+++ /dev/null
-From f37356d0a41499f9222f9f2b9c0147b500ae4285 Mon Sep 17 00:00:00 2001
-From: John Johansen <john.johansen@canonical.com>
-Date: Tue, 18 Jul 2017 23:04:47 -0700
-Subject: [PATCH 07/17] apparmor: add mount mediation
-
-Add basic mount mediation. That allows controlling based on basic
-mount parameters. It does not include special mount parameters for
-apparmor, super block labeling, or any triggers for apparmor namespace
-parameter modifications on pivot root.
-
-default userspace policy rules have the form of
- MOUNT RULE = ( MOUNT | REMOUNT | UMOUNT )
-
- MOUNT = [ QUALIFIERS ] 'mount' [ MOUNT CONDITIONS ] [ SOURCE FILEGLOB ]
- [ '->' MOUNTPOINT FILEGLOB ]
-
- REMOUNT = [ QUALIFIERS ] 'remount' [ MOUNT CONDITIONS ]
- MOUNTPOINT FILEGLOB
-
- UMOUNT = [ QUALIFIERS ] 'umount' [ MOUNT CONDITIONS ] MOUNTPOINT FILEGLOB
-
- MOUNT CONDITIONS = [ ( 'fstype' | 'vfstype' ) ( '=' | 'in' )
- MOUNT FSTYPE EXPRESSION ]
- [ 'options' ( '=' | 'in' ) MOUNT FLAGS EXPRESSION ]
-
- MOUNT FSTYPE EXPRESSION = ( MOUNT FSTYPE LIST | MOUNT EXPRESSION )
-
- MOUNT FSTYPE LIST = Comma separated list of valid filesystem and
- virtual filesystem types (eg ext4, debugfs, etc)
-
- MOUNT FLAGS EXPRESSION = ( MOUNT FLAGS LIST | MOUNT EXPRESSION )
-
- MOUNT FLAGS LIST = Comma separated list of MOUNT FLAGS.
-
- MOUNT FLAGS = ( 'ro' | 'rw' | 'nosuid' | 'suid' | 'nodev' | 'dev' |
- 'noexec' | 'exec' | 'sync' | 'async' | 'remount' |
- 'mand' | 'nomand' | 'dirsync' | 'noatime' | 'atime' |
- 'nodiratime' | 'diratime' | 'bind' | 'rbind' | 'move' |
- 'verbose' | 'silent' | 'loud' | 'acl' | 'noacl' |
- 'unbindable' | 'runbindable' | 'private' | 'rprivate' |
- 'slave' | 'rslave' | 'shared' | 'rshared' |
- 'relatime' | 'norelatime' | 'iversion' | 'noiversion' |
- 'strictatime' | 'nouser' | 'user' )
-
- MOUNT EXPRESSION = ( ALPHANUMERIC | AARE ) ...
-
- PIVOT ROOT RULE = [ QUALIFIERS ] pivot_root [ oldroot=OLD PUT FILEGLOB ]
- [ NEW ROOT FILEGLOB ]
-
- SOURCE FILEGLOB = FILEGLOB
-
- MOUNTPOINT FILEGLOB = FILEGLOB
-
-eg.
- mount,
- mount /dev/foo,
- mount options=ro /dev/foo -> /mnt/,
- mount options in (ro,atime) /dev/foo -> /mnt/,
- mount options=ro options=atime,
-
-Signed-off-by: John Johansen <john.johansen@canonical.com>
-Acked-by: Seth Arnold <seth.arnold@canonical.com>
-(cherry picked from commit fa488437d0f95b2e5db1e624341fe0d5a233f729)
----
- security/apparmor/Makefile | 2 +-
- security/apparmor/apparmorfs.c | 8 +-
- security/apparmor/domain.c | 4 +-
- security/apparmor/include/apparmor.h | 1 +
- security/apparmor/include/audit.h | 11 +
- security/apparmor/include/domain.h | 5 +
- security/apparmor/include/mount.h | 54 +++
- security/apparmor/lsm.c | 64 ++++
- security/apparmor/mount.c | 696 +++++++++++++++++++++++++++++++++++
- 9 files changed, 841 insertions(+), 4 deletions(-)
- create mode 100644 security/apparmor/include/mount.h
- create mode 100644 security/apparmor/mount.c
-
-diff --git a/security/apparmor/Makefile b/security/apparmor/Makefile
-index a16b195274de..81a34426d024 100644
---- a/security/apparmor/Makefile
-+++ b/security/apparmor/Makefile
-@@ -4,7 +4,7 @@ obj-$(CONFIG_SECURITY_APPARMOR) += apparmor.o
-
- apparmor-y := apparmorfs.o audit.o capability.o context.o ipc.o lib.o match.o \
- path.o domain.o policy.o policy_unpack.o procattr.o lsm.o \
-- resource.o secid.o file.o policy_ns.o label.o
-+ resource.o secid.o file.o policy_ns.o label.o mount.o
- apparmor-$(CONFIG_SECURITY_APPARMOR_HASH) += crypto.o
-
- clean-files := capability_names.h rlim_names.h
-diff --git a/security/apparmor/apparmorfs.c b/security/apparmor/apparmorfs.c
-index a5f9e1aa51f7..8fa6c898c44b 100644
---- a/security/apparmor/apparmorfs.c
-+++ b/security/apparmor/apparmorfs.c
-@@ -2159,9 +2159,14 @@ static struct aa_sfs_entry aa_sfs_entry_policy[] = {
- { }
- };
-
-+static struct aa_sfs_entry aa_sfs_entry_mount[] = {
-+ AA_SFS_FILE_STRING("mask", "mount umount pivot_root"),
-+ { }
-+};
-+
- static struct aa_sfs_entry aa_sfs_entry_ns[] = {
- AA_SFS_FILE_BOOLEAN("profile", 1),
-- AA_SFS_FILE_BOOLEAN("pivot_root", 1),
-+ AA_SFS_FILE_BOOLEAN("pivot_root", 0),
- { }
- };
-
-@@ -2180,6 +2185,7 @@ static struct aa_sfs_entry aa_sfs_entry_features[] = {
- AA_SFS_DIR("policy", aa_sfs_entry_policy),
- AA_SFS_DIR("domain", aa_sfs_entry_domain),
- AA_SFS_DIR("file", aa_sfs_entry_file),
-+ AA_SFS_DIR("mount", aa_sfs_entry_mount),
- AA_SFS_DIR("namespaces", aa_sfs_entry_ns),
- AA_SFS_FILE_U64("capability", VFS_CAP_FLAGS_MASK),
- AA_SFS_DIR("rlimit", aa_sfs_entry_rlimit),
-diff --git a/security/apparmor/domain.c b/security/apparmor/domain.c
-index d0594446ae3f..ffc8c75a6785 100644
---- a/security/apparmor/domain.c
-+++ b/security/apparmor/domain.c
-@@ -374,8 +374,8 @@ static const char *next_name(int xtype, const char *name)
- *
- * Returns: refcounted label, or NULL on failure (MAYBE NULL)
- */
--static struct aa_label *x_table_lookup(struct aa_profile *profile, u32 xindex,
-- const char **name)
-+struct aa_label *x_table_lookup(struct aa_profile *profile, u32 xindex,
-+ const char **name)
- {
- struct aa_label *label = NULL;
- u32 xtype = xindex & AA_X_TYPE_MASK;
-diff --git a/security/apparmor/include/apparmor.h b/security/apparmor/include/apparmor.h
-index 962a20a75e01..829082c35faa 100644
---- a/security/apparmor/include/apparmor.h
-+++ b/security/apparmor/include/apparmor.h
-@@ -27,6 +27,7 @@
- #define AA_CLASS_NET 4
- #define AA_CLASS_RLIMITS 5
- #define AA_CLASS_DOMAIN 6
-+#define AA_CLASS_MOUNT 7
- #define AA_CLASS_PTRACE 9
- #define AA_CLASS_SIGNAL 10
- #define AA_CLASS_LABEL 16
-diff --git a/security/apparmor/include/audit.h b/security/apparmor/include/audit.h
-index d9a156ae11b9..c3fe1c5ef3bc 100644
---- a/security/apparmor/include/audit.h
-+++ b/security/apparmor/include/audit.h
-@@ -71,6 +71,10 @@ enum audit_type {
- #define OP_FMPROT "file_mprotect"
- #define OP_INHERIT "file_inherit"
-
-+#define OP_PIVOTROOT "pivotroot"
-+#define OP_MOUNT "mount"
-+#define OP_UMOUNT "umount"
-+
- #define OP_CREATE "create"
- #define OP_POST_CREATE "post_create"
- #define OP_BIND "bind"
-@@ -132,6 +136,13 @@ struct apparmor_audit_data {
- int rlim;
- unsigned long max;
- } rlim;
-+ struct {
-+ const char *src_name;
-+ const char *type;
-+ const char *trans;
-+ const char *data;
-+ unsigned long flags;
-+ } mnt;
- };
- };
-
-diff --git a/security/apparmor/include/domain.h b/security/apparmor/include/domain.h
-index bab5810b6e9a..db27403346c5 100644
---- a/security/apparmor/include/domain.h
-+++ b/security/apparmor/include/domain.h
-@@ -15,6 +15,8 @@
- #include <linux/binfmts.h>
- #include <linux/types.h>
-
-+#include "label.h"
-+
- #ifndef __AA_DOMAIN_H
- #define __AA_DOMAIN_H
-
-@@ -29,6 +31,9 @@ struct aa_domain {
- #define AA_CHANGE_ONEXEC 4
- #define AA_CHANGE_STACK 8
-
-+struct aa_label *x_table_lookup(struct aa_profile *profile, u32 xindex,
-+ const char **name);
-+
- int apparmor_bprm_set_creds(struct linux_binprm *bprm);
- int apparmor_bprm_secureexec(struct linux_binprm *bprm);
-
-diff --git a/security/apparmor/include/mount.h b/security/apparmor/include/mount.h
-new file mode 100644
-index 000000000000..25d6067fa6ef
---- /dev/null
-+++ b/security/apparmor/include/mount.h
-@@ -0,0 +1,54 @@
-+/*
-+ * AppArmor security module
-+ *
-+ * This file contains AppArmor file mediation function definitions.
-+ *
-+ * Copyright 2017 Canonical Ltd.
-+ *
-+ * This program is free software; you can redistribute it and/or
-+ * modify it under the terms of the GNU General Public License as
-+ * published by the Free Software Foundation, version 2 of the
-+ * License.
-+ */
-+
-+#ifndef __AA_MOUNT_H
-+#define __AA_MOUNT_H
-+
-+#include <linux/fs.h>
-+#include <linux/path.h>
-+
-+#include "domain.h"
-+#include "policy.h"
-+
-+/* mount perms */
-+#define AA_MAY_PIVOTROOT 0x01
-+#define AA_MAY_MOUNT 0x02
-+#define AA_MAY_UMOUNT 0x04
-+#define AA_AUDIT_DATA 0x40
-+#define AA_MNT_CONT_MATCH 0x40
-+
-+#define AA_MS_IGNORE_MASK (MS_KERNMOUNT | MS_NOSEC | MS_ACTIVE | MS_BORN)
-+
-+int aa_remount(struct aa_label *label, const struct path *path,
-+ unsigned long flags, void *data);
-+
-+int aa_bind_mount(struct aa_label *label, const struct path *path,
-+ const char *old_name, unsigned long flags);
-+
-+
-+int aa_mount_change_type(struct aa_label *label, const struct path *path,
-+ unsigned long flags);
-+
-+int aa_move_mount(struct aa_label *label, const struct path *path,
-+ const char *old_name);
-+
-+int aa_new_mount(struct aa_label *label, const char *dev_name,
-+ const struct path *path, const char *type, unsigned long flags,
-+ void *data);
-+
-+int aa_umount(struct aa_label *label, struct vfsmount *mnt, int flags);
-+
-+int aa_pivotroot(struct aa_label *label, const struct path *old_path,
-+ const struct path *new_path);
-+
-+#endif /* __AA_MOUNT_H */
-diff --git a/security/apparmor/lsm.c b/security/apparmor/lsm.c
-index af22f3dfbcce..4ad0b3a45142 100644
---- a/security/apparmor/lsm.c
-+++ b/security/apparmor/lsm.c
-@@ -38,6 +38,7 @@
- #include "include/policy.h"
- #include "include/policy_ns.h"
- #include "include/procattr.h"
-+#include "include/mount.h"
-
- /* Flag indicating whether initialization completed */
- int apparmor_initialized;
-@@ -511,6 +512,65 @@ static int apparmor_file_mprotect(struct vm_area_struct *vma,
- !(vma->vm_flags & VM_SHARED) ? MAP_PRIVATE : 0);
- }
-
-+static int apparmor_sb_mount(const char *dev_name, const struct path *path,
-+ const char *type, unsigned long flags, void *data)
-+{
-+ struct aa_label *label;
-+ int error = 0;
-+
-+ /* Discard magic */
-+ if ((flags & MS_MGC_MSK) == MS_MGC_VAL)
-+ flags &= ~MS_MGC_MSK;
-+
-+ flags &= ~AA_MS_IGNORE_MASK;
-+
-+ label = __begin_current_label_crit_section();
-+ if (!unconfined(label)) {
-+ if (flags & MS_REMOUNT)
-+ error = aa_remount(label, path, flags, data);
-+ else if (flags & MS_BIND)
-+ error = aa_bind_mount(label, path, dev_name, flags);
-+ else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE |
-+ MS_UNBINDABLE))
-+ error = aa_mount_change_type(label, path, flags);
-+ else if (flags & MS_MOVE)
-+ error = aa_move_mount(label, path, dev_name);
-+ else
-+ error = aa_new_mount(label, dev_name, path, type,
-+ flags, data);
-+ }
-+ __end_current_label_crit_section(label);
-+
-+ return error;
-+}
-+
-+static int apparmor_sb_umount(struct vfsmount *mnt, int flags)
-+{
-+ struct aa_label *label;
-+ int error = 0;
-+
-+ label = __begin_current_label_crit_section();
-+ if (!unconfined(label))
-+ error = aa_umount(label, mnt, flags);
-+ __end_current_label_crit_section(label);
-+
-+ return error;
-+}
-+
-+static int apparmor_sb_pivotroot(const struct path *old_path,
-+ const struct path *new_path)
-+{
-+ struct aa_label *label;
-+ int error = 0;
-+
-+ label = aa_get_current_label();
-+ if (!unconfined(label))
-+ error = aa_pivotroot(label, old_path, new_path);
-+ aa_put_label(label);
-+
-+ return error;
-+}
-+
- static int apparmor_getprocattr(struct task_struct *task, char *name,
- char **value)
- {
-@@ -682,6 +742,10 @@ static struct security_hook_list apparmor_hooks[] __lsm_ro_after_init = {
- LSM_HOOK_INIT(capget, apparmor_capget),
- LSM_HOOK_INIT(capable, apparmor_capable),
-
-+ LSM_HOOK_INIT(sb_mount, apparmor_sb_mount),
-+ LSM_HOOK_INIT(sb_umount, apparmor_sb_umount),
-+ LSM_HOOK_INIT(sb_pivotroot, apparmor_sb_pivotroot),
-+
- LSM_HOOK_INIT(path_link, apparmor_path_link),
- LSM_HOOK_INIT(path_unlink, apparmor_path_unlink),
- LSM_HOOK_INIT(path_symlink, apparmor_path_symlink),
-diff --git a/security/apparmor/mount.c b/security/apparmor/mount.c
-new file mode 100644
-index 000000000000..82a64b58041d
---- /dev/null
-+++ b/security/apparmor/mount.c
-@@ -0,0 +1,696 @@
-+/*
-+ * AppArmor security module
-+ *
-+ * This file contains AppArmor mediation of files
-+ *
-+ * Copyright (C) 1998-2008 Novell/SUSE
-+ * Copyright 2009-2017 Canonical Ltd.
-+ *
-+ * This program is free software; you can redistribute it and/or
-+ * modify it under the terms of the GNU General Public License as
-+ * published by the Free Software Foundation, version 2 of the
-+ * License.
-+ */
-+
-+#include <linux/fs.h>
-+#include <linux/mount.h>
-+#include <linux/namei.h>
-+
-+#include "include/apparmor.h"
-+#include "include/audit.h"
-+#include "include/context.h"
-+#include "include/domain.h"
-+#include "include/file.h"
-+#include "include/match.h"
-+#include "include/mount.h"
-+#include "include/path.h"
-+#include "include/policy.h"
-+
-+
-+static void audit_mnt_flags(struct audit_buffer *ab, unsigned long flags)
-+{
-+ if (flags & MS_RDONLY)
-+ audit_log_format(ab, "ro");
-+ else
-+ audit_log_format(ab, "rw");
-+ if (flags & MS_NOSUID)
-+ audit_log_format(ab, ", nosuid");
-+ if (flags & MS_NODEV)
-+ audit_log_format(ab, ", nodev");
-+ if (flags & MS_NOEXEC)
-+ audit_log_format(ab, ", noexec");
-+ if (flags & MS_SYNCHRONOUS)
-+ audit_log_format(ab, ", sync");
-+ if (flags & MS_REMOUNT)
-+ audit_log_format(ab, ", remount");
-+ if (flags & MS_MANDLOCK)
-+ audit_log_format(ab, ", mand");
-+ if (flags & MS_DIRSYNC)
-+ audit_log_format(ab, ", dirsync");
-+ if (flags & MS_NOATIME)
-+ audit_log_format(ab, ", noatime");
-+ if (flags & MS_NODIRATIME)
-+ audit_log_format(ab, ", nodiratime");
-+ if (flags & MS_BIND)
-+ audit_log_format(ab, flags & MS_REC ? ", rbind" : ", bind");
-+ if (flags & MS_MOVE)
-+ audit_log_format(ab, ", move");
-+ if (flags & MS_SILENT)
-+ audit_log_format(ab, ", silent");
-+ if (flags & MS_POSIXACL)
-+ audit_log_format(ab, ", acl");
-+ if (flags & MS_UNBINDABLE)
-+ audit_log_format(ab, flags & MS_REC ? ", runbindable" :
-+ ", unbindable");
-+ if (flags & MS_PRIVATE)
-+ audit_log_format(ab, flags & MS_REC ? ", rprivate" :
-+ ", private");
-+ if (flags & MS_SLAVE)
-+ audit_log_format(ab, flags & MS_REC ? ", rslave" :
-+ ", slave");
-+ if (flags & MS_SHARED)
-+ audit_log_format(ab, flags & MS_REC ? ", rshared" :
-+ ", shared");
-+ if (flags & MS_RELATIME)
-+ audit_log_format(ab, ", relatime");
-+ if (flags & MS_I_VERSION)
-+ audit_log_format(ab, ", iversion");
-+ if (flags & MS_STRICTATIME)
-+ audit_log_format(ab, ", strictatime");
-+ if (flags & MS_NOUSER)
-+ audit_log_format(ab, ", nouser");
-+}
-+
-+/**
-+ * audit_cb - call back for mount specific audit fields
-+ * @ab: audit_buffer (NOT NULL)
-+ * @va: audit struct to audit values of (NOT NULL)
-+ */
-+static void audit_cb(struct audit_buffer *ab, void *va)
-+{
-+ struct common_audit_data *sa = va;
-+
-+ if (aad(sa)->mnt.type) {
-+ audit_log_format(ab, " fstype=");
-+ audit_log_untrustedstring(ab, aad(sa)->mnt.type);
-+ }
-+ if (aad(sa)->mnt.src_name) {
-+ audit_log_format(ab, " srcname=");
-+ audit_log_untrustedstring(ab, aad(sa)->mnt.src_name);
-+ }
-+ if (aad(sa)->mnt.trans) {
-+ audit_log_format(ab, " trans=");
-+ audit_log_untrustedstring(ab, aad(sa)->mnt.trans);
-+ }
-+ if (aad(sa)->mnt.flags) {
-+ audit_log_format(ab, " flags=\"");
-+ audit_mnt_flags(ab, aad(sa)->mnt.flags);
-+ audit_log_format(ab, "\"");
-+ }
-+ if (aad(sa)->mnt.data) {
-+ audit_log_format(ab, " options=");
-+ audit_log_untrustedstring(ab, aad(sa)->mnt.data);
-+ }
-+}
-+
-+/**
-+ * audit_mount - handle the auditing of mount operations
-+ * @profile: the profile being enforced (NOT NULL)
-+ * @op: operation being mediated (NOT NULL)
-+ * @name: name of object being mediated (MAYBE NULL)
-+ * @src_name: src_name of object being mediated (MAYBE_NULL)
-+ * @type: type of filesystem (MAYBE_NULL)
-+ * @trans: name of trans (MAYBE NULL)
-+ * @flags: filesystem idependent mount flags
-+ * @data: filesystem mount flags
-+ * @request: permissions requested
-+ * @perms: the permissions computed for the request (NOT NULL)
-+ * @info: extra information message (MAYBE NULL)
-+ * @error: 0 if operation allowed else failure error code
-+ *
-+ * Returns: %0 or error on failure
-+ */
-+static int audit_mount(struct aa_profile *profile, const char *op,
-+ const char *name, const char *src_name,
-+ const char *type, const char *trans,
-+ unsigned long flags, const void *data, u32 request,
-+ struct aa_perms *perms, const char *info, int error)
-+{
-+ int audit_type = AUDIT_APPARMOR_AUTO;
-+ DEFINE_AUDIT_DATA(sa, LSM_AUDIT_DATA_NONE, op);
-+
-+ if (likely(!error)) {
-+ u32 mask = perms->audit;
-+
-+ if (unlikely(AUDIT_MODE(profile) == AUDIT_ALL))
-+ mask = 0xffff;
-+
-+ /* mask off perms that are not being force audited */
-+ request &= mask;
-+
-+ if (likely(!request))
-+ return 0;
-+ audit_type = AUDIT_APPARMOR_AUDIT;
-+ } else {
-+ /* only report permissions that were denied */
-+ request = request & ~perms->allow;
-+
-+ if (request & perms->kill)
-+ audit_type = AUDIT_APPARMOR_KILL;
-+
-+ /* quiet known rejects, assumes quiet and kill do not overlap */
-+ if ((request & perms->quiet) &&
-+ AUDIT_MODE(profile) != AUDIT_NOQUIET &&
-+ AUDIT_MODE(profile) != AUDIT_ALL)
-+ request &= ~perms->quiet;
-+
-+ if (!request)
-+ return error;
-+ }
-+
-+ aad(&sa)->name = name;
-+ aad(&sa)->mnt.src_name = src_name;
-+ aad(&sa)->mnt.type = type;
-+ aad(&sa)->mnt.trans = trans;
-+ aad(&sa)->mnt.flags = flags;
-+ if (data && (perms->audit & AA_AUDIT_DATA))
-+ aad(&sa)->mnt.data = data;
-+ aad(&sa)->info = info;
-+ aad(&sa)->error = error;
-+
-+ return aa_audit(audit_type, profile, &sa, audit_cb);
-+}
-+
-+/**
-+ * match_mnt_flags - Do an ordered match on mount flags
-+ * @dfa: dfa to match against
-+ * @state: state to start in
-+ * @flags: mount flags to match against
-+ *
-+ * Mount flags are encoded as an ordered match. This is done instead of
-+ * checking against a simple bitmask, to allow for logical operations
-+ * on the flags.
-+ *
-+ * Returns: next state after flags match
-+ */
-+static unsigned int match_mnt_flags(struct aa_dfa *dfa, unsigned int state,
-+ unsigned long flags)
-+{
-+ unsigned int i;
-+
-+ for (i = 0; i <= 31 ; ++i) {
-+ if ((1 << i) & flags)
-+ state = aa_dfa_next(dfa, state, i + 1);
-+ }
-+
-+ return state;
-+}
-+
-+/**
-+ * compute_mnt_perms - compute mount permission associated with @state
-+ * @dfa: dfa to match against (NOT NULL)
-+ * @state: state match finished in
-+ *
-+ * Returns: mount permissions
-+ */
-+static struct aa_perms compute_mnt_perms(struct aa_dfa *dfa,
-+ unsigned int state)
-+{
-+ struct aa_perms perms;
-+
-+ perms.kill = 0;
-+ perms.allow = dfa_user_allow(dfa, state);
-+ perms.audit = dfa_user_audit(dfa, state);
-+ perms.quiet = dfa_user_quiet(dfa, state);
-+ perms.xindex = dfa_user_xindex(dfa, state);
-+
-+ return perms;
-+}
-+
-+static const char * const mnt_info_table[] = {
-+ "match succeeded",
-+ "failed mntpnt match",
-+ "failed srcname match",
-+ "failed type match",
-+ "failed flags match",
-+ "failed data match"
-+};
-+
-+/*
-+ * Returns 0 on success else element that match failed in, this is the
-+ * index into the mnt_info_table above
-+ */
-+static int do_match_mnt(struct aa_dfa *dfa, unsigned int start,
-+ const char *mntpnt, const char *devname,
-+ const char *type, unsigned long flags,
-+ void *data, bool binary, struct aa_perms *perms)
-+{
-+ unsigned int state;
-+
-+ AA_BUG(!dfa);
-+ AA_BUG(!perms);
-+
-+ state = aa_dfa_match(dfa, start, mntpnt);
-+ state = aa_dfa_null_transition(dfa, state);
-+ if (!state)
-+ return 1;
-+
-+ if (devname)
-+ state = aa_dfa_match(dfa, state, devname);
-+ state = aa_dfa_null_transition(dfa, state);
-+ if (!state)
-+ return 2;
-+
-+ if (type)
-+ state = aa_dfa_match(dfa, state, type);
-+ state = aa_dfa_null_transition(dfa, state);
-+ if (!state)
-+ return 3;
-+
-+ state = match_mnt_flags(dfa, state, flags);
-+ if (!state)
-+ return 4;
-+ *perms = compute_mnt_perms(dfa, state);
-+ if (perms->allow & AA_MAY_MOUNT)
-+ return 0;
-+
-+ /* only match data if not binary and the DFA flags data is expected */
-+ if (data && !binary && (perms->allow & AA_MNT_CONT_MATCH)) {
-+ state = aa_dfa_null_transition(dfa, state);
-+ if (!state)
-+ return 4;
-+
-+ state = aa_dfa_match(dfa, state, data);
-+ if (!state)
-+ return 5;
-+ *perms = compute_mnt_perms(dfa, state);
-+ if (perms->allow & AA_MAY_MOUNT)
-+ return 0;
-+ }
-+
-+ /* failed at end of flags match */
-+ return 4;
-+}
-+
-+
-+static int path_flags(struct aa_profile *profile, const struct path *path)
-+{
-+ AA_BUG(!profile);
-+ AA_BUG(!path);
-+
-+ return profile->path_flags |
-+ (S_ISDIR(path->dentry->d_inode->i_mode) ? PATH_IS_DIR : 0);
-+}
-+
-+/**
-+ * match_mnt_path_str - handle path matching for mount
-+ * @profile: the confining profile
-+ * @mntpath: for the mntpnt (NOT NULL)
-+ * @buffer: buffer to be used to lookup mntpath
-+ * @devnme: string for the devname/src_name (MAY BE NULL OR ERRPTR)
-+ * @type: string for the dev type (MAYBE NULL)
-+ * @flags: mount flags to match
-+ * @data: fs mount data (MAYBE NULL)
-+ * @binary: whether @data is binary
-+ * @devinfo: error str if (IS_ERR(@devname))
-+ *
-+ * Returns: 0 on success else error
-+ */
-+static int match_mnt_path_str(struct aa_profile *profile,
-+ const struct path *mntpath, char *buffer,
-+ const char *devname, const char *type,
-+ unsigned long flags, void *data, bool binary,
-+ const char *devinfo)
-+{
-+ struct aa_perms perms = { };
-+ const char *mntpnt = NULL, *info = NULL;
-+ int pos, error;
-+
-+ AA_BUG(!profile);
-+ AA_BUG(!mntpath);
-+ AA_BUG(!buffer);
-+
-+ error = aa_path_name(mntpath, path_flags(profile, mntpath), buffer,
-+ &mntpnt, &info, profile->disconnected);
-+ if (error)
-+ goto audit;
-+ if (IS_ERR(devname)) {
-+ error = PTR_ERR(devname);
-+ devname = NULL;
-+ info = devinfo;
-+ goto audit;
-+ }
-+
-+ error = -EACCES;
-+ pos = do_match_mnt(profile->policy.dfa,
-+ profile->policy.start[AA_CLASS_MOUNT],
-+ mntpnt, devname, type, flags, data, binary, &perms);
-+ if (pos) {
-+ info = mnt_info_table[pos];
-+ goto audit;
-+ }
-+ error = 0;
-+
-+audit:
-+ return audit_mount(profile, OP_MOUNT, mntpnt, devname, type, NULL,
-+ flags, data, AA_MAY_MOUNT, &perms, info, error);
-+}
-+
-+/**
-+ * match_mnt - handle path matching for mount
-+ * @profile: the confining profile
-+ * @mntpath: for the mntpnt (NOT NULL)
-+ * @buffer: buffer to be used to lookup mntpath
-+ * @devpath: path devname/src_name (MAYBE NULL)
-+ * @devbuffer: buffer to be used to lookup devname/src_name
-+ * @type: string for the dev type (MAYBE NULL)
-+ * @flags: mount flags to match
-+ * @data: fs mount data (MAYBE NULL)
-+ * @binary: whether @data is binary
-+ *
-+ * Returns: 0 on success else error
-+ */
-+static int match_mnt(struct aa_profile *profile, const struct path *path,
-+ char *buffer, struct path *devpath, char *devbuffer,
-+ const char *type, unsigned long flags, void *data,
-+ bool binary)
-+{
-+ const char *devname = NULL, *info = NULL;
-+ int error = -EACCES;
-+
-+ AA_BUG(!profile);
-+ AA_BUG(devpath && !devbuffer);
-+
-+ if (devpath) {
-+ error = aa_path_name(devpath, path_flags(profile, devpath),
-+ devbuffer, &devname, &info,
-+ profile->disconnected);
-+ if (error)
-+ devname = ERR_PTR(error);
-+ }
-+
-+ return match_mnt_path_str(profile, path, buffer, devname, type, flags,
-+ data, binary, info);
-+}
-+
-+int aa_remount(struct aa_label *label, const struct path *path,
-+ unsigned long flags, void *data)
-+{
-+ struct aa_profile *profile;
-+ char *buffer = NULL;
-+ bool binary;
-+ int error;
-+
-+ AA_BUG(!label);
-+ AA_BUG(!path);
-+
-+ binary = path->dentry->d_sb->s_type->fs_flags & FS_BINARY_MOUNTDATA;
-+
-+ get_buffers(buffer);
-+ error = fn_for_each_confined(label, profile,
-+ match_mnt(profile, path, buffer, NULL, NULL, NULL,
-+ flags, data, binary));
-+ put_buffers(buffer);
-+
-+ return error;
-+}
-+
-+int aa_bind_mount(struct aa_label *label, const struct path *path,
-+ const char *dev_name, unsigned long flags)
-+{
-+ struct aa_profile *profile;
-+ char *buffer = NULL, *old_buffer = NULL;
-+ struct path old_path;
-+ int error;
-+
-+ AA_BUG(!label);
-+ AA_BUG(!path);
-+
-+ if (!dev_name || !*dev_name)
-+ return -EINVAL;
-+
-+ flags &= MS_REC | MS_BIND;
-+
-+ error = kern_path(dev_name, LOOKUP_FOLLOW|LOOKUP_AUTOMOUNT, &old_path);
-+ if (error)
-+ return error;
-+
-+ get_buffers(buffer, old_buffer);
-+ error = fn_for_each_confined(label, profile,
-+ match_mnt(profile, path, buffer, &old_path, old_buffer,
-+ NULL, flags, NULL, false));
-+ put_buffers(buffer, old_buffer);
-+ path_put(&old_path);
-+
-+ return error;
-+}
-+
-+int aa_mount_change_type(struct aa_label *label, const struct path *path,
-+ unsigned long flags)
-+{
-+ struct aa_profile *profile;
-+ char *buffer = NULL;
-+ int error;
-+
-+ AA_BUG(!label);
-+ AA_BUG(!path);
-+
-+ /* These are the flags allowed by do_change_type() */
-+ flags &= (MS_REC | MS_SILENT | MS_SHARED | MS_PRIVATE | MS_SLAVE |
-+ MS_UNBINDABLE);
-+
-+ get_buffers(buffer);
-+ error = fn_for_each_confined(label, profile,
-+ match_mnt(profile, path, buffer, NULL, NULL, NULL,
-+ flags, NULL, false));
-+ put_buffers(buffer);
-+
-+ return error;
-+}
-+
-+int aa_move_mount(struct aa_label *label, const struct path *path,
-+ const char *orig_name)
-+{
-+ struct aa_profile *profile;
-+ char *buffer = NULL, *old_buffer = NULL;
-+ struct path old_path;
-+ int error;
-+
-+ AA_BUG(!label);
-+ AA_BUG(!path);
-+
-+ if (!orig_name || !*orig_name)
-+ return -EINVAL;
-+
-+ error = kern_path(orig_name, LOOKUP_FOLLOW, &old_path);
-+ if (error)
-+ return error;
-+
-+ get_buffers(buffer, old_buffer);
-+ error = fn_for_each_confined(label, profile,
-+ match_mnt(profile, path, buffer, &old_path, old_buffer,
-+ NULL, MS_MOVE, NULL, false));
-+ put_buffers(buffer, old_buffer);
-+ path_put(&old_path);
-+
-+ return error;
-+}
-+
-+int aa_new_mount(struct aa_label *label, const char *dev_name,
-+ const struct path *path, const char *type, unsigned long flags,
-+ void *data)
-+{
-+ struct aa_profile *profile;
-+ char *buffer = NULL, *dev_buffer = NULL;
-+ bool binary = true;
-+ int error;
-+ int requires_dev = 0;
-+ struct path tmp_path, *dev_path = NULL;
-+
-+ AA_BUG(!label);
-+ AA_BUG(!path);
-+
-+ if (type) {
-+ struct file_system_type *fstype;
-+
-+ fstype = get_fs_type(type);
-+ if (!fstype)
-+ return -ENODEV;
-+ binary = fstype->fs_flags & FS_BINARY_MOUNTDATA;
-+ requires_dev = fstype->fs_flags & FS_REQUIRES_DEV;
-+ put_filesystem(fstype);
-+
-+ if (requires_dev) {
-+ if (!dev_name || !*dev_name)
-+ return -ENOENT;
-+
-+ error = kern_path(dev_name, LOOKUP_FOLLOW, &tmp_path);
-+ if (error)
-+ return error;
-+ dev_path = &tmp_path;
-+ }
-+ }
-+
-+ get_buffers(buffer, dev_buffer);
-+ if (dev_path) {
-+ error = fn_for_each_confined(label, profile,
-+ match_mnt(profile, path, buffer, dev_path, dev_buffer,
-+ type, flags, data, binary));
-+ } else {
-+ error = fn_for_each_confined(label, profile,
-+ match_mnt_path_str(profile, path, buffer, dev_name,
-+ type, flags, data, binary, NULL));
-+ }
-+ put_buffers(buffer, dev_buffer);
-+ if (dev_path)
-+ path_put(dev_path);
-+
-+ return error;
-+}
-+
-+static int profile_umount(struct aa_profile *profile, struct path *path,
-+ char *buffer)
-+{
-+ struct aa_perms perms = { };
-+ const char *name = NULL, *info = NULL;
-+ unsigned int state;
-+ int error;
-+
-+ AA_BUG(!profile);
-+ AA_BUG(!path);
-+
-+ error = aa_path_name(path, path_flags(profile, path), buffer, &name,
-+ &info, profile->disconnected);
-+ if (error)
-+ goto audit;
-+
-+ state = aa_dfa_match(profile->policy.dfa,
-+ profile->policy.start[AA_CLASS_MOUNT],
-+ name);
-+ perms = compute_mnt_perms(profile->policy.dfa, state);
-+ if (AA_MAY_UMOUNT & ~perms.allow)
-+ error = -EACCES;
-+
-+audit:
-+ return audit_mount(profile, OP_UMOUNT, name, NULL, NULL, NULL, 0, NULL,
-+ AA_MAY_UMOUNT, &perms, info, error);
-+}
-+
-+int aa_umount(struct aa_label *label, struct vfsmount *mnt, int flags)
-+{
-+ struct aa_profile *profile;
-+ char *buffer = NULL;
-+ int error;
-+ struct path path = { .mnt = mnt, .dentry = mnt->mnt_root };
-+
-+ AA_BUG(!label);
-+ AA_BUG(!mnt);
-+
-+ get_buffers(buffer);
-+ error = fn_for_each_confined(label, profile,
-+ profile_umount(profile, &path, buffer));
-+ put_buffers(buffer);
-+
-+ return error;
-+}
-+
-+/* helper fn for transition on pivotroot
-+ *
-+ * Returns: label for transition or ERR_PTR. Does not return NULL
-+ */
-+static struct aa_label *build_pivotroot(struct aa_profile *profile,
-+ const struct path *new_path,
-+ char *new_buffer,
-+ const struct path *old_path,
-+ char *old_buffer)
-+{
-+ const char *old_name, *new_name = NULL, *info = NULL;
-+ const char *trans_name = NULL;
-+ struct aa_perms perms = { };
-+ unsigned int state;
-+ int error;
-+
-+ AA_BUG(!profile);
-+ AA_BUG(!new_path);
-+ AA_BUG(!old_path);
-+
-+ if (profile_unconfined(profile))
-+ return aa_get_newest_label(&profile->label);
-+
-+ error = aa_path_name(old_path, path_flags(profile, old_path),
-+ old_buffer, &old_name, &info,
-+ profile->disconnected);
-+ if (error)
-+ goto audit;
-+ error = aa_path_name(new_path, path_flags(profile, new_path),
-+ new_buffer, &new_name, &info,
-+ profile->disconnected);
-+ if (error)
-+ goto audit;
-+
-+ error = -EACCES;
-+ state = aa_dfa_match(profile->policy.dfa,
-+ profile->policy.start[AA_CLASS_MOUNT],
-+ new_name);
-+ state = aa_dfa_null_transition(profile->policy.dfa, state);
-+ state = aa_dfa_match(profile->policy.dfa, state, old_name);
-+ perms = compute_mnt_perms(profile->policy.dfa, state);
-+
-+ if (AA_MAY_PIVOTROOT & perms.allow)
-+ error = 0;
-+
-+audit:
-+ error = audit_mount(profile, OP_PIVOTROOT, new_name, old_name,
-+ NULL, trans_name, 0, NULL, AA_MAY_PIVOTROOT,
-+ &perms, info, error);
-+ if (error)
-+ return ERR_PTR(error);
-+
-+ return aa_get_newest_label(&profile->label);
-+}
-+
-+int aa_pivotroot(struct aa_label *label, const struct path *old_path,
-+ const struct path *new_path)
-+{
-+ struct aa_profile *profile;
-+ struct aa_label *target = NULL;
-+ char *old_buffer = NULL, *new_buffer = NULL, *info = NULL;
-+ int error;
-+
-+ AA_BUG(!label);
-+ AA_BUG(!old_path);
-+ AA_BUG(!new_path);
-+
-+ get_buffers(old_buffer, new_buffer);
-+ target = fn_label_build(label, profile, GFP_ATOMIC,
-+ build_pivotroot(profile, new_path, new_buffer,
-+ old_path, old_buffer));
-+ if (!target) {
-+ info = "label build failed";
-+ error = -ENOMEM;
-+ goto fail;
-+ } else if (!IS_ERR(target)) {
-+ error = aa_replace_current_label(target);
-+ if (error) {
-+ /* TODO: audit target */
-+ aa_put_label(target);
-+ goto out;
-+ }
-+ } else
-+ /* already audited error */
-+ error = PTR_ERR(target);
-+out:
-+ put_buffers(old_buffer, new_buffer);
-+
-+ return error;
-+
-+fail:
-+ /* TODO: add back in auditing of new_name and old_name */
-+ error = fn_for_each(label, profile,
-+ audit_mount(profile, OP_PIVOTROOT, NULL /*new_name */,
-+ NULL /* old_name */,
-+ NULL, NULL,
-+ 0, NULL, AA_MAY_PIVOTROOT, &nullperms, info,
-+ error));
-+ goto out;
-+}
---
-2.11.0
-
aufs4.x-rcN base patch
diff --git a/MAINTAINERS b/MAINTAINERS
-index 1c3feff..1a12137 100644
+index af0cb69..d360d2e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
-@@ -2392,6 +2392,19 @@ F: include/linux/audit.h
+@@ -2465,6 +2465,19 @@ F: include/linux/audit.h
F: include/uapi/linux/audit.h
F: kernel/audit*
M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com>
W: http://miguelojeda.es/auxdisplay.htm
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
-index f321b96..10707c3 100644
+index 85de673..d44de9d 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
-@@ -700,6 +700,24 @@ static inline int is_loop_device(struct file *file)
+@@ -686,6 +686,24 @@ static inline int is_loop_device(struct file *file)
return i && S_ISBLK(i->i_mode) && MAJOR(i->i_rdev) == LOOP_MAJOR;
}
void (*finish)(void *))
{
diff --git a/fs/fcntl.c b/fs/fcntl.c
-index 3b01b64..659760e 100644
+index 448a111..f51c2cf 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -31,7 +31,7 @@
return error;
diff --git a/fs/inode.c b/fs/inode.c
-index 5037059..73820bf 100644
+index d1e35b5..f7800d6 100644
--- a/fs/inode.c
+++ b/fs/inode.c
-@@ -1641,7 +1641,7 @@ EXPORT_SYMBOL(generic_update_time);
+@@ -1655,7 +1655,7 @@ EXPORT_SYMBOL(generic_update_time);
* This does the actual work of updating an inodes time or version. Must have
* had called mnt_want_write() before calling this.
*/
{
int (*update_time)(struct inode *, struct timespec *, int);
+diff --git a/fs/namespace.c b/fs/namespace.c
+index d18deb4..e5a4a7f 100644
+--- a/fs/namespace.c
++++ b/fs/namespace.c
+@@ -846,6 +846,12 @@ static inline int check_mnt(struct mount *mnt)
+ return mnt->mnt_ns == current->nsproxy->mnt_ns;
+ }
+
++/* for aufs, CONFIG_AUFS_BR_FUSE */
++int is_current_mnt_ns(struct vfsmount *mnt)
++{
++ return check_mnt(real_mount(mnt));
++}
++
+ /*
+ * vfsmount lock must be held for write
+ */
diff --git a/fs/read_write.c b/fs/read_write.c
-index 0cc7033..6e542f0 100644
+index f0d4b16..6aa8c7a 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
-@@ -473,6 +473,28 @@ ssize_t __vfs_write(struct file *file, const char __user *p, size_t count,
+@@ -483,6 +483,28 @@ ssize_t __vfs_write(struct file *file, const char __user *p, size_t count,
+ return -EINVAL;
}
- EXPORT_SYMBOL(__vfs_write);
+vfs_readf_t vfs_readf(struct file *file)
+{
+ return ERR_PTR(-ENOSYS);
+}
+
- ssize_t __kernel_write(struct file *file, const char *buf, size_t count, loff_t *pos)
+ ssize_t __kernel_write(struct file *file, const void *buf, size_t count, loff_t *pos)
{
mm_segment_t old_fs;
diff --git a/fs/splice.c b/fs/splice.c
-index ae41201..9753304 100644
+index f3084cc..eb888c6 100644
--- a/fs/splice.c
+++ b/fs/splice.c
-@@ -853,8 +853,8 @@ EXPORT_SYMBOL(generic_splice_sendpage);
+@@ -837,8 +837,8 @@ EXPORT_SYMBOL(generic_splice_sendpage);
/*
* Attempt to initiate a splice from pipe to file.
*/
{
ssize_t (*splice_write)(struct pipe_inode_info *, struct file *,
loff_t *, size_t, unsigned int);
-@@ -870,9 +870,9 @@ static long do_splice_from(struct pipe_inode_info *pipe, struct file *out,
+@@ -854,9 +854,9 @@ static long do_splice_from(struct pipe_inode_info *pipe, struct file *out,
/*
* Attempt to initiate a splice from a file to a pipe.
*/
ssize_t (*splice_read)(struct file *, loff_t *,
struct pipe_inode_info *, size_t, unsigned int);
diff --git a/fs/sync.c b/fs/sync.c
-index 2a54c1f..7a5fa3f 100644
+index a576aa2..eb61780 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -27,7 +27,7 @@
static inline void fput_light(struct file *file, int fput_needed)
{
diff --git a/include/linux/fs.h b/include/linux/fs.h
-index cbfe127..9b21bb5 100644
+index 13dab19..8ab6566 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
-@@ -1262,6 +1262,7 @@ extern void fasync_free(struct fasync_struct *);
+@@ -1264,6 +1264,7 @@ extern void fasync_free(struct fasync_struct *);
/* can be called from interrupts */
extern void kill_fasync(struct fasync_struct **, int, int);
extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
extern int f_setown(struct file *filp, unsigned long arg, int force);
extern void f_delown(struct file *filp);
-@@ -1683,6 +1684,7 @@ struct file_operations {
+@@ -1710,6 +1711,7 @@ struct file_operations {
ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
int (*check_flags)(int);
int (*flock) (struct file *, int, struct file_lock *);
ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
-@@ -1753,6 +1755,12 @@ ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
+@@ -1780,6 +1782,12 @@ ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
struct iovec *fast_pointer,
struct iovec **ret_pointer);
+vfs_writef_t vfs_writef(struct file *file);
+
extern ssize_t __vfs_read(struct file *, char __user *, size_t, loff_t *);
- extern ssize_t __vfs_write(struct file *, const char __user *, size_t, loff_t *);
extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
-@@ -2157,6 +2165,7 @@ extern int current_umask(void);
+ extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
+@@ -2182,6 +2190,7 @@ extern int current_umask(void);
extern void ihold(struct inode * inode);
extern void iput(struct inode *);
extern int generic_update_time(struct inode *, struct timespec *, int);
/* /sys/fs */
extern struct kobject *fs_kobj;
-@@ -2437,6 +2446,7 @@ static inline bool sb_is_blkdev_sb(struct super_block *sb)
+@@ -2462,6 +2471,7 @@ static inline bool sb_is_blkdev_sb(struct super_block *sb)
return false;
}
#endif
extern int sync_filesystem(struct super_block *);
extern const struct file_operations def_blk_fops;
extern const struct file_operations def_chr_fops;
+diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
+index bfa8e0b..728d810 100644
+--- a/include/linux/lockdep.h
++++ b/include/linux/lockdep.h
+@@ -405,6 +405,8 @@ static inline int lockdep_match_key(struct lockdep_map *lock,
+ return lock->key == key;
+ }
+
++struct lock_class *lockdep_hlock_class(struct held_lock *hlock);
++
+ /*
+ * Acquire a lock.
+ *
+@@ -529,6 +531,7 @@ struct lock_class_key { };
+
+ #define lockdep_depth(tsk) (0)
+
++#define lockdep_is_held(lock) (1)
+ #define lockdep_is_held_type(l, r) (1)
+
+ #define lockdep_assert_held(l) do { (void)(l); } while (0)
+diff --git a/include/linux/mnt_namespace.h b/include/linux/mnt_namespace.h
+index 12b2ab5..8b810d1 100644
+--- a/include/linux/mnt_namespace.h
++++ b/include/linux/mnt_namespace.h
+@@ -5,11 +5,14 @@
+ struct mnt_namespace;
+ struct fs_struct;
+ struct user_namespace;
++struct vfsmount;
+
+ extern struct mnt_namespace *copy_mnt_ns(unsigned long, struct mnt_namespace *,
+ struct user_namespace *, struct fs_struct *);
+ extern void put_mnt_ns(struct mnt_namespace *ns);
+
++extern int is_current_mnt_ns(struct vfsmount *mnt);
++
+ extern const struct file_operations proc_mounts_operations;
+ extern const struct file_operations proc_mountinfo_operations;
+ extern const struct file_operations proc_mountstats_operations;
diff --git a/include/linux/splice.h b/include/linux/splice.h
index db42746..12f3a5a 100644
--- a/include/linux/splice.h
+ struct pipe_inode_info *pipe, size_t len,
+ unsigned int flags);
#endif
+diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c
+index e36e652..bc97a97 100644
+--- a/kernel/locking/lockdep.c
++++ b/kernel/locking/lockdep.c
+@@ -144,7 +144,7 @@ static struct lock_list list_entries[MAX_LOCKDEP_ENTRIES];
+ unsigned long nr_lock_classes;
+ static struct lock_class lock_classes[MAX_LOCKDEP_KEYS];
+
+-static inline struct lock_class *hlock_class(struct held_lock *hlock)
++inline struct lock_class *lockdep_hlock_class(struct held_lock *hlock)
+ {
+ if (!hlock->class_idx) {
+ /*
+@@ -155,6 +155,7 @@ static inline struct lock_class *hlock_class(struct held_lock *hlock)
+ }
+ return lock_classes + hlock->class_idx - 1;
+ }
++#define hlock_class(hlock) lockdep_hlock_class(hlock)
+
+ #ifdef CONFIG_LOCK_STAT
+ static DEFINE_PER_CPU(struct lock_class_stats[MAX_LOCKDEP_KEYS], cpu_lock_stats);
aufs4.x-rcN mmap patch
diff --git a/fs/proc/base.c b/fs/proc/base.c
-index 719c2e9..a1b7968 100644
+index ad3b076..ad4a50d 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
-@@ -1986,7 +1986,7 @@ static int map_files_get_link(struct dentry *dentry, struct path *path)
+@@ -1987,7 +1987,7 @@ static int map_files_get_link(struct dentry *dentry, struct path *path)
down_read(&mm->mmap_sem);
vma = find_exact_vma(mm, vm_start, vm_end);
if (vma && vma->vm_file) {
ino = inode->i_ino;
}
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
-index fe8f326..b2f7f1a 100644
+index 5589b4b..f60aea2 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
-@@ -293,7 +293,10 @@ show_map_vma(struct seq_file *m, struct vm_area_struct *vma, int is_pid)
+@@ -309,7 +309,10 @@ show_map_vma(struct seq_file *m, struct vm_area_struct *vma, int is_pid)
const char *name = NULL;
if (file) {
dev = inode->i_sb->s_dev;
ino = inode->i_ino;
pgoff = ((loff_t)vma->vm_pgoff) << PAGE_SHIFT;
-@@ -1640,7 +1643,7 @@ static int show_numa_map(struct seq_file *m, void *v, int is_pid)
+@@ -1734,7 +1737,7 @@ static int show_numa_map(struct seq_file *m, void *v, int is_pid)
struct proc_maps_private *proc_priv = &numa_priv->proc_maps;
struct vm_area_struct *vma = v;
struct numa_maps *md = &numa_priv->md;
struct mm_walk walk = {
.hugetlb_entry = gather_hugetlb_stats,
diff --git a/fs/proc/task_nommu.c b/fs/proc/task_nommu.c
-index 23266694..58e59b6 100644
+index b00b7660..93e8a86 100644
--- a/fs/proc/task_nommu.c
+++ b/fs/proc/task_nommu.c
-@@ -157,7 +157,10 @@ static int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma,
+@@ -155,7 +155,10 @@ static int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma,
file = vma->vm_file;
if (file) {
ino = inode->i_ino;
pgoff = (loff_t)vma->vm_pgoff << PAGE_SHIFT;
diff --git a/include/linux/mm.h b/include/linux/mm.h
-index 46b9ac5..62ba1c3 100644
+index 065d99d..04486c3 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
-@@ -1306,6 +1306,28 @@ static inline int fixup_user_fault(struct task_struct *tsk,
+@@ -1348,6 +1348,28 @@ static inline int fixup_user_fault(struct task_struct *tsk,
}
#endif
unsigned int gup_flags);
extern int access_remote_vm(struct mm_struct *mm, unsigned long addr,
diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
-index 3cadee0..d0142c1 100644
+index 1861ea8..d85a914 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
-@@ -259,6 +259,7 @@ struct vm_region {
+@@ -260,6 +260,7 @@ struct vm_region {
unsigned long vm_top; /* region allocated to here */
unsigned long vm_pgoff; /* the offset in vm_file corresponding to vm_start */
struct file *vm_file; /* the backing file or NULL */
int vm_usage; /* region usage count (access under nommu_region_sem) */
bool vm_icache_flushed : 1; /* true if the icache has been flushed for
-@@ -333,6 +334,7 @@ struct vm_area_struct {
+@@ -334,6 +335,7 @@ struct vm_area_struct {
unsigned long vm_pgoff; /* Offset (within vm_file) in PAGE_SIZE
units */
struct file * vm_file; /* File we map to (can be NULL). */
+ struct file *vm_prfile; /* shadow of vm_file */
void * vm_private_data; /* was vm_pte (shared mem) */
- #ifndef CONFIG_MMU
+ atomic_long_t swap_readahead_info;
diff --git a/kernel/fork.c b/kernel/fork.c
-index cbbea27..f0fd8a1 100644
+index 07cc743..b1d2b43 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
-@@ -663,7 +663,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
+@@ -676,7 +676,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
struct inode *inode = file_inode(file);
struct address_space *mapping = file->f_mapping;
atomic_dec(&inode->i_writecount);
i_mmap_lock_write(mapping);
diff --git a/mm/Makefile b/mm/Makefile
-index 411bd24..e7de927 100644
+index e3ac3ae..745b26c 100644
--- a/mm/Makefile
+++ b/mm/Makefile
@@ -39,7 +39,7 @@ obj-y := filemap.o mempool.o oom_kill.o \
obj-y += init-mm.o
diff --git a/mm/filemap.c b/mm/filemap.c
-index 0b41c8c..c5262ff 100644
+index 594d73f..7183aef 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
-@@ -2543,7 +2543,7 @@ int filemap_page_mkwrite(struct vm_fault *vmf)
+@@ -2590,7 +2590,7 @@ int filemap_page_mkwrite(struct vm_fault *vmf)
int ret = VM_FAULT_LOCKED;
sb_start_pagefault(inode->i_sb);
if (page->mapping != inode->i_mapping) {
unlock_page(page);
diff --git a/mm/mmap.c b/mm/mmap.c
-index f19efcf..7fdd59e 100644
+index 680506f..081406a 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
-@@ -170,7 +170,7 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
+@@ -171,7 +171,7 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
if (vma->vm_ops && vma->vm_ops->close)
vma->vm_ops->close(vma);
if (vma->vm_file)
mpol_put(vma_policy(vma));
kmem_cache_free(vm_area_cachep, vma);
return next;
-@@ -895,7 +895,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+@@ -896,7 +896,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
if (remove_next) {
if (file) {
uprobe_munmap(next, next->vm_start, next->vm_end);
}
if (next->anon_vma)
anon_vma_merge(vma, next);
-@@ -1745,8 +1745,8 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
+@@ -1746,8 +1746,8 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
return addr;
unmap_and_free_vma:
/* Undo any partial mapping done by a device driver. */
unmap_region(mm, vma, prev, vma->vm_start, vma->vm_end);
-@@ -2568,7 +2568,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
+@@ -2569,7 +2569,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
goto out_free_mpol;
if (new->vm_file)
if (new->vm_ops && new->vm_ops->open)
new->vm_ops->open(new);
-@@ -2587,7 +2587,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
+@@ -2588,7 +2588,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
if (new->vm_ops && new->vm_ops->close)
new->vm_ops->close(new);
if (new->vm_file)
unlink_anon_vmas(new);
out_free_mpol:
mpol_put(vma_policy(new));
-@@ -2741,7 +2741,7 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
+@@ -2750,7 +2750,7 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
struct vm_area_struct *vma;
unsigned long populate = 0;
unsigned long ret = -EINVAL;
pr_warn_once("%s (%d) uses deprecated remap_file_pages() syscall. See Documentation/vm/remap_file_pages.txt.\n",
current->comm, current->pid);
-@@ -2816,10 +2816,27 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
+@@ -2825,10 +2825,27 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
}
}
out:
up_write(&mm->mmap_sem);
if (populate)
-@@ -3110,7 +3127,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
+@@ -3136,7 +3153,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
if (anon_vma_clone(new_vma, vma))
goto out_free_mempol;
if (new_vma->vm_file)
new_vma->vm_ops->open(new_vma);
vma_link(mm, new_vma, prev, rb_link, rb_parent);
diff --git a/mm/nommu.c b/mm/nommu.c
-index fc184f5..637ea81 100644
+index 17c00d9..4bcdf94 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -641,7 +641,7 @@ static void __put_nommu_region(struct vm_region *region)
put_nommu_region(vma->vm_region);
kmem_cache_free(vm_area_cachep, vma);
}
-@@ -1326,7 +1326,7 @@ unsigned long do_mmap(struct file *file,
+@@ -1321,7 +1321,7 @@ unsigned long do_mmap(struct file *file,
goto error_just_free;
}
}
kmem_cache_free(vm_region_jar, region);
region = pregion;
result = start;
-@@ -1401,10 +1401,10 @@ unsigned long do_mmap(struct file *file,
+@@ -1396,10 +1396,10 @@ unsigned long do_mmap(struct file *file,
up_write(&nommu_region_sem);
error:
if (region->vm_file)
/**
* d_ancestor - search for an ancestor
diff --git a/fs/exec.c b/fs/exec.c
-index 62175cb..f0b6fdd 100644
+index 3e14ba2..6818b01 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -109,6 +109,7 @@ bool path_noexec(const struct path *path)
#ifdef CONFIG_USELIB
/*
diff --git a/fs/fcntl.c b/fs/fcntl.c
-index 659760e..5c37087 100644
+index f51c2cf..58bf222 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -84,6 +84,7 @@ int setfl(int fd, struct file * filp, unsigned long arg)
static void f_modown(struct file *filp, struct pid *pid, enum pid_type type,
int force)
diff --git a/fs/file_table.c b/fs/file_table.c
-index 72e861a..01ae52f 100644
+index 61517f5..c6bab39c 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -148,6 +148,7 @@ struct file *get_empty_filp(void)
/**
* alloc_file - allocate and initialize a 'struct file'
-@@ -260,6 +261,7 @@ void flush_delayed_fput(void)
+@@ -258,6 +259,7 @@ void flush_delayed_fput(void)
{
delayed_fput(NULL);
}
static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
-@@ -302,6 +304,7 @@ void __fput_sync(struct file *file)
+@@ -300,6 +302,7 @@ void __fput_sync(struct file *file)
}
EXPORT_SYMBOL(fput);
void put_filp(struct file *file)
{
-@@ -310,6 +313,7 @@ void put_filp(struct file *file)
+@@ -308,6 +311,7 @@ void put_filp(struct file *file)
file_free(file);
}
}
+EXPORT_SYMBOL_GPL(put_filp);
void __init files_init(void)
- {
+ {
diff --git a/fs/inode.c b/fs/inode.c
-index 73820bf..7db829e 100644
+index f7800d6..f31a6c7 100644
--- a/fs/inode.c
+++ b/fs/inode.c
-@@ -1650,6 +1650,7 @@ int update_time(struct inode *inode, struct timespec *time, int flags)
+@@ -1664,6 +1664,7 @@ int update_time(struct inode *inode, struct timespec *time, int flags)
return update_time(inode, time, flags);
}
/**
* touch_atime - update the access time
diff --git a/fs/namespace.c b/fs/namespace.c
-index f8893dc..c55d949 100644
+index e5a4a7f..6d0c376 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
-@@ -463,6 +463,7 @@ void __mnt_drop_write(struct vfsmount *mnt)
+@@ -517,6 +517,7 @@ void __mnt_drop_write(struct vfsmount *mnt)
mnt_dec_writers(real_mount(mnt));
preempt_enable();
}
/**
* mnt_drop_write - give up write access to a mount
-@@ -1823,6 +1824,7 @@ int iterate_mounts(int (*f)(struct vfsmount *, void *), void *arg,
+@@ -851,6 +852,7 @@ int is_current_mnt_ns(struct vfsmount *mnt)
+ {
+ return check_mnt(real_mount(mnt));
+ }
++EXPORT_SYMBOL_GPL(is_current_mnt_ns);
+
+ /*
+ * vfsmount lock must be held for write
+@@ -1887,6 +1889,7 @@ int iterate_mounts(int (*f)(struct vfsmount *, void *), void *arg,
}
return 0;
}
/*
* Destroy all marks in destroy_list, waits for SRCU period to finish before
diff --git a/fs/open.c b/fs/open.c
-index 35bb784..92e08c5 100644
+index 7ea1184..6e2e241 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -64,6 +64,7 @@ int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
static int do_dentry_open(struct file *f,
struct inode *inode,
diff --git a/fs/read_write.c b/fs/read_write.c
-index 6e542f0..c6fa090 100644
+index 6aa8c7a..b5d392e 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
-@@ -483,6 +483,7 @@ vfs_readf_t vfs_readf(struct file *file)
+@@ -453,6 +453,7 @@ ssize_t vfs_read(struct file *file, char __user *buf, size_t count, loff_t *pos)
+
+ return ret;
+ }
++EXPORT_SYMBOL_GPL(vfs_read);
+
+ static ssize_t new_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos)
+ {
+@@ -493,6 +494,7 @@ vfs_readf_t vfs_readf(struct file *file)
return new_sync_read;
return ERR_PTR(-ENOSYS);
}
vfs_writef_t vfs_writef(struct file *file)
{
-@@ -494,6 +495,7 @@ vfs_writef_t vfs_writef(struct file *file)
+@@ -504,6 +506,7 @@ vfs_writef_t vfs_writef(struct file *file)
return new_sync_write;
return ERR_PTR(-ENOSYS);
}
+EXPORT_SYMBOL_GPL(vfs_writef);
- ssize_t __kernel_write(struct file *file, const char *buf, size_t count, loff_t *pos)
+ ssize_t __kernel_write(struct file *file, const void *buf, size_t count, loff_t *pos)
+ {
+@@ -573,6 +576,7 @@ ssize_t vfs_write(struct file *file, const char __user *buf, size_t count, loff_
+
+ return ret;
+ }
++EXPORT_SYMBOL_GPL(vfs_write);
+
+ static inline loff_t file_pos_read(struct file *file)
{
diff --git a/fs/splice.c b/fs/splice.c
-index 9753304..b38e036 100644
+index eb888c6..7ab89d2 100644
--- a/fs/splice.c
+++ b/fs/splice.c
-@@ -866,6 +866,7 @@ long do_splice_from(struct pipe_inode_info *pipe, struct file *out,
+@@ -850,6 +850,7 @@ long do_splice_from(struct pipe_inode_info *pipe, struct file *out,
return splice_write(pipe, out, ppos, len, flags);
}
/*
* Attempt to initiate a splice from a file to a pipe.
-@@ -895,6 +896,7 @@ long do_splice_to(struct file *in, loff_t *ppos,
+@@ -879,6 +880,7 @@ long do_splice_to(struct file *in, loff_t *ppos,
return splice_read(in, ppos, pipe, len, flags);
}
/**
* splice_direct_to_actor - splices data directly between two non-pipes
diff --git a/fs/sync.c b/fs/sync.c
-index 7a5fa3f..c9b9d46 100644
+index eb61780..32c5a05 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -38,6 +38,7 @@ int __sync_filesystem(struct super_block *sb, int wait)
/*
* Write out and wait upon all dirty data associated with this
diff --git a/fs/xattr.c b/fs/xattr.c
-index 464c94b..0234d49 100644
+index 61cd28b..35570cd 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
-@@ -296,6 +296,7 @@ vfs_getxattr_alloc(struct dentry *dentry, const char *name, char **xattr_value,
+@@ -297,6 +297,7 @@ vfs_getxattr_alloc(struct dentry *dentry, const char *name, char **xattr_value,
*xattr_value = value;
return error;
}
ssize_t
__vfs_getxattr(struct dentry *dentry, struct inode *inode, const char *name,
+diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c
+index bc97a97..895a1ba 100644
+--- a/kernel/locking/lockdep.c
++++ b/kernel/locking/lockdep.c
+@@ -155,6 +155,7 @@ inline struct lock_class *lockdep_hlock_class(struct held_lock *hlock)
+ }
+ return lock_classes + hlock->class_idx - 1;
+ }
++EXPORT_SYMBOL_GPL(lockdep_hlock_class);
+ #define hlock_class(hlock) lockdep_hlock_class(hlock)
+
+ #ifdef CONFIG_LOCK_STAT
diff --git a/kernel/task_work.c b/kernel/task_work.c
-index d513051..e056d54 100644
+index 836a72a..aa00d49 100644
--- a/kernel/task_work.c
+++ b/kernel/task_work.c
-@@ -119,3 +119,4 @@ void task_work_run(void)
+@@ -115,3 +115,4 @@ void task_work_run(void)
} while (work);
}
}
+EXPORT_SYMBOL_GPL(task_work_run);
diff --git a/security/commoncap.c b/security/commoncap.c
-index 7abebd7..c079ce4 100644
+index fc46f5b..90543ef 100644
--- a/security/commoncap.c
+++ b/security/commoncap.c
-@@ -1062,12 +1062,14 @@ int cap_mmap_addr(unsigned long addr)
+@@ -1270,12 +1270,14 @@ int cap_mmap_addr(unsigned long addr)
}
return ret;
}
int devcgroup_inode_mknod(int mode, dev_t dev)
{
diff --git a/security/security.c b/security/security.c
-index 3013237..342ce8b 100644
+index 4bf0f57..b30d1e1 100644
--- a/security/security.c
+++ b/security/security.c
-@@ -535,6 +535,7 @@ int security_path_rmdir(const struct path *dir, struct dentry *dentry)
+@@ -530,6 +530,7 @@ int security_path_rmdir(const struct path *dir, struct dentry *dentry)
return 0;
return call_int_hook(path_rmdir, 0, dir, dentry);
}
int security_path_unlink(const struct path *dir, struct dentry *dentry)
{
-@@ -551,6 +552,7 @@ int security_path_symlink(const struct path *dir, struct dentry *dentry,
+@@ -546,6 +547,7 @@ int security_path_symlink(const struct path *dir, struct dentry *dentry,
return 0;
return call_int_hook(path_symlink, 0, dir, dentry, old_name);
}
int security_path_link(struct dentry *old_dentry, const struct path *new_dir,
struct dentry *new_dentry)
-@@ -559,6 +561,7 @@ int security_path_link(struct dentry *old_dentry, const struct path *new_dir,
+@@ -554,6 +556,7 @@ int security_path_link(struct dentry *old_dentry, const struct path *new_dir,
return 0;
return call_int_hook(path_link, 0, old_dentry, new_dir, new_dentry);
}
int security_path_rename(const struct path *old_dir, struct dentry *old_dentry,
const struct path *new_dir, struct dentry *new_dentry,
-@@ -586,6 +589,7 @@ int security_path_truncate(const struct path *path)
+@@ -581,6 +584,7 @@ int security_path_truncate(const struct path *path)
return 0;
return call_int_hook(path_truncate, 0, path);
}
int security_path_chmod(const struct path *path, umode_t mode)
{
-@@ -593,6 +597,7 @@ int security_path_chmod(const struct path *path, umode_t mode)
+@@ -588,6 +592,7 @@ int security_path_chmod(const struct path *path, umode_t mode)
return 0;
return call_int_hook(path_chmod, 0, path, mode);
}
int security_path_chown(const struct path *path, kuid_t uid, kgid_t gid)
{
-@@ -600,6 +605,7 @@ int security_path_chown(const struct path *path, kuid_t uid, kgid_t gid)
+@@ -595,6 +600,7 @@ int security_path_chown(const struct path *path, kuid_t uid, kgid_t gid)
return 0;
return call_int_hook(path_chown, 0, path, uid, gid);
}
int security_path_chroot(const struct path *path)
{
-@@ -685,6 +691,7 @@ int security_inode_readlink(struct dentry *dentry)
+@@ -680,6 +686,7 @@ int security_inode_readlink(struct dentry *dentry)
return 0;
return call_int_hook(inode_readlink, 0, dentry);
}
int security_inode_follow_link(struct dentry *dentry, struct inode *inode,
bool rcu)
-@@ -700,6 +707,7 @@ int security_inode_permission(struct inode *inode, int mask)
+@@ -695,6 +702,7 @@ int security_inode_permission(struct inode *inode, int mask)
return 0;
return call_int_hook(inode_permission, 0, inode, mask);
}
int security_inode_setattr(struct dentry *dentry, struct iattr *attr)
{
-@@ -871,6 +879,7 @@ int security_file_permission(struct file *file, int mask)
+@@ -866,6 +874,7 @@ int security_file_permission(struct file *file, int mask)
return fsnotify_perm(file, mask);
}
int security_file_alloc(struct file *file)
{
-@@ -930,6 +939,7 @@ int security_mmap_file(struct file *file, unsigned long prot,
+@@ -925,6 +934,7 @@ int security_mmap_file(struct file *file, unsigned long prot,
return ret;
return ima_file_mmap(file, prot);
}
+ will be empty. About XINO files, see the aufs manual.
diff -urN /usr/share/empty/Documentation/filesystems/aufs/design/01intro.txt linux/Documentation/filesystems/aufs/design/01intro.txt
--- /usr/share/empty/Documentation/filesystems/aufs/design/01intro.txt 1970-01-01 01:00:00.000000000 +0100
-+++ linux/Documentation/filesystems/aufs/design/01intro.txt 2017-09-05 10:42:11.038754821 +0200
++++ linux/Documentation/filesystems/aufs/design/01intro.txt 2017-11-12 22:24:42.257509799 +0100
@@ -0,0 +1,171 @@
+
+# Copyright (C) 2005-2017 Junjiro R. Okajima
+ where the source and the target exists and selects the higher
+ one. If the selected branch is readonly, then aufs follows the
+ copyup policy.
+diff -urN /usr/share/empty/Documentation/filesystems/aufs/design/06dirren.dot linux/Documentation/filesystems/aufs/design/06dirren.dot
+--- /usr/share/empty/Documentation/filesystems/aufs/design/06dirren.dot 1970-01-01 01:00:00.000000000 +0100
++++ linux/Documentation/filesystems/aufs/design/06dirren.dot 2017-11-12 22:24:44.694244127 +0100
+@@ -0,0 +1,31 @@
++
++// to view this graph, run dot(1) command in GRAPHVIZ.
++
++digraph G {
++node [shape=box];
++whinfo [label="detailed info file\n(lower_brid_root-hinum, h_inum, namelen, old name)"];
++
++node [shape=oval];
++
++aufs_rename -> whinfo [label="store/remove"];
++
++node [shape=oval];
++inode_list [label="h_inum list in branch\ncache"];
++
++node [shape=box];
++whinode [label="h_inum list file"];
++
++node [shape=oval];
++brmgmt [label="br_add/del/mod/umount"];
++
++brmgmt -> inode_list [label="create/remove"];
++brmgmt -> whinode [label="load/store"];
++
++inode_list -> whinode [style=dashed,dir=both];
++
++aufs_rename -> inode_list [label="add/del"];
++
++aufs_lookup -> inode_list [label="search"];
++
++aufs_lookup -> whinfo [label="load/remove"];
++}
+diff -urN /usr/share/empty/Documentation/filesystems/aufs/design/06dirren.txt linux/Documentation/filesystems/aufs/design/06dirren.txt
+--- /usr/share/empty/Documentation/filesystems/aufs/design/06dirren.txt 1970-01-01 01:00:00.000000000 +0100
++++ linux/Documentation/filesystems/aufs/design/06dirren.txt 2017-11-12 22:24:44.694244127 +0100
+@@ -0,0 +1,102 @@
++
++# Copyright (C) 2017 Junjiro R. Okajima
++#
++# This program is free software; you can redistribute it and/or modify
++# it under the terms of the GNU General Public License as published by
++# the Free Software Foundation; either version 2 of the License, or
++# (at your option) any later version.
++#
++# This program is distributed in the hope that it will be useful,
++# but WITHOUT ANY WARRANTY; without even the implied warranty of
++# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
++# GNU General Public License for more details.
++#
++# You should have received a copy of the GNU General Public License
++# along with this program. If not, see <http://www.gnu.org/licenses/>.
++
++Special handling for renaming a directory (DIRREN)
++----------------------------------------------------------------------
++First, let's assume we have a simple usecase.
++
++- /u = /rw + /ro
++- /rw/dirA exists
++- /ro/dirA and /ro/dirA/file exist too
++- there is no dirB on both branches
++- a user issues rename("dirA", "dirB")
++
++Now, what should aufs behave against this rename(2)?
++There are a few possible cases.
++
++A. returns EROFS.
++ since dirA exists on a readonly branch which cannot be renamed.
++B. returns EXDEV.
++ it is possible to copy-up dirA (only the dir itself), but the child
++ entries ("file" in this case) should not be. it must be a bad
++ approach to copy-up recursively.
++C. returns a success.
++ even the branch /ro is readonly, aufs tries renaming it. Obviously it
++ is a violation of aufs' policy.
++D. construct an extra information which indicates that /ro/dirA should
++ be handled as the name of dirB.
++ overlayfs has a similar feature called REDIRECT.
++
++Until now, aufs implements the case B only which returns EXDEV, and
++expects the userspace application behaves like mv(1) which tries
++issueing rename(2) recursively.
++
++A new aufs feature called DIRREN is introduced which implements the case
++D. There are several "extra information" added.
++
++1. detailed info per renamed directory
++ path: /rw/dirB/$AUFS_WH_DR_INFO_PFX.<lower branch-id>
++2. the inode-number list of directories on a branch
++ path: /rw/dirB/$AUFS_WH_DR_BRHINO
++
++The filename of "detailed info per directory" represents the lower
++branch, and its format is
++- a type of the branch id
++ one of these.
++ + uuid (not implemented yet)
++ + fsid
++ + dev
++- the inode-number of the branch root dir
++
++And it contains these info in a single regular file.
++- magic number
++- branch's inode-number of the logically renamed dir
++- the name of the before-renamed dir
++
++The "detailed info per directory" file is created in aufs rename(2), and
++loaded in any lookup.
++The info is considered in lookup for the matching case only. Here
++"matching" means that the root of branch (in the info filename) is same
++to the current looking-up branch. After looking-up the before-renamed
++name, the inode-number is compared. And the matched dentry is used.
++
++The "inode-number list of directories" is a regular file which contains
++simply the inode-numbers on the branch. The file is created or updated
++in removing the branch, and loaded in adding the branch. Its lifetime is
++equal to the branch.
++The list is refered in lookup, and when the current target inode is
++found in the list, the aufs tries loading the "detailed info per
++directory" and get the changed and valid name of the dir.
++
++Theoretically these "extra informaiton" may be able to be put into XATTR
++in the dir inode. But aufs doesn't choose this way because
++1. XATTR may not be supported by the branch (or its configuration)
++2. XATTR may have its size limit.
++3. XATTR may be less easy to convert than a regular file, when the
++ format of the info is changed in the future.
++At the same time, I agree that the regular file approach is much slower
++than XATTR approach. So, in the future, aufs may take the XATTR or other
++better approach.
++
++This DIRREN feature is enabled by aufs configuration, and is activated
++by a new mount option.
++
++For the more complicated case, there is a work with UDBA option, which
++is to dected the direct access to the branches (by-passing aufs) and to
++maintain the cashes in aufs. Since a single cached aufs dentry may
++contains two names, before- and after-rename, the name comparision in
++UDBA handler may not work correctly. In this case, the behaviour will be
++equivalen to udba=reval case.
diff -urN /usr/share/empty/Documentation/filesystems/aufs/design/06fhsm.txt linux/Documentation/filesystems/aufs/design/06fhsm.txt
--- /usr/share/empty/Documentation/filesystems/aufs/design/06fhsm.txt 1970-01-01 01:00:00.000000000 +0100
+++ linux/Documentation/filesystems/aufs/design/06fhsm.txt 2017-07-29 12:14:25.896375188 +0200
+# End: ;
diff -urN /usr/share/empty/fs/aufs/aufs.h linux/fs/aufs/aufs.h
--- /usr/share/empty/fs/aufs/aufs.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/aufs.h 2017-07-29 12:14:25.896375188 +0200
-@@ -0,0 +1,59 @@
++++ linux/fs/aufs/aufs.h 2017-11-12 22:24:44.697577553 +0100
+@@ -0,0 +1,60 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#include "dbgaufs.h"
+#include "dentry.h"
+#include "dir.h"
++#include "dirren.h"
+#include "dynop.h"
+#include "file.h"
+#include "fstype.h"
++#include "hbl.h"
+#include "inode.h"
+#include "loop.h"
+#include "module.h"
+#include "opts.h"
+#include "rwsem.h"
-+#include "spl.h"
+#include "super.h"
+#include "sysaufs.h"
+#include "vfsub.h"
+#endif /* __AUFS_H__ */
diff -urN /usr/share/empty/fs/aufs/branch.c linux/fs/aufs/branch.c
--- /usr/share/empty/fs/aufs/branch.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/branch.c 2017-07-29 12:14:25.896375188 +0200
-@@ -0,0 +1,1422 @@
++++ linux/fs/aufs/branch.c 2017-11-12 22:24:44.697577553 +0100
+@@ -0,0 +1,1432 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ struct au_dykey **key;
+
+ au_hnotify_fin_br(br);
++ /* always, regardless the mount option */
++ au_dr_hino_free(&br->br_dirren);
+
+ if (br->br_xino.xi_file)
+ fput(br->br_xino.xi_file);
+ goto out;
+ add_branch->br_xino.xi_nondir.total = 8; /* initial size */
+ add_branch->br_xino.xi_nondir.array
-+ = kzalloc(sizeof(ino_t) * add_branch->br_xino.xi_nondir.total,
++ = kcalloc(add_branch->br_xino.xi_nondir.total, sizeof(ino_t),
+ GFP_NOFS);
+ if (unlikely(!add_branch->br_xino.xi_nondir.array))
+ goto out_br;
+ br->br_id = au_new_br_id(sb);
+ AuDebugOn(br->br_id < 0);
+
++ /* always, regardless the given option */
++ err = au_dr_br_init(sb, br, &add->path);
++ if (unlikely(err))
++ goto out_err;
++
+ if (au_br_writable(add->perm)) {
+ err = au_wbr_init(br, sb, add->perm);
+ if (unlikely(err))
+{
+ unsigned long long n;
+ struct file **p, *f;
-+ struct au_sphlhead *files;
++ struct hlist_bl_head *files;
++ struct hlist_bl_node *pos;
+ struct au_finfo *finfo;
+
+ n = 0;
+ p = a;
+ files = &au_sbi(sb)->si_files;
-+ spin_lock(&files->spin);
-+ hlist_for_each_entry(finfo, &files->head, fi_hlist) {
++ hlist_bl_lock(files);
++ hlist_bl_for_each_entry(finfo, pos, files, fi_hlist) {
+ f = finfo->fi_file;
+ if (file_count(f)
+ && !special_file(file_inode(f)->i_mode)) {
+ AuDebugOn(n > max);
+ }
+ }
-+ spin_unlock(&files->spin);
++ hlist_bl_unlock(files);
+
+ return n;
+}
+ au_br_do_del_hip(au_ii(inode), bindex, bbot);
+ au_sbilist_unlock();
+
++ /* ignore an error */
++ au_dr_br_fin(sb, br); /* always, regardless the mount option */
++
+ dput(h_root);
+ iput(h_inode);
+ au_br_do_free(br);
+ goto out; /* success */
+
+out_bf:
-+ if (bf)
-+ kfree(bf);
++ kfree(bf);
+out:
+ AuTraceErr(err);
+ return err;
+}
diff -urN /usr/share/empty/fs/aufs/branch.h linux/fs/aufs/branch.h
--- /usr/share/empty/fs/aufs/branch.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/branch.h 2017-07-29 12:14:25.896375188 +0200
-@@ -0,0 +1,321 @@
++++ linux/fs/aufs/branch.h 2017-11-12 22:24:44.697577553 +0100
+@@ -0,0 +1,333 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#ifdef __KERNEL__
+
+#include <linux/mount.h>
++#include "dirren.h"
+#include "dynop.h"
+#include "rwsem.h"
+#include "super.h"
+ /* entries under sysfs per mount-point */
+ struct au_brsysfs br_sysfs[AuBrSysfs_Last];
+#endif
++
++ struct au_dr_br br_dirren;
+};
+
+/* ---------------------------------------------------------------------- */
+
+static inline int au_br_rdonly(struct au_branch *br)
+{
-+ return ((au_br_sb(br)->s_flags & MS_RDONLY)
++ return (sb_rdonly(au_br_sb(br))
+ || !au_br_writable(br->br_perm))
+ ? -EROFS : 0;
+}
+
+/* ---------------------------------------------------------------------- */
+
++#define wbr_wh_read_lock(wbr) au_rw_read_lock(&(wbr)->wbr_wh_rwsem)
++#define wbr_wh_write_lock(wbr) au_rw_write_lock(&(wbr)->wbr_wh_rwsem)
++#define wbr_wh_read_trylock(wbr) au_rw_read_trylock(&(wbr)->wbr_wh_rwsem)
++#define wbr_wh_write_trylock(wbr) au_rw_write_trylock(&(wbr)->wbr_wh_rwsem)
+/*
-+ * wbr_wh_read_lock, wbr_wh_write_lock
-+ * wbr_wh_read_unlock, wbr_wh_write_unlock, wbr_wh_downgrade_lock
-+ */
-+AuSimpleRwsemFuncs(wbr_wh, struct au_wbr *wbr, &wbr->wbr_wh_rwsem);
++#define wbr_wh_read_trylock_nested(wbr) \
++ au_rw_read_trylock_nested(&(wbr)->wbr_wh_rwsem)
++#define wbr_wh_write_trylock_nested(wbr) \
++ au_rw_write_trylock_nested(&(wbr)->wbr_wh_rwsem)
++*/
+
-+#define WbrWhMustNoWaiters(wbr) AuRwMustNoWaiters(&wbr->wbr_wh_rwsem)
-+#define WbrWhMustAnyLock(wbr) AuRwMustAnyLock(&wbr->wbr_wh_rwsem)
-+#define WbrWhMustWriteLock(wbr) AuRwMustWriteLock(&wbr->wbr_wh_rwsem)
++#define wbr_wh_read_unlock(wbr) au_rw_read_unlock(&(wbr)->wbr_wh_rwsem)
++#define wbr_wh_write_unlock(wbr) au_rw_write_unlock(&(wbr)->wbr_wh_rwsem)
++#define wbr_wh_downgrade_lock(wbr) au_rw_dgrade_lock(&(wbr)->wbr_wh_rwsem)
++
++#define WbrWhMustNoWaiters(wbr) AuRwMustNoWaiters(&(wbr)->wbr_wh_rwsem)
++#define WbrWhMustAnyLock(wbr) AuRwMustAnyLock(&(wbr)->wbr_wh_rwsem)
++#define WbrWhMustWriteLock(wbr) AuRwMustWriteLock(&(wbr)->wbr_wh_rwsem)
+
+/* ---------------------------------------------------------------------- */
+
+#endif /* __AUFS_BRANCH_H__ */
diff -urN /usr/share/empty/fs/aufs/conf.mk linux/fs/aufs/conf.mk
--- /usr/share/empty/fs/aufs/conf.mk 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/conf.mk 2017-07-29 12:14:25.899708630 +0200
-@@ -0,0 +1,38 @@
++++ linux/fs/aufs/conf.mk 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,39 @@
+
+AuConfStr = CONFIG_AUFS_FS=${CONFIG_AUFS_FS}
+
+ XATTR \
+ FHSM \
+ RDU \
++ DIRREN \
+ SHWH \
+ BR_RAMFS \
+ BR_FUSE POLL \
+-include ${srctree}/${src}/conf_priv.mk
diff -urN /usr/share/empty/fs/aufs/cpup.c linux/fs/aufs/cpup.c
--- /usr/share/empty/fs/aufs/cpup.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/cpup.c 2017-09-05 10:42:11.055421928 +0200
-@@ -0,0 +1,1442 @@
++++ linux/fs/aufs/cpup.c 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,1443 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ else {
+ inode_unlock_shared(h_src_inode);
+ err = vfsub_getattr(&h_path, &h_src_attr->st);
-+ vfsub_inode_lock_shared_nested(h_src_inode, AuLsc_I_CHILD);
++ vfsub_inode_lock_shared_nested(h_src_inode,
++ AuLsc_I_CHILD);
+ }
+ if (unlikely(err)) {
+ inode_unlock_shared(h_src_inode);
+}
diff -urN /usr/share/empty/fs/aufs/cpup.h linux/fs/aufs/cpup.h
--- /usr/share/empty/fs/aufs/cpup.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/cpup.h 2017-07-29 12:14:25.899708630 +0200
-@@ -0,0 +1,94 @@
++++ linux/fs/aufs/cpup.h 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,99 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#define AuCpup_RWDST (1 << 5) /* force write target even if
+ the branch is marked as RO */
+
++#ifndef CONFIG_AUFS_BR_HFSPLUS
++#undef AuCpup_HOPEN
++#define AuCpup_HOPEN 0
++#endif
++
+#define au_ftest_cpup(flags, name) ((flags) & AuCpup_##name)
+#define au_fset_cpup(flags, name) \
+ do { (flags) |= AuCpup_##name; } while (0)
+#endif /* __AUFS_CPUP_H__ */
diff -urN /usr/share/empty/fs/aufs/dbgaufs.c linux/fs/aufs/dbgaufs.c
--- /usr/share/empty/fs/aufs/dbgaufs.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dbgaufs.c 2017-07-29 12:14:25.899708630 +0200
-@@ -0,0 +1,438 @@
++++ linux/fs/aufs/dbgaufs.c 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,437 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ struct dbgaufs_plink_arg *p;
+ struct au_sbinfo *sbinfo;
+ struct super_block *sb;
-+ struct au_sphlhead *sphl;
++ struct hlist_bl_head *hbl;
+
+ err = -ENOMEM;
+ p = (void *)get_zeroed_page(GFP_NOFS);
+ limit -= n;
+
+ sum = 0;
-+ for (i = 0, sphl = sbinfo->si_plink;
-+ i < AuPlink_NHASH;
-+ i++, sphl++) {
-+ n = au_sphl_count(sphl);
++ for (i = 0, hbl = sbinfo->si_plink; i < AuPlink_NHASH;
++ i++, hbl++) {
++ n = au_hbl_count(hbl);
+ sum += n;
+
+ n = snprintf(p->a + p->n, limit, "%lu ", n);
+#endif /* __DBGAUFS_H__ */
diff -urN /usr/share/empty/fs/aufs/dcsub.c linux/fs/aufs/dcsub.c
--- /usr/share/empty/fs/aufs/dcsub.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dcsub.c 2017-07-29 12:14:25.899708630 +0200
++++ linux/fs/aufs/dcsub.c 2017-11-12 22:24:42.267510077 +0100
@@ -0,0 +1,225 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+#endif /* __AUFS_DCSUB_H__ */
diff -urN /usr/share/empty/fs/aufs/debug.c linux/fs/aufs/debug.c
--- /usr/share/empty/fs/aufs/debug.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/debug.c 2017-07-29 12:14:25.899708630 +0200
++++ linux/fs/aufs/debug.c 2017-11-12 22:24:42.267510077 +0100
@@ -0,0 +1,440 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+#endif /* __AUFS_DEBUG_H__ */
diff -urN /usr/share/empty/fs/aufs/dentry.c linux/fs/aufs/dentry.c
--- /usr/share/empty/fs/aufs/dentry.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dentry.c 2017-09-05 10:42:11.055421928 +0200
-@@ -0,0 +1,1130 @@
++++ linux/fs/aufs/dentry.c 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,1152 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#include <linux/namei.h>
+#include "aufs.h"
+
-+struct au_do_lookup_args {
-+ unsigned int flags;
-+ mode_t type;
-+};
-+
+/*
+ * returns positive/negative dentry, NULL or an error.
+ * NULL means whiteout-ed or not-found.
+ */
+static struct dentry*
+au_do_lookup(struct dentry *h_parent, struct dentry *dentry,
-+ aufs_bindex_t bindex, struct qstr *wh_name,
-+ struct au_do_lookup_args *args)
++ aufs_bindex_t bindex, struct au_do_lookup_args *args)
+{
+ struct dentry *h_dentry;
+ struct inode *h_inode;
+ br = au_sbr(dentry->d_sb, bindex);
+ wh_able = !!au_br_whable(br->br_perm);
+ if (wh_able)
-+ wh_found = au_wh_test(h_parent, wh_name, ignore_perm);
++ wh_found = au_wh_test(h_parent, &args->whname, ignore_perm);
+ h_dentry = ERR_PTR(wh_found);
+ if (!wh_found)
+ goto real_lookup;
+
+real_lookup:
+ if (!ignore_perm)
-+ h_dentry = vfsub_lkup_one(&dentry->d_name, h_parent);
++ h_dentry = vfsub_lkup_one(args->name, h_parent);
+ else
-+ h_dentry = au_sio_lkup_one(&dentry->d_name, h_parent);
++ h_dentry = au_sio_lkup_one(args->name, h_parent);
+ if (IS_ERR(h_dentry)) {
+ if (PTR_ERR(h_dentry) == -ENAMETOOLONG
+ && !allow_neg)
+ } else if (wh_found
+ || (args->type && args->type != (h_inode->i_mode & S_IFMT)))
+ goto out_neg;
++ else if (au_ftest_lkup(args->flags, DIRREN)
++ /* && h_inode */
++ && !au_dr_lkup_h_ino(args, bindex, h_inode->i_ino)) {
++ AuDbg("b%d %pd ignored hi%llu\n", bindex, h_dentry,
++ (unsigned long long)h_inode->i_ino);
++ goto out_neg;
++ }
+
+ if (au_dbbot(dentry) <= bindex)
+ au_set_dbbot(dentry, bindex);
+{
+ int npositive, err;
+ aufs_bindex_t bindex, btail, bdiropq;
-+ unsigned char isdir, dirperm1;
-+ struct qstr whname;
++ unsigned char isdir, dirperm1, dirren;
+ struct au_do_lookup_args args = {
-+ .flags = flags
++ .flags = flags,
++ .name = &dentry->d_name
+ };
-+ const struct qstr *name = &dentry->d_name;
+ struct dentry *parent;
+ struct super_block *sb;
+
+ sb = dentry->d_sb;
-+ err = au_test_shwh(sb, name);
++ err = au_test_shwh(sb, args.name);
+ if (unlikely(err))
+ goto out;
+
-+ err = au_wh_name_alloc(&whname, name);
++ err = au_wh_name_alloc(&args.whname, args.name);
+ if (unlikely(err))
+ goto out;
+
+ isdir = !!d_is_dir(dentry);
+ dirperm1 = !!au_opt_test(au_mntflags(sb), DIRPERM1);
++ dirren = !!au_opt_test(au_mntflags(sb), DIRREN);
++ if (dirren)
++ au_fset_lkup(args.flags, DIRREN);
+
+ npositive = 0;
+ parent = dget_parent(dentry);
+ for (bindex = btop; bindex <= btail; bindex++) {
+ struct dentry *h_parent, *h_dentry;
+ struct inode *h_inode, *h_dir;
++ struct au_branch *br;
+
+ h_dentry = au_h_dptr(dentry, bindex);
+ if (h_dentry) {
+ if (!h_parent || !d_is_dir(h_parent))
+ continue;
+
++ if (dirren) {
++ /* if the inum matches, then use the prepared name */
++ err = au_dr_lkup_name(&args, bindex);
++ if (unlikely(err))
++ goto out_parent;
++ }
++
+ h_dir = d_inode(h_parent);
+ vfsub_inode_lock_shared_nested(h_dir, AuLsc_I_PARENT);
-+ h_dentry = au_do_lookup(h_parent, dentry, bindex, &whname,
-+ &args);
++ h_dentry = au_do_lookup(h_parent, dentry, bindex, &args);
+ inode_unlock_shared(h_dir);
+ err = PTR_ERR(h_dentry);
+ if (IS_ERR(h_dentry))
+ if (bdiropq >= 0 && bdiropq <= bindex)
+ break;
+ }
++ br = au_sbr(sb, bindex);
++ if (dirren
++ && au_dr_hino_test_add(&br->br_dirren, h_inode->i_ino,
++ /*add_ent*/NULL)) {
++ /* prepare next name to lookup */
++ err = au_dr_lkup(&args, dentry, bindex);
++ if (unlikely(err))
++ goto out_parent;
++ }
+ }
+
+ if (npositive) {
+
+out_parent:
+ dput(parent);
-+ kfree(whname.name);
++ kfree(args.whname.name);
++ if (dirren)
++ au_dr_lkup_fin(&args);
+out:
+ return err;
+}
+
+/* todo: remove this */
+static int h_d_revalidate(struct dentry *dentry, struct inode *inode,
-+ unsigned int flags, int do_udba)
++ unsigned int flags, int do_udba, int dirren)
+{
+ int err;
+ umode_t mode, h_mode;
+ && !is_root
+ && ((!h_nfs
+ && (unhashed != !!d_unhashed(h_dentry)
-+ || (!tmpfile
++ || (!tmpfile && !dirren
+ && !au_qstreq(name, h_name))
+ ))
+ || (h_nfs
+{
+ int valid, err;
+ unsigned int sigen;
-+ unsigned char do_udba;
++ unsigned char do_udba, dirren;
+ struct super_block *sb;
+ struct inode *inode;
+
+ }
+ }
+
-+ err = h_d_revalidate(dentry, inode, flags, do_udba);
++ dirren = !!au_opt_test(au_mntflags(sb), DIRREN);
++ err = h_d_revalidate(dentry, inode, flags, do_udba, dirren);
+ if (unlikely(!err && do_udba && au_dbtop(dentry) < 0)) {
+ err = -EIO;
+ AuDbg("both of real entry and whiteout found, %p, err %d\n",
+};
diff -urN /usr/share/empty/fs/aufs/dentry.h linux/fs/aufs/dentry.h
--- /usr/share/empty/fs/aufs/dentry.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dentry.h 2017-07-29 12:14:25.899708630 +0200
-@@ -0,0 +1,252 @@
++++ linux/fs/aufs/dentry.h 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,266 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#ifdef __KERNEL__
+
+#include <linux/dcache.h>
++#include "dirren.h"
+#include "rwsem.h"
+
+struct au_hdentry {
+/* flags for au_lkup_dentry() */
+#define AuLkup_ALLOW_NEG 1
+#define AuLkup_IGNORE_PERM (1 << 1)
++#define AuLkup_DIRREN (1 << 2)
+#define au_ftest_lkup(flags, name) ((flags) & AuLkup_##name)
+#define au_fset_lkup(flags, name) \
+ do { (flags) |= AuLkup_##name; } while (0)
+#define au_fclr_lkup(flags, name) \
+ do { (flags) &= ~AuLkup_##name; } while (0)
+
++#ifndef CONFIG_AUFS_DIRREN
++#undef AuLkup_DIRREN
++#define AuLkup_DIRREN 0
++#endif
++
++struct au_do_lookup_args {
++ unsigned int flags;
++ mode_t type;
++ struct qstr whname, *name;
++ struct au_dr_lookup dirren;
++};
++
+/* ---------------------------------------------------------------------- */
+
+/* dentry.c */
+#endif /* __AUFS_DENTRY_H__ */
diff -urN /usr/share/empty/fs/aufs/dinfo.c linux/fs/aufs/dinfo.c
--- /usr/share/empty/fs/aufs/dinfo.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dinfo.c 2017-07-29 12:14:25.899708630 +0200
++++ linux/fs/aufs/dinfo.c 2017-11-12 22:24:42.267510077 +0100
@@ -0,0 +1,553 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/dir.c linux/fs/aufs/dir.c
--- /usr/share/empty/fs/aufs/dir.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dir.c 2017-09-05 10:42:11.058755349 +0200
++++ linux/fs/aufs/dir.c 2017-11-12 22:24:44.704244405 +0100
@@ -0,0 +1,759 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ finfo = au_fi(file);
+ fidir = finfo->fi_hdir;
+ if (fidir) {
-+ au_sphl_del(&finfo->fi_hlist,
-+ &au_sbi(file->f_path.dentry->d_sb)->si_files);
++ au_hbl_del(&finfo->fi_hlist,
++ &au_sbi(file->f_path.dentry->d_sb)->si_files);
+ vdir_cache = fidir->fd_vdir_cache; /* lock-free */
+ if (vdir_cache)
+ au_vdir_free(vdir_cache);
+};
diff -urN /usr/share/empty/fs/aufs/dir.h linux/fs/aufs/dir.h
--- /usr/share/empty/fs/aufs/dir.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dir.h 2017-07-29 12:14:25.899708630 +0200
++++ linux/fs/aufs/dir.h 2017-11-12 22:24:42.267510077 +0100
@@ -0,0 +1,131 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+
+#endif /* __KERNEL__ */
+#endif /* __AUFS_DIR_H__ */
+diff -urN /usr/share/empty/fs/aufs/dirren.c linux/fs/aufs/dirren.c
+--- /usr/share/empty/fs/aufs/dirren.c 1970-01-01 01:00:00.000000000 +0100
++++ linux/fs/aufs/dirren.c 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,1315 @@
++/*
++ * Copyright (C) 2017 Junjiro R. Okajima
++ *
++ * This program, aufs is free software; you can redistribute it and/or modify
++ * it under the terms of the GNU General Public License as published by
++ * the Free Software Foundation; either version 2 of the License, or
++ * (at your option) any later version.
++ *
++ * This program is distributed in the hope that it will be useful,
++ * but WITHOUT ANY WARRANTY; without even the implied warranty of
++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
++ * GNU General Public License for more details.
++ *
++ * You should have received a copy of the GNU General Public License
++ * along with this program. If not, see <http://www.gnu.org/licenses/>.
++ */
++
++/*
++ * special handling in renaming a directoy
++ * in order to support looking-up the before-renamed name on the lower readonly
++ * branches
++ */
++
++#include <linux/byteorder/generic.h>
++#include "aufs.h"
++
++static void au_dr_hino_del(struct au_dr_br *dr, struct au_dr_hino *ent)
++{
++ int idx;
++
++ idx = au_dr_ihash(ent->dr_h_ino);
++ au_hbl_del(&ent->dr_hnode, dr->dr_h_ino + idx);
++}
++
++static int au_dr_hino_test_empty(struct au_dr_br *dr)
++{
++ int ret, i;
++ struct hlist_bl_head *hbl;
++
++ ret = 1;
++ for (i = 0; ret && i < AuDirren_NHASH; i++) {
++ hbl = dr->dr_h_ino + i;
++ hlist_bl_lock(hbl);
++ ret &= hlist_bl_empty(hbl);
++ hlist_bl_unlock(hbl);
++ }
++
++ return ret;
++}
++
++static struct au_dr_hino *au_dr_hino_find(struct au_dr_br *dr, ino_t ino)
++{
++ struct au_dr_hino *found, *ent;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos;
++ int idx;
++
++ found = NULL;
++ idx = au_dr_ihash(ino);
++ hbl = dr->dr_h_ino + idx;
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(ent, pos, hbl, dr_hnode)
++ if (ent->dr_h_ino == ino) {
++ found = ent;
++ break;
++ }
++ hlist_bl_unlock(hbl);
++
++ return found;
++}
++
++int au_dr_hino_test_add(struct au_dr_br *dr, ino_t ino,
++ struct au_dr_hino *add_ent)
++{
++ int found, idx;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos;
++ struct au_dr_hino *ent;
++
++ found = 0;
++ idx = au_dr_ihash(ino);
++ hbl = dr->dr_h_ino + idx;
++#if 0
++ {
++ struct hlist_bl_node *tmp;
++
++ hlist_bl_for_each_entry_safe(ent, pos, tmp, hbl, dr_hnode)
++ AuDbg("hi%llu\n", (unsigned long long)ent->dr_h_ino);
++ }
++#endif
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(ent, pos, hbl, dr_hnode)
++ if (ent->dr_h_ino == ino) {
++ found = 1;
++ break;
++ }
++ if (!found && add_ent)
++ hlist_bl_add_head(&add_ent->dr_hnode, hbl);
++ hlist_bl_unlock(hbl);
++
++ if (!found && add_ent)
++ AuDbg("i%llu added\n", (unsigned long long)add_ent->dr_h_ino);
++
++ return found;
++}
++
++void au_dr_hino_free(struct au_dr_br *dr)
++{
++ int i;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos, *tmp;
++ struct au_dr_hino *ent;
++
++ /* SiMustWriteLock(sb); */
++
++ for (i = 0; i < AuDirren_NHASH; i++) {
++ hbl = dr->dr_h_ino + i;
++ /* no spinlock since sbinfo must be write-locked */
++ hlist_bl_for_each_entry_safe(ent, pos, tmp, hbl, dr_hnode)
++ kfree(ent);
++ INIT_HLIST_BL_HEAD(hbl);
++ }
++}
++
++/* returns the number of inodes or an error */
++static int au_dr_hino_store(struct super_block *sb, struct au_branch *br,
++ struct file *hinofile)
++{
++ int err, i;
++ ssize_t ssz;
++ loff_t pos, oldsize;
++ __be64 u64;
++ struct inode *hinoinode;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *n1, *n2;
++ struct au_dr_hino *ent;
++
++ SiMustWriteLock(sb);
++ AuDebugOn(!au_br_writable(br->br_perm));
++
++ hinoinode = file_inode(hinofile);
++ oldsize = i_size_read(hinoinode);
++
++ err = 0;
++ pos = 0;
++ hbl = br->br_dirren.dr_h_ino;
++ for (i = 0; !err && i < AuDirren_NHASH; i++, hbl++) {
++ /* no bit-lock since sbinfo must be write-locked */
++ hlist_bl_for_each_entry_safe(ent, n1, n2, hbl, dr_hnode) {
++ AuDbg("hi%llu, %pD2\n",
++ (unsigned long long)ent->dr_h_ino, hinofile);
++ u64 = cpu_to_be64(ent->dr_h_ino);
++ ssz = vfsub_write_k(hinofile, &u64, sizeof(u64), &pos);
++ if (ssz == sizeof(u64))
++ continue;
++
++ /* write error */
++ pr_err("ssz %zd, %pD2\n", ssz, hinofile);
++ err = -ENOSPC;
++ if (ssz < 0)
++ err = ssz;
++ break;
++ }
++ }
++ /* regardless the error */
++ if (pos < oldsize) {
++ err = vfsub_trunc(&hinofile->f_path, pos, /*attr*/0, hinofile);
++ AuTraceErr(err);
++ }
++
++ AuTraceErr(err);
++ return err;
++}
++
++static int au_dr_hino_load(struct au_dr_br *dr, struct file *hinofile)
++{
++ int err, hidx;
++ ssize_t ssz;
++ size_t sz, n;
++ loff_t pos;
++ uint64_t u64;
++ struct au_dr_hino *ent;
++ struct inode *hinoinode;
++ struct hlist_bl_head *hbl;
++
++ err = 0;
++ pos = 0;
++ hbl = dr->dr_h_ino;
++ hinoinode = file_inode(hinofile);
++ sz = i_size_read(hinoinode);
++ AuDebugOn(sz % sizeof(u64));
++ n = sz / sizeof(u64);
++ while (n--) {
++ ssz = vfsub_read_k(hinofile, &u64, sizeof(u64), &pos);
++ if (unlikely(ssz != sizeof(u64))) {
++ pr_err("ssz %zd, %pD2\n", ssz, hinofile);
++ err = -EINVAL;
++ if (ssz < 0)
++ err = ssz;
++ goto out_free;
++ }
++
++ ent = kmalloc(sizeof(*ent), GFP_NOFS);
++ if (!ent) {
++ err = -ENOMEM;
++ AuTraceErr(err);
++ goto out_free;
++ }
++ ent->dr_h_ino = be64_to_cpu((__force __be64)u64);
++ AuDbg("hi%llu, %pD2\n",
++ (unsigned long long)ent->dr_h_ino, hinofile);
++ hidx = au_dr_ihash(ent->dr_h_ino);
++ au_hbl_add(&ent->dr_hnode, hbl + hidx);
++ }
++ goto out; /* success */
++
++out_free:
++ au_dr_hino_free(dr);
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++/*
++ * @bindex/@br is a switch to distinguish whether suspending hnotify or not.
++ * @path is a switch to distinguish load and store.
++ */
++static int au_dr_hino(struct super_block *sb, aufs_bindex_t bindex,
++ struct au_branch *br, const struct path *path)
++{
++ int err, flags;
++ unsigned char load, suspend;
++ struct file *hinofile;
++ struct au_hinode *hdir;
++ struct inode *dir, *delegated;
++ struct path hinopath;
++ struct qstr hinoname = QSTR_INIT(AUFS_WH_DR_BRHINO,
++ sizeof(AUFS_WH_DR_BRHINO) - 1);
++
++ AuDebugOn(bindex < 0 && !br);
++ AuDebugOn(bindex >= 0 && br);
++
++ err = -EINVAL;
++ suspend = !br;
++ if (suspend)
++ br = au_sbr(sb, bindex);
++ load = !!path;
++ if (!load) {
++ path = &br->br_path;
++ AuDebugOn(!au_br_writable(br->br_perm));
++ if (unlikely(!au_br_writable(br->br_perm)))
++ goto out;
++ }
++
++ hdir = NULL;
++ if (suspend) {
++ dir = d_inode(sb->s_root);
++ hdir = au_hinode(au_ii(dir), bindex);
++ dir = hdir->hi_inode;
++ au_hn_inode_lock_nested(hdir, AuLsc_I_CHILD);
++ } else {
++ dir = d_inode(path->dentry);
++ inode_lock_nested(dir, AuLsc_I_CHILD);
++ }
++ hinopath.dentry = vfsub_lkup_one(&hinoname, path->dentry);
++ err = PTR_ERR(hinopath.dentry);
++ if (IS_ERR(hinopath.dentry))
++ goto out_unlock;
++
++ err = 0;
++ flags = O_RDONLY;
++ if (load) {
++ if (d_is_negative(hinopath.dentry))
++ goto out_dput; /* success */
++ } else {
++ if (au_dr_hino_test_empty(&br->br_dirren)) {
++ if (d_is_positive(hinopath.dentry)) {
++ delegated = NULL;
++ err = vfsub_unlink(dir, &hinopath, &delegated,
++ /*force*/0);
++ AuTraceErr(err);
++ if (unlikely(err))
++ pr_err("ignored err %d, %pd2\n",
++ err, hinopath.dentry);
++ if (unlikely(err == -EWOULDBLOCK))
++ iput(delegated);
++ err = 0;
++ }
++ goto out_dput;
++ } else if (!d_is_positive(hinopath.dentry)) {
++ err = vfsub_create(dir, &hinopath, 0600,
++ /*want_excl*/false);
++ AuTraceErr(err);
++ if (unlikely(err))
++ goto out_dput;
++ }
++ flags = O_WRONLY;
++ }
++ hinopath.mnt = path->mnt;
++ hinofile = vfsub_dentry_open(&hinopath, flags);
++ if (suspend)
++ au_hn_inode_unlock(hdir);
++ else
++ inode_unlock(dir);
++ dput(hinopath.dentry);
++ AuTraceErrPtr(hinofile);
++ if (IS_ERR(hinofile)) {
++ err = PTR_ERR(hinofile);
++ goto out;
++ }
++
++ if (load)
++ err = au_dr_hino_load(&br->br_dirren, hinofile);
++ else
++ err = au_dr_hino_store(sb, br, hinofile);
++ fput(hinofile);
++ goto out;
++
++out_dput:
++ dput(hinopath.dentry);
++out_unlock:
++ if (suspend)
++ au_hn_inode_unlock(hdir);
++ else
++ inode_unlock(dir);
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++/* ---------------------------------------------------------------------- */
++
++static int au_dr_brid_init(struct au_dr_brid *brid, const struct path *path)
++{
++ int err;
++ struct kstatfs kstfs;
++ dev_t dev;
++ struct dentry *dentry;
++ struct super_block *sb;
++
++ err = vfs_statfs((void *)path, &kstfs);
++ AuTraceErr(err);
++ if (unlikely(err))
++ goto out;
++
++ /* todo: support for UUID */
++
++ if (kstfs.f_fsid.val[0] || kstfs.f_fsid.val[1]) {
++ brid->type = AuBrid_FSID;
++ brid->fsid = kstfs.f_fsid;
++ } else {
++ dentry = path->dentry;
++ sb = dentry->d_sb;
++ dev = sb->s_dev;
++ if (dev) {
++ brid->type = AuBrid_DEV;
++ brid->dev = dev;
++ }
++ }
++
++out:
++ return err;
++}
++
++int au_dr_br_init(struct super_block *sb, struct au_branch *br,
++ const struct path *path)
++{
++ int err, i;
++ struct au_dr_br *dr;
++ struct hlist_bl_head *hbl;
++
++ dr = &br->br_dirren;
++ hbl = dr->dr_h_ino;
++ for (i = 0; i < AuDirren_NHASH; i++, hbl++)
++ INIT_HLIST_BL_HEAD(hbl);
++
++ err = au_dr_brid_init(&dr->dr_brid, path);
++ if (unlikely(err))
++ goto out;
++
++ if (au_opt_test(au_mntflags(sb), DIRREN))
++ err = au_dr_hino(sb, /*bindex*/-1, br, path);
++
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++int au_dr_br_fin(struct super_block *sb, struct au_branch *br)
++{
++ int err;
++
++ err = 0;
++ if (au_br_writable(br->br_perm))
++ err = au_dr_hino(sb, /*bindex*/-1, br, /*path*/NULL);
++ if (!err)
++ au_dr_hino_free(&br->br_dirren);
++
++ return err;
++}
++
++/* ---------------------------------------------------------------------- */
++
++static int au_brid_str(struct au_dr_brid *brid, struct inode *h_inode,
++ char *buf, size_t sz)
++{
++ int err;
++ unsigned int major, minor;
++ char *p;
++
++ p = buf;
++ err = snprintf(p, sz, "%d_", brid->type);
++ AuDebugOn(err > sz);
++ p += err;
++ sz -= err;
++ switch (brid->type) {
++ case AuBrid_Unset:
++ return -EINVAL;
++ case AuBrid_UUID:
++ err = snprintf(p, sz, "%pU", brid->uuid.b);
++ break;
++ case AuBrid_FSID:
++ err = snprintf(p, sz, "%08x-%08x",
++ brid->fsid.val[0], brid->fsid.val[1]);
++ break;
++ case AuBrid_DEV:
++ major = MAJOR(brid->dev);
++ minor = MINOR(brid->dev);
++ if (major <= 0xff && minor <= 0xff)
++ err = snprintf(p, sz, "%02x%02x", major, minor);
++ else
++ err = snprintf(p, sz, "%03x:%05x", major, minor);
++ break;
++ }
++ AuDebugOn(err > sz);
++ p += err;
++ sz -= err;
++ err = snprintf(p, sz, "_%llu", (unsigned long long)h_inode->i_ino);
++ AuDebugOn(err > sz);
++ p += err;
++ sz -= err;
++
++ return p - buf;
++}
++
++static int au_drinfo_name(struct au_branch *br, char *name, int len)
++{
++ int rlen;
++ struct dentry *br_dentry;
++ struct inode *br_inode;
++
++ br_dentry = au_br_dentry(br);
++ br_inode = d_inode(br_dentry);
++ rlen = au_brid_str(&br->br_dirren.dr_brid, br_inode, name, len);
++ AuDebugOn(rlen >= AUFS_DIRREN_ENV_VAL_SZ);
++ AuDebugOn(rlen > len);
++
++ return rlen;
++}
++
++/* ---------------------------------------------------------------------- */
++
++/*
++ * from the given @h_dentry, construct drinfo at @*fdata.
++ * when the size of @*fdata is not enough, reallocate and return new @fdata and
++ * @allocated.
++ */
++static int au_drinfo_construct(struct au_drinfo_fdata **fdata,
++ struct dentry *h_dentry,
++ unsigned char *allocated)
++{
++ int err, v;
++ struct au_drinfo_fdata *f, *p;
++ struct au_drinfo *drinfo;
++ struct inode *h_inode;
++ struct qstr *qname;
++
++ err = 0;
++ f = *fdata;
++ h_inode = d_inode(h_dentry);
++ qname = &h_dentry->d_name;
++ drinfo = &f->drinfo;
++ drinfo->ino = (__force uint64_t)cpu_to_be64(h_inode->i_ino);
++ drinfo->oldnamelen = qname->len;
++ if (*allocated < sizeof(*f) + qname->len) {
++ v = roundup_pow_of_two(*allocated + qname->len);
++ p = au_krealloc(f, v, GFP_NOFS, /*may_shrink*/0);
++ if (unlikely(!p)) {
++ err = -ENOMEM;
++ AuTraceErr(err);
++ goto out;
++ }
++ f = p;
++ *fdata = f;
++ *allocated = v;
++ drinfo = &f->drinfo;
++ }
++ memcpy(drinfo->oldname, qname->name, qname->len);
++ AuDbg("i%llu, %.*s\n",
++ be64_to_cpu((__force __be64)drinfo->ino), drinfo->oldnamelen,
++ drinfo->oldname);
++
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++/* callers have to free the return value */
++static struct au_drinfo *au_drinfo_read_k(struct file *file, ino_t h_ino)
++{
++ struct au_drinfo *ret, *drinfo;
++ struct au_drinfo_fdata fdata;
++ int len;
++ loff_t pos;
++ ssize_t ssz;
++
++ ret = ERR_PTR(-EIO);
++ pos = 0;
++ ssz = vfsub_read_k(file, &fdata, sizeof(fdata), &pos);
++ if (unlikely(ssz != sizeof(fdata))) {
++ AuIOErr("ssz %zd, %u, %pD2\n",
++ ssz, (unsigned int)sizeof(fdata), file);
++ goto out;
++ }
++
++ fdata.magic = ntohl((__force __be32)fdata.magic);
++ switch (fdata.magic) {
++ case AUFS_DRINFO_MAGIC_V1:
++ break;
++ default:
++ AuIOErr("magic-num 0x%x, 0x%x, %pD2\n",
++ fdata.magic, AUFS_DRINFO_MAGIC_V1, file);
++ goto out;
++ }
++
++ drinfo = &fdata.drinfo;
++ len = drinfo->oldnamelen;
++ if (!len) {
++ AuIOErr("broken drinfo %pD2\n", file);
++ goto out;
++ }
++
++ ret = NULL;
++ drinfo->ino = be64_to_cpu((__force __be64)drinfo->ino);
++ if (unlikely(h_ino && drinfo->ino != h_ino)) {
++ AuDbg("ignored i%llu, i%llu, %pD2\n",
++ (unsigned long long)drinfo->ino,
++ (unsigned long long)h_ino, file);
++ goto out; /* success */
++ }
++
++ ret = kmalloc(sizeof(*ret) + len, GFP_NOFS);
++ if (unlikely(!ret)) {
++ ret = ERR_PTR(-ENOMEM);
++ AuTraceErrPtr(ret);
++ goto out;
++ }
++
++ *ret = *drinfo;
++ ssz = vfsub_read_k(file, (void *)ret->oldname, len, &pos);
++ if (unlikely(ssz != len)) {
++ kfree(ret);
++ ret = ERR_PTR(-EIO);
++ AuIOErr("ssz %zd, %u, %pD2\n", ssz, len, file);
++ goto out;
++ }
++
++ AuDbg("oldname %.*s\n", ret->oldnamelen, ret->oldname);
++
++out:
++ return ret;
++}
++
++/* ---------------------------------------------------------------------- */
++
++/* in order to be revertible */
++struct au_drinfo_rev_elm {
++ int created;
++ struct dentry *info_dentry;
++ struct au_drinfo *info_last;
++};
++
++struct au_drinfo_rev {
++ unsigned char already;
++ aufs_bindex_t nelm;
++ struct au_drinfo_rev_elm elm[0];
++};
++
++/* todo: isn't it too large? */
++struct au_drinfo_store {
++ struct path h_ppath;
++ struct dentry *h_dentry;
++ struct au_drinfo_fdata *fdata;
++ char *infoname; /* inside of whname, just after PFX */
++ char whname[sizeof(AUFS_WH_DR_INFO_PFX) + AUFS_DIRREN_ENV_VAL_SZ];
++ aufs_bindex_t btgt, btail;
++ unsigned char no_sio,
++ allocated, /* current size of *fdata */
++ infonamelen, /* room size for p */
++ whnamelen, /* length of the genarated name */
++ renameback; /* renamed back */
++};
++
++/* on rename(2) error, the caller should revert it using @elm */
++static int au_drinfo_do_store(struct au_drinfo_store *w,
++ struct au_drinfo_rev_elm *elm)
++{
++ int err, len;
++ ssize_t ssz;
++ loff_t pos;
++ struct path infopath = {
++ .mnt = w->h_ppath.mnt
++ };
++ struct inode *h_dir, *h_inode, *delegated;
++ struct file *infofile;
++ struct qstr *qname;
++
++ AuDebugOn(elm
++ && memcmp(elm, page_address(ZERO_PAGE(0)), sizeof(*elm)));
++
++ infopath.dentry = vfsub_lookup_one_len(w->whname, w->h_ppath.dentry,
++ w->whnamelen);
++ AuTraceErrPtr(infopath.dentry);
++ if (IS_ERR(infopath.dentry)) {
++ err = PTR_ERR(infopath.dentry);
++ goto out;
++ }
++
++ err = 0;
++ h_dir = d_inode(w->h_ppath.dentry);
++ if (elm && d_is_negative(infopath.dentry)) {
++ err = vfsub_create(h_dir, &infopath, 0600, /*want_excl*/true);
++ AuTraceErr(err);
++ if (unlikely(err))
++ goto out_dput;
++ elm->created = 1;
++ elm->info_dentry = dget(infopath.dentry);
++ }
++
++ infofile = vfsub_dentry_open(&infopath, O_RDWR);
++ AuTraceErrPtr(infofile);
++ if (IS_ERR(infofile)) {
++ err = PTR_ERR(infofile);
++ goto out_dput;
++ }
++
++ h_inode = d_inode(infopath.dentry);
++ if (elm && i_size_read(h_inode)) {
++ h_inode = d_inode(w->h_dentry);
++ elm->info_last = au_drinfo_read_k(infofile, h_inode->i_ino);
++ AuTraceErrPtr(elm->info_last);
++ if (IS_ERR(elm->info_last)) {
++ err = PTR_ERR(elm->info_last);
++ elm->info_last = NULL;
++ AuDebugOn(elm->info_dentry);
++ goto out_fput;
++ }
++ }
++
++ if (elm && w->renameback) {
++ delegated = NULL;
++ err = vfsub_unlink(h_dir, &infopath, &delegated, /*force*/0);
++ AuTraceErr(err);
++ if (unlikely(err == -EWOULDBLOCK))
++ iput(delegated);
++ goto out_fput;
++ }
++
++ pos = 0;
++ qname = &w->h_dentry->d_name;
++ len = sizeof(*w->fdata) + qname->len;
++ if (!elm)
++ len = sizeof(*w->fdata) + w->fdata->drinfo.oldnamelen;
++ ssz = vfsub_write_k(infofile, w->fdata, len, &pos);
++ if (ssz == len) {
++ AuDbg("hi%llu, %.*s\n", w->fdata->drinfo.ino,
++ w->fdata->drinfo.oldnamelen, w->fdata->drinfo.oldname);
++ goto out_fput; /* success */
++ } else {
++ err = -EIO;
++ if (ssz < 0)
++ err = ssz;
++ /* the caller should revert it using @elm */
++ }
++
++out_fput:
++ fput(infofile);
++out_dput:
++ dput(infopath.dentry);
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++struct au_call_drinfo_do_store_args {
++ int *errp;
++ struct au_drinfo_store *w;
++ struct au_drinfo_rev_elm *elm;
++};
++
++static void au_call_drinfo_do_store(void *args)
++{
++ struct au_call_drinfo_do_store_args *a = args;
++
++ *a->errp = au_drinfo_do_store(a->w, a->elm);
++}
++
++static int au_drinfo_store_sio(struct au_drinfo_store *w,
++ struct au_drinfo_rev_elm *elm)
++{
++ int err, wkq_err;
++
++ if (w->no_sio)
++ err = au_drinfo_do_store(w, elm);
++ else {
++ struct au_call_drinfo_do_store_args a = {
++ .errp = &err,
++ .w = w,
++ .elm = elm
++ };
++ wkq_err = au_wkq_wait(au_call_drinfo_do_store, &a);
++ if (unlikely(wkq_err))
++ err = wkq_err;
++ }
++ AuTraceErr(err);
++
++ return err;
++}
++
++static int au_drinfo_store_work_init(struct au_drinfo_store *w,
++ aufs_bindex_t btgt)
++{
++ int err;
++
++ memset(w, 0, sizeof(*w));
++ w->allocated = roundup_pow_of_two(sizeof(*w->fdata) + 40);
++ strcpy(w->whname, AUFS_WH_DR_INFO_PFX);
++ w->infoname = w->whname + sizeof(AUFS_WH_DR_INFO_PFX) - 1;
++ w->infonamelen = sizeof(w->whname) - sizeof(AUFS_WH_DR_INFO_PFX);
++ w->btgt = btgt;
++ w->no_sio = !!uid_eq(current_fsuid(), GLOBAL_ROOT_UID);
++
++ err = -ENOMEM;
++ w->fdata = kcalloc(1, w->allocated, GFP_NOFS);
++ if (unlikely(!w->fdata)) {
++ AuTraceErr(err);
++ goto out;
++ }
++ w->fdata->magic = (__force uint32_t)htonl(AUFS_DRINFO_MAGIC_V1);
++ err = 0;
++
++out:
++ return err;
++}
++
++static void au_drinfo_store_work_fin(struct au_drinfo_store *w)
++{
++ kfree(w->fdata);
++}
++
++static void au_drinfo_store_rev(struct au_drinfo_rev *rev,
++ struct au_drinfo_store *w)
++{
++ struct au_drinfo_rev_elm *elm;
++ struct inode *h_dir, *delegated;
++ int err, nelm;
++ struct path infopath = {
++ .mnt = w->h_ppath.mnt
++ };
++
++ h_dir = d_inode(w->h_ppath.dentry);
++ IMustLock(h_dir);
++
++ err = 0;
++ elm = rev->elm;
++ for (nelm = rev->nelm; nelm > 0; nelm--, elm++) {
++ AuDebugOn(elm->created && elm->info_last);
++ if (elm->created) {
++ AuDbg("here\n");
++ delegated = NULL;
++ infopath.dentry = elm->info_dentry;
++ err = vfsub_unlink(h_dir, &infopath, &delegated,
++ !w->no_sio);
++ AuTraceErr(err);
++ if (unlikely(err == -EWOULDBLOCK))
++ iput(delegated);
++ dput(elm->info_dentry);
++ } else if (elm->info_last) {
++ AuDbg("here\n");
++ w->fdata->drinfo = *elm->info_last;
++ memcpy(w->fdata->drinfo.oldname,
++ elm->info_last->oldname,
++ elm->info_last->oldnamelen);
++ err = au_drinfo_store_sio(w, /*elm*/NULL);
++ kfree(elm->info_last);
++ }
++ if (unlikely(err))
++ AuIOErr("%d, %s\n", err, w->whname);
++ /* go on even if err */
++ }
++}
++
++/* caller has to call au_dr_rename_fin() later */
++static int au_drinfo_store(struct dentry *dentry, aufs_bindex_t btgt,
++ struct qstr *dst_name, void *_rev)
++{
++ int err, sz, nelm;
++ aufs_bindex_t bindex, btail;
++ struct au_drinfo_store work;
++ struct au_drinfo_rev *rev, **p;
++ struct au_drinfo_rev_elm *elm;
++ struct super_block *sb;
++ struct au_branch *br;
++ struct au_hinode *hdir;
++
++ err = au_drinfo_store_work_init(&work, btgt);
++ AuTraceErr(err);
++ if (unlikely(err))
++ goto out;
++
++ err = -ENOMEM;
++ btail = au_dbtaildir(dentry);
++ nelm = btail - btgt;
++ sz = sizeof(*rev) + sizeof(*elm) * nelm;
++ rev = kcalloc(1, sz, GFP_NOFS);
++ if (unlikely(!rev)) {
++ AuTraceErr(err);
++ goto out_args;
++ }
++ rev->nelm = nelm;
++ elm = rev->elm;
++ p = _rev;
++ *p = rev;
++
++ err = 0;
++ sb = dentry->d_sb;
++ work.h_ppath.dentry = au_h_dptr(dentry, btgt);
++ work.h_ppath.mnt = au_sbr_mnt(sb, btgt);
++ hdir = au_hi(d_inode(dentry), btgt);
++ au_hn_inode_lock_nested(hdir, AuLsc_I_CHILD);
++ for (bindex = btgt + 1; bindex <= btail; bindex++, elm++) {
++ work.h_dentry = au_h_dptr(dentry, bindex);
++ if (!work.h_dentry)
++ continue;
++
++ err = au_drinfo_construct(&work.fdata, work.h_dentry,
++ &work.allocated);
++ AuTraceErr(err);
++ if (unlikely(err))
++ break;
++
++ work.renameback = au_qstreq(&work.h_dentry->d_name, dst_name);
++ br = au_sbr(sb, bindex);
++ work.whnamelen = sizeof(AUFS_WH_DR_INFO_PFX) - 1;
++ work.whnamelen += au_drinfo_name(br, work.infoname,
++ work.infonamelen);
++ AuDbg("whname %.*s, i%llu, %.*s\n",
++ work.whnamelen, work.whname,
++ be64_to_cpu((__force __be64)work.fdata->drinfo.ino),
++ work.fdata->drinfo.oldnamelen,
++ work.fdata->drinfo.oldname);
++
++ err = au_drinfo_store_sio(&work, elm);
++ AuTraceErr(err);
++ if (unlikely(err))
++ break;
++ }
++ if (unlikely(err)) {
++ /* revert all drinfo */
++ au_drinfo_store_rev(rev, &work);
++ kfree(rev);
++ *p = NULL;
++ }
++ au_hn_inode_unlock(hdir);
++
++out_args:
++ au_drinfo_store_work_fin(&work);
++out:
++ return err;
++}
++
++/* ---------------------------------------------------------------------- */
++
++int au_dr_rename(struct dentry *src, aufs_bindex_t bindex,
++ struct qstr *dst_name, void *_rev)
++{
++ int err, already;
++ ino_t ino;
++ struct super_block *sb;
++ struct au_branch *br;
++ struct au_dr_br *dr;
++ struct dentry *h_dentry;
++ struct inode *h_inode;
++ struct au_dr_hino *ent;
++ struct au_drinfo_rev *rev, **p;
++
++ AuDbg("bindex %d\n", bindex);
++
++ err = -ENOMEM;
++ ent = kmalloc(sizeof(*ent), GFP_NOFS);
++ if (unlikely(!ent))
++ goto out;
++
++ sb = src->d_sb;
++ br = au_sbr(sb, bindex);
++ dr = &br->br_dirren;
++ h_dentry = au_h_dptr(src, bindex);
++ h_inode = d_inode(h_dentry);
++ ino = h_inode->i_ino;
++ ent->dr_h_ino = ino;
++ already = au_dr_hino_test_add(dr, ino, ent);
++ AuDbg("b%d, hi%llu, already %d\n",
++ bindex, (unsigned long long)ino, already);
++
++ err = au_drinfo_store(src, bindex, dst_name, _rev);
++ AuTraceErr(err);
++ if (!err) {
++ p = _rev;
++ rev = *p;
++ rev->already = already;
++ goto out; /* success */
++ }
++
++ /* revert */
++ if (!already)
++ au_dr_hino_del(dr, ent);
++ kfree(ent);
++
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++void au_dr_rename_fin(struct dentry *src, aufs_bindex_t btgt, void *_rev)
++{
++ struct au_drinfo_rev *rev;
++ struct au_drinfo_rev_elm *elm;
++ int nelm;
++
++ rev = _rev;
++ elm = rev->elm;
++ for (nelm = rev->nelm; nelm > 0; nelm--, elm++) {
++ dput(elm->info_dentry);
++ kfree(elm->info_last);
++ }
++ kfree(rev);
++}
++
++void au_dr_rename_rev(struct dentry *src, aufs_bindex_t btgt, void *_rev)
++{
++ int err;
++ struct au_drinfo_store work;
++ struct au_drinfo_rev *rev = _rev;
++ struct super_block *sb;
++ struct au_branch *br;
++ struct inode *h_inode;
++ struct au_dr_br *dr;
++ struct au_dr_hino *ent;
++
++ err = au_drinfo_store_work_init(&work, btgt);
++ if (unlikely(err))
++ goto out;
++
++ sb = src->d_sb;
++ br = au_sbr(sb, btgt);
++ work.h_ppath.dentry = au_h_dptr(src, btgt);
++ work.h_ppath.mnt = au_br_mnt(br);
++ au_drinfo_store_rev(rev, &work);
++ au_drinfo_store_work_fin(&work);
++ if (rev->already)
++ goto out;
++
++ dr = &br->br_dirren;
++ h_inode = d_inode(work.h_ppath.dentry);
++ ent = au_dr_hino_find(dr, h_inode->i_ino);
++ BUG_ON(!ent);
++ au_dr_hino_del(dr, ent);
++ kfree(ent);
++
++out:
++ kfree(rev);
++ if (unlikely(err))
++ pr_err("failed to remove dirren info\n");
++}
++
++/* ---------------------------------------------------------------------- */
++
++static struct au_drinfo *au_drinfo_do_load(struct path *h_ppath,
++ char *whname, int whnamelen,
++ struct dentry **info_dentry)
++{
++ struct au_drinfo *drinfo;
++ struct file *f;
++ struct inode *h_dir;
++ struct path infopath;
++ int unlocked;
++
++ AuDbg("%pd/%.*s\n", h_ppath->dentry, whnamelen, whname);
++
++ *info_dentry = NULL;
++ drinfo = NULL;
++ unlocked = 0;
++ h_dir = d_inode(h_ppath->dentry);
++ vfsub_inode_lock_shared_nested(h_dir, AuLsc_I_PARENT);
++ infopath.dentry = vfsub_lookup_one_len(whname, h_ppath->dentry,
++ whnamelen);
++ if (IS_ERR(infopath.dentry)) {
++ drinfo = (void *)infopath.dentry;
++ goto out;
++ }
++
++ if (d_is_negative(infopath.dentry))
++ goto out_dput; /* success */
++
++ infopath.mnt = h_ppath->mnt;
++ f = vfsub_dentry_open(&infopath, O_RDONLY);
++ inode_unlock_shared(h_dir);
++ unlocked = 1;
++ if (IS_ERR(f)) {
++ drinfo = (void *)f;
++ goto out_dput;
++ }
++
++ drinfo = au_drinfo_read_k(f, /*h_ino*/0);
++ if (IS_ERR_OR_NULL(drinfo))
++ goto out_fput;
++
++ AuDbg("oldname %.*s\n", drinfo->oldnamelen, drinfo->oldname);
++ *info_dentry = dget(infopath.dentry); /* keep it alive */
++
++out_fput:
++ fput(f);
++out_dput:
++ dput(infopath.dentry);
++out:
++ if (!unlocked)
++ inode_unlock_shared(h_dir);
++ AuTraceErrPtr(drinfo);
++ return drinfo;
++}
++
++struct au_drinfo_do_load_args {
++ struct au_drinfo **drinfop;
++ struct path *h_ppath;
++ char *whname;
++ int whnamelen;
++ struct dentry **info_dentry;
++};
++
++static void au_call_drinfo_do_load(void *args)
++{
++ struct au_drinfo_do_load_args *a = args;
++
++ *a->drinfop = au_drinfo_do_load(a->h_ppath, a->whname, a->whnamelen,
++ a->info_dentry);
++}
++
++struct au_drinfo_load {
++ struct path h_ppath;
++ struct qstr *qname;
++ unsigned char no_sio;
++
++ aufs_bindex_t ninfo;
++ struct au_drinfo **drinfo;
++};
++
++static int au_drinfo_load(struct au_drinfo_load *w, aufs_bindex_t bindex,
++ struct au_branch *br)
++{
++ int err, wkq_err, whnamelen, e;
++ char whname[sizeof(AUFS_WH_DR_INFO_PFX) + AUFS_DIRREN_ENV_VAL_SZ]
++ = AUFS_WH_DR_INFO_PFX;
++ struct au_drinfo *drinfo;
++ struct qstr oldname;
++ struct inode *h_dir, *delegated;
++ struct dentry *info_dentry;
++ struct path infopath;
++
++ whnamelen = sizeof(AUFS_WH_DR_INFO_PFX) - 1;
++ whnamelen += au_drinfo_name(br, whname + whnamelen,
++ sizeof(whname) - whnamelen);
++ if (w->no_sio)
++ drinfo = au_drinfo_do_load(&w->h_ppath, whname, whnamelen,
++ &info_dentry);
++ else {
++ struct au_drinfo_do_load_args args = {
++ .drinfop = &drinfo,
++ .h_ppath = &w->h_ppath,
++ .whname = whname,
++ .whnamelen = whnamelen,
++ .info_dentry = &info_dentry
++ };
++ wkq_err = au_wkq_wait(au_call_drinfo_do_load, &args);
++ if (unlikely(wkq_err))
++ drinfo = ERR_PTR(wkq_err);
++ }
++ err = PTR_ERR(drinfo);
++ if (IS_ERR_OR_NULL(drinfo))
++ goto out;
++
++ err = 0;
++ oldname.len = drinfo->oldnamelen;
++ oldname.name = drinfo->oldname;
++ if (au_qstreq(w->qname, &oldname)) {
++ /* the name is renamed back */
++ kfree(drinfo);
++ drinfo = NULL;
++
++ infopath.dentry = info_dentry;
++ infopath.mnt = w->h_ppath.mnt;
++ h_dir = d_inode(w->h_ppath.dentry);
++ delegated = NULL;
++ inode_lock_nested(h_dir, AuLsc_I_PARENT);
++ e = vfsub_unlink(h_dir, &infopath, &delegated, !w->no_sio);
++ inode_unlock(h_dir);
++ if (unlikely(e))
++ AuIOErr("ignored %d, %pd2\n", e, &infopath.dentry);
++ if (unlikely(e == -EWOULDBLOCK))
++ iput(delegated);
++ }
++ kfree(w->drinfo[bindex]);
++ w->drinfo[bindex] = drinfo;
++ dput(info_dentry);
++
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++/* ---------------------------------------------------------------------- */
++
++static void au_dr_lkup_free(struct au_drinfo **drinfo, int n)
++{
++ struct au_drinfo **p = drinfo;
++
++ while (n-- > 0)
++ kfree(*drinfo++);
++ kfree(p);
++}
++
++int au_dr_lkup(struct au_do_lookup_args *lkup, struct dentry *dentry,
++ aufs_bindex_t btgt)
++{
++ int err, ninfo;
++ struct au_drinfo_load w;
++ aufs_bindex_t bindex, bbot;
++ struct au_branch *br;
++ struct inode *h_dir;
++ struct au_dr_hino *ent;
++ struct super_block *sb;
++
++ AuDbg("%.*s, name %.*s, whname %.*s, b%d\n",
++ AuLNPair(&dentry->d_name), AuLNPair(&lkup->dirren.dr_name),
++ AuLNPair(&lkup->whname), btgt);
++
++ sb = dentry->d_sb;
++ bbot = au_sbbot(sb);
++ w.ninfo = bbot + 1;
++ if (!lkup->dirren.drinfo) {
++ lkup->dirren.drinfo = kcalloc(w.ninfo,
++ sizeof(*lkup->dirren.drinfo),
++ GFP_NOFS);
++ if (unlikely(!lkup->dirren.drinfo)) {
++ err = -ENOMEM;
++ goto out;
++ }
++ lkup->dirren.ninfo = w.ninfo;
++ }
++ w.drinfo = lkup->dirren.drinfo;
++ w.no_sio = !!uid_eq(current_fsuid(), GLOBAL_ROOT_UID);
++ w.h_ppath.dentry = au_h_dptr(dentry, btgt);
++ AuDebugOn(!w.h_ppath.dentry);
++ w.h_ppath.mnt = au_sbr_mnt(sb, btgt);
++ w.qname = &dentry->d_name;
++
++ ninfo = 0;
++ for (bindex = btgt + 1; bindex <= bbot; bindex++) {
++ br = au_sbr(sb, bindex);
++ err = au_drinfo_load(&w, bindex, br);
++ if (unlikely(err))
++ goto out_free;
++ if (w.drinfo[bindex])
++ ninfo++;
++ }
++ if (!ninfo) {
++ br = au_sbr(sb, btgt);
++ h_dir = d_inode(w.h_ppath.dentry);
++ ent = au_dr_hino_find(&br->br_dirren, h_dir->i_ino);
++ AuDebugOn(!ent);
++ au_dr_hino_del(&br->br_dirren, ent);
++ kfree(ent);
++ }
++ goto out; /* success */
++
++out_free:
++ au_dr_lkup_free(lkup->dirren.drinfo, lkup->dirren.ninfo);
++ lkup->dirren.ninfo = 0;
++ lkup->dirren.drinfo = NULL;
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++void au_dr_lkup_fin(struct au_do_lookup_args *lkup)
++{
++ au_dr_lkup_free(lkup->dirren.drinfo, lkup->dirren.ninfo);
++}
++
++int au_dr_lkup_name(struct au_do_lookup_args *lkup, aufs_bindex_t btgt)
++{
++ int err;
++ struct au_drinfo *drinfo;
++
++ err = 0;
++ if (!lkup->dirren.drinfo)
++ goto out;
++ AuDebugOn(lkup->dirren.ninfo < btgt + 1);
++ drinfo = lkup->dirren.drinfo[btgt + 1];
++ if (!drinfo)
++ goto out;
++
++ kfree(lkup->whname.name);
++ lkup->whname.name = NULL;
++ lkup->dirren.dr_name.len = drinfo->oldnamelen;
++ lkup->dirren.dr_name.name = drinfo->oldname;
++ lkup->name = &lkup->dirren.dr_name;
++ err = au_wh_name_alloc(&lkup->whname, lkup->name);
++ if (!err)
++ AuDbg("name %.*s, whname %.*s, b%d\n",
++ AuLNPair(lkup->name), AuLNPair(&lkup->whname),
++ btgt);
++
++out:
++ AuTraceErr(err);
++ return err;
++}
++
++int au_dr_lkup_h_ino(struct au_do_lookup_args *lkup, aufs_bindex_t bindex,
++ ino_t h_ino)
++{
++ int match;
++ struct au_drinfo *drinfo;
++
++ match = 1;
++ if (!lkup->dirren.drinfo)
++ goto out;
++ AuDebugOn(lkup->dirren.ninfo < bindex + 1);
++ drinfo = lkup->dirren.drinfo[bindex + 1];
++ if (!drinfo)
++ goto out;
++
++ match = (drinfo->ino == h_ino);
++ AuDbg("match %d\n", match);
++
++out:
++ return match;
++}
++
++/* ---------------------------------------------------------------------- */
++
++int au_dr_opt_set(struct super_block *sb)
++{
++ int err;
++ aufs_bindex_t bindex, bbot;
++ struct au_branch *br;
++
++ err = 0;
++ bbot = au_sbbot(sb);
++ for (bindex = 0; !err && bindex <= bbot; bindex++) {
++ br = au_sbr(sb, bindex);
++ err = au_dr_hino(sb, bindex, /*br*/NULL, &br->br_path);
++ }
++
++ return err;
++}
++
++int au_dr_opt_flush(struct super_block *sb)
++{
++ int err;
++ aufs_bindex_t bindex, bbot;
++ struct au_branch *br;
++
++ err = 0;
++ bbot = au_sbbot(sb);
++ for (bindex = 0; !err && bindex <= bbot; bindex++) {
++ br = au_sbr(sb, bindex);
++ if (au_br_writable(br->br_perm))
++ err = au_dr_hino(sb, bindex, /*br*/NULL, /*path*/NULL);
++ }
++
++ return err;
++}
++
++int au_dr_opt_clr(struct super_block *sb, int no_flush)
++{
++ int err;
++ aufs_bindex_t bindex, bbot;
++ struct au_branch *br;
++
++ err = 0;
++ if (!no_flush) {
++ err = au_dr_opt_flush(sb);
++ if (unlikely(err))
++ goto out;
++ }
++
++ bbot = au_sbbot(sb);
++ for (bindex = 0; bindex <= bbot; bindex++) {
++ br = au_sbr(sb, bindex);
++ au_dr_hino_free(&br->br_dirren);
++ }
++
++out:
++ return err;
++}
+diff -urN /usr/share/empty/fs/aufs/dirren.h linux/fs/aufs/dirren.h
+--- /usr/share/empty/fs/aufs/dirren.h 1970-01-01 01:00:00.000000000 +0100
++++ linux/fs/aufs/dirren.h 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,139 @@
++/*
++ * Copyright (C) 2017 Junjiro R. Okajima
++ *
++ * This program, aufs is free software; you can redistribute it and/or modify
++ * it under the terms of the GNU General Public License as published by
++ * the Free Software Foundation; either version 2 of the License, or
++ * (at your option) any later version.
++ *
++ * This program is distributed in the hope that it will be useful,
++ * but WITHOUT ANY WARRANTY; without even the implied warranty of
++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
++ * GNU General Public License for more details.
++ *
++ * You should have received a copy of the GNU General Public License
++ * along with this program. If not, see <http://www.gnu.org/licenses/>.
++ */
++
++/*
++ * renamed dir info
++ */
++
++#ifndef __AUFS_DIRREN_H__
++#define __AUFS_DIRREN_H__
++
++#ifdef __KERNEL__
++
++#include <linux/dcache.h>
++#include <linux/statfs.h>
++#include <linux/uuid.h>
++#include "hbl.h"
++
++#define AuDirren_NHASH 100
++
++#ifdef CONFIG_AUFS_DIRREN
++enum au_brid_type {
++ AuBrid_Unset,
++ AuBrid_UUID,
++ AuBrid_FSID,
++ AuBrid_DEV
++};
++
++struct au_dr_brid {
++ enum au_brid_type type;
++ union {
++ uuid_t uuid; /* unimplemented yet */
++ fsid_t fsid;
++ dev_t dev;
++ };
++};
++
++/* 20 is the max digits length of ulong 64 */
++/* brid-type "_" uuid "_" inum */
++#define AUFS_DIRREN_FNAME_SZ (1 + 1 + UUID_STRING_LEN + 20)
++#define AUFS_DIRREN_ENV_VAL_SZ (AUFS_DIRREN_FNAME_SZ + 1 + 20)
++
++struct au_dr_hino {
++ struct hlist_bl_node dr_hnode;
++ ino_t dr_h_ino;
++};
++
++struct au_dr_br {
++ struct hlist_bl_head dr_h_ino[AuDirren_NHASH];
++ struct au_dr_brid dr_brid;
++};
++
++struct au_dr_lookup {
++ /* dr_name is pointed by struct au_do_lookup_args.name */
++ struct qstr dr_name; /* subset of dr_info */
++ aufs_bindex_t ninfo;
++ struct au_drinfo **drinfo;
++};
++#else
++struct au_dr_hino;
++/* empty */
++struct au_dr_br { };
++struct au_dr_lookup { };
++#endif
++
++/* ---------------------------------------------------------------------- */
++
++struct au_branch;
++struct au_do_lookup_args;
++struct au_hinode;
++#ifdef CONFIG_AUFS_DIRREN
++int au_dr_hino_test_add(struct au_dr_br *dr, ino_t h_ino,
++ struct au_dr_hino *add_ent);
++void au_dr_hino_free(struct au_dr_br *dr);
++int au_dr_br_init(struct super_block *sb, struct au_branch *br,
++ const struct path *path);
++int au_dr_br_fin(struct super_block *sb, struct au_branch *br);
++int au_dr_rename(struct dentry *src, aufs_bindex_t bindex,
++ struct qstr *dst_name, void *_rev);
++void au_dr_rename_fin(struct dentry *src, aufs_bindex_t btgt, void *rev);
++void au_dr_rename_rev(struct dentry *src, aufs_bindex_t bindex, void *rev);
++int au_dr_lkup(struct au_do_lookup_args *lkup, struct dentry *dentry,
++ aufs_bindex_t bindex);
++int au_dr_lkup_name(struct au_do_lookup_args *lkup, aufs_bindex_t btgt);
++int au_dr_lkup_h_ino(struct au_do_lookup_args *lkup, aufs_bindex_t bindex,
++ ino_t h_ino);
++void au_dr_lkup_fin(struct au_do_lookup_args *lkup);
++int au_dr_opt_set(struct super_block *sb);
++int au_dr_opt_flush(struct super_block *sb);
++int au_dr_opt_clr(struct super_block *sb, int no_flush);
++#else
++AuStubInt0(au_dr_hino_test_add, struct au_dr_br *dr, ino_t h_ino,
++ struct au_dr_hino *add_ent);
++AuStubVoid(au_dr_hino_free, struct au_dr_br *dr);
++AuStubInt0(au_dr_br_init, struct super_block *sb, struct au_branch *br,
++ const struct path *path);
++AuStubInt0(au_dr_br_fin, struct super_block *sb, struct au_branch *br);
++AuStubInt0(au_dr_rename, struct dentry *src, aufs_bindex_t bindex,
++ struct qstr *dst_name, void *_rev);
++AuStubVoid(au_dr_rename_fin, struct dentry *src, aufs_bindex_t btgt, void *rev);
++AuStubVoid(au_dr_rename_rev, struct dentry *src, aufs_bindex_t bindex,
++ void *rev);
++AuStubInt0(au_dr_lkup, struct au_do_lookup_args *lkup, struct dentry *dentry,
++ aufs_bindex_t bindex);
++AuStubInt0(au_dr_lkup_name, struct au_do_lookup_args *lkup, aufs_bindex_t btgt);
++AuStubInt0(au_dr_lkup_h_ino, struct au_do_lookup_args *lkup,
++ aufs_bindex_t bindex, ino_t h_ino);
++AuStubVoid(au_dr_lkup_fin, struct au_do_lookup_args *lkup);
++AuStubInt0(au_dr_opt_set, struct super_block *sb);
++AuStubInt0(au_dr_opt_flush, struct super_block *sb);
++AuStubInt0(au_dr_opt_clr, struct super_block *sb, int no_flush);
++#endif
++
++/* ---------------------------------------------------------------------- */
++
++#ifdef CONFIG_AUFS_DIRREN
++static inline int au_dr_ihash(ino_t h_ino)
++{
++ return h_ino % AuDirren_NHASH;
++}
++#else
++AuStubInt0(au_dr_ihash, ino_t h_ino);
++#endif
++
++#endif /* __KERNEL__ */
++#endif /* __AUFS_DIRREN_H__ */
diff -urN /usr/share/empty/fs/aufs/dynop.c linux/fs/aufs/dynop.c
--- /usr/share/empty/fs/aufs/dynop.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dynop.c 2017-07-29 12:14:25.899708630 +0200
-@@ -0,0 +1,371 @@
++++ linux/fs/aufs/dynop.c 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,369 @@
+/*
+ * Copyright (C) 2010-2017 Junjiro R. Okajima
+ *
+ * How large will these lists be?
+ * Usually just a few elements, 20-30 at most for each, I guess.
+ */
-+static struct au_sphlhead dynop[AuDyLast];
++static struct hlist_bl_head dynop[AuDyLast];
+
-+static struct au_dykey *dy_gfind_get(struct au_sphlhead *sphl, const void *h_op)
++static struct au_dykey *dy_gfind_get(struct hlist_bl_head *hbl,
++ const void *h_op)
+{
+ struct au_dykey *key, *tmp;
-+ struct hlist_head *head;
++ struct hlist_bl_node *pos;
+
+ key = NULL;
-+ head = &sphl->head;
-+ rcu_read_lock();
-+ hlist_for_each_entry_rcu(tmp, head, dk_hnode)
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(tmp, pos, hbl, dk_hnode)
+ if (tmp->dk_op.dy_hop == h_op) {
+ key = tmp;
+ kref_get(&key->dk_kref);
+ break;
+ }
-+ rcu_read_unlock();
++ hlist_bl_unlock(hbl);
+
+ return key;
+}
+}
+
+/* kref_get() if @key is already added */
-+static struct au_dykey *dy_gadd(struct au_sphlhead *sphl, struct au_dykey *key)
++static struct au_dykey *dy_gadd(struct hlist_bl_head *hbl, struct au_dykey *key)
+{
+ struct au_dykey *tmp, *found;
-+ struct hlist_head *head;
++ struct hlist_bl_node *pos;
+ const void *h_op = key->dk_op.dy_hop;
+
+ found = NULL;
-+ head = &sphl->head;
-+ spin_lock(&sphl->spin);
-+ hlist_for_each_entry(tmp, head, dk_hnode)
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(tmp, pos, hbl, dk_hnode)
+ if (tmp->dk_op.dy_hop == h_op) {
+ kref_get(&tmp->dk_kref);
+ found = tmp;
+ break;
+ }
+ if (!found)
-+ hlist_add_head_rcu(&key->dk_hnode, head);
-+ spin_unlock(&sphl->spin);
++ hlist_bl_add_head(&key->dk_hnode, hbl);
++ hlist_bl_unlock(hbl);
+
+ if (!found)
+ DyPrSym(key);
+static void dy_free(struct kref *kref)
+{
+ struct au_dykey *key;
-+ struct au_sphlhead *sphl;
++ struct hlist_bl_head *hbl;
+
+ key = container_of(kref, struct au_dykey, dk_kref);
-+ sphl = dynop + key->dk_op.dy_type;
-+ au_sphl_del_rcu(&key->dk_hnode, sphl);
++ hbl = dynop + key->dk_op.dy_type;
++ au_hbl_del(&key->dk_hnode, hbl);
+ call_rcu(&key->dk_rcu, dy_free_rcu);
+}
+
+static struct au_dykey *dy_get(struct au_dynop *op, struct au_branch *br)
+{
+ struct au_dykey *key, *old;
-+ struct au_sphlhead *sphl;
++ struct hlist_bl_head *hbl;
+ struct op {
+ unsigned int sz;
+ void (*set)(struct au_dykey *key, const void *h_op,
+ };
+ const struct op *p;
+
-+ sphl = dynop + op->dy_type;
-+ key = dy_gfind_get(sphl, op->dy_hop);
++ hbl = dynop + op->dy_type;
++ key = dy_gfind_get(hbl, op->dy_hop);
+ if (key)
+ goto out_add; /* success */
+
+ key->dk_op.dy_hop = op->dy_hop;
+ kref_init(&key->dk_kref);
+ p->set(key, op->dy_hop, au_br_sb(br));
-+ old = dy_gadd(sphl, key);
++ old = dy_gadd(hbl, key);
+ if (old) {
+ kfree(key);
+ key = old;
+
+void au_dy_arefresh(int do_dx)
+{
-+ struct au_sphlhead *sphl;
-+ struct hlist_head *head;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos;
+ struct au_dykey *key;
+
-+ sphl = dynop + AuDy_AOP;
-+ head = &sphl->head;
-+ spin_lock(&sphl->spin);
-+ hlist_for_each_entry(key, head, dk_hnode)
++ hbl = dynop + AuDy_AOP;
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(key, pos, hbl, dk_hnode)
+ dy_adx((void *)key, do_dx);
-+ spin_unlock(&sphl->spin);
++ hlist_bl_unlock(hbl);
+}
+
+/* ---------------------------------------------------------------------- */
+ BUILD_BUG_ON(offsetof(struct au_dyaop, da_key));
+
+ for (i = 0; i < AuDyLast; i++)
-+ au_sphl_init(dynop + i);
++ INIT_HLIST_BL_HEAD(dynop + i);
+}
+
+void au_dy_fin(void)
+ int i;
+
+ for (i = 0; i < AuDyLast; i++)
-+ WARN_ON(!hlist_empty(&dynop[i].head));
++ WARN_ON(!hlist_bl_empty(dynop + i));
+}
diff -urN /usr/share/empty/fs/aufs/dynop.h linux/fs/aufs/dynop.h
--- /usr/share/empty/fs/aufs/dynop.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/dynop.h 2017-07-29 12:14:25.899708630 +0200
++++ linux/fs/aufs/dynop.h 2017-11-12 22:24:44.704244405 +0100
@@ -0,0 +1,74 @@
+/*
+ * Copyright (C) 2010-2017 Junjiro R. Okajima
+
+struct au_dykey {
+ union {
-+ struct hlist_node dk_hnode;
++ struct hlist_bl_node dk_hnode;
+ struct rcu_head dk_rcu;
+ };
+ struct au_dynop dk_op;
+#endif /* __AUFS_DYNOP_H__ */
diff -urN /usr/share/empty/fs/aufs/export.c linux/fs/aufs/export.c
--- /usr/share/empty/fs/aufs/export.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/export.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/export.c 2017-11-12 22:24:42.267510077 +0100
@@ -0,0 +1,836 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/file.c linux/fs/aufs/file.c
--- /usr/share/empty/fs/aufs/file.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/file.c 2017-07-29 12:14:25.903042072 +0200
-@@ -0,0 +1,858 @@
++++ linux/fs/aufs/file.c 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,856 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+
+static int au_cmoo(struct dentry *dentry)
+{
-+ int err, cmoo;
++ int err, cmoo, matched;
+ unsigned int udba;
+ struct path h_path;
+ struct au_pin pin;
+ sbinfo = au_sbi(sb);
+ fhsm = &sbinfo->si_fhsm;
+ pid = au_fhsm_pid(fhsm);
-+ if (pid
-+ && (current->pid == pid
-+ || current->real_parent->pid == pid))
++ rcu_read_lock();
++ matched = (pid
++ && (current->pid == pid
++ || rcu_dereference(current->real_parent)->pid == pid));
++ rcu_read_unlock();
++ if (matched)
+ goto out;
+
+ br = au_sbr(sb, cpg.bsrc);
+
+int au_do_open(struct file *file, struct au_do_open_args *args)
+{
-+ int err, no_lock = args->no_lock;
++ int err, aopen = args->aopen;
+ struct dentry *dentry;
+ struct au_finfo *finfo;
+
-+ if (!no_lock)
++ if (!aopen)
+ err = au_finfo_init(file, args->fidir);
+ else {
+ lockdep_off();
+
+ dentry = file->f_path.dentry;
+ AuDebugOn(IS_ERR_OR_NULL(dentry));
-+ if (!no_lock) {
-+ di_write_lock_child(dentry);
-+ err = au_cmoo(dentry);
-+ di_downgrade_lock(dentry, AuLock_IR);
-+ if (!err)
++ di_write_lock_child(dentry);
++ err = au_cmoo(dentry);
++ di_downgrade_lock(dentry, AuLock_IR);
++ if (!err) {
++ if (!aopen)
+ err = args->open(file, vfsub_file_flags(file), NULL);
-+ di_read_unlock(dentry, AuLock_IR);
-+ } else {
-+ err = au_cmoo(dentry);
-+ if (!err)
-+ err = args->open(file, vfsub_file_flags(file),
-+ args->h_file);
-+ if (!err && au_fbtop(file) != au_dbtop(dentry))
-+ /*
-+ * cmoo happens after h_file was opened.
-+ * need to refresh file later.
-+ */
-+ atomic_dec(&au_fi(file)->fi_generation);
++ else {
++ lockdep_off();
++ err = args->open(file, vfsub_file_flags(file), NULL);
++ lockdep_on();
++ }
+ }
++ di_read_unlock(dentry, AuLock_IR);
+
+ finfo = au_fi(file);
+ if (!err) {
+ finfo->fi_file = file;
-+ au_sphl_add(&finfo->fi_hlist,
-+ &au_sbi(file->f_path.dentry->d_sb)->si_files);
++ au_hbl_add(&finfo->fi_hlist,
++ &au_sbi(file->f_path.dentry->d_sb)->si_files);
+ }
-+ if (!no_lock)
++ if (!aopen)
+ fi_write_unlock(file);
+ else {
+ lockdep_off();
+ }
+
+out:
++ AuTraceErr(err);
+ return err;
+}
+
+};
diff -urN /usr/share/empty/fs/aufs/file.h linux/fs/aufs/file.h
--- /usr/share/empty/fs/aufs/file.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/file.h 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,331 @@
++++ linux/fs/aufs/file.h 2017-11-12 22:24:44.704244405 +0100
+@@ -0,0 +1,340 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ };
+ struct au_fidir *fi_hdir; /* for dir only */
+
-+ struct hlist_node fi_hlist;
++ struct hlist_bl_node fi_hlist;
+ struct file *fi_file; /* very ugly */
+} ____cacheline_aligned_in_smp;
+
+struct file *au_h_open(struct dentry *dentry, aufs_bindex_t bindex, int flags,
+ struct file *file, int force_wr);
+struct au_do_open_args {
-+ int no_lock;
++ int aopen;
+ int (*open)(struct file *file, int flags,
+ struct file *h_file);
+ struct au_fidir *fidir;
+
+/* ---------------------------------------------------------------------- */
+
++#define fi_read_lock(f) au_rw_read_lock(&au_fi(f)->fi_rwsem)
++#define fi_write_lock(f) au_rw_write_lock(&au_fi(f)->fi_rwsem)
++#define fi_read_trylock(f) au_rw_read_trylock(&au_fi(f)->fi_rwsem)
++#define fi_write_trylock(f) au_rw_write_trylock(&au_fi(f)->fi_rwsem)
+/*
-+ * fi_read_lock, fi_write_lock,
-+ * fi_read_unlock, fi_write_unlock, fi_downgrade_lock
-+ */
-+AuSimpleRwsemFuncs(fi, struct file *f, &au_fi(f)->fi_rwsem);
++#define fi_read_trylock_nested(f) \
++ au_rw_read_trylock_nested(&au_fi(f)->fi_rwsem)
++#define fi_write_trylock_nested(f) \
++ au_rw_write_trylock_nested(&au_fi(f)->fi_rwsem)
++*/
++
++#define fi_read_unlock(f) au_rw_read_unlock(&au_fi(f)->fi_rwsem)
++#define fi_write_unlock(f) au_rw_write_unlock(&au_fi(f)->fi_rwsem)
++#define fi_downgrade_lock(f) au_rw_dgrade_lock(&au_fi(f)->fi_rwsem)
+
+/* lock subclass for finfo */
+enum {
+#endif /* __AUFS_FILE_H__ */
diff -urN /usr/share/empty/fs/aufs/finfo.c linux/fs/aufs/finfo.c
--- /usr/share/empty/fs/aufs/finfo.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/finfo.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/finfo.c 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,148 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/f_op.c linux/fs/aufs/f_op.c
--- /usr/share/empty/fs/aufs/f_op.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/f_op.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/f_op.c 2017-11-12 22:24:44.704244405 +0100
@@ -0,0 +1,817 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ aufs_bindex_t bindex;
+
+ finfo = au_fi(file);
-+ au_sphl_del(&finfo->fi_hlist,
-+ &au_sbi(file->f_path.dentry->d_sb)->si_files);
++ au_hbl_del(&finfo->fi_hlist,
++ &au_sbi(file->f_path.dentry->d_sb)->si_files);
+ bindex = finfo->fi_btop;
+ if (bindex >= 0)
+ au_set_h_fptr(file, bindex, NULL);
+
+#endif /* __KERNEL__ */
+#endif /* __AUFS_FSTYPE_H__ */
+diff -urN /usr/share/empty/fs/aufs/hbl.h linux/fs/aufs/hbl.h
+--- /usr/share/empty/fs/aufs/hbl.h 1970-01-01 01:00:00.000000000 +0100
++++ linux/fs/aufs/hbl.h 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,64 @@
++/*
++ * Copyright (C) 2017 Junjiro R. Okajima
++ *
++ * This program, aufs is free software; you can redistribute it and/or modify
++ * it under the terms of the GNU General Public License as published by
++ * the Free Software Foundation; either version 2 of the License, or
++ * (at your option) any later version.
++ *
++ * This program is distributed in the hope that it will be useful,
++ * but WITHOUT ANY WARRANTY; without even the implied warranty of
++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
++ * GNU General Public License for more details.
++ *
++ * You should have received a copy of the GNU General Public License
++ * along with this program. If not, see <http://www.gnu.org/licenses/>.
++ */
++
++/*
++ * helpers for hlist_bl.h
++ */
++
++#ifndef __AUFS_HBL_H__
++#define __AUFS_HBL_H__
++
++#ifdef __KERNEL__
++
++#include <linux/list_bl.h>
++
++static inline void au_hbl_add(struct hlist_bl_node *node,
++ struct hlist_bl_head *hbl)
++{
++ hlist_bl_lock(hbl);
++ hlist_bl_add_head(node, hbl);
++ hlist_bl_unlock(hbl);
++}
++
++static inline void au_hbl_del(struct hlist_bl_node *node,
++ struct hlist_bl_head *hbl)
++{
++ hlist_bl_lock(hbl);
++ hlist_bl_del(node);
++ hlist_bl_unlock(hbl);
++}
++
++#define au_hbl_for_each(pos, head) \
++ for (pos = hlist_bl_first(head); \
++ pos; \
++ pos = pos->next)
++
++static inline unsigned long au_hbl_count(struct hlist_bl_head *hbl)
++{
++ unsigned long cnt;
++ struct hlist_bl_node *pos;
++
++ cnt = 0;
++ hlist_bl_lock(hbl);
++ au_hbl_for_each(pos, hbl)
++ cnt++;
++ hlist_bl_unlock(hbl);
++ return cnt;
++}
++
++#endif /* __KERNEL__ */
++#endif /* __AUFS_HBL_H__ */
diff -urN /usr/share/empty/fs/aufs/hfsnotify.c linux/fs/aufs/hfsnotify.c
--- /usr/share/empty/fs/aufs/hfsnotify.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/hfsnotify.c 2017-07-31 10:27:18.853311720 +0200
++++ linux/fs/aufs/hfsnotify.c 2017-11-12 22:24:44.707577830 +0100
@@ -0,0 +1,289 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ hn_mark);
+ /* AuDbg("here\n"); */
+ au_cache_free_hnotify(hn);
-+ smp_mb__before_atomic();
++ smp_mb__before_atomic(); /* for atomic64_dec */
+ if (atomic64_dec_and_test(&au_hfsn_ifree))
+ wake_up(&au_hfsn_wq);
+}
+}
diff -urN /usr/share/empty/fs/aufs/hnotify.c linux/fs/aufs/hnotify.c
--- /usr/share/empty/fs/aufs/hnotify.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/hnotify.c 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,711 @@
++++ linux/fs/aufs/hnotify.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,719 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ AuDebugOn(!sbinfo);
+ si_write_lock(sb, AuLock_NOPLMW);
+
++ if (au_opt_test(sbinfo->si_mntflags, DIRREN))
++ switch (a->mask & FS_EVENTS_POSS_ON_CHILD) {
++ case FS_MOVED_FROM:
++ case FS_MOVED_TO:
++ AuWarn1("DIRREN with UDBA may not work correctly "
++ "for the direct rename(2)\n");
++ }
++
+ ii_read_lock_parent(a->dir);
+ bfound = -1;
+ bbot = au_ibbot(a->dir);
+}
diff -urN /usr/share/empty/fs/aufs/iinfo.c linux/fs/aufs/iinfo.c
--- /usr/share/empty/fs/aufs/iinfo.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/iinfo.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/iinfo.c 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,285 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/inode.c linux/fs/aufs/inode.c
--- /usr/share/empty/fs/aufs/inode.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/inode.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/inode.c 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,527 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/inode.h linux/fs/aufs/inode.h
--- /usr/share/empty/fs/aufs/inode.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/inode.h 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,694 @@
++++ linux/fs/aufs/inode.h 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,695 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+
+#include <linux/fsnotify.h>
+#include "rwsem.h"
-+#include "vfsub.h"
+
+struct vfsmount;
+
+struct au_icntnr {
+ struct au_iinfo iinfo;
+ struct inode vfs_inode;
-+ struct hlist_node plink;
++ struct hlist_bl_node plink;
+} ____cacheline_aligned_in_smp;
+
+/* au_pin flags */
+#undef AuWriteLockFunc
+#undef AuRWLockFuncs
+
-+/*
-+ * ii_read_unlock, ii_write_unlock, ii_downgrade_lock
-+ */
-+AuSimpleUnlockRwsemFuncs(ii, struct inode *i, &au_ii(i)->ii_rwsem);
++#define ii_read_unlock(i) au_rw_read_unlock(&au_ii(i)->ii_rwsem)
++#define ii_write_unlock(i) au_rw_write_unlock(&au_ii(i)->ii_rwsem)
++#define ii_downgrade_lock(i) au_rw_dgrade_lock(&au_ii(i)->ii_rwsem)
+
+#define IiMustNoWaiters(i) AuRwMustNoWaiters(&au_ii(i)->ii_rwsem)
+#define IiMustAnyLock(i) AuRwMustAnyLock(&au_ii(i)->ii_rwsem)
+ au_hn_suspend(hdir);
+}
+
++#if 0 /* unused */
++#include "vfsub.h"
+static inline void au_hn_inode_lock_shared_nested(struct au_hinode *hdir,
+ unsigned int sc)
+{
+ vfsub_inode_lock_shared_nested(hdir->hi_inode, sc);
+ au_hn_suspend(hdir);
+}
++#endif
+
+static inline void au_hn_inode_unlock(struct au_hinode *hdir)
+{
+#endif
diff -urN /usr/share/empty/fs/aufs/i_op_add.c linux/fs/aufs/i_op_add.c
--- /usr/share/empty/fs/aufs/i_op_add.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/i_op_add.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/i_op_add.c 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,920 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/i_op.c linux/fs/aufs/i_op.c
--- /usr/share/empty/fs/aufs/i_op.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/i_op.c 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,1452 @@
++++ linux/fs/aufs/i_op.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,1459 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+/* ---------------------------------------------------------------------- */
+
+struct aopen_node {
-+ struct hlist_node hlist;
++ struct hlist_bl_node hblist;
+ struct file *file, *h_file;
+};
+
+static int au_do_aopen(struct inode *inode, struct file *file)
+{
-+ struct au_sphlhead *aopen;
++ struct hlist_bl_head *aopen;
++ struct hlist_bl_node *pos;
+ struct aopen_node *node;
+ struct au_do_open_args args = {
-+ .no_lock = 1,
-+ .open = au_do_open_nondir
++ .aopen = 1,
++ .open = au_do_open_nondir
+ };
+
+ aopen = &au_sbi(inode->i_sb)->si_aopen;
-+ spin_lock(&aopen->spin);
-+ hlist_for_each_entry(node, &aopen->head, hlist)
++ hlist_bl_lock(aopen);
++ hlist_bl_for_each_entry(node, pos, aopen, hblist)
+ if (node->file == file) {
+ args.h_file = node->h_file;
+ break;
+ }
-+ spin_unlock(&aopen->spin);
++ hlist_bl_unlock(aopen);
+ /* AuDebugOn(!args.h_file); */
+
+ return au_do_open(file, &args);
+ struct file *file, unsigned int open_flag,
+ umode_t create_mode, int *opened)
+{
-+ int err, h_opened = *opened;
++ int err, unlocked, h_opened = *opened;
+ unsigned int lkup_flags;
+ struct dentry *parent, *d;
-+ struct au_sphlhead *aopen;
++ struct hlist_bl_head *aopen;
+ struct vfsub_aopen_args args = {
+ .open_flag = open_flag,
+ .create_mode = create_mode,
+ || !(open_flag & O_CREAT))
+ goto out_no_open;
+
++ unlocked = 0;
+ err = aufs_read_lock(dentry, AuLock_DW | AuLock_FLUSH | AuLock_GEN);
+ if (unlikely(err))
+ goto out;
+ put_filp(args.file);
+ goto out_unlock;
+ }
++ di_write_unlock(parent);
++ di_write_unlock(dentry);
++ unlocked = 1;
+
+ /* some filesystems don't set FILE_CREATED while succeeded? */
+ *opened |= FILE_CREATED;
+ args.file = NULL;
+ }
+ aopen = &au_sbi(dir->i_sb)->si_aopen;
-+ au_sphl_add(&aopen_node.hlist, aopen);
++ au_hbl_add(&aopen_node.hblist, aopen);
+ err = finish_open(file, dentry, au_do_aopen, opened);
-+ au_sphl_del(&aopen_node.hlist, aopen);
++ au_hbl_del(&aopen_node.hblist, aopen);
+ AuTraceErr(err);
+ AuDbgFile(file);
+ if (aopen_node.h_file)
+ fput(aopen_node.h_file);
+
+out_unlock:
-+ di_write_unlock(parent);
-+ aufs_read_unlock(dentry, AuLock_DW);
++ if (unlocked)
++ si_read_unlock(dentry->d_sb);
++ else {
++ di_write_unlock(parent);
++ aufs_read_unlock(dentry, AuLock_DW);
++ }
+ AuDbgDentry(dentry);
+ if (unlikely(err < 0))
+ goto out;
+ lockdep_off();
+ si_read_lock(sb, AuLock_FLUSH);
+ ii_write_lock_child(inode);
-+ lockdep_on();
+
+ err = 0;
+ bindex = au_ibtop(inode);
+ AuDebugOn(1);
+ }
+
-+ lockdep_off();
+ if (!err)
+ au_cpup_attr_timesizes(inode);
+ ii_write_unlock(inode);
+};
diff -urN /usr/share/empty/fs/aufs/i_op_del.c linux/fs/aufs/i_op_del.c
--- /usr/share/empty/fs/aufs/i_op_del.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/i_op_del.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/i_op_del.c 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,511 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/i_op_ren.c linux/fs/aufs/i_op_ren.c
--- /usr/share/empty/fs/aufs/i_op_ren.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/i_op_ren.c 2017-07-29 12:14:25.903042072 +0200
-@@ -0,0 +1,1165 @@
++++ linux/fs/aufs/i_op_ren.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,1246 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#define AuRen_DT_DSTDIR (1 << 6)
+#define AuRen_DIROPQ_SRC (1 << 7)
+#define AuRen_DIROPQ_DST (1 << 8)
++#define AuRen_DIRREN (1 << 9)
++#define AuRen_DROPPED_SRC (1 << 10)
++#define AuRen_DROPPED_DST (1 << 11)
+#define au_ftest_ren(flags, name) ((flags) & AuRen_##name)
+#define au_fset_ren(flags, name) \
+ do { (flags) |= AuRen_##name; } while (0)
+#define au_fclr_ren(flags, name) \
+ do { (flags) &= ~AuRen_##name; } while (0)
+
++#ifndef CONFIG_AUFS_DIRREN
++#undef AuRen_DIRREN
++#define AuRen_DIRREN 0
++#endif
++
+struct au_ren_args {
+ struct {
+ struct dentry *dentry, *h_dentry, *parent, *h_parent,
+
+ struct au_whtmp_rmdir *thargs;
+ struct dentry *h_dst;
++ struct au_hinode *h_root;
+};
+
+/* ---------------------------------------------------------------------- */
+ d = a->dst_dentry; /* already renamed on the branch */
+ always = !!au_opt_test(au_mntflags(d->d_sb), ALWAYS_DIROPQ);
+ if (au_ftest_ren(a->auren_flags, ISDIR_SRC)
++ && !au_ftest_ren(a->auren_flags, DIRREN)
+ && a->btgt != au_dbdiropq(a->src_dentry)
+ && (a->dst_wh_dentry
+ || a->btgt <= au_dbdiropq(d)
+ /* prepare workqueue args for asynchronous rmdir */
+ h_d = a->dst_h_dentry;
+ if (au_ftest_ren(a->auren_flags, ISDIR_DST)
++ /* && !au_ftest_ren(a->auren_flags, DIRREN) */
+ && d_is_positive(h_d)) {
+ err = -ENOMEM;
+ a->thargs = au_whtmp_rmdir_alloc(a->src_dentry->d_sb,
+ }
+
+ BUG_ON(d_is_positive(a->dst_h_dentry) && a->src_btop != a->btgt);
++#if 0
++ BUG_ON(!au_ftest_ren(a->auren_flags, DIRREN)
++ && d_is_positive(a->dst_h_dentry)
++ && a->src_btop != a->btgt);
++#endif
+
+ /* rename by vfs_rename or cpup */
+ err = au_ren_or_cpup(a);
+}
+
+/*
-+ * test if @dentry dir can be rename source or not.
-+ * if it can, return 0 and @children is filled.
++ * test if @a->src_dentry dir can be rename source or not.
++ * if it can, return 0.
+ * success means,
+ * - it is a logically empty dir.
+ * - or, it exists on writable branch and has no children including whiteouts
-+ * on the lower branch.
++ * on the lower branch unless DIRREN is on.
+ */
-+static int may_rename_srcdir(struct dentry *dentry, aufs_bindex_t btgt)
++static int may_rename_srcdir(struct au_ren_args *a)
+{
+ int err;
+ unsigned int rdhash;
-+ aufs_bindex_t btop;
++ aufs_bindex_t btop, btgt;
++ struct dentry *dentry;
++ struct super_block *sb;
++ struct au_sbinfo *sbinfo;
+
++ dentry = a->src_dentry;
++ sb = dentry->d_sb;
++ sbinfo = au_sbi(sb);
++ if (au_opt_test(sbinfo->si_mntflags, DIRREN))
++ au_fset_ren(a->auren_flags, DIRREN);
++
++ btgt = a->btgt;
+ btop = au_dbtop(dentry);
+ if (btop != btgt) {
+ struct au_nhash whlist;
+
-+ SiMustAnyLock(dentry->d_sb);
-+ rdhash = au_sbi(dentry->d_sb)->si_rdhash;
++ SiMustAnyLock(sb);
++ rdhash = sbinfo->si_rdhash;
+ if (!rdhash)
+ rdhash = au_rdhash_est(au_dir_size(/*file*/NULL,
+ dentry));
+
+out:
+ if (err == -ENOTEMPTY) {
-+ AuWarn1("renaming dir who has child(ren) on multiple branches,"
-+ " is not supported\n");
-+ err = -EXDEV;
++ if (au_ftest_ren(a->auren_flags, DIRREN)) {
++ err = 0;
++ } else {
++ AuWarn1("renaming dir who has child(ren) on multiple "
++ "branches, is not supported\n");
++ err = -EXDEV;
++ }
+ }
+ return err;
+}
+ err = may_rename_dstdir(d, &a->whlist);
+ au_set_dbtop(d, a->btgt);
+ } else
-+ err = may_rename_srcdir(d, a->btgt);
++ err = may_rename_srcdir(a);
+ }
+ a->dst_h_dentry = au_h_dptr(d, au_dbtop(d));
+ if (unlikely(err))
+ d = a->src_dentry;
+ a->src_h_dentry = au_h_dptr(d, au_dbtop(d));
+ if (au_ftest_ren(a->auren_flags, ISDIR_SRC)) {
-+ err = may_rename_srcdir(d, a->btgt);
++ err = may_rename_srcdir(a);
+ if (unlikely(err)) {
+ au_nhash_wh_free(&a->whlist);
+ a->whlist.nh_num = 0;
+{
+ vfsub_unlock_rename(a->src_h_parent, a->src_hdir,
+ a->dst_h_parent, a->dst_hdir);
++ if (au_ftest_ren(a->auren_flags, DIRREN)
++ && a->h_root)
++ au_hn_inode_unlock(a->h_root);
+ if (au_ftest_ren(a->auren_flags, MNT_WRITE))
+ vfsub_mnt_drop_write(au_br_mnt(a->br));
+}
+ if (unlikely(err))
+ goto out;
+ au_fset_ren(a->auren_flags, MNT_WRITE);
++ if (au_ftest_ren(a->auren_flags, DIRREN)) {
++ struct dentry *root;
++ struct inode *dir;
++
++ /*
++ * sbinfo is already locked, so this ii_read_lock is
++ * unnecessary. but our debugging feature checks it.
++ */
++ root = a->src_inode->i_sb->s_root;
++ if (root != a->src_parent && root != a->dst_parent) {
++ dir = d_inode(root);
++ ii_read_lock_parent3(dir);
++ a->h_root = au_hi(dir, a->btgt);
++ ii_read_unlock(dir);
++ au_hn_inode_lock_nested(a->h_root, AuLsc_I_PARENT3);
++ }
++ }
+ a->h_trap = vfsub_lock_rename(a->src_h_parent, a->src_hdir,
+ a->dst_h_parent, a->dst_hdir);
+ udba = au_opt_udba(a->src_dentry->d_sb);
+ au_update_dbrange(d, /*do_put_zero*/0);
+ }
+
++ if (a->exchange
++ || au_ftest_ren(a->auren_flags, DIRREN)) {
++ d_drop(a->src_dentry);
++ if (au_ftest_ren(a->auren_flags, DIRREN))
++ au_set_dbwh(a->src_dentry, -1);
++ return;
++ }
++
+ d = a->src_dentry;
-+ if (!a->exchange) {
-+ au_set_dbwh(d, -1);
-+ bbot = au_dbbot(d);
-+ for (bindex = a->btgt + 1; bindex <= bbot; bindex++) {
-+ h_d = au_h_dptr(d, bindex);
-+ if (h_d)
-+ au_set_h_dptr(d, bindex, NULL);
-+ }
-+ au_set_dbbot(d, a->btgt);
++ au_set_dbwh(d, -1);
++ bbot = au_dbbot(d);
++ for (bindex = a->btgt + 1; bindex <= bbot; bindex++) {
++ h_d = au_h_dptr(d, bindex);
++ if (h_d)
++ au_set_h_dptr(d, bindex, NULL);
++ }
++ au_set_dbbot(d, a->btgt);
+
-+ sb = d->d_sb;
-+ i = a->src_inode;
-+ if (au_opt_test(au_mntflags(sb), PLINK) && au_plink_test(i))
-+ return; /* success */
++ sb = d->d_sb;
++ i = a->src_inode;
++ if (au_opt_test(au_mntflags(sb), PLINK) && au_plink_test(i))
++ return; /* success */
+
-+ bbot = au_ibbot(i);
-+ for (bindex = a->btgt + 1; bindex <= bbot; bindex++) {
-+ h_i = au_h_iptr(i, bindex);
-+ if (h_i) {
-+ au_xino_write(sb, bindex, h_i->i_ino, /*ino*/0);
-+ /* ignore this error */
-+ au_set_h_iptr(i, bindex, NULL, 0);
-+ }
++ bbot = au_ibbot(i);
++ for (bindex = a->btgt + 1; bindex <= bbot; bindex++) {
++ h_i = au_h_iptr(i, bindex);
++ if (h_i) {
++ au_xino_write(sb, bindex, h_i->i_ino, /*ino*/0);
++ /* ignore this error */
++ au_set_h_iptr(i, bindex, NULL, 0);
+ }
-+ au_set_ibbot(i, a->btgt);
+ }
-+ d_drop(a->src_dentry);
++ au_set_ibbot(i, a->btgt);
+}
+
+/* ---------------------------------------------------------------------- */
+ unsigned int _flags)
+{
+ int err, lock_flags;
++ void *rev;
+ /* reduce stack space */
+ struct au_ren_args *a;
+ struct au_pin pin;
+ goto out_free;
+ lock_flags |= AuLock_DIRS;
+ }
-+ err = aufs_read_and_write_lock2(a->dst_dentry, a->src_dentry, lock_flags);
++ err = aufs_read_and_write_lock2(a->dst_dentry, a->src_dentry,
++ lock_flags);
+ if (unlikely(err))
+ goto out_free;
+
+ /* store timestamps to be revertible */
+ au_ren_dt(a);
+
++ /* store dirren info */
++ if (au_ftest_ren(a->auren_flags, DIRREN)) {
++ err = au_dr_rename(a->src_dentry, a->btgt,
++ &a->dst_dentry->d_name, &rev);
++ AuTraceErr(err);
++ if (unlikely(err))
++ goto out_dt;
++ }
++
+ /* here we go */
+ err = do_rename(a);
+ if (unlikely(err))
-+ goto out_dt;
++ goto out_dirren;
++
++ if (au_ftest_ren(a->auren_flags, DIRREN))
++ au_dr_rename_fin(a->src_dentry, a->btgt, rev);
+
+ /* update dir attributes */
+ au_ren_refresh_dir(a);
+
+ goto out_hdir; /* success */
+
++out_dirren:
++ if (au_ftest_ren(a->auren_flags, DIRREN))
++ au_dr_rename_rev(a->src_dentry, a->btgt, rev);
+out_dt:
+ au_ren_rev_dt(err, a);
+out_hdir:
+ }
+out_parent:
+ if (!err) {
++ if (d_unhashed(a->src_dentry))
++ au_fset_ren(a->auren_flags, DROPPED_SRC);
++ if (d_unhashed(a->dst_dentry))
++ au_fset_ren(a->auren_flags, DROPPED_DST);
+ if (!a->exchange)
+ d_move(a->src_dentry, a->dst_dentry);
-+ else
++ else {
+ d_exchange(a->src_dentry, a->dst_dentry);
++ if (au_ftest_ren(a->auren_flags, DROPPED_DST))
++ d_drop(a->dst_dentry);
++ }
++ if (au_ftest_ren(a->auren_flags, DROPPED_SRC))
++ d_drop(a->src_dentry);
+ } else {
+ au_update_dbtop(a->dst_dentry);
+ if (!a->dst_inode)
+}
diff -urN /usr/share/empty/fs/aufs/Kconfig linux/fs/aufs/Kconfig
--- /usr/share/empty/fs/aufs/Kconfig 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/Kconfig 2017-07-29 12:14:25.896375188 +0200
-@@ -0,0 +1,185 @@
++++ linux/fs/aufs/Kconfig 2017-11-12 22:24:44.697577553 +0100
+@@ -0,0 +1,198 @@
+config AUFS_FS
+ tristate "Aufs (Advanced multi layered unification filesystem) support"
+ help
+ shows better performance in most cases.
+ See detail in aufs.5.
+
++config AUFS_DIRREN
++ bool "Workaround for rename(2)-ing a directory"
++ help
++ By default, aufs returns EXDEV error in renameing a dir who has
++ his child on the lower branch, since it is a bad idea to issue
++ rename(2) internally for every lower branch. But user may not
++ accept this behaviour. So here is a workaround to allow such
++ rename(2) and store some extra infromation on the writable
++ branch. Obviously this costs high (and I don't like it).
++ To use this feature, you need to enable this configuration AND
++ to specify the mount option `dirren.'
++ See details in aufs.5 and the design documents.
++
+config AUFS_SHWH
+ bool "Show whiteouts"
+ help
+endif
diff -urN /usr/share/empty/fs/aufs/loop.c linux/fs/aufs/loop.c
--- /usr/share/empty/fs/aufs/loop.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/loop.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/loop.c 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,147 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+endif
diff -urN /usr/share/empty/fs/aufs/Makefile linux/fs/aufs/Makefile
--- /usr/share/empty/fs/aufs/Makefile 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/Makefile 2017-07-29 12:14:25.896375188 +0200
-@@ -0,0 +1,44 @@
++++ linux/fs/aufs/Makefile 2017-11-12 22:24:44.697577553 +0100
+@@ -0,0 +1,45 @@
+
+include ${src}/magic.mk
+ifeq (${CONFIG_AUFS_FS},m)
+aufs-$(CONFIG_AUFS_EXPORT) += export.o
+aufs-$(CONFIG_AUFS_XATTR) += xattr.o
+aufs-$(CONFIG_FS_POSIX_ACL) += posix_acl.o
++aufs-$(CONFIG_AUFS_DIRREN) += dirren.o
+aufs-$(CONFIG_AUFS_FHSM) += fhsm.o
+aufs-$(CONFIG_AUFS_POLL) += poll.o
+aufs-$(CONFIG_AUFS_RDU) += rdu.o
+aufs-$(CONFIG_AUFS_MAGIC_SYSRQ) += sysrq.o
diff -urN /usr/share/empty/fs/aufs/module.c linux/fs/aufs/module.c
--- /usr/share/empty/fs/aufs/module.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/module.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/module.c 2017-11-12 22:24:44.707577830 +0100
@@ -0,0 +1,266 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ * iterate_supers_type() doesn't protect us from
+ * remounting (branch management)
+ */
-+struct au_sphlhead au_sbilist;
++struct hlist_bl_head au_sbilist;
+#endif
+
+/*
+module_exit(aufs_exit);
diff -urN /usr/share/empty/fs/aufs/module.h linux/fs/aufs/module.h
--- /usr/share/empty/fs/aufs/module.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/module.h 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/module.h 2017-11-12 22:24:42.270843503 +0100
@@ -0,0 +1,101 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+#endif /* __AUFS_MODULE_H__ */
diff -urN /usr/share/empty/fs/aufs/mvdown.c linux/fs/aufs/mvdown.c
--- /usr/share/empty/fs/aufs/mvdown.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/mvdown.c 2017-07-29 12:14:25.903042072 +0200
++++ linux/fs/aufs/mvdown.c 2017-11-12 22:24:44.707577830 +0100
@@ -0,0 +1,704 @@
+/*
+ * Copyright (C) 2011-2017 Junjiro R. Okajima
+ for (bindex++; bindex <= bbot; bindex++) {
+ br = au_sbr(sb, bindex);
+ if (au_br_fhsm(br->br_perm)
-+ && (!(au_br_sb(br)->s_flags & MS_RDONLY)))
++ && !sb_rdonly(au_br_sb(br)))
+ return bindex;
+ }
+ else if (!(a->mvdown.flags & AUFS_MVDOWN_ROLOWER))
+ else
+ for (bindex++; bindex <= bbot; bindex++) {
+ br = au_sbr(sb, bindex);
-+ if (!(au_br_sb(br)->s_flags & MS_RDONLY)) {
++ if (!sb_rdonly(au_br_sb(br))) {
+ if (au_br_rdonly(br))
+ a->mvdown.flags
+ |= AUFS_MVDOWN_ROLOWER_R;
+}
diff -urN /usr/share/empty/fs/aufs/opts.c linux/fs/aufs/opts.c
--- /usr/share/empty/fs/aufs/opts.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/opts.c 2017-07-29 12:14:25.903042072 +0200
-@@ -0,0 +1,1846 @@
++++ linux/fs/aufs/opts.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,1891 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ Opt_verbose, Opt_noverbose,
+ Opt_sum, Opt_nosum, Opt_wsum,
+ Opt_dirperm1, Opt_nodirperm1,
++ Opt_dirren, Opt_nodirren,
+ Opt_acl, Opt_noacl,
+ Opt_tail, Opt_ignore, Opt_ignore_silent, Opt_err
+};
+ {Opt_dio, "dio"},
+ {Opt_nodio, "nodio"},
+
++#ifdef CONFIG_AUFS_DIRREN
++ {Opt_dirren, "dirren"},
++ {Opt_nodirren, "nodirren"},
++#else
++ {Opt_ignore, "dirren"},
++ {Opt_ignore_silent, "nodirren"},
++#endif
++
+#ifdef CONFIG_AUFS_FHSM
+ {Opt_fhsm_sec, "fhsm_sec=%d"},
+#else
-+ {Opt_ignore_silent, "fhsm_sec=%d"},
++ {Opt_ignore, "fhsm_sec=%d"},
+#endif
+
+ {Opt_diropq_a, "diropq=always"},
+
+ /* keep them temporary */
+ {Opt_ignore_silent, "nodlgt"},
-+ {Opt_ignore_silent, "clean_plink"},
++ {Opt_ignore, "clean_plink"},
+
+#ifdef CONFIG_AUFS_SHWH
+ {Opt_shwh, "shwh"},
+ {Opt_acl, "acl"},
+ {Opt_noacl, "noacl"},
+#else
-+ {Opt_ignore_silent, "acl"},
++ {Opt_ignore, "acl"},
+ {Opt_ignore_silent, "noacl"},
+#endif
+
+ case Opt_fhsm_sec:
+ AuDbg("fhsm_sec %u\n", opt->fhsm_second);
+ break;
++ case Opt_dirren:
++ AuLabel(dirren);
++ break;
++ case Opt_nodirren:
++ AuLabel(nodirren);
++ break;
+ case Opt_acl:
+ AuLabel(acl);
+ break;
+ case Opt_wsum:
+ case Opt_rdblk_def:
+ case Opt_rdhash_def:
++ case Opt_dirren:
++ case Opt_nodirren:
+ case Opt_acl:
+ case Opt_noacl:
+ err = 0;
+ au_fclr_opts(opts->flags, TRUNC_XIB);
+ break;
+
++ case Opt_dirren:
++ err = 1;
++ if (!au_opt_test(sbinfo->si_mntflags, DIRREN)) {
++ err = au_dr_opt_set(sb);
++ if (!err)
++ err = 1;
++ }
++ if (err == 1)
++ au_opt_set(sbinfo->si_mntflags, DIRREN);
++ break;
++ case Opt_nodirren:
++ err = 1;
++ if (au_opt_test(sbinfo->si_mntflags, DIRREN)) {
++ err = au_dr_opt_clr(sb, au_ftest_opts(opts->flags,
++ DR_FLUSHED));
++ if (!err)
++ err = 1;
++ }
++ if (err == 1)
++ au_opt_clr(sbinfo->si_mntflags, DIRREN);
++ break;
++
+ case Opt_acl:
+ sb->s_flags |= MS_POSIXACL;
+ break;
+
+ SiMustWriteLock(sb);
+
-+ err = 0;
++ err = au_dr_opt_flush(sb);
++ if (unlikely(err))
++ goto out;
++ au_fset_opts(opts->flags, DR_FLUSHED);
++
+ dir = d_inode(sb->s_root);
+ sbinfo = au_sbi(sb);
+ opt_xino = NULL;
+ au_fset_opts(opts->flags, REFRESH);
+
+ AuDbg("status 0x%x\n", opts->flags);
++
++out:
+ return err;
+}
+
+}
diff -urN /usr/share/empty/fs/aufs/opts.h linux/fs/aufs/opts.h
--- /usr/share/empty/fs/aufs/opts.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/opts.h 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,213 @@
++++ linux/fs/aufs/opts.h 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,224 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+#define AuOpt_WARN_PERM (1 << 12) /* warn when add-branch */
+#define AuOpt_VERBOSE (1 << 13) /* busy inode when del-branch */
+#define AuOpt_DIO (1 << 14) /* direct io */
++#define AuOpt_DIRREN (1 << 15) /* directory rename */
+
+#ifndef CONFIG_AUFS_HNOTIFY
+#undef AuOpt_UDBA_HNOTIFY
+#define AuOpt_UDBA_HNOTIFY 0
+#endif
++#ifndef CONFIG_AUFS_DIRREN
++#undef AuOpt_DIRREN
++#define AuOpt_DIRREN 0
++#endif
+#ifndef CONFIG_AUFS_SHWH
+#undef AuOpt_SHWH
+#define AuOpt_SHWH 0
+#define AuOpts_TRUNC_XIB (1 << 2)
+#define AuOpts_REFRESH_DYAOP (1 << 3)
+#define AuOpts_REFRESH_IDOP (1 << 4)
++#define AuOpts_DR_FLUSHED (1 << 5)
+#define au_ftest_opts(flags, name) ((flags) & AuOpts_##name)
+#define au_fset_opts(flags, name) \
+ do { (flags) |= AuOpts_##name; } while (0)
+#define au_fclr_opts(flags, name) \
+ do { (flags) &= ~AuOpts_##name; } while (0)
+
++#ifndef CONFIG_AUFS_DIRREN
++#undef AuOpts_DR_FLUSHED
++#define AuOpts_DR_FLUSHED 0
++#endif
++
+struct au_opts {
+ struct au_opt *opt;
+ int max_opt;
+#endif /* __AUFS_OPTS_H__ */
diff -urN /usr/share/empty/fs/aufs/plink.c linux/fs/aufs/plink.c
--- /usr/share/empty/fs/aufs/plink.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/plink.c 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,514 @@
++++ linux/fs/aufs/plink.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,515 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+{
+ int i;
+ struct au_sbinfo *sbinfo;
-+ struct hlist_head *plink_hlist;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos;
+ struct au_icntnr *icntnr;
+
+ SiMustAnyLock(sb);
+ AuDebugOn(au_plink_maint(sb, AuLock_NOPLM));
+
+ for (i = 0; i < AuPlink_NHASH; i++) {
-+ plink_hlist = &sbinfo->si_plink[i].head;
-+ rcu_read_lock();
-+ hlist_for_each_entry_rcu(icntnr, plink_hlist, plink)
++ hbl = sbinfo->si_plink + i;
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(icntnr, pos, hbl, plink)
+ AuDbg("%lu\n", icntnr->vfs_inode.i_ino);
-+ rcu_read_unlock();
++ hlist_bl_unlock(hbl);
+ }
+}
+#endif
+{
+ int found, i;
+ struct au_sbinfo *sbinfo;
-+ struct hlist_head *plink_hlist;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos;
+ struct au_icntnr *icntnr;
+
+ sbinfo = au_sbi(inode->i_sb);
+
+ found = 0;
+ i = au_plink_hash(inode->i_ino);
-+ plink_hlist = &sbinfo->si_plink[i].head;
-+ rcu_read_lock();
-+ hlist_for_each_entry_rcu(icntnr, plink_hlist, plink)
++ hbl = sbinfo->si_plink + i;
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(icntnr, pos, hbl, plink)
+ if (&icntnr->vfs_inode == inode) {
+ found = 1;
+ break;
+ }
-+ rcu_read_unlock();
++ hlist_bl_unlock(hbl);
+ return found;
+}
+
+{
+ struct super_block *sb;
+ struct au_sbinfo *sbinfo;
-+ struct hlist_head *plink_hlist;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos;
+ struct au_icntnr *icntnr;
-+ struct au_sphlhead *sphl;
+ int found, err, cnt, i;
+
+ sb = inode->i_sb;
+ return;
+
+ i = au_plink_hash(inode->i_ino);
-+ sphl = sbinfo->si_plink + i;
-+ plink_hlist = &sphl->head;
++ hbl = sbinfo->si_plink + i;
+ au_igrab(inode);
+
-+ spin_lock(&sphl->spin);
-+ hlist_for_each_entry(icntnr, plink_hlist, plink) {
++ hlist_bl_lock(hbl);
++ hlist_bl_for_each_entry(icntnr, pos, hbl, plink) {
+ if (&icntnr->vfs_inode == inode) {
+ found = 1;
+ break;
+ }
+ if (!found) {
+ icntnr = container_of(inode, struct au_icntnr, vfs_inode);
-+ hlist_add_head_rcu(&icntnr->plink, plink_hlist);
++ hlist_bl_add_head(&icntnr->plink, hbl);
+ }
-+ spin_unlock(&sphl->spin);
++ hlist_bl_unlock(hbl);
+ if (!found) {
-+ cnt = au_sphl_count(sphl);
++ cnt = au_hbl_count(hbl);
+#define msg "unexpectedly unblanced or too many pseudo-links"
+ if (cnt > AUFS_PLINK_WARN)
+ AuWarn1(msg ", %d\n", cnt);
+ err = whplink(h_dentry, inode, bindex, au_sbr(sb, bindex));
+ if (unlikely(err)) {
+ pr_warn("err %d, damaged pseudo link.\n", err);
-+ au_sphl_del_rcu(&icntnr->plink, sphl);
++ au_hbl_del(&icntnr->plink, hbl);
+ iput(&icntnr->vfs_inode);
+ }
+ } else
+{
+ int i, warned;
+ struct au_sbinfo *sbinfo;
-+ struct hlist_head *plink_hlist;
-+ struct hlist_node *tmp;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos, *tmp;
+ struct au_icntnr *icntnr;
+
+ SiMustWriteLock(sb);
+ /* no spin_lock since sbinfo is write-locked */
+ warned = 0;
+ for (i = 0; i < AuPlink_NHASH; i++) {
-+ plink_hlist = &sbinfo->si_plink[i].head;
-+ if (!warned && verbose && !hlist_empty(plink_hlist)) {
++ hbl = sbinfo->si_plink + i;
++ if (!warned && verbose && !hlist_bl_empty(hbl)) {
+ pr_warn("pseudo-link is not flushed");
+ warned = 1;
+ }
-+ hlist_for_each_entry_safe(icntnr, tmp, plink_hlist, plink)
++ hlist_bl_for_each_entry_safe(icntnr, pos, tmp, hbl, plink)
+ iput(&icntnr->vfs_inode);
-+ INIT_HLIST_HEAD(plink_hlist);
++ INIT_HLIST_BL_HEAD(hbl);
+ }
+}
+
+void au_plink_half_refresh(struct super_block *sb, aufs_bindex_t br_id)
+{
+ struct au_sbinfo *sbinfo;
-+ struct hlist_head *plink_hlist;
-+ struct hlist_node *tmp;
++ struct hlist_bl_head *hbl;
++ struct hlist_bl_node *pos, *tmp;
+ struct au_icntnr *icntnr;
+ struct inode *inode;
+ int i, do_put;
+ AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK));
+ AuDebugOn(au_plink_maint(sb, AuLock_NOPLM));
+
-+ /* no spin_lock since sbinfo is write-locked */
++ /* no bit_lock since sbinfo is write-locked */
+ for (i = 0; i < AuPlink_NHASH; i++) {
-+ plink_hlist = &sbinfo->si_plink[i].head;
-+ hlist_for_each_entry_safe(icntnr, tmp, plink_hlist, plink) {
++ hbl = sbinfo->si_plink + i;
++ hlist_bl_for_each_entry_safe(icntnr, pos, tmp, hbl, plink) {
+ inode = au_igrab(&icntnr->vfs_inode);
+ ii_write_lock_child(inode);
+ do_put = au_plink_do_half_refresh(inode, br_id);
+ if (do_put) {
-+ hlist_del(&icntnr->plink);
++ hlist_bl_del(&icntnr->plink);
+ iput(inode);
+ }
+ ii_write_unlock(inode);
+}
diff -urN /usr/share/empty/fs/aufs/poll.c linux/fs/aufs/poll.c
--- /usr/share/empty/fs/aufs/poll.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/poll.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/poll.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,52 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/posix_acl.c linux/fs/aufs/posix_acl.c
--- /usr/share/empty/fs/aufs/posix_acl.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/posix_acl.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/posix_acl.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,102 @@
+/*
+ * Copyright (C) 2014-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/procfs.c linux/fs/aufs/procfs.c
--- /usr/share/empty/fs/aufs/procfs.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/procfs.c 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,169 @@
++++ linux/fs/aufs/procfs.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,170 @@
+/*
+ * Copyright (C) 2010-2017 Junjiro R. Okajima
+ *
+ int err;
+ struct super_block *sb;
+ struct au_sbinfo *sbinfo;
++ struct hlist_bl_node *pos;
+
+ err = -EBUSY;
+ if (unlikely(file->private_data))
+
+ sb = NULL;
+ /* don't use au_sbilist_lock() here */
-+ spin_lock(&au_sbilist.spin);
-+ hlist_for_each_entry(sbinfo, &au_sbilist.head, si_list)
++ hlist_bl_lock(&au_sbilist);
++ hlist_bl_for_each_entry(sbinfo, pos, &au_sbilist, si_list)
+ if (id == sysaufs_si_id(sbinfo)) {
+ kobject_get(&sbinfo->si_kobj);
+ sb = sbinfo->si_sb;
+ break;
+ }
-+ spin_unlock(&au_sbilist.spin);
++ hlist_bl_unlock(&au_sbilist);
+
+ err = -EINVAL;
+ if (unlikely(!sb))
+}
diff -urN /usr/share/empty/fs/aufs/rdu.c linux/fs/aufs/rdu.c
--- /usr/share/empty/fs/aufs/rdu.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/rdu.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/rdu.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,381 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+#endif
diff -urN /usr/share/empty/fs/aufs/rwsem.h linux/fs/aufs/rwsem.h
--- /usr/share/empty/fs/aufs/rwsem.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/rwsem.h 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,198 @@
++++ linux/fs/aufs/rwsem.h 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,72 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+
+#include "debug.h"
+
-+struct au_rwsem {
-+ struct rw_semaphore rwsem;
-+#ifdef CONFIG_AUFS_DEBUG
-+ /* just for debugging, not almighty counter */
-+ atomic_t rcnt, wcnt;
-+#endif
-+};
-+
-+#ifdef CONFIG_LOCKDEP
-+#define au_lockdep_set_name(rw) \
-+ lockdep_set_class_and_name(&(rw)->rwsem, \
-+ /*original key*/(rw)->rwsem.dep_map.key, \
-+ /*name*/#rw)
-+#else
-+#define au_lockdep_set_name(rw) do {} while (0)
-+#endif
-+
-+#ifdef CONFIG_AUFS_DEBUG
-+#define AuDbgCntInit(rw) do { \
-+ atomic_set(&(rw)->rcnt, 0); \
-+ atomic_set(&(rw)->wcnt, 0); \
-+ smp_mb(); /* atomic set */ \
-+} while (0)
-+
-+#define AuDbgCnt(rw, cnt) atomic_read(&(rw)->cnt)
-+#define AuDbgCntInc(rw, cnt) atomic_inc(&(rw)->cnt)
-+#define AuDbgCntDec(rw, cnt) WARN_ON(atomic_dec_return(&(rw)->cnt) < 0)
-+#define AuDbgRcntInc(rw) AuDbgCntInc(rw, rcnt)
-+#define AuDbgRcntDec(rw) AuDbgCntDec(rw, rcnt)
-+#define AuDbgWcntInc(rw) AuDbgCntInc(rw, wcnt)
-+#define AuDbgWcntDec(rw) AuDbgCntDec(rw, wcnt)
-+#else
-+#define AuDbgCnt(rw, cnt) 0
-+#define AuDbgCntInit(rw) do {} while (0)
-+#define AuDbgRcntInc(rw) do {} while (0)
-+#define AuDbgRcntDec(rw) do {} while (0)
-+#define AuDbgWcntInc(rw) do {} while (0)
-+#define AuDbgWcntDec(rw) do {} while (0)
-+#endif /* CONFIG_AUFS_DEBUG */
++/* in the futre, the name 'au_rwsem' will be totally gone */
++#define au_rwsem rw_semaphore
+
+/* to debug easier, do not make them inlined functions */
-+#define AuRwMustNoWaiters(rw) AuDebugOn(rwsem_is_contended(&(rw)->rwsem))
++#define AuRwMustNoWaiters(rw) AuDebugOn(rwsem_is_contended(rw))
+/* rwsem_is_locked() is unusable */
-+#define AuRwMustReadLock(rw) AuDebugOn(AuDbgCnt(rw, rcnt) <= 0)
-+#define AuRwMustWriteLock(rw) AuDebugOn(AuDbgCnt(rw, wcnt) <= 0)
-+#define AuRwMustAnyLock(rw) AuDebugOn(AuDbgCnt(rw, rcnt) <= 0 \
-+ && AuDbgCnt(rw, wcnt) <= 0)
-+#define AuRwDestroy(rw) AuDebugOn(AuDbgCnt(rw, rcnt) \
-+ || AuDbgCnt(rw, wcnt))
-+
-+#define au_rw_init(rw) do { \
-+ AuDbgCntInit(rw); \
-+ init_rwsem(&(rw)->rwsem); \
-+ au_lockdep_set_name(rw); \
-+ } while (0)
++#define AuRwMustReadLock(rw) AuDebugOn(!lockdep_recursing(current) \
++ && debug_locks \
++ && !lockdep_is_held_type(rw, 1))
++#define AuRwMustWriteLock(rw) AuDebugOn(!lockdep_recursing(current) \
++ && debug_locks \
++ && !lockdep_is_held_type(rw, 0))
++#define AuRwMustAnyLock(rw) AuDebugOn(!lockdep_recursing(current) \
++ && debug_locks \
++ && !lockdep_is_held(rw))
++#define AuRwDestroy(rw) AuDebugOn(!lockdep_recursing(current) \
++ && debug_locks \
++ && lockdep_is_held(rw))
++
++#define au_rw_init(rw) init_rwsem(rw)
+
+#define au_rw_init_wlock(rw) do { \
+ au_rw_init(rw); \
-+ down_write(&(rw)->rwsem); \
-+ AuDbgWcntInc(rw); \
++ down_write(rw); \
+ } while (0)
+
-+#define au_rw_init_wlock_nested(rw, lsc) do { \
-+ au_rw_init(rw); \
-+ down_write_nested(&(rw)->rwsem, lsc); \
-+ AuDbgWcntInc(rw); \
++#define au_rw_init_wlock_nested(rw, lsc) do { \
++ au_rw_init(rw); \
++ down_write_nested(rw, lsc); \
+ } while (0)
+
-+static inline void au_rw_read_lock(struct au_rwsem *rw)
-+{
-+ down_read(&rw->rwsem);
-+ AuDbgRcntInc(rw);
-+}
-+
-+static inline void au_rw_read_lock_nested(struct au_rwsem *rw, unsigned int lsc)
-+{
-+ down_read_nested(&rw->rwsem, lsc);
-+ AuDbgRcntInc(rw);
-+}
-+
-+static inline void au_rw_read_unlock(struct au_rwsem *rw)
-+{
-+ AuRwMustReadLock(rw);
-+ AuDbgRcntDec(rw);
-+ up_read(&rw->rwsem);
-+}
-+
-+static inline void au_rw_dgrade_lock(struct au_rwsem *rw)
-+{
-+ AuRwMustWriteLock(rw);
-+ AuDbgRcntInc(rw);
-+ AuDbgWcntDec(rw);
-+ downgrade_write(&rw->rwsem);
-+}
-+
-+static inline void au_rw_write_lock(struct au_rwsem *rw)
-+{
-+ down_write(&rw->rwsem);
-+ AuDbgWcntInc(rw);
-+}
-+
-+static inline void au_rw_write_lock_nested(struct au_rwsem *rw,
-+ unsigned int lsc)
-+{
-+ down_write_nested(&rw->rwsem, lsc);
-+ AuDbgWcntInc(rw);
-+}
-+
-+static inline void au_rw_write_unlock(struct au_rwsem *rw)
-+{
-+ AuRwMustWriteLock(rw);
-+ AuDbgWcntDec(rw);
-+ up_write(&rw->rwsem);
-+}
-+
-+/* why is not _nested version defined */
-+static inline int au_rw_read_trylock(struct au_rwsem *rw)
-+{
-+ int ret;
-+
-+ ret = down_read_trylock(&rw->rwsem);
-+ if (ret)
-+ AuDbgRcntInc(rw);
-+ return ret;
-+}
-+
-+static inline int au_rw_write_trylock(struct au_rwsem *rw)
-+{
-+ int ret;
-+
-+ ret = down_write_trylock(&rw->rwsem);
-+ if (ret)
-+ AuDbgWcntInc(rw);
-+ return ret;
-+}
-+
-+#undef AuDbgCntDec
-+#undef AuDbgRcntInc
-+#undef AuDbgRcntDec
-+#undef AuDbgWcntDec
-+
-+#define AuSimpleLockRwsemFuncs(prefix, param, rwsem) \
-+static inline void prefix##_read_lock(param) \
-+{ au_rw_read_lock(rwsem); } \
-+static inline void prefix##_write_lock(param) \
-+{ au_rw_write_lock(rwsem); } \
-+static inline int prefix##_read_trylock(param) \
-+{ return au_rw_read_trylock(rwsem); } \
-+static inline int prefix##_write_trylock(param) \
-+{ return au_rw_write_trylock(rwsem); }
-+/* why is not _nested version defined */
-+/* static inline void prefix##_read_trylock_nested(param, lsc)
-+{ au_rw_read_trylock_nested(rwsem, lsc)); }
-+static inline void prefix##_write_trylock_nestd(param, lsc)
-+{ au_rw_write_trylock_nested(rwsem, lsc); } */
-+
-+#define AuSimpleUnlockRwsemFuncs(prefix, param, rwsem) \
-+static inline void prefix##_read_unlock(param) \
-+{ au_rw_read_unlock(rwsem); } \
-+static inline void prefix##_write_unlock(param) \
-+{ au_rw_write_unlock(rwsem); } \
-+static inline void prefix##_downgrade_lock(param) \
-+{ au_rw_dgrade_lock(rwsem); }
-+
-+#define AuSimpleRwsemFuncs(prefix, param, rwsem) \
-+ AuSimpleLockRwsemFuncs(prefix, param, rwsem) \
-+ AuSimpleUnlockRwsemFuncs(prefix, param, rwsem)
++#define au_rw_read_lock(rw) down_read(rw)
++#define au_rw_read_lock_nested(rw, lsc) down_read_nested(rw, lsc)
++#define au_rw_read_unlock(rw) up_read(rw)
++#define au_rw_dgrade_lock(rw) downgrade_write(rw)
++#define au_rw_write_lock(rw) down_write(rw)
++#define au_rw_write_lock_nested(rw, lsc) down_write_nested(rw, lsc)
++#define au_rw_write_unlock(rw) up_write(rw)
++/* why is not _nested version defined? */
++#define au_rw_read_trylock(rw) down_read_trylock(rw)
++#define au_rw_write_trylock(rw) down_write_trylock(rw)
+
+#endif /* __KERNEL__ */
+#endif /* __AUFS_RWSEM_H__ */
diff -urN /usr/share/empty/fs/aufs/sbinfo.c linux/fs/aufs/sbinfo.c
--- /usr/share/empty/fs/aufs/sbinfo.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/sbinfo.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/sbinfo.c 2017-11-12 22:24:44.707577830 +0100
@@ -0,0 +1,304 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+
+ sbinfo = container_of(kobj, struct au_sbinfo, si_kobj);
+ for (i = 0; i < AuPlink_NHASH; i++)
-+ AuDebugOn(!hlist_empty(&sbinfo->si_plink[i].head));
++ AuDebugOn(!hlist_bl_empty(sbinfo->si_plink + i));
+ AuDebugOn(atomic_read(&sbinfo->si_nowait.nw_len));
+
+ AuDebugOn(percpu_counter_sum(&sbinfo->si_ninodes));
+ sbinfo->si_xino_brid = -1;
+ /* leave si_xib_last_pindex and si_xib_next_bit */
+
-+ au_sphl_init(&sbinfo->si_aopen);
++ INIT_HLIST_BL_HEAD(&sbinfo->si_aopen);
+
+ sbinfo->si_rdcache = msecs_to_jiffies(AUFS_RDCACHE_DEF * MSEC_PER_SEC);
+ sbinfo->si_rdblk = AUFS_RDBLK_DEF;
+ sbinfo->si_dirwh = AUFS_DIRWH_DEF;
+
+ for (i = 0; i < AuPlink_NHASH; i++)
-+ au_sphl_init(sbinfo->si_plink + i);
++ INIT_HLIST_BL_HEAD(sbinfo->si_plink + i);
+ init_waitqueue_head(&sbinfo->si_plink_wq);
+ spin_lock_init(&sbinfo->si_plink_maint_lock);
+
-+ au_sphl_init(&sbinfo->si_files);
++ INIT_HLIST_BL_HEAD(&sbinfo->si_files);
+
+ /* with getattr by default */
+ sbinfo->si_iop_array = aufs_iop;
+ di_write_unlock2(d1, d2);
+ si_read_unlock(d1->d_sb);
+}
-diff -urN /usr/share/empty/fs/aufs/spl.h linux/fs/aufs/spl.h
---- /usr/share/empty/fs/aufs/spl.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/spl.h 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,113 @@
-+/*
-+ * Copyright (C) 2005-2017 Junjiro R. Okajima
-+ *
-+ * This program, aufs is free software; you can redistribute it and/or modify
-+ * it under the terms of the GNU General Public License as published by
-+ * the Free Software Foundation; either version 2 of the License, or
-+ * (at your option) any later version.
-+ *
-+ * This program is distributed in the hope that it will be useful,
-+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
-+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-+ * GNU General Public License for more details.
-+ *
-+ * You should have received a copy of the GNU General Public License
-+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
-+ */
-+
-+/*
-+ * simple list protected by a spinlock
-+ */
-+
-+#ifndef __AUFS_SPL_H__
-+#define __AUFS_SPL_H__
-+
-+#ifdef __KERNEL__
-+
-+#if 0
-+struct au_splhead {
-+ spinlock_t spin;
-+ struct list_head head;
-+};
-+
-+static inline void au_spl_init(struct au_splhead *spl)
-+{
-+ spin_lock_init(&spl->spin);
-+ INIT_LIST_HEAD(&spl->head);
-+}
-+
-+static inline void au_spl_add(struct list_head *list, struct au_splhead *spl)
-+{
-+ spin_lock(&spl->spin);
-+ list_add(list, &spl->head);
-+ spin_unlock(&spl->spin);
-+}
-+
-+static inline void au_spl_del(struct list_head *list, struct au_splhead *spl)
-+{
-+ spin_lock(&spl->spin);
-+ list_del(list);
-+ spin_unlock(&spl->spin);
-+}
-+
-+static inline void au_spl_del_rcu(struct list_head *list,
-+ struct au_splhead *spl)
-+{
-+ spin_lock(&spl->spin);
-+ list_del_rcu(list);
-+ spin_unlock(&spl->spin);
-+}
-+#endif
-+
-+/* ---------------------------------------------------------------------- */
-+
-+struct au_sphlhead {
-+ spinlock_t spin;
-+ struct hlist_head head;
-+};
-+
-+static inline void au_sphl_init(struct au_sphlhead *sphl)
-+{
-+ spin_lock_init(&sphl->spin);
-+ INIT_HLIST_HEAD(&sphl->head);
-+}
-+
-+static inline void au_sphl_add(struct hlist_node *hlist,
-+ struct au_sphlhead *sphl)
-+{
-+ spin_lock(&sphl->spin);
-+ hlist_add_head(hlist, &sphl->head);
-+ spin_unlock(&sphl->spin);
-+}
-+
-+static inline void au_sphl_del(struct hlist_node *hlist,
-+ struct au_sphlhead *sphl)
-+{
-+ spin_lock(&sphl->spin);
-+ hlist_del(hlist);
-+ spin_unlock(&sphl->spin);
-+}
-+
-+static inline void au_sphl_del_rcu(struct hlist_node *hlist,
-+ struct au_sphlhead *sphl)
-+{
-+ spin_lock(&sphl->spin);
-+ hlist_del_rcu(hlist);
-+ spin_unlock(&sphl->spin);
-+}
-+
-+static inline unsigned long au_sphl_count(struct au_sphlhead *sphl)
-+{
-+ unsigned long cnt;
-+ struct hlist_node *pos;
-+
-+ cnt = 0;
-+ spin_lock(&sphl->spin);
-+ hlist_for_each(pos, &sphl->head)
-+ cnt++;
-+ spin_unlock(&sphl->spin);
-+ return cnt;
-+}
-+
-+#endif /* __KERNEL__ */
-+#endif /* __AUFS_SPL_H__ */
diff -urN /usr/share/empty/fs/aufs/super.c linux/fs/aufs/super.c
--- /usr/share/empty/fs/aufs/super.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/super.c 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,1044 @@
++++ linux/fs/aufs/super.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,1046 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+
+ au_fhsm_show(m, sbinfo);
+
++ AuBool(DIRREN, dirren);
+ AuBool(SUM, sum);
+ /* AuBool(SUM_W, wsum); */
+ AuBool(WARN_PERM, warn_perm);
+ if (au_opt_test(sbinfo->si_mntflags, PLINK))
+ au_plink_put(sb, /*verbose*/1);
+ au_xino_clr(sb);
++ au_dr_opt_flush(sb);
+ sbinfo->si_sb = NULL;
+ aufs_write_unlock(sb->s_root);
+ au_nwt_flush(&sbinfo->si_nowait);
+};
diff -urN /usr/share/empty/fs/aufs/super.h linux/fs/aufs/super.h
--- /usr/share/empty/fs/aufs/super.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/super.h 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,617 @@
++++ linux/fs/aufs/super.h 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,626 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+
+#include <linux/fs.h>
+#include <linux/kobject.h>
++#include "hbl.h"
+#include "rwsem.h"
-+#include "spl.h"
+#include "wkq.h"
+
+/* policies to select one among multiple writable branches */
+#endif
+
+ /* dirty trick to suppoer atomic_open */
-+ struct au_sphlhead si_aopen;
++ struct hlist_bl_head si_aopen;
+
+ /* vdir parameters */
+ unsigned long si_rdcache; /* max cache time in jiffies */
+ unsigned int si_dirwh;
+
+ /* pseudo_link list */
-+ struct au_sphlhead si_plink[AuPlink_NHASH];
++ struct hlist_bl_head si_plink[AuPlink_NHASH];
+ wait_queue_head_t si_plink_wq;
+ spinlock_t si_plink_maint_lock;
+ pid_t si_plink_maint_pid;
+
+ /* file list */
-+ struct au_sphlhead si_files;
++ struct hlist_bl_head si_files;
+
+ /* with/without getattr, brother of sb->s_d_op */
+ struct inode_operations *si_iop_array;
+#endif
+
+#ifdef CONFIG_AUFS_SBILIST
-+ struct hlist_node si_list;
++ struct hlist_bl_node si_list;
+#endif
+
+ /* dirty, necessary for unmounting, sysfs and sysrq */
+
+#ifdef CONFIG_AUFS_SBILIST
+/* module.c */
-+extern struct au_sphlhead au_sbilist;
++extern struct hlist_bl_head au_sbilist;
+
+static inline void au_sbilist_init(void)
+{
-+ au_sphl_init(&au_sbilist);
++ INIT_HLIST_BL_HEAD(&au_sbilist);
+}
+
+static inline void au_sbilist_add(struct super_block *sb)
+{
-+ au_sphl_add(&au_sbi(sb)->si_list, &au_sbilist);
++ au_hbl_add(&au_sbi(sb)->si_list, &au_sbilist);
+}
+
+static inline void au_sbilist_del(struct super_block *sb)
+{
-+ au_sphl_del(&au_sbi(sb)->si_list, &au_sbilist);
++ au_hbl_del(&au_sbi(sb)->si_list, &au_sbilist);
+}
+
+#ifdef CONFIG_AUFS_MAGIC_SYSRQ
+static inline void au_sbilist_lock(void)
+{
-+ spin_lock(&au_sbilist.spin);
++ hlist_bl_lock(&au_sbilist);
+}
+
+static inline void au_sbilist_unlock(void)
+{
-+ spin_unlock(&au_sbilist.spin);
++ hlist_bl_unlock(&au_sbilist);
+}
+#define AuGFP_SBILIST GFP_ATOMIC
+#else
+/* ---------------------------------------------------------------------- */
+
+/* lock superblock. mainly for entry point functions */
-+/*
-+ * __si_read_lock, __si_write_lock,
-+ * __si_read_unlock, __si_write_unlock, __si_downgrade_lock
-+ */
-+AuSimpleRwsemFuncs(__si, struct super_block *sb, &au_sbi(sb)->si_rwsem);
++#define __si_read_lock(sb) au_rw_read_lock(&au_sbi(sb)->si_rwsem)
++#define __si_write_lock(sb) au_rw_write_lock(&au_sbi(sb)->si_rwsem)
++#define __si_read_trylock(sb) au_rw_read_trylock(&au_sbi(sb)->si_rwsem)
++#define __si_write_trylock(sb) au_rw_write_trylock(&au_sbi(sb)->si_rwsem)
++/*
++#define __si_read_trylock_nested(sb) \
++ au_rw_read_trylock_nested(&au_sbi(sb)->si_rwsem)
++#define __si_write_trylock_nested(sb) \
++ au_rw_write_trylock_nested(&au_sbi(sb)->si_rwsem)
++*/
++
++#define __si_read_unlock(sb) au_rw_read_unlock(&au_sbi(sb)->si_rwsem)
++#define __si_write_unlock(sb) au_rw_write_unlock(&au_sbi(sb)->si_rwsem)
++#define __si_downgrade_lock(sb) au_rw_dgrade_lock(&au_sbi(sb)->si_rwsem)
+
+#define SiMustNoWaiters(sb) AuRwMustNoWaiters(&au_sbi(sb)->si_rwsem)
+#define SiMustAnyLock(sb) AuRwMustAnyLock(&au_sbi(sb)->si_rwsem)
+#endif /* __SYSAUFS_H__ */
diff -urN /usr/share/empty/fs/aufs/sysfs.c linux/fs/aufs/sysfs.c
--- /usr/share/empty/fs/aufs/sysfs.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/sysfs.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/sysfs.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,376 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/sysrq.c linux/fs/aufs/sysrq.c
--- /usr/share/empty/fs/aufs/sysrq.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/sysrq.c 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,157 @@
++++ linux/fs/aufs/sysrq.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,159 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ char *plevel;
+ struct au_sbinfo *sbinfo;
+ struct file *file;
-+ struct au_sphlhead *files;
++ struct hlist_bl_head *files;
++ struct hlist_bl_node *pos;
+ struct au_finfo *finfo;
+
+ plevel = au_plevel;
+#endif
+ pr("files\n");
+ files = &au_sbi(sb)->si_files;
-+ spin_lock(&files->spin);
-+ hlist_for_each_entry(finfo, &files->head, fi_hlist) {
++ hlist_bl_lock(files);
++ hlist_bl_for_each_entry(finfo, pos, files, fi_hlist) {
+ umode_t mode;
+
+ file = finfo->fi_file;
+ if (!special_file(mode))
+ au_dpri_file(file);
+ }
-+ spin_unlock(&files->spin);
++ hlist_bl_unlock(files);
+ pr("done\n");
+
+#undef pr
+static void au_sysrq(int key __maybe_unused)
+{
+ struct au_sbinfo *sbinfo;
++ struct hlist_bl_node *pos;
+
+ lockdep_off();
+ au_sbilist_lock();
-+ hlist_for_each_entry(sbinfo, &au_sbilist.head, si_list)
++ hlist_bl_for_each_entry(sbinfo, pos, &au_sbilist, si_list)
+ sysrq_sb(sbinfo->si_sb);
+ au_sbilist_unlock();
+ lockdep_on();
+}
diff -urN /usr/share/empty/fs/aufs/vdir.c linux/fs/aufs/vdir.c
--- /usr/share/empty/fs/aufs/vdir.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/vdir.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/vdir.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,892 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+}
diff -urN /usr/share/empty/fs/aufs/vfsub.c linux/fs/aufs/vfsub.c
--- /usr/share/empty/fs/aufs/vfsub.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/vfsub.c 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,900 @@
++++ linux/fs/aufs/vfsub.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,894 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ * sub-routines for VFS
+ */
+
++#include <linux/mnt_namespace.h>
+#include <linux/namei.h>
+#include <linux/nsproxy.h>
+#include <linux/security.h>
+#include <linux/splice.h>
-+#ifdef CONFIG_AUFS_BR_FUSE
-+#include "../fs/mount.h"
-+#endif
+#include "aufs.h"
+
+#ifdef CONFIG_AUFS_BR_FUSE
+int vfsub_test_mntns(struct vfsmount *mnt, struct super_block *h_sb)
+{
-+ struct nsproxy *ns;
-+
+ if (!au_test_fuse(h_sb) || !au_userns)
+ return 0;
+
-+ ns = current->nsproxy;
-+ /* no {get,put}_nsproxy(ns) */
-+ return real_mount(mnt)->mnt_ns == ns->mnt_ns ? 0 : -EACCES;
++ return is_current_mnt_ns(mnt) ? 0 : -EACCES;
+}
+#endif
+
+}
diff -urN /usr/share/empty/fs/aufs/vfsub.h linux/fs/aufs/vfsub.h
--- /usr/share/empty/fs/aufs/vfsub.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/vfsub.h 2017-09-05 10:42:11.058755349 +0200
++++ linux/fs/aufs/vfsub.h 2017-11-12 22:24:44.707577830 +0100
@@ -0,0 +1,360 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+
+static inline int vfsub_native_ro(struct inode *inode)
+{
-+ return (inode->i_sb->s_flags & MS_RDONLY)
++ return sb_rdonly(inode->i_sb)
+ || IS_RDONLY(inode)
+ /* || IS_APPEND(inode) */
+ || IS_IMMUTABLE(inode);
+#endif /* __AUFS_VFSUB_H__ */
diff -urN /usr/share/empty/fs/aufs/wbr_policy.c linux/fs/aufs/wbr_policy.c
--- /usr/share/empty/fs/aufs/wbr_policy.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/wbr_policy.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/wbr_policy.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,830 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+};
diff -urN /usr/share/empty/fs/aufs/whout.c linux/fs/aufs/whout.c
--- /usr/share/empty/fs/aufs/whout.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/whout.c 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/whout.c 2017-11-12 22:24:42.274176929 +0100
@@ -0,0 +1,1061 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+#endif /* __AUFS_WHOUT_H__ */
diff -urN /usr/share/empty/fs/aufs/wkq.c linux/fs/aufs/wkq.c
--- /usr/share/empty/fs/aufs/wkq.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/wkq.c 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,213 @@
++++ linux/fs/aufs/wkq.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,390 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ au_wkq_func_t func;
+ void *args;
+
++#ifdef CONFIG_LOCKDEP
++ int dont_check;
++ struct held_lock **hlock;
++#endif
++
+ struct completion *comp;
+};
+
+/* ---------------------------------------------------------------------- */
++/*
++ * Aufs passes some operations to the workqueue such as the internal copyup.
++ * This scheme looks rather unnatural for LOCKDEP debugging feature, since the
++ * job run by workqueue depends upon the locks acquired in the other task.
++ * Delegating a small operation to the workqueue, aufs passes its lockdep
++ * information too. And the job in the workqueue restores the info in order to
++ * pretend as if it acquired those locks. This is just to make LOCKDEP work
++ * correctly and expectedly.
++ */
++
++#ifndef CONFIG_LOCKDEP
++AuStubInt0(au_wkq_lockdep_alloc, struct au_wkinfo *wkinfo);
++AuStubVoid(au_wkq_lockdep_free, struct au_wkinfo *wkinfo);
++AuStubVoid(au_wkq_lockdep_pre, struct au_wkinfo *wkinfo);
++AuStubVoid(au_wkq_lockdep_post, struct au_wkinfo *wkinfo);
++AuStubVoid(au_wkq_lockdep_init, struct au_wkinfo *wkinfo);
++#else
++static void au_wkq_lockdep_init(struct au_wkinfo *wkinfo)
++{
++ wkinfo->hlock = NULL;
++ wkinfo->dont_check = 0;
++}
++
++/*
++ * 1: matched
++ * 0: unmatched
++ */
++static int au_wkq_lockdep_test(struct lock_class_key *key, const char *name)
++{
++ static DEFINE_SPINLOCK(spin);
++ static struct {
++ char *name;
++ struct lock_class_key *key;
++ } a[] = {
++ { .name = "&sbinfo->si_rwsem" },
++ { .name = "&finfo->fi_rwsem" },
++ { .name = "&dinfo->di_rwsem" },
++ { .name = "&iinfo->ii_rwsem" }
++ };
++ static int set;
++ int i;
++
++ /* lockless read from 'set.' see below */
++ if (set == ARRAY_SIZE(a)) {
++ for (i = 0; i < ARRAY_SIZE(a); i++)
++ if (a[i].key == key)
++ goto match;
++ goto unmatch;
++ }
++
++ spin_lock(&spin);
++ if (set)
++ for (i = 0; i < ARRAY_SIZE(a); i++)
++ if (a[i].key == key) {
++ spin_unlock(&spin);
++ goto match;
++ }
++ for (i = 0; i < ARRAY_SIZE(a); i++) {
++ if (a[i].key) {
++ if (unlikely(a[i].key == key)) { /* rare but possible */
++ spin_unlock(&spin);
++ goto match;
++ } else
++ continue;
++ }
++ if (strstr(a[i].name, name)) {
++ /*
++ * the order of these three lines is important for the
++ * lockless read above.
++ */
++ a[i].key = key;
++ spin_unlock(&spin);
++ set++;
++ /* AuDbg("%d, %s\n", set, name); */
++ goto match;
++ }
++ }
++ spin_unlock(&spin);
++ goto unmatch;
++
++match:
++ return 1;
++unmatch:
++ return 0;
++}
++
++static int au_wkq_lockdep_alloc(struct au_wkinfo *wkinfo)
++{
++ int err, n;
++ struct task_struct *curr;
++ struct held_lock **hl, *held_locks, *p;
++
++ err = 0;
++ curr = current;
++ wkinfo->dont_check = lockdep_recursing(curr);
++ if (wkinfo->dont_check)
++ goto out;
++ n = curr->lockdep_depth;
++ if (!n)
++ goto out;
++
++ err = -ENOMEM;
++ wkinfo->hlock = kmalloc_array(n + 1, sizeof(*wkinfo->hlock), GFP_NOFS);
++ if (unlikely(!wkinfo->hlock))
++ goto out;
++
++ err = 0;
++#if 0
++ if (0 && au_debug_test()) /* left for debugging */
++ lockdep_print_held_locks(curr);
++#endif
++ held_locks = curr->held_locks;
++ hl = wkinfo->hlock;
++ while (n--) {
++ p = held_locks++;
++ if (au_wkq_lockdep_test(p->instance->key, p->instance->name))
++ *hl++ = p;
++ }
++ *hl = NULL;
++
++out:
++ return err;
++}
++
++static void au_wkq_lockdep_free(struct au_wkinfo *wkinfo)
++{
++ kfree(wkinfo->hlock);
++}
++
++static void au_wkq_lockdep_pre(struct au_wkinfo *wkinfo)
++{
++ struct held_lock *p, **hl = wkinfo->hlock;
++ int subclass;
++
++ if (wkinfo->dont_check)
++ lockdep_off();
++ if (!hl)
++ return;
++ while ((p = *hl++)) { /* assignment */
++ subclass = lockdep_hlock_class(p)->subclass;
++ /* AuDbg("%s, %d\n", p->instance->name, subclass); */
++ if (p->read)
++ rwsem_acquire_read(p->instance, subclass, 0,
++ /*p->acquire_ip*/_RET_IP_);
++ else
++ rwsem_acquire(p->instance, subclass, 0,
++ /*p->acquire_ip*/_RET_IP_);
++ }
++}
++
++static void au_wkq_lockdep_post(struct au_wkinfo *wkinfo)
++{
++ struct held_lock *p, **hl = wkinfo->hlock;
++
++ if (wkinfo->dont_check)
++ lockdep_on();
++ if (!hl)
++ return;
++ while ((p = *hl++)) /* assignment */
++ rwsem_release(p->instance, 0, /*p->acquire_ip*/_RET_IP_);
++}
++#endif
+
+static void wkq_func(struct work_struct *wk)
+{
+ AuDebugOn(!uid_eq(current_fsuid(), GLOBAL_ROOT_UID));
+ AuDebugOn(rlimit(RLIMIT_FSIZE) != RLIM_INFINITY);
+
++ au_wkq_lockdep_pre(wkinfo);
+ wkinfo->func(wkinfo->args);
++ au_wkq_lockdep_post(wkinfo);
+ if (au_ftest_wkq(wkinfo->flags, WAIT))
+ complete(wkinfo->comp);
+ else {
+ };
+
+ err = au_wkq_comp_alloc(&wkinfo, &comp);
++ if (unlikely(err))
++ goto out;
++ err = au_wkq_lockdep_alloc(&wkinfo);
++ if (unlikely(err))
++ goto out_comp;
+ if (!err) {
+ au_wkq_run(&wkinfo);
+ /* no timeout, no interrupt */
+ wait_for_completion(wkinfo.comp);
-+ au_wkq_comp_free(comp);
-+ destroy_work_on_stack(&wkinfo.wk);
+ }
++ au_wkq_lockdep_free(&wkinfo);
+
++out_comp:
++ au_wkq_comp_free(comp);
++out:
++ destroy_work_on_stack(&wkinfo.wk);
+ return err;
-+
+}
+
+/*
+ wkinfo->func = func;
+ wkinfo->args = args;
+ wkinfo->comp = NULL;
++ au_wkq_lockdep_init(wkinfo);
+ kobject_get(wkinfo->kobj);
+ __module_get(THIS_MODULE); /* todo: ?? */
+
+}
diff -urN /usr/share/empty/fs/aufs/wkq.h linux/fs/aufs/wkq.h
--- /usr/share/empty/fs/aufs/wkq.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/wkq.h 2017-07-29 12:14:25.906375514 +0200
++++ linux/fs/aufs/wkq.h 2017-11-12 22:24:44.707577830 +0100
@@ -0,0 +1,93 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+
+#ifdef __KERNEL__
+
-+#include <linux/percpu_counter.h>
++#include <linux/wait.h>
+
+struct super_block;
+
+#endif /* __AUFS_WKQ_H__ */
diff -urN /usr/share/empty/fs/aufs/xattr.c linux/fs/aufs/xattr.c
--- /usr/share/empty/fs/aufs/xattr.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/xattr.c 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,357 @@
++++ linux/fs/aufs/xattr.c 2017-11-12 22:24:44.707577830 +0100
+@@ -0,0 +1,355 @@
+/*
+ * Copyright (C) 2014-2017 Junjiro R. Okajima
+ *
+ AuTraceErr(err);
+ }
+
-+ if (value)
-+ kfree(value);
++ kfree(value);
+
+out_free:
-+ if (o)
-+ kfree(o);
++ kfree(o);
+out:
+ if (!unlocked)
+ inode_unlock_shared(h_isrc);
+}
diff -urN /usr/share/empty/fs/aufs/xino.c linux/fs/aufs/xino.c
--- /usr/share/empty/fs/aufs/xino.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux/fs/aufs/xino.c 2017-07-29 12:14:25.906375514 +0200
-@@ -0,0 +1,1415 @@
++++ linux/fs/aufs/xino.c 2017-11-12 22:24:44.710911257 +0100
+@@ -0,0 +1,1418 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+ lockdep_off();
+ err = do_xino_fwrite(func, file, buf, size, pos);
+ lockdep_on();
-+ } else
++ } else {
++ lockdep_off();
+ err = xino_fwrite_wkq(func, file, buf, size, pos);
++ lockdep_on();
++ }
+
+ return err;
+}
+}
diff -urN /usr/share/empty/include/uapi/linux/aufs_type.h linux/include/uapi/linux/aufs_type.h
--- /usr/share/empty/include/uapi/linux/aufs_type.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux/include/uapi/linux/aufs_type.h 2017-09-05 10:42:11.058755349 +0200
-@@ -0,0 +1,419 @@
++++ linux/include/uapi/linux/aufs_type.h 2017-11-12 22:24:44.710911257 +0100
+@@ -0,0 +1,447 @@
+/*
+ * Copyright (C) 2005-2017 Junjiro R. Okajima
+ *
+
+#include <linux/limits.h>
+
-+#define AUFS_VERSION "4.x-rcN-20170904"
++#define AUFS_VERSION "4.x-rcN-20171106"
+
+/* todo? move this to linux-2.6.19/include/magic.h */
+#define AUFS_SUPER_MAGIC ('a' << 24 | 'u' << 16 | 'f' << 8 | 's')
+#define AUFS_PLINK_MAINT_DIR "fs/" AUFS_NAME
+#define AUFS_PLINK_MAINT_PATH AUFS_PLINK_MAINT_DIR "/" AUFS_PLINK_MAINT_NAME
+
++/* dirren, renamed dir */
++#define AUFS_DR_INFO_PFX AUFS_WH_PFX ".dr."
++#define AUFS_DR_BRHINO_NAME AUFS_WH_PFX "hino"
++/* whiteouted doubly */
++#define AUFS_WH_DR_INFO_PFX AUFS_WH_PFX AUFS_DR_INFO_PFX
++#define AUFS_WH_DR_BRHINO AUFS_WH_PFX AUFS_DR_BRHINO_NAME
++
+#define AUFS_DIROPQ_NAME AUFS_WH_PFX ".opq" /* whiteouted doubly */
+#define AUFS_WH_DIROPQ AUFS_WH_PFX AUFS_DIROPQ_NAME
+
+
+/* ---------------------------------------------------------------------- */
+
++/* dirren. the branch is identified by the filename who contains this */
++struct au_drinfo {
++ uint64_t ino;
++ union {
++ uint8_t oldnamelen;
++ uint64_t _padding;
++ };
++ uint8_t oldname[0];
++} __aligned(8);
++
++struct au_drinfo_fdata {
++ uint32_t magic;
++ struct au_drinfo drinfo;
++} __aligned(8);
++
++#define AUFS_DRINFO_MAGIC_V1 ('a' << 24 | 'd' << 16 | 'r' << 8 | 0x01)
++/* future */
++#define AUFS_DRINFO_MAGIC_V2 ('a' << 24 | 'd' << 16 | 'r' << 8 | 0x02)
++
++/* ---------------------------------------------------------------------- */
++
+struct aufs_wbr_fd {
+ uint32_t oflags;
+ int16_t brid;
aufs4.x-rcN loopback patch
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
-index 10707c3..af32e47 100644
+index d44de9d..095672b 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
-@@ -547,7 +547,7 @@ static int do_req_filebacked(struct loop_device *lo, struct request *rq)
+@@ -595,6 +595,15 @@ static inline void loop_update_dio(struct loop_device *lo)
+ lo->use_dio);
}
- struct switch_request {
-- struct file *file;
-+ struct file *file, *virt_file;
- struct completion wait;
- };
-
-@@ -573,6 +573,7 @@ static void do_loop_switch(struct loop_device *lo, struct switch_request *p)
- mapping = file->f_mapping;
- mapping_set_gfp_mask(old_file->f_mapping, lo->old_gfp_mask);
- lo->lo_backing_file = file;
-+ lo->lo_backing_virt_file = p->virt_file;
- lo->lo_blocksize = S_ISBLK(mapping->host->i_mode) ?
- mapping->host->i_bdev->bd_block_size : PAGE_SIZE;
- lo->old_gfp_mask = mapping_gfp_mask(mapping);
-@@ -585,11 +586,13 @@ static void do_loop_switch(struct loop_device *lo, struct switch_request *p)
- * First it needs to flush existing IO, it does this by sending a magic
- * BIO down the pipe. The completion of this BIO does the actual switch.
- */
--static int loop_switch(struct loop_device *lo, struct file *file)
-+static int loop_switch(struct loop_device *lo, struct file *file,
-+ struct file *virt_file)
- {
- struct switch_request w;
-
- w.file = file;
-+ w.virt_file = virt_file;
-
- /* freeze queue and wait for completion of scheduled requests */
- blk_mq_freeze_queue(lo->lo_queue);
-@@ -611,7 +614,16 @@ static int loop_flush(struct loop_device *lo)
- /* loop not yet configured, no running thread, nothing to flush */
- if (lo->lo_state != Lo_bound)
- return 0;
-- return loop_switch(lo, NULL);
-+ return loop_switch(lo, NULL, NULL);
-+}
-+
+static struct file *loop_real_file(struct file *file)
+{
+ struct file *f = NULL;
+ if (file->f_path.dentry->d_sb->s_op->real_loop)
+ f = file->f_path.dentry->d_sb->s_op->real_loop(file);
+ return f;
- }
-
++}
++
static void loop_reread_partitions(struct loop_device *lo,
-@@ -648,6 +660,7 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
+ struct block_device *bdev)
+ {
+@@ -629,6 +638,7 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
unsigned int arg)
{
struct file *file, *old_file;
struct inode *inode;
int error;
-@@ -664,9 +677,16 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
+@@ -645,9 +655,16 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
file = fget(arg);
if (!file)
goto out;
error = -EINVAL;
-@@ -678,17 +698,21 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
- goto out_putf;
-
- /* and ... switch */
-- error = loop_switch(lo, file);
-+ error = loop_switch(lo, file, virt_file);
- if (error)
- goto out_putf;
+@@ -662,6 +679,7 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
+ blk_mq_freeze_queue(lo->lo_queue);
+ mapping_set_gfp_mask(old_file->f_mapping, lo->old_gfp_mask);
+ lo->lo_backing_file = file;
++ lo->lo_backing_virt_file = virt_file;
+ lo->old_gfp_mask = mapping_gfp_mask(file->f_mapping);
+ mapping_set_gfp_mask(file->f_mapping,
+ lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS));
+@@ -669,12 +687,16 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
+ blk_mq_unfreeze_queue(lo->lo_queue);
fput(old_file);
+ if (old_virt_file)
out:
return error;
}
-@@ -882,7 +906,7 @@ static int loop_prepare_queue(struct loop_device *lo)
+@@ -868,7 +890,7 @@ static int loop_prepare_queue(struct loop_device *lo)
static int loop_set_fd(struct loop_device *lo, fmode_t mode,
struct block_device *bdev, unsigned int arg)
{
+ struct file *file, *f, *virt_file = NULL;
struct inode *inode;
struct address_space *mapping;
- unsigned lo_blocksize;
-@@ -897,6 +921,12 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
+ int lo_flags = 0;
+@@ -882,6 +904,12 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
file = fget(arg);
if (!file)
goto out;
error = -EBUSY;
if (lo->lo_state != Lo_unbound)
-@@ -949,6 +979,7 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
+@@ -930,6 +958,7 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
lo->lo_device = bdev;
lo->lo_flags = lo_flags;
lo->lo_backing_file = file;
lo->transfer = NULL;
lo->ioctl = NULL;
lo->lo_sizelimit = 0;
-@@ -981,6 +1012,8 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
+@@ -963,6 +992,8 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
out_putf:
fput(file);
out:
/* This is safe: open() is still holding a reference. */
module_put(THIS_MODULE);
-@@ -1027,6 +1060,7 @@ loop_init_xfer(struct loop_device *lo, struct loop_func_table *xfer,
+@@ -1009,6 +1040,7 @@ loop_init_xfer(struct loop_device *lo, struct loop_func_table *xfer,
static int loop_clr_fd(struct loop_device *lo)
{
struct file *filp = lo->lo_backing_file;
gfp_t gfp = lo->old_gfp_mask;
struct block_device *bdev = lo->lo_device;
-@@ -1058,6 +1092,7 @@ static int loop_clr_fd(struct loop_device *lo)
+@@ -1040,6 +1072,7 @@ static int loop_clr_fd(struct loop_device *lo)
spin_lock_irq(&lo->lo_lock);
lo->lo_state = Lo_rundown;
lo->lo_backing_file = NULL;
spin_unlock_irq(&lo->lo_lock);
loop_release_xfer(lo);
-@@ -1102,6 +1137,8 @@ static int loop_clr_fd(struct loop_device *lo)
+@@ -1087,6 +1120,8 @@ static int loop_clr_fd(struct loop_device *lo)
* bd_mutex which is usually taken before lo_ctl_mutex.
*/
fput(filp);
}
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
-index fecd3f9..6b3a7c9 100644
+index 1f39567..128b137 100644
--- a/drivers/block/loop.h
+++ b/drivers/block/loop.h
@@ -46,7 +46,7 @@ struct loop_device {
- struct file * lo_backing_file;
+ struct file * lo_backing_file, *lo_backing_virt_file;
struct block_device *lo_device;
- unsigned lo_blocksize;
void *key_data;
+
diff --git a/fs/aufs/f_op.c b/fs/aufs/f_op.c
-index 870717e..ea7fee0 100644
+index 8aff060..e853272 100644
--- a/fs/aufs/f_op.c
+++ b/fs/aufs/f_op.c
@@ -357,7 +357,7 @@ static ssize_t aufs_read_iter(struct kiocb *kio, struct iov_iter *iov_iter)
#endif /* __KERNEL__ */
diff --git a/fs/aufs/super.c b/fs/aufs/super.c
-index 5455fb1..8b9df60 100644
+index 3c300125..128d790 100644
--- a/fs/aufs/super.c
+++ b/fs/aufs/super.c
-@@ -837,7 +837,10 @@ static const struct super_operations aufs_sop = {
+@@ -838,7 +838,10 @@ static const struct super_operations aufs_sop = {
.statfs = aufs_statfs,
.put_super = aufs_put_super,
.sync_fs = aufs_sync_fs,
/* ---------------------------------------------------------------------- */
diff --git a/include/linux/fs.h b/include/linux/fs.h
-index 9b21bb5..f7124fa 100644
+index 8ab6566..8dbaa52 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
-@@ -1814,6 +1814,10 @@ struct super_operations {
+@@ -1838,6 +1838,10 @@ struct super_operations {
struct shrink_control *);
long (*free_cached_objects)(struct super_block *,
struct shrink_control *);
projects / packages / kernel.git / commitdiff