 +}
23272 +
23273 +static int proc_virtual_status(char *buffer)
23274 +{
23275 +       return sprintf(buffer,
23276 +               "#CTotal:\t%d\n"
23277 +               "#CActive:\t%d\n"
23278 +               "#NSProxy:\t%d\t%d %d %d %d %d %d\n"
23279 +               "#InitTask:\t%d\t%d %d\n",
23280 +               atomic_read(&vx_global_ctotal),
23281 +               atomic_read(&vx_global_cactive),
23282 +               atomic_read(&vs_global_nsproxy),
23283 +               atomic_read(&vs_global_fs),
23284 +               atomic_read(&vs_global_mnt_ns),
23285 +               atomic_read(&vs_global_uts_ns),
23286 +               atomic_read(&nr_ipc_ns),
23287 +               atomic_read(&vs_global_user_ns),
23288 +               atomic_read(&vs_global_pid_ns),
23289 +               atomic_read(&init_task.usage),
23290 +               atomic_read(&init_task.nsproxy->count),
23291 +               init_task.fs->users);
23292 +}
23293 +
23294 +
23295 +int proc_vxi_info(struct vx_info *vxi, char *buffer)
23296 +{
23297 +       int length;
23298 +
23299 +       length = sprintf(buffer,
23300 +               "ID:\t%d\n"
23301 +               "Info:\t%p\n"
23302 +               "Init:\t%d\n"
23303 +               "OOM:\t%lld\n",
23304 +               vxi->vx_id,
23305 +               vxi,
23306 +               vxi->vx_initpid,
23307 +               vxi->vx_badness_bias);
23308 +       return length;
23309 +}
23310 +
23311 +int proc_vxi_status(struct vx_info *vxi, char *buffer)
23312 +{
23313 +       char *orig = buffer;
23314 +
23315 +       buffer += sprintf(buffer,
23316 +               "UseCnt:\t%d\n"
23317 +               "Tasks:\t%d\n"
23318 +               "Flags:\t%016llx\n",
23319 +               atomic_read(&vxi->vx_usecnt),
23320 +               atomic_read(&vxi->vx_tasks),
23321 +               (unsigned long long)vxi->vx_flags);
23322 +
23323 +       buffer += sprintf(buffer, "BCaps:\t");
23324 +       buffer = print_cap_t(buffer, &vxi->vx_bcaps);
23325 +       buffer += sprintf(buffer, "\n");
23326 +
23327 +       buffer += sprintf(buffer,
23328 +               "CCaps:\t%016llx\n"
23329 +               "Umask:\t%16llx\n"
23330 +               "Spaces:\t%08lx %08lx\n",
23331 +               (unsigned long long)vxi->vx_ccaps,
23332 +               (unsigned long long)vxi->vx_umask,
23333 +               vxi->space[0].vx_nsmask, vxi->space[1].vx_nsmask);
23334 +       return buffer - orig;
23335 +}
23336 +
23337 +int proc_vxi_limit(struct vx_info *vxi, char *buffer)
23338 +{
23339 +       return vx_info_proc_limit(&vxi->limit, buffer);
23340 +}
23341 +
23342 +int proc_vxi_sched(struct vx_info *vxi, char *buffer)
23343 +{
23344 +       int cpu, length;
23345 +
23346 +       length = vx_info_proc_sched(&vxi->sched, buffer);
23347 +       for_each_online_cpu(cpu) {
23348 +               length += vx_info_proc_sched_pc(
23349 +                       &vx_per_cpu(vxi, sched_pc, cpu),
23350 +                       buffer + length, cpu);
23351 +       }
23352 +       return length;
23353 +}
23354 +
23355 +int proc_vxi_nsproxy0(struct vx_info *vxi, char *buffer)
23356 +{
23357 +       return vx_info_proc_nsproxy(vxi->space[0].vx_nsproxy, buffer);
23358 +}
23359 +
23360 +int proc_vxi_nsproxy1(struct vx_info *vxi, char *buffer)
23361 +{
23362 +       return vx_info_proc_nsproxy(vxi->space[1].vx_nsproxy, buffer);
23363 +}
23364 +
23365 +int proc_vxi_cvirt(struct vx_info *vxi, char *buffer)
23366 +{
23367 +       int cpu, length;
23368 +
23369 +       vx_update_load(vxi);
23370 +       length = vx_info_proc_cvirt(&vxi->cvirt, buffer);
23371 +       for_each_online_cpu(cpu) {
23372 +               length += vx_info_proc_cvirt_pc(
23373 +                       &vx_per_cpu(vxi, cvirt_pc, cpu),
23374 +                       buffer + length, cpu);
23375 +       }
23376 +       return length;
23377 +}
23378 +
23379 +int proc_vxi_cacct(struct vx_info *vxi, char *buffer)
23380 +{
23381 +       return vx_info_proc_cacct(&vxi->cacct, buffer);
23382 +}
23383 +
23384 +
23385 +static int proc_virtnet_info(char *buffer)
23386 +{
23387 +       return proc_vci(buffer);
23388 +}
23389 +
23390 +static int proc_virtnet_status(char *buffer)
23391 +{
23392 +       return sprintf(buffer,
23393 +               "#CTotal:\t%d\n"
23394 +               "#CActive:\t%d\n",
23395 +               atomic_read(&nx_global_ctotal),
23396 +               atomic_read(&nx_global_cactive));
23397 +}
23398 +
23399 +int proc_nxi_info(struct nx_info *nxi, char *buffer)
23400 +{
23401 +       struct nx_addr_v4 *v4a;
23402 +#ifdef CONFIG_IPV6
23403 +       struct nx_addr_v6 *v6a;
23404 +#endif
23405 +       int length, i;
23406 +
23407 +       length = sprintf(buffer,
23408 +               "ID:\t%d\n"
23409 +               "Info:\t%p\n"
23410 +               "Bcast:\t" NIPQUAD_FMT "\n"
23411 +               "Lback:\t" NIPQUAD_FMT "\n",
23412 +               nxi->nx_id,
23413 +               nxi,
23414 +               NIPQUAD(nxi->v4_bcast.s_addr),
23415 +               NIPQUAD(nxi->v4_lback.s_addr));
23416 +
23417 +       if (!NX_IPV4(nxi))
23418 +               goto skip_v4;
23419 +       for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
23420 +               length += sprintf(buffer + length, "%d:\t" NXAV4_FMT "\n",
23421 +                       i, NXAV4(v4a));
23422 +skip_v4:
23423 +#ifdef CONFIG_IPV6
23424 +       if (!NX_IPV6(nxi))
23425 +               goto skip_v6;
23426 +       for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
23427 +               length += sprintf(buffer + length, "%d:\t" NXAV6_FMT "\n",
23428 +                       i, NXAV6(v6a));
23429 +skip_v6:
23430 +#endif
23431 +       return length;
23432 +}
23433 +
23434 +int proc_nxi_status(struct nx_info *nxi, char *buffer)
23435 +{
23436 +       int length;
23437 +
23438 +       length = sprintf(buffer,
23439 +               "UseCnt:\t%d\n"
23440 +               "Tasks:\t%d\n"
23441 +               "Flags:\t%016llx\n"
23442 +               "NCaps:\t%016llx\n",
23443 +               atomic_read(&nxi->nx_usecnt),
23444 +               atomic_read(&nxi->nx_tasks),
23445 +               (unsigned long long)nxi->nx_flags,
23446 +               (unsigned long long)nxi->nx_ncaps);
23447 +       return length;
23448 +}
23449 +
23450 +
23451 +
23452 +/* here the inode helpers */
23453 +
23454 +struct vs_entry {
23455 +       int len;
23456 +       char *name;
23457 +       mode_t mode;
23458 +       struct inode_operations *iop;
23459 +       struct file_operations *fop;
23460 +       union proc_op op;
23461 +};
23462 +
23463 +static struct inode *vs_proc_make_inode(struct super_block *sb, struct vs_entry *p)
23464 +{
23465 +       struct inode *inode = new_inode(sb);
23466 +
23467 +       if (!inode)
23468 +               goto out;
23469 +
23470 +       inode->i_mode = p->mode;
23471 +       if (p->iop)
23472 +               inode->i_op = p->iop;
23473 +       if (p->fop)
23474 +               inode->i_fop = p->fop;
23475 +
23476 +       inode->i_nlink = (p->mode & S_IFDIR) ? 2 : 1;
23477 +       inode->i_flags |= S_IMMUTABLE;
23478 +
23479 +       inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
23480 +
23481 +       inode->i_uid = 0;
23482 +       inode->i_gid = 0;
23483 +       inode->i_tag = 0;
23484 +out:
23485 +       return inode;
23486 +}
23487 +
23488 +static struct dentry *vs_proc_instantiate(struct inode *dir,
23489 +       struct dentry *dentry, int id, void *ptr)
23490 +{
23491 +       struct vs_entry *p = ptr;
23492 +       struct inode *inode = vs_proc_make_inode(dir->i_sb, p);
23493 +       struct dentry *error = ERR_PTR(-EINVAL);
23494 +
23495 +       if (!inode)
23496 +               goto out;
23497 +
23498 +       PROC_I(inode)->op = p->op;
23499 +       PROC_I(inode)->fd = id;
23500 +       d_add(dentry, inode);
23501 +       error = NULL;
23502 +out:
23503 +       return error;
23504 +}
23505 +
23506 +/* Lookups */
23507 +
23508 +typedef struct dentry *instantiate_t(struct inode *, struct dentry *, int, void *);
23509 +
23510 +/*
23511 + * Fill a directory entry.
23512 + *
23513 + * If possible create the dcache entry and derive our inode number and
23514 + * file type from dcache entry.
23515 + *
23516 + * Since all of the proc inode numbers are dynamically generated, the inode
23517 + * numbers do not exist until the inode is cache.  This means creating the
23518 + * the dcache entry in readdir is necessary to keep the inode numbers
23519 + * reported by readdir in sync with the inode numbers reported
23520 + * by stat.
23521 + */
23522 +static int proc_fill_cache(struct file *filp, void *dirent, filldir_t filldir,
23523 +       char *name, int len, instantiate_t instantiate, int id, void *ptr)
23524 +{
23525 +       struct dentry *child, *dir = filp->f_dentry;
23526 +       struct inode *inode;
23527 +       struct qstr qname;
23528 +       ino_t ino = 0;
23529 +       unsigned type = DT_UNKNOWN;
23530 +
23531 +       qname.name = name;
23532 +       qname.len  = len;
23533 +       qname.hash = full_name_hash(name, len);
23534 +
23535 +       child = d_lookup(dir, &qname);
23536 +       if (!child) {
23537 +               struct dentry *new;
23538 +               new = d_alloc(dir, &qname);
23539 +               if (new) {
23540 +                       child = instantiate(dir->d_inode, new, id, ptr);
23541 +                       if (child)
23542 +                               dput(new);
23543 +                       else
23544 +                               child = new;
23545 +               }
23546 +       }
23547 +       if (!child || IS_ERR(child) || !child->d_inode)
23548 +               goto end_instantiate;
23549 +       inode = child->d_inode;
23550 +       if (inode) {
23551 +               ino = inode->i_ino;
23552 +               type = inode->i_mode >> 12;
23553 +       }
23554 +       dput(child);
23555 +end_instantiate:
23556 +       if (!ino)
23557 +               ino = find_inode_number(dir, &qname);
23558 +       if (!ino)
23559 +               ino = 1;
23560 +       return filldir(dirent, name, len, filp->f_pos, ino, type);
23561 +}
23562 +
23563 +
23564 +
23565 +/* get and revalidate vx_info/xid */
23566 +
23567 +static inline
23568 +struct vx_info *get_proc_vx_info(struct inode *inode)
23569 +{
23570 +       return lookup_vx_info(PROC_I(inode)->fd);
23571 +}
23572 +
23573 +static int proc_xid_revalidate(struct dentry *dentry, struct nameidata *nd)
23574 +{
23575 +       struct inode *inode = dentry->d_inode;
23576 +       xid_t xid = PROC_I(inode)->fd;
23577 +
23578 +       if (!xid || xid_is_hashed(xid))
23579 +               return 1;
23580 +       d_drop(dentry);
23581 +       return 0;
23582 +}
23583 +
23584 +
23585 +/* get and revalidate nx_info/nid */
23586 +
23587 +static int proc_nid_revalidate(struct dentry *dentry, struct nameidata *nd)
23588 +{
23589 +       struct inode *inode = dentry->d_inode;
23590 +       nid_t nid = PROC_I(inode)->fd;
23591 +
23592 +       if (!nid || nid_is_hashed(nid))
23593 +               return 1;
23594 +       d_drop(dentry);
23595 +       return 0;
23596 +}
23597 +
23598 +
23599 +
23600 +#define PROC_BLOCK_SIZE (PAGE_SIZE - 1024)
23601 +
23602 +static ssize_t proc_vs_info_read(struct file *file, char __user *buf,
23603 +                         size_t count, loff_t *ppos)
23604 +{
23605 +       struct inode *inode = file->f_dentry->d_inode;
23606 +       unsigned long page;
23607 +       ssize_t length = 0;
23608 +
23609 +       if (count > PROC_BLOCK_SIZE)
23610 +               count = PROC_BLOCK_SIZE;
23611 +
23612 +       /* fade that out as soon as stable */
23613 +       WARN_ON(PROC_I(inode)->fd);
23614 +
23615 +       if (!(page = __get_free_page(GFP_KERNEL)))
23616 +               return -ENOMEM;
23617 +
23618 +       BUG_ON(!PROC_I(inode)->op.proc_vs_read);
23619 +       length = PROC_I(inode)->op.proc_vs_read((char *)page);
23620 +
23621 +       if (length >= 0)
23622 +               length = simple_read_from_buffer(buf, count, ppos,
23623 +                       (char *)page, length);
23624 +
23625 +       free_page(page);
23626 +       return length;
23627 +}
23628 +
23629 +static ssize_t proc_vx_info_read(struct file *file, char __user *buf,
23630 +                         size_t count, loff_t *ppos)
23631 +{
23632 +       struct inode *inode = file->f_dentry->d_inode;
23633 +       struct vx_info *vxi = NULL;
23634 +       xid_t xid = PROC_I(inode)->fd;
23635 +       unsigned long page;
23636 +       ssize_t length = 0;
23637 +
23638 +       if (count > PROC_BLOCK_SIZE)
23639 +               count = PROC_BLOCK_SIZE;
23640 +
23641 +       /* fade that out as soon as stable */
23642 +       WARN_ON(!xid);
23643 +       vxi = lookup_vx_info(xid);
23644 +       if (!vxi)
23645 +               goto out;
23646 +
23647 +       length = -ENOMEM;
23648 +       if (!(page = __get_free_page(GFP_KERNEL)))
23649 +               goto out_put;
23650 +
23651 +       BUG_ON(!PROC_I(inode)->op.proc_vxi_read);
23652 +       length = PROC_I(inode)->op.proc_vxi_read(vxi, (char *)page);
23653 +
23654 +       if (length >= 0)
23655 +               length = simple_read_from_buffer(buf, count, ppos,
23656 +                       (char *)page, length);
23657 +
23658 +       free_page(page);
23659 +out_put:
23660 +       put_vx_info(vxi);
23661 +out:
23662 +       return length;
23663 +}
23664 +
23665 +static ssize_t proc_nx_info_read(struct file *file, char __user *buf,
23666 +                         size_t count, loff_t *ppos)
23667 +{
23668 +       struct inode *inode = file->f_dentry->d_inode;
23669 +       struct nx_info *nxi = NULL;
23670 +       nid_t nid = PROC_I(inode)->fd;
23671 +       unsigned long page;
23672 +       ssize_t length = 0;
23673 +
23674 +       if (count > PROC_BLOCK_SIZE)
23675 +               count = PROC_BLOCK_SIZE;
23676 +
23677 +       /* fade that out as soon as stable */
23678 +       WARN_ON(!nid);
23679 +       nxi = lookup_nx_info(nid);
23680 +       if (!nxi)
23681 +               goto out;
23682 +
23683 +       length = -ENOMEM;
23684 +       if (!(page = __get_free_page(GFP_KERNEL)))
23685 +               goto out_put;
23686 +
23687 +       BUG_ON(!PROC_I(inode)->op.proc_nxi_read);
23688 +       length = PROC_I(inode)->op.proc_nxi_read(nxi, (char *)page);
23689 +
23690 +       if (length >= 0)
23691 +               length = simple_read_from_buffer(buf, count, ppos,
23692 +                       (char *)page, length);
23693 +
23694 +       free_page(page);
23695 +out_put:
23696 +       put_nx_info(nxi);
23697 +out:
23698 +       return length;
23699 +}
23700 +
23701 +
23702 +
23703 +/* here comes the lower level */
23704 +
23705 +
23706 +#define NOD(NAME, MODE, IOP, FOP, OP) {        \
23707 +       .len  = sizeof(NAME) - 1,       \
23708 +       .name = (NAME),                 \
23709 +       .mode = MODE,                   \
23710 +       .iop  = IOP,                    \
23711 +       .fop  = FOP,                    \
23712 +       .op   = OP,                     \
23713 +}
23714 +
23715 +
23716 +#define DIR(NAME, MODE, OTYPE)                         \
23717 +       NOD(NAME, (S_IFDIR | (MODE)),                   \
23718 +               &proc_ ## OTYPE ## _inode_operations,   \
23719 +               &proc_ ## OTYPE ## _file_operations, { } )
23720 +
23721 +#define INF(NAME, MODE, OTYPE)                         \
23722 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
23723 +               &proc_vs_info_file_operations,          \
23724 +               { .proc_vs_read = &proc_##OTYPE } )
23725 +
23726 +#define VINF(NAME, MODE, OTYPE)                                \
23727 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
23728 +               &proc_vx_info_file_operations,          \
23729 +               { .proc_vxi_read = &proc_##OTYPE } )
23730 +
23731 +#define NINF(NAME, MODE, OTYPE)                                \
23732 +       NOD(NAME, (S_IFREG | (MODE)), NULL,             \
23733 +               &proc_nx_info_file_operations,          \
23734 +               { .proc_nxi_read = &proc_##OTYPE } )
23735 +
23736 +
23737 +static struct file_operations proc_vs_info_file_operations = {
23738 +       .read =         proc_vs_info_read,
23739 +};
23740 +
23741 +static struct file_operations proc_vx_info_file_operations = {
23742 +       .read =         proc_vx_info_read,
23743 +};
23744 +
23745 +static struct dentry_operations proc_xid_dentry_operations = {
23746 +       .d_revalidate = proc_xid_revalidate,
23747 +};
23748 +
23749 +static struct vs_entry vx_base_stuff[] = {
23750 +       VINF("info",    S_IRUGO, vxi_info),
23751 +       VINF("status",  S_IRUGO, vxi_status),
23752 +       VINF("limit",   S_IRUGO, vxi_limit),
23753 +       VINF("sched",   S_IRUGO, vxi_sched),
23754 +       VINF("nsproxy", S_IRUGO, vxi_nsproxy0),
23755 +       VINF("nsproxy1",S_IRUGO, vxi_nsproxy1),
23756 +       VINF("cvirt",   S_IRUGO, vxi_cvirt),
23757 +       VINF("cacct",   S_IRUGO, vxi_cacct),
23758 +       {}
23759 +};
23760 +
23761 +
23762 +
23763 +
23764 +static struct dentry *proc_xid_instantiate(struct inode *dir,
23765 +       struct dentry *dentry, int id, void *ptr)
23766 +{
23767 +       dentry->d_op = &proc_xid_dentry_operations;
23768 +       return vs_proc_instantiate(dir, dentry, id, ptr);
23769 +}
23770 +
23771 +static struct dentry *proc_xid_lookup(struct inode *dir,
23772 +       struct dentry *dentry, struct nameidata *nd)
23773 +{
23774 +       struct vs_entry *p = vx_base_stuff;
23775 +       struct dentry *error = ERR_PTR(-ENOENT);
23776 +
23777 +       for (; p->name; p++) {
23778 +               if (p->len != dentry->d_name.len)
23779 +                       continue;
23780 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
23781 +                       break;
23782 +       }
23783 +       if (!p->name)
23784 +               goto out;
23785 +
23786 +       error = proc_xid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
23787 +out:
23788 +       return error;
23789 +}
23790 +
23791 +static int proc_xid_readdir(struct file *filp,
23792 +       void *dirent, filldir_t filldir)
23793 +{
23794 +       struct dentry *dentry = filp->f_dentry;
23795 +       struct inode *inode = dentry->d_inode;
23796 +       struct vs_entry *p = vx_base_stuff;
23797 +       int size = sizeof(vx_base_stuff) / sizeof(struct vs_entry);
23798 +       int pos, index;
23799 +       u64 ino;
23800 +
23801 +       pos = filp->f_pos;
23802 +       switch (pos) {
23803 +       case 0:
23804 +               ino = inode->i_ino;
23805 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
23806 +                       goto out;
23807 +               pos++;
23808 +               /* fall through */
23809 +       case 1:
23810 +               ino = parent_ino(dentry);
23811 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
23812 +                       goto out;
23813 +               pos++;
23814 +               /* fall through */
23815 +       default:
23816 +               index = pos - 2;
23817 +               if (index >= size)
23818 +                       goto out;
23819 +               for (p += index; p->name; p++) {
23820 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
23821 +                               vs_proc_instantiate, PROC_I(inode)->fd, p))
23822 +                               goto out;
23823 +                       pos++;
23824 +               }
23825 +       }
23826 +out:
23827 +       filp->f_pos = pos;
23828 +       return 1;
23829 +}
23830 +
23831 +
23832 +
23833 +static struct file_operations proc_nx_info_file_operations = {
23834 +       .read =         proc_nx_info_read,
23835 +};
23836 +
23837 +static struct dentry_operations proc_nid_dentry_operations = {
23838 +       .d_revalidate = proc_nid_revalidate,
23839 +};
23840 +
23841 +static struct vs_entry nx_base_stuff[] = {
23842 +       NINF("info",    S_IRUGO, nxi_info),
23843 +       NINF("status",  S_IRUGO, nxi_status),
23844 +       {}
23845 +};
23846 +
23847 +
23848 +static struct dentry *proc_nid_instantiate(struct inode *dir,
23849 +       struct dentry *dentry, int id, void *ptr)
23850 +{
23851 +       dentry->d_op = &proc_nid_dentry_operations;
23852 +       return vs_proc_instantiate(dir, dentry, id, ptr);
23853 +}
23854 +
23855 +static struct dentry *proc_nid_lookup(struct inode *dir,
23856 +       struct dentry *dentry, struct nameidata *nd)
23857 +{
23858 +       struct vs_entry *p = nx_base_stuff;
23859 +       struct dentry *error = ERR_PTR(-ENOENT);
23860 +
23861 +       for (; p->name; p++) {
23862 +               if (p->len != dentry->d_name.len)
23863 +                       continue;
23864 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
23865 +                       break;
23866 +       }
23867 +       if (!p->name)
23868 +               goto out;
23869 +
23870 +       error = proc_nid_instantiate(dir, dentry, PROC_I(dir)->fd, p);
23871 +out:
23872 +       return error;
23873 +}
23874 +
23875 +static int proc_nid_readdir(struct file *filp,
23876 +       void *dirent, filldir_t filldir)
23877 +{
23878 +       struct dentry *dentry = filp->f_dentry;
23879 +       struct inode *inode = dentry->d_inode;
23880 +       struct vs_entry *p = nx_base_stuff;
23881 +       int size = sizeof(nx_base_stuff) / sizeof(struct vs_entry);
23882 +       int pos, index;
23883 +       u64 ino;
23884 +
23885 +       pos = filp->f_pos;
23886 +       switch (pos) {
23887 +       case 0:
23888 +               ino = inode->i_ino;
23889 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
23890 +                       goto out;
23891 +               pos++;
23892 +               /* fall through */
23893 +       case 1:
23894 +               ino = parent_ino(dentry);
23895 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
23896 +                       goto out;
23897 +               pos++;
23898 +               /* fall through */
23899 +       default:
23900 +               index = pos - 2;
23901 +               if (index >= size)
23902 +                       goto out;
23903 +               for (p += index; p->name; p++) {
23904 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
23905 +                               vs_proc_instantiate, PROC_I(inode)->fd, p))
23906 +                               goto out;
23907 +                       pos++;
23908 +               }
23909 +       }
23910 +out:
23911 +       filp->f_pos = pos;
23912 +       return 1;
23913 +}
23914 +
23915 +
23916 +#define MAX_MULBY10    ((~0U - 9) / 10)
23917 +
23918 +static inline int atovid(const char *str, int len)
23919 +{
23920 +       int vid, c;
23921 +
23922 +       vid = 0;
23923 +       while (len-- > 0) {
23924 +               c = *str - '0';
23925 +               str++;
23926 +               if (c > 9)
23927 +                       return -1;
23928 +               if (vid >= MAX_MULBY10)
23929 +                       return -1;
23930 +               vid *= 10;
23931 +               vid += c;
23932 +               if (!vid)
23933 +                       return -1;
23934 +       }
23935 +       return vid;
23936 +}
23937 +
23938 +/* now the upper level (virtual) */
23939 +
23940 +
23941 +static struct file_operations proc_xid_file_operations = {
23942 +       .read =         generic_read_dir,
23943 +       .readdir =      proc_xid_readdir,
23944 +};
23945 +
23946 +static struct inode_operations proc_xid_inode_operations = {
23947 +       .lookup =       proc_xid_lookup,
23948 +};
23949 +
23950 +static struct vs_entry vx_virtual_stuff[] = {
23951 +       INF("info",     S_IRUGO, virtual_info),
23952 +       INF("status",   S_IRUGO, virtual_status),
23953 +       DIR(NULL,       S_IRUGO | S_IXUGO, xid),
23954 +};
23955 +
23956 +
23957 +static struct dentry *proc_virtual_lookup(struct inode *dir,
23958 +       struct dentry *dentry, struct nameidata *nd)
23959 +{
23960 +       struct vs_entry *p = vx_virtual_stuff;
23961 +       struct dentry *error = ERR_PTR(-ENOENT);
23962 +       int id = 0;
23963 +
23964 +       for (; p->name; p++) {
23965 +               if (p->len != dentry->d_name.len)
23966 +                       continue;
23967 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
23968 +                       break;
23969 +       }
23970 +       if (p->name)
23971 +               goto instantiate;
23972 +
23973 +       id = atovid(dentry->d_name.name, dentry->d_name.len);
23974 +       if ((id < 0) || !xid_is_hashed(id))
23975 +               goto out;
23976 +
23977 +instantiate:
23978 +       error = proc_xid_instantiate(dir, dentry, id, p);
23979 +out:
23980 +       return error;
23981 +}
23982 +
23983 +static struct file_operations proc_nid_file_operations = {
23984 +       .read =         generic_read_dir,
23985 +       .readdir =      proc_nid_readdir,
23986 +};
23987 +
23988 +static struct inode_operations proc_nid_inode_operations = {
23989 +       .lookup =       proc_nid_lookup,
23990 +};
23991 +
23992 +static struct vs_entry nx_virtnet_stuff[] = {
23993 +       INF("info",     S_IRUGO, virtnet_info),
23994 +       INF("status",   S_IRUGO, virtnet_status),
23995 +       DIR(NULL,       S_IRUGO | S_IXUGO, nid),
23996 +};
23997 +
23998 +
23999 +static struct dentry *proc_virtnet_lookup(struct inode *dir,
24000 +       struct dentry *dentry, struct nameidata *nd)
24001 +{
24002 +       struct vs_entry *p = nx_virtnet_stuff;
24003 +       struct dentry *error = ERR_PTR(-ENOENT);
24004 +       int id = 0;
24005 +
24006 +       for (; p->name; p++) {
24007 +               if (p->len != dentry->d_name.len)
24008 +                       continue;
24009 +               if (!memcmp(dentry->d_name.name, p->name, p->len))
24010 +                       break;
24011 +       }
24012 +       if (p->name)
24013 +               goto instantiate;
24014 +
24015 +       id = atovid(dentry->d_name.name, dentry->d_name.len);
24016 +       if ((id < 0) || !nid_is_hashed(id))
24017 +               goto out;
24018 +
24019 +instantiate:
24020 +       error = proc_nid_instantiate(dir, dentry, id, p);
24021 +out:
24022 +       return error;
24023 +}
24024 +
24025 +
24026 +#define PROC_MAXVIDS 32
24027 +
24028 +int proc_virtual_readdir(struct file *filp,
24029 +       void *dirent, filldir_t filldir)
24030 +{
24031 +       struct dentry *dentry = filp->f_dentry;
24032 +       struct inode *inode = dentry->d_inode;
24033 +       struct vs_entry *p = vx_virtual_stuff;
24034 +       int size = sizeof(vx_virtual_stuff) / sizeof(struct vs_entry);
24035 +       int pos, index;
24036 +       unsigned int xid_array[PROC_MAXVIDS];
24037 +       char buf[PROC_NUMBUF];
24038 +       unsigned int nr_xids, i;
24039 +       u64 ino;
24040 +
24041 +       pos = filp->f_pos;
24042 +       switch (pos) {
24043 +       case 0:
24044 +               ino = inode->i_ino;
24045 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
24046 +                       goto out;
24047 +               pos++;
24048 +               /* fall through */
24049 +       case 1:
24050 +               ino = parent_ino(dentry);
24051 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
24052 +                       goto out;
24053 +               pos++;
24054 +               /* fall through */
24055 +       default:
24056 +               index = pos - 2;
24057 +               if (index >= size)
24058 +                       goto entries;
24059 +               for (p += index; p->name; p++) {
24060 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
24061 +                               vs_proc_instantiate, 0, p))
24062 +                               goto out;
24063 +                       pos++;
24064 +               }
24065 +       entries:
24066 +               index = pos - size;
24067 +               p = &vx_virtual_stuff[size - 1];
24068 +               nr_xids = get_xid_list(index, xid_array, PROC_MAXVIDS);
24069 +               for (i = 0; i < nr_xids; i++) {
24070 +                       int n, xid = xid_array[i];
24071 +                       unsigned int j = PROC_NUMBUF;
24072 +
24073 +                       n = xid;
24074 +                       do
24075 +                               buf[--j] = '0' + (n % 10);
24076 +                       while (n /= 10);
24077 +
24078 +                       if (proc_fill_cache(filp, dirent, filldir,
24079 +                               buf + j, PROC_NUMBUF - j,
24080 +                               vs_proc_instantiate, xid, p))
24081 +                               goto out;
24082 +                       pos++;
24083 +               }
24084 +       }
24085 +out:
24086 +       filp->f_pos = pos;
24087 +       return 0;
24088 +}
24089 +
24090 +static int proc_virtual_getattr(struct vfsmount *mnt,
24091 +       struct dentry *dentry, struct kstat *stat)
24092 +{
24093 +       struct inode *inode = dentry->d_inode;
24094 +
24095 +       generic_fillattr(inode, stat);
24096 +       stat->nlink = 2 + atomic_read(&vx_global_cactive);
24097 +       return 0;
24098 +}
24099 +
24100 +static struct file_operations proc_virtual_dir_operations = {
24101 +       .read =         generic_read_dir,
24102 +       .readdir =      proc_virtual_readdir,
24103 +};
24104 +
24105 +static struct inode_operations proc_virtual_dir_inode_operations = {
24106 +       .getattr =      proc_virtual_getattr,
24107 +       .lookup =       proc_virtual_lookup,
24108 +};
24109 +
24110 +
24111 +
24112 +
24113 +
24114 +int proc_virtnet_readdir(struct file *filp,
24115 +       void *dirent, filldir_t filldir)
24116 +{
24117 +       struct dentry *dentry = filp->f_dentry;
24118 +       struct inode *inode = dentry->d_inode;
24119 +       struct vs_entry *p = nx_virtnet_stuff;
24120 +       int size = sizeof(nx_virtnet_stuff) / sizeof(struct vs_entry);
24121 +       int pos, index;
24122 +       unsigned int nid_array[PROC_MAXVIDS];
24123 +       char buf[PROC_NUMBUF];
24124 +       unsigned int nr_nids, i;
24125 +       u64 ino;
24126 +
24127 +       pos = filp->f_pos;
24128 +       switch (pos) {
24129 +       case 0:
24130 +               ino = inode->i_ino;
24131 +               if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
24132 +                       goto out;
24133 +               pos++;
24134 +               /* fall through */
24135 +       case 1:
24136 +               ino = parent_ino(dentry);
24137 +               if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
24138 +                       goto out;
24139 +               pos++;
24140 +               /* fall through */
24141 +       default:
24142 +               index = pos - 2;
24143 +               if (index >= size)
24144 +                       goto entries;
24145 +               for (p += index; p->name; p++) {
24146 +                       if (proc_fill_cache(filp, dirent, filldir, p->name, p->len,
24147 +                               vs_proc_instantiate, 0, p))
24148 +                               goto out;
24149 +                       pos++;
24150 +               }
24151 +       entries:
24152 +               index = pos - size;
24153 +               p = &nx_virtnet_stuff[size - 1];
24154 +               nr_nids = get_nid_list(index, nid_array, PROC_MAXVIDS);
24155 +               for (i = 0; i < nr_nids; i++) {
24156 +                       int n, nid = nid_array[i];
24157 +                       unsigned int j = PROC_NUMBUF;
24158 +
24159 +                       n = nid;
24160 +                       do
24161 +                               buf[--j] = '0' + (n % 10);
24162 +                       while (n /= 10);
24163 +
24164 +                       if (proc_fill_cache(filp, dirent, filldir,
24165 +                               buf + j, PROC_NUMBUF - j,
24166 +                               vs_proc_instantiate, nid, p))
24167 +                               goto out;
24168 +                       pos++;
24169 +               }
24170 +       }
24171 +out:
24172 +       filp->f_pos = pos;
24173 +       return 0;
24174 +}
24175 +
24176 +static int proc_virtnet_getattr(struct vfsmount *mnt,
24177 +       struct dentry *dentry, struct kstat *stat)
24178 +{
24179 +       struct inode *inode = dentry->d_inode;
24180 +
24181 +       generic_fillattr(inode, stat);
24182 +       stat->nlink = 2 + atomic_read(&nx_global_cactive);
24183 +       return 0;
24184 +}
24185 +
24186 +static struct file_operations proc_virtnet_dir_operations = {
24187 +       .read =         generic_read_dir,
24188 +       .readdir =      proc_virtnet_readdir,
24189 +};
24190 +
24191 +static struct inode_operations proc_virtnet_dir_inode_operations = {
24192 +       .getattr =      proc_virtnet_getattr,
24193 +       .lookup =       proc_virtnet_lookup,
24194 +};
24195 +
24196 +
24197 +
24198 +void proc_vx_init(void)
24199 +{
24200 +       struct proc_dir_entry *ent;
24201 +
24202 +       ent = proc_mkdir("virtual", 0);
24203 +       if (ent) {
24204 +               ent->proc_fops = &proc_virtual_dir_operations;
24205 +               ent->proc_iops = &proc_virtual_dir_inode_operations;
24206 +       }
24207 +       proc_virtual = ent;
24208 +
24209 +       ent = proc_mkdir("virtnet", 0);
24210 +       if (ent) {
24211 +               ent->proc_fops = &proc_virtnet_dir_operations;
24212 +               ent->proc_iops = &proc_virtnet_dir_inode_operations;
24213 +       }
24214 +       proc_virtnet = ent;
24215 +}
24216 +
24217 +
24218 +
24219 +
24220 +/* per pid info */
24221 +
24222 +
24223 +int proc_pid_vx_info(struct task_struct *p, char *buffer)
24224 +{
24225 +       struct vx_info *vxi;
24226 +       char *orig = buffer;
24227 +
24228 +       buffer += sprintf(buffer, "XID:\t%d\n", vx_task_xid(p));
24229 +
24230 +       vxi = task_get_vx_info(p);
24231 +       if (!vxi)
24232 +               goto out;
24233 +
24234 +       buffer += sprintf(buffer, "BCaps:\t");
24235 +       buffer = print_cap_t(buffer, &vxi->vx_bcaps);
24236 +       buffer += sprintf(buffer, "\n");
24237 +       buffer += sprintf(buffer, "CCaps:\t%016llx\n",
24238 +               (unsigned long long)vxi->vx_ccaps);
24239 +       buffer += sprintf(buffer, "CFlags:\t%016llx\n",
24240 +               (unsigned long long)vxi->vx_flags);
24241 +       buffer += sprintf(buffer, "CIPid:\t%d\n", vxi->vx_initpid);
24242 +
24243 +       put_vx_info(vxi);
24244 +out:
24245 +       return buffer - orig;
24246 +}
24247 +
24248 +
24249 +int proc_pid_nx_info(struct task_struct *p, char *buffer)
24250 +{
24251 +       struct nx_info *nxi;
24252 +       struct nx_addr_v4 *v4a;
24253 +#ifdef CONFIG_IPV6
24254 +       struct nx_addr_v6 *v6a;
24255 +#endif
24256 +       char *orig = buffer;
24257 +       int i;
24258 +
24259 +       buffer += sprintf(buffer, "NID:\t%d\n", nx_task_nid(p));
24260 +
24261 +       nxi = task_get_nx_info(p);
24262 +       if (!nxi)
24263 +               goto out;
24264 +
24265 +       buffer += sprintf(buffer, "NCaps:\t%016llx\n",
24266 +               (unsigned long long)nxi->nx_ncaps);
24267 +       buffer += sprintf(buffer, "NFlags:\t%016llx\n",
24268 +               (unsigned long long)nxi->nx_flags);
24269 +
24270 +       buffer += sprintf(buffer,
24271 +               "V4Root[bcast]:\t" NIPQUAD_FMT "\n",
24272 +               NIPQUAD(nxi->v4_bcast.s_addr));
24273 +       buffer += sprintf (buffer,
24274 +               "V4Root[lback]:\t" NIPQUAD_FMT "\n",
24275 +               NIPQUAD(nxi->v4_lback.s_addr));
24276 +       if (!NX_IPV4(nxi))
24277 +               goto skip_v4;
24278 +       for (i = 0, v4a = &nxi->v4; v4a; i++, v4a = v4a->next)
24279 +               buffer += sprintf(buffer, "V4Root[%d]:\t" NXAV4_FMT "\n",
24280 +                       i, NXAV4(v4a));
24281 +skip_v4:
24282 +#ifdef CONFIG_IPV6
24283 +       if (!NX_IPV6(nxi))
24284 +               goto skip_v6;
24285 +       for (i = 0, v6a = &nxi->v6; v6a; i++, v6a = v6a->next)
24286 +               buffer += sprintf(buffer, "V6Root[%d]:\t" NXAV6_FMT "\n",
24287 +                       i, NXAV6(v6a));
24288 +skip_v6:
24289 +#endif
24290 +       put_nx_info(nxi);
24291 +out:
24292 +       return buffer - orig;
24293 +}
24294 +
24295 diff -NurpP --minimal linux-3.0/kernel/vserver/sched.c linux-3.0-vs2.3.1-pre7/kernel/vserver/sched.c
24296 --- linux-3.0/kernel/vserver/sched.c    1970-01-01 01:00:00.000000000 +0100
24297 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/sched.c       2011-06-10 22:11:24.000000000 +0200
24298 @@ -0,0 +1,82 @@
24299 +/*
24300 + *  linux/kernel/vserver/sched.c
24301 + *
24302 + *  Virtual Server: Scheduler Support
24303 + *
24304 + *  Copyright (C) 2004-2010  Herbert Pötzl
24305 + *
24306 + *  V0.01  adapted Sam Vilains version to 2.6.3
24307 + *  V0.02  removed legacy interface
24308 + *  V0.03  changed vcmds to vxi arg
24309 + *  V0.04  removed older and legacy interfaces
24310 + *  V0.05  removed scheduler code/commands
24311 + *
24312 + */
24313 +
24314 +#include <linux/vs_context.h>
24315 +#include <linux/vs_sched.h>
24316 +#include <linux/vserver/sched_cmd.h>
24317 +
24318 +#include <asm/uaccess.h>
24319 +
24320 +
24321 +void vx_update_sched_param(struct _vx_sched *sched,
24322 +       struct _vx_sched_pc *sched_pc)
24323 +{
24324 +       sched_pc->prio_bias = sched->prio_bias;
24325 +}
24326 +
24327 +static int do_set_prio_bias(struct vx_info *vxi, struct vcmd_prio_bias *data)
24328 +{
24329 +       int cpu;
24330 +
24331 +       if (data->prio_bias > MAX_PRIO_BIAS)
24332 +               data->prio_bias = MAX_PRIO_BIAS;
24333 +       if (data->prio_bias < MIN_PRIO_BIAS)
24334 +               data->prio_bias = MIN_PRIO_BIAS;
24335 +
24336 +       if (data->cpu_id != ~0) {
24337 +               vxi->sched.update = cpumask_of_cpu(data->cpu_id);
24338 +               cpus_and(vxi->sched.update, cpu_online_map,
24339 +                       vxi->sched.update);
24340 +       } else
24341 +               vxi->sched.update = cpu_online_map;
24342 +
24343 +       for_each_cpu_mask(cpu, vxi->sched.update)
24344 +               vx_update_sched_param(&vxi->sched,
24345 +                       &vx_per_cpu(vxi, sched_pc, cpu));
24346 +       return 0;
24347 +}
24348 +
24349 +int vc_set_prio_bias(struct vx_info *vxi, void __user *data)
24350 +{
24351 +       struct vcmd_prio_bias vc_data;
24352 +
24353 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
24354 +               return -EFAULT;
24355 +
24356 +       return do_set_prio_bias(vxi, &vc_data);
24357 +}
24358 +
24359 +int vc_get_prio_bias(struct vx_info *vxi, void __user *data)
24360 +{
24361 +       struct vcmd_prio_bias vc_data;
24362 +       struct _vx_sched_pc *pcd;
24363 +       int cpu;
24364 +
24365 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
24366 +               return -EFAULT;
24367 +
24368 +       cpu = vc_data.cpu_id;
24369 +
24370 +       if (!cpu_possible(cpu))
24371 +               return -EINVAL;
24372 +
24373 +       pcd = &vx_per_cpu(vxi, sched_pc, cpu);
24374 +       vc_data.prio_bias = pcd->prio_bias;
24375 +
24376 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
24377 +               return -EFAULT;
24378 +       return 0;
24379 +}
24380 +
24381 diff -NurpP --minimal linux-3.0/kernel/vserver/sched_init.h linux-3.0-vs2.3.1-pre7/kernel/vserver/sched_init.h
24382 --- linux-3.0/kernel/vserver/sched_init.h       1970-01-01 01:00:00.000000000 +0100
24383 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/sched_init.h  2011-06-10 22:11:24.000000000 +0200
24384 @@ -0,0 +1,27 @@
24385 +
24386 +static inline void vx_info_init_sched(struct _vx_sched *sched)
24387 +{
24388 +       /* scheduling; hard code starting values as constants */
24389 +       sched->prio_bias = 0;
24390 +}
24391 +
24392 +static inline
24393 +void vx_info_init_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
24394 +{
24395 +       sched_pc->prio_bias = 0;
24396 +
24397 +       sched_pc->user_ticks = 0;
24398 +       sched_pc->sys_ticks = 0;
24399 +       sched_pc->hold_ticks = 0;
24400 +}
24401 +
24402 +static inline void vx_info_exit_sched(struct _vx_sched *sched)
24403 +{
24404 +       return;
24405 +}
24406 +
24407 +static inline
24408 +void vx_info_exit_sched_pc(struct _vx_sched_pc *sched_pc, int cpu)
24409 +{
24410 +       return;
24411 +}
24412 diff -NurpP --minimal linux-3.0/kernel/vserver/sched_proc.h linux-3.0-vs2.3.1-pre7/kernel/vserver/sched_proc.h
24413 --- linux-3.0/kernel/vserver/sched_proc.h       1970-01-01 01:00:00.000000000 +0100
24414 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/sched_proc.h  2011-06-10 22:11:24.000000000 +0200
24415 @@ -0,0 +1,32 @@
24416 +#ifndef _VX_SCHED_PROC_H
24417 +#define _VX_SCHED_PROC_H
24418 +
24419 +
24420 +static inline
24421 +int vx_info_proc_sched(struct _vx_sched *sched, char *buffer)
24422 +{
24423 +       int length = 0;
24424 +
24425 +       length += sprintf(buffer,
24426 +               "PrioBias:\t%8d\n",
24427 +               sched->prio_bias);
24428 +       return length;
24429 +}
24430 +
24431 +static inline
24432 +int vx_info_proc_sched_pc(struct _vx_sched_pc *sched_pc,
24433 +       char *buffer, int cpu)
24434 +{
24435 +       int length = 0;
24436 +
24437 +       length += sprintf(buffer + length,
24438 +               "cpu %d: %lld %lld %lld", cpu,
24439 +               (unsigned long long)sched_pc->user_ticks,
24440 +               (unsigned long long)sched_pc->sys_ticks,
24441 +               (unsigned long long)sched_pc->hold_ticks);
24442 +       length += sprintf(buffer + length,
24443 +               " %d\n", sched_pc->prio_bias);
24444 +       return length;
24445 +}
24446 +
24447 +#endif /* _VX_SCHED_PROC_H */
24448 diff -NurpP --minimal linux-3.0/kernel/vserver/signal.c linux-3.0-vs2.3.1-pre7/kernel/vserver/signal.c
24449 --- linux-3.0/kernel/vserver/signal.c   1970-01-01 01:00:00.000000000 +0100
24450 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/signal.c      2011-06-10 22:11:24.000000000 +0200
24451 @@ -0,0 +1,134 @@
24452 +/*
24453 + *  linux/kernel/vserver/signal.c
24454 + *
24455 + *  Virtual Server: Signal Support
24456 + *
24457 + *  Copyright (C) 2003-2007  Herbert Pötzl
24458 + *
24459 + *  V0.01  broken out from vcontext V0.05
24460 + *  V0.02  changed vcmds to vxi arg
24461 + *  V0.03  adjusted siginfo for kill
24462 + *
24463 + */
24464 +
24465 +#include <asm/uaccess.h>
24466 +
24467 +#include <linux/vs_context.h>
24468 +#include <linux/vs_pid.h>
24469 +#include <linux/vserver/signal_cmd.h>
24470 +
24471 +
24472 +int vx_info_kill(struct vx_info *vxi, int pid, int sig)
24473 +{
24474 +       int retval, count = 0;
24475 +       struct task_struct *p;
24476 +       struct siginfo *sip = SEND_SIG_PRIV;
24477 +
24478 +       retval = -ESRCH;
24479 +       vxdprintk(VXD_CBIT(misc, 4),
24480 +               "vx_info_kill(%p[#%d],%d,%d)*",
24481 +               vxi, vxi->vx_id, pid, sig);
24482 +       read_lock(&tasklist_lock);
24483 +       switch (pid) {
24484 +       case  0:
24485 +       case -1:
24486 +               for_each_process(p) {
24487 +                       int err = 0;
24488 +
24489 +                       if (vx_task_xid(p) != vxi->vx_id || p->pid <= 1 ||
24490 +                               (pid && vxi->vx_initpid == p->pid))
24491 +                               continue;
24492 +
24493 +                       err = group_send_sig_info(sig, sip, p);
24494 +                       ++count;
24495 +                       if (err != -EPERM)
24496 +                               retval = err;
24497 +               }
24498 +               break;
24499 +
24500 +       case 1:
24501 +               if (vxi->vx_initpid) {
24502 +                       pid = vxi->vx_initpid;
24503 +                       /* for now, only SIGINT to private init ... */
24504 +                       if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
24505 +                               /* ... as long as there are tasks left */
24506 +                               (atomic_read(&vxi->vx_tasks) > 1))
24507 +                               sig = SIGINT;
24508 +               }
24509 +               /* fallthrough */
24510 +       default:
24511 +               rcu_read_lock();
24512 +               p = find_task_by_real_pid(pid);
24513 +               rcu_read_unlock();
24514 +               if (p) {
24515 +                       if (vx_task_xid(p) == vxi->vx_id)
24516 +                               retval = group_send_sig_info(sig, sip, p);
24517 +               }
24518 +               break;
24519 +       }
24520 +       read_unlock(&tasklist_lock);
24521 +       vxdprintk(VXD_CBIT(misc, 4),
24522 +               "vx_info_kill(%p[#%d],%d,%d,%ld) = %d",
24523 +               vxi, vxi->vx_id, pid, sig, (long)sip, retval);
24524 +       return retval;
24525 +}
24526 +
24527 +int vc_ctx_kill(struct vx_info *vxi, void __user *data)
24528 +{
24529 +       struct vcmd_ctx_kill_v0 vc_data;
24530 +
24531 +       if (copy_from_user(&vc_data, data, sizeof(vc_data)))
24532 +               return -EFAULT;
24533 +
24534 +       /* special check to allow guest shutdown */
24535 +       if (!vx_info_flags(vxi, VXF_STATE_ADMIN, 0) &&
24536 +               /* forbid killall pid=0 when init is present */
24537 +               (((vc_data.pid < 1) && vxi->vx_initpid) ||
24538 +               (vc_data.pid > 1)))
24539 +               return -EACCES;
24540 +
24541 +       return vx_info_kill(vxi, vc_data.pid, vc_data.sig);
24542 +}
24543 +
24544 +
24545 +static int __wait_exit(struct vx_info *vxi)
24546 +{
24547 +       DECLARE_WAITQUEUE(wait, current);
24548 +       int ret = 0;
24549 +
24550 +       add_wait_queue(&vxi->vx_wait, &wait);
24551 +       set_current_state(TASK_INTERRUPTIBLE);
24552 +
24553 +wait:
24554 +       if (vx_info_state(vxi,
24555 +               VXS_SHUTDOWN | VXS_HASHED | VXS_HELPER) == VXS_SHUTDOWN)
24556 +               goto out;
24557 +       if (signal_pending(current)) {
24558 +               ret = -ERESTARTSYS;
24559 +               goto out;
24560 +       }
24561 +       schedule();
24562 +       goto wait;
24563 +
24564 +out:
24565 +       set_current_state(TASK_RUNNING);
24566 +       remove_wait_queue(&vxi->vx_wait, &wait);
24567 +       return ret;
24568 +}
24569 +
24570 +
24571 +
24572 +int vc_wait_exit(struct vx_info *vxi, void __user *data)
24573 +{
24574 +       struct vcmd_wait_exit_v0 vc_data;
24575 +       int ret;
24576 +
24577 +       ret = __wait_exit(vxi);
24578 +       vc_data.reboot_cmd = vxi->reboot_cmd;
24579 +       vc_data.exit_code = vxi->exit_code;
24580 +
24581 +       if (copy_to_user(data, &vc_data, sizeof(vc_data)))
24582 +               ret = -EFAULT;
24583 +       return ret;
24584 +}
24585 +
24586 diff -NurpP --minimal linux-3.0/kernel/vserver/space.c linux-3.0-vs2.3.1-pre7/kernel/vserver/space.c
24587 --- linux-3.0/kernel/vserver/space.c    1970-01-01 01:00:00.000000000 +0100
24588 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/space.c       2011-07-20 02:11:49.000000000 +0200
24589 @@ -0,0 +1,435 @@
24590 +/*
24591 + *  linux/kernel/vserver/space.c
24592 + *
24593 + *  Virtual Server: Context Space Support
24594 + *
24595 + *  Copyright (C) 2003-2010  Herbert Pötzl
24596 + *
24597 + *  V0.01  broken out from context.c 0.07
24598 + *  V0.02  added task locking for namespace
24599 + *  V0.03  broken out vx_enter_namespace
24600 + *  V0.04  added *space support and commands
24601 + *  V0.05  added credential support
24602 + *
24603 + */
24604 +
24605 +#include <linux/utsname.h>
24606 +#include <linux/nsproxy.h>
24607 +#include <linux/err.h>
24608 +#include <linux/fs_struct.h>
24609 +#include <linux/cred.h>
24610 +#include <asm/uaccess.h>
24611 +
24612 +#include <linux/vs_context.h>
24613 +#include <linux/vserver/space.h>
24614 +#include <linux/vserver/space_cmd.h>
24615 +
24616 +atomic_t vs_global_nsproxy     = ATOMIC_INIT(0);
24617 +atomic_t vs_global_fs          = ATOMIC_INIT(0);
24618 +atomic_t vs_global_mnt_ns      = ATOMIC_INIT(0);
24619 +atomic_t vs_global_uts_ns      = ATOMIC_INIT(0);
24620 +atomic_t vs_global_user_ns     = ATOMIC_INIT(0);
24621 +atomic_t vs_global_pid_ns      = ATOMIC_INIT(0);
24622 +
24623 +
24624 +/* namespace functions */
24625 +
24626 +#include <linux/mnt_namespace.h>
24627 +#include <linux/user_namespace.h>
24628 +#include <linux/pid_namespace.h>
24629 +#include <linux/ipc_namespace.h>
24630 +#include <net/net_namespace.h>
24631 +
24632 +
24633 +static const struct vcmd_space_mask_v1 space_mask_v0 = {
24634 +       .mask = CLONE_FS |
24635 +               CLONE_NEWNS |
24636 +#ifdef CONFIG_UTS_NS
24637 +               CLONE_NEWUTS |
24638 +#endif
24639 +#ifdef CONFIG_IPC_NS
24640 +               CLONE_NEWIPC |
24641 +#endif
24642 +#ifdef CONFIG_USER_NS
24643 +               CLONE_NEWUSER |
24644 +#endif
24645 +               0
24646 +};
24647 +
24648 +static const struct vcmd_space_mask_v1 space_mask = {
24649 +       .mask = CLONE_FS |
24650 +               CLONE_NEWNS |
24651 +#ifdef CONFIG_UTS_NS
24652 +               CLONE_NEWUTS |
24653 +#endif
24654 +#ifdef CONFIG_IPC_NS
24655 +               CLONE_NEWIPC |
24656 +#endif
24657 +#ifdef CONFIG_USER_NS
24658 +               CLONE_NEWUSER |
24659 +#endif
24660 +#ifdef CONFIG_PID_NS
24661 +               CLONE_NEWPID |
24662 +#endif
24663 +#ifdef CONFIG_NET_NS
24664 +               CLONE_NEWNET |
24665 +#endif
24666 +               0
24667 +};
24668 +
24669 +static const struct vcmd_space_mask_v1 default_space_mask = {
24670 +       .mask = CLONE_FS |
24671 +               CLONE_NEWNS |
24672 +#ifdef CONFIG_UTS_NS
24673 +               CLONE_NEWUTS |
24674 +#endif
24675 +#ifdef CONFIG_IPC_NS
24676 +               CLONE_NEWIPC |
24677 +#endif
24678 +#ifdef CONFIG_USER_NS
24679 +               CLONE_NEWUSER |
24680 +#endif
24681 +#ifdef CONFIG_PID_NS
24682 +//             CLONE_NEWPID |
24683 +#endif
24684 +               0
24685 +};
24686 +
24687 +/*
24688 + *     build a new nsproxy mix
24689 + *      assumes that both proxies are 'const'
24690 + *     does not touch nsproxy refcounts
24691 + *     will hold a reference on the result.
24692 + */
24693 +
24694 +struct nsproxy *vs_mix_nsproxy(struct nsproxy *old_nsproxy,
24695 +       struct nsproxy *new_nsproxy, unsigned long mask)
24696 +{
24697 +       struct mnt_namespace *old_ns;
24698 +       struct uts_namespace *old_uts;
24699 +       struct ipc_namespace *old_ipc;
24700 +#ifdef CONFIG_PID_NS
24701 +       struct pid_namespace *old_pid;
24702 +#endif
24703 +#ifdef CONFIG_NET_NS
24704 +       struct net *old_net;
24705 +#endif
24706 +       struct nsproxy *nsproxy;
24707 +
24708 +       nsproxy = copy_nsproxy(old_nsproxy);
24709 +       if (!nsproxy)
24710 +               goto out;
24711 +
24712 +       if (mask & CLONE_NEWNS) {
24713 +               old_ns = nsproxy->mnt_ns;
24714 +               nsproxy->mnt_ns = new_nsproxy->mnt_ns;
24715 +               if (nsproxy->mnt_ns)
24716 +                       get_mnt_ns(nsproxy->mnt_ns);
24717 +       } else
24718 +               old_ns = NULL;
24719 +
24720 +       if (mask & CLONE_NEWUTS) {
24721 +               old_uts = nsproxy->uts_ns;
24722 +               nsproxy->uts_ns = new_nsproxy->uts_ns;
24723 +               if (nsproxy->uts_ns)
24724 +                       get_uts_ns(nsproxy->uts_ns);
24725 +       } else
24726 +               old_uts = NULL;
24727 +
24728 +       if (mask & CLONE_NEWIPC) {
24729 +               old_ipc = nsproxy->ipc_ns;
24730 +               nsproxy->ipc_ns = new_nsproxy->ipc_ns;
24731 +               if (nsproxy->ipc_ns)
24732 +                       get_ipc_ns(nsproxy->ipc_ns);
24733 +       } else
24734 +               old_ipc = NULL;
24735 +
24736 +#ifdef CONFIG_PID_NS
24737 +       if (mask & CLONE_NEWPID) {
24738 +               old_pid = nsproxy->pid_ns;
24739 +               nsproxy->pid_ns = new_nsproxy->pid_ns;
24740 +               if (nsproxy->pid_ns)
24741 +                       get_pid_ns(nsproxy->pid_ns);
24742 +       } else
24743 +               old_pid = NULL;
24744 +#endif
24745 +#ifdef CONFIG_NET_NS
24746 +       if (mask & CLONE_NEWNET) {
24747 +               old_net = nsproxy->net_ns;
24748 +               nsproxy->net_ns = new_nsproxy->net_ns;
24749 +               if (nsproxy->net_ns)
24750 +                       get_net(nsproxy->net_ns);
24751 +       } else
24752 +               old_net = NULL;
24753 +#endif
24754 +       if (old_ns)
24755 +               put_mnt_ns(old_ns);
24756 +       if (old_uts)
24757 +               put_uts_ns(old_uts);
24758 +       if (old_ipc)
24759 +               put_ipc_ns(old_ipc);
24760 +#ifdef CONFIG_PID_NS
24761 +       if (old_pid)
24762 +               put_pid_ns(old_pid);
24763 +#endif
24764 +#ifdef CONFIG_NET_NS
24765 +       if (old_net)
24766 +               put_net(old_net);
24767 +#endif
24768 +out:
24769 +       return nsproxy;
24770 +}
24771 +
24772 +
24773 +/*
24774 + *     merge two nsproxy structs into a new one.
24775 + *     will hold a reference on the result.
24776 + */
24777 +
24778 +static inline
24779 +struct nsproxy *__vs_merge_nsproxy(struct nsproxy *old,
24780 +       struct nsproxy *proxy, unsigned long mask)
24781 +{
24782 +       struct nsproxy null_proxy = { .mnt_ns = NULL };
24783 +
24784 +       if (!proxy)
24785 +               return NULL;
24786 +
24787 +       if (mask) {
24788 +               /* vs_mix_nsproxy returns with reference */
24789 +               return vs_mix_nsproxy(old ? old : &null_proxy,
24790 +                       proxy, mask);
24791 +       }
24792 +       get_nsproxy(proxy);
24793 +       return proxy;
24794 +}
24795 +
24796 +
24797 +int vx_enter_space(struct vx_info *vxi, unsigned long mask, unsigned index)
24798 +{
24799 +       struct nsproxy *proxy, *proxy_cur, *proxy_new;
24800 +       struct fs_struct *fs_cur, *fs = NULL;
24801 +       struct _vx_space *space;
24802 +       int ret, kill = 0;
24803 +
24804 +       vxdprintk(VXD_CBIT(space, 8), "vx_enter_space(%p[#%u],0x%08lx,%d)",
24805 +               vxi, vxi->vx_id, mask, index);
24806 +
24807 +       if (vx_info_flags(vxi, VXF_INFO_PRIVATE, 0))
24808 +               return -EACCES;
24809 +
24810 +       if (index >= VX_SPACES)
24811 +               return -EINVAL;
24812 +
24813 +       space = &vxi->space[index];
24814 +
24815 +       if (!mask)
24816 +               mask = space->vx_nsmask;
24817 +
24818 +       if ((mask & space->vx_nsmask) != mask)
24819 +               return -EINVAL;
24820 +
24821 +       if (mask & CLONE_FS) {
24822 +               fs = copy_fs_struct(space->vx_fs);
24823 +               if (!fs)
24824 +                       return -ENOMEM;
24825 +       }
24826 +       proxy = space->vx_nsproxy;
24827 +
24828 +       vxdprintk(VXD_CBIT(space, 9),
24829 +               "vx_enter_space(%p[#%u],0x%08lx,%d) -> (%p,%p)",
24830 +               vxi, vxi->vx_id, mask, index, proxy, fs);
24831 +
24832 +       task_lock(current);
24833 +       fs_cur = current->fs;
24834 +
24835 +       if (mask & CLONE_FS) {
24836 +               spin_lock(&fs_cur->lock);
24837 +               current->fs = fs;
24838 +               kill = !--fs_cur->users;
24839 +               spin_unlock(&fs_cur->lock);
24840 +       }
24841 +
24842 +       proxy_cur = current->nsproxy;
24843 +       get_nsproxy(proxy_cur);
24844 +       task_unlock(current);
24845 +
24846 +       if (kill)
24847 +               free_fs_struct(fs_cur);
24848 +
24849 +       proxy_new = __vs_merge_nsproxy(proxy_cur, proxy, mask);
24850 +       if (IS_ERR(proxy_new)) {
24851 +               ret = PTR_ERR(proxy_new);
24852 +               goto out_put;
24853 +       }
24854 +
24855 +       proxy_new = xchg(&current->nsproxy, proxy_new);
24856 +
24857 +       if (mask & CLONE_NEWUSER) {
24858 +               struct cred *cred;
24859 +
24860 +               vxdprintk(VXD_CBIT(space, 10),
24861 +                       "vx_enter_space(%p[#%u],%p) cred (%p,%p)",
24862 +                       vxi, vxi->vx_id, space->vx_cred,
24863 +                       current->real_cred, current->cred);
24864 +
24865 +               if (space->vx_cred) {
24866 +                       cred = __prepare_creds(space->vx_cred);
24867 +                       if (cred)
24868 +                               commit_creds(cred);
24869 +               }
24870 +       }
24871 +
24872 +       ret = 0;
24873 +
24874 +       if (proxy_new)
24875 +               put_nsproxy(proxy_new);
24876 +out_put:
24877 +       if (proxy_cur)
24878 +               put_nsproxy(proxy_cur);
24879 +       return ret;
24880 +}
24881 +
24882 +
24883 +int vx_set_space(struct vx_info *vxi, unsigned long mask, unsigned index)
24884 +{
24885 +       struct nsproxy *proxy_vxi, *proxy_cur, *proxy_new;
24886 +       struct fs_struct *fs_vxi, *fs;
24887 +       struct _vx_space *space;
24888 +       int ret, kill = 0;
24889 +
24890 +       vxdprintk(VXD_CBIT(space, 8), "vx_set_space(%p[#%u],0x%08lx,%d)",
24891 +               vxi, vxi->vx_id, mask, index);
24892 +
24893 +       if ((mask & space_mask.mask) != mask)
24894 +               return -EINVAL;
24895 +
24896 +       if (index >= VX_SPACES)
24897 +               return -EINVAL;
24898 +
24899 +       space = &vxi->space[index];
24900 +
24901 +       proxy_vxi = space->vx_nsproxy;
24902 +       fs_vxi = space->vx_fs;
24903 +
24904 +       if (mask & CLONE_FS) {
24905 +               fs = copy_fs_struct(current->fs);
24906 +               if (!fs)
24907 +                       return -ENOMEM;
24908 +       }
24909 +
24910 +       task_lock(current);
24911 +
24912 +       if (mask & CLONE_FS) {
24913 +               spin_lock(&fs_vxi->lock);
24914 +               space->vx_fs = fs;
24915 +               kill = !--fs_vxi->users;
24916 +               spin_unlock(&fs_vxi->lock);
24917 +       }
24918 +
24919 +       proxy_cur = current->nsproxy;
24920 +       get_nsproxy(proxy_cur);
24921 +       task_unlock(current);
24922 +
24923 +       if (kill)
24924 +               free_fs_struct(fs_vxi);
24925 +
24926 +       proxy_new = __vs_merge_nsproxy(proxy_vxi, proxy_cur, mask);
24927 +       if (IS_ERR(proxy_new)) {
24928 +               ret = PTR_ERR(proxy_new);
24929 +               goto out_put;
24930 +       }
24931 +
24932 +       proxy_new = xchg(&space->vx_nsproxy, proxy_new);
24933 +       space->vx_nsmask |= mask;
24934 +
24935 +       if (mask & CLONE_NEWUSER) {
24936 +               struct cred *cred;
24937 +
24938 +               vxdprintk(VXD_CBIT(space, 10),
24939 +                       "vx_set_space(%p[#%u],%p) cred (%p,%p)",
24940 +                       vxi, vxi->vx_id, space->vx_cred,
24941 +                       current->real_cred, current->cred);
24942 +
24943 +               cred = prepare_creds();
24944 +               cred = (struct cred *)xchg(&space->vx_cred, cred);
24945 +               if (cred)
24946 +                       abort_creds(cred);
24947 +       }
24948 +
24949 +       ret = 0;
24950 +
24951 +       if (proxy_new)
24952 +               put_nsproxy(proxy_new);
24953 +out_put:
24954 +       if (proxy_cur)
24955 +               put_nsproxy(proxy_cur);
24956 +       return ret;
24957 +}
24958 +
24959 +
24960 +int vc_enter_space_v1(struct vx_info *vxi, void __user *data)
24961 +{
24962 +       struct vcmd_space_mask_v1 vc_data = { .mask = 0 };
24963 +
24964 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
24965 +               return -EFAULT;
24966 +
24967 +       return vx_enter_space(vxi, vc_data.mask, 0);
24968 +}
24969 +
24970 +int vc_enter_space(struct vx_info *vxi, void __user *data)
24971 +{
24972 +       struct vcmd_space_mask_v2 vc_data = { .mask = 0 };
24973 +
24974 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
24975 +               return -EFAULT;
24976 +
24977 +       if (vc_data.index >= VX_SPACES)
24978 +               return -EINVAL;
24979 +
24980 +       return vx_enter_space(vxi, vc_data.mask, vc_data.index);
24981 +}
24982 +
24983 +int vc_set_space_v1(struct vx_info *vxi, void __user *data)
24984 +{
24985 +       struct vcmd_space_mask_v1 vc_data = { .mask = 0 };
24986 +
24987 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
24988 +               return -EFAULT;
24989 +
24990 +       return vx_set_space(vxi, vc_data.mask, 0);
24991 +}
24992 +
24993 +int vc_set_space(struct vx_info *vxi, void __user *data)
24994 +{
24995 +       struct vcmd_space_mask_v2 vc_data = { .mask = 0 };
24996 +
24997 +       if (data && copy_from_user(&vc_data, data, sizeof(vc_data)))
24998 +               return -EFAULT;
24999 +
25000 +       if (vc_data.index >= VX_SPACES)
25001 +               return -EINVAL;
25002 +
25003 +       return vx_set_space(vxi, vc_data.mask, vc_data.index);
25004 +}
25005 +
25006 +int vc_get_space_mask(void __user *data, int type)
25007 +{
25008 +       const struct vcmd_space_mask_v1 *mask;
25009 +
25010 +       if (type == 0)
25011 +               mask = &space_mask_v0;
25012 +       else if (type == 1)
25013 +               mask = &space_mask;
25014 +       else
25015 +               mask = &default_space_mask;
25016 +
25017 +       vxdprintk(VXD_CBIT(space, 10),
25018 +               "vc_get_space_mask(%d) = %08llx", type, mask->mask);
25019 +
25020 +       if (copy_to_user(data, mask, sizeof(*mask)))
25021 +               return -EFAULT;
25022 +       return 0;
25023 +}
25024 +
25025 diff -NurpP --minimal linux-3.0/kernel/vserver/switch.c linux-3.0-vs2.3.1-pre7/kernel/vserver/switch.c
25026 --- linux-3.0/kernel/vserver/switch.c   1970-01-01 01:00:00.000000000 +0100
25027 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/switch.c      2011-07-20 02:40:51.000000000 +0200
25028 @@ -0,0 +1,548 @@
25029 +/*
25030 + *  linux/kernel/vserver/switch.c
25031 + *
25032 + *  Virtual Server: Syscall Switch
25033 + *
25034 + *  Copyright (C) 2003-2007  Herbert Pötzl
25035 + *
25036 + *  V0.01  syscall switch
25037 + *  V0.02  added signal to context
25038 + *  V0.03  added rlimit functions
25039 + *  V0.04  added iattr, task/xid functions
25040 + *  V0.05  added debug/history stuff
25041 + *  V0.06  added compat32 layer
25042 + *  V0.07  vcmd args and perms
25043 + *  V0.08  added status commands
25044 + *  V0.09  added tag commands
25045 + *  V0.10  added oom bias
25046 + *  V0.11  added device commands
25047 + *
25048 + */
25049 +
25050 +#include <linux/vs_context.h>
25051 +#include <linux/vs_network.h>
25052 +#include <linux/vserver/switch.h>
25053 +
25054 +#include "vci_config.h"
25055 +
25056 +
25057 +static inline
25058 +int vc_get_version(uint32_t id)
25059 +{
25060 +       return VCI_VERSION;
25061 +}
25062 +
25063 +static inline
25064 +int vc_get_vci(uint32_t id)
25065 +{
25066 +       return vci_kernel_config();
25067 +}
25068 +
25069 +#include <linux/vserver/context_cmd.h>
25070 +#include <linux/vserver/cvirt_cmd.h>
25071 +#include <linux/vserver/cacct_cmd.h>
25072 +#include <linux/vserver/limit_cmd.h>
25073 +#include <linux/vserver/network_cmd.h>
25074 +#include <linux/vserver/sched_cmd.h>
25075 +#include <linux/vserver/debug_cmd.h>
25076 +#include <linux/vserver/inode_cmd.h>
25077 +#include <linux/vserver/dlimit_cmd.h>
25078 +#include <linux/vserver/signal_cmd.h>
25079 +#include <linux/vserver/space_cmd.h>
25080 +#include <linux/vserver/tag_cmd.h>
25081 +#include <linux/vserver/device_cmd.h>
25082 +
25083 +#include <linux/vserver/inode.h>
25084 +#include <linux/vserver/dlimit.h>
25085 +
25086 +
25087 +#ifdef CONFIG_COMPAT
25088 +#define __COMPAT(name, id, data, compat)       \
25089 +       (compat) ? name ## _x32(id, data) : name(id, data)
25090 +#define __COMPAT_NO_ID(name, data, compat)     \
25091 +       (compat) ? name ## _x32(data) : name(data)
25092 +#else
25093 +#define __COMPAT(name, id, data, compat)       \
25094 +       name(id, data)
25095 +#define __COMPAT_NO_ID(name, data, compat)     \
25096 +       name(data)
25097 +#endif
25098 +
25099 +
25100 +static inline
25101 +long do_vcmd(uint32_t cmd, uint32_t id,
25102 +       struct vx_info *vxi, struct nx_info *nxi,
25103 +       void __user *data, int compat)
25104 +{
25105 +       switch (cmd) {
25106 +
25107 +       case VCMD_get_version:
25108 +               return vc_get_version(id);
25109 +       case VCMD_get_vci:
25110 +               return vc_get_vci(id);
25111 +
25112 +       case VCMD_task_xid:
25113 +               return vc_task_xid(id);
25114 +       case VCMD_vx_info:
25115 +               return vc_vx_info(vxi, data);
25116 +
25117 +       case VCMD_task_nid:
25118 +               return vc_task_nid(id);
25119 +       case VCMD_nx_info:
25120 +               return vc_nx_info(nxi, data);
25121 +
25122 +       case VCMD_task_tag:
25123 +               return vc_task_tag(id);
25124 +
25125 +       case VCMD_set_space_v1:
25126 +               return vc_set_space_v1(vxi, data);
25127 +       /* this is version 2 */
25128 +       case VCMD_set_space:
25129 +               return vc_set_space(vxi, data);
25130 +
25131 +       case VCMD_get_space_mask_v0:
25132 +               return vc_get_space_mask(data, 0);
25133 +       /* this is version 1 */
25134 +       case VCMD_get_space_mask:
25135 +               return vc_get_space_mask(data, 1);
25136 +
25137 +       case VCMD_get_space_default:
25138 +               return vc_get_space_mask(data, -1);
25139 +
25140 +       case VCMD_set_umask:
25141 +               return vc_set_umask(vxi, data);
25142 +
25143 +       case VCMD_get_umask:
25144 +               return vc_get_umask(vxi, data);
25145 +
25146 +#ifdef CONFIG_IA32_EMULATION
25147 +       case VCMD_get_rlimit:
25148 +               return __COMPAT(vc_get_rlimit, vxi, data, compat);
25149 +       case VCMD_set_rlimit:
25150 +               return __COMPAT(vc_set_rlimit, vxi, data, compat);
25151 +#else
25152 +       case VCMD_get_rlimit:
25153 +               return vc_get_rlimit(vxi, data);
25154 +       case VCMD_set_rlimit:
25155 +               return vc_set_rlimit(vxi, data);
25156 +#endif
25157 +       case VCMD_get_rlimit_mask:
25158 +               return vc_get_rlimit_mask(id, data);
25159 +       case VCMD_reset_hits:
25160 +               return vc_reset_hits(vxi, data);
25161 +       case VCMD_reset_minmax:
25162 +               return vc_reset_minmax(vxi, data);
25163 +
25164 +       case VCMD_get_vhi_name:
25165 +               return vc_get_vhi_name(vxi, data);
25166 +       case VCMD_set_vhi_name:
25167 +               return vc_set_vhi_name(vxi, data);
25168 +
25169 +       case VCMD_ctx_stat:
25170 +               return vc_ctx_stat(vxi, data);
25171 +       case VCMD_virt_stat:
25172 +               return vc_virt_stat(vxi, data);
25173 +       case VCMD_sock_stat:
25174 +               return vc_sock_stat(vxi, data);
25175 +       case VCMD_rlimit_stat:
25176 +               return vc_rlimit_stat(vxi, data);
25177 +
25178 +       case VCMD_set_cflags:
25179 +               return vc_set_cflags(vxi, data);
25180 +       case VCMD_get_cflags:
25181 +               return vc_get_cflags(vxi, data);
25182 +
25183 +       /* this is version 1 */
25184 +       case VCMD_set_ccaps:
25185 +               return vc_set_ccaps(vxi, data);
25186 +       /* this is version 1 */
25187 +       case VCMD_get_ccaps:
25188 +               return vc_get_ccaps(vxi, data);
25189 +       case VCMD_set_bcaps:
25190 +               return vc_set_bcaps(vxi, data);
25191 +       case VCMD_get_bcaps:
25192 +               return vc_get_bcaps(vxi, data);
25193 +
25194 +       case VCMD_set_badness:
25195 +               return vc_set_badness(vxi, data);
25196 +       case VCMD_get_badness:
25197 +               return vc_get_badness(vxi, data);
25198 +
25199 +       case VCMD_set_nflags:
25200 +               return vc_set_nflags(nxi, data);
25201 +       case VCMD_get_nflags:
25202 +               return vc_get_nflags(nxi, data);
25203 +
25204 +       case VCMD_set_ncaps:
25205 +               return vc_set_ncaps(nxi, data);
25206 +       case VCMD_get_ncaps:
25207 +               return vc_get_ncaps(nxi, data);
25208 +
25209 +       case VCMD_set_prio_bias:
25210 +               return vc_set_prio_bias(vxi, data);
25211 +       case VCMD_get_prio_bias:
25212 +               return vc_get_prio_bias(vxi, data);
25213 +       case VCMD_add_dlimit:
25214 +               return __COMPAT(vc_add_dlimit, id, data, compat);
25215 +       case VCMD_rem_dlimit:
25216 +               return __COMPAT(vc_rem_dlimit, id, data, compat);
25217 +       case VCMD_set_dlimit:
25218 +               return __COMPAT(vc_set_dlimit, id, data, compat);
25219 +       case VCMD_get_dlimit:
25220 +               return __COMPAT(vc_get_dlimit, id, data, compat);
25221 +
25222 +       case VCMD_ctx_kill:
25223 +               return vc_ctx_kill(vxi, data);
25224 +
25225 +       case VCMD_wait_exit:
25226 +               return vc_wait_exit(vxi, data);
25227 +
25228 +       case VCMD_get_iattr:
25229 +               return __COMPAT_NO_ID(vc_get_iattr, data, compat);
25230 +       case VCMD_set_iattr:
25231 +               return __COMPAT_NO_ID(vc_set_iattr, data, compat);
25232 +
25233 +       case VCMD_fget_iattr:
25234 +               return vc_fget_iattr(id, data);
25235 +       case VCMD_fset_iattr:
25236 +               return vc_fset_iattr(id, data);
25237 +
25238 +       case VCMD_enter_space_v0:
25239 +               return vc_enter_space_v1(vxi, NULL);
25240 +       case VCMD_enter_space_v1:
25241 +               return vc_enter_space_v1(vxi, data);
25242 +       /* this is version 2 */
25243 +       case VCMD_enter_space:
25244 +               return vc_enter_space(vxi, data);
25245 +
25246 +       case VCMD_ctx_create_v0:
25247 +               return vc_ctx_create(id, NULL);
25248 +       case VCMD_ctx_create:
25249 +               return vc_ctx_create(id, data);
25250 +       case VCMD_ctx_migrate_v0:
25251 +               return vc_ctx_migrate(vxi, NULL);
25252 +       case VCMD_ctx_migrate:
25253 +               return vc_ctx_migrate(vxi, data);
25254 +
25255 +       case VCMD_net_create_v0:
25256 +               return vc_net_create(id, NULL);
25257 +       case VCMD_net_create:
25258 +               return vc_net_create(id, data);
25259 +       case VCMD_net_migrate:
25260 +               return vc_net_migrate(nxi, data);
25261 +
25262 +       case VCMD_tag_migrate:
25263 +               return vc_tag_migrate(id);
25264 +
25265 +       case VCMD_net_add:
25266 +               return vc_net_add(nxi, data);
25267 +       case VCMD_net_remove:
25268 +               return vc_net_remove(nxi, data);
25269 +
25270 +       case VCMD_net_add_ipv4_v1:
25271 +               return vc_net_add_ipv4_v1(nxi, data);
25272 +       /* this is version 2 */
25273 +       case VCMD_net_add_ipv4:
25274 +               return vc_net_add_ipv4(nxi, data);
25275 +
25276 +       case VCMD_net_rem_ipv4_v1:
25277 +               return vc_net_rem_ipv4_v1(nxi, data);
25278 +       /* this is version 2 */
25279 +       case VCMD_net_rem_ipv4:
25280 +               return vc_net_rem_ipv4(nxi, data);
25281 +#ifdef CONFIG_IPV6
25282 +       case VCMD_net_add_ipv6:
25283 +               return vc_net_add_ipv6(nxi, data);
25284 +       case VCMD_net_remove_ipv6:
25285 +               return vc_net_remove_ipv6(nxi, data);
25286 +#endif
25287 +/*     case VCMD_add_match_ipv4:
25288 +               return vc_add_match_ipv4(nxi, data);
25289 +       case VCMD_get_match_ipv4:
25290 +               return vc_get_match_ipv4(nxi, data);
25291 +#ifdef CONFIG_IPV6
25292 +       case VCMD_add_match_ipv6:
25293 +               return vc_add_match_ipv6(nxi, data);
25294 +       case VCMD_get_match_ipv6:
25295 +               return vc_get_match_ipv6(nxi, data);
25296 +#endif */
25297 +
25298 +#ifdef CONFIG_VSERVER_DEVICE
25299 +       case VCMD_set_mapping:
25300 +               return __COMPAT(vc_set_mapping, vxi, data, compat);
25301 +       case VCMD_unset_mapping:
25302 +               return __COMPAT(vc_unset_mapping, vxi, data, compat);
25303 +#endif
25304 +#ifdef CONFIG_VSERVER_HISTORY
25305 +       case VCMD_dump_history:
25306 +               return vc_dump_history(id);
25307 +       case VCMD_read_history:
25308 +               return __COMPAT(vc_read_history, id, data, compat);
25309 +#endif
25310 +       default:
25311 +               vxwprintk_task(1, "unimplemented VCMD_%02d_%d[%d]",
25312 +                       VC_CATEGORY(cmd), VC_COMMAND(cmd), VC_VERSION(cmd));
25313 +       }
25314 +       return -ENOSYS;
25315 +}
25316 +
25317 +
25318 +#define        __VCMD(vcmd, _perm, _args, _flags)              \
25319 +       case VCMD_ ## vcmd: perm = _perm;               \
25320 +               args = _args; flags = _flags; break
25321 +
25322 +
25323 +#define VCA_NONE       0x00
25324 +#define VCA_VXI                0x01
25325 +#define VCA_NXI                0x02
25326 +
25327 +#define VCF_NONE       0x00
25328 +#define VCF_INFO       0x01
25329 +#define VCF_ADMIN      0x02
25330 +#define VCF_ARES       0x06    /* includes admin */
25331 +#define VCF_SETUP      0x08
25332 +
25333 +#define VCF_ZIDOK      0x10    /* zero id okay */
25334 +
25335 +
25336 +static inline
25337 +long do_vserver(uint32_t cmd, uint32_t id, void __user *data, int compat)
25338 +{
25339 +       long ret;
25340 +       int permit = -1, state = 0;
25341 +       int perm = -1, args = 0, flags = 0;
25342 +       struct vx_info *vxi = NULL;
25343 +       struct nx_info *nxi = NULL;
25344 +
25345 +       switch (cmd) {
25346 +       /* unpriviledged commands */
25347 +       __VCMD(get_version,      0, VCA_NONE,   0);
25348 +       __VCMD(get_vci,          0, VCA_NONE,   0);
25349 +       __VCMD(get_rlimit_mask,  0, VCA_NONE,   0);
25350 +       __VCMD(get_space_mask_v0,0, VCA_NONE,   0);
25351 +       __VCMD(get_space_mask,   0, VCA_NONE,   0);
25352 +       __VCMD(get_space_default,0, VCA_NONE,   0);
25353 +
25354 +       /* info commands */
25355 +       __VCMD(task_xid,         2, VCA_NONE,   0);
25356 +       __VCMD(reset_hits,       2, VCA_VXI,    0);
25357 +       __VCMD(reset_minmax,     2, VCA_VXI,    0);
25358 +       __VCMD(vx_info,          3, VCA_VXI,    VCF_INFO);
25359 +       __VCMD(get_bcaps,        3, VCA_VXI,    VCF_INFO);
25360 +       __VCMD(get_ccaps,        3, VCA_VXI,    VCF_INFO);
25361 +       __VCMD(get_cflags,       3, VCA_VXI,    VCF_INFO);
25362 +       __VCMD(get_umask,        3, VCA_VXI,    VCF_INFO);
25363 +       __VCMD(get_badness,      3, VCA_VXI,    VCF_INFO);
25364 +       __VCMD(get_vhi_name,     3, VCA_VXI,    VCF_INFO);
25365 +       __VCMD(get_rlimit,       3, VCA_VXI,    VCF_INFO);
25366 +
25367 +       __VCMD(ctx_stat,         3, VCA_VXI,    VCF_INFO);
25368 +       __VCMD(virt_stat,        3, VCA_VXI,    VCF_INFO);
25369 +       __VCMD(sock_stat,        3, VCA_VXI,    VCF_INFO);
25370 +       __VCMD(rlimit_stat,      3, VCA_VXI,    VCF_INFO);
25371 +
25372 +       __VCMD(task_nid,         2, VCA_NONE,   0);
25373 +       __VCMD(nx_info,          3, VCA_NXI,    VCF_INFO);
25374 +       __VCMD(get_ncaps,        3, VCA_NXI,    VCF_INFO);
25375 +       __VCMD(get_nflags,       3, VCA_NXI,    VCF_INFO);
25376 +
25377 +       __VCMD(task_tag,         2, VCA_NONE,   0);
25378 +
25379 +       __VCMD(get_iattr,        2, VCA_NONE,   0);
25380 +       __VCMD(fget_iattr,       2, VCA_NONE,   0);
25381 +       __VCMD(get_dlimit,       3, VCA_NONE,   VCF_INFO);
25382 +       __VCMD(get_prio_bias,    3, VCA_VXI,    VCF_INFO);
25383 +
25384 +       /* lower admin commands */
25385 +       __VCMD(wait_exit,        4, VCA_VXI,    VCF_INFO);
25386 +       __VCMD(ctx_create_v0,    5, VCA_NONE,   0);
25387 +       __VCMD(ctx_create,       5, VCA_NONE,   0);
25388 +       __VCMD(ctx_migrate_v0,   5, VCA_VXI,    VCF_ADMIN);
25389 +       __VCMD(ctx_migrate,      5, VCA_VXI,    VCF_ADMIN);
25390 +       __VCMD(enter_space_v0,   5, VCA_VXI,    VCF_ADMIN);
25391 +       __VCMD(enter_space_v1,   5, VCA_VXI,    VCF_ADMIN);
25392 +       __VCMD(enter_space,      5, VCA_VXI,    VCF_ADMIN);
25393 +
25394 +       __VCMD(net_create_v0,    5, VCA_NONE,   0);
25395 +       __VCMD(net_create,       5, VCA_NONE,   0);
25396 +       __VCMD(net_migrate,      5, VCA_NXI,    VCF_ADMIN);
25397 +
25398 +       __VCMD(tag_migrate,      5, VCA_NONE,   VCF_ADMIN);
25399 +
25400 +       /* higher admin commands */
25401 +       __VCMD(ctx_kill,         6, VCA_VXI,    VCF_ARES);
25402 +       __VCMD(set_space_v1,     7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25403 +       __VCMD(set_space,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25404 +
25405 +       __VCMD(set_ccaps,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25406 +       __VCMD(set_bcaps,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25407 +       __VCMD(set_cflags,       7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25408 +       __VCMD(set_umask,        7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25409 +       __VCMD(set_badness,      7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25410 +
25411 +       __VCMD(set_vhi_name,     7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25412 +       __VCMD(set_rlimit,       7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25413 +       __VCMD(set_prio_bias,    7, VCA_VXI,    VCF_ARES | VCF_SETUP);
25414 +
25415 +       __VCMD(set_ncaps,        7, VCA_NXI,    VCF_ARES | VCF_SETUP);
25416 +       __VCMD(set_nflags,       7, VCA_NXI,    VCF_ARES | VCF_SETUP);
25417 +       __VCMD(net_add,          8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25418 +       __VCMD(net_remove,       8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25419 +       __VCMD(net_add_ipv4_v1,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25420 +       __VCMD(net_rem_ipv4_v1,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25421 +       __VCMD(net_add_ipv4,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25422 +       __VCMD(net_rem_ipv4,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25423 +#ifdef CONFIG_IPV6
25424 +       __VCMD(net_add_ipv6,     8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25425 +       __VCMD(net_remove_ipv6,  8, VCA_NXI,    VCF_ARES | VCF_SETUP);
25426 +#endif
25427 +       __VCMD(set_iattr,        7, VCA_NONE,   0);
25428 +       __VCMD(fset_iattr,       7, VCA_NONE,   0);
25429 +       __VCMD(set_dlimit,       7, VCA_NONE,   VCF_ARES);
25430 +       __VCMD(add_dlimit,       8, VCA_NONE,   VCF_ARES);
25431 +       __VCMD(rem_dlimit,       8, VCA_NONE,   VCF_ARES);
25432 +
25433 +#ifdef CONFIG_VSERVER_DEVICE
25434 +       __VCMD(set_mapping,      8, VCA_VXI,    VCF_ARES|VCF_ZIDOK);
25435 +       __VCMD(unset_mapping,    8, VCA_VXI,    VCF_ARES|VCF_ZIDOK);
25436 +#endif
25437 +       /* debug level admin commands */
25438 +#ifdef CONFIG_VSERVER_HISTORY
25439 +       __VCMD(dump_history,     9, VCA_NONE,   0);
25440 +       __VCMD(read_history,     9, VCA_NONE,   0);
25441 +#endif
25442 +
25443 +       default:
25444 +               perm = -1;
25445 +       }
25446 +
25447 +       vxdprintk(VXD_CBIT(switch, 0),
25448 +               "vc: VCMD_%02d_%d[%d], %d,%p [%d,%d,%x,%x]",
25449 +               VC_CATEGORY(cmd), VC_COMMAND(cmd),
25450 +               VC_VERSION(cmd), id, data, compat,
25451 +               perm, args, flags);
25452 +
25453 +       ret = -ENOSYS;
25454 +       if (perm < 0)
25455 +               goto out;
25456 +
25457 +       state = 1;
25458 +       if (!capable(CAP_CONTEXT))
25459 +               goto out;
25460 +
25461 +       state = 2;
25462 +       /* moved here from the individual commands */
25463 +       ret = -EPERM;
25464 +       if ((perm > 1) && !capable(CAP_SYS_ADMIN))
25465 +               goto out;
25466 +
25467 +       state = 3;
25468 +       /* vcmd involves resource management  */
25469 +       ret = -EPERM;
25470 +       if ((flags & VCF_ARES) && !capable(CAP_SYS_RESOURCE))
25471 +               goto out;
25472 +
25473 +       state = 4;
25474 +       /* various legacy exceptions */
25475 +       switch (cmd) {
25476 +       /* will go away when spectator is a cap */
25477 +       case VCMD_ctx_migrate_v0:
25478 +       case VCMD_ctx_migrate:
25479 +               if (id == 1) {
25480 +                       current->xid = 1;
25481 +                       ret = 1;
25482 +                       goto out;
25483 +               }
25484 +               break;
25485 +
25486 +       /* will go away when spectator is a cap */
25487 +       case VCMD_net_migrate:
25488 +               if (id == 1) {
25489 +                       current->nid = 1;
25490 +                       ret = 1;
25491 +                       goto out;
25492 +               }
25493 +               break;
25494 +       }
25495 +
25496 +       /* vcmds are fine by default */
25497 +       permit = 1;
25498 +
25499 +       /* admin type vcmds require admin ... */
25500 +       if (flags & VCF_ADMIN)
25501 +               permit = vx_check(0, VS_ADMIN) ? 1 : 0;
25502 +
25503 +       /* ... but setup type vcmds override that */
25504 +       if (!permit && (flags & VCF_SETUP))
25505 +               permit = vx_flags(VXF_STATE_SETUP, 0) ? 2 : 0;
25506 +
25507 +       state = 5;
25508 +       ret = -EPERM;
25509 +       if (!permit)
25510 +               goto out;
25511 +
25512 +       state = 6;
25513 +       if (!id && (flags & VCF_ZIDOK))
25514 +               goto skip_id;
25515 +
25516 +       ret = -ESRCH;
25517 +       if (args & VCA_VXI) {
25518 +               vxi = lookup_vx_info(id);
25519 +               if (!vxi)
25520 +                       goto out;
25521 +
25522 +               if ((flags & VCF_ADMIN) &&
25523 +                       /* special case kill for shutdown */
25524 +                       (cmd != VCMD_ctx_kill) &&
25525 +                       /* can context be administrated? */
25526 +                       !vx_info_flags(vxi, VXF_STATE_ADMIN, 0)) {
25527 +                       ret = -EACCES;
25528 +                       goto out_vxi;
25529 +               }
25530 +       }
25531 +       state = 7;
25532 +       if (args & VCA_NXI) {
25533 +               nxi = lookup_nx_info(id);
25534 +               if (!nxi)
25535 +                       goto out_vxi;
25536 +
25537 +               if ((flags & VCF_ADMIN) &&
25538 +                       /* can context be administrated? */
25539 +                       !nx_info_flags(nxi, NXF_STATE_ADMIN, 0)) {
25540 +                       ret = -EACCES;
25541 +                       goto out_nxi;
25542 +               }
25543 +       }
25544 +skip_id:
25545 +       state = 8;
25546 +       ret = do_vcmd(cmd, id, vxi, nxi, data, compat);
25547 +
25548 +out_nxi:
25549 +       if ((args & VCA_NXI) && nxi)
25550 +               put_nx_info(nxi);
25551 +out_vxi:
25552 +       if ((args & VCA_VXI) && vxi)
25553 +               put_vx_info(vxi);
25554 +out:
25555 +       vxdprintk(VXD_CBIT(switch, 1),
25556 +               "vc: VCMD_%02d_%d[%d] = %08lx(%ld) [%d,%d]",
25557 +               VC_CATEGORY(cmd), VC_COMMAND(cmd),
25558 +               VC_VERSION(cmd), ret, ret, state, permit);
25559 +       return ret;
25560 +}
25561 +
25562 +asmlinkage long
25563 +sys_vserver(uint32_t cmd, uint32_t id, void __user *data)
25564 +{
25565 +       return do_vserver(cmd, id, data, 0);
25566 +}
25567 +
25568 +#ifdef CONFIG_COMPAT
25569 +
25570 +asmlinkage long
25571 +sys32_vserver(uint32_t cmd, uint32_t id, void __user *data)
25572 +{
25573 +       return do_vserver(cmd, id, data, 1);
25574 +}
25575 +
25576 +#endif /* CONFIG_COMPAT */
25577 diff -NurpP --minimal linux-3.0/kernel/vserver/sysctl.c linux-3.0-vs2.3.1-pre7/kernel/vserver/sysctl.c
25578 --- linux-3.0/kernel/vserver/sysctl.c   1970-01-01 01:00:00.000000000 +0100
25579 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/sysctl.c      2011-06-10 22:11:24.000000000 +0200
25580 @@ -0,0 +1,241 @@
25581 +/*
25582 + *  kernel/vserver/sysctl.c
25583 + *
25584 + *  Virtual Context Support
25585 + *
25586 + *  Copyright (C) 2004-2007  Herbert Pötzl
25587 + *
25588 + *  V0.01  basic structure
25589 + *
25590 + */
25591 +
25592 +#include <linux/module.h>
25593 +#include <linux/ctype.h>
25594 +#include <linux/sysctl.h>
25595 +#include <linux/parser.h>
25596 +#include <asm/uaccess.h>
25597 +
25598 +enum {
25599 +       CTL_DEBUG_ERROR         = 0,
25600 +       CTL_DEBUG_SWITCH        = 1,
25601 +       CTL_DEBUG_XID,
25602 +       CTL_DEBUG_NID,
25603 +       CTL_DEBUG_TAG,
25604 +       CTL_DEBUG_NET,
25605 +       CTL_DEBUG_LIMIT,
25606 +       CTL_DEBUG_CRES,
25607 +       CTL_DEBUG_DLIM,
25608 +       CTL_DEBUG_QUOTA,
25609 +       CTL_DEBUG_CVIRT,
25610 +       CTL_DEBUG_SPACE,
25611 +       CTL_DEBUG_MISC,
25612 +};
25613 +
25614 +
25615 +unsigned int vx_debug_switch   = 0;
25616 +unsigned int vx_debug_xid      = 0;
25617 +unsigned int vx_debug_nid      = 0;
25618 +unsigned int vx_debug_tag      = 0;
25619 +unsigned int vx_debug_net      = 0;
25620 +unsigned int vx_debug_limit    = 0;
25621 +unsigned int vx_debug_cres     = 0;
25622 +unsigned int vx_debug_dlim     = 0;
25623 +unsigned int vx_debug_quota    = 0;
25624 +unsigned int vx_debug_cvirt    = 0;
25625 +unsigned int vx_debug_space    = 0;
25626 +unsigned int vx_debug_misc     = 0;
25627 +
25628 +
25629 +static struct ctl_table_header *vserver_table_header;
25630 +static ctl_table vserver_root_table[];
25631 +
25632 +
25633 +void vserver_register_sysctl(void)
25634 +{
25635 +       if (!vserver_table_header) {
25636 +               vserver_table_header = register_sysctl_table(vserver_root_table);
25637 +       }
25638 +
25639 +}
25640 +
25641 +void vserver_unregister_sysctl(void)
25642 +{
25643 +       if (vserver_table_header) {
25644 +               unregister_sysctl_table(vserver_table_header);
25645 +               vserver_table_header = NULL;
25646 +       }
25647 +}
25648 +
25649 +
25650 +static int proc_dodebug(ctl_table *table, int write,
25651 +       void __user *buffer, size_t *lenp, loff_t *ppos)
25652 +{
25653 +       char            tmpbuf[20], *p, c;
25654 +       unsigned int    value;
25655 +       size_t          left, len;
25656 +
25657 +       if ((*ppos && !write) || !*lenp) {
25658 +               *lenp = 0;
25659 +               return 0;
25660 +       }
25661 +
25662 +       left = *lenp;
25663 +
25664 +       if (write) {
25665 +               if (!access_ok(VERIFY_READ, buffer, left))
25666 +                       return -EFAULT;
25667 +               p = (char *)buffer;
25668 +               while (left && __get_user(c, p) >= 0 && isspace(c))
25669 +                       left--, p++;
25670 +               if (!left)
25671 +                       goto done;
25672 +
25673 +               if (left > sizeof(tmpbuf) - 1)
25674 +                       return -EINVAL;
25675 +               if (copy_from_user(tmpbuf, p, left))
25676 +                       return -EFAULT;
25677 +               tmpbuf[left] = '\0';
25678 +
25679 +               for (p = tmpbuf, value = 0; '0' <= *p && *p <= '9'; p++, left--)
25680 +                       value = 10 * value + (*p - '0');
25681 +               if (*p && !isspace(*p))
25682 +                       return -EINVAL;
25683 +               while (left && isspace(*p))
25684 +                       left--, p++;
25685 +               *(unsigned int *)table->data = value;
25686 +       } else {
25687 +               if (!access_ok(VERIFY_WRITE, buffer, left))
25688 +                       return -EFAULT;
25689 +               len = sprintf(tmpbuf, "%d", *(unsigned int *)table->data);
25690 +               if (len > left)
25691 +                       len = left;
25692 +               if (__copy_to_user(buffer, tmpbuf, len))
25693 +                       return -EFAULT;
25694 +               if ((left -= len) > 0) {
25695 +                       if (put_user('\n', (char *)buffer + len))
25696 +                               return -EFAULT;
25697 +                       left--;
25698 +               }
25699 +       }
25700 +
25701 +done:
25702 +       *lenp -= left;
25703 +       *ppos += *lenp;
25704 +       return 0;
25705 +}
25706 +
25707 +static int zero;
25708 +
25709 +#define        CTL_ENTRY(ctl, name)                            \
25710 +       {                                               \
25711 +               .procname       = #name,                \
25712 +               .data           = &vx_ ## name,         \
25713 +               .maxlen         = sizeof(int),          \
25714 +               .mode           = 0644,                 \
25715 +               .proc_handler   = &proc_dodebug,        \
25716 +               .extra1         = &zero,                \
25717 +               .extra2         = &zero,                \
25718 +       }
25719 +
25720 +static ctl_table vserver_debug_table[] = {
25721 +       CTL_ENTRY(CTL_DEBUG_SWITCH,     debug_switch),
25722 +       CTL_ENTRY(CTL_DEBUG_XID,        debug_xid),
25723 +       CTL_ENTRY(CTL_DEBUG_NID,        debug_nid),
25724 +       CTL_ENTRY(CTL_DEBUG_TAG,        debug_tag),
25725 +       CTL_ENTRY(CTL_DEBUG_NET,        debug_net),
25726 +       CTL_ENTRY(CTL_DEBUG_LIMIT,      debug_limit),
25727 +       CTL_ENTRY(CTL_DEBUG_CRES,       debug_cres),
25728 +       CTL_ENTRY(CTL_DEBUG_DLIM,       debug_dlim),
25729 +       CTL_ENTRY(CTL_DEBUG_QUOTA,      debug_quota),
25730 +       CTL_ENTRY(CTL_DEBUG_CVIRT,      debug_cvirt),
25731 +       CTL_ENTRY(CTL_DEBUG_SPACE,      debug_space),
25732 +       CTL_ENTRY(CTL_DEBUG_MISC,       debug_misc),
25733 +       { 0 }
25734 +};
25735 +
25736 +static ctl_table vserver_root_table[] = {
25737 +       {
25738 +               .procname       = "vserver",
25739 +               .mode           = 0555,
25740 +               .child          = vserver_debug_table
25741 +       },
25742 +       { 0 }
25743 +};
25744 +
25745 +
25746 +static match_table_t tokens = {
25747 +       { CTL_DEBUG_SWITCH,     "switch=%x"     },
25748 +       { CTL_DEBUG_XID,        "xid=%x"        },
25749 +       { CTL_DEBUG_NID,        "nid=%x"        },
25750 +       { CTL_DEBUG_TAG,        "tag=%x"        },
25751 +       { CTL_DEBUG_NET,        "net=%x"        },
25752 +       { CTL_DEBUG_LIMIT,      "limit=%x"      },
25753 +       { CTL_DEBUG_CRES,       "cres=%x"       },
25754 +       { CTL_DEBUG_DLIM,       "dlim=%x"       },
25755 +       { CTL_DEBUG_QUOTA,      "quota=%x"      },
25756 +       { CTL_DEBUG_CVIRT,      "cvirt=%x"      },
25757 +       { CTL_DEBUG_SPACE,      "space=%x"      },
25758 +       { CTL_DEBUG_MISC,       "misc=%x"       },
25759 +       { CTL_DEBUG_ERROR,      NULL            }
25760 +};
25761 +
25762 +#define        HANDLE_CASE(id, name, val)                              \
25763 +       case CTL_DEBUG_ ## id:                                  \
25764 +               vx_debug_ ## name = val;                        \
25765 +               printk("vs_debug_" #name "=0x%x\n", val);       \
25766 +               break
25767 +
25768 +
25769 +static int __init vs_debug_setup(char *str)
25770 +{
25771 +       char *p;
25772 +       int token;
25773 +
25774 +       printk("vs_debug_setup(%s)\n", str);
25775 +       while ((p = strsep(&str, ",")) != NULL) {
25776 +               substring_t args[MAX_OPT_ARGS];
25777 +               unsigned int value;
25778 +
25779 +               if (!*p)
25780 +                       continue;
25781 +
25782 +               token = match_token(p, tokens, args);
25783 +               value = (token > 0) ? simple_strtoul(args[0].from, NULL, 0) : 0;
25784 +
25785 +               switch (token) {
25786 +               HANDLE_CASE(SWITCH, switch, value);
25787 +               HANDLE_CASE(XID,    xid,    value);
25788 +               HANDLE_CASE(NID,    nid,    value);
25789 +               HANDLE_CASE(TAG,    tag,    value);
25790 +               HANDLE_CASE(NET,    net,    value);
25791 +               HANDLE_CASE(LIMIT,  limit,  value);
25792 +               HANDLE_CASE(CRES,   cres,   value);
25793 +               HANDLE_CASE(DLIM,   dlim,   value);
25794 +               HANDLE_CASE(QUOTA,  quota,  value);
25795 +               HANDLE_CASE(CVIRT,  cvirt,  value);
25796 +               HANDLE_CASE(SPACE,  space,  value);
25797 +               HANDLE_CASE(MISC,   misc,   value);
25798 +               default:
25799 +                       return -EINVAL;
25800 +                       break;
25801 +               }
25802 +       }
25803 +       return 1;
25804 +}
25805 +
25806 +__setup("vsdebug=", vs_debug_setup);
25807 +
25808 +
25809 +
25810 +EXPORT_SYMBOL_GPL(vx_debug_switch);
25811 +EXPORT_SYMBOL_GPL(vx_debug_xid);
25812 +EXPORT_SYMBOL_GPL(vx_debug_nid);
25813 +EXPORT_SYMBOL_GPL(vx_debug_net);
25814 +EXPORT_SYMBOL_GPL(vx_debug_limit);
25815 +EXPORT_SYMBOL_GPL(vx_debug_cres);
25816 +EXPORT_SYMBOL_GPL(vx_debug_dlim);
25817 +EXPORT_SYMBOL_GPL(vx_debug_quota);
25818 +EXPORT_SYMBOL_GPL(vx_debug_cvirt);
25819 +EXPORT_SYMBOL_GPL(vx_debug_space);
25820 +EXPORT_SYMBOL_GPL(vx_debug_misc);
25821 +
25822 diff -NurpP --minimal linux-3.0/kernel/vserver/tag.c linux-3.0-vs2.3.1-pre7/kernel/vserver/tag.c
25823 --- linux-3.0/kernel/vserver/tag.c      1970-01-01 01:00:00.000000000 +0100
25824 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/tag.c 2011-06-10 22:11:24.000000000 +0200
25825 @@ -0,0 +1,63 @@
25826 +/*
25827 + *  linux/kernel/vserver/tag.c
25828 + *
25829 + *  Virtual Server: Shallow Tag Space
25830 + *
25831 + *  Copyright (C) 2007  Herbert Pötzl
25832 + *
25833 + *  V0.01  basic implementation
25834 + *
25835 + */
25836 +
25837 +#include <linux/sched.h>
25838 +#include <linux/vserver/debug.h>
25839 +#include <linux/vs_pid.h>
25840 +#include <linux/vs_tag.h>
25841 +
25842 +#include <linux/vserver/tag_cmd.h>
25843 +
25844 +
25845 +int dx_migrate_task(struct task_struct *p, tag_t tag)
25846 +{
25847 +       if (!p)
25848 +               BUG();
25849 +
25850 +       vxdprintk(VXD_CBIT(tag, 5),
25851 +               "dx_migrate_task(%p[#%d],#%d)", p, p->tag, tag);
25852 +
25853 +       task_lock(p);
25854 +       p->tag = tag;
25855 +       task_unlock(p);
25856 +
25857 +       vxdprintk(VXD_CBIT(tag, 5),
25858 +               "moved task %p into [#%d]", p, tag);
25859 +       return 0;
25860 +}
25861 +
25862 +/* vserver syscall commands below here */
25863 +
25864 +/* taks xid and vx_info functions */
25865 +
25866 +
25867 +int vc_task_tag(uint32_t id)
25868 +{
25869 +       tag_t tag;
25870 +
25871 +       if (id) {
25872 +               struct task_struct *tsk;
25873 +               rcu_read_lock();
25874 +               tsk = find_task_by_real_pid(id);
25875 +               tag = (tsk) ? tsk->tag : -ESRCH;
25876 +               rcu_read_unlock();
25877 +       } else
25878 +               tag = dx_current_tag();
25879 +       return tag;
25880 +}
25881 +
25882 +
25883 +int vc_tag_migrate(uint32_t tag)
25884 +{
25885 +       return dx_migrate_task(current, tag & 0xFFFF);
25886 +}
25887 +
25888 +
25889 diff -NurpP --minimal linux-3.0/kernel/vserver/vci_config.h linux-3.0-vs2.3.1-pre7/kernel/vserver/vci_config.h
25890 --- linux-3.0/kernel/vserver/vci_config.h       1970-01-01 01:00:00.000000000 +0100
25891 +++ linux-3.0-vs2.3.1-pre7/kernel/vserver/vci_config.h  2011-06-10 22:11:24.000000000 +0200
25892 @@ -0,0 +1,76 @@
25893 +
25894 +/*  interface version */
25895 +
25896 +#define VCI_VERSION            0x00020308
25897 +
25898 +
25899 +enum {
25900 +       VCI_KCBIT_NO_DYNAMIC = 0,
25901 +
25902 +       VCI_KCBIT_PROC_SECURE = 4,
25903 +       /* VCI_KCBIT_HARDCPU = 5, */
25904 +       /* VCI_KCBIT_IDLELIMIT = 6, */
25905 +       /* VCI_KCBIT_IDLETIME = 7, */
25906 +
25907 +       VCI_KCBIT_COWBL = 8,
25908 +       VCI_KCBIT_FULLCOWBL = 9,
25909 +       VCI_KCBIT_SPACES = 10,
25910 +       VCI_KCBIT_NETV2 = 11,
25911 +       VCI_KCBIT_MEMCG = 12,
25912 +
25913 +       VCI_KCBIT_DEBUG = 16,
25914 +       VCI_KCBIT_HISTORY = 20,
25915 +       VCI_KCBIT_TAGGED = 24,
25916 +       VCI_KCBIT_PPTAG = 28,
25917 +
25918 +       VCI_KCBIT_MORE = 31,
25919 +};
25920 +
25921 +
25922 +static inline uint32_t vci_kernel_config(void)
25923 +{
25924 +       return
25925 +       (1 << VCI_KCBIT_NO_DYNAMIC) |
25926 +
25927 +       /* configured features */
25928 +#ifdef CONFIG_VSERVER_PROC_SECURE
25929 +       (1 << VCI_KCBIT_PROC_SECURE) |
25930 +#endif
25931 +#ifdef CONFIG_VSERVER_COWBL
25932 +       (1 << VCI_KCBIT_COWBL) |
25933 +       (1 << VCI_KCBIT_FULLCOWBL) |
25934 +#endif
25935 +       (1 << VCI_KCBIT_SPACES) |
25936 +       (1 << VCI_KCBIT_NETV2) |
25937 +#ifdef CONFIG_CGROUP_MEM_RES_CTLR
25938 +       (1 << VCI_KCBIT_MEMCG) |
25939 +#endif
25940 +
25941 +       /* debug options */
25942 +#ifdef CONFIG_VSERVER_DEBUG
25943 +       (1 << VCI_KCBIT_DEBUG) |
25944 +#endif
25945 +#ifdef CONFIG_VSERVER_HISTORY
25946 +       (1 << VCI_KCBIT_HISTORY) |
25947 +#endif
25948 +
25949 +       /* inode context tagging */
25950 +#if    defined(CONFIG_TAGGING_NONE)
25951 +       (0 << VCI_KCBIT_TAGGED) |
25952 +#elif  defined(CONFIG_TAGGING_UID16)
25953 +       (1 << VCI_KCBIT_TAGGED) |
25954 +#elif  defined(CONFIG_TAGGING_GID16)
25955 +       (2 << VCI_KCBIT_TAGGED) |
25956 +#elif  defined(CONFIG_TAGGING_ID24)
25957 +       (3 << VCI_KCBIT_TAGGED) |
25958 +#elif  defined(CONFIG_TAGGING_INTERN)
25959 +       (4 << VCI_KCBIT_TAGGED) |
25960 +#elif  defined(CONFIG_TAGGING_RUNTIME)
25961 +       (5 << VCI_KCBIT_TAGGED) |
25962 +#else
25963 +       (7 << VCI_KCBIT_TAGGED) |
25964 +#endif
25965 +       (1 << VCI_KCBIT_PPTAG) |
25966 +       0;
25967 +}
25968 +
25969 diff -NurpP --minimal linux-3.0/mm/filemap_xip.c linux-3.0-vs2.3.1-pre7/mm/filemap_xip.c
25970 --- linux-3.0/mm/filemap_xip.c  2011-07-22 11:18:12.000000000 +0200
25971 +++ linux-3.0-vs2.3.1-pre7/mm/filemap_xip.c     2011-06-10 22:11:24.000000000 +0200
25972 @@ -18,6 +18,7 @@
25973  #include <linux/seqlock.h>
25974  #include <linux/mutex.h>
25975  #include <linux/gfp.h>
25976 +#include <linux/vs_memory.h>
25977  #include <asm/tlbflush.h>
25978  #include <asm/io.h>
25979
25980 diff -NurpP --minimal linux-3.0/mm/fremap.c linux-3.0-vs2.3.1-pre7/mm/fremap.c
25981 --- linux-3.0/mm/fremap.c       2011-07-22 11:18:12.000000000 +0200
25982 +++ linux-3.0-vs2.3.1-pre7/mm/fremap.c  2011-06-10 22:11:24.000000000 +0200
25983 @@ -16,6 +16,7 @@
25984  #include <linux/module.h>
25985  #include <linux/syscalls.h>
25986  #include <linux/mmu_notifier.h>
25987 +#include <linux/vs_memory.h>
25988
25989  #include <asm/mmu_context.h>
25990  #include <asm/cacheflush.h>
25991 diff -NurpP --minimal linux-3.0/mm/hugetlb.c linux-3.0-vs2.3.1-pre7/mm/hugetlb.c
25992 --- linux-3.0/mm/hugetlb.c      2011-07-22 11:18:12.000000000 +0200
25993 +++ linux-3.0-vs2.3.1-pre7/mm/hugetlb.c 2011-06-22 12:39:16.000000000 +0200
25994 @@ -28,6 +28,7 @@
25995
25996  #include <linux/hugetlb.h>
25997  #include <linux/node.h>
25998 +#include <linux/vs_memory.h>
25999  #include "internal.h"
26000
26001  const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
26002 diff -NurpP --minimal linux-3.0/mm/memcontrol.c linux-3.0-vs2.3.1-pre7/mm/memcontrol.c
26003 --- linux-3.0/mm/memcontrol.c   2011-07-22 11:18:12.000000000 +0200
26004 +++ linux-3.0-vs2.3.1-pre7/mm/memcontrol.c      2011-07-19 00:44:39.000000000 +0200
26005 @@ -741,6 +741,31 @@ struct mem_cgroup *mem_cgroup_from_task(
26006                                 struct mem_cgroup, css);
26007  }
26008
26009 +u64 mem_cgroup_res_read_u64(struct mem_cgroup *mem, int member)
26010 +{
26011 +       return res_counter_read_u64(&mem->res, member);
26012 +}
26013 +
26014 +u64 mem_cgroup_memsw_read_u64(struct mem_cgroup *mem, int member)
26015 +{
26016 +       return res_counter_read_u64(&mem->memsw, member);
26017 +}
26018 +
26019 +s64 mem_cgroup_stat_read_cache(struct mem_cgroup *mem)
26020 +{
26021 +       return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_CACHE);
26022 +}
26023 +
26024 +s64 mem_cgroup_stat_read_anon(struct mem_cgroup *mem)
26025 +{
26026 +       return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_RSS);
26027 +}
26028 +
26029 +s64 mem_cgroup_stat_read_mapped(struct mem_cgroup *mem)
26030 +{
26031 +       return mem_cgroup_read_stat(mem, MEM_CGROUP_STAT_FILE_MAPPED);
26032 +}
26033 +
26034  struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm)
26035  {
26036         struct mem_cgroup *mem = NULL;
26037 diff -NurpP --minimal linux-3.0/mm/memory.c linux-3.0-vs2.3.1-pre7/mm/memory.c
26038 --- linux-3.0/mm/memory.c       2011-07-22 11:18:12.000000000 +0200
26039 +++ linux-3.0-vs2.3.1-pre7/mm/memory.c  2011-07-19 00:44:39.000000000 +0200
26040 @@ -3332,6 +3332,7 @@ int handle_pte_fault(struct mm_struct *m
26041  {
26042         pte_t entry;
26043         spinlock_t *ptl;
26044 +       int ret = 0, type = VXPT_UNKNOWN;
26045
26046         entry = *pte;
26047         if (!pte_present(entry)) {
26048 @@ -3356,9 +3357,12 @@ int handle_pte_fault(struct mm_struct *m
26049         if (unlikely(!pte_same(*pte, entry)))
26050                 goto unlock;
26051         if (flags & FAULT_FLAG_WRITE) {
26052 -               if (!pte_write(entry))
26053 -                       return do_wp_page(mm, vma, address,
26054 +               if (!pte_write(entry)) {
26055 +                       ret = do_wp_page(mm, vma, address,
26056                                         pte, pmd, ptl, entry);
26057 +                       type = VXPT_WRITE;
26058 +                       goto out;
26059 +               }
26060                 entry = pte_mkdirty(entry);
26061         }
26062         entry = pte_mkyoung(entry);
26063 @@ -3376,7 +3380,10 @@ int handle_pte_fault(struct mm_struct *m
26064         }
26065  unlock:
26066         pte_unmap_unlock(pte, ptl);
26067 -       return 0;
26068 +       ret = 0;
26069 +out:
26070 +       vx_page_fault(mm, vma, type, ret);
26071 +       return ret;
26072  }
26073
26074  /*
26075 diff -NurpP --minimal linux-3.0/mm/mremap.c linux-3.0-vs2.3.1-pre7/mm/mremap.c
26076 --- linux-3.0/mm/mremap.c       2011-07-22 11:18:12.000000000 +0200
26077 +++ linux-3.0-vs2.3.1-pre7/mm/mremap.c  2011-06-10 22:11:24.000000000 +0200
26078 @@ -19,6 +19,7 @@
26079  #include <linux/security.h>
26080  #include <linux/syscalls.h>
26081  #include <linux/mmu_notifier.h>
26082 +#include <linux/vs_memory.h>
26083
26084  #include <asm/uaccess.h>
26085  #include <asm/cacheflush.h>
26086 diff -NurpP --minimal linux-3.0/mm/oom_kill.c linux-3.0-vs2.3.1-pre7/mm/oom_kill.c
26087 --- linux-3.0/mm/oom_kill.c     2011-07-22 11:18:12.000000000 +0200
26088 +++ linux-3.0-vs2.3.1-pre7/mm/oom_kill.c        2011-06-10 22:11:24.000000000 +0200
26089 @@ -32,6 +32,9 @@
26090  #include <linux/mempolicy.h>
26091  #include <linux/security.h>
26092  #include <linux/ptrace.h>
26093 +#include <linux/reboot.h>
26094 +#include <linux/vs_memory.h>
26095 +#include <linux/vs_context.h>
26096
26097  int sysctl_panic_on_oom;
26098  int sysctl_oom_kill_allocating_task;
26099 @@ -134,11 +137,18 @@ struct task_struct *find_lock_task_mm(st
26100  static bool oom_unkillable_task(struct task_struct *p,
26101                 const struct mem_cgroup *mem, const nodemask_t *nodemask)
26102  {
26103 -       if (is_global_init(p))
26104 +       unsigned xid = vx_current_xid();
26105 +
26106 +       /* skip the init task, global and per guest */
26107 +       if (task_is_init(p))
26108                 return true;
26109         if (p->flags & PF_KTHREAD)
26110                 return true;
26111
26112 +       /* skip other guest and host processes if oom in guest */
26113 +       if (xid && vx_task_xid(p) != xid)
26114 +               return true;
26115 +
26116         /* When mem_cgroup_out_of_memory() and p is not member of the group */
26117         if (mem && !task_in_mem_cgroup(p, mem))
26118                 return true;
26119 @@ -214,6 +224,18 @@ unsigned int oom_badness(struct task_str
26120         points += p->signal->oom_score_adj;
26121
26122         /*
26123 +        * add points for context badness and
26124 +        * reduce badness for processes belonging to
26125 +        * a different context
26126 +        */
26127 +
26128 +       points += vx_badness(p, p->mm);
26129 +
26130 +       if ((vx_current_xid() > 1) &&
26131 +               vx_current_xid() != vx_task_xid(p))
26132 +               points /= 16;
26133 +
26134 +       /*
26135          * Never return 0 for an eligible task that may be killed since it's
26136          * possible that no single user task uses more than 0.1% of memory and
26137          * no single admin tasks uses more than 3.0%.
26138 @@ -427,8 +449,8 @@ static int oom_kill_task(struct task_str
26139         /* mm cannot be safely dereferenced after task_unlock(p) */
26140         mm = p->mm;
26141
26142 -       pr_err("Killed process %d (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB\n",
26143 -               task_pid_nr(p), p->comm, K(p->mm->total_vm),
26144 +       pr_err("Killed process %d:#%u (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB\n",
26145 +               task_pid_nr(p), p->xid, p->comm, K(p->mm->total_vm),
26146                 K(get_mm_counter(p->mm, MM_ANONPAGES)),
26147                 K(get_mm_counter(p->mm, MM_FILEPAGES)));
26148         task_unlock(p);
26149 @@ -482,8 +504,8 @@ static int oom_kill_process(struct task_
26150         }
26151
26152         task_lock(p);
26153 -       pr_err("%s: Kill process %d (%s) score %d or sacrifice child\n",
26154 -               message, task_pid_nr(p), p->comm, points);
26155 +       pr_err("%s: Kill process %d:#%u (%s) score %d or sacrifice child\n",
26156 +               message, task_pid_nr(p), p->xid, p->comm, points);
26157         task_unlock(p);
26158
26159         /*
26160 @@ -584,6 +606,8 @@ int unregister_oom_notifier(struct notif
26161  }
26162  EXPORT_SYMBOL_GPL(unregister_oom_notifier);
26163
26164 +long vs_oom_action(unsigned int);
26165 +
26166  /*
26167   * Try to acquire the OOM killer lock for the zones in zonelist.  Returns zero
26168   * if a parallel OOM killing is already taking place that includes a zone in
26169 @@ -742,7 +766,12 @@ retry:
26170         if (!p) {
26171                 dump_header(NULL, gfp_mask, order, NULL, mpol_mask);
26172                 read_unlock(&tasklist_lock);
26173 -               panic("Out of memory and no killable processes...\n");
26174 +
26175 +               /* avoid panic for guest OOM */
26176 +               if (current->xid)
26177 +                       vs_oom_action(LINUX_REBOOT_CMD_OOM);
26178 +               else
26179 +                       panic("Out of memory and no killable processes...\n");
26180         }
26181
26182         if (oom_kill_process(p, gfp_mask, order, points, totalpages, NULL,
26183 diff -NurpP --minimal linux-3.0/mm/page_alloc.c linux-3.0-vs2.3.1-pre7/mm/page_alloc.c
26184 --- linux-3.0/mm/page_alloc.c   2011-07-22 11:18:12.000000000 +0200
26185 +++ linux-3.0-vs2.3.1-pre7/mm/page_alloc.c      2011-06-10 22:24:12.000000000 +0200
26186 @@ -57,6 +57,8 @@
26187  #include <linux/ftrace_event.h>
26188  #include <linux/memcontrol.h>
26189  #include <linux/prefetch.h>
26190 +#include <linux/vs_base.h>
26191 +#include <linux/vs_limit.h>
26192
26193  #include <asm/tlbflush.h>
26194  #include <asm/div64.h>
26195 @@ -2470,6 +2472,9 @@ void si_meminfo(struct sysinfo *val)
26196         val->totalhigh = totalhigh_pages;
26197         val->freehigh = nr_free_highpages();
26198         val->mem_unit = PAGE_SIZE;
26199 +
26200 +       if (vx_flags(VXF_VIRT_MEM, 0))
26201 +               vx_vsi_meminfo(val);
26202  }
26203
26204  EXPORT_SYMBOL(si_meminfo);
26205 @@ -2490,6 +2495,9 @@ void si_meminfo_node(struct sysinfo *val
26206         val->freehigh = 0;
26207  #endif
26208         val->mem_unit = PAGE_SIZE;
26209 +
26210 +       if (vx_flags(VXF_VIRT_MEM, 0))
26211 +               vx_vsi_meminfo(val);
26212  }
26213  #endif
26214
26215 diff -NurpP --minimal linux-3.0/mm/pgtable-generic.c linux-3.0-vs2.3.1-pre7/mm/pgtable-generic.c
26216 --- linux-3.0/mm/pgtable-generic.c      2011-03-15 18:07:42.000000000 +0100
26217 +++ linux-3.0-vs2.3.1-pre7/mm/pgtable-generic.c 2011-06-10 22:11:24.000000000 +0200
26218 @@ -6,6 +6,8 @@
26219   *  Copyright (C) 2010  Linus Torvalds
26220   */
26221
26222 +#include <linux/mm.h>
26223 +
26224  #include <linux/pagemap.h>
26225  #include <asm/tlb.h>
26226  #include <asm-generic/pgtable.h>
26227 diff -NurpP --minimal linux-3.0/mm/rmap.c linux-3.0-vs2.3.1-pre7/mm/rmap.c
26228 --- linux-3.0/mm/rmap.c 2011-07-22 11:18:12.000000000 +0200
26229 +++ linux-3.0-vs2.3.1-pre7/mm/rmap.c    2011-07-01 11:35:35.000000000 +0200
26230 @@ -57,6 +57,7 @@
26231  #include <linux/mmu_notifier.h>
26232  #include <linux/migrate.h>
26233  #include <linux/hugetlb.h>
26234 +#include <linux/vs_memory.h>
26235
26236  #include <asm/tlbflush.h>
26237
26238 diff -NurpP --minimal linux-3.0/mm/shmem.c linux-3.0-vs2.3.1-pre7/mm/shmem.c
26239 --- linux-3.0/mm/shmem.c        2011-07-22 11:18:12.000000000 +0200
26240 +++ linux-3.0-vs2.3.1-pre7/mm/shmem.c   2011-07-01 11:35:35.000000000 +0200
26241 @@ -1850,7 +1850,7 @@ static int shmem_statfs(struct dentry *d
26242  {
26243         struct shmem_sb_info *sbinfo = SHMEM_SB(dentry->d_sb);
26244
26245 -       buf->f_type = TMPFS_MAGIC;
26246 +       buf->f_type = TMPFS_SUPER_MAGIC;
26247         buf->f_bsize = PAGE_CACHE_SIZE;
26248         buf->f_namelen = NAME_MAX;
26249         if (sbinfo->max_blocks) {
26250 @@ -2605,7 +2605,7 @@ int shmem_fill_super(struct super_block
26251         sb->s_maxbytes = SHMEM_MAX_BYTES;
26252         sb->s_blocksize = PAGE_CACHE_SIZE;
26253         sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
26254 -       sb->s_magic = TMPFS_MAGIC;
26255 +       sb->s_magic = TMPFS_SUPER_MAGIC;
26256         sb->s_op = &shmem_ops;
26257         sb->s_time_gran = 1;
26258  #ifdef CONFIG_TMPFS_XATTR
26259 diff -NurpP --minimal linux-3.0/mm/slab.c linux-3.0-vs2.3.1-pre7/mm/slab.c
26260 --- linux-3.0/mm/slab.c 2011-07-22 11:18:12.000000000 +0200
26261 +++ linux-3.0-vs2.3.1-pre7/mm/slab.c    2011-06-15 02:41:23.000000000 +0200
26262 @@ -411,6 +411,8 @@ static void kmem_list3_init(struct kmem_
26263  #define STATS_INC_FREEMISS(x)  do { } while (0)
26264  #endif
26265
26266 +#include "slab_vs.h"
26267 +
26268  #if DEBUG
26269
26270  /*
26271 @@ -3348,6 +3350,7 @@ retry:
26272
26273         obj = slab_get_obj(cachep, slabp, nodeid);
26274         check_slabp(cachep, slabp);
26275 +       vx_slab_alloc(cachep, flags);
26276         l3->free_objects--;
26277         /* move slabp to correct slabp list: */
26278         list_del(&slabp->list);
26279 @@ -3425,6 +3428,7 @@ __cache_alloc_node(struct kmem_cache *ca
26280         /* ___cache_alloc_node can fall back to other nodes */
26281         ptr = ____cache_alloc_node(cachep, flags, nodeid);
26282    out:
26283 +       vx_slab_alloc(cachep, flags);
26284         local_irq_restore(save_flags);
26285         ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller);
26286         kmemleak_alloc_recursive(ptr, obj_size(cachep), 1, cachep->flags,
26287 @@ -3612,6 +3616,7 @@ static inline void __cache_free(struct k
26288         check_irq_off();
26289         kmemleak_free_recursive(objp, cachep->flags);
26290         objp = cache_free_debugcheck(cachep, objp, caller);
26291 +       vx_slab_free(cachep);
26292
26293         kmemcheck_slab_free(cachep, objp, obj_size(cachep));
26294
26295 diff -NurpP --minimal linux-3.0/mm/slab_vs.h linux-3.0-vs2.3.1-pre7/mm/slab_vs.h
26296 --- linux-3.0/mm/slab_vs.h      1970-01-01 01:00:00.000000000 +0100
26297 +++ linux-3.0-vs2.3.1-pre7/mm/slab_vs.h 2011-06-10 22:11:24.000000000 +0200
26298 @@ -0,0 +1,29 @@
26299 +
26300 +#include <linux/vserver/context.h>
26301 +
26302 +#include <linux/vs_context.h>
26303 +
26304 +static inline
26305 +void vx_slab_alloc(struct kmem_cache *cachep, gfp_t flags)
26306 +{
26307 +       int what = gfp_zone(cachep->gfpflags);
26308 +       struct vx_info *vxi = current_vx_info();
26309 +
26310 +       if (!vxi)
26311 +               return;
26312 +
26313 +       atomic_add(cachep->buffer_size, &vxi->cacct.slab[what]);
26314 +}
26315 +
26316 +static inline
26317 +void vx_slab_free(struct kmem_cache *cachep)
26318 +{
26319 +       int what = gfp_zone(cachep->gfpflags);
26320 +       struct vx_info *vxi = current_vx_info();
26321 +
26322 +       if (!vxi)
26323 +               return;
26324 +
26325 +       atomic_sub(cachep->buffer_size, &vxi->cacct.slab[what]);
26326 +}
26327 +
26328 diff -NurpP --minimal linux-3.0/mm/swapfile.c linux-3.0-vs2.3.1-pre7/mm/swapfile.c
26329 --- linux-3.0/mm/swapfile.c     2011-07-22 11:18:12.000000000 +0200
26330 +++ linux-3.0-vs2.3.1-pre7/mm/swapfile.c        2011-07-01 11:35:35.000000000 +0200
26331 @@ -37,6 +37,8 @@
26332  #include <asm/tlbflush.h>
26333  #include <linux/swapops.h>
26334  #include <linux/page_cgroup.h>
26335 +#include <linux/vs_base.h>
26336 +#include <linux/vs_memory.h>
26337
26338  static bool swap_count_continued(struct swap_info_struct *, pgoff_t,
26339                                  unsigned char);
26340 @@ -1759,6 +1761,16 @@ static int swap_show(struct seq_file *sw
26341
26342         if (si == SEQ_START_TOKEN) {
26343                 seq_puts(swap,"Filename\t\t\t\tType\t\tSize\tUsed\tPriority\n");
26344 +               if (vx_flags(VXF_VIRT_MEM, 0)) {
26345 +                       struct sysinfo si;
26346 +
26347 +                       vx_vsi_swapinfo(&si);
26348 +                       if (si.totalswap < (1 << 10))
26349 +                               return 0;
26350 +                       seq_printf(swap, "%s\t\t\t\t\t%s\t%lu\t%lu\t%d\n",
26351 +                               "hdv0", "partition", si.totalswap >> 10,
26352 +                               (si.totalswap - si.freeswap) >> 10, -1);
26353 +               }
26354                 return 0;
26355         }
26356
26357 @@ -2186,6 +2198,8 @@ void si_swapinfo(struct sysinfo *val)
26358         val->freeswap = nr_swap_pages + nr_to_be_unused;
26359         val->totalswap = total_swap_pages + nr_to_be_unused;
26360         spin_unlock(&swap_lock);
26361 +       if (vx_flags(VXF_VIRT_MEM, 0))
26362 +               vx_vsi_swapinfo(val);
26363  }
26364
26365  /*
26366 diff -NurpP --minimal linux-3.0/net/bridge/br_multicast.c linux-3.0-vs2.3.1-pre7/net/bridge/br_multicast.c
26367 --- linux-3.0/net/bridge/br_multicast.c 2011-07-22 11:18:12.000000000 +0200
26368 +++ linux-3.0-vs2.3.1-pre7/net/bridge/br_multicast.c    2011-07-19 00:44:39.000000000 +0200
26369 @@ -447,7 +447,7 @@ static struct sk_buff *br_ip6_multicast_
26370         ip6h->hop_limit = 1;
26371         ipv6_addr_set(&ip6h->daddr, htonl(0xff020000), 0, 0, htonl(1));
26372         ipv6_dev_get_saddr(dev_net(br->dev), br->dev, &ip6h->daddr, 0,
26373 -                          &ip6h->saddr);
26374 +                          &ip6h->saddr, NULL);
26375         ipv6_eth_mc_map(&ip6h->daddr, eth->h_dest);
26376
26377         hopopt = (u8 *)(ip6h + 1);
26378 diff -NurpP --minimal linux-3.0/net/core/dev.c linux-3.0-vs2.3.1-pre7/net/core/dev.c
26379 --- linux-3.0/net/core/dev.c    2011-07-22 11:18:13.000000000 +0200
26380 +++ linux-3.0-vs2.3.1-pre7/net/core/dev.c       2011-06-15 02:40:14.000000000 +0200
26381 @@ -127,6 +127,7 @@
26382  #include <linux/in.h>
26383  #include <linux/jhash.h>
26384  #include <linux/random.h>
26385 +#include <linux/vs_inet.h>
26386  #include <trace/events/napi.h>
26387  #include <trace/events/net.h>
26388  #include <trace/events/skb.h>
26389 @@ -609,7 +610,8 @@ struct net_device *__dev_get_by_name(str
26390         struct hlist_head *head = dev_name_hash(net, name);
26391
26392         hlist_for_each_entry(dev, p, head, name_hlist)
26393 -               if (!strncmp(dev->name, name, IFNAMSIZ))
26394 +               if (!strncmp(dev->name, name, IFNAMSIZ) &&
26395 +                   nx_dev_visible(current_nx_info(), dev))
26396                         return dev;
26397
26398         return NULL;
26399 @@ -635,7 +637,8 @@ struct net_device *dev_get_by_name_rcu(s
26400         struct hlist_head *head = dev_name_hash(net, name);
26401
26402         hlist_for_each_entry_rcu(dev, p, head, name_hlist)
26403 -               if (!strncmp(dev->name, name, IFNAMSIZ))
26404 +               if (!strncmp(dev->name, name, IFNAMSIZ) &&
26405 +                   nx_dev_visible(current_nx_info(), dev))
26406                         return dev;
26407
26408         return NULL;
26409 @@ -686,7 +689,8 @@ struct net_device *__dev_get_by_index(st
26410         struct hlist_head *head = dev_index_hash(net, ifindex);
26411
26412         hlist_for_each_entry(dev, p, head, index_hlist)
26413 -               if (dev->ifindex == ifindex)
26414 +               if ((dev->ifindex == ifindex) &&
26415 +                   nx_dev_visible(current_nx_info(), dev))
26416                         return dev;
26417
26418         return NULL;
26419 @@ -711,7 +715,8 @@ struct net_device *dev_get_by_index_rcu(
26420         struct hlist_head *head = dev_index_hash(net, ifindex);
26421
26422         hlist_for_each_entry_rcu(dev, p, head, index_hlist)
26423 -               if (dev->ifindex == ifindex)
26424 +               if ((dev->ifindex == ifindex) &&
26425 +                   nx_dev_visible(current_nx_info(), dev))
26426                         return dev;
26427
26428         return NULL;
26429 @@ -764,7 +769,8 @@ struct net_device *dev_getbyhwaddr_rcu(s
26430
26431         for_each_netdev_rcu(net, dev)
26432                 if (dev->type == type &&
26433 -                   !memcmp(dev->dev_addr, ha, dev->addr_len))
26434 +                   !memcmp(dev->dev_addr, ha, dev->addr_len) &&
26435 +                   nx_dev_visible(current_nx_info(), dev))
26436                         return dev;
26437
26438         return NULL;
26439 @@ -776,9 +782,11 @@ struct net_device *__dev_getfirstbyhwtyp
26440         struct net_device *dev;
26441
26442         ASSERT_RTNL();
26443 -       for_each_netdev(net, dev)
26444 -               if (dev->type == type)
26445 +       for_each_netdev(net, dev) {
26446 +               if ((dev->type == type) &&
26447 +                   nx_dev_visible(current_nx_info(), dev))
26448                         return dev;
26449 +       }
26450
26451         return NULL;
26452  }
26453 @@ -896,6 +904,8 @@ static int __dev_alloc_name(struct net *
26454                                 continue;
26455                         if (i < 0 || i >= max_netdevices)
26456                                 continue;
26457 +                       if (!nx_dev_visible(current_nx_info(), d))
26458 +                               continue;
26459
26460                         /*  avoid cases where sscanf is not exact inverse of printf */
26461                         snprintf(buf, IFNAMSIZ, name, i);
26462 @@ -3937,6 +3947,8 @@ static int dev_ifconf(struct net *net, c
26463
26464         total = 0;
26465         for_each_netdev(net, dev) {
26466 +               if (!nx_dev_visible(current_nx_info(), dev))
26467 +                       continue;
26468                 for (i = 0; i < NPROTO; i++) {
26469                         if (gifconf_list[i]) {
26470                                 int done;
26471 @@ -4011,6 +4023,10 @@ static void dev_seq_printf_stats(struct
26472         struct rtnl_link_stats64 temp;
26473         const struct rtnl_link_stats64 *stats = dev_get_stats(dev, &temp);
26474
26475 +       /* device visible inside network context? */
26476 +       if (!nx_dev_visible(current_nx_info(), dev))
26477 +               return;
26478 +
26479         seq_printf(seq, "%6s: %7llu %7llu %4llu %4llu %4llu %5llu %10llu %9llu "
26480                    "%8llu %7llu %4llu %4llu %4llu %5llu %7llu %10llu\n",
26481                    dev->name, stats->rx_bytes, stats->rx_packets,
26482 diff -NurpP --minimal linux-3.0/net/core/rtnetlink.c linux-3.0-vs2.3.1-pre7/net/core/rtnetlink.c
26483 --- linux-3.0/net/core/rtnetlink.c      2011-07-22 11:18:13.000000000 +0200
26484 +++ linux-3.0-vs2.3.1-pre7/net/core/rtnetlink.c 2011-06-10 22:11:24.000000000 +0200
26485 @@ -1015,6 +1015,8 @@ static int rtnl_dump_ifinfo(struct sk_bu
26486                 hlist_for_each_entry_rcu(dev, node, head, index_hlist) {
26487                         if (idx < s_idx)
26488                                 goto cont;
26489 +                       if (!nx_dev_visible(skb->sk->sk_nx_info, dev))
26490 +                               continue;
26491                         if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
26492                                              NETLINK_CB(cb->skb).pid,
26493                                              cb->nlh->nlmsg_seq, 0,
26494 @@ -1848,6 +1850,9 @@ void rtmsg_ifinfo(int type, struct net_d
26495         struct sk_buff *skb;
26496         int err = -ENOBUFS;
26497
26498 +       if (!nx_dev_visible(current_nx_info(), dev))
26499 +               return;
26500 +
26501         skb = nlmsg_new(if_nlmsg_size(dev), GFP_KERNEL);
26502         if (skb == NULL)
26503                 goto errout;
26504 diff -NurpP --minimal linux-3.0/net/core/sock.c linux-3.0-vs2.3.1-pre7/net/core/sock.c
26505 --- linux-3.0/net/core/sock.c   2011-05-22 16:18:00.000000000 +0200
26506 +++ linux-3.0-vs2.3.1-pre7/net/core/sock.c      2011-06-10 22:11:24.000000000 +0200
26507 @@ -127,6 +127,10 @@
26508  #include <net/cls_cgroup.h>
26509
26510  #include <linux/filter.h>
26511 +#include <linux/vs_socket.h>
26512 +#include <linux/vs_limit.h>
26513 +#include <linux/vs_context.h>
26514 +#include <linux/vs_network.h>
26515
26516  #ifdef CONFIG_INET
26517  #include <net/tcp.h>
26518 @@ -1070,6 +1074,8 @@ static struct sock *sk_prot_alloc(struct
26519                         goto out_free_sec;
26520                 sk_tx_queue_clear(sk);
26521         }
26522 +               sock_vx_init(sk);
26523 +               sock_nx_init(sk);
26524
26525         return sk;
26526
26527 @@ -1169,6 +1175,11 @@ static void __sk_free(struct sock *sk)
26528                 put_cred(sk->sk_peer_cred);
26529         put_pid(sk->sk_peer_pid);
26530         put_net(sock_net(sk));
26531 +       vx_sock_dec(sk);
26532 +       clr_vx_info(&sk->sk_vx_info);
26533 +       sk->sk_xid = -1;
26534 +       clr_nx_info(&sk->sk_nx_info);
26535 +       sk->sk_nid = -1;
26536         sk_prot_free(sk->sk_prot_creator, sk);
26537  }
26538
26539 @@ -1216,6 +1227,8 @@ struct sock *sk_clone(const struct sock
26540
26541                 /* SANITY */
26542                 get_net(sock_net(newsk));
26543 +               sock_vx_init(newsk);
26544 +               sock_nx_init(newsk);
26545                 sk_node_init(&newsk->sk_node);
26546                 sock_lock_init(newsk);
26547                 bh_lock_sock(newsk);
26548 @@ -1271,6 +1284,12 @@ struct sock *sk_clone(const struct sock
26549                 smp_wmb();
26550                 atomic_set(&newsk->sk_refcnt, 2);
26551
26552 +               set_vx_info(&newsk->sk_vx_info, sk->sk_vx_info);
26553 +               newsk->sk_xid = sk->sk_xid;
26554 +               vx_sock_inc(newsk);
26555 +               set_nx_info(&newsk->sk_nx_info, sk->sk_nx_info);
26556 +               newsk->sk_nid = sk->sk_nid;
26557 +
26558                 /*
26559                  * Increment the counter in the same struct proto as the master
26560                  * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that
26561 @@ -2017,6 +2036,12 @@ void sock_init_data(struct socket *sock,
26562
26563         sk->sk_stamp = ktime_set(-1L, 0);
26564
26565 +       set_vx_info(&sk->sk_vx_info, current_vx_info());
26566 +       sk->sk_xid = vx_current_xid();
26567 +       vx_sock_inc(sk);
26568 +       set_nx_info(&sk->sk_nx_info, current_nx_info());
26569 +       sk->sk_nid = nx_current_nid();
26570 +
26571         /*
26572          * Before updating sk_refcnt, we must commit prior changes to memory
26573          * (Documentation/RCU/rculist_nulls.txt for details)
26574 diff -NurpP --minimal linux-3.0/net/ipv4/af_inet.c linux-3.0-vs2.3.1-pre7/net/ipv4/af_inet.c
26575 --- linux-3.0/net/ipv4/af_inet.c        2011-07-22 11:18:13.000000000 +0200
26576 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/af_inet.c   2011-07-19 00:44:39.000000000 +0200
26577 @@ -117,6 +117,7 @@
26578  #ifdef CONFIG_IP_MROUTE
26579  #include <linux/mroute.h>
26580  #endif
26581 +#include <linux/vs_limit.h>
26582
26583
26584  /* The inetsw table contains everything that inet_create needs to
26585 @@ -326,9 +327,13 @@ lookup_protocol:
26586         }
26587
26588         err = -EPERM;
26589 +       if ((protocol == IPPROTO_ICMP) &&
26590 +               nx_capable(CAP_NET_RAW, NXC_RAW_ICMP))
26591 +               goto override;
26592 +
26593         if (sock->type == SOCK_RAW && !kern && !capable(CAP_NET_RAW))
26594                 goto out_rcu_unlock;
26595 -
26596 +override:
26597         err = -EAFNOSUPPORT;
26598         if (!inet_netns_ok(net, protocol))
26599                 goto out_rcu_unlock;
26600 @@ -452,6 +457,7 @@ int inet_bind(struct socket *sock, struc
26601         struct sockaddr_in *addr = (struct sockaddr_in *)uaddr;
26602         struct sock *sk = sock->sk;
26603         struct inet_sock *inet = inet_sk(sk);
26604 +       struct nx_v4_sock_addr nsa;
26605         unsigned short snum;
26606         int chk_addr_ret;
26607         int err;
26608 @@ -470,7 +476,11 @@ int inet_bind(struct socket *sock, struc
26609                 goto out;
26610         }
26611
26612 -       chk_addr_ret = inet_addr_type(sock_net(sk), addr->sin_addr.s_addr);
26613 +       err = v4_map_sock_addr(inet, addr, &nsa);
26614 +       if (err)
26615 +               goto out;
26616 +
26617 +       chk_addr_ret = inet_addr_type(sock_net(sk), nsa.saddr);
26618
26619         /* Not specified by any standard per-se, however it breaks too
26620          * many applications when removed.  It is unfortunate since
26621 @@ -482,7 +492,7 @@ int inet_bind(struct socket *sock, struc
26622         err = -EADDRNOTAVAIL;
26623         if (!sysctl_ip_nonlocal_bind &&
26624             !(inet->freebind || inet->transparent) &&
26625 -           addr->sin_addr.s_addr != htonl(INADDR_ANY) &&
26626 +           nsa.saddr != htonl(INADDR_ANY) &&
26627             chk_addr_ret != RTN_LOCAL &&
26628             chk_addr_ret != RTN_MULTICAST &&
26629             chk_addr_ret != RTN_BROADCAST)
26630 @@ -507,7 +517,7 @@ int inet_bind(struct socket *sock, struc
26631         if (sk->sk_state != TCP_CLOSE || inet->inet_num)
26632                 goto out_release_sock;
26633
26634 -       inet->inet_rcv_saddr = inet->inet_saddr = addr->sin_addr.s_addr;
26635 +       v4_set_sock_addr(inet, &nsa);
26636         if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
26637                 inet->inet_saddr = 0;  /* Use device */
26638
26639 @@ -710,11 +720,13 @@ int inet_getname(struct socket *sock, st
26640                      peer == 1))
26641                         return -ENOTCONN;
26642                 sin->sin_port = inet->inet_dport;
26643 -               sin->sin_addr.s_addr = inet->inet_daddr;
26644 +               sin->sin_addr.s_addr =
26645 +                       nx_map_sock_lback(sk->sk_nx_info, inet->inet_daddr);
26646         } else {
26647                 __be32 addr = inet->inet_rcv_saddr;
26648                 if (!addr)
26649                         addr = inet->inet_saddr;
26650 +               addr = nx_map_sock_lback(sk->sk_nx_info, addr);
26651                 sin->sin_port = inet->inet_sport;
26652                 sin->sin_addr.s_addr = addr;
26653         }
26654 diff -NurpP --minimal linux-3.0/net/ipv4/devinet.c linux-3.0-vs2.3.1-pre7/net/ipv4/devinet.c
26655 --- linux-3.0/net/ipv4/devinet.c        2011-07-22 11:18:13.000000000 +0200
26656 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/devinet.c   2011-06-10 22:11:24.000000000 +0200
26657 @@ -518,6 +518,7 @@ struct in_device *inetdev_by_index(struc
26658  }
26659  EXPORT_SYMBOL(inetdev_by_index);
26660
26661 +
26662  /* Called only from RTNL semaphored context. No locks. */
26663
26664  struct in_ifaddr *inet_ifa_byprefix(struct in_device *in_dev, __be32 prefix,
26665 @@ -759,6 +760,8 @@ int devinet_ioctl(struct net *net, unsig
26666
26667         in_dev = __in_dev_get_rtnl(dev);
26668         if (in_dev) {
26669 +               struct nx_info *nxi = current_nx_info();
26670 +
26671                 if (tryaddrmatch) {
26672                         /* Matthias Andree */
26673                         /* compare label and address (4.4BSD style) */
26674 @@ -767,6 +770,8 @@ int devinet_ioctl(struct net *net, unsig
26675                            This is checked above. */
26676                         for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
26677                              ifap = &ifa->ifa_next) {
26678 +                               if (!nx_v4_ifa_visible(nxi, ifa))
26679 +                                       continue;
26680                                 if (!strcmp(ifr.ifr_name, ifa->ifa_label) &&
26681                                     sin_orig.sin_addr.s_addr ==
26682                                                         ifa->ifa_local) {
26683 @@ -779,9 +784,12 @@ int devinet_ioctl(struct net *net, unsig
26684                    comparing just the label */
26685                 if (!ifa) {
26686                         for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL;
26687 -                            ifap = &ifa->ifa_next)
26688 +                            ifap = &ifa->ifa_next) {
26689 +                               if (!nx_v4_ifa_visible(nxi, ifa))
26690 +                                       continue;
26691                                 if (!strcmp(ifr.ifr_name, ifa->ifa_label))
26692                                         break;
26693 +                       }
26694                 }
26695         }
26696
26697 @@ -934,6 +942,8 @@ static int inet_gifconf(struct net_devic
26698                 goto out;
26699
26700         for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
26701 +               if (!nx_v4_ifa_visible(current_nx_info(), ifa))
26702 +                       continue;
26703                 if (!buf) {
26704                         done += sizeof(ifr);
26705                         continue;
26706 @@ -1294,6 +1304,7 @@ static int inet_dump_ifaddr(struct sk_bu
26707         struct net_device *dev;
26708         struct in_device *in_dev;
26709         struct in_ifaddr *ifa;
26710 +       struct sock *sk = skb->sk;
26711         struct hlist_head *head;
26712         struct hlist_node *node;
26713
26714 @@ -1316,6 +1327,8 @@ static int inet_dump_ifaddr(struct sk_bu
26715
26716                         for (ifa = in_dev->ifa_list, ip_idx = 0; ifa;
26717                              ifa = ifa->ifa_next, ip_idx++) {
26718 +                       if (sk && !nx_v4_ifa_visible(sk->sk_nx_info, ifa))
26719 +                               continue;
26720                                 if (ip_idx < s_ip_idx)
26721                                         continue;
26722                                 if (inet_fill_ifaddr(skb, ifa,
26723 diff -NurpP --minimal linux-3.0/net/ipv4/inet_connection_sock.c linux-3.0-vs2.3.1-pre7/net/ipv4/inet_connection_sock.c
26724 --- linux-3.0/net/ipv4/inet_connection_sock.c   2011-07-22 11:18:13.000000000 +0200
26725 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/inet_connection_sock.c      2011-06-10 22:11:24.000000000 +0200
26726 @@ -52,6 +52,37 @@ void inet_get_local_port_range(int *low,
26727  }
26728  EXPORT_SYMBOL(inet_get_local_port_range);
26729
26730 +int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
26731 +{
26732 +       __be32  sk1_rcv_saddr = sk_rcv_saddr(sk1),
26733 +               sk2_rcv_saddr = sk_rcv_saddr(sk2);
26734 +
26735 +       if (inet_v6_ipv6only(sk2))
26736 +               return 0;
26737 +
26738 +       if (sk1_rcv_saddr &&
26739 +           sk2_rcv_saddr &&
26740 +           sk1_rcv_saddr == sk2_rcv_saddr)
26741 +               return 1;
26742 +
26743 +       if (sk1_rcv_saddr &&
26744 +           !sk2_rcv_saddr &&
26745 +           v4_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr, NXA_MASK_BIND))
26746 +               return 1;
26747 +
26748 +       if (sk2_rcv_saddr &&
26749 +           !sk1_rcv_saddr &&
26750 +           v4_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr, NXA_MASK_BIND))
26751 +               return 1;
26752 +
26753 +       if (!sk1_rcv_saddr &&
26754 +           !sk2_rcv_saddr &&
26755 +           nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info))
26756 +               return 1;
26757 +
26758 +       return 0;
26759 +}
26760 +
26761  int inet_csk_bind_conflict(const struct sock *sk,
26762                            const struct inet_bind_bucket *tb)
26763  {
26764 @@ -74,9 +105,7 @@ int inet_csk_bind_conflict(const struct
26765                      sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
26766                         if (!reuse || !sk2->sk_reuse ||
26767                             sk2->sk_state == TCP_LISTEN) {
26768 -                               const __be32 sk2_rcv_saddr = sk_rcv_saddr(sk2);
26769 -                               if (!sk2_rcv_saddr || !sk_rcv_saddr(sk) ||
26770 -                                   sk2_rcv_saddr == sk_rcv_saddr(sk))
26771 +                               if (ipv4_rcv_saddr_equal(sk, sk2))
26772                                         break;
26773                         }
26774                 }
26775 diff -NurpP --minimal linux-3.0/net/ipv4/inet_diag.c linux-3.0-vs2.3.1-pre7/net/ipv4/inet_diag.c
26776 --- linux-3.0/net/ipv4/inet_diag.c      2011-07-22 11:18:13.000000000 +0200
26777 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/inet_diag.c 2011-06-22 12:39:16.000000000 +0200
26778 @@ -33,6 +33,8 @@
26779  #include <linux/stddef.h>
26780
26781  #include <linux/inet_diag.h>
26782 +#include <linux/vs_network.h>
26783 +#include <linux/vs_inet.h>
26784
26785  static const struct inet_diag_handler **inet_diag_table;
26786
26787 @@ -119,8 +121,10 @@ static int inet_csk_diag_fill(struct soc
26788
26789         r->id.idiag_sport = inet->inet_sport;
26790         r->id.idiag_dport = inet->inet_dport;
26791 -       r->id.idiag_src[0] = inet->inet_rcv_saddr;
26792 -       r->id.idiag_dst[0] = inet->inet_daddr;
26793 +       r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info,
26794 +               inet->inet_rcv_saddr);
26795 +       r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info,
26796 +               inet->inet_daddr);
26797
26798  #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
26799         if (r->idiag_family == AF_INET6) {
26800 @@ -205,8 +209,8 @@ static int inet_twsk_diag_fill(struct in
26801         r->id.idiag_cookie[1] = (u32)(((unsigned long)tw >> 31) >> 1);
26802         r->id.idiag_sport     = tw->tw_sport;
26803         r->id.idiag_dport     = tw->tw_dport;
26804 -       r->id.idiag_src[0]    = tw->tw_rcv_saddr;
26805 -       r->id.idiag_dst[0]    = tw->tw_daddr;
26806 +       r->id.idiag_src[0]    = nx_map_sock_lback(tw->tw_nx_info, tw->tw_rcv_saddr);
26807 +       r->id.idiag_dst[0]    = nx_map_sock_lback(tw->tw_nx_info, tw->tw_daddr);
26808         r->idiag_state        = tw->tw_substate;
26809         r->idiag_timer        = 3;
26810         r->idiag_expires      = DIV_ROUND_UP(tmo * 1000, HZ);
26811 @@ -263,6 +267,7 @@ static int inet_diag_get_exact(struct sk
26812         err = -EINVAL;
26813
26814         if (req->idiag_family == AF_INET) {
26815 +               /* TODO: lback */
26816                 sk = inet_lookup(&init_net, hashinfo, req->id.idiag_dst[0],
26817                                  req->id.idiag_dport, req->id.idiag_src[0],
26818                                  req->id.idiag_sport, req->id.idiag_if);
26819 @@ -505,6 +510,7 @@ static int inet_csk_diag_dump(struct soc
26820                 } else
26821  #endif
26822                 {
26823 +                       /* TODO: lback */
26824                         entry.saddr = &inet->inet_rcv_saddr;
26825                         entry.daddr = &inet->inet_daddr;
26826                 }
26827 @@ -543,6 +549,7 @@ static int inet_twsk_diag_dump(struct in
26828                 } else
26829  #endif
26830                 {
26831 +                       /* TODO: lback */
26832                         entry.saddr = &tw->tw_rcv_saddr;
26833                         entry.daddr = &tw->tw_daddr;
26834                 }
26835 @@ -589,8 +596,8 @@ static int inet_diag_fill_req(struct sk_
26836
26837         r->id.idiag_sport = inet->inet_sport;
26838         r->id.idiag_dport = ireq->rmt_port;
26839 -       r->id.idiag_src[0] = ireq->loc_addr;
26840 -       r->id.idiag_dst[0] = ireq->rmt_addr;
26841 +       r->id.idiag_src[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->loc_addr);
26842 +       r->id.idiag_dst[0] = nx_map_sock_lback(sk->sk_nx_info, ireq->rmt_addr);
26843         r->idiag_expires = jiffies_to_msecs(tmo);
26844         r->idiag_rqueue = 0;
26845         r->idiag_wqueue = 0;
26846 @@ -661,6 +668,7 @@ static int inet_diag_dump_reqs(struct sk
26847                                 continue;
26848
26849                         if (bc) {
26850 +                               /* TODO: lback */
26851                                 entry.saddr =
26852  #if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
26853                                         (entry.family == AF_INET6) ?
26854 @@ -731,6 +739,8 @@ static int inet_diag_dump(struct sk_buff
26855                         sk_nulls_for_each(sk, node, &ilb->head) {
26856                                 struct inet_sock *inet = inet_sk(sk);
26857
26858 +                               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
26859 +                                       continue;
26860                                 if (num < s_num) {
26861                                         num++;
26862                                         continue;
26863 @@ -797,6 +807,8 @@ skip_listen_ht:
26864                 sk_nulls_for_each(sk, node, &head->chain) {
26865                         struct inet_sock *inet = inet_sk(sk);
26866
26867 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
26868 +                               continue;
26869                         if (num < s_num)
26870                                 goto next_normal;
26871                         if (!(r->idiag_states & (1 << sk->sk_state)))
26872 @@ -821,6 +833,8 @@ next_normal:
26873                         inet_twsk_for_each(tw, node,
26874                                     &head->twchain) {
26875
26876 +                               if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
26877 +                                       continue;
26878                                 if (num < s_num)
26879                                         goto next_dying;
26880                                 if (r->id.idiag_sport != tw->tw_sport &&
26881 diff -NurpP --minimal linux-3.0/net/ipv4/inet_hashtables.c linux-3.0-vs2.3.1-pre7/net/ipv4/inet_hashtables.c
26882 --- linux-3.0/net/ipv4/inet_hashtables.c        2011-01-05 21:50:42.000000000 +0100
26883 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/inet_hashtables.c   2011-06-10 22:11:24.000000000 +0200
26884 @@ -21,6 +21,7 @@
26885
26886  #include <net/inet_connection_sock.h>
26887  #include <net/inet_hashtables.h>
26888 +#include <net/route.h>
26889  #include <net/ip.h>
26890
26891  /*
26892 @@ -155,6 +156,11 @@ static inline int compute_score(struct s
26893                         if (rcv_saddr != daddr)
26894                                 return -1;
26895                         score += 2;
26896 +               } else {
26897 +                       /* block non nx_info ips */
26898 +                       if (!v4_addr_in_nx_info(sk->sk_nx_info,
26899 +                               daddr, NXA_MASK_BIND))
26900 +                               return -1;
26901                 }
26902                 if (sk->sk_bound_dev_if) {
26903                         if (sk->sk_bound_dev_if != dif)
26904 @@ -172,7 +178,6 @@ static inline int compute_score(struct s
26905   * wildcarded during the search since they can never be otherwise.
26906   */
26907
26908 -
26909  struct sock *__inet_lookup_listener(struct net *net,
26910                                     struct inet_hashinfo *hashinfo,
26911                                     const __be32 daddr, const unsigned short hnum,
26912 @@ -195,6 +200,7 @@ begin:
26913                         hiscore = score;
26914                 }
26915         }
26916 +
26917         /*
26918          * if the nulls value we got at the end of this lookup is
26919          * not the expected one, we must restart lookup.
26920 diff -NurpP --minimal linux-3.0/net/ipv4/netfilter/nf_nat_helper.c linux-3.0-vs2.3.1-pre7/net/ipv4/netfilter/nf_nat_helper.c
26921 --- linux-3.0/net/ipv4/netfilter/nf_nat_helper.c        2011-07-22 11:18:13.000000000 +0200
26922 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/netfilter/nf_nat_helper.c   2011-06-15 02:40:14.000000000 +0200
26923 @@ -20,6 +20,7 @@
26924  #include <net/route.h>
26925
26926  #include <linux/netfilter_ipv4.h>
26927 +#include <net/route.h>
26928  #include <net/netfilter/nf_conntrack.h>
26929  #include <net/netfilter/nf_conntrack_helper.h>
26930  #include <net/netfilter/nf_conntrack_ecache.h>
26931 diff -NurpP --minimal linux-3.0/net/ipv4/netfilter.c linux-3.0-vs2.3.1-pre7/net/ipv4/netfilter.c
26932 --- linux-3.0/net/ipv4/netfilter.c      2011-07-22 11:18:13.000000000 +0200
26933 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/netfilter.c 2011-07-19 00:44:39.000000000 +0200
26934 @@ -5,7 +5,7 @@
26935  #include <linux/ip.h>
26936  #include <linux/skbuff.h>
26937  #include <linux/gfp.h>
26938 -#include <net/route.h>
26939 +// #include <net/route.h>
26940  #include <net/xfrm.h>
26941  #include <net/ip.h>
26942  #include <net/netfilter/nf_queue.h>
26943 diff -NurpP --minimal linux-3.0/net/ipv4/raw.c linux-3.0-vs2.3.1-pre7/net/ipv4/raw.c
26944 --- linux-3.0/net/ipv4/raw.c    2011-07-22 11:18:13.000000000 +0200
26945 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/raw.c       2011-06-10 22:11:24.000000000 +0200
26946 @@ -117,7 +117,7 @@ static struct sock *__raw_v4_lookup(stru
26947
26948                 if (net_eq(sock_net(sk), net) && inet->inet_num == num  &&
26949                     !(inet->inet_daddr && inet->inet_daddr != raddr)    &&
26950 -                   !(inet->inet_rcv_saddr && inet->inet_rcv_saddr != laddr) &&
26951 +                   v4_sock_addr_match(sk->sk_nx_info, inet, laddr)     &&
26952                     !(sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif))
26953                         goto found; /* gotcha */
26954         }
26955 @@ -385,6 +385,12 @@ static int raw_send_hdrinc(struct sock *
26956                 icmp_out_count(net, ((struct icmphdr *)
26957                         skb_transport_header(skb))->type);
26958
26959 +       err = -EPERM;
26960 +       if (!nx_check(0, VS_ADMIN) && !capable(CAP_NET_RAW) &&
26961 +               sk->sk_nx_info &&
26962 +               !v4_addr_in_nx_info(sk->sk_nx_info, iph->saddr, NXA_MASK_BIND))
26963 +               goto error_free;
26964 +
26965         err = NF_HOOK(NFPROTO_IPV4, NF_INET_LOCAL_OUT, skb, NULL,
26966                       rt->dst.dev, dst_output);
26967         if (err > 0)
26968 @@ -647,17 +653,19 @@ static int raw_bind(struct sock *sk, str
26969  {
26970         struct inet_sock *inet = inet_sk(sk);
26971         struct sockaddr_in *addr = (struct sockaddr_in *) uaddr;
26972 +       struct nx_v4_sock_addr nsa = { 0 };
26973         int ret = -EINVAL;
26974         int chk_addr_ret;
26975
26976         if (sk->sk_state != TCP_CLOSE || addr_len < sizeof(struct sockaddr_in))
26977                 goto out;
26978 -       chk_addr_ret = inet_addr_type(sock_net(sk), addr->sin_addr.s_addr);
26979 +       v4_map_sock_addr(inet, addr, &nsa);
26980 +       chk_addr_ret = inet_addr_type(sock_net(sk), nsa.saddr);
26981         ret = -EADDRNOTAVAIL;
26982 -       if (addr->sin_addr.s_addr && chk_addr_ret != RTN_LOCAL &&
26983 +       if (nsa.saddr && chk_addr_ret != RTN_LOCAL &&
26984             chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST)
26985                 goto out;
26986 -       inet->inet_rcv_saddr = inet->inet_saddr = addr->sin_addr.s_addr;
26987 +       v4_set_sock_addr(inet, &nsa);
26988         if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST)
26989                 inet->inet_saddr = 0;  /* Use device */
26990         sk_dst_reset(sk);
26991 @@ -709,7 +717,8 @@ static int raw_recvmsg(struct kiocb *ioc
26992         /* Copy the address. */
26993         if (sin) {
26994                 sin->sin_family = AF_INET;
26995 -               sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
26996 +               sin->sin_addr.s_addr =
26997 +                       nx_map_sock_lback(sk->sk_nx_info, ip_hdr(skb)->saddr);
26998                 sin->sin_port = 0;
26999                 memset(&sin->sin_zero, 0, sizeof(sin->sin_zero));
27000         }
27001 @@ -905,7 +914,8 @@ static struct sock *raw_get_first(struct
27002                 struct hlist_node *node;
27003
27004                 sk_for_each(sk, node, &state->h->ht[state->bucket])
27005 -                       if (sock_net(sk) == seq_file_net(seq))
27006 +                       if ((sock_net(sk) == seq_file_net(seq)) &&
27007 +                               nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
27008                                 goto found;
27009         }
27010         sk = NULL;
27011 @@ -921,7 +931,8 @@ static struct sock *raw_get_next(struct
27012                 sk = sk_next(sk);
27013  try_again:
27014                 ;
27015 -       } while (sk && sock_net(sk) != seq_file_net(seq));
27016 +       } while (sk && ((sock_net(sk) != seq_file_net(seq)) ||
27017 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
27018
27019         if (!sk && ++state->bucket < RAW_HTABLE_SIZE) {
27020                 sk = sk_head(&state->h->ht[state->bucket]);
27021 diff -NurpP --minimal linux-3.0/net/ipv4/tcp.c linux-3.0-vs2.3.1-pre7/net/ipv4/tcp.c
27022 --- linux-3.0/net/ipv4/tcp.c    2011-07-22 11:18:13.000000000 +0200
27023 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/tcp.c       2011-07-19 00:44:39.000000000 +0200
27024 @@ -266,6 +266,7 @@
27025  #include <linux/crypto.h>
27026  #include <linux/time.h>
27027  #include <linux/slab.h>
27028 +#include <linux/in.h>
27029
27030  #include <net/icmp.h>
27031  #include <net/tcp.h>
27032 diff -NurpP --minimal linux-3.0/net/ipv4/tcp_ipv4.c linux-3.0-vs2.3.1-pre7/net/ipv4/tcp_ipv4.c
27033 --- linux-3.0/net/ipv4/tcp_ipv4.c       2011-07-22 11:18:13.000000000 +0200
27034 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/tcp_ipv4.c  2011-06-22 12:39:16.000000000 +0200
27035 @@ -2002,6 +2002,12 @@ static void *listening_get_next(struct s
27036                 req = req->dl_next;
27037                 while (1) {
27038                         while (req) {
27039 +                               vxdprintk(VXD_CBIT(net, 6),
27040 +                                       "sk,req: %p [#%d] (from %d)", req->sk,
27041 +                                       (req->sk)?req->sk->sk_nid:0, nx_current_nid());
27042 +                               if (req->sk &&
27043 +                                       !nx_check(req->sk->sk_nid, VS_WATCH_P | VS_IDENT))
27044 +                                       continue;
27045                                 if (req->rsk_ops->family == st->family) {
27046                                         cur = req;
27047                                         goto out;
27048 @@ -2026,6 +2032,10 @@ get_req:
27049         }
27050  get_sk:
27051         sk_nulls_for_each_from(sk, node) {
27052 +               vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)",
27053 +                       sk, sk->sk_nid, nx_current_nid());
27054 +               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
27055 +                       continue;
27056                 if (!net_eq(sock_net(sk), net))
27057                         continue;
27058                 if (sk->sk_family == st->family) {
27059 @@ -2102,6 +2112,11 @@ static void *established_get_first(struc
27060
27061                 spin_lock_bh(lock);
27062                 sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
27063 +                       vxdprintk(VXD_CBIT(net, 6),
27064 +                               "sk,egf: %p [#%d] (from %d)",
27065 +                               sk, sk->sk_nid, nx_current_nid());
27066 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
27067 +                               continue;
27068                         if (sk->sk_family != st->family ||
27069                             !net_eq(sock_net(sk), net)) {
27070                                 continue;
27071 @@ -2112,6 +2127,11 @@ static void *established_get_first(struc
27072                 st->state = TCP_SEQ_STATE_TIME_WAIT;
27073                 inet_twsk_for_each(tw, node,
27074                                    &tcp_hashinfo.ehash[st->bucket].twchain) {
27075 +                       vxdprintk(VXD_CBIT(net, 6),
27076 +                               "tw: %p [#%d] (from %d)",
27077 +                               tw, tw->tw_nid, nx_current_nid());
27078 +                       if (!nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))
27079 +                               continue;
27080                         if (tw->tw_family != st->family ||
27081                             !net_eq(twsk_net(tw), net)) {
27082                                 continue;
27083 @@ -2141,7 +2161,9 @@ static void *established_get_next(struct
27084                 tw = cur;
27085                 tw = tw_next(tw);
27086  get_tw:
27087 -               while (tw && (tw->tw_family != st->family || !net_eq(twsk_net(tw), net))) {
27088 +               while (tw && (tw->tw_family != st->family ||
27089 +                       !net_eq(twsk_net(tw), net) ||
27090 +                       !nx_check(tw->tw_nid, VS_WATCH_P | VS_IDENT))) {
27091                         tw = tw_next(tw);
27092                 }
27093                 if (tw) {
27094 @@ -2165,6 +2187,11 @@ get_tw:
27095                 sk = sk_nulls_next(sk);
27096
27097         sk_nulls_for_each_from(sk, node) {
27098 +               vxdprintk(VXD_CBIT(net, 6),
27099 +                       "sk,egn: %p [#%d] (from %d)",
27100 +                       sk, sk->sk_nid, nx_current_nid());
27101 +               if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
27102 +                       continue;
27103                 if (sk->sk_family == st->family && net_eq(sock_net(sk), net))
27104                         goto found;
27105         }
27106 @@ -2374,9 +2401,9 @@ static void get_openreq4(struct sock *sk
27107         seq_printf(f, "%4d: %08X:%04X %08X:%04X"
27108                 " %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %pK%n",
27109                 i,
27110 -               ireq->loc_addr,
27111 +               nx_map_sock_lback(current_nx_info(), ireq->loc_addr),
27112                 ntohs(inet_sk(sk)->inet_sport),
27113 -               ireq->rmt_addr,
27114 +               nx_map_sock_lback(current_nx_info(), ireq->rmt_addr),
27115                 ntohs(ireq->rmt_port),
27116                 TCP_SYN_RECV,
27117                 0, 0, /* could print option size, but that is af dependent. */
27118 diff -NurpP --minimal linux-3.0/net/ipv4/tcp_minisocks.c linux-3.0-vs2.3.1-pre7/net/ipv4/tcp_minisocks.c
27119 --- linux-3.0/net/ipv4/tcp_minisocks.c  2011-03-15 18:07:45.000000000 +0100
27120 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/tcp_minisocks.c     2011-06-10 22:11:24.000000000 +0200
27121 @@ -23,6 +23,9 @@
27122  #include <linux/slab.h>
27123  #include <linux/sysctl.h>
27124  #include <linux/workqueue.h>
27125 +#include <linux/vs_limit.h>
27126 +#include <linux/vs_socket.h>
27127 +#include <linux/vs_context.h>
27128  #include <net/tcp.h>
27129  #include <net/inet_common.h>
27130  #include <net/xfrm.h>
27131 @@ -335,6 +338,11 @@ void tcp_time_wait(struct sock *sk, int
27132                 tcptw->tw_ts_recent     = tp->rx_opt.ts_recent;
27133                 tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
27134
27135 +               tw->tw_xid              = sk->sk_xid;
27136 +               tw->tw_vx_info          = NULL;
27137 +               tw->tw_nid              = sk->sk_nid;
27138 +               tw->tw_nx_info          = NULL;
27139 +
27140  #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
27141                 if (tw->tw_family == PF_INET6) {
27142                         struct ipv6_pinfo *np = inet6_sk(sk);
27143 diff -NurpP --minimal linux-3.0/net/ipv4/udp.c linux-3.0-vs2.3.1-pre7/net/ipv4/udp.c
27144 --- linux-3.0/net/ipv4/udp.c    2011-07-22 11:18:13.000000000 +0200
27145 +++ linux-3.0-vs2.3.1-pre7/net/ipv4/udp.c       2011-07-19 00:44:39.000000000 +0200
27146 @@ -296,14 +296,7 @@ fail:
27147  }
27148  EXPORT_SYMBOL(udp_lib_get_port);
27149
27150 -static int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
27151 -{
27152 -       struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);
27153 -
27154 -       return  (!ipv6_only_sock(sk2)  &&
27155 -                (!inet1->inet_rcv_saddr || !inet2->inet_rcv_saddr ||
27156 -                  inet1->inet_rcv_saddr == inet2->inet_rcv_saddr));
27157 -}
27158 +extern int ipv4_rcv_saddr_equal(const struct sock *, const struct sock *);
27159
27160  static unsigned int udp4_portaddr_hash(struct net *net, __be32 saddr,
27161                                        unsigned int port)
27162 @@ -338,6 +331,11 @@ static inline int compute_score(struct s
27163                         if (inet->inet_rcv_saddr != daddr)
27164                                 return -1;
27165                         score += 2;
27166 +               } else {
27167 +                       /* block non nx_info ips */
27168 +                       if (!v4_addr_in_nx_info(sk->sk_nx_info,
27169 +                               daddr, NXA_MASK_BIND))
27170 +                               return -1;
27171                 }
27172                 if (inet->inet_daddr) {
27173                         if (inet->inet_daddr != saddr)
27174 @@ -441,6 +439,7 @@ exact_match:
27175         return result;
27176  }
27177
27178 +
27179  /* UDP is nearly always wildcards out the wazoo, it makes no sense to try
27180   * harder than this. -DaveM
27181   */
27182 @@ -486,6 +485,11 @@ begin:
27183         sk_nulls_for_each_rcu(sk, node, &hslot->head) {
27184                 score = compute_score(sk, net, saddr, hnum, sport,
27185                                       daddr, dport, dif);
27186 +               /* FIXME: disabled?
27187 +               if (score == 9) {
27188 +                       result = sk;
27189 +                       break;
27190 +               } else */
27191                 if (score > badness) {
27192                         result = sk;
27193                         badness = score;
27194 @@ -499,6 +503,7 @@ begin:
27195         if (get_nulls_value(node) != slot)
27196                 goto begin;
27197
27198 +
27199         if (result) {
27200                 if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
27201                         result = NULL;
27202 @@ -508,6 +513,7 @@ begin:
27203                         goto begin;
27204                 }
27205         }
27206 +
27207         rcu_read_unlock();
27208         return result;
27209  }
27210 @@ -550,8 +556,7 @@ static inline struct sock *udp_v4_mcast_
27211                     udp_sk(s)->udp_port_hash != hnum ||
27212                     (inet->inet_daddr && inet->inet_daddr != rmt_addr) ||
27213                     (inet->inet_dport != rmt_port && inet->inet_dport) ||
27214 -                   (inet->inet_rcv_saddr &&
27215 -                    inet->inet_rcv_saddr != loc_addr) ||
27216 +                   !v4_sock_addr_match(sk->sk_nx_info, inet, loc_addr) ||
27217                     ipv6_only_sock(s) ||
27218                     (s->sk_bound_dev_if && s->sk_bound_dev_if != dif))
27219                         continue;
27220 @@ -1226,7 +1231,8 @@ try_again:
27221         if (sin) {
27222                 sin->sin_family = AF_INET;
27223                 sin->sin_port = udp_hdr(skb)->source;
27224 -               sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
27225 +               sin->sin_addr.s_addr = nx_map_sock_lback(
27226 +                       skb->sk->sk_nx_info, ip_hdr(skb)->saddr);
27227                 memset(sin->sin_zero, 0, sizeof(sin->sin_zero));
27228         }
27229         if (inet->cmsg_flags)
27230 @@ -1970,6 +1976,8 @@ static struct sock *udp_get_first(struct
27231                 sk_nulls_for_each(sk, node, &hslot->head) {
27232                         if (!net_eq(sock_net(sk), net))
27233                                 continue;
27234 +                       if (!nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT))
27235 +                               continue;
27236                         if (sk->sk_family == state->family)
27237                                 goto found;
27238                 }
27239 @@ -1987,7 +1995,9 @@ static struct sock *udp_get_next(struct
27240
27241         do {
27242                 sk = sk_nulls_next(sk);
27243 -       } while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
27244 +       } while (sk && (!net_eq(sock_net(sk), net) ||
27245 +               sk->sk_family != state->family ||
27246 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)));
27247
27248         if (!sk) {
27249                 if (state->bucket <= state->udp_table->mask)
27250 diff -NurpP --minimal linux-3.0/net/ipv6/Kconfig linux-3.0-vs2.3.1-pre7/net/ipv6/Kconfig
27251 --- linux-3.0/net/ipv6/Kconfig  2010-08-02 16:52:59.000000000 +0200
27252 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/Kconfig     2011-06-10 22:11:24.000000000 +0200
27253 @@ -4,8 +4,8 @@
27254
27255  #   IPv6 as module will cause a CRASH if you try to unload it
27256  menuconfig IPV6
27257 -       tristate "The IPv6 protocol"
27258 -       default m
27259 +       bool "The IPv6 protocol"
27260 +       default n
27261         ---help---
27262           This is complemental support for the IP version 6.
27263           You will still be able to do traditional IPv4 networking as well.
27264 diff -NurpP --minimal linux-3.0/net/ipv6/addrconf.c linux-3.0-vs2.3.1-pre7/net/ipv6/addrconf.c
27265 --- linux-3.0/net/ipv6/addrconf.c       2011-07-22 11:18:13.000000000 +0200
27266 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/addrconf.c  2011-06-10 22:11:24.000000000 +0200
27267 @@ -87,6 +87,8 @@
27268
27269  #include <linux/proc_fs.h>
27270  #include <linux/seq_file.h>
27271 +#include <linux/vs_network.h>
27272 +#include <linux/vs_inet6.h>
27273
27274  /* Set to 3 to get tracing... */
27275  #define ACONF_DEBUG 2
27276 @@ -1108,7 +1110,7 @@ out:
27277
27278  int ipv6_dev_get_saddr(struct net *net, struct net_device *dst_dev,
27279                        const struct in6_addr *daddr, unsigned int prefs,
27280 -                      struct in6_addr *saddr)
27281 +                      struct in6_addr *saddr, struct nx_info *nxi)
27282  {
27283         struct ipv6_saddr_score scores[2],
27284                                 *score = &scores[0], *hiscore = &scores[1];
27285 @@ -1180,6 +1182,8 @@ int ipv6_dev_get_saddr(struct net *net,
27286                                                dev->name);
27287                                 continue;
27288                         }
27289 +                       if (!v6_addr_in_nx_info(nxi, &score->ifa->addr, -1))
27290 +                               continue;
27291
27292                         score->rule = -1;
27293                         bitmap_zero(score->scorebits, IPV6_SADDR_RULE_MAX);
27294 @@ -3048,7 +3052,10 @@ static void if6_seq_stop(struct seq_file
27295  static int if6_seq_show(struct seq_file *seq, void *v)
27296  {
27297         struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
27298 -       seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
27299 +
27300 +       if (nx_check(0, VS_ADMIN|VS_WATCH) ||
27301 +           v6_addr_in_nx_info(current_nx_info(), &ifp->addr, -1))
27302 +               seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
27303                    &ifp->addr,
27304                    ifp->idev->dev->ifindex,
27305                    ifp->prefix_len,
27306 @@ -3554,6 +3561,11 @@ static int in6_dump_addrs(struct inet6_d
27307         struct ifacaddr6 *ifaca;
27308         int err = 1;
27309         int ip_idx = *p_ip_idx;
27310 +       struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
27311 +
27312 +       /* disable ipv6 on non v6 guests */
27313 +       if (nxi && !nx_info_has_v6(nxi))
27314 +               return skb->len;
27315
27316         read_lock_bh(&idev->lock);
27317         switch (type) {
27318 @@ -3564,6 +3576,8 @@ static int in6_dump_addrs(struct inet6_d
27319                 list_for_each_entry(ifa, &idev->addr_list, if_list) {
27320                         if (++ip_idx < s_ip_idx)
27321                                 continue;
27322 +                               if (!v6_addr_in_nx_info(nxi, &ifa->addr, -1))
27323 +                                       continue;
27324                         err = inet6_fill_ifaddr(skb, ifa,
27325                                                 NETLINK_CB(cb->skb).pid,
27326                                                 cb->nlh->nlmsg_seq,
27327 @@ -3580,6 +3594,8 @@ static int in6_dump_addrs(struct inet6_d
27328                      ifmca = ifmca->next, ip_idx++) {
27329                         if (ip_idx < s_ip_idx)
27330                                 continue;
27331 +                               if (!v6_addr_in_nx_info(nxi, &ifmca->mca_addr, -1))
27332 +                                       continue;
27333                         err = inet6_fill_ifmcaddr(skb, ifmca,
27334                                                   NETLINK_CB(cb->skb).pid,
27335                                                   cb->nlh->nlmsg_seq,
27336 @@ -3595,6 +3611,8 @@ static int in6_dump_addrs(struct inet6_d
27337                      ifaca = ifaca->aca_next, ip_idx++) {
27338                         if (ip_idx < s_ip_idx)
27339                                 continue;
27340 +                               if (!v6_addr_in_nx_info(nxi, &ifaca->aca_addr, -1))
27341 +                                       continue;
27342                         err = inet6_fill_ifacaddr(skb, ifaca,
27343                                                   NETLINK_CB(cb->skb).pid,
27344                                                   cb->nlh->nlmsg_seq,
27345 @@ -3980,6 +3998,11 @@ static int inet6_dump_ifinfo(struct sk_b
27346         struct inet6_dev *idev;
27347         struct hlist_head *head;
27348         struct hlist_node *node;
27349 +       struct nx_info *nxi = skb->sk ? skb->sk->sk_nx_info : NULL;
27350 +
27351 +       /* FIXME: maybe disable ipv6 on non v6 guests?
27352 +       if (skb->sk && skb->sk->sk_vx_info)
27353 +               return skb->len; */
27354
27355         s_h = cb->args[0];
27356         s_idx = cb->args[1];
27357 @@ -3991,6 +4014,8 @@ static int inet6_dump_ifinfo(struct sk_b
27358                 hlist_for_each_entry_rcu(dev, node, head, index_hlist) {
27359                         if (idx < s_idx)
27360                                 goto cont;
27361 +                       if (!v6_dev_in_nx_info(dev, nxi))
27362 +                               goto cont;
27363                         idev = __in6_dev_get(dev);
27364                         if (!idev)
27365                                 goto cont;
27366 diff -NurpP --minimal linux-3.0/net/ipv6/af_inet6.c linux-3.0-vs2.3.1-pre7/net/ipv6/af_inet6.c
27367 --- linux-3.0/net/ipv6/af_inet6.c       2011-07-22 11:18:13.000000000 +0200
27368 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/af_inet6.c  2011-07-19 00:44:39.000000000 +0200
27369 @@ -42,6 +42,8 @@
27370  #include <linux/netdevice.h>
27371  #include <linux/icmpv6.h>
27372  #include <linux/netfilter_ipv6.h>
27373 +#include <linux/vs_inet.h>
27374 +#include <linux/vs_inet6.h>
27375
27376  #include <net/ip.h>
27377  #include <net/ipv6.h>
27378 @@ -160,9 +162,12 @@ lookup_protocol:
27379         }
27380
27381         err = -EPERM;
27382 +       if ((protocol == IPPROTO_ICMPV6) &&
27383 +               nx_capable(CAP_NET_RAW, NXC_RAW_ICMP))
27384 +               goto override;
27385         if (sock->type == SOCK_RAW && !kern && !capable(CAP_NET_RAW))
27386                 goto out_rcu_unlock;
27387 -
27388 +override:
27389         sock->ops = answer->ops;
27390         answer_prot = answer->prot;
27391         answer_no_check = answer->no_check;
27392 @@ -261,6 +266,7 @@ int inet6_bind(struct socket *sock, stru
27393         struct inet_sock *inet = inet_sk(sk);
27394         struct ipv6_pinfo *np = inet6_sk(sk);
27395         struct net *net = sock_net(sk);
27396 +       struct nx_v6_sock_addr nsa;
27397         __be32 v4addr = 0;
27398         unsigned short snum;
27399         int addr_type = 0;
27400 @@ -276,6 +282,10 @@ int inet6_bind(struct socket *sock, stru
27401         if (addr->sin6_family != AF_INET6)
27402                 return -EAFNOSUPPORT;
27403
27404 +       err = v6_map_sock_addr(inet, addr, &nsa);
27405 +       if (err)
27406 +               return err;
27407 +
27408         addr_type = ipv6_addr_type(&addr->sin6_addr);
27409         if ((addr_type & IPV6_ADDR_MULTICAST) && sock->type == SOCK_STREAM)
27410                 return -EINVAL;
27411 @@ -307,6 +317,7 @@ int inet6_bind(struct socket *sock, stru
27412                 /* Reproduce AF_INET checks to make the bindings consistent */
27413                 v4addr = addr->sin6_addr.s6_addr32[3];
27414                 chk_addr_ret = inet_addr_type(net, v4addr);
27415 +
27416                 if (!sysctl_ip_nonlocal_bind &&
27417                     !(inet->freebind || inet->transparent) &&
27418                     v4addr != htonl(INADDR_ANY) &&
27419 @@ -316,6 +327,10 @@ int inet6_bind(struct socket *sock, stru
27420                         err = -EADDRNOTAVAIL;
27421                         goto out;
27422                 }
27423 +               if (!v4_addr_in_nx_info(sk->sk_nx_info, v4addr, NXA_MASK_BIND)) {
27424 +                       err = -EADDRNOTAVAIL;
27425 +                       goto out;
27426 +               }
27427         } else {
27428                 if (addr_type != IPV6_ADDR_ANY) {
27429                         struct net_device *dev = NULL;
27430 @@ -342,6 +357,11 @@ int inet6_bind(struct socket *sock, stru
27431                                 }
27432                         }
27433
27434 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
27435 +                               err = -EADDRNOTAVAIL;
27436 +                               goto out;
27437 +                       }
27438 +
27439                         /* ipv4 addr of the socket is invalid.  Only the
27440                          * unspecified and mapped address have a v4 equivalent.
27441                          */
27442 @@ -358,6 +378,9 @@ int inet6_bind(struct socket *sock, stru
27443                 }
27444         }
27445
27446 +       /* what's that for? */
27447 +       v6_set_sock_addr(inet, &nsa);
27448 +
27449         inet->inet_rcv_saddr = v4addr;
27450         inet->inet_saddr = v4addr;
27451
27452 @@ -459,9 +482,11 @@ int inet6_getname(struct socket *sock, s
27453                         return -ENOTCONN;
27454                 sin->sin6_port = inet->inet_dport;
27455                 ipv6_addr_copy(&sin->sin6_addr, &np->daddr);
27456 +               /* FIXME: remap lback? */
27457                 if (np->sndflow)
27458                         sin->sin6_flowinfo = np->flow_label;
27459         } else {
27460 +               /* FIXME: remap lback? */
27461                 if (ipv6_addr_any(&np->rcv_saddr))
27462                         ipv6_addr_copy(&sin->sin6_addr, &np->saddr);
27463                 else
27464 diff -NurpP --minimal linux-3.0/net/ipv6/fib6_rules.c linux-3.0-vs2.3.1-pre7/net/ipv6/fib6_rules.c
27465 --- linux-3.0/net/ipv6/fib6_rules.c     2011-05-22 16:18:00.000000000 +0200
27466 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/fib6_rules.c        2011-06-10 22:11:24.000000000 +0200
27467 @@ -90,7 +90,7 @@ static int fib6_rule_action(struct fib_r
27468                                                ip6_dst_idev(&rt->dst)->dev,
27469                                                &flp6->daddr,
27470                                                rt6_flags2srcprefs(flags),
27471 -                                              &saddr))
27472 +                                              &saddr, NULL))
27473                                 goto again;
27474                         if (!ipv6_prefix_equal(&saddr, &r->src.addr,
27475                                                r->src.plen))
27476 diff -NurpP --minimal linux-3.0/net/ipv6/inet6_hashtables.c linux-3.0-vs2.3.1-pre7/net/ipv6/inet6_hashtables.c
27477 --- linux-3.0/net/ipv6/inet6_hashtables.c       2011-05-22 16:18:00.000000000 +0200
27478 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/inet6_hashtables.c  2011-06-10 22:11:24.000000000 +0200
27479 @@ -16,6 +16,7 @@
27480
27481  #include <linux/module.h>
27482  #include <linux/random.h>
27483 +#include <linux/vs_inet6.h>
27484
27485  #include <net/inet_connection_sock.h>
27486  #include <net/inet_hashtables.h>
27487 @@ -82,7 +83,6 @@ struct sock *__inet6_lookup_established(
27488         unsigned int slot = hash & hashinfo->ehash_mask;
27489         struct inet_ehash_bucket *head = &hashinfo->ehash[slot];
27490
27491 -
27492         rcu_read_lock();
27493  begin:
27494         sk_nulls_for_each_rcu(sk, node, &head->chain) {
27495 @@ -94,7 +94,7 @@ begin:
27496                                 sock_put(sk);
27497                                 goto begin;
27498                         }
27499 -               goto out;
27500 +                       goto out;
27501                 }
27502         }
27503         if (get_nulls_value(node) != slot)
27504 @@ -140,6 +140,9 @@ static inline int compute_score(struct s
27505                         if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
27506                                 return -1;
27507                         score++;
27508 +               } else {
27509 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
27510 +                               return -1;
27511                 }
27512                 if (sk->sk_bound_dev_if) {
27513                         if (sk->sk_bound_dev_if != dif)
27514 diff -NurpP --minimal linux-3.0/net/ipv6/ip6_output.c linux-3.0-vs2.3.1-pre7/net/ipv6/ip6_output.c
27515 --- linux-3.0/net/ipv6/ip6_output.c     2011-07-22 11:18:13.000000000 +0200
27516 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/ip6_output.c        2011-06-16 14:16:51.000000000 +0200
27517 @@ -933,7 +933,8 @@ static int ip6_dst_lookup_tail(struct so
27518                 struct rt6_info *rt = (struct rt6_info *) *dst;
27519                 err = ip6_route_get_saddr(net, rt, &fl6->daddr,
27520                                           sk ? inet6_sk(sk)->srcprefs : 0,
27521 -                                         &fl6->saddr);
27522 +                                         &fl6->saddr,
27523 +                                         sk ? sk->sk_nx_info : NULL);
27524                 if (err)
27525                         goto out_err_release;
27526         }
27527 diff -NurpP --minimal linux-3.0/net/ipv6/ndisc.c linux-3.0-vs2.3.1-pre7/net/ipv6/ndisc.c
27528 --- linux-3.0/net/ipv6/ndisc.c  2011-07-22 11:18:13.000000000 +0200
27529 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/ndisc.c     2011-06-10 22:11:24.000000000 +0200
27530 @@ -597,7 +597,7 @@ static void ndisc_send_na(struct net_dev
27531         } else {
27532                 if (ipv6_dev_get_saddr(dev_net(dev), dev, daddr,
27533                                        inet6_sk(dev_net(dev)->ipv6.ndisc_sk)->srcprefs,
27534 -                                      &tmpaddr))
27535 +                                      &tmpaddr, NULL))
27536                         return;
27537                 src_addr = &tmpaddr;
27538         }
27539 diff -NurpP --minimal linux-3.0/net/ipv6/raw.c linux-3.0-vs2.3.1-pre7/net/ipv6/raw.c
27540 --- linux-3.0/net/ipv6/raw.c    2011-07-22 11:18:13.000000000 +0200
27541 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/raw.c       2011-06-10 22:11:24.000000000 +0200
27542 @@ -30,6 +30,7 @@
27543  #include <linux/icmpv6.h>
27544  #include <linux/netfilter.h>
27545  #include <linux/netfilter_ipv6.h>
27546 +#include <linux/vs_inet6.h>
27547  #include <linux/skbuff.h>
27548  #include <linux/compat.h>
27549  #include <asm/uaccess.h>
27550 @@ -284,6 +285,13 @@ static int rawv6_bind(struct sock *sk, s
27551                                 goto out_unlock;
27552                 }
27553
27554 +               if (!v6_addr_in_nx_info(sk->sk_nx_info, &addr->sin6_addr, -1)) {
27555 +                       err = -EADDRNOTAVAIL;
27556 +                       if (dev)
27557 +                               dev_put(dev);
27558 +                       goto out;
27559 +               }
27560 +
27561                 /* ipv4 addr of the socket is invalid.  Only the
27562                  * unspecified and mapped address have a v4 equivalent.
27563                  */
27564 diff -NurpP --minimal linux-3.0/net/ipv6/route.c linux-3.0-vs2.3.1-pre7/net/ipv6/route.c
27565 --- linux-3.0/net/ipv6/route.c  2011-07-22 11:18:13.000000000 +0200
27566 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/route.c     2011-07-19 00:44:39.000000000 +0200
27567 @@ -54,6 +54,7 @@
27568  #include <net/xfrm.h>
27569  #include <net/netevent.h>
27570  #include <net/netlink.h>
27571 +#include <linux/vs_inet6.h>
27572
27573  #include <asm/uaccess.h>
27574
27575 @@ -2050,15 +2051,17 @@ int ip6_route_get_saddr(struct net *net,
27576                         struct rt6_info *rt,
27577                         const struct in6_addr *daddr,
27578                         unsigned int prefs,
27579 -                       struct in6_addr *saddr)
27580 +                       struct in6_addr *saddr,
27581 +                       struct nx_info *nxi)
27582  {
27583         struct inet6_dev *idev = ip6_dst_idev((struct dst_entry*)rt);
27584         int err = 0;
27585 -       if (rt->rt6i_prefsrc.plen)
27586 +       if (rt->rt6i_prefsrc.plen && (!nxi ||
27587 +           v6_addr_in_nx_info(nxi, &rt->rt6i_prefsrc.addr, NXA_TYPE_ADDR)))
27588                 ipv6_addr_copy(saddr, &rt->rt6i_prefsrc.addr);
27589         else
27590                 err = ipv6_dev_get_saddr(net, idev ? idev->dev : NULL,
27591 -                                        daddr, prefs, saddr);
27592 +                                        daddr, prefs, saddr, nxi);
27593         return err;
27594  }
27595
27596 @@ -2387,7 +2390,8 @@ static int rt6_fill_node(struct net *net
27597                         NLA_PUT_U32(skb, RTA_IIF, iif);
27598         } else if (dst) {
27599                 struct in6_addr saddr_buf;
27600 -               if (ip6_route_get_saddr(net, rt, dst, 0, &saddr_buf) == 0)
27601 +               if (ip6_route_get_saddr(net, rt, dst, 0, &saddr_buf,
27602 +                                       (skb->sk ? skb->sk->sk_nx_info : NULL)) == 0)
27603                         NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
27604         }
27605
27606 diff -NurpP --minimal linux-3.0/net/ipv6/tcp_ipv6.c linux-3.0-vs2.3.1-pre7/net/ipv6/tcp_ipv6.c
27607 --- linux-3.0/net/ipv6/tcp_ipv6.c       2011-07-22 11:18:13.000000000 +0200
27608 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/tcp_ipv6.c  2011-06-22 12:39:16.000000000 +0200
27609 @@ -69,6 +69,7 @@
27610
27611  #include <linux/crypto.h>
27612  #include <linux/scatterlist.h>
27613 +#include <linux/vs_inet6.h>
27614
27615  static void    tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb);
27616  static void    tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
27617 @@ -161,8 +162,15 @@ static int tcp_v6_connect(struct sock *s
27618          *      connect() to INADDR_ANY means loopback (BSD'ism).
27619          */
27620
27621 -       if(ipv6_addr_any(&usin->sin6_addr))
27622 -               usin->sin6_addr.s6_addr[15] = 0x1;
27623 +       if(ipv6_addr_any(&usin->sin6_addr)) {
27624 +               struct nx_info *nxi =  sk->sk_nx_info;
27625 +
27626 +               if (nxi && nx_info_has_v6(nxi))
27627 +                       /* FIXME: remap lback? */
27628 +                       usin->sin6_addr = nxi->v6.ip;
27629 +               else
27630 +                       usin->sin6_addr.s6_addr[15] = 0x1;
27631 +       }
27632
27633         addr_type = ipv6_addr_type(&usin->sin6_addr);
27634
27635 diff -NurpP --minimal linux-3.0/net/ipv6/udp.c linux-3.0-vs2.3.1-pre7/net/ipv6/udp.c
27636 --- linux-3.0/net/ipv6/udp.c    2011-07-22 11:18:13.000000000 +0200
27637 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/udp.c       2011-07-19 00:44:39.000000000 +0200
27638 @@ -45,41 +45,67 @@
27639  #include <net/tcp_states.h>
27640  #include <net/ip6_checksum.h>
27641  #include <net/xfrm.h>
27642 +#include <linux/vs_inet6.h>
27643
27644  #include <linux/proc_fs.h>
27645  #include <linux/seq_file.h>
27646  #include "udp_impl.h"
27647
27648 -int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
27649 +int ipv6_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
27650  {
27651 -       const struct in6_addr *sk_rcv_saddr6 = &inet6_sk(sk)->rcv_saddr;
27652 +       const struct in6_addr *sk1_rcv_saddr6 = &inet6_sk(sk1)->rcv_saddr;
27653         const struct in6_addr *sk2_rcv_saddr6 = inet6_rcv_saddr(sk2);
27654 -       __be32 sk1_rcv_saddr = sk_rcv_saddr(sk);
27655 +       __be32 sk1_rcv_saddr = sk_rcv_saddr(sk1);
27656         __be32 sk2_rcv_saddr = sk_rcv_saddr(sk2);
27657 -       int sk_ipv6only = ipv6_only_sock(sk);
27658 +       int sk1_ipv6only = ipv6_only_sock(sk1);
27659         int sk2_ipv6only = inet_v6_ipv6only(sk2);
27660 -       int addr_type = ipv6_addr_type(sk_rcv_saddr6);
27661 +       int addr_type = ipv6_addr_type(sk1_rcv_saddr6);
27662         int addr_type2 = sk2_rcv_saddr6 ? ipv6_addr_type(sk2_rcv_saddr6) : IPV6_ADDR_MAPPED;
27663
27664         /* if both are mapped, treat as IPv4 */
27665 -       if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED)
27666 -               return (!sk2_ipv6only &&
27667 +       if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED) {
27668 +               if (!sk2_ipv6only &&
27669                         (!sk1_rcv_saddr || !sk2_rcv_saddr ||
27670 -                         sk1_rcv_saddr == sk2_rcv_saddr));
27671 +                         sk1_rcv_saddr == sk2_rcv_saddr))
27672 +                       goto vs_v4;
27673 +               else
27674 +                       return 0;
27675 +       }
27676
27677         if (addr_type2 == IPV6_ADDR_ANY &&
27678             !(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED))
27679 -               return 1;
27680 +               goto vs;
27681
27682         if (addr_type == IPV6_ADDR_ANY &&
27683 -           !(sk_ipv6only && addr_type2 == IPV6_ADDR_MAPPED))
27684 -               return 1;
27685 +           !(sk1_ipv6only && addr_type2 == IPV6_ADDR_MAPPED))
27686 +               goto vs;
27687
27688         if (sk2_rcv_saddr6 &&
27689 -           ipv6_addr_equal(sk_rcv_saddr6, sk2_rcv_saddr6))
27690 -               return 1;
27691 +           ipv6_addr_equal(sk1_rcv_saddr6, sk2_rcv_saddr6))
27692 +               goto vs;
27693
27694         return 0;
27695 +
27696 +vs_v4:
27697 +       if (!sk1_rcv_saddr && !sk2_rcv_saddr)
27698 +               return nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info);
27699 +       if (!sk2_rcv_saddr)
27700 +               return v4_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr, -1);
27701 +       if (!sk1_rcv_saddr)
27702 +               return v4_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr, -1);
27703 +       return 1;
27704 +vs:
27705 +       if (addr_type2 == IPV6_ADDR_ANY && addr_type == IPV6_ADDR_ANY)
27706 +               return nx_v6_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info);
27707 +       else if (addr_type2 == IPV6_ADDR_ANY)
27708 +               return v6_addr_in_nx_info(sk2->sk_nx_info, sk1_rcv_saddr6, -1);
27709 +       else if (addr_type == IPV6_ADDR_ANY) {
27710 +               if (addr_type2 == IPV6_ADDR_MAPPED)
27711 +                       return nx_v4_addr_conflict(sk1->sk_nx_info, sk2->sk_nx_info);
27712 +               else
27713 +                       return v6_addr_in_nx_info(sk1->sk_nx_info, sk2_rcv_saddr6, -1);
27714 +       }
27715 +       return 1;
27716  }
27717
27718  static unsigned int udp6_portaddr_hash(struct net *net,
27719 @@ -143,6 +169,10 @@ static inline int compute_score(struct s
27720                         if (!ipv6_addr_equal(&np->rcv_saddr, daddr))
27721                                 return -1;
27722                         score++;
27723 +               } else {
27724 +                       /* block non nx_info ips */
27725 +                       if (!v6_addr_in_nx_info(sk->sk_nx_info, daddr, -1))
27726 +                               return -1;
27727                 }
27728                 if (!ipv6_addr_any(&np->daddr)) {
27729                         if (!ipv6_addr_equal(&np->daddr, saddr))
27730 diff -NurpP --minimal linux-3.0/net/ipv6/xfrm6_policy.c linux-3.0-vs2.3.1-pre7/net/ipv6/xfrm6_policy.c
27731 --- linux-3.0/net/ipv6/xfrm6_policy.c   2011-07-22 11:18:13.000000000 +0200
27732 +++ linux-3.0-vs2.3.1-pre7/net/ipv6/xfrm6_policy.c      2011-06-10 22:11:24.000000000 +0200
27733 @@ -63,7 +63,7 @@ static int xfrm6_get_saddr(struct net *n
27734         dev = ip6_dst_idev(dst)->dev;
27735         ipv6_dev_get_saddr(dev_net(dev), dev,
27736                            (struct in6_addr *)&daddr->a6, 0,
27737 -                          (struct in6_addr *)&saddr->a6);
27738 +                          (struct in6_addr *)&saddr->a6, NULL);
27739         dst_release(dst);
27740         return 0;
27741  }
27742 diff -NurpP --minimal linux-3.0/net/netfilter/ipvs/ip_vs_xmit.c linux-3.0-vs2.3.1-pre7/net/netfilter/ipvs/ip_vs_xmit.c
27743 --- linux-3.0/net/netfilter/ipvs/ip_vs_xmit.c   2011-07-22 11:18:13.000000000 +0200
27744 +++ linux-3.0-vs2.3.1-pre7/net/netfilter/ipvs/ip_vs_xmit.c      2011-07-22 15:27:52.000000000 +0200
27745 @@ -226,7 +226,7 @@ __ip_vs_route_output_v6(struct net *net,
27746                 return dst;
27747         if (ipv6_addr_any(&fl6.saddr) &&
27748             ipv6_dev_get_saddr(net, ip6_dst_idev(dst)->dev,
27749 -                              &fl6.daddr, 0, &fl6.saddr) < 0)
27750 +                              &fl6.daddr, 0, &fl6.saddr, NULL) < 0)
27751                 goto out_err;
27752         if (do_xfrm) {
27753                 dst = xfrm_lookup(net, dst, flowi6_to_flowi(&fl6), NULL, 0);
27754 diff -NurpP --minimal linux-3.0/net/netlink/af_netlink.c linux-3.0-vs2.3.1-pre7/net/netlink/af_netlink.c
27755 --- linux-3.0/net/netlink/af_netlink.c  2011-07-22 11:18:13.000000000 +0200
27756 +++ linux-3.0-vs2.3.1-pre7/net/netlink/af_netlink.c     2011-06-10 22:11:24.000000000 +0200
27757 @@ -55,6 +55,9 @@
27758  #include <linux/types.h>
27759  #include <linux/audit.h>
27760  #include <linux/mutex.h>
27761 +#include <linux/vs_context.h>
27762 +#include <linux/vs_network.h>
27763 +#include <linux/vs_limit.h>
27764
27765  #include <net/net_namespace.h>
27766  #include <net/sock.h>
27767 @@ -1907,6 +1910,8 @@ static struct sock *netlink_seq_socket_i
27768                         sk_for_each(s, node, &hash->table[j]) {
27769                                 if (sock_net(s) != seq_file_net(seq))
27770                                         continue;
27771 +                               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
27772 +                                       continue;
27773                                 if (off == pos) {
27774                                         iter->link = i;
27775                                         iter->hash_idx = j;
27776 @@ -1941,7 +1946,8 @@ static void *netlink_seq_next(struct seq
27777         s = v;
27778         do {
27779                 s = sk_next(s);
27780 -       } while (s && sock_net(s) != seq_file_net(seq));
27781 +       } while (s && (sock_net(s) != seq_file_net(seq) ||
27782 +               !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)));
27783         if (s)
27784                 return s;
27785
27786 @@ -1953,7 +1959,8 @@ static void *netlink_seq_next(struct seq
27787
27788                 for (; j <= hash->mask; j++) {
27789                         s = sk_head(&hash->table[j]);
27790 -                       while (s && sock_net(s) != seq_file_net(seq))
27791 +                       while (s && (sock_net(s) != seq_file_net(seq) ||
27792 +                               !nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT)))
27793                                 s = sk_next(s);
27794                         if (s) {
27795                                 iter->link = i;
27796 diff -NurpP --minimal linux-3.0/net/socket.c linux-3.0-vs2.3.1-pre7/net/socket.c
27797 --- linux-3.0/net/socket.c      2011-07-22 11:18:13.000000000 +0200
27798 +++ linux-3.0-vs2.3.1-pre7/net/socket.c 2011-06-16 14:16:51.000000000 +0200
27799 @@ -98,6 +98,10 @@
27800
27801  #include <net/sock.h>
27802  #include <linux/netfilter.h>
27803 +#include <linux/vs_base.h>
27804 +#include <linux/vs_socket.h>
27805 +#include <linux/vs_inet.h>
27806 +#include <linux/vs_inet6.h>
27807
27808  #include <linux/if_tun.h>
27809  #include <linux/ipv6_route.h>
27810 @@ -546,6 +550,7 @@ static inline int __sock_sendmsg_nosec(s
27811                                        struct msghdr *msg, size_t size)
27812  {
27813         struct sock_iocb *si = kiocb_to_siocb(iocb);
27814 +       size_t len;
27815
27816         sock_update_classid(sock->sk);
27817
27818 @@ -554,7 +559,22 @@ static inline int __sock_sendmsg_nosec(s
27819         si->msg = msg;
27820         si->size = size;
27821
27822 -       return sock->ops->sendmsg(iocb, sock, msg, size);
27823 +       len = sock->ops->sendmsg(iocb, sock, msg, size);
27824 +       if (sock->sk) {
27825 +               if (len == size)
27826 +                       vx_sock_send(sock->sk, size);
27827 +               else
27828 +                       vx_sock_fail(sock->sk, size);
27829 +       }
27830 +       vxdprintk(VXD_CBIT(net, 7),
27831 +               "__sock_sendmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
27832 +               sock, sock->sk,
27833 +               (sock->sk)?sock->sk->sk_nx_info:0,
27834 +               (sock->sk)?sock->sk->sk_vx_info:0,
27835 +               (sock->sk)?sock->sk->sk_xid:0,
27836 +               (sock->sk)?sock->sk->sk_nid:0,
27837 +               (unsigned int)size, len);
27838 +       return len;
27839  }
27840
27841  static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
27842 @@ -694,6 +714,7 @@ static inline int __sock_recvmsg_nosec(s
27843                                        struct msghdr *msg, size_t size, int flags)
27844  {
27845         struct sock_iocb *si = kiocb_to_siocb(iocb);
27846 +       int len;
27847
27848         sock_update_classid(sock->sk);
27849
27850 @@ -703,7 +724,18 @@ static inline int __sock_recvmsg_nosec(s
27851         si->size = size;
27852         si->flags = flags;
27853
27854 -       return sock->ops->recvmsg(iocb, sock, msg, size, flags);
27855 +       len = sock->ops->recvmsg(iocb, sock, msg, size, flags);
27856 +       if ((len >= 0) && sock->sk)
27857 +               vx_sock_recv(sock->sk, len);
27858 +       vxdprintk(VXD_CBIT(net, 7),
27859 +               "__sock_recvmsg: %p[%p,%p,%p;%d/%d]:%d/%d",
27860 +               sock, sock->sk,
27861 +               (sock->sk)?sock->sk->sk_nx_info:0,
27862 +               (sock->sk)?sock->sk->sk_vx_info:0,
27863 +               (sock->sk)?sock->sk->sk_xid:0,
27864 +               (sock->sk)?sock->sk->sk_nid:0,
27865 +               (unsigned int)size, len);
27866 +       return len;
27867  }
27868
27869  static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
27870 @@ -1188,6 +1220,13 @@ int __sock_create(struct net *net, int f
27871         if (type < 0 || type >= SOCK_MAX)
27872                 return -EINVAL;
27873
27874 +       if (!nx_check(0, VS_ADMIN)) {
27875 +               if (family == PF_INET && !current_nx_info_has_v4())
27876 +                       return -EAFNOSUPPORT;
27877 +               if (family == PF_INET6 && !current_nx_info_has_v6())
27878 +                       return -EAFNOSUPPORT;
27879 +       }
27880 +
27881         /* Compatibility.
27882
27883            This uglymoron is moved from INET layer to here to avoid
27884 @@ -1323,6 +1362,7 @@ SYSCALL_DEFINE3(socket, int, family, int
27885         if (retval < 0)
27886                 goto out;
27887
27888 +       set_bit(SOCK_USER_SOCKET, &sock->flags);
27889         retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
27890         if (retval < 0)
27891                 goto out_release;
27892 @@ -1364,10 +1404,12 @@ SYSCALL_DEFINE4(socketpair, int, family,
27893         err = sock_create(family, type, protocol, &sock1);
27894         if (err < 0)
27895                 goto out;
27896 +       set_bit(SOCK_USER_SOCKET, &sock1->flags);
27897
27898         err = sock_create(family, type, protocol, &sock2);
27899         if (err < 0)
27900                 goto out_release_1;
27901 +       set_bit(SOCK_USER_SOCKET, &sock2->flags);
27902
27903         err = sock1->ops->socketpair(sock1, sock2);
27904         if (err < 0)
27905 diff -NurpP --minimal linux-3.0/net/sunrpc/auth.c linux-3.0-vs2.3.1-pre7/net/sunrpc/auth.c
27906 --- linux-3.0/net/sunrpc/auth.c 2011-07-22 11:18:13.000000000 +0200
27907 +++ linux-3.0-vs2.3.1-pre7/net/sunrpc/auth.c    2011-06-10 22:11:24.000000000 +0200
27908 @@ -14,6 +14,7 @@
27909  #include <linux/hash.h>
27910  #include <linux/sunrpc/clnt.h>
27911  #include <linux/spinlock.h>
27912 +#include <linux/vs_tag.h>
27913
27914  #ifdef RPC_DEBUG
27915  # define RPCDBG_FACILITY       RPCDBG_AUTH
27916 @@ -427,6 +428,7 @@ rpcauth_lookupcred(struct rpc_auth *auth
27917         memset(&acred, 0, sizeof(acred));
27918         acred.uid = cred->fsuid;
27919         acred.gid = cred->fsgid;
27920 +       acred.tag = dx_current_tag();
27921         acred.group_info = get_group_info(((struct cred *)cred)->group_info);
27922
27923         ret = auth->au_ops->lookup_cred(auth, &acred, flags);
27924 @@ -467,6 +469,7 @@ rpcauth_bind_root_cred(struct rpc_task *
27925         struct auth_cred acred = {
27926                 .uid = 0,
27927                 .gid = 0,
27928 +               .tag = dx_current_tag(),
27929         };
27930
27931         dprintk("RPC: %5u looking up %s cred\n",
27932 diff -NurpP --minimal linux-3.0/net/sunrpc/auth_unix.c linux-3.0-vs2.3.1-pre7/net/sunrpc/auth_unix.c
27933 --- linux-3.0/net/sunrpc/auth_unix.c    2010-10-21 13:08:01.000000000 +0200
27934 +++ linux-3.0-vs2.3.1-pre7/net/sunrpc/auth_unix.c       2011-06-10 22:11:24.000000000 +0200
27935 @@ -12,12 +12,14 @@
27936  #include <linux/module.h>
27937  #include <linux/sunrpc/clnt.h>
27938  #include <linux/sunrpc/auth.h>
27939 +#include <linux/vs_tag.h>
27940
27941  #define NFS_NGROUPS    16
27942
27943  struct unx_cred {
27944         struct rpc_cred         uc_base;
27945         gid_t                   uc_gid;
27946 +       tag_t                   uc_tag;
27947         gid_t                   uc_gids[NFS_NGROUPS];
27948  };
27949  #define uc_uid                 uc_base.cr_uid
27950 @@ -78,6 +80,7 @@ unx_create_cred(struct rpc_auth *auth, s
27951                 groups = NFS_NGROUPS;
27952
27953         cred->uc_gid = acred->gid;
27954 +       cred->uc_tag = acred->tag;
27955         for (i = 0; i < groups; i++)
27956                 cred->uc_gids[i] = GROUP_AT(acred->group_info, i);
27957         if (i < NFS_NGROUPS)
27958 @@ -119,7 +122,9 @@ unx_match(struct auth_cred *acred, struc
27959         unsigned int i;
27960
27961
27962 -       if (cred->uc_uid != acred->uid || cred->uc_gid != acred->gid)
27963 +       if (cred->uc_uid != acred->uid ||
27964 +               cred->uc_gid != acred->gid ||
27965 +               cred->uc_tag != acred->tag)
27966                 return 0;
27967
27968         if (acred->group_info != NULL)
27969 @@ -142,7 +147,7 @@ unx_marshal(struct rpc_task *task, __be3
27970         struct rpc_clnt *clnt = task->tk_client;
27971         struct unx_cred *cred = container_of(task->tk_rqstp->rq_cred, struct unx_cred, uc_base);
27972         __be32          *base, *hold;
27973 -       int             i;
27974 +       int             i, tag;
27975
27976         *p++ = htonl(RPC_AUTH_UNIX);
27977         base = p++;
27978 @@ -152,9 +157,12 @@ unx_marshal(struct rpc_task *task, __be3
27979          * Copy the UTS nodename captured when the client was created.
27980          */
27981         p = xdr_encode_array(p, clnt->cl_nodename, clnt->cl_nodelen);
27982 +       tag = task->tk_client->cl_tag;
27983
27984 -       *p++ = htonl((u32) cred->uc_uid);
27985 -       *p++ = htonl((u32) cred->uc_gid);
27986 +       *p++ = htonl((u32) TAGINO_UID(tag,
27987 +               cred->uc_uid, cred->uc_tag));
27988 +       *p++ = htonl((u32) TAGINO_GID(tag,
27989 +               cred->uc_gid, cred->uc_tag));
27990         hold = p++;
27991         for (i = 0; i < 16 && cred->uc_gids[i] != (gid_t) NOGROUP; i++)
27992                 *p++ = htonl((u32) cred->uc_gids[i]);
27993 diff -NurpP --minimal linux-3.0/net/sunrpc/clnt.c linux-3.0-vs2.3.1-pre7/net/sunrpc/clnt.c
27994 --- linux-3.0/net/sunrpc/clnt.c 2011-07-22 11:18:13.000000000 +0200
27995 +++ linux-3.0-vs2.3.1-pre7/net/sunrpc/clnt.c    2011-07-01 11:35:35.000000000 +0200
27996 @@ -31,6 +31,7 @@
27997  #include <linux/in.h>
27998  #include <linux/in6.h>
27999  #include <linux/un.h>
28000 +#include <linux/vs_cvirt.h>
28001
28002  #include <linux/sunrpc/clnt.h>
28003  #include <linux/sunrpc/rpc_pipe_fs.h>
28004 @@ -362,6 +363,9 @@ struct rpc_clnt *rpc_create(struct rpc_c
28005         if (!(args->flags & RPC_CLNT_CREATE_QUIET))
28006                 clnt->cl_chatty = 1;
28007
28008 +       /* TODO: handle RPC_CLNT_CREATE_TAGGED
28009 +       if (args->flags & RPC_CLNT_CREATE_TAGGED)
28010 +               clnt->cl_tag = 1; */
28011         return clnt;
28012  }
28013  EXPORT_SYMBOL_GPL(rpc_create);
28014 diff -NurpP --minimal linux-3.0/net/unix/af_unix.c linux-3.0-vs2.3.1-pre7/net/unix/af_unix.c
28015 --- linux-3.0/net/unix/af_unix.c        2011-07-22 11:18:13.000000000 +0200
28016 +++ linux-3.0-vs2.3.1-pre7/net/unix/af_unix.c   2011-06-10 22:11:24.000000000 +0200
28017 @@ -114,6 +114,8 @@
28018  #include <linux/mount.h>
28019  #include <net/checksum.h>
28020  #include <linux/security.h>
28021 +#include <linux/vs_context.h>
28022 +#include <linux/vs_limit.h>
28023
28024  static struct hlist_head unix_socket_table[UNIX_HASH_SIZE + 1];
28025  static DEFINE_SPINLOCK(unix_table_lock);
28026 @@ -258,6 +260,8 @@ static struct sock *__unix_find_socket_b
28027                 if (!net_eq(sock_net(s), net))
28028                         continue;
28029
28030 +               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
28031 +                       continue;
28032                 if (u->addr->len == len &&
28033                     !memcmp(u->addr->name, sunname, len))
28034                         goto found;
28035 @@ -2208,6 +2212,8 @@ static struct sock *unix_seq_idx(struct
28036         for (s = first_unix_socket(&iter->i); s; s = next_unix_socket(&iter->i, s)) {
28037                 if (sock_net(s) != seq_file_net(seq))
28038                         continue;
28039 +               if (!nx_check(s->sk_nid, VS_WATCH_P | VS_IDENT))
28040 +                       continue;
28041                 if (off == pos)
28042                         return s;
28043                 ++off;
28044 @@ -2232,7 +2238,8 @@ static void *unix_seq_next(struct seq_fi
28045                 sk = first_unix_socket(&iter->i);
28046         else
28047                 sk = next_unix_socket(&iter->i, sk);
28048 -       while (sk && (sock_net(sk) != seq_file_net(seq)))
28049 +       while (sk && (sock_net(sk) != seq_file_net(seq) ||
28050 +               !nx_check(sk->sk_nid, VS_WATCH_P | VS_IDENT)))
28051                 sk = next_unix_socket(&iter->i, sk);
28052         return sk;
28053  }
28054 diff -NurpP --minimal linux-3.0/scripts/checksyscalls.sh linux-3.0-vs2.3.1-pre7/scripts/checksyscalls.sh
28055 --- linux-3.0/scripts/checksyscalls.sh  2011-03-15 18:07:46.000000000 +0100
28056 +++ linux-3.0-vs2.3.1-pre7/scripts/checksyscalls.sh     2011-06-10 22:11:24.000000000 +0200
28057 @@ -193,7 +193,6 @@ cat << EOF
28058  #define __IGNORE_afs_syscall
28059  #define __IGNORE_getpmsg
28060  #define __IGNORE_putpmsg
28061 -#define __IGNORE_vserver
28062  EOF
28063  }
28064
28065 diff -NurpP --minimal linux-3.0/security/commoncap.c linux-3.0-vs2.3.1-pre7/security/commoncap.c
28066 --- linux-3.0/security/commoncap.c      2011-07-22 11:18:14.000000000 +0200
28067 +++ linux-3.0-vs2.3.1-pre7/security/commoncap.c 2011-06-16 14:16:51.000000000 +0200
28068 @@ -62,6 +62,7 @@ int cap_netlink_recv(struct sk_buff *skb
28069                 return -EPERM;
28070         return 0;
28071  }
28072 +
28073  EXPORT_SYMBOL(cap_netlink_recv);
28074
28075  /**
28076 @@ -83,14 +84,20 @@ EXPORT_SYMBOL(cap_netlink_recv);
28077  int cap_capable(struct task_struct *tsk, const struct cred *cred,
28078                 struct user_namespace *targ_ns, int cap, int audit)
28079  {
28080 +       struct vx_info *vxi = tsk->vx_info;
28081 +
28082         for (;;) {
28083                 /* The creator of the user namespace has all caps. */
28084                 if (targ_ns != &init_user_ns && targ_ns->creator == cred->user)
28085                         return 0;
28086
28087                 /* Do we have the necessary capabilities? */
28088 -               if (targ_ns == cred->user->user_ns)
28089 -                       return cap_raised(cred->cap_effective, cap) ? 0 : -EPERM;
28090 +               if (targ_ns == cred->user->user_ns) {
28091 +                       if (vx_info_flags(vxi, VXF_STATE_SETUP, 0) &&
28092 +                           cap_raised(cred->cap_effective, cap))
28093 +                               return 0;
28094 +                       return vx_cap_raised(vxi, cred->cap_effective, cap) ? 0 : -EPERM;
28095 +               }
28096
28097                 /* Have we tried all of the parent namespaces? */
28098                 if (targ_ns == &init_user_ns)
28099 @@ -611,7 +618,7 @@ int cap_inode_setxattr(struct dentry *de
28100
28101         if (!strncmp(name, XATTR_SECURITY_PREFIX,
28102                      sizeof(XATTR_SECURITY_PREFIX) - 1) &&
28103 -           !capable(CAP_SYS_ADMIN))
28104 +               !vx_capable(CAP_SYS_ADMIN, VXC_FS_SECURITY))
28105                 return -EPERM;
28106         return 0;
28107  }
28108 @@ -637,7 +644,7 @@ int cap_inode_removexattr(struct dentry
28109
28110         if (!strncmp(name, XATTR_SECURITY_PREFIX,
28111                      sizeof(XATTR_SECURITY_PREFIX) - 1) &&
28112 -           !capable(CAP_SYS_ADMIN))
28113 +               !vx_capable(CAP_SYS_ADMIN, VXC_FS_SECURITY))
28114                 return -EPERM;
28115         return 0;
28116  }
28117 @@ -971,3 +978,4 @@ int cap_file_mmap(struct file *file, uns
28118         }
28119         return ret;
28120  }
28121 +
28122 diff -NurpP --minimal linux-3.0/security/selinux/hooks.c linux-3.0-vs2.3.1-pre7/security/selinux/hooks.c
28123 --- linux-3.0/security/selinux/hooks.c  2011-07-22 11:18:14.000000000 +0200
28124 +++ linux-3.0-vs2.3.1-pre7/security/selinux/hooks.c     2011-06-15 02:40:14.000000000 +0200
28125 @@ -67,7 +67,6 @@
28126  #include <linux/dccp.h>
28127  #include <linux/quota.h>
28128  #include <linux/un.h>          /* for Unix socket types */
28129 -#include <net/af_unix.h>       /* for Unix socket types */
28130  #include <linux/parser.h>
28131  #include <linux/nfs_mount.h>
28132  #include <net/ipv6.h>