From 73422811d290c628b4ddbf6830e5cd6fa42e84f1 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Sun, 10 May 2009 16:05:39 -0400 Subject: [PATCH 01/87] reiserfs: allow exposing privroot w/ xattrs enabled This patch adds an -oexpose_privroot option to allow access to the privroot. Signed-off-by: Jeff Mahoney Signed-off-by: Al Viro --- fs/reiserfs/dir.c | 10 ++++------ fs/reiserfs/super.c | 1 + fs/reiserfs/xattr.c | 3 ++- include/linux/reiserfs_fs_sb.h | 2 ++ 4 files changed, 9 insertions(+), 7 deletions(-) diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c index 45ee3d357c70..6d2668fdc384 100644 --- a/fs/reiserfs/dir.c +++ b/fs/reiserfs/dir.c @@ -44,13 +44,11 @@ static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry, static inline bool is_privroot_deh(struct dentry *dir, struct reiserfs_de_head *deh) { - int ret = 0; -#ifdef CONFIG_REISERFS_FS_XATTR struct dentry *privroot = REISERFS_SB(dir->d_sb)->priv_root; - ret = (dir == dir->d_parent && privroot->d_inode && - deh->deh_objectid == INODE_PKEY(privroot->d_inode)->k_objectid); -#endif - return ret; + if (reiserfs_expose_privroot(dir->d_sb)) + return 0; + return (dir == dir->d_parent && privroot->d_inode && + deh->deh_objectid == INODE_PKEY(privroot->d_inode)->k_objectid); } int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent, diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 3567fb9e3fb1..9dbdcfb5d314 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -898,6 +898,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options, /* strin {"conv",.setmask = 1 << REISERFS_CONVERT}, {"attrs",.setmask = 1 << REISERFS_ATTRS}, {"noattrs",.clrmask = 1 << REISERFS_ATTRS}, + {"expose_privroot", .setmask = 1 << REISERFS_EXPOSE_PRIVROOT}, #ifdef CONFIG_REISERFS_FS_XATTR {"user_xattr",.setmask = 1 << REISERFS_XATTRS_USER}, {"nouser_xattr",.clrmask = 1 << REISERFS_XATTRS_USER}, diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 8e7deb0e6964..f3d47d856848 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -981,7 +981,8 @@ int reiserfs_lookup_privroot(struct super_block *s) strlen(PRIVROOT_NAME)); if (!IS_ERR(dentry)) { REISERFS_SB(s)->priv_root = dentry; - s->s_root->d_op = &xattr_lookup_poison_ops; + if (!reiserfs_expose_privroot(s)) + s->s_root->d_op = &xattr_lookup_poison_ops; if (dentry->d_inode) dentry->d_inode->i_flags |= S_PRIVATE; } else diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index 6473650c28f1..dab68bbed675 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -453,6 +453,7 @@ enum reiserfs_mount_options { REISERFS_ATTRS, REISERFS_XATTRS_USER, REISERFS_POSIXACL, + REISERFS_EXPOSE_PRIVROOT, REISERFS_BARRIER_NONE, REISERFS_BARRIER_FLUSH, @@ -490,6 +491,7 @@ enum reiserfs_mount_options { #define reiserfs_data_writeback(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_DATA_WRITEBACK)) #define reiserfs_xattrs_user(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_XATTRS_USER)) #define reiserfs_posixacl(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_POSIXACL)) +#define reiserfs_expose_privroot(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_EXPOSE_PRIVROOT)) #define reiserfs_xattrs_optional(s) (reiserfs_xattrs_user(s) || reiserfs_posixacl(s)) #define reiserfs_barrier_none(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_BARRIER_NONE)) #define reiserfs_barrier_flush(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_BARRIER_FLUSH)) From 4e44b6852e03c915618ca6776b6697b436246b00 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 11:08:56 -0400 Subject: [PATCH 02/87] Get rid of path_lookup in autofs4 Signed-off-by: Al Viro --- fs/autofs4/dev-ioctl.c | 195 +++++++++++++---------------------------- 1 file changed, 60 insertions(+), 135 deletions(-) diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c index 84168c0dcc2d..f71dac9986f0 100644 --- a/fs/autofs4/dev-ioctl.c +++ b/fs/autofs4/dev-ioctl.c @@ -192,77 +192,42 @@ static int autofs_dev_ioctl_protosubver(struct file *fp, return 0; } -/* - * Walk down the mount stack looking for an autofs mount that - * has the requested device number (aka. new_encode_dev(sb->s_dev). - */ -static int autofs_dev_ioctl_find_super(struct nameidata *nd, dev_t devno) +static int find_autofs_mount(const char *pathname, + struct path *res, + int test(struct path *path, void *data), + void *data) { - struct dentry *dentry; - struct inode *inode; - struct super_block *sb; - dev_t s_dev; - unsigned int err; - + struct path path; + int err = kern_path(pathname, 0, &path); + if (err) + return err; err = -ENOENT; - - /* Lookup the dentry name at the base of our mount point */ - dentry = d_lookup(nd->path.dentry, &nd->last); - if (!dentry) - goto out; - - dput(nd->path.dentry); - nd->path.dentry = dentry; - - /* And follow the mount stack looking for our autofs mount */ - while (follow_down(&nd->path.mnt, &nd->path.dentry)) { - inode = nd->path.dentry->d_inode; - if (!inode) - break; - - sb = inode->i_sb; - s_dev = new_encode_dev(sb->s_dev); - if (devno == s_dev) { - if (sb->s_magic == AUTOFS_SUPER_MAGIC) { + while (path.dentry == path.mnt->mnt_root) { + if (path.mnt->mnt_sb->s_magic == AUTOFS_SUPER_MAGIC) { + if (test(&path, data)) { + path_get(&path); + if (!err) /* already found some */ + path_put(res); + *res = path; err = 0; - break; } } + if (!follow_up(&path.mnt, &path.dentry)) + break; } -out: + path_put(&path); return err; } -/* - * Walk down the mount stack looking for an autofs mount that - * has the requested mount type (ie. indirect, direct or offset). - */ -static int autofs_dev_ioctl_find_sbi_type(struct nameidata *nd, unsigned int type) +static int test_by_dev(struct path *path, void *p) { - struct dentry *dentry; - struct autofs_info *ino; - unsigned int err; + return path->mnt->mnt_sb->s_dev == *(dev_t *)p; +} - err = -ENOENT; - - /* Lookup the dentry name at the base of our mount point */ - dentry = d_lookup(nd->path.dentry, &nd->last); - if (!dentry) - goto out; - - dput(nd->path.dentry); - nd->path.dentry = dentry; - - /* And follow the mount stack looking for our autofs mount */ - while (follow_down(&nd->path.mnt, &nd->path.dentry)) { - ino = autofs4_dentry_ino(nd->path.dentry); - if (ino && ino->sbi->type & type) { - err = 0; - break; - } - } -out: - return err; +static int test_by_type(struct path *path, void *p) +{ + struct autofs_info *ino = autofs4_dentry_ino(path->dentry); + return ino && ino->sbi->type & *(unsigned *)p; } static void autofs_dev_ioctl_fd_install(unsigned int fd, struct file *file) @@ -283,31 +248,25 @@ static void autofs_dev_ioctl_fd_install(unsigned int fd, struct file *file) * Open a file descriptor on the autofs mount point corresponding * to the given path and device number (aka. new_encode_dev(sb->s_dev)). */ -static int autofs_dev_ioctl_open_mountpoint(const char *path, dev_t devid) +static int autofs_dev_ioctl_open_mountpoint(const char *name, dev_t devid) { - struct file *filp; - struct nameidata nd; int err, fd; fd = get_unused_fd(); if (likely(fd >= 0)) { - /* Get nameidata of the parent directory */ - err = path_lookup(path, LOOKUP_PARENT, &nd); + struct file *filp; + struct path path; + + err = find_autofs_mount(name, &path, test_by_dev, &devid); if (err) goto out; /* - * Search down, within the parent, looking for an - * autofs super block that has the device number + * Find autofs super block that has the device number * corresponding to the autofs fs we want to open. */ - err = autofs_dev_ioctl_find_super(&nd, devid); - if (err) { - path_put(&nd.path); - goto out; - } - filp = dentry_open(nd.path.dentry, nd.path.mnt, O_RDONLY, + filp = dentry_open(path.dentry, path.mnt, O_RDONLY, current_cred()); if (IS_ERR(filp)) { err = PTR_ERR(filp); @@ -340,7 +299,7 @@ static int autofs_dev_ioctl_openmount(struct file *fp, param->ioctlfd = -1; path = param->path; - devid = param->openmount.devid; + devid = new_decode_dev(param->openmount.devid); err = 0; fd = autofs_dev_ioctl_open_mountpoint(path, devid); @@ -475,8 +434,7 @@ static int autofs_dev_ioctl_requester(struct file *fp, struct autofs_dev_ioctl *param) { struct autofs_info *ino; - struct nameidata nd; - const char *path; + struct path path; dev_t devid; int err = -ENOENT; @@ -485,32 +443,24 @@ static int autofs_dev_ioctl_requester(struct file *fp, goto out; } - path = param->path; - devid = new_encode_dev(sbi->sb->s_dev); + devid = sbi->sb->s_dev; param->requester.uid = param->requester.gid = -1; - /* Get nameidata of the parent directory */ - err = path_lookup(path, LOOKUP_PARENT, &nd); + err = find_autofs_mount(param->path, &path, test_by_dev, &devid); if (err) goto out; - err = autofs_dev_ioctl_find_super(&nd, devid); - if (err) - goto out_release; - - ino = autofs4_dentry_ino(nd.path.dentry); + ino = autofs4_dentry_ino(path.dentry); if (ino) { err = 0; - autofs4_expire_wait(nd.path.dentry); + autofs4_expire_wait(path.dentry); spin_lock(&sbi->fs_lock); param->requester.uid = ino->uid; param->requester.gid = ino->gid; spin_unlock(&sbi->fs_lock); } - -out_release: - path_put(&nd.path); + path_put(&path); out: return err; } @@ -569,8 +519,8 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp, struct autofs_sb_info *sbi, struct autofs_dev_ioctl *param) { - struct nameidata nd; - const char *path; + struct path path; + const char *name; unsigned int type; unsigned int devid, magic; int err = -ENOENT; @@ -580,71 +530,46 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp, goto out; } - path = param->path; + name = param->path; type = param->ismountpoint.in.type; param->ismountpoint.out.devid = devid = 0; param->ismountpoint.out.magic = magic = 0; if (!fp || param->ioctlfd == -1) { - if (autofs_type_any(type)) { - struct super_block *sb; - - err = path_lookup(path, LOOKUP_FOLLOW, &nd); - if (err) - goto out; - - sb = nd.path.dentry->d_sb; - devid = new_encode_dev(sb->s_dev); - } else { - struct autofs_info *ino; - - err = path_lookup(path, LOOKUP_PARENT, &nd); - if (err) - goto out; - - err = autofs_dev_ioctl_find_sbi_type(&nd, type); - if (err) - goto out_release; - - ino = autofs4_dentry_ino(nd.path.dentry); - devid = autofs4_get_dev(ino->sbi); - } - + if (autofs_type_any(type)) + err = kern_path(name, LOOKUP_FOLLOW, &path); + else + err = find_autofs_mount(name, &path, test_by_type, &type); + if (err) + goto out; + devid = new_encode_dev(path.mnt->mnt_sb->s_dev); err = 0; - if (nd.path.dentry->d_inode && - nd.path.mnt->mnt_root == nd.path.dentry) { + if (path.dentry->d_inode && + path.mnt->mnt_root == path.dentry) { err = 1; - magic = nd.path.dentry->d_inode->i_sb->s_magic; + magic = path.dentry->d_inode->i_sb->s_magic; } } else { - dev_t dev = autofs4_get_dev(sbi); + dev_t dev = sbi->sb->s_dev; - err = path_lookup(path, LOOKUP_PARENT, &nd); + err = find_autofs_mount(name, &path, test_by_dev, &dev); if (err) goto out; - err = autofs_dev_ioctl_find_super(&nd, dev); - if (err) - goto out_release; + devid = new_encode_dev(dev); - devid = dev; + err = have_submounts(path.dentry); - err = have_submounts(nd.path.dentry); - - if (nd.path.mnt->mnt_mountpoint != nd.path.mnt->mnt_root) { - if (follow_down(&nd.path.mnt, &nd.path.dentry)) { - struct inode *inode = nd.path.dentry->d_inode; - magic = inode->i_sb->s_magic; - } + if (path.mnt->mnt_mountpoint != path.mnt->mnt_root) { + if (follow_down(&path.mnt, &path.dentry)) + magic = path.mnt->mnt_sb->s_magic; } } param->ismountpoint.out.devid = devid; param->ismountpoint.out.magic = magic; - -out_release: - path_put(&nd.path); + path_put(&path); out: return err; } From 9b4a9b14a793bc69b505ed916051f6f32db13bb8 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 11:44:16 -0400 Subject: [PATCH 03/87] Preparations to caching root in path_walk() Split do_path_lookup(), opencode the call from do_filp_open() do_filp_open() is the only caller of do_path_lookup() that cares about root afterwards (it keeps resolving symlinks on O_CREAT path after it'd done LOOKUP_PARENT walk). So when we start caching fs->root in path_walk(), it'll need a different treatment. Signed-off-by: Al Viro --- fs/namei.c | 35 +++++++++++++++++++++++------------ 1 file changed, 23 insertions(+), 12 deletions(-) diff --git a/fs/namei.c b/fs/namei.c index c82805d088e1..895733efc6b9 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1017,9 +1017,7 @@ static int path_walk(const char *name, struct nameidata *nd) return link_path_walk(name, nd); } -/* Returns 0 and nd will be valid on success; Retuns error, otherwise. */ -static int do_path_lookup(int dfd, const char *name, - unsigned int flags, struct nameidata *nd) +static int path_init(int dfd, const char *name, unsigned int flags, struct nameidata *nd) { int retval = 0; int fput_needed; @@ -1063,17 +1061,25 @@ static int do_path_lookup(int dfd, const char *name, fput_light(file, fput_needed); } - - retval = path_walk(name, nd); - if (unlikely(!retval && !audit_dummy_context() && nd->path.dentry && - nd->path.dentry->d_inode)) - audit_inode(name, nd->path.dentry); -out_fail: - return retval; + return 0; fput_fail: fput_light(file, fput_needed); - goto out_fail; +out_fail: + return retval; +} + +/* Returns 0 and nd will be valid on success; Retuns error, otherwise. */ +static int do_path_lookup(int dfd, const char *name, + unsigned int flags, struct nameidata *nd) +{ + int retval = path_init(dfd, name, flags, nd); + if (!retval) + retval = path_walk(name, nd); + if (unlikely(!retval && !audit_dummy_context() && nd->path.dentry && + nd->path.dentry->d_inode)) + audit_inode(name, nd->path.dentry); + return retval; } int path_lookup(const char *name, unsigned int flags, @@ -1676,9 +1682,14 @@ struct file *do_filp_open(int dfd, const char *pathname, /* * Create - we need to know the parent. */ - error = do_path_lookup(dfd, pathname, LOOKUP_PARENT, &nd); + error = path_init(dfd, pathname, LOOKUP_PARENT, &nd); if (error) return ERR_PTR(error); + error = path_walk(pathname, &nd); + if (error) + return ERR_PTR(error); + if (unlikely(!audit_dummy_context())) + audit_inode(pathname, nd.path.dentry); /* * We have the parent and last component. First of all, check From 2a737871108de9ba8930f7650d549f1383767f8b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 11:49:53 -0400 Subject: [PATCH 04/87] Cache root in nameidata New field: nd->root. When pathname resolution wants to know the root, check if nd->root.mnt is non-NULL; use nd->root if it is, otherwise copy current->fs->root there. After path_walk() is finished, we check if we'd got a cached value in nd->root and drop it. Before calling path_walk() we should either set nd->root.mnt to NULL *or* copy (and pin down) some path to nd->root. In the latter case we won't be looking at current->fs->root at all. Signed-off-by: Al Viro --- fs/namei.c | 53 +++++++++++++++++++++++++++---------------- include/linux/namei.h | 1 + 2 files changed, 35 insertions(+), 19 deletions(-) diff --git a/fs/namei.c b/fs/namei.c index 895733efc6b9..88baaf2b9167 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -552,6 +552,17 @@ static __always_inline int link_path_walk(const char *name, struct nameidata *nd return result; } +static __always_inline void set_root(struct nameidata *nd) +{ + if (!nd->root.mnt) { + struct fs_struct *fs = current->fs; + read_lock(&fs->lock); + nd->root = fs->root; + path_get(&nd->root); + read_unlock(&fs->lock); + } +} + static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *link) { int res = 0; @@ -560,14 +571,10 @@ static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *l goto fail; if (*link == '/') { - struct fs_struct *fs = current->fs; - + set_root(nd); path_put(&nd->path); - - read_lock(&fs->lock); - nd->path = fs->root; - path_get(&fs->root); - read_unlock(&fs->lock); + nd->path = nd->root; + path_get(&nd->root); } res = link_path_walk(link, nd); @@ -741,19 +748,16 @@ int follow_down(struct vfsmount **mnt, struct dentry **dentry) static __always_inline void follow_dotdot(struct nameidata *nd) { - struct fs_struct *fs = current->fs; + set_root(nd); while(1) { struct vfsmount *parent; struct dentry *old = nd->path.dentry; - read_lock(&fs->lock); - if (nd->path.dentry == fs->root.dentry && - nd->path.mnt == fs->root.mnt) { - read_unlock(&fs->lock); + if (nd->path.dentry == nd->root.dentry && + nd->path.mnt == nd->root.mnt) { break; } - read_unlock(&fs->lock); spin_lock(&dcache_lock); if (nd->path.dentry != nd->path.mnt->mnt_root) { nd->path.dentry = dget(nd->path.dentry->d_parent); @@ -1022,18 +1026,18 @@ static int path_init(int dfd, const char *name, unsigned int flags, struct namei int retval = 0; int fput_needed; struct file *file; - struct fs_struct *fs = current->fs; nd->last_type = LAST_ROOT; /* if there are only slashes... */ nd->flags = flags; nd->depth = 0; + nd->root.mnt = NULL; if (*name=='/') { - read_lock(&fs->lock); - nd->path = fs->root; - path_get(&fs->root); - read_unlock(&fs->lock); + set_root(nd); + nd->path = nd->root; + path_get(&nd->root); } else if (dfd == AT_FDCWD) { + struct fs_struct *fs = current->fs; read_lock(&fs->lock); nd->path = fs->pwd; path_get(&fs->pwd); @@ -1079,6 +1083,10 @@ static int do_path_lookup(int dfd, const char *name, if (unlikely(!retval && !audit_dummy_context() && nd->path.dentry && nd->path.dentry->d_inode)) audit_inode(name, nd->path.dentry); + if (nd->root.mnt) { + path_put(&nd->root); + nd->root.mnt = NULL; + } return retval; } @@ -1115,6 +1123,7 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt, nd->last_type = LAST_ROOT; nd->flags = flags; nd->depth = 0; + nd->root.mnt = NULL; nd->path.dentry = dentry; nd->path.mnt = mnt; @@ -1125,8 +1134,12 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt, nd->path.dentry->d_inode)) audit_inode(name, nd->path.dentry); - return retval; + if (nd->root.mnt) { + path_put(&nd->root); + nd->root.mnt = NULL; + } + return retval; } /** @@ -1817,6 +1830,8 @@ exit: if (!IS_ERR(nd.intent.open.file)) release_open_intent(&nd); exit_parent: + if (nd.root.mnt) + path_put(&nd.root); path_put(&nd.path); return ERR_PTR(error); diff --git a/include/linux/namei.h b/include/linux/namei.h index 518098fe63af..325dd3ad39a0 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -18,6 +18,7 @@ enum { MAX_NESTED_LINKS = 8 }; struct nameidata { struct path path; struct qstr last; + struct path root; unsigned int flags; int last_type; unsigned depth; From 5b857119538daac7118c1364d7ff3613f12b84d3 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 11:53:49 -0400 Subject: [PATCH 05/87] Make vfs_path_lookup() use starting point as root Signed-off-by: Al Viro --- fs/namei.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/fs/namei.c b/fs/namei.c index 88baaf2b9167..4379ef989709 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1123,21 +1123,20 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt, nd->last_type = LAST_ROOT; nd->flags = flags; nd->depth = 0; - nd->root.mnt = NULL; nd->path.dentry = dentry; nd->path.mnt = mnt; path_get(&nd->path); + nd->root = nd->path; + path_get(&nd->root); retval = path_walk(name, nd); if (unlikely(!retval && !audit_dummy_context() && nd->path.dentry && nd->path.dentry->d_inode)) audit_inode(name, nd->path.dentry); - if (nd->root.mnt) { - path_put(&nd->root); - nd->root.mnt = NULL; - } + path_put(&nd->root); + nd->root.mnt = NULL; return retval; } From dd5cae6e9772ecc62fd374f7a8ec10eb51c96c4d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 12:21:18 -0400 Subject: [PATCH 06/87] Don't bother with check_mnt() in do_add_mount() on shrinkable ones These guys are what we add as submounts; checks for "is that attached in our namespace" are simply irrelevant for those and counterproductive for use of private vfsmount trees a-la what NFS folks want. Signed-off-by: Al Viro --- fs/namespace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/namespace.c b/fs/namespace.c index 134d494158d9..88a904d5aa23 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1698,7 +1698,7 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path, follow_down(&path->mnt, &path->dentry)) ; err = -EINVAL; - if (!check_mnt(path->mnt)) + if (!(mnt_flags & MNT_SHRINKABLE) && !check_mnt(path->mnt)) goto unlock; /* Refuse the same filesystem on the same mount point */ From 55430e2ecee574e729c12d4063b3ecabfa98fa82 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 02:04:46 -0400 Subject: [PATCH 07/87] nfsd struct path use: exp_get_by_name() Signed-off-by: Al Viro --- fs/nfsd/export.c | 38 ++++++++++++++++++-------------------- 1 file changed, 18 insertions(+), 20 deletions(-) diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 5839b229cd0e..3f6d51b8c3ef 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -847,9 +847,8 @@ exp_get_fsid_key(svc_client *clp, int fsid) return exp_find_key(clp, FSID_NUM, fsidv, NULL); } -static svc_export *exp_get_by_name(svc_client *clp, struct vfsmount *mnt, - struct dentry *dentry, - struct cache_req *reqp) +static svc_export *exp_get_by_name(svc_client *clp, const struct path *path, + struct cache_req *reqp) { struct svc_export *exp, key; int err; @@ -858,8 +857,7 @@ static svc_export *exp_get_by_name(svc_client *clp, struct vfsmount *mnt, return ERR_PTR(-ENOENT); key.ex_client = clp; - key.ex_path.mnt = mnt; - key.ex_path.dentry = dentry; + key.ex_path = *path; exp = svc_export_lookup(&key); if (exp == NULL) @@ -877,20 +875,19 @@ static struct svc_export *exp_parent(svc_client *clp, struct vfsmount *mnt, struct dentry *dentry, struct cache_req *reqp) { + struct path path = {.mnt = mnt, .dentry = dentry}; svc_export *exp; - dget(dentry); - exp = exp_get_by_name(clp, mnt, dentry, reqp); + dget(path.dentry); + exp = exp_get_by_name(clp, &path, reqp); - while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(dentry)) { - struct dentry *parent; - - parent = dget_parent(dentry); - dput(dentry); - dentry = parent; - exp = exp_get_by_name(clp, mnt, dentry, reqp); + while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(path.dentry)) { + struct dentry *parent = dget_parent(path.dentry); + dput(path.dentry); + path.dentry = parent; + exp = exp_get_by_name(clp, &path, reqp); } - dput(dentry); + dput(path.dentry); return exp; } @@ -1018,7 +1015,7 @@ exp_export(struct nfsctl_export *nxp) goto out_put_clp; err = -EINVAL; - exp = exp_get_by_name(clp, path.mnt, path.dentry, NULL); + exp = exp_get_by_name(clp, &path, NULL); memset(&new, 0, sizeof(new)); @@ -1135,7 +1132,7 @@ exp_unexport(struct nfsctl_export *nxp) goto out_domain; err = -EINVAL; - exp = exp_get_by_name(dom, path.mnt, path.dentry, NULL); + exp = exp_get_by_name(dom, &path, NULL); path_put(&path); if (IS_ERR(exp)) goto out_domain; @@ -1207,7 +1204,7 @@ static struct svc_export *exp_find(struct auth_domain *clp, int fsid_type, if (IS_ERR(ek)) return ERR_CAST(ek); - exp = exp_get_by_name(clp, ek->ek_path.mnt, ek->ek_path.dentry, reqp); + exp = exp_get_by_name(clp, &ek->ek_path, reqp); cache_put(&ek->h, &svc_expkey_cache); if (IS_ERR(exp)) @@ -1251,12 +1248,13 @@ rqst_exp_get_by_name(struct svc_rqst *rqstp, struct vfsmount *mnt, struct dentry *dentry) { struct svc_export *gssexp, *exp = ERR_PTR(-ENOENT); + struct path path = {.mnt = mnt, .dentry = dentry}; if (rqstp->rq_client == NULL) goto gss; /* First try the auth_unix client: */ - exp = exp_get_by_name(rqstp->rq_client, mnt, dentry, + exp = exp_get_by_name(rqstp->rq_client, &path, &rqstp->rq_chandle); if (PTR_ERR(exp) == -ENOENT) goto gss; @@ -1269,7 +1267,7 @@ gss: /* Otherwise, try falling back on gss client */ if (rqstp->rq_gssclient == NULL) return exp; - gssexp = exp_get_by_name(rqstp->rq_gssclient, mnt, dentry, + gssexp = exp_get_by_name(rqstp->rq_gssclient, &path, &rqstp->rq_chandle); if (PTR_ERR(gssexp) == -ENOENT) return exp; From 5bf3bd2b5cb68ba43c91f5bd0ac043543fba2558 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 02:14:32 -0400 Subject: [PATCH 08/87] switch exp_parent() to struct path ... and lose the always-NULL last argument (non-NULL case had been split off a while ago). Signed-off-by: Al Viro --- fs/nfsd/export.c | 26 +++++++++++--------------- 1 file changed, 11 insertions(+), 15 deletions(-) diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 3f6d51b8c3ef..5149dabde555 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -871,23 +871,19 @@ static svc_export *exp_get_by_name(svc_client *clp, const struct path *path, /* * Find the export entry for a given dentry. */ -static struct svc_export *exp_parent(svc_client *clp, struct vfsmount *mnt, - struct dentry *dentry, - struct cache_req *reqp) +static struct svc_export *exp_parent(svc_client *clp, struct path *path) { - struct path path = {.mnt = mnt, .dentry = dentry}; - svc_export *exp; + struct dentry *saved = dget(path->dentry); + svc_export *exp = exp_get_by_name(clp, path, NULL); - dget(path.dentry); - exp = exp_get_by_name(clp, &path, reqp); - - while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(path.dentry)) { - struct dentry *parent = dget_parent(path.dentry); - dput(path.dentry); - path.dentry = parent; - exp = exp_get_by_name(clp, &path, reqp); + while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(path->dentry)) { + struct dentry *parent = dget_parent(path->dentry); + dput(path->dentry); + path->dentry = parent; + exp = exp_get_by_name(clp, path, NULL); } - dput(path.dentry); + dput(path->dentry); + path->dentry = saved; return exp; } @@ -1174,7 +1170,7 @@ exp_rootfh(svc_client *clp, char *name, struct knfsd_fh *f, int maxsize) dprintk("nfsd: exp_rootfh(%s [%p] %s:%s/%ld)\n", name, path.dentry, clp->name, inode->i_sb->s_id, inode->i_ino); - exp = exp_parent(clp, path.mnt, path.dentry, NULL); + exp = exp_parent(clp, &path); if (IS_ERR(exp)) { err = PTR_ERR(exp); goto out; From 91c9fa8f75877c0c1e455c23e8f8206c91c8f77f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 02:42:05 -0400 Subject: [PATCH 09/87] switch rqst_exp_get_by_name() Signed-off-by: Al Viro --- fs/nfsd/export.c | 15 ++++++--------- fs/nfsd/vfs.c | 32 ++++++++++++++++---------------- include/linux/nfsd/export.h | 3 +-- 3 files changed, 23 insertions(+), 27 deletions(-) diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 5149dabde555..84f5e5cb0863 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -1240,18 +1240,15 @@ __be32 check_nfsd_access(struct svc_export *exp, struct svc_rqst *rqstp) * use exp_get_by_name() or exp_find(). */ struct svc_export * -rqst_exp_get_by_name(struct svc_rqst *rqstp, struct vfsmount *mnt, - struct dentry *dentry) +rqst_exp_get_by_name(struct svc_rqst *rqstp, struct path *path) { struct svc_export *gssexp, *exp = ERR_PTR(-ENOENT); - struct path path = {.mnt = mnt, .dentry = dentry}; if (rqstp->rq_client == NULL) goto gss; /* First try the auth_unix client: */ - exp = exp_get_by_name(rqstp->rq_client, &path, - &rqstp->rq_chandle); + exp = exp_get_by_name(rqstp->rq_client, path, &rqstp->rq_chandle); if (PTR_ERR(exp) == -ENOENT) goto gss; if (IS_ERR(exp)) @@ -1263,8 +1260,7 @@ gss: /* Otherwise, try falling back on gss client */ if (rqstp->rq_gssclient == NULL) return exp; - gssexp = exp_get_by_name(rqstp->rq_gssclient, &path, - &rqstp->rq_chandle); + gssexp = exp_get_by_name(rqstp->rq_gssclient, path, &rqstp->rq_chandle); if (PTR_ERR(gssexp) == -ENOENT) return exp; if (!IS_ERR(exp)) @@ -1307,9 +1303,10 @@ rqst_exp_parent(struct svc_rqst *rqstp, struct vfsmount *mnt, struct dentry *dentry) { struct svc_export *exp; + struct path path = {.mnt = mnt, .dentry = dentry}; dget(dentry); - exp = rqst_exp_get_by_name(rqstp, mnt, dentry); + exp = rqst_exp_get_by_name(rqstp, &path); while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(dentry)) { struct dentry *parent; @@ -1317,7 +1314,7 @@ rqst_exp_parent(struct svc_rqst *rqstp, struct vfsmount *mnt, parent = dget_parent(dentry); dput(dentry); dentry = parent; - exp = rqst_exp_get_by_name(rqstp, mnt, dentry); + exp = rqst_exp_get_by_name(rqstp, &path); } dput(dentry); return exp; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index bd584bcf1d9f..d84c4eaa526b 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -101,36 +101,36 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp, { struct svc_export *exp = *expp, *exp2 = NULL; struct dentry *dentry = *dpp; - struct vfsmount *mnt = mntget(exp->ex_path.mnt); - struct dentry *mounts = dget(dentry); + struct path path = {.mnt = mntget(exp->ex_path.mnt), + .dentry = dget(dentry)}; int err = 0; - while (follow_down(&mnt,&mounts)&&d_mountpoint(mounts)); + while (follow_down(&path.mnt, &path.dentry) && + d_mountpoint(path.dentry)) + ; - exp2 = rqst_exp_get_by_name(rqstp, mnt, mounts); + exp2 = rqst_exp_get_by_name(rqstp, &path); if (IS_ERR(exp2)) { if (PTR_ERR(exp2) != -ENOENT) err = PTR_ERR(exp2); - dput(mounts); - mntput(mnt); + path_put(&path); goto out; } if ((exp->ex_flags & NFSEXP_CROSSMOUNT) || EX_NOHIDE(exp2)) { /* successfully crossed mount point */ /* - * This is subtle: dentry is *not* under mnt at this point. - * The only reason we are safe is that original mnt is pinned - * down by exp, so we should dput before putting exp. + * This is subtle: path.dentry is *not* on path.mnt + * at this point. The only reason we are safe is that + * original mnt is pinned down by exp, so we should + * put path *before* putting exp */ - dput(dentry); - *dpp = mounts; - exp_put(exp); + *dpp = path.dentry; + path.dentry = dentry; *expp = exp2; - } else { - exp_put(exp2); - dput(mounts); + exp2 = exp; } - mntput(mnt); + path_put(&path); + exp_put(exp2); out: return err; } diff --git a/include/linux/nfsd/export.h b/include/linux/nfsd/export.h index bcd0201589f8..98f6fd584d53 100644 --- a/include/linux/nfsd/export.h +++ b/include/linux/nfsd/export.h @@ -125,8 +125,7 @@ void nfsd_export_flush(void); void exp_readlock(void); void exp_readunlock(void); struct svc_export * rqst_exp_get_by_name(struct svc_rqst *, - struct vfsmount *, - struct dentry *); + struct path *); struct svc_export * rqst_exp_parent(struct svc_rqst *, struct vfsmount *mnt, struct dentry *dentry); From e64c390ca0b60fd2119331ef1fa888d7ea27e424 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 03:00:46 -0400 Subject: [PATCH 10/87] switch rqst_exp_parent() Signed-off-by: Al Viro --- fs/nfsd/export.c | 25 ++++++++++--------------- fs/nfsd/vfs.c | 25 +++++++++++++------------ include/linux/nfsd/export.h | 3 +-- 3 files changed, 24 insertions(+), 29 deletions(-) diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 84f5e5cb0863..8b1f8efb4690 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -1299,24 +1299,19 @@ gss: } struct svc_export * -rqst_exp_parent(struct svc_rqst *rqstp, struct vfsmount *mnt, - struct dentry *dentry) +rqst_exp_parent(struct svc_rqst *rqstp, struct path *path) { - struct svc_export *exp; - struct path path = {.mnt = mnt, .dentry = dentry}; + struct dentry *saved = dget(path->dentry); + struct svc_export *exp = rqst_exp_get_by_name(rqstp, path); - dget(dentry); - exp = rqst_exp_get_by_name(rqstp, &path); - - while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(dentry)) { - struct dentry *parent; - - parent = dget_parent(dentry); - dput(dentry); - dentry = parent; - exp = rqst_exp_get_by_name(rqstp, &path); + while (PTR_ERR(exp) == -ENOENT && !IS_ROOT(path->dentry)) { + struct dentry *parent = dget_parent(path->dentry); + dput(path->dentry); + path->dentry = parent; + exp = rqst_exp_get_by_name(rqstp, path); } - dput(dentry); + dput(path->dentry); + path->dentry = saved; return exp; } diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index d84c4eaa526b..9f1ea3127f5d 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -169,28 +169,29 @@ nfsd_lookup_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp, /* checking mountpoint crossing is very different when stepping up */ struct svc_export *exp2 = NULL; struct dentry *dp; - struct vfsmount *mnt = mntget(exp->ex_path.mnt); - dentry = dget(dparent); - while(dentry == mnt->mnt_root && follow_up(&mnt, &dentry)) - ; - dp = dget_parent(dentry); - dput(dentry); - dentry = dp; + struct path path = {.mnt = mntget(exp->ex_path.mnt), + .dentry = dget(dparent)}; - exp2 = rqst_exp_parent(rqstp, mnt, dentry); + while (path.dentry == path.mnt->mnt_root && + follow_up(&path.mnt, &path.dentry)) + ; + dp = dget_parent(path.dentry); + dput(path.dentry); + path.dentry = dp; + + exp2 = rqst_exp_parent(rqstp, &path); if (PTR_ERR(exp2) == -ENOENT) { - dput(dentry); dentry = dget(dparent); } else if (IS_ERR(exp2)) { host_err = PTR_ERR(exp2); - dput(dentry); - mntput(mnt); + path_put(&path); goto out_nfserr; } else { + dentry = dget(path.dentry); exp_put(exp); exp = exp2; } - mntput(mnt); + path_put(&path); } } else { fh_lock(fhp); diff --git a/include/linux/nfsd/export.h b/include/linux/nfsd/export.h index 98f6fd584d53..a6d9ef2bb34a 100644 --- a/include/linux/nfsd/export.h +++ b/include/linux/nfsd/export.h @@ -127,8 +127,7 @@ void exp_readunlock(void); struct svc_export * rqst_exp_get_by_name(struct svc_rqst *, struct path *); struct svc_export * rqst_exp_parent(struct svc_rqst *, - struct vfsmount *mnt, - struct dentry *dentry); + struct path *); int exp_rootfh(struct auth_domain *, char *path, struct knfsd_fh *, int maxsize); __be32 exp_pseudoroot(struct svc_rqst *, struct svc_fh *); From bab77ebf51e3902f608ecf08c9d34a0a52ac35a9 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 03:26:48 -0400 Subject: [PATCH 11/87] switch follow_up() to struct path Signed-off-by: Al Viro --- fs/autofs4/dev-ioctl.c | 2 +- fs/namei.c | 16 ++++++++-------- fs/nfsd/vfs.c | 2 +- include/linux/namei.h | 2 +- 4 files changed, 11 insertions(+), 11 deletions(-) diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c index f71dac9986f0..670407576b25 100644 --- a/fs/autofs4/dev-ioctl.c +++ b/fs/autofs4/dev-ioctl.c @@ -212,7 +212,7 @@ static int find_autofs_mount(const char *pathname, err = 0; } } - if (!follow_up(&path.mnt, &path.dentry)) + if (!follow_up(&path)) break; } path_put(&path); diff --git a/fs/namei.c b/fs/namei.c index 4379ef989709..8c1f48ae68e7 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -675,23 +675,23 @@ loop: return err; } -int follow_up(struct vfsmount **mnt, struct dentry **dentry) +int follow_up(struct path *path) { struct vfsmount *parent; struct dentry *mountpoint; spin_lock(&vfsmount_lock); - parent=(*mnt)->mnt_parent; - if (parent == *mnt) { + parent = path->mnt->mnt_parent; + if (parent == path->mnt) { spin_unlock(&vfsmount_lock); return 0; } mntget(parent); - mountpoint=dget((*mnt)->mnt_mountpoint); + mountpoint = dget(path->mnt->mnt_mountpoint); spin_unlock(&vfsmount_lock); - dput(*dentry); - *dentry = mountpoint; - mntput(*mnt); - *mnt = parent; + dput(path->dentry); + path->dentry = mountpoint; + mntput(path->mnt); + path->mnt = parent; return 1; } diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 9f1ea3127f5d..7b2b3f775326 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -173,7 +173,7 @@ nfsd_lookup_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp, .dentry = dget(dparent)}; while (path.dentry == path.mnt->mnt_root && - follow_up(&path.mnt, &path.dentry)) + follow_up(&path)) ; dp = dget_parent(path.dentry); dput(path.dentry); diff --git a/include/linux/namei.h b/include/linux/namei.h index 325dd3ad39a0..9cd5a717be3b 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -79,7 +79,7 @@ extern struct dentry *lookup_one_len(const char *, struct dentry *, int); extern struct dentry *lookup_one_noperm(const char *, struct dentry *); extern int follow_down(struct vfsmount **, struct dentry **); -extern int follow_up(struct vfsmount **, struct dentry **); +extern int follow_up(struct path *); extern struct dentry *lock_rename(struct dentry *, struct dentry *); extern void unlock_rename(struct dentry *, struct dentry *); From 589ff870ed60a9ebdd5ec99ec3f5afe1282fe151 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 03:28:19 -0400 Subject: [PATCH 12/87] Switch collect_mounts() to struct path Signed-off-by: Al Viro --- fs/namespace.c | 4 ++-- include/linux/fs.h | 2 +- kernel/audit_tree.c | 6 +++--- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/namespace.c b/fs/namespace.c index 88a904d5aa23..c85962206aad 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1253,11 +1253,11 @@ Enomem: return NULL; } -struct vfsmount *collect_mounts(struct vfsmount *mnt, struct dentry *dentry) +struct vfsmount *collect_mounts(struct path *path) { struct vfsmount *tree; down_write(&namespace_sem); - tree = copy_tree(mnt, dentry, CL_COPY_ALL | CL_PRIVATE); + tree = copy_tree(path->mnt, path->dentry, CL_COPY_ALL | CL_PRIVATE); up_write(&namespace_sem); return tree; } diff --git a/include/linux/fs.h b/include/linux/fs.h index 323b5ce474c1..03fb2102b8f3 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1800,7 +1800,7 @@ extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data); extern int may_umount_tree(struct vfsmount *); extern int may_umount(struct vfsmount *); extern long do_mount(char *, char *, char *, unsigned long, void *); -extern struct vfsmount *collect_mounts(struct vfsmount *, struct dentry *); +extern struct vfsmount *collect_mounts(struct path *); extern void drop_collected_mounts(struct vfsmount *); extern int vfs_statfs(struct dentry *, struct kstatfs *); diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index 6e7351739a82..1f6396d76687 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -568,7 +568,7 @@ void audit_trim_trees(void) if (err) goto skip_it; - root_mnt = collect_mounts(path.mnt, path.dentry); + root_mnt = collect_mounts(&path); path_put(&path); if (!root_mnt) goto skip_it; @@ -660,7 +660,7 @@ int audit_add_tree_rule(struct audit_krule *rule) err = kern_path(tree->pathname, 0, &path); if (err) goto Err; - mnt = collect_mounts(path.mnt, path.dentry); + mnt = collect_mounts(&path); path_put(&path); if (!mnt) { err = -ENOMEM; @@ -720,7 +720,7 @@ int audit_tag_tree(char *old, char *new) err = kern_path(new, 0, &path); if (err) return err; - tagged = collect_mounts(path.mnt, path.dentry); + tagged = collect_mounts(&path); path_put(&path); if (!tagged) return -ENOMEM; From 9393bd07cf218ca51d0e627653f906a9d76a9131 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 13:58:15 -0400 Subject: [PATCH 13/87] switch follow_down() Signed-off-by: Al Viro --- fs/afs/mntpt.c | 2 +- fs/autofs/dirhash.c | 5 ++--- fs/autofs4/autofs_i.h | 6 +++--- fs/autofs4/dev-ioctl.c | 2 +- fs/autofs4/expire.c | 15 +++++++-------- fs/autofs4/root.c | 7 +++---- fs/cifs/cifs_dfs_ref.c | 2 +- fs/namei.c | 12 ++++++------ fs/namespace.c | 4 ++-- fs/nfs/namespace.c | 2 +- fs/nfsd/vfs.c | 3 +-- include/linux/namei.h | 2 +- 12 files changed, 29 insertions(+), 33 deletions(-) diff --git a/fs/afs/mntpt.c b/fs/afs/mntpt.c index 2b9e2d03a390..c52be53f6946 100644 --- a/fs/afs/mntpt.c +++ b/fs/afs/mntpt.c @@ -244,7 +244,7 @@ static void *afs_mntpt_follow_link(struct dentry *dentry, struct nameidata *nd) case -EBUSY: /* someone else made a mount here whilst we were busy */ while (d_mountpoint(nd->path.dentry) && - follow_down(&nd->path.mnt, &nd->path.dentry)) + follow_down(&nd->path)) ; err = 0; default: diff --git a/fs/autofs/dirhash.c b/fs/autofs/dirhash.c index 4eb4d8dfb2f1..2316e944a109 100644 --- a/fs/autofs/dirhash.c +++ b/fs/autofs/dirhash.c @@ -85,13 +85,12 @@ struct autofs_dir_ent *autofs_expire(struct super_block *sb, } path.mnt = mnt; path_get(&path); - if (!follow_down(&path.mnt, &path.dentry)) { + if (!follow_down(&path)) { path_put(&path); DPRINTK(("autofs: not expirable (not a mounted directory): %s\n", ent->name)); continue; } - while (d_mountpoint(path.dentry) && - follow_down(&path.mnt, &path.dentry)) + while (d_mountpoint(path.dentry) && follow_down(&path)); ; umount_ok = may_umount(path.mnt); path_put(&path); diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index b7ff33c63101..8f7cdde41733 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -223,12 +223,12 @@ int autofs4_wait(struct autofs_sb_info *,struct dentry *, enum autofs_notify); int autofs4_wait_release(struct autofs_sb_info *,autofs_wqt_t,int); void autofs4_catatonic_mode(struct autofs_sb_info *); -static inline int autofs4_follow_mount(struct vfsmount **mnt, struct dentry **dentry) +static inline int autofs4_follow_mount(struct path *path) { int res = 0; - while (d_mountpoint(*dentry)) { - int followed = follow_down(mnt, dentry); + while (d_mountpoint(path->dentry)) { + int followed = follow_down(path); if (!followed) break; res = 1; diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c index 670407576b25..f3da2eb51f56 100644 --- a/fs/autofs4/dev-ioctl.c +++ b/fs/autofs4/dev-ioctl.c @@ -562,7 +562,7 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp, err = have_submounts(path.dentry); if (path.mnt->mnt_mountpoint != path.mnt->mnt_root) { - if (follow_down(&path.mnt, &path.dentry)) + if (follow_down(&path)) magic = path.mnt->mnt_sb->s_magic; } } diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c index 3077d8f16523..aa39ae83f019 100644 --- a/fs/autofs4/expire.c +++ b/fs/autofs4/expire.c @@ -48,19 +48,19 @@ static inline int autofs4_can_expire(struct dentry *dentry, static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry) { struct dentry *top = dentry; + struct path path = {.mnt = mnt, .dentry = dentry}; int status = 1; DPRINTK("dentry %p %.*s", dentry, (int)dentry->d_name.len, dentry->d_name.name); - mntget(mnt); - dget(dentry); + path_get(&path); - if (!follow_down(&mnt, &dentry)) + if (!follow_down(&path)) goto done; - if (is_autofs4_dentry(dentry)) { - struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); + if (is_autofs4_dentry(path.dentry)) { + struct autofs_sb_info *sbi = autofs4_sbi(path.dentry->d_sb); /* This is an autofs submount, we can't expire it */ if (autofs_type_indirect(sbi->type)) @@ -70,7 +70,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry) * Otherwise it's an offset mount and we need to check * if we can umount its mount, if there is one. */ - if (!d_mountpoint(dentry)) { + if (!d_mountpoint(path.dentry)) { status = 0; goto done; } @@ -86,8 +86,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry) status = 0; done: DPRINTK("returning = %d", status); - dput(dentry); - mntput(mnt); + path_put(&path); return status; } diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index e383bf0334f1..b96a3c57359d 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -181,7 +181,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) nd->flags); /* * For an expire of a covered direct or offset mount we need - * to beeak out of follow_down() at the autofs mount trigger + * to break out of follow_down() at the autofs mount trigger * (d_mounted--), so we can see the expiring flag, and manage * the blocking and following here until the expire is completed. */ @@ -190,7 +190,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd) if (ino->flags & AUTOFS_INF_EXPIRING) { spin_unlock(&sbi->fs_lock); /* Follow down to our covering mount. */ - if (!follow_down(&nd->path.mnt, &nd->path.dentry)) + if (!follow_down(&nd->path)) goto done; goto follow; } @@ -230,8 +230,7 @@ follow: * to follow it. */ if (d_mountpoint(dentry)) { - if (!autofs4_follow_mount(&nd->path.mnt, - &nd->path.dentry)) { + if (!autofs4_follow_mount(&nd->path)) { status = -ENOENT; goto out_error; } diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c index 83d62759c7c7..3bb11be8b6a8 100644 --- a/fs/cifs/cifs_dfs_ref.c +++ b/fs/cifs/cifs_dfs_ref.c @@ -275,7 +275,7 @@ static int add_mount_helper(struct vfsmount *newmnt, struct nameidata *nd, case -EBUSY: /* someone else made a mount here whilst we were busy */ while (d_mountpoint(nd->path.dentry) && - follow_down(&nd->path.mnt, &nd->path.dentry)) + follow_down(&nd->path)) ; err = 0; default: diff --git a/fs/namei.c b/fs/namei.c index 8c1f48ae68e7..4d49a3eee6d4 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -731,16 +731,16 @@ static void follow_mount(struct vfsmount **mnt, struct dentry **dentry) /* no need for dcache_lock, as serialization is taken care in * namespace.c */ -int follow_down(struct vfsmount **mnt, struct dentry **dentry) +int follow_down(struct path *path) { struct vfsmount *mounted; - mounted = lookup_mnt(*mnt, *dentry); + mounted = lookup_mnt(path->mnt, path->dentry); if (mounted) { - dput(*dentry); - mntput(*mnt); - *mnt = mounted; - *dentry = dget(mounted->mnt_root); + dput(path->dentry); + mntput(path->mnt); + path->mnt = mounted; + path->dentry = dget(mounted->mnt_root); return 1; } return 0; diff --git a/fs/namespace.c b/fs/namespace.c index c85962206aad..ba5237be1cf9 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1601,7 +1601,7 @@ static int do_move_mount(struct path *path, char *old_name) down_write(&namespace_sem); while (d_mountpoint(path->dentry) && - follow_down(&path->mnt, &path->dentry)) + follow_down(path)) ; err = -EINVAL; if (!check_mnt(path->mnt) || !check_mnt(old_path.mnt)) @@ -1695,7 +1695,7 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path, down_write(&namespace_sem); /* Something was mounted here while we slept */ while (d_mountpoint(path->dentry) && - follow_down(&path->mnt, &path->dentry)) + follow_down(path)) ; err = -EINVAL; if (!(mnt_flags & MNT_SHRINKABLE) && !check_mnt(path->mnt)) diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c index 64a288ee046d..f01caec84463 100644 --- a/fs/nfs/namespace.c +++ b/fs/nfs/namespace.c @@ -154,7 +154,7 @@ out_err: goto out; out_follow: while (d_mountpoint(nd->path.dentry) && - follow_down(&nd->path.mnt, &nd->path.dentry)) + follow_down(&nd->path)) ; err = 0; goto out; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 7b2b3f775326..99f835753596 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -105,8 +105,7 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp, .dentry = dget(dentry)}; int err = 0; - while (follow_down(&path.mnt, &path.dentry) && - d_mountpoint(path.dentry)) + while (d_mountpoint(path.dentry) && follow_down(&path)) ; exp2 = rqst_exp_get_by_name(rqstp, &path); diff --git a/include/linux/namei.h b/include/linux/namei.h index 9cd5a717be3b..d870ae2faedc 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -78,7 +78,7 @@ extern void release_open_intent(struct nameidata *); extern struct dentry *lookup_one_len(const char *, struct dentry *, int); extern struct dentry *lookup_one_noperm(const char *, struct dentry *); -extern int follow_down(struct vfsmount **, struct dentry **); +extern int follow_down(struct path *); extern int follow_up(struct path *); extern struct dentry *lock_rename(struct dentry *, struct dentry *); From 79ed0226198c628133530b179a90dbf42b1c2eba Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 13:59:41 -0400 Subject: [PATCH 14/87] switch follow_mount() Signed-off-by: Al Viro --- fs/namei.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/fs/namei.c b/fs/namei.c index 4d49a3eee6d4..c006bc61d1ea 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -715,16 +715,16 @@ static int __follow_mount(struct path *path) return res; } -static void follow_mount(struct vfsmount **mnt, struct dentry **dentry) +static void follow_mount(struct path *path) { - while (d_mountpoint(*dentry)) { - struct vfsmount *mounted = lookup_mnt(*mnt, *dentry); + while (d_mountpoint(path->dentry)) { + struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry); if (!mounted) break; - dput(*dentry); - mntput(*mnt); - *mnt = mounted; - *dentry = dget(mounted->mnt_root); + dput(path->dentry); + mntput(path->mnt); + path->mnt = mounted; + path->dentry = dget(mounted->mnt_root); } } @@ -779,7 +779,7 @@ static __always_inline void follow_dotdot(struct nameidata *nd) mntput(nd->path.mnt); nd->path.mnt = parent; } - follow_mount(&nd->path.mnt, &nd->path.dentry); + follow_mount(&nd->path); } /* From 1c755af4df75996b0dd4b7e6cacaf9d57a6ef2ef Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 18 Apr 2009 14:06:57 -0400 Subject: [PATCH 15/87] switch lookup_mnt() Signed-off-by: Al Viro --- fs/namei.c | 6 +++--- fs/namespace.c | 4 ++-- include/linux/dcache.h | 2 +- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/namei.c b/fs/namei.c index c006bc61d1ea..527119afb6a5 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -702,7 +702,7 @@ static int __follow_mount(struct path *path) { int res = 0; while (d_mountpoint(path->dentry)) { - struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry); + struct vfsmount *mounted = lookup_mnt(path); if (!mounted) break; dput(path->dentry); @@ -718,7 +718,7 @@ static int __follow_mount(struct path *path) static void follow_mount(struct path *path) { while (d_mountpoint(path->dentry)) { - struct vfsmount *mounted = lookup_mnt(path->mnt, path->dentry); + struct vfsmount *mounted = lookup_mnt(path); if (!mounted) break; dput(path->dentry); @@ -735,7 +735,7 @@ int follow_down(struct path *path) { struct vfsmount *mounted; - mounted = lookup_mnt(path->mnt, path->dentry); + mounted = lookup_mnt(path); if (mounted) { dput(path->dentry); mntput(path->mnt); diff --git a/fs/namespace.c b/fs/namespace.c index ba5237be1cf9..b94ad3d685ff 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -442,11 +442,11 @@ struct vfsmount *__lookup_mnt(struct vfsmount *mnt, struct dentry *dentry, * lookup_mnt increments the ref count before returning * the vfsmount struct. */ -struct vfsmount *lookup_mnt(struct vfsmount *mnt, struct dentry *dentry) +struct vfsmount *lookup_mnt(struct path *path) { struct vfsmount *child_mnt; spin_lock(&vfsmount_lock); - if ((child_mnt = __lookup_mnt(mnt, dentry, 1))) + if ((child_mnt = __lookup_mnt(path->mnt, path->dentry, 1))) mntget(child_mnt); spin_unlock(&vfsmount_lock); return child_mnt; diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 97978004338d..72ce2ae88591 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -370,7 +370,7 @@ static inline int d_mountpoint(struct dentry *dentry) return dentry->d_mounted; } -extern struct vfsmount *lookup_mnt(struct vfsmount *, struct dentry *); +extern struct vfsmount *lookup_mnt(struct path *); extern struct dentry *lookup_create(struct nameidata *nd, int is_dir); extern int sysctl_vfs_cache_pressure; From 3174c21b74b56c6a53fddd41a30fd6f757a32bd0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 7 Apr 2009 13:19:18 -0400 Subject: [PATCH 16/87] Move junk from proc_fs.h to fs/proc/internal.h Signed-off-by: Al Viro --- fs/proc/internal.h | 25 +++++++++++++++++++++++++ fs/proc/proc_devtree.c | 1 + include/linux/proc_fs.h | 24 ------------------------ 3 files changed, 26 insertions(+), 24 deletions(-) diff --git a/fs/proc/internal.h b/fs/proc/internal.h index f6db9618a888..753ca37002c8 100644 --- a/fs/proc/internal.h +++ b/fs/proc/internal.h @@ -92,3 +92,28 @@ struct pde_opener { struct list_head lh; }; void pde_users_dec(struct proc_dir_entry *pde); + +extern spinlock_t proc_subdir_lock; + +struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *); +int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir); +unsigned long task_vsize(struct mm_struct *); +int task_statm(struct mm_struct *, int *, int *, int *, int *); +void task_mem(struct seq_file *, struct mm_struct *); + +struct proc_dir_entry *de_get(struct proc_dir_entry *de); +void de_put(struct proc_dir_entry *de); + +extern struct vfsmount *proc_mnt; +int proc_fill_super(struct super_block *); +struct inode *proc_get_inode(struct super_block *, unsigned int, struct proc_dir_entry *); + +/* + * These are generic /proc routines that use the internal + * "struct proc_dir_entry" tree to traverse the filesystem. + * + * The /proc root directory has extended versions to take care + * of the /proc/ subdirectories. + */ +int proc_readdir(struct file *, void *, filldir_t); +struct dentry *proc_lookup(struct inode *, struct dentry *, struct nameidata *); diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c index de2bba5a3440..fc6c3025befd 100644 --- a/fs/proc/proc_devtree.c +++ b/fs/proc/proc_devtree.c @@ -11,6 +11,7 @@ #include #include #include +#include "internal.h" #ifndef HAVE_ARCH_DEVTREE_FIXUPS static inline void set_node_proc_entry(struct device_node *np, diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index fbfa3d44d33d..e6e77d31c418 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -93,20 +93,9 @@ struct vmcore { #ifdef CONFIG_PROC_FS -extern spinlock_t proc_subdir_lock; - extern void proc_root_init(void); void proc_flush_task(struct task_struct *task); -struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *); -int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir); -unsigned long task_vsize(struct mm_struct *); -int task_statm(struct mm_struct *, int *, int *, int *, int *); -void task_mem(struct seq_file *, struct mm_struct *); -void clear_refs_smap(struct mm_struct *mm); - -struct proc_dir_entry *de_get(struct proc_dir_entry *de); -void de_put(struct proc_dir_entry *de); extern struct proc_dir_entry *create_proc_entry(const char *name, mode_t mode, struct proc_dir_entry *parent); @@ -116,20 +105,7 @@ struct proc_dir_entry *proc_create_data(const char *name, mode_t mode, void *data); extern void remove_proc_entry(const char *name, struct proc_dir_entry *parent); -extern struct vfsmount *proc_mnt; struct pid_namespace; -extern int proc_fill_super(struct super_block *); -extern struct inode *proc_get_inode(struct super_block *, unsigned int, struct proc_dir_entry *); - -/* - * These are generic /proc routines that use the internal - * "struct proc_dir_entry" tree to traverse the filesystem. - * - * The /proc root directory has extended versions to take care - * of the /proc/ subdirectories. - */ -extern int proc_readdir(struct file *, void *, filldir_t); -extern struct dentry *proc_lookup(struct inode *, struct dentry *, struct nameidata *); extern int pid_ns_prepare_proc(struct pid_namespace *ns); extern void pid_ns_release_proc(struct pid_namespace *ns); From d3ef3d7351ccfbef3e5d926efc5ee332136f40d4 Mon Sep 17 00:00:00 2001 From: "npiggin@suse.de" Date: Sun, 26 Apr 2009 20:25:54 +1000 Subject: [PATCH 17/87] fs: mnt_want_write speedup This patch speeds up lmbench lat_mmap test by about 8%. lat_mmap is set up basically to mmap a 64MB file on tmpfs, fault in its pages, then unmap it. A microbenchmark yes, but it exercises some important paths in the mm. Before: avg = 501.9 std = 14.7773 After: avg = 462.286 std = 5.46106 (50 runs of each, stddev gives a reasonable confidence, but there is quite a bit of variation there still) It does this by removing the complex per-cpu locking and counter-cache and replaces it with a percpu counter in struct vfsmount. This makes the code much simpler, and avoids spinlocks (although the msync is still pretty costly, unfortunately). It results in about 900 bytes smaller code too. It does increase the size of a vfsmount, however. It should also give a speedup on large systems if CPUs are frequently operating on different mounts (because the existing scheme has to operate on an atomic in the struct vfsmount when switching between mounts). But I'm most interested in the single threaded path performance for the moment. [AV: minor cleanup] Cc: Dave Hansen Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/namespace.c | 276 +++++++++++++++--------------------------- include/linux/mount.h | 21 +++- 2 files changed, 110 insertions(+), 187 deletions(-) diff --git a/fs/namespace.c b/fs/namespace.c index b94ad3d685ff..22ae06ad751d 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -131,10 +131,20 @@ struct vfsmount *alloc_vfsmnt(const char *name) INIT_LIST_HEAD(&mnt->mnt_share); INIT_LIST_HEAD(&mnt->mnt_slave_list); INIT_LIST_HEAD(&mnt->mnt_slave); - atomic_set(&mnt->__mnt_writers, 0); +#ifdef CONFIG_SMP + mnt->mnt_writers = alloc_percpu(int); + if (!mnt->mnt_writers) + goto out_free_devname; +#else + mnt->mnt_writers = 0; +#endif } return mnt; +#ifdef CONFIG_SMP +out_free_devname: + kfree(mnt->mnt_devname); +#endif out_free_id: mnt_free_id(mnt); out_free_cache: @@ -171,65 +181,38 @@ int __mnt_is_readonly(struct vfsmount *mnt) } EXPORT_SYMBOL_GPL(__mnt_is_readonly); -struct mnt_writer { - /* - * If holding multiple instances of this lock, they - * must be ordered by cpu number. - */ - spinlock_t lock; - struct lock_class_key lock_class; /* compiles out with !lockdep */ - unsigned long count; - struct vfsmount *mnt; -} ____cacheline_aligned_in_smp; -static DEFINE_PER_CPU(struct mnt_writer, mnt_writers); - -static int __init init_mnt_writers(void) +static inline void inc_mnt_writers(struct vfsmount *mnt) { - int cpu; - for_each_possible_cpu(cpu) { - struct mnt_writer *writer = &per_cpu(mnt_writers, cpu); - spin_lock_init(&writer->lock); - lockdep_set_class(&writer->lock, &writer->lock_class); - writer->count = 0; - } - return 0; +#ifdef CONFIG_SMP + (*per_cpu_ptr(mnt->mnt_writers, smp_processor_id()))++; +#else + mnt->mnt_writers++; +#endif } -fs_initcall(init_mnt_writers); -static void unlock_mnt_writers(void) +static inline void dec_mnt_writers(struct vfsmount *mnt) { +#ifdef CONFIG_SMP + (*per_cpu_ptr(mnt->mnt_writers, smp_processor_id()))--; +#else + mnt->mnt_writers--; +#endif +} + +static unsigned int count_mnt_writers(struct vfsmount *mnt) +{ +#ifdef CONFIG_SMP + unsigned int count = 0; int cpu; - struct mnt_writer *cpu_writer; for_each_possible_cpu(cpu) { - cpu_writer = &per_cpu(mnt_writers, cpu); - spin_unlock(&cpu_writer->lock); + count += *per_cpu_ptr(mnt->mnt_writers, cpu); } -} -static inline void __clear_mnt_count(struct mnt_writer *cpu_writer) -{ - if (!cpu_writer->mnt) - return; - /* - * This is in case anyone ever leaves an invalid, - * old ->mnt and a count of 0. - */ - if (!cpu_writer->count) - return; - atomic_add(cpu_writer->count, &cpu_writer->mnt->__mnt_writers); - cpu_writer->count = 0; -} - /* - * must hold cpu_writer->lock - */ -static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer, - struct vfsmount *mnt) -{ - if (cpu_writer->mnt == mnt) - return; - __clear_mnt_count(cpu_writer); - cpu_writer->mnt = mnt; + return count; +#else + return mnt->mnt_writers; +#endif } /* @@ -253,75 +236,34 @@ static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer, int mnt_want_write(struct vfsmount *mnt) { int ret = 0; - struct mnt_writer *cpu_writer; - cpu_writer = &get_cpu_var(mnt_writers); - spin_lock(&cpu_writer->lock); + preempt_disable(); + inc_mnt_writers(mnt); + /* + * The store to inc_mnt_writers must be visible before we pass + * MNT_WRITE_HOLD loop below, so that the slowpath can see our + * incremented count after it has set MNT_WRITE_HOLD. + */ + smp_mb(); + while (mnt->mnt_flags & MNT_WRITE_HOLD) + cpu_relax(); + /* + * After the slowpath clears MNT_WRITE_HOLD, mnt_is_readonly will + * be set to match its requirements. So we must not load that until + * MNT_WRITE_HOLD is cleared. + */ + smp_rmb(); if (__mnt_is_readonly(mnt)) { + dec_mnt_writers(mnt); ret = -EROFS; goto out; } - use_cpu_writer_for_mount(cpu_writer, mnt); - cpu_writer->count++; out: - spin_unlock(&cpu_writer->lock); - put_cpu_var(mnt_writers); + preempt_enable(); return ret; } EXPORT_SYMBOL_GPL(mnt_want_write); -static void lock_mnt_writers(void) -{ - int cpu; - struct mnt_writer *cpu_writer; - - for_each_possible_cpu(cpu) { - cpu_writer = &per_cpu(mnt_writers, cpu); - spin_lock(&cpu_writer->lock); - __clear_mnt_count(cpu_writer); - cpu_writer->mnt = NULL; - } -} - -/* - * These per-cpu write counts are not guaranteed to have - * matched increments and decrements on any given cpu. - * A file open()ed for write on one cpu and close()d on - * another cpu will imbalance this count. Make sure it - * does not get too far out of whack. - */ -static void handle_write_count_underflow(struct vfsmount *mnt) -{ - if (atomic_read(&mnt->__mnt_writers) >= - MNT_WRITER_UNDERFLOW_LIMIT) - return; - /* - * It isn't necessary to hold all of the locks - * at the same time, but doing it this way makes - * us share a lot more code. - */ - lock_mnt_writers(); - /* - * vfsmount_lock is for mnt_flags. - */ - spin_lock(&vfsmount_lock); - /* - * If coalescing the per-cpu writer counts did not - * get us back to a positive writer count, we have - * a bug. - */ - if ((atomic_read(&mnt->__mnt_writers) < 0) && - !(mnt->mnt_flags & MNT_IMBALANCED_WRITE_COUNT)) { - WARN(1, KERN_DEBUG "leak detected on mount(%p) writers " - "count: %d\n", - mnt, atomic_read(&mnt->__mnt_writers)); - /* use the flag to keep the dmesg spam down */ - mnt->mnt_flags |= MNT_IMBALANCED_WRITE_COUNT; - } - spin_unlock(&vfsmount_lock); - unlock_mnt_writers(); -} - /** * mnt_drop_write - give up write access to a mount * @mnt: the mount on which to give up write access @@ -332,37 +274,9 @@ static void handle_write_count_underflow(struct vfsmount *mnt) */ void mnt_drop_write(struct vfsmount *mnt) { - int must_check_underflow = 0; - struct mnt_writer *cpu_writer; - - cpu_writer = &get_cpu_var(mnt_writers); - spin_lock(&cpu_writer->lock); - - use_cpu_writer_for_mount(cpu_writer, mnt); - if (cpu_writer->count > 0) { - cpu_writer->count--; - } else { - must_check_underflow = 1; - atomic_dec(&mnt->__mnt_writers); - } - - spin_unlock(&cpu_writer->lock); - /* - * Logically, we could call this each time, - * but the __mnt_writers cacheline tends to - * be cold, and makes this expensive. - */ - if (must_check_underflow) - handle_write_count_underflow(mnt); - /* - * This could be done right after the spinlock - * is taken because the spinlock keeps us on - * the cpu, and disables preemption. However, - * putting it here bounds the amount that - * __mnt_writers can underflow. Without it, - * we could theoretically wrap __mnt_writers. - */ - put_cpu_var(mnt_writers); + preempt_disable(); + dec_mnt_writers(mnt); + preempt_enable(); } EXPORT_SYMBOL_GPL(mnt_drop_write); @@ -370,24 +284,41 @@ static int mnt_make_readonly(struct vfsmount *mnt) { int ret = 0; - lock_mnt_writers(); - /* - * With all the locks held, this value is stable - */ - if (atomic_read(&mnt->__mnt_writers) > 0) { - ret = -EBUSY; - goto out; - } - /* - * nobody can do a successful mnt_want_write() with all - * of the counts in MNT_DENIED_WRITE and the locks held. - */ spin_lock(&vfsmount_lock); - if (!ret) + mnt->mnt_flags |= MNT_WRITE_HOLD; + /* + * After storing MNT_WRITE_HOLD, we'll read the counters. This store + * should be visible before we do. + */ + smp_mb(); + + /* + * With writers on hold, if this value is zero, then there are + * definitely no active writers (although held writers may subsequently + * increment the count, they'll have to wait, and decrement it after + * seeing MNT_READONLY). + * + * It is OK to have counter incremented on one CPU and decremented on + * another: the sum will add up correctly. The danger would be when we + * sum up each counter, if we read a counter before it is incremented, + * but then read another CPU's count which it has been subsequently + * decremented from -- we would see more decrements than we should. + * MNT_WRITE_HOLD protects against this scenario, because + * mnt_want_write first increments count, then smp_mb, then spins on + * MNT_WRITE_HOLD, so it can't be decremented by another CPU while + * we're counting up here. + */ + if (count_mnt_writers(mnt) > 0) + ret = -EBUSY; + else mnt->mnt_flags |= MNT_READONLY; + /* + * MNT_READONLY must become visible before ~MNT_WRITE_HOLD, so writers + * that become unheld will see MNT_READONLY. + */ + smp_wmb(); + mnt->mnt_flags &= ~MNT_WRITE_HOLD; spin_unlock(&vfsmount_lock); -out: - unlock_mnt_writers(); return ret; } @@ -410,6 +341,9 @@ void free_vfsmnt(struct vfsmount *mnt) { kfree(mnt->mnt_devname); mnt_free_id(mnt); +#ifdef CONFIG_SMP + free_percpu(mnt->mnt_writers); +#endif kmem_cache_free(mnt_cache, mnt); } @@ -604,38 +538,18 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root, static inline void __mntput(struct vfsmount *mnt) { - int cpu; struct super_block *sb = mnt->mnt_sb; - /* - * We don't have to hold all of the locks at the - * same time here because we know that we're the - * last reference to mnt and that no new writers - * can come in. - */ - for_each_possible_cpu(cpu) { - struct mnt_writer *cpu_writer = &per_cpu(mnt_writers, cpu); - spin_lock(&cpu_writer->lock); - if (cpu_writer->mnt != mnt) { - spin_unlock(&cpu_writer->lock); - continue; - } - atomic_add(cpu_writer->count, &mnt->__mnt_writers); - cpu_writer->count = 0; - /* - * Might as well do this so that no one - * ever sees the pointer and expects - * it to be valid. - */ - cpu_writer->mnt = NULL; - spin_unlock(&cpu_writer->lock); - } /* * This probably indicates that somebody messed * up a mnt_want/drop_write() pair. If this * happens, the filesystem was probably unable * to make r/w->r/o transitions. */ - WARN_ON(atomic_read(&mnt->__mnt_writers)); + /* + * atomic_dec_and_lock() used to deal with ->mnt_count decrements + * provides barriers, so count_mnt_writers() below is safe. AV + */ + WARN_ON(count_mnt_writers(mnt)); dput(mnt->mnt_root); free_vfsmnt(mnt); deactivate_super(sb); diff --git a/include/linux/mount.h b/include/linux/mount.h index 51f55f903aff..ac49c1f8e5c0 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -30,7 +30,7 @@ struct mnt_namespace; #define MNT_STRICTATIME 0x80 #define MNT_SHRINKABLE 0x100 -#define MNT_IMBALANCED_WRITE_COUNT 0x200 /* just for debugging */ +#define MNT_WRITE_HOLD 0x200 #define MNT_SHARED 0x1000 /* if the vfsmount is a shared mount */ #define MNT_UNBINDABLE 0x2000 /* if the vfsmount is a unbindable mount */ @@ -65,13 +65,22 @@ struct vfsmount { int mnt_expiry_mark; /* true if marked for expiry */ int mnt_pinned; int mnt_ghosts; - /* - * This value is not stable unless all of the mnt_writers[] spinlocks - * are held, and all mnt_writer[]s on this mount have 0 as their ->count - */ - atomic_t __mnt_writers; +#ifdef CONFIG_SMP + int *mnt_writers; +#else + int mnt_writers; +#endif }; +static inline int *get_mnt_writers_ptr(struct vfsmount *mnt) +{ +#ifdef CONFIG_SMP + return mnt->mnt_writers; +#else + return &mnt->mnt_writers; +#endif +} + static inline struct vfsmount *mntget(struct vfsmount *mnt) { if (mnt) From 96029c4e09ccbd73a6d0ed2b29e80bf2586ad7ef Mon Sep 17 00:00:00 2001 From: "npiggin@suse.de" Date: Sun, 26 Apr 2009 20:25:55 +1000 Subject: [PATCH 18/87] fs: introduce mnt_clone_write This patch speeds up lmbench lat_mmap test by about another 2% after the first patch. Before: avg = 462.286 std = 5.46106 After: avg = 453.12 std = 9.58257 (50 runs of each, stddev gives a reasonable confidence) It does this by introducing mnt_clone_write, which avoids some heavyweight operations of mnt_want_write if called on a vfsmount which we know already has a write count; and mnt_want_write_file, which can call mnt_clone_write if the file is open for write. After these two patches, mnt_want_write and mnt_drop_write go from 7% on the profile down to 1.3% (including mnt_clone_write). [AV: mnt_want_write_file() should take file alone and derive mnt from it; not only all callers have that form, but that's the only mnt about which we know that it's already held for write if file is opened for write] Cc: Dave Hansen Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/file_table.c | 2 +- fs/inode.c | 2 +- fs/namespace.c | 40 ++++++++++++++++++++++++++++++++++++++++ fs/open.c | 4 ++-- fs/xattr.c | 4 ++-- include/linux/mount.h | 4 ++++ 6 files changed, 50 insertions(+), 6 deletions(-) diff --git a/fs/file_table.c b/fs/file_table.c index 54018fe48840..3d66dbcebef6 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -214,7 +214,7 @@ int init_file(struct file *file, struct vfsmount *mnt, struct dentry *dentry, */ if ((mode & FMODE_WRITE) && !special_file(dentry->d_inode->i_mode)) { file_take_write(file); - error = mnt_want_write(mnt); + error = mnt_clone_write(mnt); WARN_ON(error); } return error; diff --git a/fs/inode.c b/fs/inode.c index ca337014ae29..a88baebf77cf 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1422,7 +1422,7 @@ void file_update_time(struct file *file) if (IS_NOCMTIME(inode)) return; - err = mnt_want_write(file->f_path.mnt); + err = mnt_want_write_file(file); if (err) return; diff --git a/fs/namespace.c b/fs/namespace.c index 22ae06ad751d..120b8a6b99ed 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -264,6 +264,46 @@ out: } EXPORT_SYMBOL_GPL(mnt_want_write); +/** + * mnt_clone_write - get write access to a mount + * @mnt: the mount on which to take a write + * + * This is effectively like mnt_want_write, except + * it must only be used to take an extra write reference + * on a mountpoint that we already know has a write reference + * on it. This allows some optimisation. + * + * After finished, mnt_drop_write must be called as usual to + * drop the reference. + */ +int mnt_clone_write(struct vfsmount *mnt) +{ + /* superblock may be r/o */ + if (__mnt_is_readonly(mnt)) + return -EROFS; + preempt_disable(); + inc_mnt_writers(mnt); + preempt_enable(); + return 0; +} +EXPORT_SYMBOL_GPL(mnt_clone_write); + +/** + * mnt_want_write_file - get write access to a file's mount + * @file: the file who's mount on which to take a write + * + * This is like mnt_want_write, but it takes a file and can + * do some optimisations if the file is open for write already + */ +int mnt_want_write_file(struct file *file) +{ + if (!(file->f_mode & FMODE_WRITE)) + return mnt_want_write(file->f_path.mnt); + else + return mnt_clone_write(file->f_path.mnt); +} +EXPORT_SYMBOL_GPL(mnt_want_write_file); + /** * mnt_drop_write - give up write access to a mount * @mnt: the mount on which to give up write access diff --git a/fs/open.c b/fs/open.c index bdfbf03615a4..7200e23d9258 100644 --- a/fs/open.c +++ b/fs/open.c @@ -612,7 +612,7 @@ SYSCALL_DEFINE2(fchmod, unsigned int, fd, mode_t, mode) audit_inode(NULL, dentry); - err = mnt_want_write(file->f_path.mnt); + err = mnt_want_write_file(file); if (err) goto out_putf; mutex_lock(&inode->i_mutex); @@ -761,7 +761,7 @@ SYSCALL_DEFINE3(fchown, unsigned int, fd, uid_t, user, gid_t, group) if (!file) goto out; - error = mnt_want_write(file->f_path.mnt); + error = mnt_want_write_file(file); if (error) goto out_fput; dentry = file->f_path.dentry; diff --git a/fs/xattr.c b/fs/xattr.c index d51b8f9db921..1c3d0af59ddf 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -297,7 +297,7 @@ SYSCALL_DEFINE5(fsetxattr, int, fd, const char __user *, name, return error; dentry = f->f_path.dentry; audit_inode(NULL, dentry); - error = mnt_want_write(f->f_path.mnt); + error = mnt_want_write_file(f); if (!error) { error = setxattr(dentry, name, value, size, flags); mnt_drop_write(f->f_path.mnt); @@ -524,7 +524,7 @@ SYSCALL_DEFINE2(fremovexattr, int, fd, const char __user *, name) return error; dentry = f->f_path.dentry; audit_inode(NULL, dentry); - error = mnt_want_write(f->f_path.mnt); + error = mnt_want_write_file(f); if (!error) { error = removexattr(dentry, name); mnt_drop_write(f->f_path.mnt); diff --git a/include/linux/mount.h b/include/linux/mount.h index ac49c1f8e5c0..5d5275364867 100644 --- a/include/linux/mount.h +++ b/include/linux/mount.h @@ -88,7 +88,11 @@ static inline struct vfsmount *mntget(struct vfsmount *mnt) return mnt; } +struct file; /* forward dec */ + extern int mnt_want_write(struct vfsmount *mnt); +extern int mnt_want_write_file(struct file *file); +extern int mnt_clone_write(struct vfsmount *mnt); extern void mnt_drop_write(struct vfsmount *mnt); extern void mntput_no_expire(struct vfsmount *mnt); extern void mnt_pin(struct vfsmount *mnt); From 864d7c4c068f23642efe91b33be3a84afe5f71e0 Mon Sep 17 00:00:00 2001 From: "npiggin@suse.de" Date: Sun, 26 Apr 2009 20:25:56 +1000 Subject: [PATCH 19/87] fs: move mark_files_ro into file_table.c This function walks the s_files lock, and operates primarily on the files in a superblock, so it better belongs here (eg. see also fs_may_remount_ro). [AV: ... and it shouldn't be static after that move] Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/file_table.c | 38 ++++++++++++++++++++++++++++++++++++++ fs/internal.h | 5 +++++ fs/super.c | 39 --------------------------------------- 3 files changed, 43 insertions(+), 39 deletions(-) diff --git a/fs/file_table.c b/fs/file_table.c index 3d66dbcebef6..334ce39881f8 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -399,6 +399,44 @@ too_bad: return 0; } +/** + * mark_files_ro - mark all files read-only + * @sb: superblock in question + * + * All files are marked read-only. We don't care about pending + * delete files so this should be used in 'force' mode only. + */ +void mark_files_ro(struct super_block *sb) +{ + struct file *f; + +retry: + file_list_lock(); + list_for_each_entry(f, &sb->s_files, f_u.fu_list) { + struct vfsmount *mnt; + if (!S_ISREG(f->f_path.dentry->d_inode->i_mode)) + continue; + if (!file_count(f)) + continue; + if (!(f->f_mode & FMODE_WRITE)) + continue; + f->f_mode &= ~FMODE_WRITE; + if (file_check_writeable(f) != 0) + continue; + file_release_write(f); + mnt = mntget(f->f_path.mnt); + file_list_unlock(); + /* + * This can sleep, so we can't hold + * the file_list_lock() spinlock. + */ + mnt_drop_write(mnt); + mntput(mnt); + goto retry; + } + file_list_unlock(); +} + void __init files_init(unsigned long mempages) { int n; diff --git a/fs/internal.h b/fs/internal.h index b4dac4fb6b61..6d4ef208ef65 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -66,3 +66,8 @@ extern void __init mnt_init(void); * fs_struct.c */ extern void chroot_fs_refs(struct path *, struct path *); + +/* + * file_table.c + */ +extern void mark_files_ro(struct super_block *); diff --git a/fs/super.c b/fs/super.c index 1943fdf655fa..c170551c23fe 100644 --- a/fs/super.c +++ b/fs/super.c @@ -615,45 +615,6 @@ out: return err; } -/** - * mark_files_ro - mark all files read-only - * @sb: superblock in question - * - * All files are marked read-only. We don't care about pending - * delete files so this should be used in 'force' mode only. - */ - -static void mark_files_ro(struct super_block *sb) -{ - struct file *f; - -retry: - file_list_lock(); - list_for_each_entry(f, &sb->s_files, f_u.fu_list) { - struct vfsmount *mnt; - if (!S_ISREG(f->f_path.dentry->d_inode->i_mode)) - continue; - if (!file_count(f)) - continue; - if (!(f->f_mode & FMODE_WRITE)) - continue; - f->f_mode &= ~FMODE_WRITE; - if (file_check_writeable(f) != 0) - continue; - file_release_write(f); - mnt = mntget(f->f_path.mnt); - file_list_unlock(); - /* - * This can sleep, so we can't hold - * the file_list_lock() spinlock. - */ - mnt_drop_write(mnt); - mntput(mnt); - goto retry; - } - file_list_unlock(); -} - /** * do_remount_sb - asks filesystem to change mount options. * @sb: superblock in question From 876a9f76abbcb775f8d21cbc99fa161f9e5937f1 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 28 Apr 2009 18:05:55 +0200 Subject: [PATCH 20/87] remove s_async_list Remove the unused s_async_list in the superblock, a leftover of the broken async inode deletion code that leaked into mainline. Having this in the middle of the sync/unmount path is not helpful for the following cleanups. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/super.c | 8 -------- include/linux/fs.h | 5 ----- 2 files changed, 13 deletions(-) diff --git a/fs/super.c b/fs/super.c index c170551c23fe..3d9f117dd2a3 100644 --- a/fs/super.c +++ b/fs/super.c @@ -38,7 +38,6 @@ #include #include #include -#include #include #include "internal.h" @@ -72,7 +71,6 @@ static struct super_block *alloc_super(struct file_system_type *type) INIT_HLIST_HEAD(&s->s_anon); INIT_LIST_HEAD(&s->s_inodes); INIT_LIST_HEAD(&s->s_dentry_lru); - INIT_LIST_HEAD(&s->s_async_list); init_rwsem(&s->s_umount); mutex_init(&s->s_lock); lockdep_set_class(&s->s_umount, &type->s_umount_key); @@ -342,11 +340,6 @@ void generic_shutdown_super(struct super_block *sb) lock_super(sb); sb->s_flags &= ~MS_ACTIVE; - /* - * wait for asynchronous fs operations to finish before going further - */ - async_synchronize_full_domain(&sb->s_async_list); - /* bad name - it should be evict_inodes() */ invalidate_inodes(sb); lock_kernel(); @@ -517,7 +510,6 @@ restart: sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); - async_synchronize_full_domain(&sb->s_async_list); if (sb->s_root && (wait || sb->s_dirt)) sb->s_op->sync_fs(sb, wait); up_read(&sb->s_umount); diff --git a/include/linux/fs.h b/include/linux/fs.h index 03fb2102b8f3..36bcff7036ef 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1372,11 +1372,6 @@ struct super_block { * generic_show_options() */ char *s_options; - - /* - * storage for asynchronous operations - */ - struct list_head s_async_list; }; extern struct timespec current_fs_time(struct super_block *sb); From 5a3e5cb8e08bd876e2542c1451c9a93dab1b0e39 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:48 +0200 Subject: [PATCH 21/87] vfs: Fix sys_sync() and fsync_super() reliability (version 4) So far, do_sync() called: sync_inodes(0); sync_supers(); sync_filesystems(0); sync_filesystems(1); sync_inodes(1); This ordering makes it kind of hard for filesystems as sync_inodes(0) need not submit all the IO (for example it skips inodes with I_SYNC set) so e.g. forcing transaction to disk in ->sync_fs() is not really enough. Therefore sys_sync has not been completely reliable on some filesystems (ext3, ext4, reiserfs, ocfs2 and others are hit by this) when racing e.g. with background writeback. A similar problem hits also other filesystems (e.g. ext2) because of write_supers() being called before the sync_inodes(1). Change the ordering of calls in do_sync() - this requires a new function sync_blockdevs() to preserve the property that block devices are always synced after write_super() / sync_fs() call. The same issue is fixed in __fsync_super() function used on umount / remount read-only. [AV: build fixes] Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/internal.h | 9 +++++++++ fs/super.c | 29 ++++++++++++++++++++++++++++- fs/sync.c | 4 +++- 3 files changed, 40 insertions(+), 2 deletions(-) diff --git a/fs/internal.h b/fs/internal.h index 6d4ef208ef65..343a537ab809 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -71,3 +71,12 @@ extern void chroot_fs_refs(struct path *, struct path *); * file_table.c */ extern void mark_files_ro(struct super_block *); + +/* + * super.c + */ +#ifdef CONFIG_BLOCK +extern void sync_blockdevs(void); +#else +static inline void sync_blockdevs(void) { } +#endif diff --git a/fs/super.c b/fs/super.c index 3d9f117dd2a3..18d159dc1e40 100644 --- a/fs/super.c +++ b/fs/super.c @@ -293,6 +293,7 @@ void __fsync_super(struct super_block *sb) { sync_inodes_sb(sb, 0); vfs_dq_sync(sb); + sync_inodes_sb(sb, 1); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) sb->s_op->write_super(sb); @@ -300,7 +301,6 @@ void __fsync_super(struct super_block *sb) if (sb->s_op->sync_fs) sb->s_op->sync_fs(sb, 1); sync_blockdev(sb->s_bdev); - sync_inodes_sb(sb, 1); } /* @@ -522,6 +522,33 @@ restart: mutex_unlock(&mutex); } +#ifdef CONFIG_BLOCK +/* + * Sync all block devices underlying some superblock + */ +void sync_blockdevs(void) +{ + struct super_block *sb; + + spin_lock(&sb_lock); +restart: + list_for_each_entry(sb, &super_blocks, s_list) { + if (!sb->s_bdev) + continue; + sb->s_count++; + spin_unlock(&sb_lock); + down_read(&sb->s_umount); + if (sb->s_root) + sync_blockdev(sb->s_bdev); + up_read(&sb->s_umount); + spin_lock(&sb_lock); + if (__put_super_and_need_restart(sb)) + goto restart; + } + spin_unlock(&sb_lock); +} +#endif + /** * get_super - get the superblock of a device * @bdev: device to get the superblock for diff --git a/fs/sync.c b/fs/sync.c index 7abc65fbf21d..631fd5aece78 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -13,6 +13,7 @@ #include #include #include +#include "internal.h" #define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \ SYNC_FILE_RANGE_WAIT_AFTER) @@ -26,10 +27,11 @@ static void do_sync(unsigned long wait) wakeup_pdflush(0); sync_inodes(0); /* All mappings, inodes and their blockdevs */ vfs_dq_sync(NULL); + sync_inodes(wait); /* Mappings, inodes and blockdevs, again. */ sync_supers(); /* Write the superblocks */ sync_filesystems(0); /* Start syncing the filesystems */ sync_filesystems(wait); /* Waitingly sync the filesystems */ - sync_inodes(wait); /* Mappings, inodes and blockdevs, again. */ + sync_blockdevs(); if (!wait) printk("Emergency Sync complete\n"); if (unlikely(laptop_mode)) From bfe881255c74800147523b59c85328a1a826ba21 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:49 +0200 Subject: [PATCH 22/87] vfs: Call ->sync_fs() even if s_dirt is 0 (version 4) sync_filesystems() has a condition that if wait == 0 and s_dirt == 0, then ->sync_fs() isn't called. This does not really make much sence since s_dirt is generally used by a filesystem to mean that ->write_super() needs to be called. But ->sync_fs() does different things. I even suspect that some filesystems (btrfs?) sets s_dirt just to fool this logic. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/super.c b/fs/super.c index 18d159dc1e40..fae91ba38e48 100644 --- a/fs/super.c +++ b/fs/super.c @@ -510,7 +510,7 @@ restart: sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); - if (sb->s_root && (wait || sb->s_dirt)) + if (sb->s_root) sb->s_op->sync_fs(sb, wait); up_read(&sb->s_umount); /* restart only when sb is no longer on the list */ From 429479f031322a0cc5c921ffb2321a51718dc875 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:50 +0200 Subject: [PATCH 23/87] vfs: Make __fsync_super() a static function (version 4) __fsync_super() does the same thing as fsync_super(). So change the only caller to use fsync_super() and make __fsync_super() static. This removes unnecessarily duplicated call to sync_blockdev() and prepares ground for the changes to __fsync_super() in the following patches. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/block_dev.c | 2 +- fs/super.c | 7 +++---- include/linux/fs.h | 1 - 3 files changed, 4 insertions(+), 6 deletions(-) diff --git a/fs/block_dev.c b/fs/block_dev.c index 931f6b8c4b2f..fe47f7227618 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -241,7 +241,7 @@ struct super_block *freeze_bdev(struct block_device *bdev) sb->s_frozen = SB_FREEZE_WRITE; smp_wmb(); - __fsync_super(sb); + fsync_super(sb); sb->s_frozen = SB_FREEZE_TRANS; smp_wmb(); diff --git a/fs/super.c b/fs/super.c index fae91ba38e48..8dbe1ead9ddd 100644 --- a/fs/super.c +++ b/fs/super.c @@ -289,7 +289,7 @@ EXPORT_SYMBOL(unlock_super); * device. Takes the superblock lock. Requires a second blkdev * flush by the caller to complete the operation. */ -void __fsync_super(struct super_block *sb) +static int __fsync_super(struct super_block *sb) { sync_inodes_sb(sb, 0); vfs_dq_sync(sb); @@ -300,7 +300,7 @@ void __fsync_super(struct super_block *sb) unlock_super(sb); if (sb->s_op->sync_fs) sb->s_op->sync_fs(sb, 1); - sync_blockdev(sb->s_bdev); + return sync_blockdev(sb->s_bdev); } /* @@ -310,8 +310,7 @@ void __fsync_super(struct super_block *sb) */ int fsync_super(struct super_block *sb) { - __fsync_super(sb); - return sync_blockdev(sb->s_bdev); + return __fsync_super(sb); } EXPORT_SYMBOL_GPL(fsync_super); diff --git a/include/linux/fs.h b/include/linux/fs.h index 36bcff7036ef..41a9907f342e 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2078,7 +2078,6 @@ extern int filemap_fdatawrite_range(struct address_space *mapping, extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync); extern void sync_supers(void); extern void sync_filesystems(int wait); -extern void __fsync_super(struct super_block *sb); extern void emergency_sync(void); extern void emergency_remount(void); extern int do_remount_sb(struct super_block *sb, int flags, From 5cee5815d1564bbbd505fea86f4550f1efdb5cd0 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:51 +0200 Subject: [PATCH 24/87] vfs: Make sys_sync() use fsync_super() (version 4) It is unnecessarily fragile to have two places (fsync_super() and do_sync()) doing data integrity sync of the filesystem. Alter __fsync_super() to accommodate needs of both callers and use it. So after this patch __fsync_super() is the only place where we gather all the calls needed to properly send all data on a filesystem to disk. Nice bonus is that we get a complete livelock avoidance and write_supers() is now only used for periodic writeback of superblocks. sync_blockdevs() introduced a couple of patches ago is gone now. [build fixes folded] Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/block_dev.c | 15 +++++--- fs/fs-writeback.c | 49 -------------------------- fs/internal.h | 16 ++++----- fs/super.c | 72 +++++++++++++-------------------------- fs/sync.c | 33 ++++++------------ include/linux/fs.h | 2 +- include/linux/writeback.h | 1 - 7 files changed, 52 insertions(+), 136 deletions(-) diff --git a/fs/block_dev.c b/fs/block_dev.c index fe47f7227618..4b6a3b9d01ef 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -176,17 +176,22 @@ blkdev_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, iov, offset, nr_segs, blkdev_get_blocks, NULL); } +int __sync_blockdev(struct block_device *bdev, int wait) +{ + if (!bdev) + return 0; + if (!wait) + return filemap_flush(bdev->bd_inode->i_mapping); + return filemap_write_and_wait(bdev->bd_inode->i_mapping); +} + /* * Write out and wait upon all the dirty data associated with a block * device via its mapping. Does not take the superblock lock. */ int sync_blockdev(struct block_device *bdev) { - int ret = 0; - - if (bdev) - ret = filemap_write_and_wait(bdev->bd_inode->i_mapping); - return ret; + return __sync_blockdev(bdev, 1); } EXPORT_SYMBOL(sync_blockdev); diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 91013ff7dd53..e0fb2e789598 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -678,55 +678,6 @@ void sync_inodes_sb(struct super_block *sb, int wait) sync_sb_inodes(sb, &wbc); } -/** - * sync_inodes - writes all inodes to disk - * @wait: wait for completion - * - * sync_inodes() goes through each super block's dirty inode list, writes the - * inodes out, waits on the writeout and puts the inodes back on the normal - * list. - * - * This is for sys_sync(). fsync_dev() uses the same algorithm. The subtle - * part of the sync functions is that the blockdev "superblock" is processed - * last. This is because the write_inode() function of a typical fs will - * perform no I/O, but will mark buffers in the blockdev mapping as dirty. - * What we want to do is to perform all that dirtying first, and then write - * back all those inode blocks via the blockdev mapping in one sweep. So the - * additional (somewhat redundant) sync_blockdev() calls here are to make - * sure that really happens. Because if we call sync_inodes_sb(wait=1) with - * outstanding dirty inodes, the writeback goes block-at-a-time within the - * filesystem's write_inode(). This is extremely slow. - */ -static void __sync_inodes(int wait) -{ - struct super_block *sb; - - spin_lock(&sb_lock); -restart: - list_for_each_entry(sb, &super_blocks, s_list) { - sb->s_count++; - spin_unlock(&sb_lock); - down_read(&sb->s_umount); - if (sb->s_root) { - sync_inodes_sb(sb, wait); - sync_blockdev(sb->s_bdev); - } - up_read(&sb->s_umount); - spin_lock(&sb_lock); - if (__put_super_and_need_restart(sb)) - goto restart; - } - spin_unlock(&sb_lock); -} - -void sync_inodes(int wait) -{ - __sync_inodes(0); - - if (wait) - __sync_inodes(1); -} - /** * write_inode_now - write an inode to disk * @inode: inode to write to disk diff --git a/fs/internal.h b/fs/internal.h index 343a537ab809..dbec3cc28338 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -25,6 +25,8 @@ static inline int sb_is_blkdev_sb(struct super_block *sb) return sb == blockdev_superblock; } +extern int __sync_blockdev(struct block_device *bdev, int wait); + #else static inline void bdev_cache_init(void) { @@ -34,6 +36,11 @@ static inline int sb_is_blkdev_sb(struct super_block *sb) { return 0; } + +static inline int __sync_blockdev(struct block_device *bdev, int wait) +{ + return 0; +} #endif /* @@ -71,12 +78,3 @@ extern void chroot_fs_refs(struct path *, struct path *); * file_table.c */ extern void mark_files_ro(struct super_block *); - -/* - * super.c - */ -#ifdef CONFIG_BLOCK -extern void sync_blockdevs(void); -#else -static inline void sync_blockdevs(void) { } -#endif diff --git a/fs/super.c b/fs/super.c index 8dbe1ead9ddd..c8ce5ed04249 100644 --- a/fs/super.c +++ b/fs/super.c @@ -284,23 +284,23 @@ EXPORT_SYMBOL(lock_super); EXPORT_SYMBOL(unlock_super); /* - * Write out and wait upon all dirty data associated with this - * superblock. Filesystem data as well as the underlying block - * device. Takes the superblock lock. Requires a second blkdev - * flush by the caller to complete the operation. + * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0) + * just dirties buffers with inodes so we have to submit IO for these buffers + * via __sync_blockdev(). This also speeds up the wait == 1 case since in that + * case write_inode() functions do sync_dirty_buffer() and thus effectively + * write one block at a time. */ -static int __fsync_super(struct super_block *sb) +static int __fsync_super(struct super_block *sb, int wait) { - sync_inodes_sb(sb, 0); vfs_dq_sync(sb); - sync_inodes_sb(sb, 1); + sync_inodes_sb(sb, wait); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) sb->s_op->write_super(sb); unlock_super(sb); if (sb->s_op->sync_fs) - sb->s_op->sync_fs(sb, 1); - return sync_blockdev(sb->s_bdev); + sb->s_op->sync_fs(sb, wait); + return __sync_blockdev(sb->s_bdev, wait); } /* @@ -310,7 +310,12 @@ static int __fsync_super(struct super_block *sb) */ int fsync_super(struct super_block *sb) { - return __fsync_super(sb); + int ret; + + ret = __fsync_super(sb, 0); + if (ret < 0) + return ret; + return __fsync_super(sb, 1); } EXPORT_SYMBOL_GPL(fsync_super); @@ -469,20 +474,18 @@ restart: } /* - * Call the ->sync_fs super_op against all filesystems which are r/w and - * which implement it. + * Sync all the data for all the filesystems (called by sys_sync() and + * emergency sync) * * This operation is careful to avoid the livelock which could easily happen - * if two or more filesystems are being continuously dirtied. s_need_sync_fs + * if two or more filesystems are being continuously dirtied. s_need_sync * is used only here. We set it against all filesystems and then clear it as * we sync them. So redirtied filesystems are skipped. * * But if process A is currently running sync_filesystems and then process B - * calls sync_filesystems as well, process B will set all the s_need_sync_fs + * calls sync_filesystems as well, process B will set all the s_need_sync * flags again, which will cause process A to resync everything. Fix that with * a local mutex. - * - * (Fabian) Avoid sync_fs with clean fs & wait mode 0 */ void sync_filesystems(int wait) { @@ -492,25 +495,23 @@ void sync_filesystems(int wait) mutex_lock(&mutex); /* Could be down_interruptible */ spin_lock(&sb_lock); list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_op->sync_fs) - continue; if (sb->s_flags & MS_RDONLY) continue; - sb->s_need_sync_fs = 1; + sb->s_need_sync = 1; } restart: list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_need_sync_fs) + if (!sb->s_need_sync) continue; - sb->s_need_sync_fs = 0; + sb->s_need_sync = 0; if (sb->s_flags & MS_RDONLY) continue; /* hm. Was remounted r/o meanwhile */ sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); if (sb->s_root) - sb->s_op->sync_fs(sb, wait); + __fsync_super(sb, wait); up_read(&sb->s_umount); /* restart only when sb is no longer on the list */ spin_lock(&sb_lock); @@ -521,33 +522,6 @@ restart: mutex_unlock(&mutex); } -#ifdef CONFIG_BLOCK -/* - * Sync all block devices underlying some superblock - */ -void sync_blockdevs(void) -{ - struct super_block *sb; - - spin_lock(&sb_lock); -restart: - list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_bdev) - continue; - sb->s_count++; - spin_unlock(&sb_lock); - down_read(&sb->s_umount); - if (sb->s_root) - sync_blockdev(sb->s_bdev); - up_read(&sb->s_umount); - spin_lock(&sb_lock); - if (__put_super_and_need_restart(sb)) - goto restart; - } - spin_unlock(&sb_lock); -} -#endif - /** * get_super - get the superblock of a device * @bdev: device to get the superblock for diff --git a/fs/sync.c b/fs/sync.c index 631fd5aece78..be0798cc33d7 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -18,35 +18,24 @@ #define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \ SYNC_FILE_RANGE_WAIT_AFTER) -/* - * sync everything. Start out by waking pdflush, because that writes back - * all queues in parallel. - */ -static void do_sync(unsigned long wait) -{ - wakeup_pdflush(0); - sync_inodes(0); /* All mappings, inodes and their blockdevs */ - vfs_dq_sync(NULL); - sync_inodes(wait); /* Mappings, inodes and blockdevs, again. */ - sync_supers(); /* Write the superblocks */ - sync_filesystems(0); /* Start syncing the filesystems */ - sync_filesystems(wait); /* Waitingly sync the filesystems */ - sync_blockdevs(); - if (!wait) - printk("Emergency Sync complete\n"); - if (unlikely(laptop_mode)) - laptop_sync_completion(); -} - SYSCALL_DEFINE0(sync) { - do_sync(1); + sync_filesystems(0); + sync_filesystems(1); + if (unlikely(laptop_mode)) + laptop_sync_completion(); return 0; } static void do_sync_work(struct work_struct *work) { - do_sync(0); + /* + * Sync twice to reduce the possibility we skipped some inodes / pages + * because they were temporarily locked + */ + sync_filesystems(0); + sync_filesystems(0); + printk("Emergency Sync complete\n"); kfree(work); } diff --git a/include/linux/fs.h b/include/linux/fs.h index 41a9907f342e..f00df653cf2b 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1321,7 +1321,7 @@ struct super_block { struct rw_semaphore s_umount; struct mutex s_lock; int s_count; - int s_need_sync_fs; + int s_need_sync; atomic_t s_active; #ifdef CONFIG_SECURITY void *s_security; diff --git a/include/linux/writeback.h b/include/linux/writeback.h index 93445477f86a..3224820c8514 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -79,7 +79,6 @@ struct writeback_control { void writeback_inodes(struct writeback_control *wbc); int inode_wait(void *); void sync_inodes_sb(struct super_block *, int wait); -void sync_inodes(int wait); /* writeback.h requires fs.h; it, too, is not included from here. */ static inline void wait_on_inode(struct inode *inode) From c15c54f5f056ee4819da9fde59a5f2cd45445f23 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:52 +0200 Subject: [PATCH 25/87] vfs: Move syncing code from super.c to sync.c (version 4) Move sync_filesystems(), __fsync_super(), fsync_super() from super.c to sync.c where it fits better. [build fixes folded] Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/super.c | 85 ---------------------------------------------- fs/sync.c | 85 ++++++++++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 3 +- 3 files changed, 86 insertions(+), 87 deletions(-) diff --git a/fs/super.c b/fs/super.c index c8ce5ed04249..f822c74f25be 100644 --- a/fs/super.c +++ b/fs/super.c @@ -283,42 +283,6 @@ void unlock_super(struct super_block * sb) EXPORT_SYMBOL(lock_super); EXPORT_SYMBOL(unlock_super); -/* - * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0) - * just dirties buffers with inodes so we have to submit IO for these buffers - * via __sync_blockdev(). This also speeds up the wait == 1 case since in that - * case write_inode() functions do sync_dirty_buffer() and thus effectively - * write one block at a time. - */ -static int __fsync_super(struct super_block *sb, int wait) -{ - vfs_dq_sync(sb); - sync_inodes_sb(sb, wait); - lock_super(sb); - if (sb->s_dirt && sb->s_op->write_super) - sb->s_op->write_super(sb); - unlock_super(sb); - if (sb->s_op->sync_fs) - sb->s_op->sync_fs(sb, wait); - return __sync_blockdev(sb->s_bdev, wait); -} - -/* - * Write out and wait upon all dirty data associated with this - * superblock. Filesystem data as well as the underlying block - * device. Takes the superblock lock. - */ -int fsync_super(struct super_block *sb) -{ - int ret; - - ret = __fsync_super(sb, 0); - if (ret < 0) - return ret; - return __fsync_super(sb, 1); -} -EXPORT_SYMBOL_GPL(fsync_super); - /** * generic_shutdown_super - common helper for ->kill_sb() * @sb: superblock to kill @@ -473,55 +437,6 @@ restart: spin_unlock(&sb_lock); } -/* - * Sync all the data for all the filesystems (called by sys_sync() and - * emergency sync) - * - * This operation is careful to avoid the livelock which could easily happen - * if two or more filesystems are being continuously dirtied. s_need_sync - * is used only here. We set it against all filesystems and then clear it as - * we sync them. So redirtied filesystems are skipped. - * - * But if process A is currently running sync_filesystems and then process B - * calls sync_filesystems as well, process B will set all the s_need_sync - * flags again, which will cause process A to resync everything. Fix that with - * a local mutex. - */ -void sync_filesystems(int wait) -{ - struct super_block *sb; - static DEFINE_MUTEX(mutex); - - mutex_lock(&mutex); /* Could be down_interruptible */ - spin_lock(&sb_lock); - list_for_each_entry(sb, &super_blocks, s_list) { - if (sb->s_flags & MS_RDONLY) - continue; - sb->s_need_sync = 1; - } - -restart: - list_for_each_entry(sb, &super_blocks, s_list) { - if (!sb->s_need_sync) - continue; - sb->s_need_sync = 0; - if (sb->s_flags & MS_RDONLY) - continue; /* hm. Was remounted r/o meanwhile */ - sb->s_count++; - spin_unlock(&sb_lock); - down_read(&sb->s_umount); - if (sb->s_root) - __fsync_super(sb, wait); - up_read(&sb->s_umount); - /* restart only when sb is no longer on the list */ - spin_lock(&sb_lock); - if (__put_super_and_need_restart(sb)) - goto restart; - } - spin_unlock(&sb_lock); - mutex_unlock(&mutex); -} - /** * get_super - get the superblock of a device * @bdev: device to get the superblock for diff --git a/fs/sync.c b/fs/sync.c index be0798cc33d7..d5fa7b79982e 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -18,6 +18,91 @@ #define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \ SYNC_FILE_RANGE_WAIT_AFTER) +/* + * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0) + * just dirties buffers with inodes so we have to submit IO for these buffers + * via __sync_blockdev(). This also speeds up the wait == 1 case since in that + * case write_inode() functions do sync_dirty_buffer() and thus effectively + * write one block at a time. + */ +static int __fsync_super(struct super_block *sb, int wait) +{ + vfs_dq_sync(sb); + sync_inodes_sb(sb, wait); + lock_super(sb); + if (sb->s_dirt && sb->s_op->write_super) + sb->s_op->write_super(sb); + unlock_super(sb); + if (sb->s_op->sync_fs) + sb->s_op->sync_fs(sb, wait); + return __sync_blockdev(sb->s_bdev, wait); +} + +/* + * Write out and wait upon all dirty data associated with this + * superblock. Filesystem data as well as the underlying block + * device. Takes the superblock lock. + */ +int fsync_super(struct super_block *sb) +{ + int ret; + + ret = __fsync_super(sb, 0); + if (ret < 0) + return ret; + return __fsync_super(sb, 1); +} +EXPORT_SYMBOL_GPL(fsync_super); + +/* + * Sync all the data for all the filesystems (called by sys_sync() and + * emergency sync) + * + * This operation is careful to avoid the livelock which could easily happen + * if two or more filesystems are being continuously dirtied. s_need_sync + * is used only here. We set it against all filesystems and then clear it as + * we sync them. So redirtied filesystems are skipped. + * + * But if process A is currently running sync_filesystems and then process B + * calls sync_filesystems as well, process B will set all the s_need_sync + * flags again, which will cause process A to resync everything. Fix that with + * a local mutex. + */ +static void sync_filesystems(int wait) +{ + struct super_block *sb; + static DEFINE_MUTEX(mutex); + + mutex_lock(&mutex); /* Could be down_interruptible */ + spin_lock(&sb_lock); + list_for_each_entry(sb, &super_blocks, s_list) { + if (sb->s_flags & MS_RDONLY) + continue; + sb->s_need_sync = 1; + } + +restart: + list_for_each_entry(sb, &super_blocks, s_list) { + if (!sb->s_need_sync) + continue; + sb->s_need_sync = 0; + if (sb->s_flags & MS_RDONLY) + continue; /* hm. Was remounted r/o meanwhile */ + sb->s_count++; + spin_unlock(&sb_lock); + down_read(&sb->s_umount); + if (sb->s_root) + __fsync_super(sb, wait); + up_read(&sb->s_umount); + /* restart only when sb is no longer on the list */ + spin_lock(&sb_lock); + if (__put_super_and_need_restart(sb)) + goto restart; + } + spin_unlock(&sb_lock); + mutex_unlock(&mutex); +} + SYSCALL_DEFINE0(sync) { sync_filesystems(0); diff --git a/include/linux/fs.h b/include/linux/fs.h index f00df653cf2b..d3f7159993cf 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1942,7 +1942,6 @@ extern struct super_block *freeze_bdev(struct block_device *); extern void emergency_thaw_all(void); extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); extern int fsync_bdev(struct block_device *); -extern int fsync_super(struct super_block *); extern int fsync_no_super(struct block_device *); #else static inline void bd_forget(struct inode *inode) {} @@ -1959,6 +1958,7 @@ static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb) return 0; } #endif +extern int fsync_super(struct super_block *); extern const struct file_operations def_blk_fops; extern const struct file_operations def_chr_fops; extern const struct file_operations bad_sock_fops; @@ -2077,7 +2077,6 @@ extern int filemap_fdatawrite_range(struct address_space *mapping, extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync); extern void sync_supers(void); -extern void sync_filesystems(int wait); extern void emergency_sync(void); extern void emergency_remount(void); extern int do_remount_sb(struct super_block *sb, int flags, From 60b0680fa236ac4e17ce31a50048c9d75f9ec831 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:53 +0200 Subject: [PATCH 26/87] vfs: Rename fsync_super() to sync_filesystem() (version 4) Rename the function so that it better describe what it really does. Also remove the unnecessary include of buffer_head.h. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/block_dev.c | 4 ++-- fs/cachefiles/interface.c | 2 +- fs/super.c | 5 ++--- fs/sync.c | 12 ++++++------ include/linux/fs.h | 2 +- 5 files changed, 12 insertions(+), 13 deletions(-) diff --git a/fs/block_dev.c b/fs/block_dev.c index 4b6a3b9d01ef..3a6d4fb2a329 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -204,7 +204,7 @@ int fsync_bdev(struct block_device *bdev) { struct super_block *sb = get_super(bdev); if (sb) { - int res = fsync_super(sb); + int res = sync_filesystem(sb); drop_super(sb); return res; } @@ -246,7 +246,7 @@ struct super_block *freeze_bdev(struct block_device *bdev) sb->s_frozen = SB_FREEZE_WRITE; smp_wmb(); - fsync_super(sb); + sync_filesystem(sb); sb->s_frozen = SB_FREEZE_TRANS; smp_wmb(); diff --git a/fs/cachefiles/interface.c b/fs/cachefiles/interface.c index 1e962348d111..dafd484d7bda 100644 --- a/fs/cachefiles/interface.c +++ b/fs/cachefiles/interface.c @@ -354,7 +354,7 @@ static void cachefiles_sync_cache(struct fscache_cache *_cache) /* make sure all pages pinned by operations on behalf of the netfs are * written to disc */ cachefiles_begin_secure(cache, &saved_cred); - ret = fsync_super(cache->mnt->mnt_sb); + ret = sync_filesystem(cache->mnt->mnt_sb); cachefiles_end_secure(cache, saved_cred); if (ret == -EIO) diff --git a/fs/super.c b/fs/super.c index f822c74f25be..721236e6177a 100644 --- a/fs/super.c +++ b/fs/super.c @@ -28,7 +28,6 @@ #include #include #include -#include /* for fsync_super() */ #include #include #include @@ -304,7 +303,7 @@ void generic_shutdown_super(struct super_block *sb) if (sb->s_root) { shrink_dcache_for_umount(sb); - fsync_super(sb); + sync_filesystem(sb); lock_super(sb); sb->s_flags &= ~MS_ACTIVE; @@ -543,7 +542,7 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) if (flags & MS_RDONLY) acct_auto_close(sb); shrink_dcache_sb(sb); - fsync_super(sb); + sync_filesystem(sb); /* If we are remounting RDONLY and current sb is read/write, make sure there are no rw files opened */ diff --git a/fs/sync.c b/fs/sync.c index d5fa7b79982e..8aa870a4d406 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -25,7 +25,7 @@ * case write_inode() functions do sync_dirty_buffer() and thus effectively * write one block at a time. */ -static int __fsync_super(struct super_block *sb, int wait) +static int __sync_filesystem(struct super_block *sb, int wait) { vfs_dq_sync(sb); sync_inodes_sb(sb, wait); @@ -43,16 +43,16 @@ static int __fsync_super(struct super_block *sb, int wait) * superblock. Filesystem data as well as the underlying block * device. Takes the superblock lock. */ -int fsync_super(struct super_block *sb) +int sync_filesystem(struct super_block *sb) { int ret; - ret = __fsync_super(sb, 0); + ret = __sync_filesystem(sb, 0); if (ret < 0) return ret; - return __fsync_super(sb, 1); + return __sync_filesystem(sb, 1); } -EXPORT_SYMBOL_GPL(fsync_super); +EXPORT_SYMBOL_GPL(sync_filesystem); /* * Sync all the data for all the filesystems (called by sys_sync() and @@ -92,7 +92,7 @@ restart: spin_unlock(&sb_lock); down_read(&sb->s_umount); if (sb->s_root) - __fsync_super(sb, wait); + __sync_filesystem(sb, wait); up_read(&sb->s_umount); /* restart only when sb is no longer on the list */ spin_lock(&sb_lock); diff --git a/include/linux/fs.h b/include/linux/fs.h index d3f7159993cf..fb1822bed7c8 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1958,7 +1958,7 @@ static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb) return 0; } #endif -extern int fsync_super(struct super_block *); +extern int sync_filesystem(struct super_block *); extern const struct file_operations def_blk_fops; extern const struct file_operations def_chr_fops; extern const struct file_operations bad_sock_fops; From 850b201b087f5525a0a7278551c2bcd0423c3b26 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 16:43:54 +0200 Subject: [PATCH 27/87] quota: cleanup dquota sync functions (version 4) Currently the VFS calls vfs_dq_sync to sync out disk quotas for a given superblock. This is a small wrapper around sync_dquots which for the case of a non-NULL superblock is a small wrapper around quota_sync_sb. Just make quota_sync_sb global (rename it to sync_quota_sb) and call it directly. Also call it directly for those cases in quota.c that have a superblock and leave sync_dquots purely an iterator over sync_quota_sb and remove it's superblock argument. To make this nicer move the check for the lack of a quota_sync method from the callers into sync_quota_sb. [folded build fix from Alexander Beregalov ] Signed-off-by: Christoph Hellwig Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/quota/quota.c | 25 ++++++++++++++----------- fs/sync.c | 2 +- include/linux/quotaops.h | 11 +++-------- 3 files changed, 18 insertions(+), 20 deletions(-) diff --git a/fs/quota/quota.c b/fs/quota/quota.c index b7f5a468f076..95c5b42384b2 100644 --- a/fs/quota/quota.c +++ b/fs/quota/quota.c @@ -159,10 +159,14 @@ static int check_quotactl_valid(struct super_block *sb, int type, int cmd, return error; } -static void quota_sync_sb(struct super_block *sb, int type) +#ifdef CONFIG_QUOTA +void sync_quota_sb(struct super_block *sb, int type) { int cnt; + if (!sb->s_qcop->quota_sync) + return; + sb->s_qcop->quota_sync(sb, type); if (sb_dqopt(sb)->flags & DQUOT_QUOTA_SYS_FILE) @@ -191,17 +195,13 @@ static void quota_sync_sb(struct super_block *sb, int type) } mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); } +#endif -void sync_dquots(struct super_block *sb, int type) +static void sync_dquots(int type) { + struct super_block *sb; int cnt; - if (sb) { - if (sb->s_qcop->quota_sync) - quota_sync_sb(sb, type); - return; - } - spin_lock(&sb_lock); restart: list_for_each_entry(sb, &super_blocks, s_list) { @@ -222,8 +222,8 @@ restart: sb->s_count++; spin_unlock(&sb_lock); down_read(&sb->s_umount); - if (sb->s_root && sb->s_qcop->quota_sync) - quota_sync_sb(sb, type); + if (sb->s_root) + sync_quota_sb(sb, type); up_read(&sb->s_umount); spin_lock(&sb_lock); if (__put_super_and_need_restart(sb)) @@ -301,7 +301,10 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id, return sb->s_qcop->set_dqblk(sb, type, id, &idq); } case Q_SYNC: - sync_dquots(sb, type); + if (sb) + sync_quota_sb(sb, type); + else + sync_dquots(type); return 0; case Q_XQUOTAON: diff --git a/fs/sync.c b/fs/sync.c index 8aa870a4d406..d90ab7764555 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -27,7 +27,7 @@ */ static int __sync_filesystem(struct super_block *sb, int wait) { - vfs_dq_sync(sb); + sync_quota_sb(sb, -1); sync_inodes_sb(sb, wait); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 36353d95c8db..047310fa22fb 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -20,7 +20,7 @@ static inline struct quota_info *sb_dqopt(struct super_block *sb) /* * declaration of quota_function calls in kernel. */ -void sync_dquots(struct super_block *sb, int type); +void sync_quota_sb(struct super_block *sb, int type); int dquot_initialize(struct inode *inode, int type); int dquot_drop(struct inode *inode); @@ -253,12 +253,7 @@ static inline void vfs_dq_free_inode(struct inode *inode) inode->i_sb->dq_op->free_inode(inode, 1); } -/* The following two functions cannot be called inside a transaction */ -static inline void vfs_dq_sync(struct super_block *sb) -{ - sync_dquots(sb, -1); -} - +/* Cannot be called inside a transaction */ static inline int vfs_dq_off(struct super_block *sb, int remount) { int ret = -ENOSYS; @@ -334,7 +329,7 @@ static inline void vfs_dq_free_inode(struct inode *inode) { } -static inline void vfs_dq_sync(struct super_block *sb) +static inline void sync_quota_sb(struct super_block *sb, int type) { } From c3f8a40c1cd5591b882497d1d00d43d0e5bb4698 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 27 Apr 2009 16:43:55 +0200 Subject: [PATCH 28/87] quota: Introduce writeout_quota_sb() (version 4) Introduce this function which just writes all the quota structures but avoids all the syncing and cache pruning work to expose quota structures to userspace. Use this function from __sync_filesystem when wait == 0. Signed-off-by: Jan Kara Signed-off-by: Al Viro --- fs/sync.c | 6 +++++- include/linux/quotaops.h | 9 +++++++++ 2 files changed, 14 insertions(+), 1 deletion(-) diff --git a/fs/sync.c b/fs/sync.c index d90ab7764555..4487b5560dc8 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -27,7 +27,11 @@ */ static int __sync_filesystem(struct super_block *sb, int wait) { - sync_quota_sb(sb, -1); + /* Avoid doing twice syncing and cache pruning for quota sync */ + if (!wait) + writeout_quota_sb(sb, -1); + else + sync_quota_sb(sb, -1); sync_inodes_sb(sb, wait); lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 047310fa22fb..7bc457593684 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -21,6 +21,11 @@ static inline struct quota_info *sb_dqopt(struct super_block *sb) * declaration of quota_function calls in kernel. */ void sync_quota_sb(struct super_block *sb, int type); +static inline void writeout_quota_sb(struct super_block *sb, int type) +{ + if (sb->s_qcop->quota_sync) + sb->s_qcop->quota_sync(sb, type); +} int dquot_initialize(struct inode *inode, int type); int dquot_drop(struct inode *inode); @@ -333,6 +338,10 @@ static inline void sync_quota_sb(struct super_block *sb, int type) { } +static inline void writeout_quota_sb(struct super_block *sb, int type) +{ +} + static inline int vfs_dq_off(struct super_block *sb, int remount) { return 0; From 59d697b70285c348c01cfc2695c3469ba71d7539 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 09:46:41 -0400 Subject: [PATCH 29/87] btrfs: remove ->write_super and stop maintaining ->s_dirt Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/btrfs/inode.c | 7 ------- fs/btrfs/super.c | 8 -------- 2 files changed, 15 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 5b68330f8585..8612b3a09811 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2322,7 +2322,6 @@ err: btrfs_update_inode(trans, root, dir); btrfs_drop_nlink(inode); ret = btrfs_update_inode(trans, root, inode); - dir->i_sb->s_dirt = 1; out: return ret; } @@ -2806,7 +2805,6 @@ error: pending_del_nr); } btrfs_free_path(path); - inode->i_sb->s_dirt = 1; return ret; } @@ -3768,7 +3766,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, init_special_inode(inode, inode->i_mode, rdev); btrfs_update_inode(trans, root, inode); } - dir->i_sb->s_dirt = 1; btrfs_update_inode_block_group(trans, inode); btrfs_update_inode_block_group(trans, dir); out_unlock: @@ -3833,7 +3830,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } - dir->i_sb->s_dirt = 1; btrfs_update_inode_block_group(trans, inode); btrfs_update_inode_block_group(trans, dir); out_unlock: @@ -3880,7 +3876,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, if (err) drop_inode = 1; - dir->i_sb->s_dirt = 1; btrfs_update_inode_block_group(trans, dir); err = btrfs_update_inode(trans, root, inode); @@ -3962,7 +3957,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) d_instantiate(dentry, inode); drop_on_err = 0; - dir->i_sb->s_dirt = 1; btrfs_update_inode_block_group(trans, inode); btrfs_update_inode_block_group(trans, dir); @@ -4991,7 +4985,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } - dir->i_sb->s_dirt = 1; btrfs_update_inode_block_group(trans, inode); btrfs_update_inode_block_group(trans, dir); if (drop_inode) diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 708ac06b953b..52d84522c2c2 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -397,7 +397,6 @@ int btrfs_sync_fs(struct super_block *sb, int wait) if (sb->s_flags & MS_RDONLY) return 0; - sb->s_dirt = 0; if (!wait) { filemap_flush(root->fs_info->btree_inode->i_mapping); return 0; @@ -408,7 +407,6 @@ int btrfs_sync_fs(struct super_block *sb, int wait) trans = btrfs_start_transaction(root, 1); ret = btrfs_commit_transaction(trans, root); - sb->s_dirt = 0; return ret; } @@ -454,11 +452,6 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) return 0; } -static void btrfs_write_super(struct super_block *sb) -{ - sb->s_dirt = 0; -} - static int btrfs_test_super(struct super_block *s, void *data) { struct btrfs_fs_devices *test_fs_devices = data; @@ -689,7 +682,6 @@ static int btrfs_unfreeze(struct super_block *sb) static struct super_operations btrfs_super_ops = { .delete_inode = btrfs_delete_inode, .put_super = btrfs_put_super, - .write_super = btrfs_write_super, .sync_fs = btrfs_sync_fs, .show_options = btrfs_show_options, .write_inode = btrfs_write_inode, From ca41f7b918294c2a17780e057568413dcbfc6d49 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 09:46:42 -0400 Subject: [PATCH 30/87] ext3: remove ->write_super and stop maintaining ->s_dirt Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/ext3/balloc.c | 3 +-- fs/ext3/ialloc.c | 3 +-- fs/ext3/inode.c | 1 - fs/ext3/resize.c | 2 -- fs/ext3/super.c | 22 ---------------------- fs/ext3/xattr.c | 1 - 6 files changed, 2 insertions(+), 30 deletions(-) diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c index 225202db8974..27967f92e820 100644 --- a/fs/ext3/balloc.c +++ b/fs/ext3/balloc.c @@ -649,7 +649,7 @@ do_more: count = overflow; goto do_more; } - sb->s_dirt = 1; + error_return: brelse(bitmap_bh); ext3_std_error(sb, err); @@ -1708,7 +1708,6 @@ allocated: if (!fatal) fatal = err; - sb->s_dirt = 1; if (fatal) goto out; diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c index dd13d60d524b..b39991285136 100644 --- a/fs/ext3/ialloc.c +++ b/fs/ext3/ialloc.c @@ -181,7 +181,7 @@ void ext3_free_inode (handle_t *handle, struct inode * inode) err = ext3_journal_dirty_metadata(handle, bitmap_bh); if (!fatal) fatal = err; - sb->s_dirt = 1; + error_return: brelse(bitmap_bh); ext3_std_error(sb, fatal); @@ -537,7 +537,6 @@ got: percpu_counter_dec(&sbi->s_freeinodes_counter); if (S_ISDIR(mode)) percpu_counter_inc(&sbi->s_dirs_counter); - sb->s_dirt = 1; inode->i_uid = current_fsuid(); if (test_opt (sb, GRPID)) diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index fcfa24361856..b0248c6d5d4c 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -2960,7 +2960,6 @@ static int ext3_do_update_inode(handle_t *handle, ext3_update_dynamic_rev(sb); EXT3_SET_RO_COMPAT_FEATURE(sb, EXT3_FEATURE_RO_COMPAT_LARGE_FILE); - sb->s_dirt = 1; handle->h_sync = 1; err = ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh); diff --git a/fs/ext3/resize.c b/fs/ext3/resize.c index 78fdf3836370..8a0b26340b54 100644 --- a/fs/ext3/resize.c +++ b/fs/ext3/resize.c @@ -934,7 +934,6 @@ int ext3_group_add(struct super_block *sb, struct ext3_new_group_data *input) EXT3_INODES_PER_GROUP(sb)); ext3_journal_dirty_metadata(handle, sbi->s_sbh); - sb->s_dirt = 1; exit_journal: unlock_super(sb); @@ -1066,7 +1065,6 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, } es->s_blocks_count = cpu_to_le32(o_blocks_count + add); ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh); - sb->s_dirt = 1; unlock_super(sb); ext3_debug("freeing blocks %lu through "E3FSBLK"\n", o_blocks_count, o_blocks_count + add); diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 3c70d52afb10..1efd958687e9 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -67,7 +67,6 @@ static const char *ext3_decode_error(struct super_block * sb, int errno, static int ext3_remount (struct super_block * sb, int * flags, char * data); static int ext3_statfs (struct dentry * dentry, struct kstatfs * buf); static int ext3_unfreeze(struct super_block *sb); -static void ext3_write_super (struct super_block * sb); static int ext3_freeze(struct super_block *sb); /* @@ -761,7 +760,6 @@ static const struct super_operations ext3_sops = { .dirty_inode = ext3_dirty_inode, .delete_inode = ext3_delete_inode, .put_super = ext3_put_super, - .write_super = ext3_write_super, .sync_fs = ext3_sync_fs, .freeze_fs = ext3_freeze, .unfreeze_fs = ext3_unfreeze, @@ -1785,7 +1783,6 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent) #else es->s_flags |= cpu_to_le32(EXT2_FLAGS_SIGNED_HASH); #endif - sb->s_dirt = 1; } if (sbi->s_blocks_per_group > blocksize * 8) { @@ -2265,7 +2262,6 @@ static int ext3_load_journal(struct super_block *sb, if (journal_devnum && journal_devnum != le32_to_cpu(es->s_journal_dev)) { es->s_journal_dev = cpu_to_le32(journal_devnum); - sb->s_dirt = 1; /* Make sure we flush the recovery flag to disk. */ ext3_commit_super(sb, es, 1); @@ -2308,7 +2304,6 @@ static int ext3_create_journal(struct super_block * sb, EXT3_SET_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_HAS_JOURNAL); es->s_journal_inum = cpu_to_le32(journal_inum); - sb->s_dirt = 1; /* Make sure we flush the recovery flag to disk. */ ext3_commit_super(sb, es, 1); @@ -2354,7 +2349,6 @@ static void ext3_mark_recovery_complete(struct super_block * sb, if (EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER) && sb->s_flags & MS_RDONLY) { EXT3_CLEAR_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER); - sb->s_dirt = 0; ext3_commit_super(sb, es, 1); } unlock_super(sb); @@ -2413,29 +2407,14 @@ int ext3_force_commit(struct super_block *sb) return 0; journal = EXT3_SB(sb)->s_journal; - sb->s_dirt = 0; ret = ext3_journal_force_commit(journal); return ret; } -/* - * Ext3 always journals updates to the superblock itself, so we don't - * have to propagate any other updates to the superblock on disk at this - * point. (We can probably nuke this function altogether, and remove - * any mention to sb->s_dirt in all of fs/ext3; eventual cleanup...) - */ -static void ext3_write_super (struct super_block * sb) -{ - if (mutex_trylock(&sb->s_lock) != 0) - BUG(); - sb->s_dirt = 0; -} - static int ext3_sync_fs(struct super_block *sb, int wait) { tid_t target; - sb->s_dirt = 0; if (journal_start_commit(EXT3_SB(sb)->s_journal, &target)) { if (wait) log_wait_commit(EXT3_SB(sb)->s_journal, target); @@ -2451,7 +2430,6 @@ static int ext3_freeze(struct super_block *sb) { int error = 0; journal_t *journal; - sb->s_dirt = 0; if (!(sb->s_flags & MS_RDONLY)) { journal = EXT3_SB(sb)->s_journal; diff --git a/fs/ext3/xattr.c b/fs/ext3/xattr.c index 83b7be849bd5..545e37c4b91e 100644 --- a/fs/ext3/xattr.c +++ b/fs/ext3/xattr.c @@ -463,7 +463,6 @@ static void ext3_xattr_update_super_block(handle_t *handle, if (ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh) == 0) { EXT3_SET_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_EXT_ATTR); - sb->s_dirt = 1; ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh); } } From b7d245de25d1f0bb75a0d04194b647762b30d3db Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 09:46:43 -0400 Subject: [PATCH 31/87] gfs2: remove ->write_super and stop maintaining ->s_dirt Signed-off-by: Christoph Hellwig Acked-by: Steven Whitehouse Signed-off-by: Al Viro --- fs/gfs2/log.c | 2 -- fs/gfs2/super.c | 13 ------------- 2 files changed, 15 deletions(-) diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c index aa62cf5976e8..f2e449c595b4 100644 --- a/fs/gfs2/log.c +++ b/fs/gfs2/log.c @@ -764,7 +764,6 @@ void __gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl) } gfs2_log_unlock(sdp); - sdp->sd_vfs->s_dirt = 0; up_write(&sdp->sd_log_flush_lock); kfree(ai); @@ -823,7 +822,6 @@ void gfs2_log_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr) log_refund(sdp, tr); buf_lo_incore_commit(sdp, tr); - sdp->sd_vfs->s_dirt = 1; up_read(&sdp->sd_log_flush_lock); gfs2_log_lock(sdp); diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 40bcc37e5a70..0a6801336470 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -787,17 +787,6 @@ restart: gfs2_sys_fs_del(sdp); } -/** - * gfs2_write_super - * @sb: the superblock - * - */ - -static void gfs2_write_super(struct super_block *sb) -{ - sb->s_dirt = 0; -} - /** * gfs2_sync_fs - sync the filesystem * @sb: the superblock @@ -807,7 +796,6 @@ static void gfs2_write_super(struct super_block *sb) static int gfs2_sync_fs(struct super_block *sb, int wait) { - sb->s_dirt = 0; if (wait && sb->s_fs_info) gfs2_log_flush(sb->s_fs_info, NULL); return 0; @@ -1324,7 +1312,6 @@ const struct super_operations gfs2_super_ops = { .write_inode = gfs2_write_inode, .delete_inode = gfs2_delete_inode, .put_super = gfs2_put_super, - .write_super = gfs2_write_super, .sync_fs = gfs2_sync_fs, .freeze_fs = gfs2_freeze, .unfreeze_fs = gfs2_unfreeze, From 94cb993f2ee99f3a9318e7b4dbb383497c4bedea Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 09:46:44 -0400 Subject: [PATCH 32/87] ocfs2: remove ->write_super and stop maintaining ->s_dirt Signed-off-by: Christoph Hellwig Acked-by: Joel Becker Signed-off-by: Al Viro --- fs/ocfs2/super.c | 14 -------------- 1 file changed, 14 deletions(-) diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 5c6163f55039..3eb076ce4c07 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -126,7 +126,6 @@ static int ocfs2_get_sector(struct super_block *sb, struct buffer_head **bh, int block, int sect_size); -static void ocfs2_write_super(struct super_block *sb); static struct inode *ocfs2_alloc_inode(struct super_block *sb); static void ocfs2_destroy_inode(struct inode *inode); static int ocfs2_susp_quotas(struct ocfs2_super *osb, int unsuspend); @@ -141,7 +140,6 @@ static const struct super_operations ocfs2_sops = { .clear_inode = ocfs2_clear_inode, .delete_inode = ocfs2_delete_inode, .sync_fs = ocfs2_sync_fs, - .write_super = ocfs2_write_super, .put_super = ocfs2_put_super, .remount_fs = ocfs2_remount, .show_options = ocfs2_show_options, @@ -365,24 +363,12 @@ static struct file_operations ocfs2_osb_debug_fops = { .llseek = generic_file_llseek, }; -/* - * write_super and sync_fs ripped right out of ext3. - */ -static void ocfs2_write_super(struct super_block *sb) -{ - if (mutex_trylock(&sb->s_lock) != 0) - BUG(); - sb->s_dirt = 0; -} - static int ocfs2_sync_fs(struct super_block *sb, int wait) { int status; tid_t target; struct ocfs2_super *osb = OCFS2_SB(sb); - sb->s_dirt = 0; - if (ocfs2_is_hard_readonly(osb)) return -EROFS; From 517bfae28353e996160518add4d00033d3886e61 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 27 Apr 2009 09:46:45 -0400 Subject: [PATCH 33/87] qnx4: remove ->write_super Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/qnx4/inode.c | 9 --------- 1 file changed, 9 deletions(-) diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c index fe1f0f31d11c..95c12fc613f1 100644 --- a/fs/qnx4/inode.c +++ b/fs/qnx4/inode.c @@ -70,14 +70,6 @@ static void qnx4_delete_inode(struct inode *inode) unlock_kernel(); } -static void qnx4_write_super(struct super_block *sb) -{ - lock_kernel(); - QNX4DEBUG(("qnx4: write_super\n")); - sb->s_dirt = 0; - unlock_kernel(); -} - static int qnx4_write_inode(struct inode *inode, int unused) { struct qnx4_inode_entry *raw_inode; @@ -138,7 +130,6 @@ static const struct super_operations qnx4_sops = #ifdef CONFIG_QNX4FS_RW .write_inode = qnx4_write_inode, .delete_inode = qnx4_delete_inode, - .write_super = qnx4_write_super, #endif }; From 8c85e125124a473d6f3e9bb187b0b84207f81d91 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 28 Apr 2009 18:00:26 +0200 Subject: [PATCH 34/87] remove ->write_super call in generic_shutdown_super We just did a full fs writeout using sync_filesystem before, and if that's not enough for the filesystem it can perform it's own writeout in ->put_super, which many filesystems already do. Move a call to foofs_write_super into every foofs_put_super for now to guarantee identical behaviour until it's cleaned up by the individual filesystem maintainers. Exceptions: - affs already has identical copy & pasted code at the beginning of affs_put_super so no need to do it twice. - xfs does the right thing without it and I have changes pending for the xfs tree touching this are so I don't really need conflicts here.. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/bfs/inode.c | 4 ++++ fs/exofs/super.c | 3 +++ fs/ext2/super.c | 3 +++ fs/ext4/super.c | 3 +++ fs/fat/inode.c | 3 +++ fs/hfs/super.c | 2 ++ fs/hfsplus/super.c | 2 ++ fs/jffs2/super.c | 3 +++ fs/nilfs2/super.c | 4 ++++ fs/reiserfs/super.c | 3 +++ fs/super.c | 2 -- fs/sysv/inode.c | 3 +++ fs/ufs/super.c | 3 +++ 13 files changed, 36 insertions(+), 2 deletions(-) diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index cc4062d12ca2..4cf3d269e271 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -30,6 +30,7 @@ MODULE_LICENSE("GPL"); #define dprintf(x...) #endif +static void bfs_write_super(struct super_block *s); void dump_imap(const char *prefix, struct super_block *s); struct inode *bfs_iget(struct super_block *sb, unsigned long ino) @@ -216,6 +217,9 @@ static void bfs_put_super(struct super_block *s) if (!info) return; + if (s->s_dirt) + bfs_write_super(s); + brelse(info->si_sbh); mutex_destroy(&info->bfs_lock); kfree(info->si_imap); diff --git a/fs/exofs/super.c b/fs/exofs/super.c index 9f1985e857e2..3cdb761db8ad 100644 --- a/fs/exofs/super.c +++ b/fs/exofs/super.c @@ -258,6 +258,9 @@ static void exofs_put_super(struct super_block *sb) int num_pend; struct exofs_sb_info *sbi = sb->s_fs_info; + if (sb->s_dirt) + exofs_write_super(sb); + /* make sure there are no pending commands */ for (num_pend = atomic_read(&sbi->s_curr_pending); num_pend > 0; num_pend = atomic_read(&sbi->s_curr_pending)) { diff --git a/fs/ext2/super.c b/fs/ext2/super.c index e3c748faf2db..932a2bcb6908 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -114,6 +114,9 @@ static void ext2_put_super (struct super_block * sb) int i; struct ext2_sb_info *sbi = EXT2_SB(sb); + if (sb->s_dirt) + ext2_write_super(sb); + ext2_xattr_put_super(sb); if (!(sb->s_flags & MS_RDONLY)) { struct ext2_super_block *es = sbi->s_es; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index f016707597a7..c7b8f8d9b7a8 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -576,6 +576,9 @@ static void ext4_put_super(struct super_block *sb) struct ext4_super_block *es = sbi->s_es; int i, err; + if (sb->s_dirt) + ext4_write_super(sb); + ext4_release_system_zone(sb); ext4_mb_release(sb); ext4_ext_release(sb); diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 296785a0dec8..4978621511bf 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -451,6 +451,9 @@ static void fat_put_super(struct super_block *sb) { struct msdos_sb_info *sbi = MSDOS_SB(sb); + if (sb->s_dirt) + fat_write_super(sb); + if (sbi->nls_disk) { unload_nls(sbi->nls_disk); sbi->nls_disk = NULL; diff --git a/fs/hfs/super.c b/fs/hfs/super.c index a36bb749926d..e071e6d06463 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -65,6 +65,8 @@ static void hfs_write_super(struct super_block *sb) */ static void hfs_put_super(struct super_block *sb) { + if (sb->s_dirt) + hfs_write_super(sb); hfs_mdb_close(sb); /* release the MDB's resources */ hfs_mdb_put(sb); diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index f2a64020f42e..40bdab79dae8 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -199,6 +199,8 @@ static void hfsplus_put_super(struct super_block *sb) dprint(DBG_SUPER, "hfsplus_put_super\n"); if (!sb->s_fs_info) return; + if (sb->s_dirt) + hfsplus_write_super(sb); if (!(sb->s_flags & MS_RDONLY) && HFSPLUS_SB(sb).s_vhdr) { struct hfsplus_vh *vhdr = HFSPLUS_SB(sb).s_vhdr; diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index 4c4e18c54a51..5059e9633edb 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -174,6 +174,9 @@ static void jffs2_put_super (struct super_block *sb) D2(printk(KERN_DEBUG "jffs2: jffs2_put_super()\n")); + if (sb->s_dirt) + jffs2_write_super(sb); + mutex_lock(&c->alloc_sem); jffs2_flush_wbuf_pad(c); mutex_unlock(&c->alloc_sem); diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 6989b03e97ab..7901d8cbb9b1 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -65,6 +65,7 @@ MODULE_DESCRIPTION("A New Implementation of the Log-structured Filesystem " "(NILFS)"); MODULE_LICENSE("GPL"); +static void nilfs_write_super(struct super_block *sb); static int nilfs_remount(struct super_block *sb, int *flags, char *data); static int test_exclusive_mount(struct file_system_type *fs_type, struct block_device *bdev, int flags); @@ -315,6 +316,9 @@ static void nilfs_put_super(struct super_block *sb) struct nilfs_sb_info *sbi = NILFS_SB(sb); struct the_nilfs *nilfs = sbi->s_nilfs; + if (sb->s_dirt) + nilfs_write_super(sb); + nilfs_detach_segment_constructor(sbi); if (!(sb->s_flags & MS_RDONLY)) { diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 9dbdcfb5d314..1b52daa351c5 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -468,6 +468,9 @@ static void reiserfs_put_super(struct super_block *s) struct reiserfs_transaction_handle th; th.t_trans_id = 0; + if (s->s_dirt) + reiserfs_write_super(s); + /* change file system state to current state if it was mounted with read-write permissions */ if (!(s->s_flags & MS_RDONLY)) { if (!journal_begin(&th, s, 10)) { diff --git a/fs/super.c b/fs/super.c index 721236e6177a..d9a29d5b1d28 100644 --- a/fs/super.c +++ b/fs/super.c @@ -311,8 +311,6 @@ void generic_shutdown_super(struct super_block *sb) invalidate_inodes(sb); lock_kernel(); - if (sop->write_super && sb->s_dirt) - sop->write_super(sb); if (sop->put_super) sop->put_super(sb); diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index da20b48d350f..cd80316302cc 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -72,6 +72,9 @@ static void sysv_put_super(struct super_block *sb) { struct sysv_sb_info *sbi = SYSV_SB(sb); + if (sb->s_dirt) + sysv_write_super(sb); + if (!(sb->s_flags & MS_RDONLY)) { /* XXX ext2 also updates the state here */ mark_buffer_dirty(sbi->s_bh1); diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 60359291761f..74afb9fbf58e 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -1152,6 +1152,9 @@ static void ufs_put_super(struct super_block *sb) UFSD("ENTER\n"); + if (sb->s_dirt) + ufs_write_super(sb); + if (!(sb->s_flags & MS_RDONLY)) ufs_put_super_internal(sb); From f3da392e9ff14b9f388e74319e6d195848991c07 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Mon, 4 May 2009 03:32:03 +0400 Subject: [PATCH 35/87] dcache: extrace and use d_unlinked() d_unlinked() will be used in middle-term to ban checkpointing when opened but unlinked file is detected, and in long term, to detect such situation and special case on it. Signed-off-by: Alexey Dobriyan Signed-off-by: Al Viro --- fs/dcache.c | 7 +++---- fs/namespace.c | 8 ++++---- include/linux/dcache.h | 5 +++++ 3 files changed, 12 insertions(+), 8 deletions(-) diff --git a/fs/dcache.c b/fs/dcache.c index 75659a6fd1f8..9e5cd3c3a6ba 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1910,7 +1910,7 @@ char *__d_path(const struct path *path, struct path *root, spin_lock(&vfsmount_lock); prepend(&end, &buflen, "\0", 1); - if (!IS_ROOT(dentry) && d_unhashed(dentry) && + if (d_unlinked(dentry) && (prepend(&end, &buflen, " (deleted)", 10) != 0)) goto Elong; @@ -2035,7 +2035,7 @@ char *dentry_path(struct dentry *dentry, char *buf, int buflen) spin_lock(&dcache_lock); prepend(&end, &buflen, "\0", 1); - if (!IS_ROOT(dentry) && d_unhashed(dentry) && + if (d_unlinked(dentry) && (prepend(&end, &buflen, "//deleted", 9) != 0)) goto Elong; if (buflen < 1) @@ -2097,9 +2097,8 @@ SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size) read_unlock(¤t->fs->lock); error = -ENOENT; - /* Has the current directory has been unlinked? */ spin_lock(&dcache_lock); - if (IS_ROOT(pwd.dentry) || !d_unhashed(pwd.dentry)) { + if (!d_unlinked(pwd.dentry)) { unsigned long len; struct path tmp = root; char * cwd; diff --git a/fs/namespace.c b/fs/namespace.c index 120b8a6b99ed..7e537f0393b5 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1384,7 +1384,7 @@ static int graft_tree(struct vfsmount *mnt, struct path *path) goto out_unlock; err = -ENOENT; - if (IS_ROOT(path->dentry) || !d_unhashed(path->dentry)) + if (!d_unlinked(path->dentry)) err = attach_recursive_mnt(mnt, path, NULL); out_unlock: mutex_unlock(&path->dentry->d_inode->i_mutex); @@ -1566,7 +1566,7 @@ static int do_move_mount(struct path *path, char *old_name) if (IS_DEADDIR(path->dentry->d_inode)) goto out1; - if (!IS_ROOT(path->dentry) && d_unhashed(path->dentry)) + if (d_unlinked(path->dentry)) goto out1; err = -EINVAL; @@ -2129,9 +2129,9 @@ SYSCALL_DEFINE2(pivot_root, const char __user *, new_root, error = -ENOENT; if (IS_DEADDIR(new.dentry->d_inode)) goto out2; - if (d_unhashed(new.dentry) && !IS_ROOT(new.dentry)) + if (d_unlinked(new.dentry)) goto out2; - if (d_unhashed(old.dentry) && !IS_ROOT(old.dentry)) + if (d_unlinked(old.dentry)) goto out2; error = -EBUSY; if (new.mnt == root.mnt || diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 72ce2ae88591..30b93b2a01a4 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -353,6 +353,11 @@ static inline int d_unhashed(struct dentry *dentry) return (dentry->d_flags & DCACHE_UNHASHED); } +static inline int d_unlinked(struct dentry *dentry) +{ + return d_unhashed(dentry) && !IS_ROOT(dentry); +} + static inline struct dentry *dget_parent(struct dentry *dentry) { struct dentry *ret; From e5004753388dcf5e1b8a52ac0ab807d232340fbb Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 5 May 2009 16:08:56 +0200 Subject: [PATCH 36/87] cleanup sync_supers Merge the write_super helper into sync_super and move the check for ->write_super earlier so that we can avoid grabbing a reference to a superblock that doesn't have it. While we're at it also add a little comment documenting sync_supers. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/super.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) diff --git a/fs/super.c b/fs/super.c index d9a29d5b1d28..cb19fffc7681 100644 --- a/fs/super.c +++ b/fs/super.c @@ -399,16 +399,14 @@ void drop_super(struct super_block *sb) EXPORT_SYMBOL(drop_super); -static inline void write_super(struct super_block *sb) -{ - lock_super(sb); - if (sb->s_root && sb->s_dirt) - if (sb->s_op->write_super) - sb->s_op->write_super(sb); - unlock_super(sb); -} - -/* +/** + * sync_supers - helper for periodic superblock writeback + * + * Call the write_super method if present on all dirty superblocks in + * the system. This is for the periodic writeback used by most older + * filesystems. For data integrity superblock writeback use + * sync_filesystems() instead. + * * Note: check the dirty flag before waiting, so we don't * hold up the sync while mounting a device. (The newly * mounted device won't need syncing.) @@ -420,12 +418,17 @@ void sync_supers(void) spin_lock(&sb_lock); restart: list_for_each_entry(sb, &super_blocks, s_list) { - if (sb->s_dirt) { + if (sb->s_op->write_super && sb->s_dirt) { sb->s_count++; spin_unlock(&sb_lock); + down_read(&sb->s_umount); - write_super(sb); + lock_super(sb); + if (sb->s_root && sb->s_dirt) + sb->s_op->write_super(sb); + unlock_super(sb); up_read(&sb->s_umount); + spin_lock(&sb_lock); if (__put_super_and_need_restart(sb)) goto restart; From 5af7926ff33b68b3ba46531471c6e0564b285efc Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 5 May 2009 15:41:25 +0200 Subject: [PATCH 37/87] enforce ->sync_fs is only called for rw superblock Make sure a superblock really is writeable by checking MS_RDONLY under s_umount. sync_filesystems needed some re-arragement for that, but all but one sync_filesystem caller had the correct locking already so that we could add that check there. cachefiles grew s_umount locking. I've also added a WARN_ON to sync_filesystem to assert this for future callers. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/btrfs/super.c | 3 --- fs/cachefiles/interface.c | 2 ++ fs/reiserfs/super.c | 21 +++++++++------------ fs/sync.c | 23 ++++++++++++++++------- fs/ubifs/super.c | 3 --- 5 files changed, 27 insertions(+), 25 deletions(-) diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 52d84522c2c2..9f179d4832d5 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -394,9 +394,6 @@ int btrfs_sync_fs(struct super_block *sb, int wait) struct btrfs_root *root = btrfs_sb(sb); int ret; - if (sb->s_flags & MS_RDONLY) - return 0; - if (!wait) { filemap_flush(root->fs_info->btree_inode->i_mapping); return 0; diff --git a/fs/cachefiles/interface.c b/fs/cachefiles/interface.c index dafd484d7bda..431accd475a7 100644 --- a/fs/cachefiles/interface.c +++ b/fs/cachefiles/interface.c @@ -354,7 +354,9 @@ static void cachefiles_sync_cache(struct fscache_cache *_cache) /* make sure all pages pinned by operations on behalf of the netfs are * written to disc */ cachefiles_begin_secure(cache, &saved_cred); + down_read(&cache->mnt->mnt_sb->s_umount); ret = sync_filesystem(cache->mnt->mnt_sb); + up_read(&cache->mnt->mnt_sb->s_umount); cachefiles_end_secure(cache, saved_cred); if (ret == -EIO) diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 1b52daa351c5..3da0401c0a96 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -64,18 +64,15 @@ static int reiserfs_statfs(struct dentry *dentry, struct kstatfs *buf); static int reiserfs_sync_fs(struct super_block *s, int wait) { - if (!(s->s_flags & MS_RDONLY)) { - struct reiserfs_transaction_handle th; - reiserfs_write_lock(s); - if (!journal_begin(&th, s, 1)) - if (!journal_end_sync(&th, s, 1)) - reiserfs_flush_old_commits(s); - s->s_dirt = 0; /* Even if it's not true. - * We'll loop forever in sync_supers otherwise */ - reiserfs_write_unlock(s); - } else { - s->s_dirt = 0; - } + struct reiserfs_transaction_handle th; + + reiserfs_write_lock(s); + if (!journal_begin(&th, s, 1)) + if (!journal_end_sync(&th, s, 1)) + reiserfs_flush_old_commits(s); + s->s_dirt = 0; /* Even if it's not true. + * We'll loop forever in sync_supers otherwise */ + reiserfs_write_unlock(s); return 0; } diff --git a/fs/sync.c b/fs/sync.c index 4487b5560dc8..89c37f732afa 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -51,6 +51,18 @@ int sync_filesystem(struct super_block *sb) { int ret; + /* + * We need to be protected against the filesystem going from + * r/o to r/w or vice versa. + */ + WARN_ON(!rwsem_is_locked(&sb->s_umount)); + + /* + * No point in syncing out anything if the filesystem is read-only. + */ + if (sb->s_flags & MS_RDONLY) + return 0; + ret = __sync_filesystem(sb, 0); if (ret < 0) return ret; @@ -79,25 +91,22 @@ static void sync_filesystems(int wait) mutex_lock(&mutex); /* Could be down_interruptible */ spin_lock(&sb_lock); - list_for_each_entry(sb, &super_blocks, s_list) { - if (sb->s_flags & MS_RDONLY) - continue; + list_for_each_entry(sb, &super_blocks, s_list) sb->s_need_sync = 1; - } restart: list_for_each_entry(sb, &super_blocks, s_list) { if (!sb->s_need_sync) continue; sb->s_need_sync = 0; - if (sb->s_flags & MS_RDONLY) - continue; /* hm. Was remounted r/o meanwhile */ sb->s_count++; spin_unlock(&sb_lock); + down_read(&sb->s_umount); - if (sb->s_root) + if (!(sb->s_flags & MS_RDONLY) && sb->s_root) __sync_filesystem(sb, wait); up_read(&sb->s_umount); + /* restart only when sb is no longer on the list */ spin_lock(&sb_lock); if (__put_super_and_need_restart(sb)) diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index e9f7a754c4f7..84f3c7fd1552 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -447,9 +447,6 @@ static int ubifs_sync_fs(struct super_block *sb, int wait) if (!wait) return 0; - if (sb->s_flags & MS_RDONLY) - return 0; - /* * VFS calls '->sync_fs()' before synchronizing all dirty inodes and * pages, so synchronize them first, then commit the journal. Strictly From 443b94baaa16771e98b29ca7c24f1e305738ffca Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 5 May 2009 23:48:50 -0400 Subject: [PATCH 38/87] Make sure that all callers of remount hold s_umount exclusive Signed-off-by: Al Viro --- fs/super.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/fs/super.c b/fs/super.c index cb19fffc7681..49f670cb9a83 100644 --- a/fs/super.c +++ b/fs/super.c @@ -579,7 +579,7 @@ static void do_emergency_remount(struct work_struct *work) list_for_each_entry(sb, &super_blocks, s_list) { sb->s_count++; spin_unlock(&sb_lock); - down_read(&sb->s_umount); + down_write(&sb->s_umount); if (sb->s_root && sb->s_bdev && !(sb->s_flags & MS_RDONLY)) { /* * ->remount_fs needs lock_kernel(). @@ -590,7 +590,8 @@ static void do_emergency_remount(struct work_struct *work) do_remount_sb(sb, MS_RDONLY, NULL, 1); unlock_kernel(); } - drop_super(sb); + up_write(&sb->s_umount); + put_super(sb); spin_lock(&sb_lock); } spin_unlock(&sb_lock); From 62c6943b4b1e818aea60c11c5a68a50785b83119 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 7 May 2009 03:12:29 -0400 Subject: [PATCH 39/87] Trim a bit of crap from fs.h do_remount_sb() is fs/internal.h fodder, fsync_no_super() is long gone. Signed-off-by: Al Viro --- fs/internal.h | 5 +++++ include/linux/fs.h | 3 --- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/internal.h b/fs/internal.h index dbec3cc28338..d55ef562f0bb 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -78,3 +78,8 @@ extern void chroot_fs_refs(struct path *, struct path *); * file_table.c */ extern void mark_files_ro(struct super_block *); + +/* + * super.c + */ +extern int do_remount_sb(struct super_block *, int, void *, int); diff --git a/include/linux/fs.h b/include/linux/fs.h index fb1822bed7c8..e7833ef5d1d6 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1942,7 +1942,6 @@ extern struct super_block *freeze_bdev(struct block_device *); extern void emergency_thaw_all(void); extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); extern int fsync_bdev(struct block_device *); -extern int fsync_no_super(struct block_device *); #else static inline void bd_forget(struct inode *inode) {} static inline int sync_blockdev(struct block_device *bdev) { return 0; } @@ -2079,8 +2078,6 @@ extern int vfs_fsync(struct file *file, struct dentry *dentry, int datasync); extern void sync_supers(void); extern void emergency_sync(void); extern void emergency_remount(void); -extern int do_remount_sb(struct super_block *sb, int flags, - void *data, int force); #ifdef CONFIG_BLOCK extern sector_t bmap(struct inode *, sector_t); #endif From a9e220f8322e2b0e0b8903fe00265461cffad3f0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 5 May 2009 22:10:44 -0400 Subject: [PATCH 40/87] No need to do lock_super() for exclusion in generic_shutdown_super() We can't run into contention on it. All other callers of lock_super() either hold s_umount (and we have it exclusive) or hold an active reference to superblock in question, which prevents the call of generic_shutdown_super() while the reference is held. So we can replace lock_super(s) with get_fs_excl() in generic_shutdown_super() (and corresponding change for unlock_super(), of course). Since ext4 expects s_lock held for its put_super, take lock_super() into it. The rest of filesystems do not care at all. Signed-off-by: Al Viro --- fs/ext4/super.c | 2 +- fs/super.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/ext4/super.c b/fs/ext4/super.c index c7b8f8d9b7a8..0d3034c5e8a4 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -576,6 +576,7 @@ static void ext4_put_super(struct super_block *sb) struct ext4_super_block *es = sbi->s_es; int i, err; + lock_super(sb); if (sb->s_dirt) ext4_write_super(sb); @@ -645,7 +646,6 @@ static void ext4_put_super(struct super_block *sb) unlock_super(sb); kobject_put(&sbi->s_kobj); wait_for_completion(&sbi->s_kobj_unregister); - lock_super(sb); lock_kernel(); kfree(sbi->s_blockgroup_lock); kfree(sbi); diff --git a/fs/super.c b/fs/super.c index 49f670cb9a83..54fd331f0cab 100644 --- a/fs/super.c +++ b/fs/super.c @@ -304,7 +304,7 @@ void generic_shutdown_super(struct super_block *sb) if (sb->s_root) { shrink_dcache_for_umount(sb); sync_filesystem(sb); - lock_super(sb); + get_fs_excl(); sb->s_flags &= ~MS_ACTIVE; /* bad name - it should be evict_inodes() */ @@ -322,7 +322,7 @@ void generic_shutdown_super(struct super_block *sb) } unlock_kernel(); - unlock_super(sb); + put_fs_excl(); } spin_lock(&sb_lock); /* should be initialized for __put_super_and_need_restart() */ From 6cfd0148425e528b859b26e436b01f23f6926224 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 5 May 2009 15:40:36 +0200 Subject: [PATCH 41/87] push BKL down into ->put_super Move BKL into ->put_super from the only caller. A couple of filesystems had trivial enough ->put_super (only kfree and NULLing of s_fs_info + stuff in there) to not get any locking: coda, cramfs, efs, hugetlbfs, omfs, qnx4, shmem, all others got the full treatment. Most of them probably don't need it, but I'd rather sort that out individually. Preferably after all the other BKL pushdowns in that area. [AV: original used to move lock_super() down as well; these changes are removed since we don't do lock_super() at all in generic_shutdown_super() now] [AV: fuse, btrfs and xfs are known to need no damn BKL, exempt] Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/adfs/super.c | 4 ++++ fs/affs/super.c | 5 ++++- fs/afs/super.c | 4 ++++ fs/befs/linuxvfs.c | 5 ++++- fs/bfs/inode.c | 4 ++++ fs/cifs/cifsfs.c | 6 +++++- fs/ecryptfs/super.c | 5 +++++ fs/exofs/super.c | 4 ++++ fs/ext2/super.c | 4 +++- fs/ext3/super.c | 5 ++++- fs/ext4/super.c | 2 +- fs/fat/inode.c | 4 ++++ fs/freevxfs/vxfs_super.c | 4 ++++ fs/gfs2/super.c | 4 ++++ fs/hfs/super.c | 4 ++++ fs/hfsplus/super.c | 5 +++++ fs/hpfs/super.c | 5 +++++ fs/isofs/inode.c | 5 +++++ fs/jffs2/super.c | 4 ++++ fs/jfs/super.c | 5 +++++ fs/minix/inode.c | 4 +++- fs/ncpfs/inode.c | 4 ++++ fs/nilfs2/super.c | 4 ++++ fs/ntfs/super.c | 6 +++++- fs/ocfs2/super.c | 4 ++++ fs/reiserfs/super.c | 4 +++- fs/smbfs/inode.c | 4 ++++ fs/squashfs/super.c | 4 ++++ fs/super.c | 3 --- fs/sysv/inode.c | 4 ++++ fs/ubifs/super.c | 5 +++++ fs/udf/super.c | 5 +++++ fs/ufs/super.c | 6 ++++++ 33 files changed, 133 insertions(+), 12 deletions(-) diff --git a/fs/adfs/super.c b/fs/adfs/super.c index dd9becca4241..0ec5aaf47aa7 100644 --- a/fs/adfs/super.c +++ b/fs/adfs/super.c @@ -132,11 +132,15 @@ static void adfs_put_super(struct super_block *sb) int i; struct adfs_sb_info *asb = ADFS_SB(sb); + lock_kernel(); + for (i = 0; i < asb->s_map_size; i++) brelse(asb->s_map[i].dm_bh); kfree(asb->s_map); kfree(asb); sb->s_fs_info = NULL; + + unlock_kernel(); } static int adfs_show_options(struct seq_file *seq, struct vfsmount *mnt) diff --git a/fs/affs/super.c b/fs/affs/super.c index 63f5183f263b..d7386462a8e7 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -29,6 +29,8 @@ affs_put_super(struct super_block *sb) struct affs_sb_info *sbi = AFFS_SB(sb); pr_debug("AFFS: put_super()\n"); + lock_kernel(); + if (!(sb->s_flags & MS_RDONLY)) { AFFS_ROOT_TAIL(sb, sbi->s_root_bh)->bm_flag = cpu_to_be32(1); secs_to_datestamp(get_seconds(), @@ -42,7 +44,8 @@ affs_put_super(struct super_block *sb) affs_brelse(sbi->s_root_bh); kfree(sbi); sb->s_fs_info = NULL; - return; + + unlock_kernel(); } static void diff --git a/fs/afs/super.c b/fs/afs/super.c index 76828e5f8a39..ad0514d0115f 100644 --- a/fs/afs/super.c +++ b/fs/afs/super.c @@ -440,8 +440,12 @@ static void afs_put_super(struct super_block *sb) _enter(""); + lock_kernel(); + afs_put_volume(as->volume); + unlock_kernel(); + _leave(""); } diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index 76afd0d6b86c..9367b6297d84 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -737,6 +737,8 @@ parse_options(char *options, befs_mount_options * opts) static void befs_put_super(struct super_block *sb) { + lock_kernel(); + kfree(BEFS_SB(sb)->mount_opts.iocharset); BEFS_SB(sb)->mount_opts.iocharset = NULL; @@ -747,7 +749,8 @@ befs_put_super(struct super_block *sb) kfree(sb->s_fs_info); sb->s_fs_info = NULL; - return; + + unlock_kernel(); } /* Allocate private field of the superblock, fill it. diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index 4cf3d269e271..3a9a1361fdc1 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -217,6 +217,8 @@ static void bfs_put_super(struct super_block *s) if (!info) return; + lock_kernel(); + if (s->s_dirt) bfs_write_super(s); @@ -225,6 +227,8 @@ static void bfs_put_super(struct super_block *s) kfree(info->si_imap); kfree(info); s->s_fs_info = NULL; + + unlock_kernel(); } static int bfs_statfs(struct dentry *dentry, struct kstatfs *buf) diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 0a10a59b6392..0d92114195ab 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -204,6 +204,9 @@ cifs_put_super(struct super_block *sb) cFYI(1, ("Empty cifs superblock info passed to unmount")); return; } + + lock_kernel(); + rc = cifs_umount(sb, cifs_sb); if (rc) cERROR(1, ("cifs_umount failed with return code %d", rc)); @@ -216,7 +219,8 @@ cifs_put_super(struct super_block *sb) unload_nls(cifs_sb->local_nls); kfree(cifs_sb); - return; + + unlock_kernel(); } static int diff --git a/fs/ecryptfs/super.c b/fs/ecryptfs/super.c index fa4c7e7d15d9..12d649602d3a 100644 --- a/fs/ecryptfs/super.c +++ b/fs/ecryptfs/super.c @@ -27,6 +27,7 @@ #include #include #include +#include #include #include #include "ecryptfs_kernel.h" @@ -120,9 +121,13 @@ static void ecryptfs_put_super(struct super_block *sb) { struct ecryptfs_sb_info *sb_info = ecryptfs_superblock_to_private(sb); + lock_kernel(); + ecryptfs_destroy_mount_crypt_stat(&sb_info->mount_crypt_stat); kmem_cache_free(ecryptfs_sb_info_cache, sb_info); ecryptfs_set_superblock_private(sb, NULL); + + unlock_kernel(); } /** diff --git a/fs/exofs/super.c b/fs/exofs/super.c index 3cdb761db8ad..cd1f8b18a218 100644 --- a/fs/exofs/super.c +++ b/fs/exofs/super.c @@ -258,6 +258,8 @@ static void exofs_put_super(struct super_block *sb) int num_pend; struct exofs_sb_info *sbi = sb->s_fs_info; + lock_kernel(); + if (sb->s_dirt) exofs_write_super(sb); @@ -274,6 +276,8 @@ static void exofs_put_super(struct super_block *sb) osduld_put_device(sbi->s_dev); kfree(sb->s_fs_info); sb->s_fs_info = NULL; + + unlock_kernel(); } /* diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 932a2bcb6908..a44963d8edbd 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -114,6 +114,8 @@ static void ext2_put_super (struct super_block * sb) int i; struct ext2_sb_info *sbi = EXT2_SB(sb); + lock_kernel(); + if (sb->s_dirt) ext2_write_super(sb); @@ -138,7 +140,7 @@ static void ext2_put_super (struct super_block * sb) kfree(sbi->s_blockgroup_lock); kfree(sbi); - return; + unlock_kernel(); } static struct kmem_cache * ext2_inode_cachep; diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 1efd958687e9..546b8d732bf2 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -398,6 +398,8 @@ static void ext3_put_super (struct super_block * sb) struct ext3_super_block *es = sbi->s_es; int i, err; + lock_kernel(); + ext3_xattr_put_super(sb); err = journal_destroy(sbi->s_journal); sbi->s_journal = NULL; @@ -446,7 +448,8 @@ static void ext3_put_super (struct super_block * sb) sb->s_fs_info = NULL; kfree(sbi->s_blockgroup_lock); kfree(sbi); - return; + + unlock_kernel(); } static struct kmem_cache *ext3_inode_cachep; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 0d3034c5e8a4..1d4180b86772 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -577,6 +577,7 @@ static void ext4_put_super(struct super_block *sb) int i, err; lock_super(sb); + lock_kernel(); if (sb->s_dirt) ext4_write_super(sb); @@ -646,7 +647,6 @@ static void ext4_put_super(struct super_block *sb) unlock_super(sb); kobject_put(&sbi->s_kobj); wait_for_completion(&sbi->s_kobj_unregister); - lock_kernel(); kfree(sbi->s_blockgroup_lock); kfree(sbi); } diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 4978621511bf..2b88c93af227 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -451,6 +451,8 @@ static void fat_put_super(struct super_block *sb) { struct msdos_sb_info *sbi = MSDOS_SB(sb); + lock_kernel(); + if (sb->s_dirt) fat_write_super(sb); @@ -470,6 +472,8 @@ static void fat_put_super(struct super_block *sb) sb->s_fs_info = NULL; kfree(sbi); + + unlock_kernel(); } static struct kmem_cache *fat_inode_cachep; diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c index 1dacda831577..cdbd1654e4cd 100644 --- a/fs/freevxfs/vxfs_super.c +++ b/fs/freevxfs/vxfs_super.c @@ -80,12 +80,16 @@ vxfs_put_super(struct super_block *sbp) { struct vxfs_sb_info *infp = VXFS_SBI(sbp); + lock_kernel(); + vxfs_put_fake_inode(infp->vsi_fship); vxfs_put_fake_inode(infp->vsi_ilist); vxfs_put_fake_inode(infp->vsi_stilist); brelse(infp->vsi_bp); kfree(infp); + + unlock_kernel(); } /** diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 0a6801336470..c8930b31cdf0 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -719,6 +719,8 @@ static void gfs2_put_super(struct super_block *sb) int error; struct gfs2_jdesc *jd; + lock_kernel(); + /* Unfreeze the filesystem, if we need to */ mutex_lock(&sdp->sd_freeze_lock); @@ -785,6 +787,8 @@ restart: /* At this point, we're through participating in the lockspace */ gfs2_sys_fs_del(sdp); + + unlock_kernel(); } /** diff --git a/fs/hfs/super.c b/fs/hfs/super.c index e071e6d06463..9f5eaa01cc77 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -65,11 +65,15 @@ static void hfs_write_super(struct super_block *sb) */ static void hfs_put_super(struct super_block *sb) { + lock_kernel(); + if (sb->s_dirt) hfs_write_super(sb); hfs_mdb_close(sb); /* release the MDB's resources */ hfs_mdb_put(sb); + + unlock_kernel(); } /* diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 40bdab79dae8..9b292dcc39c8 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -199,6 +199,9 @@ static void hfsplus_put_super(struct super_block *sb) dprint(DBG_SUPER, "hfsplus_put_super\n"); if (!sb->s_fs_info) return; + + lock_kernel(); + if (sb->s_dirt) hfsplus_write_super(sb); if (!(sb->s_flags & MS_RDONLY) && HFSPLUS_SB(sb).s_vhdr) { @@ -220,6 +223,8 @@ static void hfsplus_put_super(struct super_block *sb) unload_nls(HFSPLUS_SB(sb).nls); kfree(sb->s_fs_info); sb->s_fs_info = NULL; + + unlock_kernel(); } static int hfsplus_statfs(struct dentry *dentry, struct kstatfs *buf) diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index fc77965be841..437a32e9deac 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -99,11 +99,16 @@ int hpfs_stop_cycles(struct super_block *s, int key, int *c1, int *c2, static void hpfs_put_super(struct super_block *s) { struct hpfs_sb_info *sbi = hpfs_sb(s); + + lock_kernel(); + kfree(sbi->sb_cp_table); kfree(sbi->sb_bmp_dir); unmark_dirty(s); s->s_fs_info = NULL; kfree(sbi); + + unlock_kernel(); } unsigned hpfs_count_one_bitmap(struct super_block *s, secno secno) diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index b4cbe9603c7d..068b34b5a107 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -42,11 +42,16 @@ static int isofs_dentry_cmp_ms(struct dentry *dentry, struct qstr *a, struct qst static void isofs_put_super(struct super_block *sb) { struct isofs_sb_info *sbi = ISOFS_SB(sb); + #ifdef CONFIG_JOLIET + lock_kernel(); + if (sbi->s_nls_iocharset) { unload_nls(sbi->s_nls_iocharset); sbi->s_nls_iocharset = NULL; } + + unlock_kernel(); #endif kfree(sbi); diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index 5059e9633edb..37b12125c127 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -174,6 +174,8 @@ static void jffs2_put_super (struct super_block *sb) D2(printk(KERN_DEBUG "jffs2: jffs2_put_super()\n")); + lock_kernel(); + if (sb->s_dirt) jffs2_write_super(sb); @@ -195,6 +197,8 @@ static void jffs2_put_super (struct super_block *sb) if (c->mtd->sync) c->mtd->sync(c->mtd); + unlock_kernel(); + D1(printk(KERN_DEBUG "jffs2_put_super returning\n")); } diff --git a/fs/jfs/super.c b/fs/jfs/super.c index d9b0e92b3602..3eb13adf3862 100644 --- a/fs/jfs/super.c +++ b/fs/jfs/super.c @@ -183,6 +183,9 @@ static void jfs_put_super(struct super_block *sb) int rc; jfs_info("In jfs_put_super"); + + lock_kernel(); + rc = jfs_umount(sb); if (rc) jfs_err("jfs_umount failed with return code %d", rc); @@ -195,6 +198,8 @@ static void jfs_put_super(struct super_block *sb) sbi->direct_inode = NULL; kfree(sbi); + + unlock_kernel(); } enum { diff --git a/fs/minix/inode.c b/fs/minix/inode.c index daad3c2740db..7eb53970f4bc 100644 --- a/fs/minix/inode.c +++ b/fs/minix/inode.c @@ -35,6 +35,8 @@ static void minix_put_super(struct super_block *sb) int i; struct minix_sb_info *sbi = minix_sb(sb); + lock_kernel(); + if (!(sb->s_flags & MS_RDONLY)) { if (sbi->s_version != MINIX_V3) /* s_state is now out from V3 sb */ sbi->s_ms->s_state = sbi->s_mount_state; @@ -49,7 +51,7 @@ static void minix_put_super(struct super_block *sb) sb->s_fs_info = NULL; kfree(sbi); - return; + unlock_kernel(); } static struct kmem_cache * minix_inode_cachep; diff --git a/fs/ncpfs/inode.c b/fs/ncpfs/inode.c index d642f0e5b365..b99ce205b1bd 100644 --- a/fs/ncpfs/inode.c +++ b/fs/ncpfs/inode.c @@ -736,6 +736,8 @@ static void ncp_put_super(struct super_block *sb) { struct ncp_server *server = NCP_SBP(sb); + lock_kernel(); + ncp_lock_server(server); ncp_disconnect(server); ncp_unlock_server(server); @@ -769,6 +771,8 @@ static void ncp_put_super(struct super_block *sb) vfree(server->packet); sb->s_fs_info = NULL; kfree(server); + + unlock_kernel(); } static int ncp_statfs(struct dentry *dentry, struct kstatfs *buf) diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 7901d8cbb9b1..7262e8427c20 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -316,6 +316,8 @@ static void nilfs_put_super(struct super_block *sb) struct nilfs_sb_info *sbi = NILFS_SB(sb); struct the_nilfs *nilfs = sbi->s_nilfs; + lock_kernel(); + if (sb->s_dirt) nilfs_write_super(sb); @@ -333,6 +335,8 @@ static void nilfs_put_super(struct super_block *sb) sbi->s_super = NULL; sb->s_fs_info = NULL; kfree(sbi); + + unlock_kernel(); } /** diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c index 6aa7c4713536..a9ec4e1084e4 100644 --- a/fs/ntfs/super.c +++ b/fs/ntfs/super.c @@ -2246,6 +2246,9 @@ static void ntfs_put_super(struct super_block *sb) ntfs_volume *vol = NTFS_SB(sb); ntfs_debug("Entering."); + + lock_kernel(); + #ifdef NTFS_RW /* * Commit all inodes while they are still open in case some of them @@ -2444,7 +2447,8 @@ static void ntfs_put_super(struct super_block *sb) } sb->s_fs_info = NULL; kfree(vol); - return; + + unlock_kernel(); } /** diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 3eb076ce4c07..02737596b597 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1536,9 +1536,13 @@ static void ocfs2_put_super(struct super_block *sb) { mlog_entry("(0x%p)\n", sb); + lock_kernel(); + ocfs2_sync_blockdev(sb); ocfs2_dismount_volume(sb, 0); + unlock_kernel(); + mlog_exit_void(); } diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 3da0401c0a96..90dcb7b033ea 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -465,6 +465,8 @@ static void reiserfs_put_super(struct super_block *s) struct reiserfs_transaction_handle th; th.t_trans_id = 0; + lock_kernel(); + if (s->s_dirt) reiserfs_write_super(s); @@ -500,7 +502,7 @@ static void reiserfs_put_super(struct super_block *s) kfree(s->s_fs_info); s->s_fs_info = NULL; - return; + unlock_kernel(); } static struct kmem_cache *reiserfs_inode_cachep; diff --git a/fs/smbfs/inode.c b/fs/smbfs/inode.c index fc27fbfc5397..1402d2d54f52 100644 --- a/fs/smbfs/inode.c +++ b/fs/smbfs/inode.c @@ -474,6 +474,8 @@ smb_put_super(struct super_block *sb) { struct smb_sb_info *server = SMB_SB(sb); + lock_kernel(); + smb_lock_server(server); server->state = CONN_INVALID; smbiod_unregister_server(server); @@ -489,6 +491,8 @@ smb_put_super(struct super_block *sb) smb_unlock_server(server); put_pid(server->conn_pid); kfree(server); + + unlock_kernel(); } static int smb_fill_super(struct super_block *sb, void *raw_data, int silent) diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c index 0adc624c956f..3b52770f46ff 100644 --- a/fs/squashfs/super.c +++ b/fs/squashfs/super.c @@ -338,6 +338,8 @@ static int squashfs_remount(struct super_block *sb, int *flags, char *data) static void squashfs_put_super(struct super_block *sb) { + lock_kernel(); + if (sb->s_fs_info) { struct squashfs_sb_info *sbi = sb->s_fs_info; squashfs_cache_delete(sbi->block_cache); @@ -350,6 +352,8 @@ static void squashfs_put_super(struct super_block *sb) kfree(sb->s_fs_info); sb->s_fs_info = NULL; } + + unlock_kernel(); } diff --git a/fs/super.c b/fs/super.c index 54fd331f0cab..bdd7158b785e 100644 --- a/fs/super.c +++ b/fs/super.c @@ -309,7 +309,6 @@ void generic_shutdown_super(struct super_block *sb) /* bad name - it should be evict_inodes() */ invalidate_inodes(sb); - lock_kernel(); if (sop->put_super) sop->put_super(sb); @@ -320,8 +319,6 @@ void generic_shutdown_super(struct super_block *sb) "Self-destruct in 5 seconds. Have a nice day...\n", sb->s_id); } - - unlock_kernel(); put_fs_excl(); } spin_lock(&sb_lock); diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index cd80316302cc..a8189864c241 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -72,6 +72,8 @@ static void sysv_put_super(struct super_block *sb) { struct sysv_sb_info *sbi = SYSV_SB(sb); + lock_kernel(); + if (sb->s_dirt) sysv_write_super(sb); @@ -87,6 +89,8 @@ static void sysv_put_super(struct super_block *sb) brelse(sbi->s_bh2); kfree(sbi); + + unlock_kernel(); } static int sysv_statfs(struct dentry *dentry, struct kstatfs *buf) diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 84f3c7fd1552..522c3fd7eb3c 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -1684,6 +1684,9 @@ static void ubifs_put_super(struct super_block *sb) ubifs_msg("un-mount UBI device %d, volume %d", c->vi.ubi_num, c->vi.vol_id); + + lock_kernel(); + /* * The following asserts are only valid if there has not been a failure * of the media. For example, there will be dirty inodes if we failed @@ -1750,6 +1753,8 @@ static void ubifs_put_super(struct super_block *sb) ubi_close_volume(c->ubi); mutex_unlock(&c->umount_mutex); kfree(c); + + unlock_kernel(); } static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data) diff --git a/fs/udf/super.c b/fs/udf/super.c index 0ba44107d8f1..04802cc39b18 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -2062,6 +2062,9 @@ static void udf_put_super(struct super_block *sb) struct udf_sb_info *sbi; sbi = UDF_SB(sb); + + lock_kernel(); + if (sbi->s_vat_inode) iput(sbi->s_vat_inode); if (sbi->s_partitions) @@ -2077,6 +2080,8 @@ static void udf_put_super(struct super_block *sb) kfree(sbi->s_partmaps); kfree(sb->s_fs_info); sb->s_fs_info = NULL; + + unlock_kernel(); } static int udf_sync_fs(struct super_block *sb, int wait) diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 74afb9fbf58e..2b4d2b6234df 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -594,6 +594,9 @@ static void ufs_put_super_internal(struct super_block *sb) UFSD("ENTER\n"); + + lock_kernel(); + ufs_put_cstotal(sb); size = uspi->s_cssize; blks = (size + uspi->s_fsize - 1) >> uspi->s_fshift; @@ -621,6 +624,9 @@ static void ufs_put_super_internal(struct super_block *sb) brelse (sbi->s_ucg[i]); kfree (sbi->s_ucg); kfree (base); + + unlock_kernel(); + UFSD("EXIT\n"); } From bbd6851a3213a525128473e978b692ab6ac11aba Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 6 May 2009 10:43:07 -0400 Subject: [PATCH 42/87] Push lock_super() into the ->remount_fs() of filesystems that care about it Note that since we can't run into contention between remount_fs and write_super (due to exclusion on s_umount), we have to care only about filesystems that touch lock_super() on their own. Out of those ext3, ext4, hpfs, sysv and ufs do need it; fat doesn't since its ->remount_fs() only accesses assign-once data (basically, it's "we have no atime on directories and only have atime on files for vfat; force nodiratime and possibly noatime into *flags"). [folded a build fix from hch] Signed-off-by: Al Viro --- fs/ext3/super.c | 3 +++ fs/ext4/super.c | 3 +++ fs/hpfs/super.c | 3 +++ fs/super.c | 2 -- fs/sysv/inode.c | 2 ++ fs/ufs/super.c | 11 ++++++++++- 6 files changed, 21 insertions(+), 3 deletions(-) diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 546b8d732bf2..e213a2613a56 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -2491,6 +2491,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data) #endif /* Store the original options */ + lock_super(sb); old_sb_flags = sb->s_flags; old_opts.s_mount_opt = sbi->s_mount_opt; old_opts.s_resuid = sbi->s_resuid; @@ -2598,6 +2599,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data) old_opts.s_qf_names[i] != sbi->s_qf_names[i]) kfree(old_opts.s_qf_names[i]); #endif + unlock_super(sb); return 0; restore_opts: sb->s_flags = old_sb_flags; @@ -2614,6 +2616,7 @@ restore_opts: sbi->s_qf_names[i] = old_opts.s_qf_names[i]; } #endif + unlock_super(sb); return err; } diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 1d4180b86772..a9c683425929 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3421,6 +3421,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) #endif /* Store the original options */ + lock_super(sb); old_sb_flags = sb->s_flags; old_opts.s_mount_opt = sbi->s_mount_opt; old_opts.s_resuid = sbi->s_resuid; @@ -3554,6 +3555,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) old_opts.s_qf_names[i] != sbi->s_qf_names[i]) kfree(old_opts.s_qf_names[i]); #endif + unlock_super(sb); return 0; restore_opts: @@ -3573,6 +3575,7 @@ restore_opts: sbi->s_qf_names[i] = old_opts.s_qf_names[i]; } #endif + unlock_super(sb); return err; } diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index 437a32e9deac..f68193cf0811 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -398,6 +398,7 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) *flags |= MS_NOATIME; + lock_super(s); uid = sbi->sb_uid; gid = sbi->sb_gid; umask = 0777 & ~sbi->sb_mode; lowercase = sbi->sb_lowercase; conv = sbi->sb_conv; @@ -430,9 +431,11 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) replace_mount_options(s, new_opts); + unlock_super(s); return 0; out_err: + unlock_super(s); kfree(new_opts); return -EINVAL; } diff --git a/fs/super.c b/fs/super.c index bdd7158b785e..2a49fed77453 100644 --- a/fs/super.c +++ b/fs/super.c @@ -556,9 +556,7 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) remount_rw = !(flags & MS_RDONLY) && (sb->s_flags & MS_RDONLY); if (sb->s_op->remount_fs) { - lock_super(sb); retval = sb->s_op->remount_fs(sb, &flags, data); - unlock_super(sb); if (retval) return retval; } diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index a8189864c241..e0a39f1fb88e 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -61,10 +61,12 @@ clean: static int sysv_remount(struct super_block *sb, int *flags, char *data) { struct sysv_sb_info *sbi = SYSV_SB(sb); + lock_super(sb); if (sbi->s_forced_ro) *flags |= MS_RDONLY; if (!(*flags & MS_RDONLY)) sb->s_dirt = 1; + unlock_super(sb); return 0; } diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 2b4d2b6234df..a5ecabfdc976 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -1181,6 +1181,7 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) unsigned new_mount_opt, ufstype; unsigned flags; + lock_super(sb); uspi = UFS_SB(sb)->s_uspi; flags = UFS_SB(sb)->s_flags; usb1 = ubh_get_usb_first(uspi); @@ -1193,17 +1194,21 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) ufstype = UFS_SB(sb)->s_mount_opt & UFS_MOUNT_UFSTYPE; new_mount_opt = 0; ufs_set_opt (new_mount_opt, ONERROR_LOCK); - if (!ufs_parse_options (data, &new_mount_opt)) + if (!ufs_parse_options (data, &new_mount_opt)) { + unlock_super(sb); return -EINVAL; + } if (!(new_mount_opt & UFS_MOUNT_UFSTYPE)) { new_mount_opt |= ufstype; } else if ((new_mount_opt & UFS_MOUNT_UFSTYPE) != ufstype) { printk("ufstype can't be changed during remount\n"); + unlock_super(sb); return -EINVAL; } if ((*mount_flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) { UFS_SB(sb)->s_mount_opt = new_mount_opt; + unlock_super(sb); return 0; } @@ -1228,6 +1233,7 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) #ifndef CONFIG_UFS_FS_WRITE printk("ufs was compiled with read-only support, " "can't be mounted as read-write\n"); + unlock_super(sb); return -EINVAL; #else if (ufstype != UFS_MOUNT_UFSTYPE_SUN && @@ -1236,16 +1242,19 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) ufstype != UFS_MOUNT_UFSTYPE_SUNx86 && ufstype != UFS_MOUNT_UFSTYPE_UFS2) { printk("this ufstype is read-only supported\n"); + unlock_super(sb); return -EINVAL; } if (!ufs_read_cylinder_structures(sb)) { printk("failed during remounting\n"); + unlock_super(sb); return -EPERM; } sb->s_flags &= ~MS_RDONLY; #endif } UFS_SB(sb)->s_mount_opt = new_mount_opt; + unlock_super(sb); return 0; } From 6fac98dd218653c6aff8a0f56305c424930cea2a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 8 May 2009 13:31:17 -0400 Subject: [PATCH 43/87] Push BKL into do_mount() Signed-off-by: Al Viro --- arch/alpha/kernel/osf_sys.c | 3 --- fs/compat.c | 2 -- fs/namespace.c | 4 ++-- 3 files changed, 2 insertions(+), 7 deletions(-) diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c index 42ee05981e71..9a3334ae282e 100644 --- a/arch/alpha/kernel/osf_sys.c +++ b/arch/alpha/kernel/osf_sys.c @@ -371,8 +371,6 @@ SYSCALL_DEFINE4(osf_mount, unsigned long, typenr, char __user *, path, int retval = -EINVAL; char *name; - lock_kernel(); - name = getname(path); retval = PTR_ERR(name); if (IS_ERR(name)) @@ -392,7 +390,6 @@ SYSCALL_DEFINE4(osf_mount, unsigned long, typenr, char __user *, path, } putname(name); out: - unlock_kernel(); return retval; } diff --git a/fs/compat.c b/fs/compat.c index bb2a9b2e8173..6aefb776dfeb 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -812,10 +812,8 @@ asmlinkage long compat_sys_mount(char __user * dev_name, char __user * dir_name, } } - lock_kernel(); retval = do_mount((char*)dev_page, dir_page, (char*)type_page, flags, (void*)data_page); - unlock_kernel(); out4: free_page(data_page); diff --git a/fs/namespace.c b/fs/namespace.c index 7e537f0393b5..4740f7bdb556 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1921,6 +1921,7 @@ long do_mount(char *dev_name, char *dir_name, char *type_page, if (retval) goto dput_out; + lock_kernel(); if (flags & MS_REMOUNT) retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags, data_page); @@ -1933,6 +1934,7 @@ long do_mount(char *dev_name, char *dir_name, char *type_page, else retval = do_new_mount(&path, type_page, flags, mnt_flags, dev_name, data_page); + unlock_kernel(); dput_out: path_put(&path); return retval; @@ -2046,10 +2048,8 @@ SYSCALL_DEFINE5(mount, char __user *, dev_name, char __user *, dir_name, if (retval < 0) goto out3; - lock_kernel(); retval = do_mount((char *)dev_page, dir_page, (char *)type_page, flags, (void *)data_page); - unlock_kernel(); free_page(data_page); out3: From 7f78d4cd4c5d01864943c22b79df1b6bde923129 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 8 May 2009 13:34:06 -0400 Subject: [PATCH 44/87] Push BKL down beyond VFS-only parts of do_mount() Signed-off-by: Al Viro --- fs/namespace.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/fs/namespace.c b/fs/namespace.c index 4740f7bdb556..b94325f00c5a 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1515,8 +1515,11 @@ static int do_remount(struct path *path, int flags, int mnt_flags, down_write(&sb->s_umount); if (flags & MS_BIND) err = change_mount_flags(path->mnt, flags); - else + else { + lock_kernel(); err = do_remount_sb(sb, flags, data, 0); + unlock_kernel(); + } if (!err) path->mnt->mnt_flags = mnt_flags; up_write(&sb->s_umount); @@ -1630,7 +1633,9 @@ static int do_new_mount(struct path *path, char *type, int flags, if (!capable(CAP_SYS_ADMIN)) return -EPERM; + lock_kernel(); mnt = do_kern_mount(type, flags, name, data); + unlock_kernel(); if (IS_ERR(mnt)) return PTR_ERR(mnt); @@ -1921,7 +1926,6 @@ long do_mount(char *dev_name, char *dir_name, char *type_page, if (retval) goto dput_out; - lock_kernel(); if (flags & MS_REMOUNT) retval = do_remount(&path, flags & ~MS_REMOUNT, mnt_flags, data_page); @@ -1934,7 +1938,6 @@ long do_mount(char *dev_name, char *dir_name, char *type_page, else retval = do_new_mount(&path, type_page, flags, mnt_flags, dev_name, data_page); - unlock_kernel(); dput_out: path_put(&path); return retval; From 4aa98cf768b6f2ea4b204620d949a665959214f6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 8 May 2009 13:36:58 -0400 Subject: [PATCH 45/87] Push BKL down into do_remount_sb() [folded fix from Jiri Slaby] Signed-off-by: Al Viro --- fs/namespace.c | 10 ++-------- fs/super.c | 16 +++++++++++----- 2 files changed, 13 insertions(+), 13 deletions(-) diff --git a/fs/namespace.c b/fs/namespace.c index b94325f00c5a..2dd333b0fe7f 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1060,11 +1060,8 @@ static int do_umount(struct vfsmount *mnt, int flags) * we just try to remount it readonly. */ down_write(&sb->s_umount); - if (!(sb->s_flags & MS_RDONLY)) { - lock_kernel(); + if (!(sb->s_flags & MS_RDONLY)) retval = do_remount_sb(sb, MS_RDONLY, NULL, 0); - unlock_kernel(); - } up_write(&sb->s_umount); return retval; } @@ -1515,11 +1512,8 @@ static int do_remount(struct path *path, int flags, int mnt_flags, down_write(&sb->s_umount); if (flags & MS_BIND) err = change_mount_flags(path->mnt, flags); - else { - lock_kernel(); + else err = do_remount_sb(sb, flags, data, 0); - unlock_kernel(); - } if (!err) path->mnt->mnt_flags = mnt_flags; up_write(&sb->s_umount); diff --git a/fs/super.c b/fs/super.c index 2a49fed77453..a64f36208797 100644 --- a/fs/super.c +++ b/fs/super.c @@ -542,25 +542,33 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) shrink_dcache_sb(sb); sync_filesystem(sb); + lock_kernel(); /* If we are remounting RDONLY and current sb is read/write, make sure there are no rw files opened */ if ((flags & MS_RDONLY) && !(sb->s_flags & MS_RDONLY)) { if (force) mark_files_ro(sb); - else if (!fs_may_remount_ro(sb)) + else if (!fs_may_remount_ro(sb)) { + unlock_kernel(); return -EBUSY; + } retval = vfs_dq_off(sb, 1); - if (retval < 0 && retval != -ENOSYS) + if (retval < 0 && retval != -ENOSYS) { + unlock_kernel(); return -EBUSY; + } } remount_rw = !(flags & MS_RDONLY) && (sb->s_flags & MS_RDONLY); if (sb->s_op->remount_fs) { retval = sb->s_op->remount_fs(sb, &flags, data); - if (retval) + if (retval) { + unlock_kernel(); return retval; + } } sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK); + unlock_kernel(); if (remount_rw) vfs_dq_quota_on_remount(sb); return 0; @@ -581,9 +589,7 @@ static void do_emergency_remount(struct work_struct *work) * * What lock protects sb->s_flags?? */ - lock_kernel(); do_remount_sb(sb, MS_RDONLY, NULL, 1); - unlock_kernel(); } up_write(&sb->s_umount); put_super(sb); From 01ba687577647beef6c5f2ea59bfb56fac9fcde2 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 11 May 2009 23:34:27 +0200 Subject: [PATCH 46/87] jffs2: move jffs2_write_super to super.c jffs2_write_super is only called from super.c and doesn't use any functionality from fs.c. So move it over to super.c and make it static there. [should go in through the vfs tree as it is a requirement for the next patch] Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/jffs2/fs.c | 15 --------------- fs/jffs2/os-linux.h | 1 - fs/jffs2/super.c | 14 ++++++++++++++ 3 files changed, 14 insertions(+), 16 deletions(-) diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c index 249305d65d5b..237b27a3d570 100644 --- a/fs/jffs2/fs.c +++ b/fs/jffs2/fs.c @@ -402,21 +402,6 @@ int jffs2_remount_fs (struct super_block *sb, int *flags, char *data) return 0; } -void jffs2_write_super (struct super_block *sb) -{ - struct jffs2_sb_info *c = JFFS2_SB_INFO(sb); - sb->s_dirt = 0; - - if (sb->s_flags & MS_RDONLY) - return; - - D1(printk(KERN_DEBUG "jffs2_write_super()\n")); - jffs2_garbage_collect_trigger(c); - jffs2_erase_pending_blocks(c, 0); - jffs2_flush_wbuf_gc(c, 0); -} - - /* jffs2_new_inode: allocate a new inode and inocache, add it to the hash, fill in the raw_inode while you're at it. */ struct inode *jffs2_new_inode (struct inode *dir_i, int mode, struct jffs2_raw_inode *ri) diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h index 5e194a5c8e29..2228380c47b9 100644 --- a/fs/jffs2/os-linux.h +++ b/fs/jffs2/os-linux.h @@ -181,7 +181,6 @@ void jffs2_dirty_inode(struct inode *inode); struct inode *jffs2_new_inode (struct inode *dir_i, int mode, struct jffs2_raw_inode *ri); int jffs2_statfs (struct dentry *, struct kstatfs *); -void jffs2_write_super (struct super_block *); int jffs2_remount_fs (struct super_block *, int *, char *); int jffs2_do_fill_super(struct super_block *sb, void *data, int silent); void jffs2_gc_release_inode(struct jffs2_sb_info *c, diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index 37b12125c127..a80a50e445e2 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -53,6 +53,20 @@ static void jffs2_i_init_once(void *foo) inode_init_once(&f->vfs_inode); } +static void jffs2_write_super(struct super_block *sb) +{ + struct jffs2_sb_info *c = JFFS2_SB_INFO(sb); + sb->s_dirt = 0; + + if (sb->s_flags & MS_RDONLY) + return; + + D1(printk(KERN_DEBUG "jffs2_write_super()\n")); + jffs2_garbage_collect_trigger(c); + jffs2_erase_pending_blocks(c, 0); + jffs2_flush_wbuf_gc(c, 0); +} + static int jffs2_sync_fs(struct super_block *sb, int wait) { struct jffs2_sb_info *c = JFFS2_SB_INFO(sb); From ebc1ac164560a241d9bf1b7519062910c3f90a01 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 11 May 2009 23:35:03 +0200 Subject: [PATCH 47/87] ->write_super lock_super pushdown Push down lock_super into ->write_super instances and remove it from the caller. Following filesystem don't need ->s_lock in ->write_super and are skipped: * bfs, nilfs2 - no other uses of s_lock and have internal locks in ->write_super * ext2 - uses BKL in ext2_write_super and has internal calls without s_lock * reiserfs - no other uses of s_lock as has reiserfs_write_lock (BKL) in ->write_super * xfs - no other uses of s_lock and uses internal lock (buffer lock on superblock buffer) to serialize ->write_super. Also xfs_fs_write_super is superflous and will go away in the next merge window Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/affs/super.c | 2 ++ fs/exofs/super.c | 2 ++ fs/ext4/super.c | 4 +++- fs/fat/inode.c | 2 ++ fs/hfs/super.c | 8 +++++--- fs/hfsplus/super.c | 6 +++++- fs/jffs2/super.c | 15 +++++++++------ fs/super.c | 2 -- fs/sync.c | 4 ---- fs/sysv/inode.c | 2 ++ fs/ufs/super.c | 2 ++ 11 files changed, 32 insertions(+), 17 deletions(-) diff --git a/fs/affs/super.c b/fs/affs/super.c index d7386462a8e7..280d361af41f 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -54,6 +54,7 @@ affs_write_super(struct super_block *sb) int clean = 2; struct affs_sb_info *sbi = AFFS_SB(sb); + lock_super(sb); if (!(sb->s_flags & MS_RDONLY)) { // if (sbi->s_bitmap[i].bm_bh) { // if (buffer_dirty(sbi->s_bitmap[i].bm_bh)) { @@ -66,6 +67,7 @@ affs_write_super(struct super_block *sb) sb->s_dirt = !clean; /* redo until bitmap synced */ } else sb->s_dirt = 0; + unlock_super(sb); pr_debug("AFFS: write_super() at %lu, clean=%d\n", get_seconds(), clean); } diff --git a/fs/exofs/super.c b/fs/exofs/super.c index cd1f8b18a218..49e16af4e619 100644 --- a/fs/exofs/super.c +++ b/fs/exofs/super.c @@ -214,6 +214,7 @@ static void exofs_write_super(struct super_block *sb) return; } + lock_super(sb); lock_kernel(); sbi = sb->s_fs_info; fscb->s_nextid = cpu_to_le64(sbi->s_nextid); @@ -246,6 +247,7 @@ out: if (or) osd_end_request(or); unlock_kernel(); + unlock_super(sb); kfree(fscb); } diff --git a/fs/ext4/super.c b/fs/ext4/super.c index a9c683425929..c17200a42301 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -579,7 +579,7 @@ static void ext4_put_super(struct super_block *sb) lock_super(sb); lock_kernel(); if (sb->s_dirt) - ext4_write_super(sb); + ext4_commit_super(sb, 1); ext4_release_system_zone(sb); ext4_mb_release(sb); @@ -3336,7 +3336,9 @@ int ext4_force_commit(struct super_block *sb) static void ext4_write_super(struct super_block *sb) { + lock_super(sb); ext4_commit_super(sb, 1); + unlock_super(sb); } static int ext4_sync_fs(struct super_block *sb, int wait) diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 2b88c93af227..2292cbf7d364 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -441,10 +441,12 @@ static void fat_clear_inode(struct inode *inode) static void fat_write_super(struct super_block *sb) { + lock_super(sb); sb->s_dirt = 0; if (!(sb->s_flags & MS_RDONLY)) fat_clusters_flush(sb); + unlock_super(sb); } static void fat_put_super(struct super_block *sb) diff --git a/fs/hfs/super.c b/fs/hfs/super.c index 9f5eaa01cc77..3aac41751030 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -49,11 +49,13 @@ MODULE_LICENSE("GPL"); */ static void hfs_write_super(struct super_block *sb) { + lock_super(sb); sb->s_dirt = 0; - if (sb->s_flags & MS_RDONLY) - return; + /* sync everything to the buffers */ - hfs_mdb_commit(sb); + if (!(sb->s_flags & MS_RDONLY)) + hfs_mdb_commit(sb); + unlock_super(sb); } /* diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 9b292dcc39c8..1aab8aa7801e 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -157,10 +157,12 @@ static void hfsplus_write_super(struct super_block *sb) struct hfsplus_vh *vhdr = HFSPLUS_SB(sb).s_vhdr; dprint(DBG_SUPER, "hfsplus_write_super\n"); + + lock_super(sb); sb->s_dirt = 0; if (sb->s_flags & MS_RDONLY) /* warn? */ - return; + goto out; vhdr->free_blocks = cpu_to_be32(HFSPLUS_SB(sb).free_blocks); vhdr->next_alloc = cpu_to_be32(HFSPLUS_SB(sb).next_alloc); @@ -192,6 +194,8 @@ static void hfsplus_write_super(struct super_block *sb) } HFSPLUS_SB(sb).flags &= ~HFSPLUS_SB_WRITEBACKUP; } + out: + unlock_super(sb); } static void hfsplus_put_super(struct super_block *sb) diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index a80a50e445e2..f7bfd3ac8bfa 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -56,15 +56,18 @@ static void jffs2_i_init_once(void *foo) static void jffs2_write_super(struct super_block *sb) { struct jffs2_sb_info *c = JFFS2_SB_INFO(sb); + + lock_super(sb); sb->s_dirt = 0; - if (sb->s_flags & MS_RDONLY) - return; + if (!(sb->s_flags & MS_RDONLY)) { + D1(printk(KERN_DEBUG "jffs2_write_super()\n")); + jffs2_garbage_collect_trigger(c); + jffs2_erase_pending_blocks(c, 0); + jffs2_flush_wbuf_gc(c, 0); + } - D1(printk(KERN_DEBUG "jffs2_write_super()\n")); - jffs2_garbage_collect_trigger(c); - jffs2_erase_pending_blocks(c, 0); - jffs2_flush_wbuf_gc(c, 0); + unlock_super(sb); } static int jffs2_sync_fs(struct super_block *sb, int wait) diff --git a/fs/super.c b/fs/super.c index a64f36208797..1905f4af01cc 100644 --- a/fs/super.c +++ b/fs/super.c @@ -420,10 +420,8 @@ restart: spin_unlock(&sb_lock); down_read(&sb->s_umount); - lock_super(sb); if (sb->s_root && sb->s_dirt) sb->s_op->write_super(sb); - unlock_super(sb); up_read(&sb->s_umount); spin_lock(&sb_lock); diff --git a/fs/sync.c b/fs/sync.c index 89c37f732afa..e9d56f6c0b74 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -33,10 +33,8 @@ static int __sync_filesystem(struct super_block *sb, int wait) else sync_quota_sb(sb, -1); sync_inodes_sb(sb, wait); - lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) sb->s_op->write_super(sb); - unlock_super(sb); if (sb->s_op->sync_fs) sb->s_op->sync_fs(sb, wait); return __sync_blockdev(sb->s_bdev, wait); @@ -164,10 +162,8 @@ int file_fsync(struct file *filp, struct dentry *dentry, int datasync) /* sync the superblock to buffers */ sb = inode->i_sb; - lock_super(sb); if (sb->s_dirt && sb->s_op->write_super) sb->s_op->write_super(sb); - unlock_super(sb); /* .. finally sync the buffers to disk */ err = sync_blockdev(sb->s_bdev); diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index e0a39f1fb88e..a3f45fc626a1 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -37,6 +37,7 @@ static void sysv_write_super(struct super_block *sb) struct sysv_sb_info *sbi = SYSV_SB(sb); unsigned long time = get_seconds(), old_time; + lock_super(sb); lock_kernel(); if (sb->s_flags & MS_RDONLY) goto clean; @@ -56,6 +57,7 @@ static void sysv_write_super(struct super_block *sb) clean: sb->s_dirt = 0; unlock_kernel(); + unlock_super(sb); } static int sysv_remount(struct super_block *sb, int *flags, char *data) diff --git a/fs/ufs/super.c b/fs/ufs/super.c index a5ecabfdc976..c97210ee0670 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -1131,6 +1131,7 @@ static void ufs_write_super(struct super_block *sb) struct ufs_super_block_third * usb3; unsigned flags; + lock_super(sb); lock_kernel(); UFSD("ENTER\n"); flags = UFS_SB(sb)->s_flags; @@ -1150,6 +1151,7 @@ static void ufs_write_super(struct super_block *sb) sb->s_dirt = 0; UFSD("EXIT\n"); unlock_kernel(); + unlock_super(sb); } static void ufs_put_super(struct super_block *sb) From 9fd5746fd3d7838bf6ff991d50f1257057d1156f Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Thu, 21 May 2009 16:01:00 -0400 Subject: [PATCH 48/87] fs: Remove i_cindex from struct inode The only user of the i_cindex element in the inode structure is used is by the firewire drivers. As part of an attempt to slim down the inode structure to save memory --- since a typical Linux system will have hundreds of thousands if not millions of inodes cached, a reduction in the size inode has high leverage. The firewire driver does not need i_cindex in any fast path, so it's simple enough to calculate when it is needed, instead of wasting space in the inode structure. Signed-off-by: "Theodore Ts'o" Cc: krh@redhat.com Cc: stefanr@s5r6.in-berlin.de Cc: linux-fsdevel@vger.kernel.org Signed-off-by: Al Viro --- drivers/ieee1394/dv1394.c | 5 +++-- drivers/ieee1394/ieee1394_core.h | 6 +++++- fs/char_dev.c | 14 +++++++++++++- include/linux/cdev.h | 2 ++ include/linux/fs.h | 1 - 5 files changed, 23 insertions(+), 5 deletions(-) diff --git a/drivers/ieee1394/dv1394.c b/drivers/ieee1394/dv1394.c index 823a6297a1af..2cd00b5b45b4 100644 --- a/drivers/ieee1394/dv1394.c +++ b/drivers/ieee1394/dv1394.c @@ -1789,12 +1789,13 @@ static int dv1394_open(struct inode *inode, struct file *file) } else { /* look up the card by ID */ unsigned long flags; + int idx = ieee1394_file_to_instance(file); spin_lock_irqsave(&dv1394_cards_lock, flags); if (!list_empty(&dv1394_cards)) { struct video_card *p; list_for_each_entry(p, &dv1394_cards, list) { - if ((p->id) == ieee1394_file_to_instance(file)) { + if ((p->id) == idx) { video = p; break; } @@ -1803,7 +1804,7 @@ static int dv1394_open(struct inode *inode, struct file *file) spin_unlock_irqrestore(&dv1394_cards_lock, flags); if (!video) { - debug_printk("dv1394: OHCI card %d not found", ieee1394_file_to_instance(file)); + debug_printk("dv1394: OHCI card %d not found", idx); return -ENODEV; } diff --git a/drivers/ieee1394/ieee1394_core.h b/drivers/ieee1394/ieee1394_core.h index 21d50f73a210..28b9f58bafd2 100644 --- a/drivers/ieee1394/ieee1394_core.h +++ b/drivers/ieee1394/ieee1394_core.h @@ -5,6 +5,7 @@ #include #include #include +#include #include #include "hosts.h" @@ -155,7 +156,10 @@ void hpsb_packet_received(struct hpsb_host *host, quadlet_t *data, size_t size, */ static inline unsigned char ieee1394_file_to_instance(struct file *file) { - return file->f_path.dentry->d_inode->i_cindex; + int idx = cdev_index(file->f_path.dentry->d_inode); + if (idx < 0) + idx = 0; + return idx; } extern int hpsb_disable_irm; diff --git a/fs/char_dev.c b/fs/char_dev.c index 38f71222a552..b7c9d5187a75 100644 --- a/fs/char_dev.c +++ b/fs/char_dev.c @@ -375,7 +375,6 @@ static int chrdev_open(struct inode *inode, struct file *filp) p = inode->i_cdev; if (!p) { inode->i_cdev = p = new; - inode->i_cindex = idx; list_add(&inode->i_devices, &p->list); new = NULL; } else if (!cdev_get(p)) @@ -405,6 +404,18 @@ static int chrdev_open(struct inode *inode, struct file *filp) return ret; } +int cdev_index(struct inode *inode) +{ + int idx; + struct kobject *kobj; + + kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx); + if (!kobj) + return -1; + kobject_put(kobj); + return idx; +} + void cd_forget(struct inode *inode) { spin_lock(&cdev_lock); @@ -557,6 +568,7 @@ EXPORT_SYMBOL(cdev_init); EXPORT_SYMBOL(cdev_alloc); EXPORT_SYMBOL(cdev_del); EXPORT_SYMBOL(cdev_add); +EXPORT_SYMBOL(cdev_index); EXPORT_SYMBOL(register_chrdev); EXPORT_SYMBOL(unregister_chrdev); EXPORT_SYMBOL(directly_mappable_cdev_bdi); diff --git a/include/linux/cdev.h b/include/linux/cdev.h index fb4591977b03..f389e319a454 100644 --- a/include/linux/cdev.h +++ b/include/linux/cdev.h @@ -28,6 +28,8 @@ int cdev_add(struct cdev *, dev_t, unsigned); void cdev_del(struct cdev *); +int cdev_index(struct inode *inode); + void cd_forget(struct inode *); extern struct backing_dev_info directly_mappable_cdev_bdi; diff --git a/include/linux/fs.h b/include/linux/fs.h index e7833ef5d1d6..bcd63706db87 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -751,7 +751,6 @@ struct inode { struct block_device *i_bdev; struct cdev *i_cdev; }; - int i_cindex; __u32 i_generation; From 28ad0c118b0ed98b042d362acfe0017591921138 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Thu, 21 May 2009 16:01:02 -0400 Subject: [PATCH 49/87] fs: Rearrange inode structure elements to avoid waste due to padding Signed-off-by: "Theodore Ts'o" Cc: linux-fsdevel@vger.kernel.org Signed-off-by: Al Viro --- include/linux/fs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/linux/fs.h b/include/linux/fs.h index bcd63706db87..d883aa1fc2eb 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -729,8 +729,8 @@ struct inode { struct timespec i_atime; struct timespec i_mtime; struct timespec i_ctime; - unsigned int i_blkbits; blkcnt_t i_blocks; + unsigned int i_blkbits; unsigned short i_bytes; umode_t i_mode; spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */ From 13205fb9260c2377438599ef0773c6a3eaeb0b07 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Mon, 25 May 2009 09:30:45 +0200 Subject: [PATCH 50/87] ntfs: remove old debug check for dirty data in ntfs_put_super() This should not trigger anymore, so kill it. Acked-by: Anton Altaparmakov Signed-off-by: Jens Axboe Signed-off-by: Al Viro --- fs/ntfs/super.c | 33 +++------------------------------ 1 file changed, 3 insertions(+), 30 deletions(-) diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c index a9ec4e1084e4..7a7b0d326395 100644 --- a/fs/ntfs/super.c +++ b/fs/ntfs/super.c @@ -2376,39 +2376,12 @@ static void ntfs_put_super(struct super_block *sb) vol->mftmirr_ino = NULL; } /* - * If any dirty inodes are left, throw away all mft data page cache - * pages to allow a clean umount. This should never happen any more - * due to mft.c::ntfs_mft_writepage() cleaning all the dirty pages as - * the underlying mft records are written out and cleaned. If it does, - * happen anyway, we want to know... + * We should have no dirty inodes left, due to + * mft.c::ntfs_mft_writepage() cleaning all the dirty pages as + * the underlying mft records are written out and cleaned. */ ntfs_commit_inode(vol->mft_ino); write_inode_now(vol->mft_ino, 1); - if (sb_has_dirty_inodes(sb)) { - const char *s1, *s2; - - mutex_lock(&vol->mft_ino->i_mutex); - truncate_inode_pages(vol->mft_ino->i_mapping, 0); - mutex_unlock(&vol->mft_ino->i_mutex); - write_inode_now(vol->mft_ino, 1); - if (sb_has_dirty_inodes(sb)) { - static const char *_s1 = "inodes"; - static const char *_s2 = ""; - s1 = _s1; - s2 = _s2; - } else { - static const char *_s1 = "mft pages"; - static const char *_s2 = "They have been thrown " - "away. "; - s1 = _s1; - s2 = _s2; - } - ntfs_error(sb, "Dirty %s found at umount time. %sYou should " - "run chkdsk. Please email " - "linux-ntfs-dev@lists.sourceforge.net and say " - "that you saw this message. Thank you.", s1, - s2); - } #endif /* NTFS_RW */ iput(vol->mft_ino); From f95022161d23ee661a48af8f280472209f513a67 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 5 Jun 2009 12:26:23 +0200 Subject: [PATCH 51/87] xfs: remove ->write_super and stop maintaining ->s_dirt the write_super method is used for (1) writing back the superblock periodically from pdflush (2) called just before ->sync_fs for data integerity syncs We don't need (1) because we have our own peridoc writeout through xfssyncd, and we don't need (2) because xfs_fs_sync_fs performs a proper synchronous superblock writeout after all other data and metadata has been written out. Also remove ->s_dirt tracking as it's only used to decide when too call ->write_super. Signed-off-by: Christoph Hellwig Reviewed-by: Eric Sandeen Signed-off-by: Al Viro --- fs/xfs/linux-2.6/xfs_super.c | 12 ------------ fs/xfs/xfs_trans.c | 2 -- 2 files changed, 14 deletions(-) diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index bb685269f832..08d6bd9a3947 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1104,15 +1104,6 @@ xfs_fs_put_super( kfree(mp); } -STATIC void -xfs_fs_write_super( - struct super_block *sb) -{ - if (!(sb->s_flags & MS_RDONLY)) - xfs_sync_fsdata(XFS_M(sb), 0); - sb->s_dirt = 0; -} - STATIC int xfs_fs_sync_super( struct super_block *sb, @@ -1137,7 +1128,6 @@ xfs_fs_sync_super( error = xfs_quiesce_data(mp); else error = xfs_sync_fsdata(mp, 0); - sb->s_dirt = 0; if (unlikely(laptop_mode)) { int prev_sync_seq = mp->m_sync_seq; @@ -1443,7 +1433,6 @@ xfs_fs_fill_super( XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, mtpt, mp->m_fsname); - sb->s_dirt = 1; sb->s_magic = XFS_SB_MAGIC; sb->s_blocksize = mp->m_sb.sb_blocksize; sb->s_blocksize_bits = ffs(sb->s_blocksize) - 1; @@ -1533,7 +1522,6 @@ static struct super_operations xfs_super_operations = { .write_inode = xfs_fs_write_inode, .clear_inode = xfs_fs_clear_inode, .put_super = xfs_fs_put_super, - .write_super = xfs_fs_write_super, .sync_fs = xfs_fs_sync_super, .freeze_fs = xfs_fs_freeze, .statfs = xfs_fs_statfs, diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 8570b826fedd..bcc39d358ad3 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -628,8 +628,6 @@ xfs_trans_apply_sb_deltas( xfs_trans_log_buf(tp, bp, offsetof(xfs_dsb_t, sb_icount), offsetof(xfs_dsb_t, sb_frextents) + sizeof(sbp->sb_frextents) - 1); - - tp->t_mountp->m_super->s_dirt = 1; } /* From 8688b8635266cf98f00c6b0350ea2dbe7c42c321 Mon Sep 17 00:00:00 2001 From: Mike Frysinger Date: Tue, 26 May 2009 05:45:04 -0400 Subject: [PATCH 52/87] linux/magic.h: move cramfs magic out of cramfs_fs.h Signed-off-by: Mike Frysinger CC: Alexander Viro Signed-off-by: Al Viro --- include/linux/cramfs_fs.h | 3 +-- include/linux/magic.h | 2 ++ 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/include/linux/cramfs_fs.h b/include/linux/cramfs_fs.h index 3be4e5a27d82..6fc2bed368b8 100644 --- a/include/linux/cramfs_fs.h +++ b/include/linux/cramfs_fs.h @@ -2,9 +2,8 @@ #define __CRAMFS_H #include +#include -#define CRAMFS_MAGIC 0x28cd3d45 /* some random number */ -#define CRAMFS_MAGIC_WEND 0x453dcd28 /* magic number with the wrong endianess */ #define CRAMFS_SIGNATURE "Compressed ROMFS" /* diff --git a/include/linux/magic.h b/include/linux/magic.h index 927138cf3050..1923327b9869 100644 --- a/include/linux/magic.h +++ b/include/linux/magic.h @@ -6,6 +6,8 @@ #define AFS_SUPER_MAGIC 0x5346414F #define AUTOFS_SUPER_MAGIC 0x0187 #define CODA_SUPER_MAGIC 0x73757245 +#define CRAMFS_MAGIC 0x28cd3d45 /* some random number */ +#define CRAMFS_MAGIC_WEND 0x453dcd28 /* magic number with the wrong endianess */ #define DEBUGFS_MAGIC 0x64626720 #define SYSFS_MAGIC 0x62656572 #define SECURITYFS_MAGIC 0x73636673 From 545b9fd3d737afc0bb5203b1e79194a471605acd Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Tue, 2 Jun 2009 12:07:47 +0200 Subject: [PATCH 53/87] fs: remove incorrect I_NEW warnings Some filesystems can call in to sync an inode that is still in the I_NEW state (eg. ext family, when mounted with -osync). This is OK because the filesystem has sole access to the new inode, so it can modify i_state without races (because no other thread should be modifying it, by definition of I_NEW). Ie. a false positive, so remove the warnings. The races are described here 7ef0d7377cb287e08f3ae94cebc919448e1f5dff, which is also where the warnings were introduced. Reported-by: Stephen Hemminger Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/fs-writeback.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index e0fb2e789598..efcedb6d9cbc 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -289,7 +289,6 @@ __sync_single_inode(struct inode *inode, struct writeback_control *wbc) int ret; BUG_ON(inode->i_state & I_SYNC); - WARN_ON(inode->i_state & I_NEW); /* Set I_SYNC, reset I_DIRTY */ dirty = inode->i_state & I_DIRTY; @@ -314,7 +313,6 @@ __sync_single_inode(struct inode *inode, struct writeback_control *wbc) } spin_lock(&inode_lock); - WARN_ON(inode->i_state & I_NEW); inode->i_state &= ~I_SYNC; if (!(inode->i_state & I_FREEING)) { if (!(inode->i_state & I_DIRTY) && From 4195f73d1329e49727bcceb028e58cb38376c2b0 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Thu, 28 May 2009 09:01:15 +0200 Subject: [PATCH 54/87] fs: block_dump missing dentry locking I think the block_dump output in __mark_inode_dirty is missing dentry locking. Surely the i_dentry list can change any time, so we may not even *get* a dentry there. If we do get one by chance, then it would appear to be able to go away or get renamed at any time... Signed-off-by: Al Viro --- fs/fs-writeback.c | 41 ++++++++++++++++++++++++----------------- 1 file changed, 24 insertions(+), 17 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index efcedb6d9cbc..40308e98c6a4 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -64,6 +64,28 @@ static void writeback_release(struct backing_dev_info *bdi) clear_bit(BDI_pdflush, &bdi->state); } +static noinline void block_dump___mark_inode_dirty(struct inode *inode) +{ + if (inode->i_ino || strcmp(inode->i_sb->s_id, "bdev")) { + struct dentry *dentry; + const char *name = "?"; + + dentry = d_find_alias(inode); + if (dentry) { + spin_lock(&dentry->d_lock); + name = (const char *) dentry->d_name.name; + } + printk(KERN_DEBUG + "%s(%d): dirtied inode %lu (%s) on %s\n", + current->comm, task_pid_nr(current), inode->i_ino, + name, inode->i_sb->s_id); + if (dentry) { + spin_unlock(&dentry->d_lock); + dput(dentry); + } + } +} + /** * __mark_inode_dirty - internal function * @inode: inode to mark @@ -114,23 +136,8 @@ void __mark_inode_dirty(struct inode *inode, int flags) if ((inode->i_state & flags) == flags) return; - if (unlikely(block_dump)) { - struct dentry *dentry = NULL; - const char *name = "?"; - - if (!list_empty(&inode->i_dentry)) { - dentry = list_entry(inode->i_dentry.next, - struct dentry, d_alias); - if (dentry && dentry->d_name.name) - name = (const char *) dentry->d_name.name; - } - - if (inode->i_ino || strcmp(inode->i_sb->s_id, "bdev")) - printk(KERN_DEBUG - "%s(%d): dirtied inode %lu (%s) on %s\n", - current->comm, task_pid_nr(current), inode->i_ino, - name, inode->i_sb->s_id); - } + if (unlikely(block_dump)) + block_dump___mark_inode_dirty(inode); spin_lock(&inode_lock); if ((inode->i_state & flags) != flags) { From 337eb00a2c3a421999c39c94ce7e33545ee8baa7 Mon Sep 17 00:00:00 2001 From: Alessio Igor Bogani Date: Tue, 12 May 2009 15:10:54 +0200 Subject: [PATCH 55/87] Push BKL down into ->remount_fs() [xfs, btrfs, capifs, shmem don't need BKL, exempt] Signed-off-by: Alessio Igor Bogani Signed-off-by: Al Viro --- drivers/usb/core/inode.c | 5 +++++ fs/affs/super.c | 7 ++++++- fs/ext2/super.c | 12 ++++++++++-- fs/ext3/super.c | 4 ++++ fs/ext4/super.c | 4 ++++ fs/hpfs/super.c | 4 ++++ fs/jffs2/fs.c | 3 +++ fs/jfs/super.c | 22 ++++++++++++++++++---- fs/nfs/super.c | 2 ++ fs/nilfs2/super.c | 4 ++++ fs/ntfs/super.c | 15 ++++++++++++++- fs/ocfs2/super.c | 4 ++++ fs/reiserfs/super.c | 4 ++++ fs/super.c | 2 -- fs/ubifs/super.c | 9 ++++++++- fs/udf/super.c | 6 +++++- fs/ufs/super.c | 11 ++++++++++- kernel/cgroup.c | 3 +++ 18 files changed, 108 insertions(+), 13 deletions(-) diff --git a/drivers/usb/core/inode.c b/drivers/usb/core/inode.c index dff5760a37f6..ffe75e83787c 100644 --- a/drivers/usb/core/inode.c +++ b/drivers/usb/core/inode.c @@ -39,6 +39,7 @@ #include #include #include +#include #include #include "usb.h" #include "hcd.h" @@ -265,9 +266,13 @@ static int remount(struct super_block *sb, int *flags, char *data) return -EINVAL; } + lock_kernel(); + if (usbfs_mount && usbfs_mount->mnt_sb) update_sb(usbfs_mount->mnt_sb); + unlock_kernel(); + return 0; } diff --git a/fs/affs/super.c b/fs/affs/super.c index 280d361af41f..c4814937c968 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -16,6 +16,7 @@ #include #include #include +#include #include "affs.h" extern struct timezone sys_tz; @@ -512,6 +513,7 @@ affs_remount(struct super_block *sb, int *flags, char *data) kfree(new_opts); return -EINVAL; } + lock_kernel(); replace_mount_options(sb, new_opts); sbi->s_flags = mount_flags; @@ -519,8 +521,10 @@ affs_remount(struct super_block *sb, int *flags, char *data) sbi->s_uid = uid; sbi->s_gid = gid; - if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) + if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) { + unlock_kernel(); return 0; + } if (*flags & MS_RDONLY) { sb->s_dirt = 1; while (sb->s_dirt) @@ -529,6 +533,7 @@ affs_remount(struct super_block *sb, int *flags, char *data) } else res = affs_init_bitmap(sb, flags); + unlock_kernel(); return res; } diff --git a/fs/ext2/super.c b/fs/ext2/super.c index a44963d8edbd..f8cbdf569190 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -1162,6 +1162,8 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data) unsigned long old_sb_flags; int err; + lock_kernel(); + /* Store the old options */ old_sb_flags = sb->s_flags; old_opts.s_mount_opt = sbi->s_mount_opt; @@ -1197,12 +1199,16 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data) sbi->s_mount_opt &= ~EXT2_MOUNT_XIP; sbi->s_mount_opt |= old_mount_opt & EXT2_MOUNT_XIP; } - if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) + if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) { + unlock_kernel(); return 0; + } if (*flags & MS_RDONLY) { if (le16_to_cpu(es->s_state) & EXT2_VALID_FS || - !(sbi->s_mount_state & EXT2_VALID_FS)) + !(sbi->s_mount_state & EXT2_VALID_FS)) { + unlock_kernel(); return 0; + } /* * OK, we are remounting a valid rw partition rdonly, so set * the rdonly flag and then mark the partition as valid again. @@ -1229,12 +1235,14 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data) sb->s_flags &= ~MS_RDONLY; } ext2_sync_super(sb, es); + unlock_kernel(); return 0; restore_opts: sbi->s_mount_opt = old_opts.s_mount_opt; sbi->s_resuid = old_opts.s_resuid; sbi->s_resgid = old_opts.s_resgid; sb->s_flags = old_sb_flags; + unlock_kernel(); return err; } diff --git a/fs/ext3/super.c b/fs/ext3/super.c index e213a2613a56..26aa64dee6aa 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -2490,6 +2490,8 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data) int i; #endif + lock_kernel(); + /* Store the original options */ lock_super(sb); old_sb_flags = sb->s_flags; @@ -2600,6 +2602,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data) kfree(old_opts.s_qf_names[i]); #endif unlock_super(sb); + unlock_kernel(); return 0; restore_opts: sb->s_flags = old_sb_flags; @@ -2617,6 +2620,7 @@ restore_opts: } #endif unlock_super(sb); + unlock_kernel(); return err; } diff --git a/fs/ext4/super.c b/fs/ext4/super.c index c17200a42301..012c4251397e 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3422,6 +3422,8 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) int i; #endif + lock_kernel(); + /* Store the original options */ lock_super(sb); old_sb_flags = sb->s_flags; @@ -3558,6 +3560,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data) kfree(old_opts.s_qf_names[i]); #endif unlock_super(sb); + unlock_kernel(); return 0; restore_opts: @@ -3578,6 +3581,7 @@ restore_opts: } #endif unlock_super(sb); + unlock_kernel(); return err; } diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index f68193cf0811..f2feaa06bf26 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -13,6 +13,7 @@ #include #include #include +#include /* Mark the filesystem dirty, so that chkdsk checks it when os/2 booted */ @@ -398,6 +399,7 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) *flags |= MS_NOATIME; + lock_kernel(); lock_super(s); uid = sbi->sb_uid; gid = sbi->sb_gid; umask = 0777 & ~sbi->sb_mode; @@ -432,10 +434,12 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) replace_mount_options(s, new_opts); unlock_super(s); + unlock_kernel(); return 0; out_err: unlock_super(s); + unlock_kernel(); kfree(new_opts); return -EINVAL; } diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c index 237b27a3d570..3451a81b2142 100644 --- a/fs/jffs2/fs.c +++ b/fs/jffs2/fs.c @@ -20,6 +20,7 @@ #include #include #include +#include #include "nodelist.h" static int jffs2_flash_setup(struct jffs2_sb_info *c); @@ -387,6 +388,7 @@ int jffs2_remount_fs (struct super_block *sb, int *flags, char *data) This also catches the case where it was stopped and this is just a remount to restart it. Flush the writebuffer, if neccecary, else we loose it */ + lock_kernel(); if (!(sb->s_flags & MS_RDONLY)) { jffs2_stop_garbage_collect_thread(c); mutex_lock(&c->alloc_sem); @@ -399,6 +401,7 @@ int jffs2_remount_fs (struct super_block *sb, int *flags, char *data) *flags |= MS_NOATIME; + unlock_kernel(); return 0; } diff --git a/fs/jfs/super.c b/fs/jfs/super.c index 3eb13adf3862..09b1b6ee2186 100644 --- a/fs/jfs/super.c +++ b/fs/jfs/super.c @@ -32,6 +32,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_filsys.h" @@ -375,19 +376,24 @@ static int jfs_remount(struct super_block *sb, int *flags, char *data) s64 newLVSize = 0; int rc = 0; int flag = JFS_SBI(sb)->flag; + int ret; if (!parse_options(data, sb, &newLVSize, &flag)) { return -EINVAL; } + lock_kernel(); if (newLVSize) { if (sb->s_flags & MS_RDONLY) { printk(KERN_ERR "JFS: resize requires volume to be mounted read-write\n"); + unlock_kernel(); return -EROFS; } rc = jfs_extendfs(sb, newLVSize, 0); - if (rc) + if (rc) { + unlock_kernel(); return rc; + } } if ((sb->s_flags & MS_RDONLY) && !(*flags & MS_RDONLY)) { @@ -398,23 +404,31 @@ static int jfs_remount(struct super_block *sb, int *flags, char *data) truncate_inode_pages(JFS_SBI(sb)->direct_inode->i_mapping, 0); JFS_SBI(sb)->flag = flag; - return jfs_mount_rw(sb, 1); + ret = jfs_mount_rw(sb, 1); + unlock_kernel(); + return ret; } if ((!(sb->s_flags & MS_RDONLY)) && (*flags & MS_RDONLY)) { rc = jfs_umount_rw(sb); JFS_SBI(sb)->flag = flag; + unlock_kernel(); return rc; } if ((JFS_SBI(sb)->flag & JFS_NOINTEGRITY) != (flag & JFS_NOINTEGRITY)) if (!(sb->s_flags & MS_RDONLY)) { rc = jfs_umount_rw(sb); - if (rc) + if (rc) { + unlock_kernel(); return rc; + } JFS_SBI(sb)->flag = flag; - return jfs_mount_rw(sb, 1); + ret = jfs_mount_rw(sb, 1); + unlock_kernel(); + return ret; } JFS_SBI(sb)->flag = flag; + unlock_kernel(); return 0; } diff --git a/fs/nfs/super.c b/fs/nfs/super.c index d2d67781c579..26127b69a275 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1813,6 +1813,7 @@ nfs_remount(struct super_block *sb, int *flags, char *raw_data) if (data == NULL) return -ENOMEM; + lock_kernel(); /* fill out struct with values from existing mount */ data->flags = nfss->flags; data->rsize = nfss->rsize; @@ -1837,6 +1838,7 @@ nfs_remount(struct super_block *sb, int *flags, char *raw_data) error = nfs_compare_remount_data(nfss, data); out: kfree(data); + unlock_kernel(); return error; } diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 7262e8427c20..11151eaa2c4a 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -906,6 +906,8 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) struct nilfs_mount_options old_opts; int err; + lock_kernel(); + old_sb_flags = sb->s_flags; old_opts.mount_opt = sbi->s_mount_opt; old_opts.snapshot_cno = sbi->s_snapshot_cno; @@ -985,6 +987,7 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) up(&sb->s_bdev->bd_mount_sem); } out: + unlock_kernel(); return 0; rw_remount_failed: @@ -993,6 +996,7 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) sb->s_flags = old_sb_flags; sbi->s_mount_opt = old_opts.mount_opt; sbi->s_snapshot_cno = old_opts.snapshot_cno; + unlock_kernel(); return err; } diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c index 7a7b0d326395..abaaa1cbf8de 100644 --- a/fs/ntfs/super.c +++ b/fs/ntfs/super.c @@ -443,6 +443,8 @@ static int ntfs_remount(struct super_block *sb, int *flags, char *opt) ntfs_volume *vol = NTFS_SB(sb); ntfs_debug("Entering with remount options string: %s", opt); + + lock_kernel(); #ifndef NTFS_RW /* For read-only compiled driver, enforce read-only flag. */ *flags |= MS_RDONLY; @@ -466,15 +468,18 @@ static int ntfs_remount(struct super_block *sb, int *flags, char *opt) if (NVolErrors(vol)) { ntfs_error(sb, "Volume has errors and is read-only%s", es); + unlock_kernel(); return -EROFS; } if (vol->vol_flags & VOLUME_IS_DIRTY) { ntfs_error(sb, "Volume is dirty and read-only%s", es); + unlock_kernel(); return -EROFS; } if (vol->vol_flags & VOLUME_MODIFIED_BY_CHKDSK) { ntfs_error(sb, "Volume has been modified by chkdsk " "and is read-only%s", es); + unlock_kernel(); return -EROFS; } if (vol->vol_flags & VOLUME_MUST_MOUNT_RO_MASK) { @@ -482,11 +487,13 @@ static int ntfs_remount(struct super_block *sb, int *flags, char *opt) "(0x%x) and is read-only%s", (unsigned)le16_to_cpu(vol->vol_flags), es); + unlock_kernel(); return -EROFS; } if (ntfs_set_volume_flags(vol, VOLUME_IS_DIRTY)) { ntfs_error(sb, "Failed to set dirty bit in volume " "information flags%s", es); + unlock_kernel(); return -EROFS; } #if 0 @@ -506,18 +513,21 @@ static int ntfs_remount(struct super_block *sb, int *flags, char *opt) ntfs_error(sb, "Failed to empty journal $LogFile%s", es); NVolSetErrors(vol); + unlock_kernel(); return -EROFS; } if (!ntfs_mark_quotas_out_of_date(vol)) { ntfs_error(sb, "Failed to mark quotas out of date%s", es); NVolSetErrors(vol); + unlock_kernel(); return -EROFS; } if (!ntfs_stamp_usnjrnl(vol)) { ntfs_error(sb, "Failed to stamp transation log " "($UsnJrnl)%s", es); NVolSetErrors(vol); + unlock_kernel(); return -EROFS; } } else if (!(sb->s_flags & MS_RDONLY) && (*flags & MS_RDONLY)) { @@ -533,8 +543,11 @@ static int ntfs_remount(struct super_block *sb, int *flags, char *opt) // TODO: Deal with *flags. - if (!parse_options(vol, opt)) + if (!parse_options(vol, opt)) { + unlock_kernel(); return -EINVAL; + } + unlock_kernel(); ntfs_debug("Done."); return 0; } diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 02737596b597..201b40a441fe 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -42,6 +42,7 @@ #include #include #include +#include #define MLOG_MASK_PREFIX ML_SUPER #include @@ -581,6 +582,8 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data) struct mount_options parsed_options; struct ocfs2_super *osb = OCFS2_SB(sb); + lock_kernel(); + if (!ocfs2_parse_options(sb, data, &parsed_options, 1)) { ret = -EINVAL; goto out; @@ -684,6 +687,7 @@ unlock_osb: ocfs2_set_journal_params(osb); } out: + unlock_kernel(); return ret; } diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 90dcb7b033ea..2969773cfc22 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -28,6 +28,7 @@ #include #include #include +#include struct file_system_type reiserfs_fs_type; @@ -1196,6 +1197,7 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) memcpy(qf_names, REISERFS_SB(s)->s_qf_names, sizeof(qf_names)); #endif + lock_kernel(); rs = SB_DISK_SUPER_BLOCK(s); if (!reiserfs_parse_options @@ -1318,10 +1320,12 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) out_ok: replace_mount_options(s, new_opts); + unlock_kernel(); return 0; out_err: kfree(new_opts); + unlock_kernel(); return err; } diff --git a/fs/super.c b/fs/super.c index 1905f4af01cc..83b47416d006 100644 --- a/fs/super.c +++ b/fs/super.c @@ -540,7 +540,6 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) shrink_dcache_sb(sb); sync_filesystem(sb); - lock_kernel(); /* If we are remounting RDONLY and current sb is read/write, make sure there are no rw files opened */ if ((flags & MS_RDONLY) && !(sb->s_flags & MS_RDONLY)) { @@ -566,7 +565,6 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) } } sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK); - unlock_kernel(); if (remount_rw) vfs_dq_quota_on_remount(sb); return 0; diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 522c3fd7eb3c..3589eab02a2f 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -36,6 +36,7 @@ #include #include #include +#include #include "ubifs.h" /* @@ -1770,17 +1771,22 @@ static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data) return err; } + lock_kernel(); if ((sb->s_flags & MS_RDONLY) && !(*flags & MS_RDONLY)) { if (c->ro_media) { ubifs_msg("cannot re-mount due to prior errors"); + unlock_kernel(); return -EROFS; } err = ubifs_remount_rw(c); - if (err) + if (err) { + unlock_kernel(); return err; + } } else if (!(sb->s_flags & MS_RDONLY) && (*flags & MS_RDONLY)) { if (c->ro_media) { ubifs_msg("cannot re-mount due to prior errors"); + unlock_kernel(); return -EROFS; } ubifs_remount_ro(c); @@ -1795,6 +1801,7 @@ static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data) } ubifs_assert(c->lst.taken_empty_lebs > 0); + unlock_kernel(); return 0; } diff --git a/fs/udf/super.c b/fs/udf/super.c index 04802cc39b18..6832135159b6 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -568,6 +568,7 @@ static int udf_remount_fs(struct super_block *sb, int *flags, char *options) if (!udf_parse_options(options, &uopt, true)) return -EINVAL; + lock_kernel(); sbi->s_flags = uopt.flags; sbi->s_uid = uopt.uid; sbi->s_gid = uopt.gid; @@ -581,13 +582,16 @@ static int udf_remount_fs(struct super_block *sb, int *flags, char *options) *flags |= MS_RDONLY; } - if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) + if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) { + unlock_kernel(); return 0; + } if (*flags & MS_RDONLY) udf_close_lvid(sb); else udf_open_lvid(sb); + unlock_kernel(); return 0; } diff --git a/fs/ufs/super.c b/fs/ufs/super.c index c97210ee0670..6560dda7b18c 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -263,6 +263,7 @@ void ufs_panic (struct super_block * sb, const char * function, struct ufs_super_block_first * usb1; va_list args; + lock_kernel(); uspi = UFS_SB(sb)->s_uspi; usb1 = ubh_get_usb_first(uspi); @@ -1182,7 +1183,8 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) struct ufs_super_block_third * usb3; unsigned new_mount_opt, ufstype; unsigned flags; - + + lock_kernel(); lock_super(sb); uspi = UFS_SB(sb)->s_uspi; flags = UFS_SB(sb)->s_flags; @@ -1198,6 +1200,7 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) ufs_set_opt (new_mount_opt, ONERROR_LOCK); if (!ufs_parse_options (data, &new_mount_opt)) { unlock_super(sb); + unlock_kernel(); return -EINVAL; } if (!(new_mount_opt & UFS_MOUNT_UFSTYPE)) { @@ -1205,12 +1208,14 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) } else if ((new_mount_opt & UFS_MOUNT_UFSTYPE) != ufstype) { printk("ufstype can't be changed during remount\n"); unlock_super(sb); + unlock_kernel(); return -EINVAL; } if ((*mount_flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) { UFS_SB(sb)->s_mount_opt = new_mount_opt; unlock_super(sb); + unlock_kernel(); return 0; } @@ -1236,6 +1241,7 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) printk("ufs was compiled with read-only support, " "can't be mounted as read-write\n"); unlock_super(sb); + unlock_kernel(); return -EINVAL; #else if (ufstype != UFS_MOUNT_UFSTYPE_SUN && @@ -1245,11 +1251,13 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) ufstype != UFS_MOUNT_UFSTYPE_UFS2) { printk("this ufstype is read-only supported\n"); unlock_super(sb); + unlock_kernel(); return -EINVAL; } if (!ufs_read_cylinder_structures(sb)) { printk("failed during remounting\n"); unlock_super(sb); + unlock_kernel(); return -EPERM; } sb->s_flags &= ~MS_RDONLY; @@ -1257,6 +1265,7 @@ static int ufs_remount (struct super_block *sb, int *mount_flags, char *data) } UFS_SB(sb)->s_mount_opt = new_mount_opt; unlock_super(sb); + unlock_kernel(); return 0; } diff --git a/kernel/cgroup.c b/kernel/cgroup.c index a7267bfd3765..3fb789f6df94 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -46,6 +46,7 @@ #include #include #include +#include #include @@ -900,6 +901,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) struct cgroup *cgrp = &root->top_cgroup; struct cgroup_sb_opts opts; + lock_kernel(); mutex_lock(&cgrp->dentry->d_inode->i_mutex); mutex_lock(&cgroup_mutex); @@ -927,6 +929,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) kfree(opts.release_agent); mutex_unlock(&cgroup_mutex); mutex_unlock(&cgrp->dentry->d_inode->i_mutex); + unlock_kernel(); return ret; } From d5aacad548db1ff547adf35d0a77eb2a8ed4fe14 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 14:56:44 -0400 Subject: [PATCH 56/87] New helper - simple_fsync() writes associated buffers, then does sync_inode() to write the inode itself (and to make it clean). Depends on ->write_inode() honouring the second argument. Signed-off-by: Al Viro --- fs/libfs.c | 25 +++++++++++++++++++++++++ include/linux/fs.h | 2 ++ 2 files changed, 27 insertions(+) diff --git a/fs/libfs.c b/fs/libfs.c index 80046ddf5063..ddfa89948c3f 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -9,6 +9,8 @@ #include #include #include +#include +#include #include @@ -807,6 +809,29 @@ struct dentry *generic_fh_to_parent(struct super_block *sb, struct fid *fid, } EXPORT_SYMBOL_GPL(generic_fh_to_parent); +int simple_fsync(struct file *file, struct dentry *dentry, int datasync) +{ + struct writeback_control wbc = { + .sync_mode = WB_SYNC_ALL, + .nr_to_write = 0, /* metadata-only; caller takes care of data */ + }; + struct inode *inode = dentry->d_inode; + int err; + int ret; + + ret = sync_mapping_buffers(inode->i_mapping); + if (!(inode->i_state & I_DIRTY)) + return ret; + if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) + return ret; + + err = sync_inode(inode, &wbc); + if (ret == 0) + ret = err; + return ret; +} +EXPORT_SYMBOL(simple_fsync); + EXPORT_SYMBOL(dcache_dir_close); EXPORT_SYMBOL(dcache_dir_lseek); EXPORT_SYMBOL(dcache_dir_open); diff --git a/include/linux/fs.h b/include/linux/fs.h index d883aa1fc2eb..ede84fa7da5d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2345,6 +2345,8 @@ extern void simple_release_fs(struct vfsmount **mount, int *count); extern ssize_t simple_read_from_buffer(void __user *to, size_t count, loff_t *ppos, const void *from, size_t available); +extern int simple_fsync(struct file *, struct dentry *, int); + #ifdef CONFIG_MIGRATION extern int buffer_migrate_page(struct address_space *, struct page *, struct page *); From 79d25767583e4e086f8309bfd1f502660a64fe7f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 09:30:08 -0400 Subject: [PATCH 57/87] Sanitize qnx4 fsync handling * have directory operations use mark_buffer_dirty_inode(), so that sync_mapping_buffers() would get those. * make qnx4_write_inode() honour its last argument. * get rid of insane copies of very ancient "walk the indirect blocks" in qnx4/fsync - they never matched the actual fs layout and, fortunately, never'd been called. Again, all this junk is not needed; ->fsync() should just do sync_mapping_buffers + sync_inode (and if we implement block allocation for qnx4, we'll need to use mark_buffer_dirty_inode() for extent blocks) Signed-off-by: Al Viro --- fs/qnx4/Makefile | 2 +- fs/qnx4/dir.c | 2 +- fs/qnx4/file.c | 2 +- fs/qnx4/fsync.c | 169 ---------------------------------------- fs/qnx4/inode.c | 38 +++------ fs/qnx4/namei.c | 4 +- include/linux/qnx4_fs.h | 2 - 7 files changed, 17 insertions(+), 202 deletions(-) delete mode 100644 fs/qnx4/fsync.c diff --git a/fs/qnx4/Makefile b/fs/qnx4/Makefile index 502d7fe98bab..e4d408cc5473 100644 --- a/fs/qnx4/Makefile +++ b/fs/qnx4/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_QNX4FS_FS) += qnx4.o -qnx4-objs := inode.o dir.o namei.o file.o bitmap.o truncate.o fsync.o +qnx4-objs := inode.o dir.o namei.o file.o bitmap.o truncate.o diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c index ea9ffefb48ad..ff6c1ba6c4e0 100644 --- a/fs/qnx4/dir.c +++ b/fs/qnx4/dir.c @@ -84,7 +84,7 @@ const struct file_operations qnx4_dir_operations = { .read = generic_read_dir, .readdir = qnx4_readdir, - .fsync = file_fsync, + .fsync = simple_fsync, }; const struct inode_operations qnx4_dir_inode_operations = diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c index 867f42b02035..e7033ea10e2f 100644 --- a/fs/qnx4/file.c +++ b/fs/qnx4/file.c @@ -29,7 +29,7 @@ const struct file_operations qnx4_file_operations = #ifdef CONFIG_QNX4FS_RW .write = do_sync_write, .aio_write = generic_file_aio_write, - .fsync = qnx4_sync_file, + .fsync = simple_fsync, #endif }; diff --git a/fs/qnx4/fsync.c b/fs/qnx4/fsync.c deleted file mode 100644 index aa3b19544bee..000000000000 --- a/fs/qnx4/fsync.c +++ /dev/null @@ -1,169 +0,0 @@ -/* - * QNX4 file system, Linux implementation. - * - * Version : 0.1 - * - * Using parts of the xiafs filesystem. - * - * History : - * - * 24-03-1998 by Richard Frowijn : first release. - */ - -#include -#include -#include -#include -#include -#include - -#include -#include - -#include - -/* - * The functions for qnx4 fs file synchronization. - */ - -#ifdef CONFIG_QNX4FS_RW - -static int sync_block(struct inode *inode, unsigned short *block, int wait) -{ - struct buffer_head *bh; - unsigned short tmp; - - if (!*block) - return 0; - tmp = *block; - bh = sb_find_get_block(inode->i_sb, *block); - if (!bh) - return 0; - if (*block != tmp) { - brelse(bh); - return 1; - } - if (wait && buffer_req(bh) && !buffer_uptodate(bh)) { - brelse(bh); - return -1; - } - if (wait || !buffer_uptodate(bh) || !buffer_dirty(bh)) { - brelse(bh); - return 0; - } - ll_rw_block(WRITE, 1, &bh); - atomic_dec(&bh->b_count); - return 0; -} - -#ifdef WTF -static int sync_iblock(struct inode *inode, unsigned short *iblock, - struct buffer_head **bh, int wait) -{ - int rc; - unsigned short tmp; - - *bh = NULL; - tmp = *iblock; - if (!tmp) - return 0; - rc = sync_block(inode, iblock, wait); - if (rc) - return rc; - *bh = sb_bread(inode->i_sb, tmp); - if (tmp != *iblock) { - brelse(*bh); - *bh = NULL; - return 1; - } - if (!*bh) - return -1; - return 0; -} -#endif - -static int sync_direct(struct inode *inode, int wait) -{ - int i; - int rc, err = 0; - - for (i = 0; i < 7; i++) { - rc = sync_block(inode, - (unsigned short *) qnx4_raw_inode(inode)->di_first_xtnt.xtnt_blk + i, wait); - if (rc > 0) - break; - if (rc) - err = rc; - } - return err; -} - -#ifdef WTF -static int sync_indirect(struct inode *inode, unsigned short *iblock, int wait) -{ - int i; - struct buffer_head *ind_bh; - int rc, err = 0; - - rc = sync_iblock(inode, iblock, &ind_bh, wait); - if (rc || !ind_bh) - return rc; - - for (i = 0; i < 512; i++) { - rc = sync_block(inode, - ((unsigned short *) ind_bh->b_data) + i, - wait); - if (rc > 0) - break; - if (rc) - err = rc; - } - brelse(ind_bh); - return err; -} - -static int sync_dindirect(struct inode *inode, unsigned short *diblock, - int wait) -{ - int i; - struct buffer_head *dind_bh; - int rc, err = 0; - - rc = sync_iblock(inode, diblock, &dind_bh, wait); - if (rc || !dind_bh) - return rc; - - for (i = 0; i < 512; i++) { - rc = sync_indirect(inode, - ((unsigned short *) dind_bh->b_data) + i, - wait); - if (rc > 0) - break; - if (rc) - err = rc; - } - brelse(dind_bh); - return err; -} -#endif - -int qnx4_sync_file(struct file *file, struct dentry *dentry, int unused) -{ - struct inode *inode = dentry->d_inode; - int wait, err = 0; - - (void) file; - if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || - S_ISLNK(inode->i_mode))) - return -EINVAL; - - lock_kernel(); - for (wait = 0; wait <= 1; wait++) { - err |= sync_direct(inode, wait); - } - err |= qnx4_sync_inode(inode); - unlock_kernel(); - return (err < 0) ? -EIO : 0; -} - -#endif diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c index 95c12fc613f1..40712867b8a8 100644 --- a/fs/qnx4/inode.c +++ b/fs/qnx4/inode.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include @@ -34,31 +35,6 @@ static const struct super_operations qnx4_sops; #ifdef CONFIG_QNX4FS_RW -int qnx4_sync_inode(struct inode *inode) -{ - int err = 0; -# if 0 - struct buffer_head *bh; - - bh = qnx4_update_inode(inode); - if (bh && buffer_dirty(bh)) - { - sync_dirty_buffer(bh); - if (buffer_req(bh) && !buffer_uptodate(bh)) - { - printk ("IO error syncing qnx4 inode [%s:%08lx]\n", - inode->i_sb->s_id, inode->i_ino); - err = -1; - } - brelse (bh); - } else if (!bh) { - err = -1; - } -# endif - - return err; -} - static void qnx4_delete_inode(struct inode *inode) { QNX4DEBUG(("qnx4: deleting inode [%lu]\n", (unsigned long) inode->i_ino)); @@ -70,7 +46,7 @@ static void qnx4_delete_inode(struct inode *inode) unlock_kernel(); } -static int qnx4_write_inode(struct inode *inode, int unused) +static int qnx4_write_inode(struct inode *inode, int do_sync) { struct qnx4_inode_entry *raw_inode; int block, ino; @@ -107,6 +83,16 @@ static int qnx4_write_inode(struct inode *inode, int unused) raw_inode->di_ctime = cpu_to_le32(inode->i_ctime.tv_sec); raw_inode->di_first_xtnt.xtnt_size = cpu_to_le32(inode->i_blocks); mark_buffer_dirty(bh); + if (do_sync) { + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) { + printk("qnx4: IO error syncing inode [%s:%08x]\n", + inode->i_sb->s_id, ino); + brelse(bh); + unlock_kernel(); + return -EIO; + } + } brelse(bh); unlock_kernel(); return 0; diff --git a/fs/qnx4/namei.c b/fs/qnx4/namei.c index 775eed3a4085..123270c53760 100644 --- a/fs/qnx4/namei.c +++ b/fs/qnx4/namei.c @@ -187,7 +187,7 @@ int qnx4_rmdir(struct inode *dir, struct dentry *dentry) de->di_status = 0; memset(de->di_fname, 0, sizeof de->di_fname); de->di_mode = 0; - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); clear_nlink(inode); mark_inode_dirty(inode); inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC; @@ -232,7 +232,7 @@ int qnx4_unlink(struct inode *dir, struct dentry *dentry) de->di_status = 0; memset(de->di_fname, 0, sizeof de->di_fname); de->di_mode = 0; - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC; mark_inode_dirty(dir); inode->i_ctime = dir->i_ctime; diff --git a/include/linux/qnx4_fs.h b/include/linux/qnx4_fs.h index 787d19ea9f46..acbaec3524e0 100644 --- a/include/linux/qnx4_fs.h +++ b/include/linux/qnx4_fs.h @@ -126,8 +126,6 @@ extern void qnx4_truncate(struct inode *inode); extern void qnx4_free_inode(struct inode *inode); extern int qnx4_unlink(struct inode *dir, struct dentry *dentry); extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry); -extern int qnx4_sync_file(struct file *file, struct dentry *dentry, int); -extern int qnx4_sync_inode(struct inode *inode); static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb) { From 964f5369667b342994fe3f384e9ba41d404ee796 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 09:47:13 -0400 Subject: [PATCH 58/87] fs/qnx4: sanitize includes fs-internal parts of qnx4_fs.h taken to fs/qnx4/qnx4.h, includes adjusted, qnx4_fs.h doesn't need unifdef anymore. Signed-off-by: Al Viro --- fs/qnx4/bitmap.c | 7 +---- fs/qnx4/dir.c | 7 +---- fs/qnx4/file.c | 3 +-- fs/qnx4/inode.c | 11 +++----- fs/qnx4/namei.c | 9 +------ fs/qnx4/qnx4.h | 57 +++++++++++++++++++++++++++++++++++++++ fs/qnx4/truncate.c | 6 +---- include/linux/Kbuild | 2 +- include/linux/qnx4_fs.h | 59 ----------------------------------------- 9 files changed, 66 insertions(+), 95 deletions(-) create mode 100644 fs/qnx4/qnx4.h diff --git a/fs/qnx4/bitmap.c b/fs/qnx4/bitmap.c index 8425cf6e9624..e1cd061a25f7 100644 --- a/fs/qnx4/bitmap.c +++ b/fs/qnx4/bitmap.c @@ -13,14 +13,9 @@ * 28-06-1998 by Frank Denis : qnx4_free_inode (to be fixed) . */ -#include -#include -#include -#include -#include -#include #include #include +#include "qnx4.h" #if 0 int qnx4_new_block(struct super_block *sb) diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c index ff6c1ba6c4e0..003c68f3238b 100644 --- a/fs/qnx4/dir.c +++ b/fs/qnx4/dir.c @@ -11,14 +11,9 @@ * 20-06-1998 by Frank Denis : Linux 2.1.99+ & dcache support. */ -#include -#include -#include -#include -#include #include #include - +#include "qnx4.h" static int qnx4_readdir(struct file *filp, void *dirent, filldir_t filldir) { diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c index e7033ea10e2f..09b170ac936c 100644 --- a/fs/qnx4/file.c +++ b/fs/qnx4/file.c @@ -12,8 +12,7 @@ * 27-06-1998 by Frank Denis : file overwriting. */ -#include -#include +#include "qnx4.h" /* * We have mostly NULL's here: the current defaults are ok for diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c index 40712867b8a8..681df5fcd161 100644 --- a/fs/qnx4/inode.c +++ b/fs/qnx4/inode.c @@ -13,20 +13,15 @@ */ #include -#include -#include -#include -#include -#include -#include #include +#include #include #include #include #include #include -#include -#include +#include +#include "qnx4.h" #define QNX4_VERSION 4 #define QNX4_BMNAME ".bitmap" diff --git a/fs/qnx4/namei.c b/fs/qnx4/namei.c index 123270c53760..5972ed214937 100644 --- a/fs/qnx4/namei.c +++ b/fs/qnx4/namei.c @@ -12,16 +12,9 @@ * 04-07-1998 by Frank Denis : first step for rmdir/unlink. */ -#include -#include -#include -#include -#include -#include -#include -#include #include #include +#include "qnx4.h" /* diff --git a/fs/qnx4/qnx4.h b/fs/qnx4/qnx4.h new file mode 100644 index 000000000000..9efc089454f6 --- /dev/null +++ b/fs/qnx4/qnx4.h @@ -0,0 +1,57 @@ +#include +#include + +#define QNX4_DEBUG 0 + +#if QNX4_DEBUG +#define QNX4DEBUG(X) printk X +#else +#define QNX4DEBUG(X) (void) 0 +#endif + +struct qnx4_sb_info { + struct buffer_head *sb_buf; /* superblock buffer */ + struct qnx4_super_block *sb; /* our superblock */ + unsigned int Version; /* may be useful */ + struct qnx4_inode_entry *BitMap; /* useful */ +}; + +struct qnx4_inode_info { + struct qnx4_inode_entry raw; + loff_t mmu_private; + struct inode vfs_inode; +}; + +extern struct inode *qnx4_iget(struct super_block *, unsigned long); +extern struct dentry *qnx4_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd); +extern unsigned long qnx4_count_free_blocks(struct super_block *sb); +extern unsigned long qnx4_block_map(struct inode *inode, long iblock); + +extern struct buffer_head *qnx4_bread(struct inode *, int, int); + +extern const struct inode_operations qnx4_file_inode_operations; +extern const struct inode_operations qnx4_dir_inode_operations; +extern const struct file_operations qnx4_file_operations; +extern const struct file_operations qnx4_dir_operations; +extern int qnx4_is_free(struct super_block *sb, long block); +extern int qnx4_set_bitmap(struct super_block *sb, long block, int busy); +extern int qnx4_create(struct inode *inode, struct dentry *dentry, int mode, struct nameidata *nd); +extern void qnx4_truncate(struct inode *inode); +extern void qnx4_free_inode(struct inode *inode); +extern int qnx4_unlink(struct inode *dir, struct dentry *dentry); +extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry); + +static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb) +{ + return sb->s_fs_info; +} + +static inline struct qnx4_inode_info *qnx4_i(struct inode *inode) +{ + return container_of(inode, struct qnx4_inode_info, vfs_inode); +} + +static inline struct qnx4_inode_entry *qnx4_raw_inode(struct inode *inode) +{ + return &qnx4_i(inode)->raw; +} diff --git a/fs/qnx4/truncate.c b/fs/qnx4/truncate.c index 6437c1c3d1dd..d94d9ee241fe 100644 --- a/fs/qnx4/truncate.c +++ b/fs/qnx4/truncate.c @@ -10,12 +10,8 @@ * 30-06-1998 by Frank DENIS : ugly filler. */ -#include -#include -#include -#include #include -#include +#include "qnx4.h" #ifdef CONFIG_QNX4FS_RW diff --git a/include/linux/Kbuild b/include/linux/Kbuild index 3f0eaa397ef5..b3afd2219ad2 100644 --- a/include/linux/Kbuild +++ b/include/linux/Kbuild @@ -135,6 +135,7 @@ header-y += posix_types.h header-y += ppdev.h header-y += prctl.h header-y += qnxtypes.h +header-y += qnx4_fs.h header-y += radeonfb.h header-y += raw.h header-y += resource.h @@ -308,7 +309,6 @@ unifdef-y += poll.h unifdef-y += ppp_defs.h unifdef-y += ppp-comp.h unifdef-y += ptrace.h -unifdef-y += qnx4_fs.h unifdef-y += quota.h unifdef-y += random.h unifdef-y += irqnr.h diff --git a/include/linux/qnx4_fs.h b/include/linux/qnx4_fs.h index acbaec3524e0..8b9aee1a9ce3 100644 --- a/include/linux/qnx4_fs.h +++ b/include/linux/qnx4_fs.h @@ -85,63 +85,4 @@ struct qnx4_super_block { struct qnx4_inode_entry AltBoot; }; -#ifdef __KERNEL__ - -#define QNX4_DEBUG 0 - -#if QNX4_DEBUG -#define QNX4DEBUG(X) printk X -#else -#define QNX4DEBUG(X) (void) 0 -#endif - -struct qnx4_sb_info { - struct buffer_head *sb_buf; /* superblock buffer */ - struct qnx4_super_block *sb; /* our superblock */ - unsigned int Version; /* may be useful */ - struct qnx4_inode_entry *BitMap; /* useful */ -}; - -struct qnx4_inode_info { - struct qnx4_inode_entry raw; - loff_t mmu_private; - struct inode vfs_inode; -}; - -extern struct inode *qnx4_iget(struct super_block *, unsigned long); -extern struct dentry *qnx4_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd); -extern unsigned long qnx4_count_free_blocks(struct super_block *sb); -extern unsigned long qnx4_block_map(struct inode *inode, long iblock); - -extern struct buffer_head *qnx4_bread(struct inode *, int, int); - -extern const struct inode_operations qnx4_file_inode_operations; -extern const struct inode_operations qnx4_dir_inode_operations; -extern const struct file_operations qnx4_file_operations; -extern const struct file_operations qnx4_dir_operations; -extern int qnx4_is_free(struct super_block *sb, long block); -extern int qnx4_set_bitmap(struct super_block *sb, long block, int busy); -extern int qnx4_create(struct inode *inode, struct dentry *dentry, int mode, struct nameidata *nd); -extern void qnx4_truncate(struct inode *inode); -extern void qnx4_free_inode(struct inode *inode); -extern int qnx4_unlink(struct inode *dir, struct dentry *dentry); -extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry); - -static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb) -{ - return sb->s_fs_info; -} - -static inline struct qnx4_inode_info *qnx4_i(struct inode *inode) -{ - return container_of(inode, struct qnx4_inode_info, vfs_inode); -} - -static inline struct qnx4_inode_entry *qnx4_raw_inode(struct inode *inode) -{ - return &qnx4_i(inode)->raw; -} - -#endif /* __KERNEL__ */ - #endif From b522412aeabadbb302fd4338eaabf09d10e2d29c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 13:44:36 -0400 Subject: [PATCH 59/87] Sanitize ->fsync() for FAT * mark directory data blocks as assoc. metadata * add new inode to deal with FAT, mark FAT blocks as assoc. metadata of that * now ->fsync() is trivial both for files and directories Signed-off-by: Al Viro --- fs/fat/dir.c | 16 ++++++++-------- fs/fat/fat.h | 6 ++++++ fs/fat/fatent.c | 13 ++++++++----- fs/fat/file.c | 14 +++++++++++++- fs/fat/inode.c | 11 ++++++++++- fs/fat/namei_msdos.c | 4 ++-- fs/fat/namei_vfat.c | 4 ++-- 7 files changed, 49 insertions(+), 19 deletions(-) diff --git a/fs/fat/dir.c b/fs/fat/dir.c index 3a7f603b6982..f3500294eec5 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -840,7 +840,7 @@ const struct file_operations fat_dir_operations = { #ifdef CONFIG_COMPAT .compat_ioctl = fat_compat_dir_ioctl, #endif - .fsync = file_fsync, + .fsync = fat_file_fsync, }; static int fat_get_short_entry(struct inode *dir, loff_t *pos, @@ -967,7 +967,7 @@ static int __fat_remove_entries(struct inode *dir, loff_t pos, int nr_slots) de++; nr_slots--; } - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); if (IS_DIRSYNC(dir)) err = sync_dirty_buffer(bh); brelse(bh); @@ -1001,7 +1001,7 @@ int fat_remove_entries(struct inode *dir, struct fat_slot_info *sinfo) de--; nr_slots--; } - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); if (IS_DIRSYNC(dir)) err = sync_dirty_buffer(bh); brelse(bh); @@ -1051,7 +1051,7 @@ static int fat_zeroed_cluster(struct inode *dir, sector_t blknr, int nr_used, } memset(bhs[n]->b_data, 0, sb->s_blocksize); set_buffer_uptodate(bhs[n]); - mark_buffer_dirty(bhs[n]); + mark_buffer_dirty_inode(bhs[n], dir); n++; blknr++; @@ -1131,7 +1131,7 @@ int fat_alloc_new_dir(struct inode *dir, struct timespec *ts) de[0].size = de[1].size = 0; memset(de + 2, 0, sb->s_blocksize - 2 * sizeof(*de)); set_buffer_uptodate(bhs[0]); - mark_buffer_dirty(bhs[0]); + mark_buffer_dirty_inode(bhs[0], dir); err = fat_zeroed_cluster(dir, blknr, 1, bhs, MAX_BUF_PER_PAGE); if (err) @@ -1193,7 +1193,7 @@ static int fat_add_new_entries(struct inode *dir, void *slots, int nr_slots, slots += copy; size -= copy; set_buffer_uptodate(bhs[n]); - mark_buffer_dirty(bhs[n]); + mark_buffer_dirty_inode(bhs[n], dir); if (!size) break; n++; @@ -1293,7 +1293,7 @@ found: for (i = 0; i < long_bhs; i++) { int copy = min_t(int, sb->s_blocksize - offset, size); memcpy(bhs[i]->b_data + offset, slots, copy); - mark_buffer_dirty(bhs[i]); + mark_buffer_dirty_inode(bhs[i], dir); offset = 0; slots += copy; size -= copy; @@ -1304,7 +1304,7 @@ found: /* Fill the short name slot. */ int copy = min_t(int, sb->s_blocksize - offset, size); memcpy(bhs[i]->b_data + offset, slots, copy); - mark_buffer_dirty(bhs[i]); + mark_buffer_dirty_inode(bhs[i], dir); if (IS_DIRSYNC(dir)) err = sync_dirty_buffer(bhs[i]); } diff --git a/fs/fat/fat.h b/fs/fat/fat.h index ea440d65819c..e4d88527b5dd 100644 --- a/fs/fat/fat.h +++ b/fs/fat/fat.h @@ -74,6 +74,7 @@ struct msdos_sb_info { int fatent_shift; struct fatent_operations *fatent_ops; + struct inode *fat_inode; spinlock_t inode_hash_lock; struct hlist_head inode_hashtable[FAT_HASH_SIZE]; @@ -251,6 +252,7 @@ struct fat_entry { } u; int nr_bhs; struct buffer_head *bhs[2]; + struct inode *fat_inode; }; static inline void fatent_init(struct fat_entry *fatent) @@ -259,6 +261,7 @@ static inline void fatent_init(struct fat_entry *fatent) fatent->entry = 0; fatent->u.ent32_p = NULL; fatent->bhs[0] = fatent->bhs[1] = NULL; + fatent->fat_inode = NULL; } static inline void fatent_set_entry(struct fat_entry *fatent, int entry) @@ -275,6 +278,7 @@ static inline void fatent_brelse(struct fat_entry *fatent) brelse(fatent->bhs[i]); fatent->nr_bhs = 0; fatent->bhs[0] = fatent->bhs[1] = NULL; + fatent->fat_inode = NULL; } extern void fat_ent_access_init(struct super_block *sb); @@ -296,6 +300,8 @@ extern int fat_setattr(struct dentry * dentry, struct iattr * attr); extern void fat_truncate(struct inode *inode); extern int fat_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); +extern int fat_file_fsync(struct file *file, struct dentry *dentry, + int datasync); /* fat/inode.c */ extern void fat_attach(struct inode *inode, loff_t i_pos); diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c index da6eea47872f..618f5305c2e4 100644 --- a/fs/fat/fatent.c +++ b/fs/fat/fatent.c @@ -73,6 +73,8 @@ static int fat12_ent_bread(struct super_block *sb, struct fat_entry *fatent, struct buffer_head **bhs = fatent->bhs; WARN_ON(blocknr < MSDOS_SB(sb)->fat_start); + fatent->fat_inode = MSDOS_SB(sb)->fat_inode; + bhs[0] = sb_bread(sb, blocknr); if (!bhs[0]) goto err; @@ -103,6 +105,7 @@ static int fat_ent_bread(struct super_block *sb, struct fat_entry *fatent, struct fatent_operations *ops = MSDOS_SB(sb)->fatent_ops; WARN_ON(blocknr < MSDOS_SB(sb)->fat_start); + fatent->fat_inode = MSDOS_SB(sb)->fat_inode; fatent->bhs[0] = sb_bread(sb, blocknr); if (!fatent->bhs[0]) { printk(KERN_ERR "FAT: FAT read failed (blocknr %llu)\n", @@ -167,9 +170,9 @@ static void fat12_ent_put(struct fat_entry *fatent, int new) } spin_unlock(&fat12_entry_lock); - mark_buffer_dirty(fatent->bhs[0]); + mark_buffer_dirty_inode(fatent->bhs[0], fatent->fat_inode); if (fatent->nr_bhs == 2) - mark_buffer_dirty(fatent->bhs[1]); + mark_buffer_dirty_inode(fatent->bhs[1], fatent->fat_inode); } static void fat16_ent_put(struct fat_entry *fatent, int new) @@ -178,7 +181,7 @@ static void fat16_ent_put(struct fat_entry *fatent, int new) new = EOF_FAT16; *fatent->u.ent16_p = cpu_to_le16(new); - mark_buffer_dirty(fatent->bhs[0]); + mark_buffer_dirty_inode(fatent->bhs[0], fatent->fat_inode); } static void fat32_ent_put(struct fat_entry *fatent, int new) @@ -189,7 +192,7 @@ static void fat32_ent_put(struct fat_entry *fatent, int new) WARN_ON(new & 0xf0000000); new |= le32_to_cpu(*fatent->u.ent32_p) & ~0x0fffffff; *fatent->u.ent32_p = cpu_to_le32(new); - mark_buffer_dirty(fatent->bhs[0]); + mark_buffer_dirty_inode(fatent->bhs[0], fatent->fat_inode); } static int fat12_ent_next(struct fat_entry *fatent) @@ -381,7 +384,7 @@ static int fat_mirror_bhs(struct super_block *sb, struct buffer_head **bhs, } memcpy(c_bh->b_data, bhs[n]->b_data, sb->s_blocksize); set_buffer_uptodate(c_bh); - mark_buffer_dirty(c_bh); + mark_buffer_dirty_inode(c_bh, sbi->fat_inode); if (sb->s_flags & MS_SYNCHRONOUS) err = sync_dirty_buffer(c_bh); brelse(c_bh); diff --git a/fs/fat/file.c b/fs/fat/file.c index 0a7f4a9918b3..e955a56b4e5e 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -133,6 +133,18 @@ static int fat_file_release(struct inode *inode, struct file *filp) return 0; } +int fat_file_fsync(struct file *filp, struct dentry *dentry, int datasync) +{ + struct inode *inode = dentry->d_inode; + int res, err; + + res = simple_fsync(filp, dentry, datasync); + err = sync_mapping_buffers(MSDOS_SB(inode->i_sb)->fat_inode->i_mapping); + + return res ? res : err; +} + + const struct file_operations fat_file_operations = { .llseek = generic_file_llseek, .read = do_sync_read, @@ -142,7 +154,7 @@ const struct file_operations fat_file_operations = { .mmap = generic_file_mmap, .release = fat_file_release, .ioctl = fat_generic_ioctl, - .fsync = file_fsync, + .fsync = fat_file_fsync, .splice_read = generic_file_splice_read, }; diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 2292cbf7d364..476f80b175fe 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -458,6 +458,8 @@ static void fat_put_super(struct super_block *sb) if (sb->s_dirt) fat_write_super(sb); + iput(sbi->fat_inode); + if (sbi->nls_disk) { unload_nls(sbi->nls_disk); sbi->nls_disk = NULL; @@ -1183,7 +1185,7 @@ static int fat_read_root(struct inode *inode) int fat_fill_super(struct super_block *sb, void *data, int silent, const struct inode_operations *fs_dir_inode_ops, int isvfat) { - struct inode *root_inode = NULL; + struct inode *root_inode = NULL, *fat_inode = NULL; struct buffer_head *bh; struct fat_boot_sector *b; struct msdos_sb_info *sbi; @@ -1423,6 +1425,11 @@ int fat_fill_super(struct super_block *sb, void *data, int silent, } error = -ENOMEM; + fat_inode = new_inode(sb); + if (!fat_inode) + goto out_fail; + MSDOS_I(fat_inode)->i_pos = 0; + sbi->fat_inode = fat_inode; root_inode = new_inode(sb); if (!root_inode) goto out_fail; @@ -1448,6 +1455,8 @@ out_invalid: " on dev %s.\n", sb->s_id); out_fail: + if (fat_inode) + iput(fat_inode); if (root_inode) iput(root_inode); if (sbi->nls_io) diff --git a/fs/fat/namei_msdos.c b/fs/fat/namei_msdos.c index da3f361a37dd..20f522861355 100644 --- a/fs/fat/namei_msdos.c +++ b/fs/fat/namei_msdos.c @@ -544,7 +544,7 @@ static int do_msdos_rename(struct inode *old_dir, unsigned char *old_name, int start = MSDOS_I(new_dir)->i_logstart; dotdot_de->start = cpu_to_le16(start); dotdot_de->starthi = cpu_to_le16(start >> 16); - mark_buffer_dirty(dotdot_bh); + mark_buffer_dirty_inode(dotdot_bh, old_inode); if (IS_DIRSYNC(new_dir)) { err = sync_dirty_buffer(dotdot_bh); if (err) @@ -586,7 +586,7 @@ error_dotdot: int start = MSDOS_I(old_dir)->i_logstart; dotdot_de->start = cpu_to_le16(start); dotdot_de->starthi = cpu_to_le16(start >> 16); - mark_buffer_dirty(dotdot_bh); + mark_buffer_dirty_inode(dotdot_bh, old_inode); corrupt |= sync_dirty_buffer(dotdot_bh); } error_inode: diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c index a0e00e3a46e9..b50ecbe97f83 100644 --- a/fs/fat/namei_vfat.c +++ b/fs/fat/namei_vfat.c @@ -965,7 +965,7 @@ static int vfat_rename(struct inode *old_dir, struct dentry *old_dentry, int start = MSDOS_I(new_dir)->i_logstart; dotdot_de->start = cpu_to_le16(start); dotdot_de->starthi = cpu_to_le16(start >> 16); - mark_buffer_dirty(dotdot_bh); + mark_buffer_dirty_inode(dotdot_bh, old_inode); if (IS_DIRSYNC(new_dir)) { err = sync_dirty_buffer(dotdot_bh); if (err) @@ -1009,7 +1009,7 @@ error_dotdot: int start = MSDOS_I(old_dir)->i_logstart; dotdot_de->start = cpu_to_le16(start); dotdot_de->starthi = cpu_to_le16(start >> 16); - mark_buffer_dirty(dotdot_bh); + mark_buffer_dirty_inode(dotdot_bh, old_inode); corrupt |= sync_dirty_buffer(dotdot_bh); } error_inode: From e1740a462ecb2eae213be15857b577cc6f6bb8b4 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 15:14:02 -0400 Subject: [PATCH 60/87] switch ext2 to simple_fsync() kill ext2_sync_file() (along with ext2/fsync.c), get rid of ext2_update_inode() - it's an alias of ext2_write_inode(). Signed-off-by: Al Viro --- fs/ext2/Makefile | 2 +- fs/ext2/dir.c | 2 +- fs/ext2/ext2.h | 3 --- fs/ext2/file.c | 4 ++-- fs/ext2/fsync.c | 50 ------------------------------------------------ fs/ext2/inode.c | 11 ++--------- 6 files changed, 6 insertions(+), 66 deletions(-) delete mode 100644 fs/ext2/fsync.c diff --git a/fs/ext2/Makefile b/fs/ext2/Makefile index e0b2b43c1fdb..f42af45cfd88 100644 --- a/fs/ext2/Makefile +++ b/fs/ext2/Makefile @@ -4,7 +4,7 @@ obj-$(CONFIG_EXT2_FS) += ext2.o -ext2-y := balloc.o dir.o file.o fsync.o ialloc.o inode.o \ +ext2-y := balloc.o dir.o file.o ialloc.o inode.o \ ioctl.o namei.o super.o symlink.o ext2-$(CONFIG_EXT2_FS_XATTR) += xattr.o xattr_user.o xattr_trusted.o diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c index 2999d72153b7..003500498c22 100644 --- a/fs/ext2/dir.c +++ b/fs/ext2/dir.c @@ -720,5 +720,5 @@ const struct file_operations ext2_dir_operations = { #ifdef CONFIG_COMPAT .compat_ioctl = ext2_compat_ioctl, #endif - .fsync = ext2_sync_file, + .fsync = simple_fsync, }; diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h index 3203042b36ef..b2bbf45039e0 100644 --- a/fs/ext2/ext2.h +++ b/fs/ext2/ext2.h @@ -113,9 +113,6 @@ extern int ext2_empty_dir (struct inode *); extern struct ext2_dir_entry_2 * ext2_dotdot (struct inode *, struct page **); extern void ext2_set_link(struct inode *, struct ext2_dir_entry_2 *, struct page *, struct inode *); -/* fsync.c */ -extern int ext2_sync_file (struct file *, struct dentry *, int); - /* ialloc.c */ extern struct inode * ext2_new_inode (struct inode *, int); extern void ext2_free_inode (struct inode *); diff --git a/fs/ext2/file.c b/fs/ext2/file.c index 45ed07122182..2b9e47dc9222 100644 --- a/fs/ext2/file.c +++ b/fs/ext2/file.c @@ -55,7 +55,7 @@ const struct file_operations ext2_file_operations = { .mmap = generic_file_mmap, .open = generic_file_open, .release = ext2_release_file, - .fsync = ext2_sync_file, + .fsync = simple_fsync, .splice_read = generic_file_splice_read, .splice_write = generic_file_splice_write, }; @@ -72,7 +72,7 @@ const struct file_operations ext2_xip_file_operations = { .mmap = xip_file_mmap, .open = generic_file_open, .release = ext2_release_file, - .fsync = ext2_sync_file, + .fsync = simple_fsync, }; #endif diff --git a/fs/ext2/fsync.c b/fs/ext2/fsync.c deleted file mode 100644 index fc66c93fcb5c..000000000000 --- a/fs/ext2/fsync.c +++ /dev/null @@ -1,50 +0,0 @@ -/* - * linux/fs/ext2/fsync.c - * - * Copyright (C) 1993 Stephen Tweedie (sct@dcs.ed.ac.uk) - * from - * Copyright (C) 1992 Remy Card (card@masi.ibp.fr) - * Laboratoire MASI - Institut Blaise Pascal - * Universite Pierre et Marie Curie (Paris VI) - * from - * linux/fs/minix/truncate.c Copyright (C) 1991, 1992 Linus Torvalds - * - * ext2fs fsync primitive - * - * Big-endian to little-endian byte-swapping/bitmaps by - * David S. Miller (davem@caip.rutgers.edu), 1995 - * - * Removed unnecessary code duplication for little endian machines - * and excessive __inline__s. - * Andi Kleen, 1997 - * - * Major simplications and cleanup - we only need to do the metadata, because - * we can depend on generic_block_fdatasync() to sync the data blocks. - */ - -#include "ext2.h" -#include /* for sync_mapping_buffers() */ - - -/* - * File may be NULL when we are called. Perhaps we shouldn't - * even pass file to fsync ? - */ - -int ext2_sync_file(struct file *file, struct dentry *dentry, int datasync) -{ - struct inode *inode = dentry->d_inode; - int err; - int ret; - - ret = sync_mapping_buffers(inode->i_mapping); - if (!(inode->i_state & I_DIRTY)) - return ret; - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return ret; - - err = ext2_sync_inode(inode); - if (ret == 0) - ret = err; - return ret; -} diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index acf678831103..29ed682061f6 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -41,8 +41,6 @@ MODULE_AUTHOR("Remy Card and others"); MODULE_DESCRIPTION("Second Extended Filesystem"); MODULE_LICENSE("GPL"); -static int ext2_update_inode(struct inode * inode, int do_sync); - /* * Test whether an inode is a fast symlink. */ @@ -66,7 +64,7 @@ void ext2_delete_inode (struct inode * inode) goto no_delete; EXT2_I(inode)->i_dtime = get_seconds(); mark_inode_dirty(inode); - ext2_update_inode(inode, inode_needs_sync(inode)); + ext2_write_inode(inode, inode_needs_sync(inode)); inode->i_size = 0; if (inode->i_blocks) @@ -1337,7 +1335,7 @@ bad_inode: return ERR_PTR(ret); } -static int ext2_update_inode(struct inode * inode, int do_sync) +int ext2_write_inode(struct inode *inode, int do_sync) { struct ext2_inode_info *ei = EXT2_I(inode); struct super_block *sb = inode->i_sb; @@ -1442,11 +1440,6 @@ static int ext2_update_inode(struct inode * inode, int do_sync) return err; } -int ext2_write_inode(struct inode *inode, int wait) -{ - return ext2_update_inode(inode, wait); -} - int ext2_sync_inode(struct inode *inode) { struct writeback_control wbc = { From 0d7916d7e985da52cdd2989c900485e17b035972 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 15:21:06 -0400 Subject: [PATCH 61/87] switch minix to simple_fsync() * get minix_write_inode() to honour the second argument * now we can use simple_fsync() for minixfs Signed-off-by: Al Viro --- fs/minix/dir.c | 2 +- fs/minix/file.c | 20 +------------------- fs/minix/inode.c | 33 ++++++++++----------------------- fs/minix/minix.h | 2 -- 4 files changed, 12 insertions(+), 45 deletions(-) diff --git a/fs/minix/dir.c b/fs/minix/dir.c index d4946c4c90e2..e5f206467e40 100644 --- a/fs/minix/dir.c +++ b/fs/minix/dir.c @@ -22,7 +22,7 @@ static int minix_readdir(struct file *, void *, filldir_t); const struct file_operations minix_dir_operations = { .read = generic_read_dir, .readdir = minix_readdir, - .fsync = minix_sync_file, + .fsync = simple_fsync, }; static inline void dir_put_page(struct page *page) diff --git a/fs/minix/file.c b/fs/minix/file.c index 17765f697e50..3eec3e607a87 100644 --- a/fs/minix/file.c +++ b/fs/minix/file.c @@ -6,15 +6,12 @@ * minix regular file handling primitives */ -#include /* for fsync_inode_buffers() */ #include "minix.h" /* * We have mostly NULLs here: the current defaults are OK for * the minix filesystem. */ -int minix_sync_file(struct file *, struct dentry *, int); - const struct file_operations minix_file_operations = { .llseek = generic_file_llseek, .read = do_sync_read, @@ -22,7 +19,7 @@ const struct file_operations minix_file_operations = { .write = do_sync_write, .aio_write = generic_file_aio_write, .mmap = generic_file_mmap, - .fsync = minix_sync_file, + .fsync = simple_fsync, .splice_read = generic_file_splice_read, }; @@ -30,18 +27,3 @@ const struct inode_operations minix_file_inode_operations = { .truncate = minix_truncate, .getattr = minix_getattr, }; - -int minix_sync_file(struct file * file, struct dentry *dentry, int datasync) -{ - struct inode *inode = dentry->d_inode; - int err; - - err = sync_mapping_buffers(inode->i_mapping); - if (!(inode->i_state & I_DIRTY)) - return err; - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return err; - - err |= minix_sync_inode(inode); - return err ? -EIO : 0; -} diff --git a/fs/minix/inode.c b/fs/minix/inode.c index 7eb53970f4bc..f91a23693597 100644 --- a/fs/minix/inode.c +++ b/fs/minix/inode.c @@ -556,38 +556,25 @@ static struct buffer_head * V2_minix_update_inode(struct inode * inode) return bh; } -static struct buffer_head *minix_update_inode(struct inode *inode) -{ - if (INODE_VERSION(inode) == MINIX_V1) - return V1_minix_update_inode(inode); - else - return V2_minix_update_inode(inode); -} - -static int minix_write_inode(struct inode * inode, int wait) -{ - brelse(minix_update_inode(inode)); - return 0; -} - -int minix_sync_inode(struct inode * inode) +static int minix_write_inode(struct inode *inode, int wait) { int err = 0; struct buffer_head *bh; - bh = minix_update_inode(inode); - if (bh && buffer_dirty(bh)) - { + if (INODE_VERSION(inode) == MINIX_V1) + bh = V1_minix_update_inode(inode); + else + bh = V2_minix_update_inode(inode); + if (!bh) + return -EIO; + if (wait && buffer_dirty(bh)) { sync_dirty_buffer(bh); - if (buffer_req(bh) && !buffer_uptodate(bh)) - { + if (buffer_req(bh) && !buffer_uptodate(bh)) { printk("IO error syncing minix inode [%s:%08lx]\n", inode->i_sb->s_id, inode->i_ino); - err = -1; + err = -EIO; } } - else if (!bh) - err = -1; brelse (bh); return err; } diff --git a/fs/minix/minix.h b/fs/minix/minix.h index e6a0b193bea4..cb7fdd11f9a5 100644 --- a/fs/minix/minix.h +++ b/fs/minix/minix.h @@ -57,7 +57,6 @@ extern int __minix_write_begin(struct file *file, struct address_space *mapping, extern void V1_minix_truncate(struct inode *); extern void V2_minix_truncate(struct inode *); extern void minix_truncate(struct inode *); -extern int minix_sync_inode(struct inode *); extern void minix_set_inode(struct inode *, dev_t); extern int V1_minix_get_block(struct inode *, long, struct buffer_head *, int); extern int V2_minix_get_block(struct inode *, long, struct buffer_head *, int); @@ -72,7 +71,6 @@ extern int minix_empty_dir(struct inode*); extern void minix_set_link(struct minix_dir_entry*, struct page*, struct inode*); extern struct minix_dir_entry *minix_dotdot(struct inode*, struct page**); extern ino_t minix_inode_by_name(struct dentry*); -extern int minix_sync_file(struct file *, struct dentry *, int); extern const struct inode_operations minix_file_inode_operations; extern const struct inode_operations minix_dir_inode_operations; From 05459ca81ac3064cb040d983342bc453cccec458 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 15:29:45 -0400 Subject: [PATCH 62/87] repair sysv_write_inode(), switch sysv to simple_fsync() Signed-off-by: Al Viro --- fs/sysv/dir.c | 2 +- fs/sysv/file.c | 17 +---------------- fs/sysv/inode.c | 47 +++++++++++++++++------------------------------ fs/sysv/sysv.h | 1 - 4 files changed, 19 insertions(+), 48 deletions(-) diff --git a/fs/sysv/dir.c b/fs/sysv/dir.c index 56f655254bfe..c7798079e644 100644 --- a/fs/sysv/dir.c +++ b/fs/sysv/dir.c @@ -24,7 +24,7 @@ static int sysv_readdir(struct file *, void *, filldir_t); const struct file_operations sysv_dir_operations = { .read = generic_read_dir, .readdir = sysv_readdir, - .fsync = sysv_sync_file, + .fsync = simple_fsync, }; static inline void dir_put_page(struct page *page) diff --git a/fs/sysv/file.c b/fs/sysv/file.c index 589be21d884e..96340c01f4a7 100644 --- a/fs/sysv/file.c +++ b/fs/sysv/file.c @@ -26,7 +26,7 @@ const struct file_operations sysv_file_operations = { .write = do_sync_write, .aio_write = generic_file_aio_write, .mmap = generic_file_mmap, - .fsync = sysv_sync_file, + .fsync = simple_fsync, .splice_read = generic_file_splice_read, }; @@ -34,18 +34,3 @@ const struct inode_operations sysv_file_inode_operations = { .truncate = sysv_truncate, .getattr = sysv_getattr, }; - -int sysv_sync_file(struct file * file, struct dentry *dentry, int datasync) -{ - struct inode *inode = dentry->d_inode; - int err; - - err = sync_mapping_buffers(inode->i_mapping); - if (!(inode->i_state & I_DIRTY)) - return err; - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return err; - - err |= sysv_sync_inode(inode); - return err ? -EIO : 0; -} diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index a3f45fc626a1..425c976cfcd2 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -247,7 +247,7 @@ bad_inode: return ERR_PTR(-EIO); } -static struct buffer_head * sysv_update_inode(struct inode * inode) +int sysv_write_inode(struct inode *inode, int wait) { struct super_block * sb = inode->i_sb; struct sysv_sb_info * sbi = SYSV_SB(sb); @@ -255,19 +255,21 @@ static struct buffer_head * sysv_update_inode(struct inode * inode) struct sysv_inode * raw_inode; struct sysv_inode_info * si; unsigned int ino, block; + int err = 0; ino = inode->i_ino; if (!ino || ino > sbi->s_ninodes) { printk("Bad inode number on dev %s: %d is out of range\n", inode->i_sb->s_id, ino); - return NULL; + return -EIO; } raw_inode = sysv_raw_inode(sb, ino, &bh); if (!raw_inode) { printk("unable to read i-node block\n"); - return NULL; + return -EIO; } + lock_kernel(); raw_inode->i_mode = cpu_to_fs16(sbi, inode->i_mode); raw_inode->i_uid = cpu_to_fs16(sbi, fs_high2lowuid(inode->i_uid)); raw_inode->i_gid = cpu_to_fs16(sbi, fs_high2lowgid(inode->i_gid)); @@ -283,38 +285,23 @@ static struct buffer_head * sysv_update_inode(struct inode * inode) for (block = 0; block < 10+1+1+1; block++) write3byte(sbi, (u8 *)&si->i_data[block], &raw_inode->i_data[3*block]); - mark_buffer_dirty(bh); - return bh; -} - -int sysv_write_inode(struct inode * inode, int wait) -{ - struct buffer_head *bh; - lock_kernel(); - bh = sysv_update_inode(inode); - brelse(bh); unlock_kernel(); + mark_buffer_dirty(bh); + if (wait) { + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) { + printk ("IO error syncing sysv inode [%s:%08x]\n", + sb->s_id, ino); + err = -EIO; + } + } + brelse(bh); return 0; } -int sysv_sync_inode(struct inode * inode) +int sysv_sync_inode(struct inode *inode) { - int err = 0; - struct buffer_head *bh; - - bh = sysv_update_inode(inode); - if (bh && buffer_dirty(bh)) { - sync_dirty_buffer(bh); - if (buffer_req(bh) && !buffer_uptodate(bh)) { - printk ("IO error syncing sysv inode [%s:%08lx]\n", - inode->i_sb->s_id, inode->i_ino); - err = -1; - } - } - else if (!bh) - err = -1; - brelse (bh); - return err; + return sysv_write_inode(inode, 1); } static void sysv_delete_inode(struct inode *inode) diff --git a/fs/sysv/sysv.h b/fs/sysv/sysv.h index 5784a318c883..53786eb5cf60 100644 --- a/fs/sysv/sysv.h +++ b/fs/sysv/sysv.h @@ -144,7 +144,6 @@ extern int __sysv_write_begin(struct file *file, struct address_space *mapping, extern struct inode *sysv_iget(struct super_block *, unsigned int); extern int sysv_write_inode(struct inode *, int); extern int sysv_sync_inode(struct inode *); -extern int sysv_sync_file(struct file *, struct dentry *, int); extern void sysv_set_inode(struct inode *, dev_t); extern int sysv_getattr(struct vfsmount *, struct dentry *, struct kstat *); extern int sysv_init_icache(void); From a932801543fe74050ebee07fde082234c46b624f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 15:35:18 -0400 Subject: [PATCH 63/87] switch ufs to simple_fsync() Signed-off-by: Al Viro --- fs/ufs/dir.c | 2 +- fs/ufs/file.c | 23 +---------------------- fs/ufs/ufs.h | 1 - 3 files changed, 2 insertions(+), 24 deletions(-) diff --git a/fs/ufs/dir.c b/fs/ufs/dir.c index 6321b797061b..6f671f1ac271 100644 --- a/fs/ufs/dir.c +++ b/fs/ufs/dir.c @@ -666,6 +666,6 @@ not_empty: const struct file_operations ufs_dir_operations = { .read = generic_read_dir, .readdir = ufs_readdir, - .fsync = ufs_sync_file, + .fsync = simple_fsync, .llseek = generic_file_llseek, }; diff --git a/fs/ufs/file.c b/fs/ufs/file.c index 2bd3a1615714..73655c61240a 100644 --- a/fs/ufs/file.c +++ b/fs/ufs/file.c @@ -24,31 +24,10 @@ */ #include -#include /* for sync_mapping_buffers() */ #include "ufs_fs.h" #include "ufs.h" - -int ufs_sync_file(struct file *file, struct dentry *dentry, int datasync) -{ - struct inode *inode = dentry->d_inode; - int err; - int ret; - - ret = sync_mapping_buffers(inode->i_mapping); - if (!(inode->i_state & I_DIRTY)) - return ret; - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return ret; - - err = ufs_sync_inode(inode); - if (ret == 0) - ret = err; - return ret; -} - - /* * We have mostly NULL's here: the current defaults are ok for * the ufs filesystem. @@ -62,6 +41,6 @@ const struct file_operations ufs_file_operations = { .aio_write = generic_file_aio_write, .mmap = generic_file_mmap, .open = generic_file_open, - .fsync = ufs_sync_file, + .fsync = simple_fsync, .splice_read = generic_file_splice_read, }; diff --git a/fs/ufs/ufs.h b/fs/ufs/ufs.h index d0c4acd4f1f3..644e77e13599 100644 --- a/fs/ufs/ufs.h +++ b/fs/ufs/ufs.h @@ -99,7 +99,6 @@ extern void ufs_set_link(struct inode *dir, struct ufs_dir_entry *de, extern const struct inode_operations ufs_file_inode_operations; extern const struct file_operations ufs_file_operations; extern const struct address_space_operations ufs_aops; -extern int ufs_sync_file(struct file *, struct dentry *, int); /* ialloc.c */ extern void ufs_free_inode (struct inode *inode); From 90de066443a8632bb42fed0a8216313d7da07aba Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 15:40:27 -0400 Subject: [PATCH 64/87] switch udf to simple_fsync() Signed-off-by: Al Viro --- fs/udf/Makefile | 2 +- fs/udf/dir.c | 2 +- fs/udf/file.c | 2 +- fs/udf/fsync.c | 52 ------------------------------------------------ fs/udf/udfdecl.h | 3 --- 5 files changed, 3 insertions(+), 58 deletions(-) delete mode 100644 fs/udf/fsync.c diff --git a/fs/udf/Makefile b/fs/udf/Makefile index 0d4503f7446d..eb880f66c23a 100644 --- a/fs/udf/Makefile +++ b/fs/udf/Makefile @@ -5,5 +5,5 @@ obj-$(CONFIG_UDF_FS) += udf.o udf-objs := balloc.o dir.o file.o ialloc.o inode.o lowlevel.o namei.o \ - partition.o super.o truncate.o symlink.o fsync.o \ + partition.o super.o truncate.o symlink.o \ directory.o misc.o udftime.o unicode.o diff --git a/fs/udf/dir.c b/fs/udf/dir.c index 2efd4d5291b6..61d9a76a3a69 100644 --- a/fs/udf/dir.c +++ b/fs/udf/dir.c @@ -210,5 +210,5 @@ const struct file_operations udf_dir_operations = { .read = generic_read_dir, .readdir = udf_readdir, .ioctl = udf_ioctl, - .fsync = udf_fsync_file, + .fsync = simple_fsync, }; diff --git a/fs/udf/file.c b/fs/udf/file.c index eb91f3b70320..7464305382b5 100644 --- a/fs/udf/file.c +++ b/fs/udf/file.c @@ -209,7 +209,7 @@ const struct file_operations udf_file_operations = { .write = do_sync_write, .aio_write = udf_file_aio_write, .release = udf_release_file, - .fsync = udf_fsync_file, + .fsync = simple_fsync, .splice_read = generic_file_splice_read, .llseek = generic_file_llseek, }; diff --git a/fs/udf/fsync.c b/fs/udf/fsync.c deleted file mode 100644 index b2c472b733b8..000000000000 --- a/fs/udf/fsync.c +++ /dev/null @@ -1,52 +0,0 @@ -/* - * fsync.c - * - * PURPOSE - * Fsync handling routines for the OSTA-UDF(tm) filesystem. - * - * COPYRIGHT - * This file is distributed under the terms of the GNU General Public - * License (GPL). Copies of the GPL can be obtained from: - * ftp://prep.ai.mit.edu/pub/gnu/GPL - * Each contributing author retains all rights to their own work. - * - * (C) 1999-2001 Ben Fennema - * (C) 1999-2000 Stelias Computing Inc - * - * HISTORY - * - * 05/22/99 blf Created. - */ - -#include "udfdecl.h" - -#include - -static int udf_fsync_inode(struct inode *, int); - -/* - * File may be NULL when we are called. Perhaps we shouldn't - * even pass file to fsync ? - */ - -int udf_fsync_file(struct file *file, struct dentry *dentry, int datasync) -{ - struct inode *inode = dentry->d_inode; - - return udf_fsync_inode(inode, datasync); -} - -static int udf_fsync_inode(struct inode *inode, int datasync) -{ - int err; - - err = sync_mapping_buffers(inode->i_mapping); - if (!(inode->i_state & I_DIRTY)) - return err; - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return err; - - err |= udf_sync_inode(inode); - - return err ? -EIO : 0; -} diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index cac51b77a5d1..8d46f4294ee7 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -223,9 +223,6 @@ extern int udf_prealloc_blocks(struct super_block *, struct inode *, uint16_t, extern int udf_new_block(struct super_block *, struct inode *, uint16_t, uint32_t, int *); -/* fsync.c */ -extern int udf_fsync_file(struct file *, struct dentry *, int); - /* directory.c */ extern struct fileIdentDesc *udf_fileident_read(struct inode *, loff_t *, struct udf_fileident_bh *, From bea6b64c277f0824cdaea6190209b26a164419d6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 7 Jun 2009 15:44:50 -0400 Subject: [PATCH 65/87] switch omfs to simple_fsync() Signed-off-by: Al Viro --- fs/omfs/file.c | 17 +---------------- 1 file changed, 1 insertion(+), 16 deletions(-) diff --git a/fs/omfs/file.c b/fs/omfs/file.c index 834b2331f6b3..d17e774eaf45 100644 --- a/fs/omfs/file.c +++ b/fs/omfs/file.c @@ -11,21 +11,6 @@ #include #include "omfs.h" -static int omfs_sync_file(struct file *file, struct dentry *dentry, - int datasync) -{ - struct inode *inode = dentry->d_inode; - int err; - - err = sync_mapping_buffers(inode->i_mapping); - if (!(inode->i_state & I_DIRTY)) - return err; - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) - return err; - err |= omfs_sync_inode(inode); - return err ? -EIO : 0; -} - static u32 omfs_max_extents(struct omfs_sb_info *sbi, int offset) { return (sbi->s_sys_blocksize - offset - @@ -344,7 +329,7 @@ struct file_operations omfs_file_operations = { .aio_read = generic_file_aio_read, .aio_write = generic_file_aio_write, .mmap = generic_file_mmap, - .fsync = omfs_sync_file, + .fsync = simple_fsync, .splice_read = generic_file_splice_read, }; From ffdc9064f8b4fa9db37a7d5180f41cce2ea2b7ad Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 8 Jun 2009 00:44:42 -0400 Subject: [PATCH 66/87] repair adfs ->write_inode(), switch to simple_fsync() Signed-off-by: Al Viro --- fs/adfs/adfs.h | 4 +++- fs/adfs/dir.c | 10 ++++++++-- fs/adfs/dir_f.c | 17 +++++++++++++++++ fs/adfs/dir_fplus.c | 17 +++++++++++++++++ fs/adfs/file.c | 2 +- fs/adfs/inode.c | 4 ++-- 6 files changed, 48 insertions(+), 6 deletions(-) diff --git a/fs/adfs/adfs.h b/fs/adfs/adfs.h index e0a85dbeeb88..a6665f37f456 100644 --- a/fs/adfs/adfs.h +++ b/fs/adfs/adfs.h @@ -53,6 +53,7 @@ struct adfs_dir_ops { int (*update)(struct adfs_dir *dir, struct object_info *obj); int (*create)(struct adfs_dir *dir, struct object_info *obj); int (*remove)(struct adfs_dir *dir, struct object_info *obj); + int (*sync)(struct adfs_dir *dir); void (*free)(struct adfs_dir *dir); }; @@ -90,7 +91,8 @@ extern const struct dentry_operations adfs_dentry_operations; extern struct adfs_dir_ops adfs_f_dir_ops; extern struct adfs_dir_ops adfs_fplus_dir_ops; -extern int adfs_dir_update(struct super_block *sb, struct object_info *obj); +extern int adfs_dir_update(struct super_block *sb, struct object_info *obj, + int wait); /* file.c */ extern const struct inode_operations adfs_file_inode_operations; diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c index e867ccf37246..4d4073447d1a 100644 --- a/fs/adfs/dir.c +++ b/fs/adfs/dir.c @@ -83,7 +83,7 @@ out: } int -adfs_dir_update(struct super_block *sb, struct object_info *obj) +adfs_dir_update(struct super_block *sb, struct object_info *obj, int wait) { int ret = -EINVAL; #ifdef CONFIG_ADFS_FS_RW @@ -106,6 +106,12 @@ adfs_dir_update(struct super_block *sb, struct object_info *obj) ret = ops->update(&dir, obj); write_unlock(&adfs_dir_lock); + if (wait) { + int err = ops->sync(&dir); + if (!ret) + ret = err; + } + ops->free(&dir); out: #endif @@ -199,7 +205,7 @@ const struct file_operations adfs_dir_operations = { .read = generic_read_dir, .llseek = generic_file_llseek, .readdir = adfs_readdir, - .fsync = file_fsync, + .fsync = simple_fsync, }; static int diff --git a/fs/adfs/dir_f.c b/fs/adfs/dir_f.c index ea7df2146921..31df6adf0de6 100644 --- a/fs/adfs/dir_f.c +++ b/fs/adfs/dir_f.c @@ -437,6 +437,22 @@ bad_dir: #endif } +static int +adfs_f_sync(struct adfs_dir *dir) +{ + int err = 0; + int i; + + for (i = dir->nr_buffers - 1; i >= 0; i--) { + struct buffer_head *bh = dir->bh[i]; + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) + err = -EIO; + } + + return err; +} + static void adfs_f_free(struct adfs_dir *dir) { @@ -456,5 +472,6 @@ struct adfs_dir_ops adfs_f_dir_ops = { .setpos = adfs_f_setpos, .getnext = adfs_f_getnext, .update = adfs_f_update, + .sync = adfs_f_sync, .free = adfs_f_free }; diff --git a/fs/adfs/dir_fplus.c b/fs/adfs/dir_fplus.c index 1ec644e32df9..139e0f345f18 100644 --- a/fs/adfs/dir_fplus.c +++ b/fs/adfs/dir_fplus.c @@ -161,6 +161,22 @@ out: return ret; } +static int +adfs_fplus_sync(struct adfs_dir *dir) +{ + int err = 0; + int i; + + for (i = dir->nr_buffers - 1; i >= 0; i--) { + struct buffer_head *bh = dir->bh[i]; + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) + err = -EIO; + } + + return err; +} + static void adfs_fplus_free(struct adfs_dir *dir) { @@ -175,5 +191,6 @@ struct adfs_dir_ops adfs_fplus_dir_ops = { .read = adfs_fplus_read, .setpos = adfs_fplus_setpos, .getnext = adfs_fplus_getnext, + .sync = adfs_fplus_sync, .free = adfs_fplus_free }; diff --git a/fs/adfs/file.c b/fs/adfs/file.c index 36e381c6a99a..8224d54a2afb 100644 --- a/fs/adfs/file.c +++ b/fs/adfs/file.c @@ -30,7 +30,7 @@ const struct file_operations adfs_file_operations = { .read = do_sync_read, .aio_read = generic_file_aio_read, .mmap = generic_file_mmap, - .fsync = file_fsync, + .fsync = simple_fsync, .write = do_sync_write, .aio_write = generic_file_aio_write, .splice_read = generic_file_splice_read, diff --git a/fs/adfs/inode.c b/fs/adfs/inode.c index e647200262a2..05b3a677201d 100644 --- a/fs/adfs/inode.c +++ b/fs/adfs/inode.c @@ -376,7 +376,7 @@ out: * The adfs-specific inode data has already been updated by * adfs_notify_change() */ -int adfs_write_inode(struct inode *inode, int unused) +int adfs_write_inode(struct inode *inode, int wait) { struct super_block *sb = inode->i_sb; struct object_info obj; @@ -391,7 +391,7 @@ int adfs_write_inode(struct inode *inode, int unused) obj.attr = ADFS_I(inode)->attr; obj.size = inode->i_size; - ret = adfs_dir_update(sb, &obj); + ret = adfs_dir_update(sb, &obj, wait); unlock_kernel(); return ret; } From 224c886643e52e6b4c1143489cd0b289b6c03976 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 8 Jun 2009 00:46:40 -0400 Subject: [PATCH 67/87] Fix adfs GET_FRAG_ID() on big-endian Missing conversion to host-endian before doing shifts Signed-off-by: Al Viro --- fs/adfs/map.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/adfs/map.c b/fs/adfs/map.c index 92ab4fbc2031..568081b93f73 100644 --- a/fs/adfs/map.c +++ b/fs/adfs/map.c @@ -62,7 +62,7 @@ static DEFINE_RWLOCK(adfs_map_lock); #define GET_FRAG_ID(_map,_start,_idmask) \ ({ \ unsigned char *_m = _map + (_start >> 3); \ - u32 _frag = get_unaligned((u32 *)_m); \ + u32 _frag = get_unaligned_le32(_m); \ _frag >>= (_start & 7); \ _frag & _idmask; \ }) From 4427f0c36e22e2cd6696b2fe7643e9756a14b3d3 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 8 Jun 2009 01:15:58 -0400 Subject: [PATCH 68/87] repair bfs_write_inode(), switch bfs to simple_fsync() Signed-off-by: Al Viro --- fs/bfs/dir.c | 8 ++++---- fs/bfs/inode.c | 12 +++++++++--- 2 files changed, 13 insertions(+), 7 deletions(-) diff --git a/fs/bfs/dir.c b/fs/bfs/dir.c index 4dd1b623f937..54bd07d44e68 100644 --- a/fs/bfs/dir.c +++ b/fs/bfs/dir.c @@ -79,7 +79,7 @@ static int bfs_readdir(struct file *f, void *dirent, filldir_t filldir) const struct file_operations bfs_dir_operations = { .read = generic_read_dir, .readdir = bfs_readdir, - .fsync = file_fsync, + .fsync = simple_fsync, .llseek = generic_file_llseek, }; @@ -205,7 +205,7 @@ static int bfs_unlink(struct inode *dir, struct dentry *dentry) inode->i_nlink = 1; } de->ino = 0; - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC; mark_inode_dirty(dir); inode->i_ctime = dir->i_ctime; @@ -267,7 +267,7 @@ static int bfs_rename(struct inode *old_dir, struct dentry *old_dentry, new_inode->i_ctime = CURRENT_TIME_SEC; inode_dec_link_count(new_inode); } - mark_buffer_dirty(old_bh); + mark_buffer_dirty_inode(old_bh, old_dir); error = 0; end_rename: @@ -320,7 +320,7 @@ static int bfs_add_entry(struct inode *dir, const unsigned char *name, for (i = 0; i < BFS_NAMELEN; i++) de->name[i] = (i < namelen) ? name[i] : 0; - mark_buffer_dirty(bh); + mark_buffer_dirty_inode(bh, dir); brelse(bh); return 0; } diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index 3a9a1361fdc1..d1d9d9088371 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -98,14 +98,15 @@ error: return ERR_PTR(-EIO); } -static int bfs_write_inode(struct inode *inode, int unused) +static int bfs_write_inode(struct inode *inode, int wait) { + struct bfs_sb_info *info = BFS_SB(inode->i_sb); unsigned int ino = (u16)inode->i_ino; unsigned long i_sblock; struct bfs_inode *di; struct buffer_head *bh; int block, off; - struct bfs_sb_info *info = BFS_SB(inode->i_sb); + int err = 0; dprintf("ino=%08x\n", ino); @@ -146,9 +147,14 @@ static int bfs_write_inode(struct inode *inode, int unused) di->i_eoffset = cpu_to_le32(i_sblock * BFS_BSIZE + inode->i_size - 1); mark_buffer_dirty(bh); + if (wait) { + sync_dirty_buffer(bh); + if (buffer_req(bh) && !buffer_uptodate(bh)) + err = -EIO; + } brelse(bh); mutex_unlock(&info->bfs_lock); - return 0; + return err; } static void bfs_delete_inode(struct inode *inode) From c475879556a8602bbe2faa9a06f6e5fcc8c05bb2 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 8 Jun 2009 01:22:00 -0400 Subject: [PATCH 69/87] sanitize ->fsync() for affs unfortunately, for affs (especially for affs directories) we have no real way to keep track of metadata ownership. So we have to do more or less what file_fsync() does, but we do *not* need to call write_super() there. Signed-off-by: Al Viro --- fs/affs/affs.h | 1 + fs/affs/dir.c | 2 +- fs/affs/file.c | 14 +++++++++++++- 3 files changed, 15 insertions(+), 2 deletions(-) diff --git a/fs/affs/affs.h b/fs/affs/affs.h index 1a2d5e3c7f4e..e511dc621a2e 100644 --- a/fs/affs/affs.h +++ b/fs/affs/affs.h @@ -182,6 +182,7 @@ extern int affs_add_entry(struct inode *dir, struct inode *inode, struct dent void affs_free_prealloc(struct inode *inode); extern void affs_truncate(struct inode *); +int affs_file_fsync(struct file *, struct dentry *, int); /* dir.c */ diff --git a/fs/affs/dir.c b/fs/affs/dir.c index 7b36904dbeac..8ca8f3a55599 100644 --- a/fs/affs/dir.c +++ b/fs/affs/dir.c @@ -21,7 +21,7 @@ const struct file_operations affs_dir_operations = { .read = generic_read_dir, .llseek = generic_file_llseek, .readdir = affs_readdir, - .fsync = file_fsync, + .fsync = affs_file_fsync, }; /* diff --git a/fs/affs/file.c b/fs/affs/file.c index 9246cb4aa018..184e55c1c9ba 100644 --- a/fs/affs/file.c +++ b/fs/affs/file.c @@ -34,7 +34,7 @@ const struct file_operations affs_file_operations = { .mmap = generic_file_mmap, .open = affs_file_open, .release = affs_file_release, - .fsync = file_fsync, + .fsync = affs_file_fsync, .splice_read = generic_file_splice_read, }; @@ -915,3 +915,15 @@ affs_truncate(struct inode *inode) } affs_free_prealloc(inode); } + +int affs_file_fsync(struct file *filp, struct dentry *dentry, int datasync) +{ + struct inode * inode = dentry->d_inode; + int ret, err; + + ret = write_inode_now(inode, 0); + err = sync_blockdev(inode->i_sb->s_bdev); + if (!ret) + ret = err; + return ret; +} From e28964365faf3b9961695eb62b48cbc9f2a2a245 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:03:15 +0200 Subject: [PATCH 70/87] affs: add ->sync_fs Add a ->sync_fs method for data integrity syncs. Factor out common code between affs_put_super, affs_write_super and the new affs_sync_fs into a helper. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/affs/super.c | 40 +++++++++++++++++++++++++++------------- 1 file changed, 27 insertions(+), 13 deletions(-) diff --git a/fs/affs/super.c b/fs/affs/super.c index c4814937c968..104fdcb3a7fc 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -24,6 +24,19 @@ extern struct timezone sys_tz; static int affs_statfs(struct dentry *dentry, struct kstatfs *buf); static int affs_remount (struct super_block *sb, int *flags, char *data); +static void +affs_commit_super(struct super_block *sb, int clean) +{ + struct affs_sb_info *sbi = AFFS_SB(sb); + struct buffer_head *bh = sbi->s_root_bh; + struct affs_root_tail *tail = AFFS_ROOT_TAIL(sb, bh); + + tail->bm_flag = cpu_to_be32(clean); + secs_to_datestamp(get_seconds(), &tail->disk_change); + affs_fix_checksum(sb, bh); + mark_buffer_dirty(bh); +} + static void affs_put_super(struct super_block *sb) { @@ -32,13 +45,8 @@ affs_put_super(struct super_block *sb) lock_kernel(); - if (!(sb->s_flags & MS_RDONLY)) { - AFFS_ROOT_TAIL(sb, sbi->s_root_bh)->bm_flag = cpu_to_be32(1); - secs_to_datestamp(get_seconds(), - &AFFS_ROOT_TAIL(sb, sbi->s_root_bh)->disk_change); - affs_fix_checksum(sb, sbi->s_root_bh); - mark_buffer_dirty(sbi->s_root_bh); - } + if (!(sb->s_flags & MS_RDONLY)) + affs_commit_super(sb, 1); kfree(sbi->s_prefix); affs_free_bitmap(sb); @@ -53,18 +61,13 @@ static void affs_write_super(struct super_block *sb) { int clean = 2; - struct affs_sb_info *sbi = AFFS_SB(sb); lock_super(sb); if (!(sb->s_flags & MS_RDONLY)) { // if (sbi->s_bitmap[i].bm_bh) { // if (buffer_dirty(sbi->s_bitmap[i].bm_bh)) { // clean = 0; - AFFS_ROOT_TAIL(sb, sbi->s_root_bh)->bm_flag = cpu_to_be32(clean); - secs_to_datestamp(get_seconds(), - &AFFS_ROOT_TAIL(sb, sbi->s_root_bh)->disk_change); - affs_fix_checksum(sb, sbi->s_root_bh); - mark_buffer_dirty(sbi->s_root_bh); + affs_commit_super(sb, clean); sb->s_dirt = !clean; /* redo until bitmap synced */ } else sb->s_dirt = 0; @@ -73,6 +76,16 @@ affs_write_super(struct super_block *sb) pr_debug("AFFS: write_super() at %lu, clean=%d\n", get_seconds(), clean); } +static int +affs_sync_fs(struct super_block *sb, int wait) +{ + lock_super(sb); + affs_commit_super(sb, 2); + sb->s_dirt = 0; + unlock_super(sb); + return 0; +} + static struct kmem_cache * affs_inode_cachep; static struct inode *affs_alloc_inode(struct super_block *sb) @@ -130,6 +143,7 @@ static const struct super_operations affs_sops = { .clear_inode = affs_clear_inode, .put_super = affs_put_super, .write_super = affs_write_super, + .sync_fs = affs_sync_fs, .statfs = affs_statfs, .remount_fs = affs_remount, .show_options = generic_show_options, From 561e47ce7244168788d4ecef9a2271df204b3c89 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:03:38 +0200 Subject: [PATCH 71/87] bfs: add ->sync_fs Add a ->sync_fs method for data integrity syncs, and reimplement ->write_super ontop of it. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/bfs/inode.c | 32 +++++++++++++++++++++----------- 1 file changed, 21 insertions(+), 11 deletions(-) diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index d1d9d9088371..6f60336c6628 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -216,6 +216,26 @@ static void bfs_delete_inode(struct inode *inode) clear_inode(inode); } +static int bfs_sync_fs(struct super_block *sb, int wait) +{ + struct bfs_sb_info *info = BFS_SB(sb); + + mutex_lock(&info->bfs_lock); + mark_buffer_dirty(info->si_sbh); + sb->s_dirt = 0; + mutex_unlock(&info->bfs_lock); + + return 0; +} + +static void bfs_write_super(struct super_block *sb) +{ + if (!(sb->s_flags & MS_RDONLY)) + bfs_sync_fs(sb, 1); + else + sb->s_dirt = 0; +} + static void bfs_put_super(struct super_block *s) { struct bfs_sb_info *info = BFS_SB(s); @@ -254,17 +274,6 @@ static int bfs_statfs(struct dentry *dentry, struct kstatfs *buf) return 0; } -static void bfs_write_super(struct super_block *s) -{ - struct bfs_sb_info *info = BFS_SB(s); - - mutex_lock(&info->bfs_lock); - if (!(s->s_flags & MS_RDONLY)) - mark_buffer_dirty(info->si_sbh); - s->s_dirt = 0; - mutex_unlock(&info->bfs_lock); -} - static struct kmem_cache *bfs_inode_cachep; static struct inode *bfs_alloc_inode(struct super_block *sb) @@ -312,6 +321,7 @@ static const struct super_operations bfs_sops = { .delete_inode = bfs_delete_inode, .put_super = bfs_put_super, .write_super = bfs_write_super, + .sync_fs = bfs_sync_fs, .statfs = bfs_statfs, }; From 80e09fb942d38beb19dcffbeb14d496beeb0a989 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:03:58 +0200 Subject: [PATCH 72/87] exofs: add ->sync_fs Add a ->sync_fs method for data integrity syncs, and reimplement ->write_super ontop of it. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/exofs/super.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/fs/exofs/super.c b/fs/exofs/super.c index 49e16af4e619..8216c5b77b53 100644 --- a/fs/exofs/super.c +++ b/fs/exofs/super.c @@ -200,18 +200,18 @@ static const struct export_operations exofs_export_ops; /* * Write the superblock to the OSD */ -static void exofs_write_super(struct super_block *sb) +static int exofs_sync_fs(struct super_block *sb, int wait) { struct exofs_sb_info *sbi; struct exofs_fscb *fscb; struct osd_request *or; struct osd_obj_id obj; - int ret; + int ret = -ENOMEM; fscb = kzalloc(sizeof(struct exofs_fscb), GFP_KERNEL); if (!fscb) { EXOFS_ERR("exofs_write_super: memory allocation failed.\n"); - return; + return -ENOMEM; } lock_super(sb); @@ -249,6 +249,15 @@ out: unlock_kernel(); unlock_super(sb); kfree(fscb); + return ret; +} + +static void exofs_write_super(struct super_block *sb) +{ + if (!(sb->s_flags & MS_RDONLY)) + exofs_sync_fs(sb, 1); + else + sb->s_dirt = 0; } /* @@ -493,6 +502,7 @@ static const struct super_operations exofs_sops = { .delete_inode = exofs_delete_inode, .put_super = exofs_put_super, .write_super = exofs_write_super, + .sync_fs = exofs_sync_fs, .statfs = exofs_statfs, }; From 40f31dd47e7c3d15af1f9845eda0fa0c4c33f32f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:04:17 +0200 Subject: [PATCH 73/87] ext2: add ->sync_fs Add a ->sync_fs method for data integrity syncs, and reimplement ->write_super ontop of it. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/ext2/super.c | 41 +++++++++++++++++++++++++++-------------- 1 file changed, 27 insertions(+), 14 deletions(-) diff --git a/fs/ext2/super.c b/fs/ext2/super.c index f8cbdf569190..458999638c3d 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -42,6 +42,7 @@ static void ext2_sync_super(struct super_block *sb, struct ext2_super_block *es); static int ext2_remount (struct super_block * sb, int * flags, char * data); static int ext2_statfs (struct dentry * dentry, struct kstatfs * buf); +static int ext2_sync_fs(struct super_block *sb, int wait); void ext2_error (struct super_block * sb, const char * function, const char * fmt, ...) @@ -309,6 +310,7 @@ static const struct super_operations ext2_sops = { .delete_inode = ext2_delete_inode, .put_super = ext2_put_super, .write_super = ext2_write_super, + .sync_fs = ext2_sync_fs, .statfs = ext2_statfs, .remount_fs = ext2_remount, .clear_inode = ext2_clear_inode, @@ -1132,25 +1134,36 @@ static void ext2_sync_super(struct super_block *sb, struct ext2_super_block *es) * set s_state to EXT2_VALID_FS after some corrections. */ -void ext2_write_super (struct super_block * sb) +static int ext2_sync_fs(struct super_block *sb, int wait) { - struct ext2_super_block * es; - lock_kernel(); - if (!(sb->s_flags & MS_RDONLY)) { - es = EXT2_SB(sb)->s_es; + struct ext2_super_block *es = EXT2_SB(sb)->s_es; - if (es->s_state & cpu_to_le16(EXT2_VALID_FS)) { - ext2_debug ("setting valid to 0\n"); - es->s_state &= cpu_to_le16(~EXT2_VALID_FS); - es->s_free_blocks_count = cpu_to_le32(ext2_count_free_blocks(sb)); - es->s_free_inodes_count = cpu_to_le32(ext2_count_free_inodes(sb)); - es->s_mtime = cpu_to_le32(get_seconds()); - ext2_sync_super(sb, es); - } else - ext2_commit_super (sb, es); + lock_kernel(); + if (es->s_state & cpu_to_le16(EXT2_VALID_FS)) { + ext2_debug("setting valid to 0\n"); + es->s_state &= cpu_to_le16(~EXT2_VALID_FS); + es->s_free_blocks_count = + cpu_to_le32(ext2_count_free_blocks(sb)); + es->s_free_inodes_count = + cpu_to_le32(ext2_count_free_inodes(sb)); + es->s_mtime = cpu_to_le32(get_seconds()); + ext2_sync_super(sb, es); + } else { + ext2_commit_super(sb, es); } sb->s_dirt = 0; unlock_kernel(); + + return 0; +} + + +void ext2_write_super(struct super_block *sb) +{ + if (!(sb->s_flags & MS_RDONLY)) + ext2_sync_fs(sb, 1); + else + sb->s_dirt = 0; } static int ext2_remount (struct super_block * sb, int * flags, char * data) From f83d6d46e7adf241a064a4a425e5cd8a8fd8925f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:04:35 +0200 Subject: [PATCH 74/87] fat: add ->sync_fs Add a ->sync_fs method for data integrity syncs. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/fat/inode.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 476f80b175fe..51a5ecf9000a 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -449,6 +449,16 @@ static void fat_write_super(struct super_block *sb) unlock_super(sb); } +static int fat_sync_fs(struct super_block *sb, int wait) +{ + lock_super(sb); + fat_clusters_flush(sb); + sb->s_dirt = 0; + unlock_super(sb); + + return 0; +} + static void fat_put_super(struct super_block *sb) { struct msdos_sb_info *sbi = MSDOS_SB(sb); @@ -643,6 +653,7 @@ static const struct super_operations fat_sops = { .delete_inode = fat_delete_inode, .put_super = fat_put_super, .write_super = fat_write_super, + .sync_fs = fat_sync_fs, .statfs = fat_statfs, .clear_inode = fat_clear_inode, .remount_fs = fat_remount, From 58bc5bbb873eb5d86126a3fd3ff02aaa69ec15d0 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:04:54 +0200 Subject: [PATCH 75/87] hfs: add ->sync_fs Add a ->sync_fs method for data integrity syncs. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/hfs/super.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/fs/hfs/super.c b/fs/hfs/super.c index 3aac41751030..6f833dc8e910 100644 --- a/fs/hfs/super.c +++ b/fs/hfs/super.c @@ -58,6 +58,16 @@ static void hfs_write_super(struct super_block *sb) unlock_super(sb); } +static int hfs_sync_fs(struct super_block *sb, int wait) +{ + lock_super(sb); + hfs_mdb_commit(sb); + sb->s_dirt = 0; + unlock_super(sb); + + return 0; +} + /* * hfs_put_super() * @@ -172,6 +182,7 @@ static const struct super_operations hfs_super_operations = { .clear_inode = hfs_clear_inode, .put_super = hfs_put_super, .write_super = hfs_write_super, + .sync_fs = hfs_sync_fs, .statfs = hfs_statfs, .remount_fs = hfs_remount, .show_options = hfs_show_options, From 7fbc6df0e7a561a313f49faa77829d5de45a97f8 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:05:12 +0200 Subject: [PATCH 76/87] hfsplus: add ->sync_fs Add a ->sync_fs method for data integrity syncs, and reimplement ->write_super ontop of it. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/hfsplus/super.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/fs/hfsplus/super.c b/fs/hfsplus/super.c index 1aab8aa7801e..9fc3af0c0dab 100644 --- a/fs/hfsplus/super.c +++ b/fs/hfsplus/super.c @@ -152,7 +152,7 @@ static void hfsplus_clear_inode(struct inode *inode) } } -static void hfsplus_write_super(struct super_block *sb) +static int hfsplus_sync_fs(struct super_block *sb, int wait) { struct hfsplus_vh *vhdr = HFSPLUS_SB(sb).s_vhdr; @@ -160,9 +160,6 @@ static void hfsplus_write_super(struct super_block *sb) lock_super(sb); sb->s_dirt = 0; - if (sb->s_flags & MS_RDONLY) - /* warn? */ - goto out; vhdr->free_blocks = cpu_to_be32(HFSPLUS_SB(sb).free_blocks); vhdr->next_alloc = cpu_to_be32(HFSPLUS_SB(sb).next_alloc); @@ -194,8 +191,16 @@ static void hfsplus_write_super(struct super_block *sb) } HFSPLUS_SB(sb).flags &= ~HFSPLUS_SB_WRITEBACKUP; } - out: unlock_super(sb); + return 0; +} + +static void hfsplus_write_super(struct super_block *sb) +{ + if (!(sb->s_flags & MS_RDONLY)) + hfsplus_sync_fs(sb, 1); + else + sb->s_dirt = 0; } static void hfsplus_put_super(struct super_block *sb) @@ -290,6 +295,7 @@ static const struct super_operations hfsplus_sops = { .clear_inode = hfsplus_clear_inode, .put_super = hfsplus_put_super, .write_super = hfsplus_write_super, + .sync_fs = hfsplus_sync_fs, .statfs = hfsplus_statfs, .remount_fs = hfsplus_remount, .show_options = hfsplus_show_options, From ad43ffdeea0a7bd3e6036c4aeec2e6699aef8ac7 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:07:45 +0200 Subject: [PATCH 77/87] sysv: add ->sync_fs Add a ->sync_fs method for data integrity syncs, and reimplement ->write_super ontop of it. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/sysv/inode.c | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c index 425c976cfcd2..479923456a54 100644 --- a/fs/sysv/inode.c +++ b/fs/sysv/inode.c @@ -31,16 +31,13 @@ #include #include "sysv.h" -/* This is only called on sync() and umount(), when s_dirt=1. */ -static void sysv_write_super(struct super_block *sb) +static int sysv_sync_fs(struct super_block *sb, int wait) { struct sysv_sb_info *sbi = SYSV_SB(sb); unsigned long time = get_seconds(), old_time; lock_super(sb); lock_kernel(); - if (sb->s_flags & MS_RDONLY) - goto clean; /* * If we are going to write out the super block, @@ -54,10 +51,19 @@ static void sysv_write_super(struct super_block *sb) *sbi->s_sb_time = cpu_to_fs32(sbi, time); mark_buffer_dirty(sbi->s_bh2); } -clean: - sb->s_dirt = 0; + unlock_kernel(); unlock_super(sb); + + return 0; +} + +static void sysv_write_super(struct super_block *sb) +{ + if (!(sb->s_flags & MS_RDONLY)) + sysv_sync_fs(sb, 1); + else + sb->s_dirt = 0; } static int sysv_remount(struct super_block *sb, int *flags, char *data) @@ -345,6 +351,7 @@ const struct super_operations sysv_sops = { .delete_inode = sysv_delete_inode, .put_super = sysv_put_super, .write_super = sysv_write_super, + .sync_fs = sysv_sync_fs, .remount_fs = sysv_remount, .statfs = sysv_statfs, }; From 8c8006564a58d0ea912bf7f2d0a758d97e4b464f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:08:05 +0200 Subject: [PATCH 78/87] ufs: add ->sync_fs Add a ->sync_fs method for data integrity syncs, and reimplement ->write_super ontop of it. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/ufs/super.c | 32 ++++++++++++++++++++++---------- 1 file changed, 22 insertions(+), 10 deletions(-) diff --git a/fs/ufs/super.c b/fs/ufs/super.c index 6560dda7b18c..5faed7954d0a 100644 --- a/fs/ufs/super.c +++ b/fs/ufs/super.c @@ -1125,7 +1125,7 @@ failed_nomem: return -ENOMEM; } -static void ufs_write_super(struct super_block *sb) +static int ufs_sync_fs(struct super_block *sb, int wait) { struct ufs_sb_private_info * uspi; struct ufs_super_block_first * usb1; @@ -1134,25 +1134,36 @@ static void ufs_write_super(struct super_block *sb) lock_super(sb); lock_kernel(); + UFSD("ENTER\n"); + flags = UFS_SB(sb)->s_flags; uspi = UFS_SB(sb)->s_uspi; usb1 = ubh_get_usb_first(uspi); usb3 = ubh_get_usb_third(uspi); - if (!(sb->s_flags & MS_RDONLY)) { - usb1->fs_time = cpu_to_fs32(sb, get_seconds()); - if ((flags & UFS_ST_MASK) == UFS_ST_SUN - || (flags & UFS_ST_MASK) == UFS_ST_SUNOS - || (flags & UFS_ST_MASK) == UFS_ST_SUNx86) - ufs_set_fs_state(sb, usb1, usb3, - UFS_FSOK - fs32_to_cpu(sb, usb1->fs_time)); - ufs_put_cstotal(sb); - } + usb1->fs_time = cpu_to_fs32(sb, get_seconds()); + if ((flags & UFS_ST_MASK) == UFS_ST_SUN || + (flags & UFS_ST_MASK) == UFS_ST_SUNOS || + (flags & UFS_ST_MASK) == UFS_ST_SUNx86) + ufs_set_fs_state(sb, usb1, usb3, + UFS_FSOK - fs32_to_cpu(sb, usb1->fs_time)); + ufs_put_cstotal(sb); sb->s_dirt = 0; + UFSD("EXIT\n"); unlock_kernel(); unlock_super(sb); + + return 0; +} + +static void ufs_write_super(struct super_block *sb) +{ + if (!(sb->s_flags & MS_RDONLY)) + ufs_sync_fs(sb, 1); + else + sb->s_dirt = 0; } static void ufs_put_super(struct super_block *sb) @@ -1381,6 +1392,7 @@ static const struct super_operations ufs_super_ops = { .delete_inode = ufs_delete_inode, .put_super = ufs_put_super, .write_super = ufs_write_super, + .sync_fs = ufs_sync_fs, .statfs = ufs_statfs, .remount_fs = ufs_remount, .show_options = ufs_show_options, From d579ed00aa96a7f7486978540a0d7cecaff742ae Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:08:21 +0200 Subject: [PATCH 79/87] jffs2: call jffs2_write_super from jffs2_sync_fs The call to ->write_super from __sync_filesystem will go away, so make sure jffs2 performs the same actions from inside ->sync_fs. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/jffs2/super.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index f7bfd3ac8bfa..07a22caf2687 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -74,6 +74,8 @@ static int jffs2_sync_fs(struct super_block *sb, int wait) { struct jffs2_sb_info *c = JFFS2_SB_INFO(sb); + jffs2_write_super(sb); + mutex_lock(&c->alloc_sem); jffs2_flush_wbuf_pad(c); mutex_unlock(&c->alloc_sem); From d731e06323cb705003e4172ec209e469be4c18e1 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:08:36 +0200 Subject: [PATCH 80/87] nilfs2: call nilfs2_write_super from nilfs2_sync_fs The call to ->write_super from __sync_filesystem will go away, so make sure nilfs2 performs the same actions from inside ->sync_fs. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/nilfs2/super.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 11151eaa2c4a..122dc1e489fb 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -391,6 +391,8 @@ static int nilfs_sync_fs(struct super_block *sb, int wait) { int err = 0; + nilfs_write_super(sb); + /* This function is called when super block should be written back */ if (wait) err = nilfs_construct_segment(sb); From 0c95ee190e1dea60c55c834d14695341085c9b7b Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 8 Jun 2009 10:08:54 +0200 Subject: [PATCH 81/87] remove the call to ->write_super in __sync_filesystem Now that all filesystems provide ->sync_fs methods we can change __sync_filesystem to only call ->sync_fs. This gives us a clear separation between periodic writeouts which are driven by ->write_super and data integrity syncs that go through ->sync_fs. (modulo file_fsync which is also going away) Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/sync.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/fs/sync.c b/fs/sync.c index e9d56f6c0b74..dd200025af85 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -33,8 +33,6 @@ static int __sync_filesystem(struct super_block *sb, int wait) else sync_quota_sb(sb, -1); sync_inodes_sb(sb, wait); - if (sb->s_dirt && sb->s_op->write_super) - sb->s_op->write_super(sb); if (sb->s_op->sync_fs) sb->s_op->sync_fs(sb, wait); return __sync_blockdev(sb->s_bdev, wait); From 81fc20bd0e75ba6357bce2403767d7c2585d8f28 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 8 Jun 2009 01:39:28 +0900 Subject: [PATCH 82/87] nilfs2: remove meaningless EBUSY case from nilfs_get_sb function The following EBUSY case in nilfs_get_sb() is meaningless. Indeed, this error code is never returned to the caller. if (!s->s_root) { ... } else if (!(s->s_flags & MS_RDONLY)) { err = -EBUSY; } This simply removes the else case. Signed-off-by: Ryusuke Konishi Signed-off-by: Al Viro --- fs/nilfs2/super.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 122dc1e489fb..1c505d0e031e 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -1186,8 +1186,6 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, s->s_flags |= MS_ACTIVE; need_to_close = 0; - } else if (!(s->s_flags & MS_RDONLY)) { - err = -EBUSY; } up(&sd.bdev->bd_mount_sem); From 33c8e57c86d1bd1548c12a4f7c4bceb94b862cca Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 8 Jun 2009 01:39:29 +0900 Subject: [PATCH 83/87] nilfs2: get rid of sget use for acquiring nilfs object This will change the way to obtain nilfs object in nilfs_get_sb() function. Previously, a preliminary sget() call was performed, and the nilfs object was acquired from a super block instance found by the sget() call. This patch, instead, instroduces a new dedicated function find_or_create_nilfs(); as the name implies, the function finds an existent nilfs object from a global list or creates a new one if no object is found on the device. Signed-off-by: Ryusuke Konishi Cc: Al Viro Signed-off-by: Al Viro --- fs/nilfs2/super.c | 80 ++++++++++--------------------------------- fs/nilfs2/the_nilfs.c | 57 ++++++++++++++++++++++++++++-- fs/nilfs2/the_nilfs.h | 4 ++- 3 files changed, 76 insertions(+), 65 deletions(-) diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 1c505d0e031e..3c9833e3e74a 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -1059,13 +1059,6 @@ static int nilfs_set_bdev_super(struct super_block *s, void *data) } static int nilfs_test_bdev_super(struct super_block *s, void *data) -{ - struct nilfs_super_data *sd = data; - - return s->s_bdev == sd->bdev; -} - -static int nilfs_test_bdev_super2(struct super_block *s, void *data) { struct nilfs_super_data *sd = data; int ret; @@ -1096,8 +1089,8 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *data, struct vfsmount *mnt) { struct nilfs_super_data sd; - struct super_block *s, *s2; - struct the_nilfs *nilfs = NULL; + struct super_block *s; + struct the_nilfs *nilfs; int err, need_to_close = 1; sd.bdev = open_bdev_exclusive(dev_name, flags, fs_type); @@ -1118,11 +1111,12 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, goto failed; } - /* - * once the super is inserted into the list by sget, s_umount - * will protect the lockfs code from trying to start a snapshot - * while we are mounting - */ + nilfs = find_or_create_nilfs(sd.bdev); + if (!nilfs) { + err = -ENOMEM; + goto failed; + } + down(&sd.bdev->bd_mount_sem); if (!sd.cno && (err = test_exclusive_mount(fs_type, sd.bdev, flags ^ MS_RDONLY))) { @@ -1131,51 +1125,22 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, } /* - * Phase-1: search any existent instance and get the_nilfs + * Search specified snapshot or R/W mode super_block */ + if (!sd.cno) + /* trying to get the latest checkpoint. */ + sd.cno = nilfs_last_cno(nilfs); + s = sget(fs_type, nilfs_test_bdev_super, nilfs_set_bdev_super, &sd); - if (IS_ERR(s)) - goto error_s; - - if (!s->s_root) { - err = -ENOMEM; - nilfs = alloc_nilfs(sd.bdev); - if (!nilfs) - goto cancel_new; - } else { - struct nilfs_sb_info *sbi = NILFS_SB(s); - - /* - * s_umount protects super_block from unmount process; - * It covers pointers of nilfs_sb_info and the_nilfs. - */ - nilfs = sbi->s_nilfs; - get_nilfs(nilfs); - up_write(&s->s_umount); - - /* - * Phase-2: search specified snapshot or R/W mode super_block - */ - if (!sd.cno) - /* trying to get the latest checkpoint. */ - sd.cno = nilfs_last_cno(nilfs); - - s2 = sget(fs_type, nilfs_test_bdev_super2, - nilfs_set_bdev_super, &sd); - deactivate_super(s); - /* - * Although deactivate_super() invokes close_bdev_exclusive() at - * kill_block_super(). Here, s is an existent mount; we need - * one more close_bdev_exclusive() call. - */ - s = s2; - if (IS_ERR(s)) - goto error_s; + if (IS_ERR(s)) { + err = PTR_ERR(s); + goto failed_unlock; } if (!s->s_root) { char b[BDEVNAME_SIZE]; + /* New superblock instance created */ s->s_flags = flags; strlcpy(s->s_id, bdevname(sd.bdev, b), sizeof(s->s_id)); sb_set_blocksize(s, block_size(sd.bdev)); @@ -1195,15 +1160,9 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, simple_set_mnt(mnt, s); return 0; - error_s: - up(&sd.bdev->bd_mount_sem); - if (nilfs) - put_nilfs(nilfs); - close_bdev_exclusive(sd.bdev, flags); - return PTR_ERR(s); - failed_unlock: up(&sd.bdev->bd_mount_sem); + put_nilfs(nilfs); failed: close_bdev_exclusive(sd.bdev, flags); @@ -1212,8 +1171,7 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, cancel_new: /* Abandoning the newly allocated superblock */ up(&sd.bdev->bd_mount_sem); - if (nilfs) - put_nilfs(nilfs); + put_nilfs(nilfs); up_write(&s->s_umount); deactivate_super(s); /* diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index a91f15b8673c..45dbf6a61744 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -35,6 +35,10 @@ #include "seglist.h" #include "segbuf.h" + +static LIST_HEAD(nilfs_objects); +static DEFINE_SPINLOCK(nilfs_lock); + void nilfs_set_last_segment(struct the_nilfs *nilfs, sector_t start_blocknr, u64 seq, __u64 cno) { @@ -55,7 +59,7 @@ void nilfs_set_last_segment(struct the_nilfs *nilfs, * Return Value: On success, pointer to the_nilfs is returned. * On error, NULL is returned. */ -struct the_nilfs *alloc_nilfs(struct block_device *bdev) +static struct the_nilfs *alloc_nilfs(struct block_device *bdev) { struct the_nilfs *nilfs; @@ -69,6 +73,7 @@ struct the_nilfs *alloc_nilfs(struct block_device *bdev) atomic_set(&nilfs->ns_ndirtyblks, 0); init_rwsem(&nilfs->ns_sem); mutex_init(&nilfs->ns_writer_mutex); + INIT_LIST_HEAD(&nilfs->ns_list); INIT_LIST_HEAD(&nilfs->ns_supers); spin_lock_init(&nilfs->ns_last_segment_lock); nilfs->ns_gc_inodes_h = NULL; @@ -77,6 +82,45 @@ struct the_nilfs *alloc_nilfs(struct block_device *bdev) return nilfs; } +/** + * find_or_create_nilfs - find or create nilfs object + * @bdev: block device to which the_nilfs is related + * + * find_nilfs() looks up an existent nilfs object created on the + * device and gets the reference count of the object. If no nilfs object + * is found on the device, a new nilfs object is allocated. + * + * Return Value: On success, pointer to the nilfs object is returned. + * On error, NULL is returned. + */ +struct the_nilfs *find_or_create_nilfs(struct block_device *bdev) +{ + struct the_nilfs *nilfs, *new = NULL; + + retry: + spin_lock(&nilfs_lock); + list_for_each_entry(nilfs, &nilfs_objects, ns_list) { + if (nilfs->ns_bdev == bdev) { + get_nilfs(nilfs); + spin_unlock(&nilfs_lock); + if (new) + put_nilfs(new); + return nilfs; /* existing object */ + } + } + if (new) { + list_add_tail(&new->ns_list, &nilfs_objects); + spin_unlock(&nilfs_lock); + return new; /* new object */ + } + spin_unlock(&nilfs_lock); + + new = alloc_nilfs(bdev); + if (new) + goto retry; + return NULL; /* insufficient memory */ +} + /** * put_nilfs - release a reference to the_nilfs * @nilfs: the_nilfs structure to be released @@ -86,13 +130,20 @@ struct the_nilfs *alloc_nilfs(struct block_device *bdev) */ void put_nilfs(struct the_nilfs *nilfs) { - if (!atomic_dec_and_test(&nilfs->ns_count)) + spin_lock(&nilfs_lock); + if (!atomic_dec_and_test(&nilfs->ns_count)) { + spin_unlock(&nilfs_lock); return; + } + list_del_init(&nilfs->ns_list); + spin_unlock(&nilfs_lock); + /* - * Increment of ns_count never occur below because the caller + * Increment of ns_count never occurs below because the caller * of get_nilfs() holds at least one reference to the_nilfs. * Thus its exclusion control is not required here. */ + might_sleep(); if (nilfs_loaded(nilfs)) { nilfs_mdt_clear(nilfs->ns_sufile); diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 30fe58778d05..116caf96e7f3 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -43,6 +43,7 @@ enum { * struct the_nilfs - struct to supervise multiple nilfs mount points * @ns_flags: flags * @ns_count: reference count + * @ns_list: list head for nilfs_list * @ns_bdev: block device * @ns_bdi: backing dev info * @ns_writer: back pointer to writable nilfs_sb_info @@ -88,6 +89,7 @@ enum { struct the_nilfs { unsigned long ns_flags; atomic_t ns_count; + struct list_head ns_list; struct block_device *ns_bdev; struct backing_dev_info *ns_bdi; @@ -191,7 +193,7 @@ THE_NILFS_FNS(DISCONTINUED, discontinued) #define NILFS_ALTSB_FREQ 60 /* spare superblock */ void nilfs_set_last_segment(struct the_nilfs *, sector_t, u64, __u64); -struct the_nilfs *alloc_nilfs(struct block_device *); +struct the_nilfs *find_or_create_nilfs(struct block_device *); void put_nilfs(struct the_nilfs *); int init_nilfs(struct the_nilfs *, struct nilfs_sb_info *, char *); int load_nilfs(struct the_nilfs *, struct nilfs_sb_info *); From 3f82ff55168e92859119bf348e9e0bd6714d2fea Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 8 Jun 2009 01:39:30 +0900 Subject: [PATCH 84/87] nilfs2: get rid of sget use for checking if current mount is present This stops using sget() for checking if an r/w-mount or an r/o-mount exists on the device. This elimination uses a back pointer to the current mount added to nilfs object. Signed-off-by: Ryusuke Konishi Cc: Al Viro Signed-off-by: Al Viro --- fs/nilfs2/super.c | 92 +++++++++++++++---------------------------- fs/nilfs2/the_nilfs.h | 3 ++ 2 files changed, 35 insertions(+), 60 deletions(-) diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 3c9833e3e74a..5a8c5e4731b3 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -67,8 +67,6 @@ MODULE_LICENSE("GPL"); static void nilfs_write_super(struct super_block *sb); static int nilfs_remount(struct super_block *sb, int *flags, char *data); -static int test_exclusive_mount(struct file_system_type *fs_type, - struct block_device *bdev, int flags); /** * nilfs_error() - report failure condition on a filesystem @@ -329,6 +327,10 @@ static void nilfs_put_super(struct super_block *sb) nilfs_commit_super(sbi, 1); up_write(&nilfs->ns_sem); } + down_write(&nilfs->ns_sem); + if (nilfs->ns_current == sbi) + nilfs->ns_current = NULL; + up_write(&nilfs->ns_sem); nilfs_detach_checkpoint(sbi); put_nilfs(sbi->s_nilfs); @@ -880,6 +882,11 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent, goto failed_root; } + down_write(&nilfs->ns_sem); + if (!nilfs_test_opt(sbi, SNAPSHOT)) + nilfs->ns_current = sbi; + up_write(&nilfs->ns_sem); + return 0; failed_root: @@ -958,14 +965,16 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) * by fsck since we originally mounted the partition.) */ down(&sb->s_bdev->bd_mount_sem); - /* Check existing RW-mount */ - if (test_exclusive_mount(sb->s_type, sb->s_bdev, 0)) { + down_read(&nilfs->ns_sem); + if (nilfs->ns_current && nilfs->ns_current != sbi) { printk(KERN_WARNING "NILFS (device %s): couldn't " - "remount because a RW-mount exists.\n", + "remount because an RW-mount exists.\n", sb->s_id); + up_read(&nilfs->ns_sem); err = -EBUSY; goto rw_remount_failed; } + up_read(&nilfs->ns_sem); if (sbi->s_snapshot_cno != nilfs_last_cno(nilfs)) { printk(KERN_WARNING "NILFS (device %s): couldn't " "remount because the current RO-mount is not " @@ -984,6 +993,7 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) down_write(&nilfs->ns_sem); nilfs_setup_super(sbi); + nilfs->ns_current = sbi; up_write(&nilfs->ns_sem); up(&sb->s_bdev->bd_mount_sem); @@ -1118,10 +1128,23 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, } down(&sd.bdev->bd_mount_sem); - if (!sd.cno && - (err = test_exclusive_mount(fs_type, sd.bdev, flags ^ MS_RDONLY))) { - err = (err < 0) ? : -EBUSY; - goto failed_unlock; + + if (!sd.cno) { + /* + * Check if an exclusive mount exists or not. + * Snapshot mounts coexist with a current mount + * (i.e. rw-mount or ro-mount), whereas rw-mount and + * ro-mount are mutually exclusive. + */ + down_read(&nilfs->ns_sem); + if (nilfs->ns_current && + ((nilfs->ns_current->s_super->s_flags ^ flags) + & MS_RDONLY)) { + up_read(&nilfs->ns_sem); + err = -EBUSY; + goto failed_unlock; + } + up_read(&nilfs->ns_sem); } /* @@ -1182,57 +1205,6 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, return err; } -static int nilfs_test_bdev_super3(struct super_block *s, void *data) -{ - struct nilfs_super_data *sd = data; - int ret; - - if (s->s_bdev != sd->bdev) - return 0; - if (down_read_trylock(&s->s_umount)) { - ret = (s->s_flags & MS_RDONLY) && s->s_root && - nilfs_test_opt(NILFS_SB(s), SNAPSHOT); - up_read(&s->s_umount); - if (ret) - return 0; /* ignore snapshot mounts */ - } - return !((sd->flags ^ s->s_flags) & MS_RDONLY); -} - -static int __false_bdev_super(struct super_block *s, void *data) -{ -#if 0 /* XXX: workaround for lock debug. This is not good idea */ - up_write(&s->s_umount); -#endif - return -EFAULT; -} - -/** - * test_exclusive_mount - check whether an exclusive RW/RO mount exists or not. - * fs_type: filesystem type - * bdev: block device - * flag: 0 (check rw-mount) or MS_RDONLY (check ro-mount) - * res: pointer to an integer to store result - * - * This function must be called within a section protected by bd_mount_mutex. - */ -static int test_exclusive_mount(struct file_system_type *fs_type, - struct block_device *bdev, int flags) -{ - struct super_block *s; - struct nilfs_super_data sd = { .flags = flags, .bdev = bdev }; - - s = sget(fs_type, nilfs_test_bdev_super3, __false_bdev_super, &sd); - if (IS_ERR(s)) { - if (PTR_ERR(s) != -EFAULT) - return PTR_ERR(s); - return 0; /* Not found */ - } - up_write(&s->s_umount); - deactivate_super(s); - return 1; /* Found */ -} - struct file_system_type nilfs_fs_type = { .owner = THIS_MODULE, .name = "nilfs2", diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 116caf96e7f3..99f7e29a5335 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -50,6 +50,7 @@ enum { * @ns_sem: semaphore for shared states * @ns_writer_mutex: mutex protecting ns_writer attach/detach * @ns_writer_refcount: number of referrers on ns_writer + * @ns_current: back pointer to current mount * @ns_sbh: buffer heads of on-disk super blocks * @ns_sbp: pointers to super block data * @ns_sbwtime: previous write time of super blocks @@ -98,6 +99,8 @@ struct the_nilfs { struct mutex ns_writer_mutex; atomic_t ns_writer_refcount; + struct nilfs_sb_info *ns_current; + /* * used for * - loading the latest checkpoint exclusively. From 6dd4740662405a68bb229ac2b9e0aeaaf2188bf2 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 8 Jun 2009 01:39:31 +0900 Subject: [PATCH 85/87] nilfs2: simplify remaining sget() use This simplifies the test function passed on the remaining sget() callsite in nilfs. Instead of checking mount type (i.e. ro-mount/rw-mount/snapshot mount) in the test function passed to sget(), this patch first looks up the nilfs_sb_info struct which the given mount type matches, and then acquires the super block instance holding the nilfs_sb_info. Signed-off-by: Ryusuke Konishi Cc: Al Viro Signed-off-by: Al Viro --- fs/nilfs2/sb.h | 1 + fs/nilfs2/super.c | 40 ++++++++++++++-------------------- fs/nilfs2/the_nilfs.c | 50 +++++++++++++++++++++++++++++++++++++++++++ fs/nilfs2/the_nilfs.h | 7 ++++++ 4 files changed, 74 insertions(+), 24 deletions(-) diff --git a/fs/nilfs2/sb.h b/fs/nilfs2/sb.h index adccd4fc654e..0776ccc2504a 100644 --- a/fs/nilfs2/sb.h +++ b/fs/nilfs2/sb.h @@ -60,6 +60,7 @@ struct nilfs_sb_info { struct super_block *s_super; /* reverse pointer to super_block */ struct the_nilfs *s_nilfs; struct list_head s_list; /* list head for nilfs->ns_supers */ + atomic_t s_count; /* reference count */ /* Segment constructor */ struct list_head s_dirty_files; /* dirty files list */ diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 5a8c5e4731b3..1d1b6e125159 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -336,7 +336,7 @@ static void nilfs_put_super(struct super_block *sb) put_nilfs(sbi->s_nilfs); sbi->s_super = NULL; sb->s_fs_info = NULL; - kfree(sbi); + nilfs_put_sbinfo(sbi); unlock_kernel(); } @@ -785,6 +785,7 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent, get_nilfs(nilfs); sbi->s_nilfs = nilfs; sbi->s_super = sb; + atomic_set(&sbi->s_count, 1); err = init_nilfs(nilfs, sbi, (char *)data); if (err) @@ -902,7 +903,7 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent, failed_sbi: put_nilfs(nilfs); sb->s_fs_info = NULL; - kfree(sbi); + nilfs_put_sbinfo(sbi); return err; } @@ -1014,6 +1015,7 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) struct nilfs_super_data { struct block_device *bdev; + struct nilfs_sb_info *sbi; __u64 cno; int flags; }; @@ -1071,27 +1073,8 @@ static int nilfs_set_bdev_super(struct super_block *s, void *data) static int nilfs_test_bdev_super(struct super_block *s, void *data) { struct nilfs_super_data *sd = data; - int ret; - if (s->s_bdev != sd->bdev) - return 0; - - if (!((s->s_flags | sd->flags) & MS_RDONLY)) - return 1; /* Reuse an old R/W-mode super_block */ - - if (s->s_flags & sd->flags & MS_RDONLY) { - if (down_read_trylock(&s->s_umount)) { - ret = s->s_root && - (sd->cno == NILFS_SB(s)->s_snapshot_cno); - up_read(&s->s_umount); - /* - * This path is locked with sb_lock by sget(). - * So, drop_super() causes deadlock. - */ - return ret; - } - } - return 0; + return sd->sbi && s->s_fs_info == (void *)sd->sbi; } static int @@ -1112,7 +1095,6 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, * much more information than normal filesystems to identify mount * instance. For snapshot mounts, not only a mount type (ro-mount * or rw-mount) but also a checkpoint number is required. - * The results are passed in sget() using nilfs_super_data. */ sd.cno = 0; sd.flags = flags; @@ -1148,13 +1130,23 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, } /* - * Search specified snapshot or R/W mode super_block + * Find existing nilfs_sb_info struct */ + sd.sbi = nilfs_find_sbinfo(nilfs, !(flags & MS_RDONLY), sd.cno); + if (!sd.cno) /* trying to get the latest checkpoint. */ sd.cno = nilfs_last_cno(nilfs); + /* + * Get super block instance holding the nilfs_sb_info struct. + * A new instance is allocated if no existing mount is present or + * existing instance has been unmounted. + */ s = sget(fs_type, nilfs_test_bdev_super, nilfs_set_bdev_super, &sd); + if (sd.sbi) + nilfs_put_sbinfo(sd.sbi); + if (IS_ERR(s)) { err = PTR_ERR(s); goto failed_unlock; diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 45dbf6a61744..221953bfc859 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -664,6 +664,56 @@ int nilfs_near_disk_full(struct the_nilfs *nilfs) return ret; } +/** + * nilfs_find_sbinfo - find existing nilfs_sb_info structure + * @nilfs: nilfs object + * @rw_mount: mount type (non-zero value for read/write mount) + * @cno: checkpoint number (zero for read-only mount) + * + * nilfs_find_sbinfo() returns the nilfs_sb_info structure which + * @rw_mount and @cno (in case of snapshots) matched. If no instance + * was found, NULL is returned. Although the super block instance can + * be unmounted after this function returns, the nilfs_sb_info struct + * is kept on memory until nilfs_put_sbinfo() is called. + */ +struct nilfs_sb_info *nilfs_find_sbinfo(struct the_nilfs *nilfs, + int rw_mount, __u64 cno) +{ + struct nilfs_sb_info *sbi; + + down_read(&nilfs->ns_sem); + /* + * The SNAPSHOT flag and sb->s_flags are supposed to be + * protected with nilfs->ns_sem. + */ + sbi = nilfs->ns_current; + if (rw_mount) { + if (sbi && !(sbi->s_super->s_flags & MS_RDONLY)) + goto found; /* read/write mount */ + else + goto out; + } else if (cno == 0) { + if (sbi && (sbi->s_super->s_flags & MS_RDONLY)) + goto found; /* read-only mount */ + else + goto out; + } + + list_for_each_entry(sbi, &nilfs->ns_supers, s_list) { + if (nilfs_test_opt(sbi, SNAPSHOT) && + sbi->s_snapshot_cno == cno) + goto found; /* snapshot mount */ + } + out: + up_read(&nilfs->ns_sem); + return NULL; + + found: + atomic_inc(&sbi->s_count); + up_read(&nilfs->ns_sem); + return sbi; +} + int nilfs_checkpoint_is_mounted(struct the_nilfs *nilfs, __u64 cno, int snapshot_mount) { diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 99f7e29a5335..be4c040fd629 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -201,6 +201,7 @@ void put_nilfs(struct the_nilfs *); int init_nilfs(struct the_nilfs *, struct nilfs_sb_info *, char *); int load_nilfs(struct the_nilfs *, struct nilfs_sb_info *); int nilfs_count_free_blocks(struct the_nilfs *, sector_t *); +struct nilfs_sb_info *nilfs_find_sbinfo(struct the_nilfs *, int, __u64); int nilfs_checkpoint_is_mounted(struct the_nilfs *, __u64, int); int nilfs_near_disk_full(struct the_nilfs *); void nilfs_fall_back_super_block(struct the_nilfs *); @@ -243,6 +244,12 @@ nilfs_detach_writer(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi) mutex_unlock(&nilfs->ns_writer_mutex); } +static inline void nilfs_put_sbinfo(struct nilfs_sb_info *sbi) +{ + if (!atomic_dec_and_test(&sbi->s_count)) + kfree(sbi); +} + static inline void nilfs_get_segment_range(struct the_nilfs *nilfs, __u64 segnum, sector_t *seg_start, sector_t *seg_end) From e59399d0102c1813cec48db5cebe1750313f88a0 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 8 Jun 2009 01:39:32 +0900 Subject: [PATCH 86/87] nilfs2: correct exclusion control in nilfs_remount function nilfs_remount() changes mount state of a superblock instance. Even though nilfs accesses other superblock instances during mount or remount, the mount state was not properly protected in nilfs_remount(). Moreover, nilfs_remount() has a lock order reversal problem; nilfs_get_sb() holds: 1. bdev->bd_mount_sem 2. sb->s_umount (sget acquires) and nilfs_remount() holds: 1. sb->s_umount (locked by the caller in vfs) 2. bdev->bd_mount_sem To avoid these problems, this patch divides a semaphore protecting super block instances from nilfs->ns_sem, and applies it to the mount state protection in nilfs_remount(). With this change, bd_mount_sem use is removed from nilfs_remount() and the lock order reversal will be resolved. And the new rw-semaphore, nilfs->ns_super_sem will properly protect the mount state except the modification from nilfs_error function. Signed-off-by: Ryusuke Konishi Signed-off-by: Al Viro --- fs/nilfs2/super.c | 44 ++++++++++++++++++++----------------------- fs/nilfs2/the_nilfs.c | 13 +++++++------ fs/nilfs2/the_nilfs.h | 7 ++++++- 3 files changed, 33 insertions(+), 31 deletions(-) diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 1d1b6e125159..f02762fa8ea0 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -327,10 +327,10 @@ static void nilfs_put_super(struct super_block *sb) nilfs_commit_super(sbi, 1); up_write(&nilfs->ns_sem); } - down_write(&nilfs->ns_sem); + down_write(&nilfs->ns_super_sem); if (nilfs->ns_current == sbi) nilfs->ns_current = NULL; - up_write(&nilfs->ns_sem); + up_write(&nilfs->ns_super_sem); nilfs_detach_checkpoint(sbi); put_nilfs(sbi->s_nilfs); @@ -408,9 +408,9 @@ int nilfs_attach_checkpoint(struct nilfs_sb_info *sbi, __u64 cno) struct buffer_head *bh_cp; int err; - down_write(&nilfs->ns_sem); + down_write(&nilfs->ns_super_sem); list_add(&sbi->s_list, &nilfs->ns_supers); - up_write(&nilfs->ns_sem); + up_write(&nilfs->ns_super_sem); sbi->s_ifile = nilfs_mdt_new( nilfs, sbi->s_super, NILFS_IFILE_INO, NILFS_IFILE_GFP); @@ -448,9 +448,9 @@ int nilfs_attach_checkpoint(struct nilfs_sb_info *sbi, __u64 cno) nilfs_mdt_destroy(sbi->s_ifile); sbi->s_ifile = NULL; - down_write(&nilfs->ns_sem); + down_write(&nilfs->ns_super_sem); list_del_init(&sbi->s_list); - up_write(&nilfs->ns_sem); + up_write(&nilfs->ns_super_sem); return err; } @@ -462,9 +462,9 @@ void nilfs_detach_checkpoint(struct nilfs_sb_info *sbi) nilfs_mdt_clear(sbi->s_ifile); nilfs_mdt_destroy(sbi->s_ifile); sbi->s_ifile = NULL; - down_write(&nilfs->ns_sem); + down_write(&nilfs->ns_super_sem); list_del_init(&sbi->s_list); - up_write(&nilfs->ns_sem); + up_write(&nilfs->ns_super_sem); } static int nilfs_mark_recovery_complete(struct nilfs_sb_info *sbi) @@ -883,10 +883,10 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent, goto failed_root; } - down_write(&nilfs->ns_sem); + down_write(&nilfs->ns_super_sem); if (!nilfs_test_opt(sbi, SNAPSHOT)) nilfs->ns_current = sbi; - up_write(&nilfs->ns_sem); + up_write(&nilfs->ns_super_sem); return 0; @@ -918,6 +918,7 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) lock_kernel(); + down_write(&nilfs->ns_super_sem); old_sb_flags = sb->s_flags; old_opts.mount_opt = sbi->s_mount_opt; old_opts.snapshot_cno = sbi->s_snapshot_cno; @@ -965,24 +966,20 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) * store the current valid flag. (It may have been changed * by fsck since we originally mounted the partition.) */ - down(&sb->s_bdev->bd_mount_sem); - down_read(&nilfs->ns_sem); if (nilfs->ns_current && nilfs->ns_current != sbi) { printk(KERN_WARNING "NILFS (device %s): couldn't " "remount because an RW-mount exists.\n", sb->s_id); - up_read(&nilfs->ns_sem); err = -EBUSY; - goto rw_remount_failed; + goto restore_opts; } - up_read(&nilfs->ns_sem); if (sbi->s_snapshot_cno != nilfs_last_cno(nilfs)) { printk(KERN_WARNING "NILFS (device %s): couldn't " "remount because the current RO-mount is not " "the latest one.\n", sb->s_id); err = -EINVAL; - goto rw_remount_failed; + goto restore_opts; } sb->s_flags &= ~MS_RDONLY; nilfs_clear_opt(sbi, SNAPSHOT); @@ -990,25 +987,24 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) err = nilfs_attach_segment_constructor(sbi); if (err) - goto rw_remount_failed; + goto restore_opts; down_write(&nilfs->ns_sem); nilfs_setup_super(sbi); - nilfs->ns_current = sbi; up_write(&nilfs->ns_sem); - up(&sb->s_bdev->bd_mount_sem); + nilfs->ns_current = sbi; } out: + up_write(&nilfs->ns_super_sem); unlock_kernel(); return 0; - rw_remount_failed: - up(&sb->s_bdev->bd_mount_sem); restore_opts: sb->s_flags = old_sb_flags; sbi->s_mount_opt = old_opts.mount_opt; sbi->s_snapshot_cno = old_opts.snapshot_cno; + up_write(&nilfs->ns_super_sem); unlock_kernel(); return err; } @@ -1118,15 +1114,15 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, * (i.e. rw-mount or ro-mount), whereas rw-mount and * ro-mount are mutually exclusive. */ - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_super_sem); if (nilfs->ns_current && ((nilfs->ns_current->s_super->s_flags ^ flags) & MS_RDONLY)) { - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_super_sem); err = -EBUSY; goto failed_unlock; } - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_super_sem); } /* diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 221953bfc859..06e8dfd538d6 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -72,6 +72,7 @@ static struct the_nilfs *alloc_nilfs(struct block_device *bdev) atomic_set(&nilfs->ns_writer_refcount, -1); atomic_set(&nilfs->ns_ndirtyblks, 0); init_rwsem(&nilfs->ns_sem); + init_rwsem(&nilfs->ns_super_sem); mutex_init(&nilfs->ns_writer_mutex); INIT_LIST_HEAD(&nilfs->ns_list); INIT_LIST_HEAD(&nilfs->ns_supers); @@ -681,10 +682,10 @@ struct nilfs_sb_info *nilfs_find_sbinfo(struct the_nilfs *nilfs, { struct nilfs_sb_info *sbi; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_super_sem); /* * The SNAPSHOT flag and sb->s_flags are supposed to be - * protected with nilfs->ns_sem. + * protected with nilfs->ns_super_sem. */ sbi = nilfs->ns_current; if (rw_mount) { @@ -705,12 +706,12 @@ struct nilfs_sb_info *nilfs_find_sbinfo(struct the_nilfs *nilfs, goto found; /* snapshot mount */ } out: - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_super_sem); return NULL; found: atomic_inc(&sbi->s_count); - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_super_sem); return sbi; } @@ -720,7 +721,7 @@ int nilfs_checkpoint_is_mounted(struct the_nilfs *nilfs, __u64 cno, struct nilfs_sb_info *sbi; int ret = 0; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_super_sem); if (cno == 0 || cno > nilfs->ns_cno) goto out_unlock; @@ -737,6 +738,6 @@ int nilfs_checkpoint_is_mounted(struct the_nilfs *nilfs, __u64 cno, ret++; out_unlock: - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_super_sem); return ret; } diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index be4c040fd629..d0cf4fb7c9ce 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -48,6 +48,7 @@ enum { * @ns_bdi: backing dev info * @ns_writer: back pointer to writable nilfs_sb_info * @ns_sem: semaphore for shared states + * @ns_super_sem: semaphore for global operations across super block instances * @ns_writer_mutex: mutex protecting ns_writer attach/detach * @ns_writer_refcount: number of referrers on ns_writer * @ns_current: back pointer to current mount @@ -96,10 +97,15 @@ struct the_nilfs { struct backing_dev_info *ns_bdi; struct nilfs_sb_info *ns_writer; struct rw_semaphore ns_sem; + struct rw_semaphore ns_super_sem; struct mutex ns_writer_mutex; atomic_t ns_writer_refcount; + /* + * components protected by ns_super_sem + */ struct nilfs_sb_info *ns_current; + struct list_head ns_supers; /* * used for @@ -113,7 +119,6 @@ struct the_nilfs { time_t ns_sbwtime[2]; unsigned ns_sbsize; unsigned ns_mount_state; - struct list_head ns_supers; /* * Following fields are dedicated to a writable FS-instance. From aa7dfb8954ccf49e026ba13d12991a4eb7defb96 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Mon, 8 Jun 2009 01:39:33 +0900 Subject: [PATCH 87/87] nilfs2: get rid of bd_mount_sem use from nilfs This will remove every bd_mount_sem use in nilfs. The intended exclusion control was replaced by the previous patch ("nilfs2: correct exclusion control in nilfs_remount function") for nilfs_remount(), and this patch will replace remains with a new mutex that this inserts in nilfs object. Signed-off-by: Ryusuke Konishi Cc: Christoph Hellwig Signed-off-by: Al Viro --- fs/nilfs2/cpfile.c | 6 +++--- fs/nilfs2/super.c | 12 ++++++------ fs/nilfs2/the_nilfs.c | 1 + fs/nilfs2/the_nilfs.h | 2 ++ 4 files changed, 12 insertions(+), 9 deletions(-) diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 300f1cdfa862..cadd36b14d07 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -864,11 +864,11 @@ int nilfs_cpfile_change_cpmode(struct inode *cpfile, __u64 cno, int mode) case NILFS_CHECKPOINT: /* * Check for protecting existing snapshot mounts: - * bd_mount_sem is used to make this operation atomic and + * ns_mount_mutex is used to make this operation atomic and * exclusive with a new mount job. Though it doesn't cover * umount, it's enough for the purpose. */ - down(&nilfs->ns_bdev->bd_mount_sem); + mutex_lock(&nilfs->ns_mount_mutex); if (nilfs_checkpoint_is_mounted(nilfs, cno, 1)) { /* Current implementation does not have to protect plain read-only mounts since they are exclusive @@ -877,7 +877,7 @@ int nilfs_cpfile_change_cpmode(struct inode *cpfile, __u64 cno, int mode) ret = -EBUSY; } else ret = nilfs_cpfile_clear_snapshot(cpfile, cno); - up(&nilfs->ns_bdev->bd_mount_sem); + mutex_unlock(&nilfs->ns_mount_mutex); return ret; case NILFS_SNAPSHOT: return nilfs_cpfile_set_snapshot(cpfile, cno); diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index f02762fa8ea0..1777a3467bd2 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -764,7 +764,7 @@ int nilfs_store_magic_and_option(struct super_block *sb, * @silent: silent mode flag * @nilfs: the_nilfs struct * - * This function is called exclusively by bd_mount_mutex. + * This function is called exclusively by nilfs->ns_mount_mutex. * So, the recovery process is protected from other simultaneous mounts. */ static int @@ -1105,7 +1105,7 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, goto failed; } - down(&sd.bdev->bd_mount_sem); + mutex_lock(&nilfs->ns_mount_mutex); if (!sd.cno) { /* @@ -1164,7 +1164,7 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, need_to_close = 0; } - up(&sd.bdev->bd_mount_sem); + mutex_unlock(&nilfs->ns_mount_mutex); put_nilfs(nilfs); if (need_to_close) close_bdev_exclusive(sd.bdev, flags); @@ -1172,7 +1172,7 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, return 0; failed_unlock: - up(&sd.bdev->bd_mount_sem); + mutex_unlock(&nilfs->ns_mount_mutex); put_nilfs(nilfs); failed: close_bdev_exclusive(sd.bdev, flags); @@ -1181,14 +1181,14 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags, cancel_new: /* Abandoning the newly allocated superblock */ - up(&sd.bdev->bd_mount_sem); + mutex_unlock(&nilfs->ns_mount_mutex); put_nilfs(nilfs); up_write(&s->s_umount); deactivate_super(s); /* * deactivate_super() invokes close_bdev_exclusive(). * We must finish all post-cleaning before this call; - * put_nilfs() and unlocking bd_mount_sem need the block device. + * put_nilfs() needs the block device. */ return err; } diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 06e8dfd538d6..e4e5c78bcc93 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -73,6 +73,7 @@ static struct the_nilfs *alloc_nilfs(struct block_device *bdev) atomic_set(&nilfs->ns_ndirtyblks, 0); init_rwsem(&nilfs->ns_sem); init_rwsem(&nilfs->ns_super_sem); + mutex_init(&nilfs->ns_mount_mutex); mutex_init(&nilfs->ns_writer_mutex); INIT_LIST_HEAD(&nilfs->ns_list); INIT_LIST_HEAD(&nilfs->ns_supers); diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index d0cf4fb7c9ce..e8adbffc626f 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -49,6 +49,7 @@ enum { * @ns_writer: back pointer to writable nilfs_sb_info * @ns_sem: semaphore for shared states * @ns_super_sem: semaphore for global operations across super block instances + * @ns_mount_mutex: mutex protecting mount process of nilfs * @ns_writer_mutex: mutex protecting ns_writer attach/detach * @ns_writer_refcount: number of referrers on ns_writer * @ns_current: back pointer to current mount @@ -98,6 +99,7 @@ struct the_nilfs { struct nilfs_sb_info *ns_writer; struct rw_semaphore ns_sem; struct rw_semaphore ns_super_sem; + struct mutex ns_mount_mutex; struct mutex ns_writer_mutex; atomic_t ns_writer_refcount;