From 67e55205ec55cc7899f1d783f217961596798419 Mon Sep 17 00:00:00 2001 From: Alessio Igor Bogani Date: Fri, 24 Apr 2009 09:06:53 +0200 Subject: [PATCH 01/22] vfs: umount_begin BKL pushdown Push BKL down into ->umount_begin() Signed-off-by: Alessio Igor Bogani Signed-off-by: Al Viro --- fs/9p/vfs_super.c | 6 +++++- fs/cifs/cifsfs.c | 3 +++ fs/fuse/inode.c | 3 +++ fs/namespace.c | 2 -- fs/nfs/super.c | 7 ++++++- 5 files changed, 17 insertions(+), 4 deletions(-) diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 5f8ab8adb5f5..7d23214e5599 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -37,6 +37,7 @@ #include #include #include +#include #include #include @@ -230,9 +231,12 @@ static int v9fs_show_options(struct seq_file *m, struct vfsmount *mnt) static void v9fs_umount_begin(struct super_block *sb) { - struct v9fs_session_info *v9ses = sb->s_fs_info; + struct v9fs_session_info *v9ses; + lock_kernel(); + v9ses = sb->s_fs_info; v9fs_session_cancel(v9ses); + unlock_kernel(); } static const struct super_operations v9fs_super_ops = { diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 0d6d8b573652..355e0efec0cf 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -35,6 +35,7 @@ #include #include #include +#include #include "cifsfs.h" #include "cifspdu.h" #define DECLARE_GLOBALS_HERE @@ -530,6 +531,7 @@ static void cifs_umount_begin(struct super_block *sb) if (tcon == NULL) return; + lock_kernel(); read_lock(&cifs_tcp_ses_lock); if (tcon->tc_count == 1) tcon->tidStatus = CifsExiting; @@ -548,6 +550,7 @@ static void cifs_umount_begin(struct super_block *sb) } /* BB FIXME - finish add checks for tidStatus BB */ + unlock_kernel(); return; } diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 459b73dd45e1..d1bc4d33ccbc 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -19,6 +19,7 @@ #include #include #include +#include MODULE_AUTHOR("Miklos Szeredi "); MODULE_DESCRIPTION("Filesystem in Userspace"); @@ -259,7 +260,9 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid, static void fuse_umount_begin(struct super_block *sb) { + lock_kernel(); fuse_abort_conn(get_fuse_conn_super(sb)); + unlock_kernel(); } static void fuse_send_destroy(struct fuse_conn *fc) diff --git a/fs/namespace.c b/fs/namespace.c index 41196209a906..0d2003fb4377 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1073,9 +1073,7 @@ static int do_umount(struct vfsmount *mnt, int flags) */ if (flags & MNT_FORCE && sb->s_op->umount_begin) { - lock_kernel(); sb->s_op->umount_begin(sb); - unlock_kernel(); } /* diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 6717200923fe..1679a164c8c9 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -683,9 +683,12 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) */ static void nfs_umount_begin(struct super_block *sb) { - struct nfs_server *server = NFS_SB(sb); + struct nfs_server *server; struct rpc_clnt *rpc; + lock_kernel(); + + server = NFS_SB(sb); /* -EIO all pending I/O */ rpc = server->client_acl; if (!IS_ERR(rpc)) @@ -693,6 +696,8 @@ static void nfs_umount_begin(struct super_block *sb) rpc = server->client; if (!IS_ERR(rpc)) rpc_killall_tasks(rpc); + + unlock_kernel(); } /* From 5a6059c358d3b3004cb6b8562e3b049b1f8564f9 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 1 May 2009 12:11:12 -0400 Subject: [PATCH 02/22] reiserfs: Expand i_mutex to enclose lookup_one_len 2.6.30-rc3 introduced some sanity checks in the VFS code to avoid NFS bugs by ensuring that lookup_one_len is always called under i_mutex. This patch expands the i_mutex locking to enclose lookup_one_len. This was always required, but not not enforced in the reiserfs code since it does locking around the xattr interactions with the xattr_sem. This is obvious enough, and it survived an overnight 50 thread ACL test. Signed-off-by: Jeff Mahoney Signed-off-by: Al Viro --- fs/reiserfs/xattr.c | 32 ++++++++++++++------------------ 1 file changed, 14 insertions(+), 18 deletions(-) diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index f83f52bae390..31a3dbb120e1 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -120,25 +120,20 @@ static struct dentry *lookup_or_create_dir(struct dentry *parent, struct dentry *dentry; BUG_ON(!parent); + mutex_lock_nested(&parent->d_inode->i_mutex, I_MUTEX_XATTR); dentry = lookup_one_len(name, parent, strlen(name)); - if (IS_ERR(dentry)) - return dentry; - else if (!dentry->d_inode) { + if (!IS_ERR(dentry) && !dentry->d_inode) { int err = -ENODATA; - if (xattr_may_create(flags)) { - mutex_lock_nested(&parent->d_inode->i_mutex, - I_MUTEX_XATTR); + if (xattr_may_create(flags)) err = xattr_mkdir(parent->d_inode, dentry, 0700); - mutex_unlock(&parent->d_inode->i_mutex); - } if (err) { dput(dentry); dentry = ERR_PTR(err); } } - + mutex_unlock(&parent->d_inode->i_mutex); return dentry; } @@ -184,6 +179,7 @@ fill_with_dentries(void *buf, const char *name, int namelen, loff_t offset, { struct reiserfs_dentry_buf *dbuf = buf; struct dentry *dentry; + WARN_ON_ONCE(!mutex_is_locked(&dbuf->xadir->d_inode->i_mutex)); if (dbuf->count == ARRAY_SIZE(dbuf->dentries)) return -ENOSPC; @@ -349,6 +345,7 @@ static struct dentry *xattr_lookup(struct inode *inode, const char *name, if (IS_ERR(xadir)) return ERR_CAST(xadir); + mutex_lock_nested(&xadir->d_inode->i_mutex, I_MUTEX_XATTR); xafile = lookup_one_len(name, xadir, strlen(name)); if (IS_ERR(xafile)) { err = PTR_ERR(xafile); @@ -360,18 +357,15 @@ static struct dentry *xattr_lookup(struct inode *inode, const char *name, if (!xafile->d_inode) { err = -ENODATA; - if (xattr_may_create(flags)) { - mutex_lock_nested(&xadir->d_inode->i_mutex, - I_MUTEX_XATTR); + if (xattr_may_create(flags)) err = xattr_create(xadir->d_inode, xafile, 0700|S_IFREG); - mutex_unlock(&xadir->d_inode->i_mutex); - } } if (err) dput(xafile); out: + mutex_unlock(&xadir->d_inode->i_mutex); dput(xadir); if (err) return ERR_PTR(err); @@ -435,6 +429,7 @@ static int lookup_and_delete_xattr(struct inode *inode, const char *name) if (IS_ERR(xadir)) return PTR_ERR(xadir); + mutex_lock_nested(&xadir->d_inode->i_mutex, I_MUTEX_XATTR); dentry = lookup_one_len(name, xadir, strlen(name)); if (IS_ERR(dentry)) { err = PTR_ERR(dentry); @@ -442,14 +437,13 @@ static int lookup_and_delete_xattr(struct inode *inode, const char *name) } if (dentry->d_inode) { - mutex_lock_nested(&xadir->d_inode->i_mutex, I_MUTEX_XATTR); err = xattr_unlink(xadir->d_inode, dentry); - mutex_unlock(&xadir->d_inode->i_mutex); update_ctime(inode); } dput(dentry); out_dput: + mutex_unlock(&xadir->d_inode->i_mutex); dput(xadir); return err; } @@ -906,9 +900,9 @@ static int create_privroot(struct dentry *dentry) { int err; struct inode *inode = dentry->d_parent->d_inode; - mutex_lock_nested(&inode->i_mutex, I_MUTEX_XATTR); + WARN_ON_ONCE(!mutex_is_locked(&inode->i_mutex)); + err = xattr_mkdir(inode, dentry, 0700); - mutex_unlock(&inode->i_mutex); if (err) { dput(dentry); dentry = NULL; @@ -980,6 +974,7 @@ int reiserfs_xattr_init(struct super_block *s, int mount_flags) /* If we don't have the privroot located yet - go find it */ if (!REISERFS_SB(s)->priv_root) { struct dentry *dentry; + mutex_lock_nested(&s->s_root->d_inode->i_mutex, I_MUTEX_CHILD); dentry = lookup_one_len(PRIVROOT_NAME, s->s_root, strlen(PRIVROOT_NAME)); if (!IS_ERR(dentry)) { @@ -993,6 +988,7 @@ int reiserfs_xattr_init(struct super_block *s, int mount_flags) } } else err = PTR_ERR(dentry); + mutex_unlock(&s->s_root->d_inode->i_mutex); if (!err && dentry) { s->s_root->d_op = &xattr_lookup_poison_ops; From edcc37a0478836b4a51eafb1bcec6a52708f681d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 3 May 2009 06:00:05 -0400 Subject: [PATCH 03/22] Always lookup priv_root on reiserfs mount and keep it ... even if it's a negative dentry. That way we can set ->d_op on root before anyone could race with us. Simplify d_compare(), while we are at it. Signed-off-by: Al Viro --- fs/reiserfs/super.c | 6 ++- fs/reiserfs/xattr.c | 86 ++++++++++++++-------------------- include/linux/reiserfs_xattr.h | 1 + 3 files changed, 41 insertions(+), 52 deletions(-) diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 0ae6486d9046..d444fe0013a4 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -1842,7 +1842,8 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) goto error; } - if ((errval = reiserfs_xattr_init(s, s->s_flags))) { + if ((errval = reiserfs_lookup_privroot(s)) || + (errval = reiserfs_xattr_init(s, s->s_flags))) { dput(s->s_root); s->s_root = NULL; goto error; @@ -1855,7 +1856,8 @@ static int reiserfs_fill_super(struct super_block *s, void *data, int silent) reiserfs_info(s, "using 3.5.x disk format\n"); } - if ((errval = reiserfs_xattr_init(s, s->s_flags))) { + if ((errval = reiserfs_lookup_privroot(s)) || + (errval = reiserfs_xattr_init(s, s->s_flags))) { dput(s->s_root); s->s_root = NULL; goto error; diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 31a3dbb120e1..2891f789f545 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -903,16 +903,19 @@ static int create_privroot(struct dentry *dentry) WARN_ON_ONCE(!mutex_is_locked(&inode->i_mutex)); err = xattr_mkdir(inode, dentry, 0700); - if (err) { - dput(dentry); - dentry = NULL; + if (err || !dentry->d_inode) { + reiserfs_warning(dentry->d_sb, "jdm-20006", + "xattrs/ACLs enabled and couldn't " + "find/create .reiserfs_priv. " + "Failing mount."); + return -EOPNOTSUPP; } - if (dentry && dentry->d_inode) - reiserfs_info(dentry->d_sb, "Created %s - reserved for xattr " - "storage.\n", PRIVROOT_NAME); + dentry->d_inode->i_flags |= S_PRIVATE; + reiserfs_info(dentry->d_sb, "Created %s - reserved for xattr " + "storage.\n", PRIVROOT_NAME); - return err; + return 0; } static int xattr_mount_check(struct super_block *s) @@ -944,11 +947,9 @@ static int xattr_lookup_poison(struct dentry *dentry, struct qstr *q1, struct qstr *name) { struct dentry *priv_root = REISERFS_SB(dentry->d_sb)->priv_root; - if (name->len == priv_root->d_name.len && - name->hash == priv_root->d_name.hash && - !memcmp(name->name, priv_root->d_name.name, name->len)) { + if (container_of(q1, struct dentry, d_name) == priv_root) return -ENOENT; - } else if (q1->len == name->len && + if (q1->len == name->len && !memcmp(q1->name, name->name, name->len)) return 0; return 1; @@ -958,6 +959,27 @@ static const struct dentry_operations xattr_lookup_poison_ops = { .d_compare = xattr_lookup_poison, }; +int reiserfs_lookup_privroot(struct super_block *s) +{ + struct dentry *dentry; + int err = 0; + + /* If we don't have the privroot located yet - go find it */ + mutex_lock(&s->s_root->d_inode->i_mutex); + dentry = lookup_one_len(PRIVROOT_NAME, s->s_root, + strlen(PRIVROOT_NAME)); + if (!IS_ERR(dentry)) { + REISERFS_SB(s)->priv_root = dentry; + s->s_root->d_op = &xattr_lookup_poison_ops; + if (dentry->d_inode) + dentry->d_inode->i_flags |= S_PRIVATE; + } else + err = PTR_ERR(dentry); + mutex_unlock(&s->s_root->d_inode->i_mutex); + + return err; +} + /* We need to take a copy of the mount flags since things like * MS_RDONLY don't get set until *after* we're called. * mount_flags != mount_options */ @@ -969,48 +991,12 @@ int reiserfs_xattr_init(struct super_block *s, int mount_flags) err = xattr_mount_check(s); if (err) goto error; -#endif - /* If we don't have the privroot located yet - go find it */ - if (!REISERFS_SB(s)->priv_root) { - struct dentry *dentry; - mutex_lock_nested(&s->s_root->d_inode->i_mutex, I_MUTEX_CHILD); - dentry = lookup_one_len(PRIVROOT_NAME, s->s_root, - strlen(PRIVROOT_NAME)); - if (!IS_ERR(dentry)) { -#ifdef CONFIG_REISERFS_FS_XATTR - if (!(mount_flags & MS_RDONLY) && !dentry->d_inode) - err = create_privroot(dentry); -#endif - if (!dentry->d_inode) { - dput(dentry); - dentry = NULL; - } - } else - err = PTR_ERR(dentry); + if (!REISERFS_SB(s)->priv_root->d_inode && !(mount_flags & MS_RDONLY)) { + mutex_lock(&s->s_root->d_inode->i_mutex); + err = create_privroot(REISERFS_SB(s)->priv_root); mutex_unlock(&s->s_root->d_inode->i_mutex); - - if (!err && dentry) { - s->s_root->d_op = &xattr_lookup_poison_ops; - dentry->d_inode->i_flags |= S_PRIVATE; - REISERFS_SB(s)->priv_root = dentry; -#ifdef CONFIG_REISERFS_FS_XATTR - /* xattrs are unavailable */ - } else if (!(mount_flags & MS_RDONLY)) { - /* If we're read-only it just means that the dir - * hasn't been created. Not an error -- just no - * xattrs on the fs. We'll check again if we - * go read-write */ - reiserfs_warning(s, "jdm-20006", - "xattrs/ACLs enabled and couldn't " - "find/create .reiserfs_priv. " - "Failing mount."); - err = -EOPNOTSUPP; -#endif - } } - -#ifdef CONFIG_REISERFS_FS_XATTR if (!err) s->s_xattr = reiserfs_xattr_handlers; diff --git a/include/linux/reiserfs_xattr.h b/include/linux/reiserfs_xattr.h index dcae01e63e40..fea1a8e65bef 100644 --- a/include/linux/reiserfs_xattr.h +++ b/include/linux/reiserfs_xattr.h @@ -38,6 +38,7 @@ struct nameidata; int reiserfs_xattr_register_handlers(void) __init; void reiserfs_xattr_unregister_handlers(void); int reiserfs_xattr_init(struct super_block *sb, int mount_flags); +int reiserfs_lookup_privroot(struct super_block *sb); int reiserfs_delete_xattrs(struct inode *inode); int reiserfs_chown_xattrs(struct inode *inode, struct iattr *attrs); From ab17c4f02156c4f75d7fa43a5aa2a7f942d47201 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 5 May 2009 15:30:15 -0400 Subject: [PATCH 04/22] reiserfs: fixup xattr_root caching The xattr_root caching was broken from my previous patch set. It wouldn't cause corruption, but could cause decreased performance due to allocating a larger chunk of the journal (~ 27 blocks) than it would actually use. This patch loads the xattr root dentry at xattr initialization and creates it on-demand. Since we're using the cached dentry, there's no point in keeping lookup_or_create_dir around, so that's removed. Signed-off-by: Jeff Mahoney Signed-off-by: Al Viro --- fs/reiserfs/xattr.c | 81 +++++++++++++++++++++------------- include/linux/reiserfs_fs_sb.h | 2 +- include/linux/reiserfs_xattr.h | 2 +- 3 files changed, 52 insertions(+), 33 deletions(-) diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 2891f789f545..c77984473db9 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -113,36 +113,28 @@ static int xattr_rmdir(struct inode *dir, struct dentry *dentry) #define xattr_may_create(flags) (!flags || flags & XATTR_CREATE) -/* Returns and possibly creates the xattr dir. */ -static struct dentry *lookup_or_create_dir(struct dentry *parent, - const char *name, int flags) -{ - struct dentry *dentry; - BUG_ON(!parent); - - mutex_lock_nested(&parent->d_inode->i_mutex, I_MUTEX_XATTR); - dentry = lookup_one_len(name, parent, strlen(name)); - if (!IS_ERR(dentry) && !dentry->d_inode) { - int err = -ENODATA; - - if (xattr_may_create(flags)) - err = xattr_mkdir(parent->d_inode, dentry, 0700); - - if (err) { - dput(dentry); - dentry = ERR_PTR(err); - } - } - mutex_unlock(&parent->d_inode->i_mutex); - return dentry; -} - static struct dentry *open_xa_root(struct super_block *sb, int flags) { struct dentry *privroot = REISERFS_SB(sb)->priv_root; - if (!privroot) + struct dentry *xaroot; + if (!privroot->d_inode) return ERR_PTR(-ENODATA); - return lookup_or_create_dir(privroot, XAROOT_NAME, flags); + + mutex_lock_nested(&privroot->d_inode->i_mutex, I_MUTEX_XATTR); + + xaroot = dget(REISERFS_SB(sb)->xattr_root); + if (!xaroot->d_inode) { + int err = -ENODATA; + if (xattr_may_create(flags)) + err = xattr_mkdir(privroot->d_inode, xaroot, 0700); + if (err) { + dput(xaroot); + xaroot = ERR_PTR(err); + } + } + + mutex_unlock(&privroot->d_inode->i_mutex); + return xaroot; } static struct dentry *open_xa_dir(const struct inode *inode, int flags) @@ -158,10 +150,22 @@ static struct dentry *open_xa_dir(const struct inode *inode, int flags) le32_to_cpu(INODE_PKEY(inode)->k_objectid), inode->i_generation); - xadir = lookup_or_create_dir(xaroot, namebuf, flags); + mutex_lock_nested(&xaroot->d_inode->i_mutex, I_MUTEX_XATTR); + + xadir = lookup_one_len(namebuf, xaroot, strlen(namebuf)); + if (!IS_ERR(xadir) && !xadir->d_inode) { + int err = -ENODATA; + if (xattr_may_create(flags)) + err = xattr_mkdir(xaroot->d_inode, xadir, 0700); + if (err) { + dput(xadir); + xadir = ERR_PTR(err); + } + } + + mutex_unlock(&xaroot->d_inode->i_mutex); dput(xaroot); return xadir; - } /* The following are side effects of other operations that aren't explicitly @@ -986,19 +990,33 @@ int reiserfs_lookup_privroot(struct super_block *s) int reiserfs_xattr_init(struct super_block *s, int mount_flags) { int err = 0; + struct dentry *privroot = REISERFS_SB(s)->priv_root; #ifdef CONFIG_REISERFS_FS_XATTR err = xattr_mount_check(s); if (err) goto error; - if (!REISERFS_SB(s)->priv_root->d_inode && !(mount_flags & MS_RDONLY)) { + if (!privroot->d_inode && !(mount_flags & MS_RDONLY)) { mutex_lock(&s->s_root->d_inode->i_mutex); err = create_privroot(REISERFS_SB(s)->priv_root); mutex_unlock(&s->s_root->d_inode->i_mutex); } - if (!err) + + if (privroot->d_inode) { s->s_xattr = reiserfs_xattr_handlers; + mutex_lock(&privroot->d_inode->i_mutex); + if (!REISERFS_SB(s)->xattr_root) { + struct dentry *dentry; + dentry = lookup_one_len(XAROOT_NAME, privroot, + strlen(XAROOT_NAME)); + if (!IS_ERR(dentry)) + REISERFS_SB(s)->xattr_root = dentry; + else + err = PTR_ERR(dentry); + } + mutex_unlock(&privroot->d_inode->i_mutex); + } error: if (err) { @@ -1008,11 +1026,12 @@ error: #endif /* The super_block MS_POSIXACL must mirror the (no)acl mount option. */ - s->s_flags = s->s_flags & ~MS_POSIXACL; #ifdef CONFIG_REISERFS_FS_POSIX_ACL if (reiserfs_posixacl(s)) s->s_flags |= MS_POSIXACL; + else #endif + s->s_flags &= ~MS_POSIXACL; return err; } diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index 6b361d23a499..8651640868a1 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -402,7 +402,7 @@ struct reiserfs_sb_info { int reserved_blocks; /* amount of blocks reserved for further allocations */ spinlock_t bitmap_lock; /* this lock on now only used to protect reserved_blocks variable */ struct dentry *priv_root; /* root of /.reiserfs_priv */ - struct dentry *xattr_root; /* root of /.reiserfs_priv/.xa */ + struct dentry *xattr_root; /* root of /.reiserfs_priv/xattrs */ int j_errno; #ifdef CONFIG_QUOTA char *s_qf_names[MAXQUOTAS]; diff --git a/include/linux/reiserfs_xattr.h b/include/linux/reiserfs_xattr.h index fea1a8e65bef..cdedc01036e4 100644 --- a/include/linux/reiserfs_xattr.h +++ b/include/linux/reiserfs_xattr.h @@ -98,7 +98,7 @@ static inline size_t reiserfs_xattr_jcreate_nblocks(struct inode *inode) if ((REISERFS_I(inode)->i_flags & i_has_xattr_dir) == 0) { nblocks += JOURNAL_BLOCKS_PER_OBJECT(inode->i_sb); - if (REISERFS_SB(inode->i_sb)->xattr_root == NULL) + if (!REISERFS_SB(inode->i_sb)->xattr_root->d_inode) nblocks += JOURNAL_BLOCKS_PER_OBJECT(inode->i_sb); } From b82bb72ba7df473461c5e2368a4e7497c8ce76e9 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 5 May 2009 15:30:16 -0400 Subject: [PATCH 05/22] reiserfs: dont associate security.* with xattr files The security.* xattrs are ignored for xattr files, so don't create them. Signed-off-by: Jeff Mahoney Signed-off-by: Al Viro --- fs/reiserfs/xattr_security.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/fs/reiserfs/xattr_security.c b/fs/reiserfs/xattr_security.c index 4d3c20e787c3..a92c8792c0f6 100644 --- a/fs/reiserfs/xattr_security.c +++ b/fs/reiserfs/xattr_security.c @@ -55,8 +55,16 @@ int reiserfs_security_init(struct inode *dir, struct inode *inode, struct reiserfs_security_handle *sec) { int blocks = 0; - int error = security_inode_init_security(inode, dir, &sec->name, - &sec->value, &sec->length); + int error; + + sec->name = NULL; + + /* Don't add selinux attributes on xattrs - they'll never get used */ + if (IS_PRIVATE(dir)) + return 0; + + error = security_inode_init_security(inode, dir, &sec->name, + &sec->value, &sec->length); if (error) { if (error == -EOPNOTSUPP) error = 0; From 677c9b2e393a0cd203bd54e9c18b012b2c73305a Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 5 May 2009 15:30:17 -0400 Subject: [PATCH 06/22] reiserfs: remove privroot hiding in lookup With Al Viro's patch to move privroot lookup to fs mount, there's no need to have special code to hide the privroot in reiserfs_lookup. I've also cleaned up the privroot hiding in reiserfs_readdir_dentry and removed the last user of reiserfs_xattrs(). Signed-off-by: Jeff Mahoney Signed-off-by: Al Viro --- fs/reiserfs/dir.c | 24 +++++++++++++----------- fs/reiserfs/namei.c | 17 ++--------------- fs/reiserfs/xattr.c | 2 +- include/linux/reiserfs_fs_sb.h | 1 - 4 files changed, 16 insertions(+), 28 deletions(-) diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c index 67a80d7e59e2..45ee3d357c70 100644 --- a/fs/reiserfs/dir.c +++ b/fs/reiserfs/dir.c @@ -41,6 +41,18 @@ static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry, #define store_ih(where,what) copy_item_head (where, what) +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; +} + int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent, filldir_t filldir, loff_t *pos) { @@ -138,18 +150,8 @@ int reiserfs_readdir_dentry(struct dentry *dentry, void *dirent, } /* Ignore the .reiserfs_priv entry */ - if (reiserfs_xattrs(inode->i_sb) && - !old_format_only(inode->i_sb) && - dentry == inode->i_sb->s_root && - REISERFS_SB(inode->i_sb)->priv_root && - REISERFS_SB(inode->i_sb)->priv_root->d_inode - && deh_objectid(deh) == - le32_to_cpu(INODE_PKEY - (REISERFS_SB(inode->i_sb)-> - priv_root->d_inode)-> - k_objectid)) { + if (is_privroot_deh(dentry, deh)) continue; - } d_off = deh_offset(deh); *pos = d_off; diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c index efd4d720718e..271579128634 100644 --- a/fs/reiserfs/namei.c +++ b/fs/reiserfs/namei.c @@ -338,21 +338,8 @@ static struct dentry *reiserfs_lookup(struct inode *dir, struct dentry *dentry, &path_to_entry, &de); pathrelse(&path_to_entry); if (retval == NAME_FOUND) { - /* Hide the .reiserfs_priv directory */ - if (reiserfs_xattrs(dir->i_sb) && - !old_format_only(dir->i_sb) && - REISERFS_SB(dir->i_sb)->priv_root && - REISERFS_SB(dir->i_sb)->priv_root->d_inode && - de.de_objectid == - le32_to_cpu(INODE_PKEY - (REISERFS_SB(dir->i_sb)->priv_root->d_inode)-> - k_objectid)) { - reiserfs_write_unlock(dir->i_sb); - return ERR_PTR(-EACCES); - } - - inode = - reiserfs_iget(dir->i_sb, (struct cpu_key *)&(de.de_dir_id)); + inode = reiserfs_iget(dir->i_sb, + (struct cpu_key *)&(de.de_dir_id)); if (!inode || IS_ERR(inode)) { reiserfs_write_unlock(dir->i_sb); return ERR_PTR(-EACCES); diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index c77984473db9..2237e10c7c7c 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -841,7 +841,7 @@ ssize_t reiserfs_listxattr(struct dentry * dentry, char *buffer, size_t size) if (!dentry->d_inode) return -EINVAL; - if (!reiserfs_xattrs(dentry->d_sb) || + if (!dentry->d_sb->s_xattr || get_inode_sd_version(dentry->d_inode) == STAT_DATA_V1) return -EOPNOTSUPP; diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h index 8651640868a1..6473650c28f1 100644 --- a/include/linux/reiserfs_fs_sb.h +++ b/include/linux/reiserfs_fs_sb.h @@ -488,7 +488,6 @@ enum reiserfs_mount_options { #define reiserfs_data_log(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_DATA_LOG)) #define reiserfs_data_ordered(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_DATA_ORDERED)) #define reiserfs_data_writeback(s) (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_DATA_WRITEBACK)) -#define reiserfs_xattrs(s) ((s)->s_xattr != NULL) #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_xattrs_optional(s) (reiserfs_xattrs_user(s) || reiserfs_posixacl(s)) From 74dbbdd7fdc11763f4698d2f3e684cf4446951e6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 6 May 2009 01:07:50 -0400 Subject: [PATCH 07/22] New helper: deactivate_locked_super() Does equivalent of up_write(&s->s_umount); deactivate_super(s); However, it does not does not unlock it until it's all over. As the result, it's safe to use to dispose of new superblock on ->get_sb() failure exits - nobody will see the sucker until it's all over. Equivalent using up_write/deactivate_super is safe for that purpose if superblock is either safe to use or has NULL ->s_root when we unlock. Normally filesystems take the required precautions, but a) we do have bugs in that area in some of them. b) up_write/deactivate_super sequence is extremely common, so the helper makes sense anyway. Signed-off-by: Al Viro --- fs/super.c | 46 ++++++++++++++++++++++++++++++++++------------ include/linux/fs.h | 1 + 2 files changed, 35 insertions(+), 12 deletions(-) diff --git a/fs/super.c b/fs/super.c index 786fe7d72790..a9dc4c33ef4d 100644 --- a/fs/super.c +++ b/fs/super.c @@ -207,6 +207,34 @@ void deactivate_super(struct super_block *s) EXPORT_SYMBOL(deactivate_super); +/** + * deactivate_locked_super - drop an active reference to superblock + * @s: superblock to deactivate + * + * Equivalent of up_write(&s->s_umount); deactivate_super(s);, except that + * it does not unlock it until it's all over. As the result, it's safe to + * use to dispose of new superblock on ->get_sb() failure exits - nobody + * will see the sucker until it's all over. Equivalent using up_write + + * deactivate_super is safe for that purpose only if superblock is either + * safe to use or has NULL ->s_root when we unlock. + */ +void deactivate_locked_super(struct super_block *s) +{ + struct file_system_type *fs = s->s_type; + if (atomic_dec_and_lock(&s->s_active, &sb_lock)) { + s->s_count -= S_BIAS-1; + spin_unlock(&sb_lock); + vfs_dq_off(s, 0); + fs->kill_sb(s); + put_filesystem(fs); + put_super(s); + } else { + up_write(&s->s_umount); + } +} + +EXPORT_SYMBOL(deactivate_locked_super); + /** * grab_super - acquire an active reference * @s: reference we are trying to make active @@ -797,8 +825,7 @@ int get_sb_ns(struct file_system_type *fs_type, int flags, void *data, sb->s_flags = flags; err = fill_super(sb, data, flags & MS_SILENT ? 1 : 0); if (err) { - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); return err; } @@ -854,8 +881,7 @@ int get_sb_bdev(struct file_system_type *fs_type, if (s->s_root) { if ((flags ^ s->s_flags) & MS_RDONLY) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); error = -EBUSY; goto error_bdev; } @@ -870,8 +896,7 @@ int get_sb_bdev(struct file_system_type *fs_type, sb_set_blocksize(s, block_size(bdev)); error = fill_super(s, data, flags & MS_SILENT ? 1 : 0); if (error) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); goto error; } @@ -921,8 +946,7 @@ int get_sb_nodev(struct file_system_type *fs_type, error = fill_super(s, data, flags & MS_SILENT ? 1 : 0); if (error) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); return error; } s->s_flags |= MS_ACTIVE; @@ -952,8 +976,7 @@ int get_sb_single(struct file_system_type *fs_type, s->s_flags = flags; error = fill_super(s, data, flags & MS_SILENT ? 1 : 0); if (error) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); return error; } s->s_flags |= MS_ACTIVE; @@ -1006,8 +1029,7 @@ vfs_kern_mount(struct file_system_type *type, int flags, const char *name, void return mnt; out_sb: dput(mnt->mnt_root); - up_write(&mnt->mnt_sb->s_umount); - deactivate_super(mnt->mnt_sb); + deactivate_locked_super(mnt->mnt_sb); out_free_secdata: free_secdata(secdata); out_mnt: diff --git a/include/linux/fs.h b/include/linux/fs.h index 5bed436f4353..11484d08042c 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1775,6 +1775,7 @@ void kill_block_super(struct super_block *sb); void kill_anon_super(struct super_block *sb); void kill_litter_super(struct super_block *sb); void deactivate_super(struct super_block *sb); +void deactivate_locked_super(struct super_block *sb); int set_anon_super(struct super_block *s, void *data); struct super_block *sget(struct file_system_type *type, int (*test)(struct super_block *,void *), From 6f5bbff9a1b7d6864a495763448a363bbfa96324 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 6 May 2009 01:34:22 -0400 Subject: [PATCH 08/22] Convert obvious places to deactivate_locked_super() Signed-off-by: Al Viro --- drivers/mtd/mtdsuper.c | 3 +-- fs/9p/vfs_super.c | 5 +---- fs/afs/super.c | 3 +-- fs/btrfs/super.c | 12 ++++-------- fs/cifs/cifsfs.c | 3 +-- fs/devpts/inode.c | 5 ++--- fs/ecryptfs/main.c | 5 ++--- fs/libfs.c | 3 +-- fs/nfs/super.c | 15 +++++---------- fs/proc/root.c | 3 +-- fs/ubifs/super.c | 3 +-- kernel/cgroup.c | 3 +-- 12 files changed, 21 insertions(+), 42 deletions(-) diff --git a/drivers/mtd/mtdsuper.c b/drivers/mtd/mtdsuper.c index 92285d0089c2..af8b42e0a55b 100644 --- a/drivers/mtd/mtdsuper.c +++ b/drivers/mtd/mtdsuper.c @@ -74,8 +74,7 @@ static int get_sb_mtd_aux(struct file_system_type *fs_type, int flags, ret = fill_super(sb, data, flags & MS_SILENT ? 1 : 0); if (ret < 0) { - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); return ret; } diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 7d23214e5599..0d29a57c63e6 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -174,10 +174,7 @@ P9_DPRINTK(P9_DEBUG_VFS, " simple set mount, return 0\n"); return 0; release_sb: - if (sb) { - up_write(&sb->s_umount); - deactivate_super(sb); - } + deactivate_locked_super(sb); free_stat: kfree(st); diff --git a/fs/afs/super.c b/fs/afs/super.c index aee239a048cb..2753f16dd315 100644 --- a/fs/afs/super.c +++ b/fs/afs/super.c @@ -405,8 +405,7 @@ static int afs_get_sb(struct file_system_type *fs_type, sb->s_flags = flags; ret = afs_fill_super(sb, ¶ms); if (ret < 0) { - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); goto error; } sb->s_options = new_opts; diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 3536bdb2d7cb..6dfae5b28f59 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -502,8 +502,7 @@ static int btrfs_get_sb(struct file_system_type *fs_type, int flags, if (s->s_root) { if ((flags ^ s->s_flags) & MS_RDONLY) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); error = -EBUSY; goto error_close_devices; } @@ -517,8 +516,7 @@ static int btrfs_get_sb(struct file_system_type *fs_type, int flags, error = btrfs_fill_super(s, fs_devices, data, flags & MS_SILENT ? 1 : 0); if (error) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); goto error_free_subvol_name; } @@ -535,15 +533,13 @@ static int btrfs_get_sb(struct file_system_type *fs_type, int flags, mutex_unlock(&s->s_root->d_inode->i_mutex); if (IS_ERR(root)) { - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); error = PTR_ERR(root); goto error_free_subvol_name; } if (!root->d_inode) { dput(root); - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); error = -ENXIO; goto error_free_subvol_name; } diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 355e0efec0cf..5e6d35804d73 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -602,8 +602,7 @@ cifs_get_sb(struct file_system_type *fs_type, rc = cifs_read_super(sb, data, dev_name, flags & MS_SILENT ? 1 : 0); if (rc) { - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); return rc; } sb->s_flags |= MS_ACTIVE; diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c index 63a4a59e4148..21165cf934ff 100644 --- a/fs/devpts/inode.c +++ b/fs/devpts/inode.c @@ -389,11 +389,10 @@ static int devpts_get_sb(struct file_system_type *fs_type, return 0; out_dput: - dput(s->s_root); + dput(s->s_root); /* undo dget() in simple_set_mnt() */ out_undo_sget: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); return error; } diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index ccabd5faa04d..9f0aa9883c28 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -614,9 +614,8 @@ static int ecryptfs_get_sb(struct file_system_type *fs_type, int flags, } goto out; out_abort: - dput(sb->s_root); - up_write(&sb->s_umount); - deactivate_super(sb); + dput(sb->s_root); /* aka mnt->mnt_root, as set by get_sb_nodev() */ + deactivate_locked_super(sb); out: return rc; } diff --git a/fs/libfs.c b/fs/libfs.c index cd223190c4e9..80046ddf5063 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -246,8 +246,7 @@ int get_sb_pseudo(struct file_system_type *fs_type, char *name, return 0; Enomem: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); return -ENOMEM; } diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 1679a164c8c9..d2d67781c579 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -2111,8 +2111,7 @@ out_err_nosb: error_splat_root: dput(mntroot); error_splat_super: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); goto out; } @@ -2208,8 +2207,7 @@ out_err_noserver: return error; error_splat_super: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); dprintk("<-- nfs_xdev_get_sb() = %d [splat]\n", error); return error; } @@ -2469,8 +2467,7 @@ out_free: error_splat_root: dput(mntroot); error_splat_super: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); goto out; } @@ -2564,8 +2561,7 @@ out_err_noserver: return error; error_splat_super: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); dprintk("<-- nfs4_xdev_get_sb() = %d [splat]\n", error); return error; } @@ -2649,8 +2645,7 @@ out_err_noserver: return error; error_splat_super: - up_write(&s->s_umount); - deactivate_super(s); + deactivate_locked_super(s); dprintk("<-- nfs4_referral_get_sb() = %d [splat]\n", error); return error; } diff --git a/fs/proc/root.c b/fs/proc/root.c index 1e15a2b176e8..b080b791d9e3 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -67,8 +67,7 @@ static int proc_get_sb(struct file_system_type *fs_type, sb->s_flags = flags; err = proc_fill_super(sb); if (err) { - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); return err; } diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index faa44f90608a..e9f7a754c4f7 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2055,8 +2055,7 @@ static int ubifs_get_sb(struct file_system_type *fs_type, int flags, return 0; out_deact: - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); out_close: ubi_close_volume(ubi); return err; diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 382109b5baeb..a7267bfd3765 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1133,8 +1133,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, free_cg_links: free_cg_links(&tmp_cg_links); drop_new_super: - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); return ret; } From c96f58573778ddf96ff67108a635f3f642ea63d3 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 6 May 2009 01:35:04 -0400 Subject: [PATCH 09/22] Fix a leak in failure exit in 9p ->get_sb() Signed-off-by: Al Viro --- fs/9p/vfs_super.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 0d29a57c63e6..ab5547ff29a1 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -156,6 +156,7 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags, root = d_alloc_root(inode); if (!root) { + iput(inode); retval = -ENOMEM; goto release_sb; } From 265e771e8197cdb22a1e2556663173fb62c9cd91 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 6 May 2009 01:36:04 -0400 Subject: [PATCH 10/22] Fix deadlock in ipathfs ->get_sb() forgot to unlock superblock before calling deactivate_super()... Signed-off-by: Al Viro --- drivers/infiniband/hw/ipath/ipath_fs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/ipath/ipath_fs.c b/drivers/infiniband/hw/ipath/ipath_fs.c index 8dc2bb781605..b3684060465e 100644 --- a/drivers/infiniband/hw/ipath/ipath_fs.c +++ b/drivers/infiniband/hw/ipath/ipath_fs.c @@ -347,7 +347,7 @@ static int ipathfs_fill_super(struct super_block *sb, void *data, spin_unlock_irqrestore(&ipath_devs_lock, flags); ret = create_device_files(sb, dd); if (ret) { - deactivate_super(sb); + deactivate_locked_super(sb); goto bail; } spin_lock_irqsave(&ipath_devs_lock, flags); From a731d12d6ddd1e703770cacb5dfecb155b03ee06 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Mon, 6 Apr 2009 16:43:42 -0700 Subject: [PATCH 11/22] ocfs2: Use nd_set_link(). ocfs2 was hand-calling vfs_follow_link(), but there's no point to that. Let's use page_follow_link_light() and nd_set_link(). Signed-off-by: Joel Becker Signed-off-by: Al Viro --- fs/ocfs2/symlink.c | 77 ++++++++++++++++++++++------------------------ 1 file changed, 37 insertions(+), 40 deletions(-) diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c index ed0a0cfd68d2..579dd1b1110f 100644 --- a/fs/ocfs2/symlink.c +++ b/fs/ocfs2/symlink.c @@ -39,6 +39,7 @@ #include #include #include +#include #define MLOG_MASK_PREFIX ML_NAMEI #include @@ -54,26 +55,6 @@ #include "buffer_head_io.h" -static char *ocfs2_page_getlink(struct dentry * dentry, - struct page **ppage); -static char *ocfs2_fast_symlink_getlink(struct inode *inode, - struct buffer_head **bh); - -/* get the link contents into pagecache */ -static char *ocfs2_page_getlink(struct dentry * dentry, - struct page **ppage) -{ - struct page * page; - struct address_space *mapping = dentry->d_inode->i_mapping; - page = read_mapping_page(mapping, 0, NULL); - if (IS_ERR(page)) - goto sync_fail; - *ppage = page; - return kmap(page); - -sync_fail: - return (char*)page; -} static char *ocfs2_fast_symlink_getlink(struct inode *inode, struct buffer_head **bh) @@ -128,40 +109,55 @@ out: return ret; } -static void *ocfs2_follow_link(struct dentry *dentry, - struct nameidata *nd) +static void *ocfs2_fast_follow_link(struct dentry *dentry, + struct nameidata *nd) { - int status; - char *link; + int status = 0; + int len; + char *target, *link = ERR_PTR(-ENOMEM); struct inode *inode = dentry->d_inode; - struct page *page = NULL; struct buffer_head *bh = NULL; - - if (ocfs2_inode_is_fast_symlink(inode)) - link = ocfs2_fast_symlink_getlink(inode, &bh); - else - link = ocfs2_page_getlink(dentry, &page); - if (IS_ERR(link)) { - status = PTR_ERR(link); + + mlog_entry_void(); + + BUG_ON(!ocfs2_inode_is_fast_symlink(inode)); + target = ocfs2_fast_symlink_getlink(inode, &bh); + if (IS_ERR(target)) { + status = PTR_ERR(target); mlog_errno(status); goto bail; } - status = vfs_follow_link(nd, link); + /* Fast symlinks can't be large */ + len = strlen(target); + link = kzalloc(len + 1, GFP_NOFS); + if (!link) { + status = -ENOMEM; + mlog_errno(status); + goto bail; + } + + memcpy(link, target, len); + nd_set_link(nd, link); bail: - if (page) { - kunmap(page); - page_cache_release(page); - } brelse(bh); - return ERR_PTR(status); + mlog_exit(status); + return status ? ERR_PTR(status) : link; +} + +static void ocfs2_fast_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie) +{ + char *link = cookie; + + kfree(link); } const struct inode_operations ocfs2_symlink_inode_operations = { .readlink = page_readlink, - .follow_link = ocfs2_follow_link, + .follow_link = page_follow_link_light, + .put_link = page_put_link, .getattr = ocfs2_getattr, .setattr = ocfs2_setattr, .setxattr = generic_setxattr, @@ -171,7 +167,8 @@ const struct inode_operations ocfs2_symlink_inode_operations = { }; const struct inode_operations ocfs2_fast_symlink_inode_operations = { .readlink = ocfs2_readlink, - .follow_link = ocfs2_follow_link, + .follow_link = ocfs2_fast_follow_link, + .put_link = ocfs2_fast_put_link, .getattr = ocfs2_getattr, .setattr = ocfs2_setattr, .setxattr = generic_setxattr, From c490d79bb70c549e096099576b1df40a8810b0d8 Mon Sep 17 00:00:00 2001 From: "npiggin@suse.de" Date: Sun, 26 Apr 2009 20:25:53 +1000 Subject: [PATCH 12/22] fs: dcache fix LRU ordering Fix ordering of LRU when moving referenced dentries to the head of the list (they should go to the head of the list in the same order as they were found from the tail, rather than reverse order). Signed-off-by: Nick Piggin Signed-off-by: Al Viro --- fs/dcache.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/dcache.c b/fs/dcache.c index 1fcffebfb44f..75659a6fd1f8 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -481,7 +481,7 @@ restart: if ((flags & DCACHE_REFERENCED) && (dentry->d_flags & DCACHE_REFERENCED)) { dentry->d_flags &= ~DCACHE_REFERENCED; - list_move_tail(&dentry->d_lru, &referenced); + list_move(&dentry->d_lru, &referenced); spin_unlock(&dentry->d_lock); } else { list_move_tail(&dentry->d_lru, &tmp); From 774e33e70b2bffa8c602d22a5d27c0061a0039cc Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Sun, 26 Apr 2009 14:51:17 +0200 Subject: [PATCH 13/22] ROMFS: romfs_dev_read() error ignored romfs_dev_read() may return -EIO, but ret is unsigned, so the errorpath isn't taken. Signed-off-by: Roel Kluin Signed-off-by: Al Viro --- fs/romfs/super.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/romfs/super.c b/fs/romfs/super.c index c53b5ef8a02f..4ab3c03d8f95 100644 --- a/fs/romfs/super.c +++ b/fs/romfs/super.c @@ -298,7 +298,8 @@ static struct inode *romfs_iget(struct super_block *sb, unsigned long pos) struct romfs_inode ri; struct inode *i; unsigned long nlen; - unsigned nextfh, ret; + unsigned nextfh; + int ret; umode_t mode; /* we might have to traverse a chain of "hard link" file entries to get From db6c1fbb92eeb4cb52c6133e0c533602f49fc4bd Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 7 Apr 2009 18:07:08 +0200 Subject: [PATCH 14/22] romfs: cleanup romfs_fs.h There's no kernel-only content in it anymore, so move it to header-y and remove the superflous #ifdef __KERNEL__. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- include/linux/Kbuild | 2 +- include/linux/romfs_fs.h | 5 ----- 2 files changed, 1 insertion(+), 6 deletions(-) diff --git a/include/linux/Kbuild b/include/linux/Kbuild index ca9b9b9bd331..3f0eaa397ef5 100644 --- a/include/linux/Kbuild +++ b/include/linux/Kbuild @@ -138,6 +138,7 @@ header-y += qnxtypes.h header-y += radeonfb.h header-y += raw.h header-y += resource.h +header-y += romfs_fs.h header-y += rose.h header-y += serial_reg.h header-y += smbno.h @@ -314,7 +315,6 @@ unifdef-y += irqnr.h unifdef-y += reboot.h unifdef-y += reiserfs_fs.h unifdef-y += reiserfs_xattr.h -unifdef-y += romfs_fs.h unifdef-y += route.h unifdef-y += rtc.h unifdef-y += rtnetlink.h diff --git a/include/linux/romfs_fs.h b/include/linux/romfs_fs.h index e20bbf9eb365..c490fbc43fe2 100644 --- a/include/linux/romfs_fs.h +++ b/include/linux/romfs_fs.h @@ -53,9 +53,4 @@ struct romfs_inode { #define ROMFH_PAD (ROMFH_SIZE-1) #define ROMFH_MASK (~ROMFH_PAD) -#ifdef __KERNEL__ - -/* Not much now */ - -#endif /* __KERNEL__ */ #endif From ddbaaf3024d764ced700efb2d818709b90ea6fdd Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Wed, 29 Apr 2009 20:14:57 -0400 Subject: [PATCH 15/22] NULL noise in fs/super.c:kill_bdev_super() Signed-off-by: H Hartley Sweeten Cc: Subrata Modak 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 a9dc4c33ef4d..1943fdf655fa 100644 --- a/fs/super.c +++ b/fs/super.c @@ -922,7 +922,7 @@ void kill_block_super(struct super_block *sb) struct block_device *bdev = sb->s_bdev; fmode_t mode = sb->s_mode; - bdev->bd_super = 0; + bdev->bd_super = NULL; generic_shutdown_super(sb); sync_blockdev(bdev); close_bdev_exclusive(bdev, mode); From 6b3304b531704711286c3359b06922b83fdba015 Mon Sep 17 00:00:00 2001 From: Manish Katiyar Date: Tue, 31 Mar 2009 19:35:54 +0530 Subject: [PATCH 16/22] Make checkpatch.pl shut up on fs/inode.c Code Quality According To Mingo(tm) has been vastly improved, no code has been damaged^Wchanged^Wdamaged. [commit message rewritten -- AV] Signed-off-by: Manish Katiyar Signed-off-by: Al Viro --- fs/inode.c | 81 +++++++++++++++++++++++------------------------------- 1 file changed, 35 insertions(+), 46 deletions(-) diff --git a/fs/inode.c b/fs/inode.c index 6ad14a1cd8c9..0571983755dc 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -99,7 +99,7 @@ static DEFINE_MUTEX(iprune_mutex); */ struct inodes_stat_t inodes_stat; -static struct kmem_cache * inode_cachep __read_mostly; +static struct kmem_cache *inode_cachep __read_mostly; static void wake_up_inode(struct inode *inode) { @@ -124,7 +124,7 @@ struct inode *inode_init_always(struct super_block *sb, struct inode *inode) static struct inode_operations empty_iops; static const struct file_operations empty_fops; - struct address_space * const mapping = &inode->i_data; + struct address_space *const mapping = &inode->i_data; inode->i_sb = sb; inode->i_blkbits = sb->s_blocksize_bits; @@ -216,7 +216,7 @@ static struct inode *alloc_inode(struct super_block *sb) return NULL; } -void destroy_inode(struct inode *inode) +void destroy_inode(struct inode *inode) { BUG_ON(inode_has_buffers(inode)); security_inode_free(inode); @@ -252,12 +252,11 @@ void inode_init_once(struct inode *inode) mutex_init(&inode->inotify_mutex); #endif } - EXPORT_SYMBOL(inode_init_once); static void init_once(void *foo) { - struct inode * inode = (struct inode *) foo; + struct inode *inode = (struct inode *) foo; inode_init_once(inode); } @@ -265,7 +264,7 @@ static void init_once(void *foo) /* * inode_lock must be held */ -void __iget(struct inode * inode) +void __iget(struct inode *inode) { if (atomic_read(&inode->i_count)) { atomic_inc(&inode->i_count); @@ -289,7 +288,7 @@ void clear_inode(struct inode *inode) { might_sleep(); invalidate_inode_buffers(inode); - + BUG_ON(inode->i_data.nrpages); BUG_ON(!(inode->i_state & I_FREEING)); BUG_ON(inode->i_state & I_CLEAR); @@ -303,7 +302,6 @@ void clear_inode(struct inode *inode) cd_forget(inode); inode->i_state = I_CLEAR; } - EXPORT_SYMBOL(clear_inode); /* @@ -351,8 +349,8 @@ static int invalidate_list(struct list_head *head, struct list_head *dispose) next = head->next; for (;;) { - struct list_head * tmp = next; - struct inode * inode; + struct list_head *tmp = next; + struct inode *inode; /* * We can reschedule here without worrying about the list's @@ -391,7 +389,7 @@ static int invalidate_list(struct list_head *head, struct list_head *dispose) * fails because there are busy inodes then a non zero value is returned. * If the discard is successful all the inodes have been discarded. */ -int invalidate_inodes(struct super_block * sb) +int invalidate_inodes(struct super_block *sb) { int busy; LIST_HEAD(throw_away); @@ -407,7 +405,6 @@ int invalidate_inodes(struct super_block * sb) return busy; } - EXPORT_SYMBOL(invalidate_inodes); static int can_unuse(struct inode *inode) @@ -504,7 +501,7 @@ static int shrink_icache_memory(int nr, gfp_t gfp_mask) * Nasty deadlock avoidance. We may hold various FS locks, * and we don't want to recurse into the FS that called us * in clear_inode() and friends.. - */ + */ if (!(gfp_mask & __GFP_FS)) return -1; prune_icache(nr); @@ -524,10 +521,13 @@ static void __wait_on_freeing_inode(struct inode *inode); * by hand after calling find_inode now! This simplifies iunique and won't * add any additional branch in the common code. */ -static struct inode * find_inode(struct super_block * sb, struct hlist_head *head, int (*test)(struct inode *, void *), void *data) +static struct inode *find_inode(struct super_block *sb, + struct hlist_head *head, + int (*test)(struct inode *, void *), + void *data) { struct hlist_node *node; - struct inode * inode = NULL; + struct inode *inode = NULL; repeat: hlist_for_each_entry(inode, node, head, i_hash) { @@ -548,10 +548,11 @@ repeat: * find_inode_fast is the fast path version of find_inode, see the comment at * iget_locked for details. */ -static struct inode * find_inode_fast(struct super_block * sb, struct hlist_head *head, unsigned long ino) +static struct inode *find_inode_fast(struct super_block *sb, + struct hlist_head *head, unsigned long ino) { struct hlist_node *node; - struct inode * inode = NULL; + struct inode *inode = NULL; repeat: hlist_for_each_entry(inode, node, head, i_hash) { @@ -631,10 +632,10 @@ struct inode *new_inode(struct super_block *sb) * here to attempt to avoid that. */ static unsigned int last_ino; - struct inode * inode; + struct inode *inode; spin_lock_prefetch(&inode_lock); - + inode = alloc_inode(sb); if (inode) { spin_lock(&inode_lock); @@ -645,7 +646,6 @@ struct inode *new_inode(struct super_block *sb) } return inode; } - EXPORT_SYMBOL(new_inode); void unlock_new_inode(struct inode *inode) @@ -674,7 +674,6 @@ void unlock_new_inode(struct inode *inode) inode->i_state &= ~(I_LOCK|I_NEW); wake_up_inode(inode); } - EXPORT_SYMBOL(unlock_new_inode); /* @@ -683,13 +682,17 @@ EXPORT_SYMBOL(unlock_new_inode); * We no longer cache the sb_flags in i_flags - see fs.h * -- rmk@arm.uk.linux.org */ -static struct inode * get_new_inode(struct super_block *sb, struct hlist_head *head, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *data) +static struct inode *get_new_inode(struct super_block *sb, + struct hlist_head *head, + int (*test)(struct inode *, void *), + int (*set)(struct inode *, void *), + void *data) { - struct inode * inode; + struct inode *inode; inode = alloc_inode(sb); if (inode) { - struct inode * old; + struct inode *old; spin_lock(&inode_lock); /* We released the lock, so.. */ @@ -731,13 +734,14 @@ set_failed: * get_new_inode_fast is the fast path version of get_new_inode, see the * comment at iget_locked for details. */ -static struct inode * get_new_inode_fast(struct super_block *sb, struct hlist_head *head, unsigned long ino) +static struct inode *get_new_inode_fast(struct super_block *sb, + struct hlist_head *head, unsigned long ino) { - struct inode * inode; + struct inode *inode; inode = alloc_inode(sb); if (inode) { - struct inode * old; + struct inode *old; spin_lock(&inode_lock); /* We released the lock, so.. */ @@ -823,7 +827,6 @@ struct inode *igrab(struct inode *inode) spin_unlock(&inode_lock); return inode; } - EXPORT_SYMBOL(igrab); /** @@ -924,7 +927,6 @@ struct inode *ilookup5_nowait(struct super_block *sb, unsigned long hashval, return ifind(sb, head, test, data, 0); } - EXPORT_SYMBOL(ilookup5_nowait); /** @@ -953,7 +955,6 @@ struct inode *ilookup5(struct super_block *sb, unsigned long hashval, return ifind(sb, head, test, data, 1); } - EXPORT_SYMBOL(ilookup5); /** @@ -976,7 +977,6 @@ struct inode *ilookup(struct super_block *sb, unsigned long ino) return ifind_fast(sb, head, ino); } - EXPORT_SYMBOL(ilookup); /** @@ -1015,7 +1015,6 @@ struct inode *iget5_locked(struct super_block *sb, unsigned long hashval, */ return get_new_inode(sb, head, test, set, data); } - EXPORT_SYMBOL(iget5_locked); /** @@ -1047,7 +1046,6 @@ struct inode *iget_locked(struct super_block *sb, unsigned long ino) */ return get_new_inode_fast(sb, head, ino); } - EXPORT_SYMBOL(iget_locked); int insert_inode_locked(struct inode *inode) @@ -1076,7 +1074,6 @@ int insert_inode_locked(struct inode *inode) iput(old); } } - EXPORT_SYMBOL(insert_inode_locked); int insert_inode_locked4(struct inode *inode, unsigned long hashval, @@ -1106,7 +1103,6 @@ int insert_inode_locked4(struct inode *inode, unsigned long hashval, iput(old); } } - EXPORT_SYMBOL(insert_inode_locked4); /** @@ -1124,7 +1120,6 @@ void __insert_inode_hash(struct inode *inode, unsigned long hashval) hlist_add_head(&inode->i_hash, head); spin_unlock(&inode_lock); } - EXPORT_SYMBOL(__insert_inode_hash); /** @@ -1139,7 +1134,6 @@ void remove_inode_hash(struct inode *inode) hlist_del_init(&inode->i_hash); spin_unlock(&inode_lock); } - EXPORT_SYMBOL(remove_inode_hash); /* @@ -1187,7 +1181,6 @@ void generic_delete_inode(struct inode *inode) BUG_ON(inode->i_state != I_CLEAR); destroy_inode(inode); } - EXPORT_SYMBOL(generic_delete_inode); static void generic_forget_inode(struct inode *inode) @@ -1237,12 +1230,11 @@ void generic_drop_inode(struct inode *inode) else generic_forget_inode(inode); } - EXPORT_SYMBOL_GPL(generic_drop_inode); /* * Called when we're dropping the last reference - * to an inode. + * to an inode. * * Call the FS "drop()" function, defaulting to * the legacy UNIX filesystem behaviour.. @@ -1262,7 +1254,7 @@ static inline void iput_final(struct inode *inode) } /** - * iput - put an inode + * iput - put an inode * @inode: inode to put * * Puts an inode, dropping its usage count. If the inode use count hits @@ -1279,7 +1271,6 @@ void iput(struct inode *inode) iput_final(inode); } } - EXPORT_SYMBOL(iput); /** @@ -1290,10 +1281,10 @@ EXPORT_SYMBOL(iput); * Returns the block number on the device holding the inode that * is the disk block number for the block of the file requested. * That is, asked for block 4 of inode 1 the function will return the - * disk block relative to the disk start that holds that block of the + * disk block relative to the disk start that holds that block of the * file. */ -sector_t bmap(struct inode * inode, sector_t block) +sector_t bmap(struct inode *inode, sector_t block) { sector_t res = 0; if (inode->i_mapping->a_ops->bmap) @@ -1425,7 +1416,6 @@ void file_update_time(struct file *file) mark_inode_dirty_sync(inode); mnt_drop_write(file->f_path.mnt); } - EXPORT_SYMBOL(file_update_time); int inode_needs_sync(struct inode *inode) @@ -1436,7 +1426,6 @@ int inode_needs_sync(struct inode *inode) return 1; return 0; } - EXPORT_SYMBOL(inode_needs_sync); int inode_wait(void *word) From e24977d45f45d1675e050dc1a0aaf4bfc4ca9866 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 2 Apr 2009 21:17:03 -0400 Subject: [PATCH 17/22] Reduce path_lookup() abuses ... use kern_path() where possible [folded a fix from rdd] Signed-off-by: Al Viro --- drivers/scsi/osd/osd_uld.c | 20 ++++++++++---------- fs/gfs2/ops_fstype.c | 8 ++++---- security/tomoyo/common.c | 6 +++--- security/tomoyo/realpath.c | 16 ++++++++-------- 4 files changed, 25 insertions(+), 25 deletions(-) diff --git a/drivers/scsi/osd/osd_uld.c b/drivers/scsi/osd/osd_uld.c index f644c9571eab..22b59e13ba83 100644 --- a/drivers/scsi/osd/osd_uld.c +++ b/drivers/scsi/osd/osd_uld.c @@ -173,26 +173,26 @@ static const struct file_operations osd_fops = { .unlocked_ioctl = osd_uld_ioctl, }; -struct osd_dev *osduld_path_lookup(const char *path) +struct osd_dev *osduld_path_lookup(const char *name) { - struct nameidata nd; + struct path path; struct inode *inode; struct cdev *cdev; struct osd_uld_device *uninitialized_var(oud); int error; - if (!path || !*path) { + if (!name || !*name) { OSD_ERR("Mount with !path || !*path\n"); return ERR_PTR(-EINVAL); } - error = path_lookup(path, LOOKUP_FOLLOW, &nd); + error = kern_path(name, LOOKUP_FOLLOW, &path); if (error) { - OSD_ERR("path_lookup of %s faild=>%d\n", path, error); + OSD_ERR("path_lookup of %s failed=>%d\n", name, error); return ERR_PTR(error); } - inode = nd.path.dentry->d_inode; + inode = path.dentry->d_inode; error = -EINVAL; /* Not the right device e.g osd_uld_device */ if (!S_ISCHR(inode->i_mode)) { OSD_DEBUG("!S_ISCHR()\n"); @@ -202,15 +202,15 @@ struct osd_dev *osduld_path_lookup(const char *path) cdev = inode->i_cdev; if (!cdev) { OSD_ERR("Before mounting an OSD Based filesystem\n"); - OSD_ERR(" user-mode must open+close the %s device\n", path); - OSD_ERR(" Example: bash: echo < %s\n", path); + OSD_ERR(" user-mode must open+close the %s device\n", name); + OSD_ERR(" Example: bash: echo < %s\n", name); goto out; } /* The Magic wand. Is it our char-dev */ /* TODO: Support sg devices */ if (cdev->owner != THIS_MODULE) { - OSD_ERR("Error mounting %s - is not an OSD device\n", path); + OSD_ERR("Error mounting %s - is not an OSD device\n", name); goto out; } @@ -220,7 +220,7 @@ struct osd_dev *osduld_path_lookup(const char *path) error = 0; out: - path_put(&nd.path); + path_put(&path); return error ? ERR_PTR(error) : &oud->od; } EXPORT_SYMBOL(osduld_path_lookup); diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 650a730707b7..1ff9473ea753 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -1282,21 +1282,21 @@ static int gfs2_get_sb(struct file_system_type *fs_type, int flags, static struct super_block *get_gfs2_sb(const char *dev_name) { struct super_block *sb; - struct nameidata nd; + struct path path; int error; - error = path_lookup(dev_name, LOOKUP_FOLLOW, &nd); + error = kern_path(dev_name, LOOKUP_FOLLOW, &path); if (error) { printk(KERN_WARNING "GFS2: path_lookup on %s returned error %d\n", dev_name, error); return NULL; } - sb = nd.path.dentry->d_inode->i_sb; + sb = path.dentry->d_inode->i_sb; if (sb && (sb->s_type == &gfs2_fs_type)) atomic_inc(&sb->s_active); else sb = NULL; - path_put(&nd.path); + path_put(&path); return sb; } diff --git a/security/tomoyo/common.c b/security/tomoyo/common.c index d4d41b3efc7c..ddfb9cccf468 100644 --- a/security/tomoyo/common.c +++ b/security/tomoyo/common.c @@ -1720,14 +1720,14 @@ static bool tomoyo_policy_loader_exists(void) * policies are not loaded yet. * Thus, let do_execve() call this function everytime. */ - struct nameidata nd; + struct path path; - if (path_lookup(tomoyo_loader, LOOKUP_FOLLOW, &nd)) { + if (kern_path(tomoyo_loader, LOOKUP_FOLLOW, &path)) { printk(KERN_INFO "Not activating Mandatory Access Control now " "since %s doesn't exist.\n", tomoyo_loader); return false; } - path_put(&nd.path); + path_put(&path); return true; } diff --git a/security/tomoyo/realpath.c b/security/tomoyo/realpath.c index bf8e2b451687..40927a84cb6e 100644 --- a/security/tomoyo/realpath.c +++ b/security/tomoyo/realpath.c @@ -165,11 +165,11 @@ char *tomoyo_realpath_from_path(struct path *path) */ char *tomoyo_realpath(const char *pathname) { - struct nameidata nd; + struct path path; - if (pathname && path_lookup(pathname, LOOKUP_FOLLOW, &nd) == 0) { - char *buf = tomoyo_realpath_from_path(&nd.path); - path_put(&nd.path); + if (pathname && kern_path(pathname, LOOKUP_FOLLOW, &path) == 0) { + char *buf = tomoyo_realpath_from_path(&path); + path_put(&path); return buf; } return NULL; @@ -184,11 +184,11 @@ char *tomoyo_realpath(const char *pathname) */ char *tomoyo_realpath_nofollow(const char *pathname) { - struct nameidata nd; + struct path path; - if (pathname && path_lookup(pathname, 0, &nd) == 0) { - char *buf = tomoyo_realpath_from_path(&nd.path); - path_put(&nd.path); + if (pathname && kern_path(pathname, 0, &path) == 0) { + char *buf = tomoyo_realpath_from_path(&path); + path_put(&path); return buf; } return NULL; From a44ddbb6d8a8ffe4e34e417048dfdd8f3dd1de4f Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 6 Apr 2009 09:38:49 -0400 Subject: [PATCH 18/22] Make open_exec() and sys_uselib() use may_open(), instead of duplicating its parts Signed-off-by: Al Viro --- fs/exec.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) diff --git a/fs/exec.c b/fs/exec.c index 639177b0eeac..41ae8e0de72d 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -126,11 +126,7 @@ SYSCALL_DEFINE1(uselib, const char __user *, library) if (nd.path.mnt->mnt_flags & MNT_NOEXEC) goto exit; - error = inode_permission(nd.path.dentry->d_inode, - MAY_READ | MAY_EXEC | MAY_OPEN); - if (error) - goto exit; - error = ima_path_check(&nd.path, MAY_READ | MAY_EXEC | MAY_OPEN); + error = may_open(&nd.path, MAY_READ | MAY_EXEC | MAY_OPEN, 0); if (error) goto exit; @@ -677,10 +673,7 @@ struct file *open_exec(const char *name) if (nd.path.mnt->mnt_flags & MNT_NOEXEC) goto out_path_put; - err = inode_permission(nd.path.dentry->d_inode, MAY_EXEC | MAY_OPEN); - if (err) - goto out_path_put; - err = ima_path_check(&nd.path, MAY_EXEC | MAY_OPEN); + err = may_open(&nd.path, MAY_EXEC | MAY_OPEN, 0); if (err) goto out_path_put; From 6e8341a11eb21826b7192d0bb88cb5b44900a9af Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 6 Apr 2009 11:16:22 -0400 Subject: [PATCH 19/22] Switch open_exec() and sys_uselib() to do_open_filp() ... and make path_lookup_open() static Signed-off-by: Al Viro --- fs/exec.c | 76 +++++++++++++++---------------------------- fs/namei.c | 13 ++++---- fs/open.c | 2 +- include/linux/fs.h | 2 +- include/linux/namei.h | 1 - 5 files changed, 36 insertions(+), 58 deletions(-) diff --git a/fs/exec.c b/fs/exec.c index 41ae8e0de72d..895823d0149d 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -105,36 +105,28 @@ static inline void put_binfmt(struct linux_binfmt * fmt) SYSCALL_DEFINE1(uselib, const char __user *, library) { struct file *file; - struct nameidata nd; char *tmp = getname(library); int error = PTR_ERR(tmp); - if (!IS_ERR(tmp)) { - error = path_lookup_open(AT_FDCWD, tmp, - LOOKUP_FOLLOW, &nd, - FMODE_READ|FMODE_EXEC); - putname(tmp); - } - if (error) + if (IS_ERR(tmp)) goto out; - error = -EINVAL; - if (!S_ISREG(nd.path.dentry->d_inode->i_mode)) - goto exit; - - error = -EACCES; - if (nd.path.mnt->mnt_flags & MNT_NOEXEC) - goto exit; - - error = may_open(&nd.path, MAY_READ | MAY_EXEC | MAY_OPEN, 0); - if (error) - goto exit; - - file = nameidata_to_filp(&nd, O_RDONLY|O_LARGEFILE); + file = do_filp_open(AT_FDCWD, tmp, + O_LARGEFILE | O_RDONLY | FMODE_EXEC, 0, + MAY_READ | MAY_EXEC | MAY_OPEN); + putname(tmp); error = PTR_ERR(file); if (IS_ERR(file)) goto out; + error = -EINVAL; + if (!S_ISREG(file->f_path.dentry->d_inode->i_mode)) + goto exit; + + error = -EACCES; + if (file->f_path.mnt->mnt_flags & MNT_NOEXEC) + goto exit; + fsnotify_open(file->f_path.dentry); error = -ENOEXEC; @@ -156,13 +148,10 @@ SYSCALL_DEFINE1(uselib, const char __user *, library) } read_unlock(&binfmt_lock); } +exit: fput(file); out: return error; -exit: - release_open_intent(&nd); - path_put(&nd.path); - goto out; } #ifdef CONFIG_MMU @@ -657,44 +646,33 @@ EXPORT_SYMBOL(setup_arg_pages); struct file *open_exec(const char *name) { - struct nameidata nd; struct file *file; int err; - err = path_lookup_open(AT_FDCWD, name, LOOKUP_FOLLOW, &nd, - FMODE_READ|FMODE_EXEC); - if (err) + file = do_filp_open(AT_FDCWD, name, + O_LARGEFILE | O_RDONLY | FMODE_EXEC, 0, + MAY_EXEC | MAY_OPEN); + if (IS_ERR(file)) goto out; err = -EACCES; - if (!S_ISREG(nd.path.dentry->d_inode->i_mode)) - goto out_path_put; + if (!S_ISREG(file->f_path.dentry->d_inode->i_mode)) + goto exit; - if (nd.path.mnt->mnt_flags & MNT_NOEXEC) - goto out_path_put; - - err = may_open(&nd.path, MAY_EXEC | MAY_OPEN, 0); - if (err) - goto out_path_put; - - file = nameidata_to_filp(&nd, O_RDONLY|O_LARGEFILE); - if (IS_ERR(file)) - return file; + if (file->f_path.mnt->mnt_flags & MNT_NOEXEC) + goto exit; fsnotify_open(file->f_path.dentry); err = deny_write_access(file); - if (err) { - fput(file); - goto out; - } + if (err) + goto exit; +out: return file; - out_path_put: - release_open_intent(&nd); - path_put(&nd.path); - out: +exit: + fput(file); return ERR_PTR(err); } EXPORT_SYMBOL(open_exec); diff --git a/fs/namei.c b/fs/namei.c index 78f253cd2d4f..967c3db92724 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1130,8 +1130,8 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt, * @nd: pointer to nameidata * @open_flags: open intent flags */ -int path_lookup_open(int dfd, const char *name, unsigned int lookup_flags, - struct nameidata *nd, int open_flags) +static int path_lookup_open(int dfd, const char *name, + unsigned int lookup_flags, struct nameidata *nd, int open_flags) { struct file *filp = get_empty_filp(); int err; @@ -1637,18 +1637,19 @@ static int open_will_write_to_fs(int flag, struct inode *inode) * open_to_namei_flags() for more details. */ struct file *do_filp_open(int dfd, const char *pathname, - int open_flag, int mode) + int open_flag, int mode, int acc_mode) { struct file *filp; struct nameidata nd; - int acc_mode, error; + int error; struct path path; struct dentry *dir; int count = 0; int will_write; int flag = open_to_namei_flags(open_flag); - acc_mode = MAY_OPEN | ACC_MODE(flag); + if (!acc_mode) + acc_mode = MAY_OPEN | ACC_MODE(flag); /* O_TRUNC implies we need access checks for write permissions */ if (flag & O_TRUNC) @@ -1869,7 +1870,7 @@ do_link: */ struct file *filp_open(const char *filename, int flags, int mode) { - return do_filp_open(AT_FDCWD, filename, flags, mode); + return do_filp_open(AT_FDCWD, filename, flags, mode, 0); } EXPORT_SYMBOL(filp_open); diff --git a/fs/open.c b/fs/open.c index 377eb25b6abf..bdfbf03615a4 100644 --- a/fs/open.c +++ b/fs/open.c @@ -1033,7 +1033,7 @@ long do_sys_open(int dfd, const char __user *filename, int flags, int mode) if (!IS_ERR(tmp)) { fd = get_unused_fd_flags(flags); if (fd >= 0) { - struct file *f = do_filp_open(dfd, tmp, flags, mode); + struct file *f = do_filp_open(dfd, tmp, flags, mode, 0); if (IS_ERR(f)) { put_unused_fd(fd); fd = PTR_ERR(f); diff --git a/include/linux/fs.h b/include/linux/fs.h index 11484d08042c..ed788426f464 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2118,7 +2118,7 @@ extern struct file *create_write_pipe(int flags); extern void free_write_pipe(struct file *); extern struct file *do_filp_open(int dfd, const char *pathname, - int open_flag, int mode); + int open_flag, int mode, int acc_mode); extern int may_open(struct path *, int, int); extern int kernel_read(struct file *, unsigned long, char *, unsigned long); diff --git a/include/linux/namei.h b/include/linux/namei.h index fc2e03579877..518098fe63af 100644 --- a/include/linux/namei.h +++ b/include/linux/namei.h @@ -69,7 +69,6 @@ extern int path_lookup(const char *, unsigned, struct nameidata *); extern int vfs_path_lookup(struct dentry *, struct vfsmount *, const char *, unsigned int, struct nameidata *); -extern int path_lookup_open(int dfd, const char *name, unsigned lookup_flags, struct nameidata *, int open_flags); extern struct file *lookup_instantiate_filp(struct nameidata *nd, struct dentry *dentry, int (*open)(struct inode *, struct file *)); extern struct file *nameidata_to_filp(struct nameidata *nd, int flags); From f9dbd05bc97d1d4f17c2057612f6a8e4dbd039e0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 6 May 2009 19:56:21 -0400 Subject: [PATCH 20/22] switch ufs directories to ufs_sync_file() Signed-off-by: Al Viro --- fs/ufs/dir.c | 2 +- fs/ufs/file.c | 2 +- fs/ufs/ufs.h | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/ufs/dir.c b/fs/ufs/dir.c index dbbbc4668769..6321b797061b 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 = file_fsync, + .fsync = ufs_sync_file, .llseek = generic_file_llseek, }; diff --git a/fs/ufs/file.c b/fs/ufs/file.c index 625ef17c6f83..2bd3a1615714 100644 --- a/fs/ufs/file.c +++ b/fs/ufs/file.c @@ -30,7 +30,7 @@ #include "ufs.h" -static int ufs_sync_file(struct file *file, struct dentry *dentry, int datasync) +int ufs_sync_file(struct file *file, struct dentry *dentry, int datasync) { struct inode *inode = dentry->d_inode; int err; diff --git a/fs/ufs/ufs.h b/fs/ufs/ufs.h index 69b3427d7885..d0c4acd4f1f3 100644 --- a/fs/ufs/ufs.h +++ b/fs/ufs/ufs.h @@ -98,8 +98,8 @@ extern void ufs_set_link(struct inode *dir, struct ufs_dir_entry *de, /* file.c */ 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 2a32cebd6cbcc43996c3e2d114fa32ba1e71192a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 8 May 2009 16:05:57 -0400 Subject: [PATCH 21/22] Fix races around the access to ->s_options Put generic_show_options read access to s_options under rcu_read_lock, split save_mount_options() into "we are setting it the first time" (uses in foo_fill_super()) and "we are relacing and freeing the old one", synchronize_rcu() before kfree() in the latter. Signed-off-by: Al Viro --- drivers/isdn/capi/capifs.c | 3 +-- fs/affs/super.c | 3 +-- fs/afs/super.c | 4 ++-- fs/hpfs/super.c | 3 +-- fs/namespace.c | 21 ++++++++++++++++++--- fs/reiserfs/super.c | 3 +-- include/linux/fs.h | 1 + 7 files changed, 25 insertions(+), 13 deletions(-) diff --git a/drivers/isdn/capi/capifs.c b/drivers/isdn/capi/capifs.c index b129409925af..8f9f3b5a3e8c 100644 --- a/drivers/isdn/capi/capifs.c +++ b/drivers/isdn/capi/capifs.c @@ -75,8 +75,7 @@ static int capifs_remount(struct super_block *s, int *flags, char *data) } } - kfree(s->s_options); - s->s_options = new_opt; + replace_mount_options(s, new_opt); config.setuid = setuid; config.setgid = setgid; diff --git a/fs/affs/super.c b/fs/affs/super.c index 5ce695e707fe..63f5183f263b 100644 --- a/fs/affs/super.c +++ b/fs/affs/super.c @@ -507,8 +507,7 @@ affs_remount(struct super_block *sb, int *flags, char *data) kfree(new_opts); return -EINVAL; } - kfree(sb->s_options); - sb->s_options = new_opts; + replace_mount_options(sb, new_opts); sbi->s_flags = mount_flags; sbi->s_mode = mode; diff --git a/fs/afs/super.c b/fs/afs/super.c index 2753f16dd315..76828e5f8a39 100644 --- a/fs/afs/super.c +++ b/fs/afs/super.c @@ -408,17 +408,17 @@ static int afs_get_sb(struct file_system_type *fs_type, deactivate_locked_super(sb); goto error; } - sb->s_options = new_opts; + save_mount_options(sb, new_opts); sb->s_flags |= MS_ACTIVE; } else { _debug("reuse"); - kfree(new_opts); ASSERTCMP(sb->s_flags, &, MS_ACTIVE); } simple_set_mnt(mnt, sb); afs_put_volume(params.volume); afs_put_cell(params.cell); + kfree(new_opts); _leave(" = 0 [%p]", sb); return 0; diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c index fecf402d7b8a..fc77965be841 100644 --- a/fs/hpfs/super.c +++ b/fs/hpfs/super.c @@ -423,8 +423,7 @@ static int hpfs_remount_fs(struct super_block *s, int *flags, char *data) if (!(*flags & MS_RDONLY)) mark_dirty(s); - kfree(s->s_options); - s->s_options = new_opts; + replace_mount_options(s, new_opts); return 0; diff --git a/fs/namespace.c b/fs/namespace.c index 0d2003fb4377..134d494158d9 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -695,12 +695,16 @@ static inline void mangle(struct seq_file *m, const char *s) */ int generic_show_options(struct seq_file *m, struct vfsmount *mnt) { - const char *options = mnt->mnt_sb->s_options; + const char *options; + + rcu_read_lock(); + options = rcu_dereference(mnt->mnt_sb->s_options); if (options != NULL && options[0]) { seq_putc(m, ','); mangle(m, options); } + rcu_read_unlock(); return 0; } @@ -721,11 +725,22 @@ EXPORT_SYMBOL(generic_show_options); */ void save_mount_options(struct super_block *sb, char *options) { - kfree(sb->s_options); - sb->s_options = kstrdup(options, GFP_KERNEL); + BUG_ON(sb->s_options); + rcu_assign_pointer(sb->s_options, kstrdup(options, GFP_KERNEL)); } EXPORT_SYMBOL(save_mount_options); +void replace_mount_options(struct super_block *sb, char *options) +{ + char *old = sb->s_options; + rcu_assign_pointer(sb->s_options, options); + if (old) { + synchronize_rcu(); + kfree(old); + } +} +EXPORT_SYMBOL(replace_mount_options); + #ifdef CONFIG_PROC_FS /* iterator */ static void *m_start(struct seq_file *m, loff_t *pos) diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index d444fe0013a4..1215a4f50cd2 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -1316,8 +1316,7 @@ static int reiserfs_remount(struct super_block *s, int *mount_flags, char *arg) } out_ok: - kfree(s->s_options); - s->s_options = new_opts; + replace_mount_options(s, new_opts); return 0; out_err: diff --git a/include/linux/fs.h b/include/linux/fs.h index ed788426f464..3b534e527e09 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2368,6 +2368,7 @@ extern void file_update_time(struct file *file); extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt); extern void save_mount_options(struct super_block *sb, char *options); +extern void replace_mount_options(struct super_block *sb, char *options); static inline ino_t parent_ino(struct dentry *dentry) { From b0c4f3222c7406f0aee9d52b7d604a1d36108463 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 8 May 2009 16:23:30 -0400 Subject: [PATCH 22/22] Fix the race between capifs remount and node creation we don't want to deal with half-updated config Signed-off-by: Al Viro --- drivers/isdn/capi/capifs.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/isdn/capi/capifs.c b/drivers/isdn/capi/capifs.c index 8f9f3b5a3e8c..bff72d81f263 100644 --- a/drivers/isdn/capi/capifs.c +++ b/drivers/isdn/capi/capifs.c @@ -75,14 +75,17 @@ static int capifs_remount(struct super_block *s, int *flags, char *data) } } - replace_mount_options(s, new_opt); + mutex_lock(&s->s_root->d_inode->i_mutex); + replace_mount_options(s, new_opt); config.setuid = setuid; config.setgid = setgid; config.uid = uid; config.gid = gid; config.mode = mode; + mutex_unlock(&s->s_root->d_inode->i_mutex); + return 0; } @@ -153,13 +156,16 @@ void capifs_new_ncci(unsigned int number, dev_t device) if (!inode) return; inode->i_ino = number+2; + + dentry = get_node(number); + + /* config contents is protected by root's i_mutex */ inode->i_uid = config.setuid ? config.uid : current_fsuid(); inode->i_gid = config.setgid ? config.gid : current_fsgid(); inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; init_special_inode(inode, S_IFCHR|config.mode, device); //inode->i_op = &capifs_file_inode_operations; - dentry = get_node(number); if (!IS_ERR(dentry) && !dentry->d_inode) d_instantiate(dentry, inode); mutex_unlock(&capifs_root->d_inode->i_mutex);