From 795858dbd253462a67e14272edeaae73c6074b17 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Mon, 15 Aug 2011 13:02:37 -0700 Subject: ceph: fix encoding of ino only (not relative) paths A 'path' consists of a starting ino and relative component. Encode even when there is no relative component. This is primarily needed by the NFS reexport code. Signed-off-by: Sage Weil --- fs/ceph/mds_client.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index fee028b5332e..86c59e16ba74 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1595,7 +1595,7 @@ static int set_request_path_attr(struct inode *rinode, struct dentry *rdentry, r = build_dentry_path(rdentry, ppath, pathlen, ino, freepath); dout(" dentry %p %llx/%.*s\n", rdentry, *ino, *pathlen, *ppath); - } else if (rpath) { + } else if (rpath || rino) { *ino = rino; *ppath = rpath; *pathlen = strlen(rpath); -- cgit v1.2.3 From 016f1c54408b1e92e2e8087bfc05ca0a9c258513 Mon Sep 17 00:00:00 2001 From: Michal Marek Date: Thu, 11 Aug 2011 12:29:46 +0200 Subject: UBIFS: not build debug messages with CONFIG_UBIFS_FS_DEBUG disabled With $ grep -e UBIFS_FS_DEBUG -e DYNAMIC_DEBUG .config # CONFIG_UBIFS_FS_DEBUG is not set CONFIG_DYNAMIC_DEBUG=y Debug messages are kept in the object files due to the dynamic_pr_debug() macro, even if they are never going to be printed: $ make fs/ubifs/super.o $ strings fs/ubifs/super.o | grep 'compiled on' compiled on: Aug 11 2011 at 12:21:38 Use plain printk to fix this. Signed-off-by: Michal Marek Signed-off-by: Artem Bityutskiy --- fs/ubifs/debug.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h index 45174b534377..feb361e252ac 100644 --- a/fs/ubifs/debug.h +++ b/fs/ubifs/debug.h @@ -335,9 +335,9 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c); #define DBGKEY(key) ((char *)(key)) #define DBGKEY1(key) ((char *)(key)) -#define ubifs_dbg_msg(fmt, ...) do { \ - if (0) \ - pr_debug(fmt "\n", ##__VA_ARGS__); \ +#define ubifs_dbg_msg(fmt, ...) do { \ + if (0) \ + printk(KERN_DEBUG fmt "\n", ##__VA_ARGS__); \ } while (0) #define dbg_dump_stack() -- cgit v1.2.3 From 259a187ade45056fd44856654f78aa9e9f0f7c75 Mon Sep 17 00:00:00 2001 From: Noah Watkins Date: Mon, 22 Aug 2011 13:49:41 -0600 Subject: ceph: fix memory leak kfree does not clean up indirect allocations in ceph_fs_client and ceph_options (e.g. snapdir_name). Signed-off-by: Noah Watkins Signed-off-by: Sage Weil --- fs/ceph/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ceph/super.c b/fs/ceph/super.c index d47c5ec7fb1f..88bacaf385d9 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -813,8 +813,8 @@ static struct dentry *ceph_mount(struct file_system_type *fs_type, fsc = create_fs_client(fsopt, opt); if (IS_ERR(fsc)) { res = ERR_CAST(fsc); - kfree(fsopt); - kfree(opt); + destroy_mount_options(fsopt); + ceph_destroy_options(opt); goto out_final; } -- cgit v1.2.3 From 242d621964dd8641df53f7d51d4c6ead655cc5a6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 24 Aug 2011 05:57:51 +0000 Subject: xfs: deprecate the nodelaylog mount option Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_super.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 9a72dda58bd0..c1b022f20d35 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -356,6 +356,8 @@ xfs_parseargs( mp->m_flags |= XFS_MOUNT_DELAYLOG; } else if (!strcmp(this_char, MNTOPT_NODELAYLOG)) { mp->m_flags &= ~XFS_MOUNT_DELAYLOG; + xfs_warn(mp, + "nodelaylog is deprecated and will be removed in Linux 3.3"); } else if (!strcmp(this_char, MNTOPT_DISCARD)) { mp->m_flags |= XFS_MOUNT_DISCARD; } else if (!strcmp(this_char, MNTOPT_NODISCARD)) { -- cgit v1.2.3 From 8c0bec2151a47906bf779c6715a10ce04453ab77 Mon Sep 17 00:00:00 2001 From: Jiaying Zhang Date: Wed, 31 Aug 2011 11:50:51 -0400 Subject: ext4: remove i_mutex lock in ext4_evict_inode to fix lockdep complaining The i_mutex lock and flush_completed_IO() added by commit 2581fdc810 in ext4_evict_inode() causes lockdep complaining about potential deadlock in several places. In most/all of these LOCKDEP complaints it looks like it's a false positive, since many of the potential circular locking cases can't take place by the time the ext4_evict_inode() is called; but since at the very least it may mask real problems, we need to address this. This change removes the flush_completed_IO() and i_mutex lock in ext4_evict_inode(). Instead, we take a different approach to resolve the software lockup that commit 2581fdc810 intends to fix. Rather than having ext4-dio-unwritten thread wait for grabing the i_mutex lock of an inode, we use mutex_trylock() instead, and simply requeue the work item if we fail to grab the inode's i_mutex lock. This should speed up work queue processing in general and also prevents the following deadlock scenario: During page fault, shrink_icache_memory is called that in turn evicts another inode B. Inode B has some pending io_end work so it calls ext4_ioend_wait() that waits for inode B's i_ioend_count to become zero. However, inode B's ioend work was queued behind some of inode A's ioend work on the same cpu's ext4-dio-unwritten workqueue. As the ext4-dio-unwritten thread on that cpu is processing inode A's ioend work, it tries to grab inode A's i_mutex lock. Since the i_mutex lock of inode A is still hold before the page fault happened, we enter a deadlock. Signed-off-by: Jiaying Zhang Signed-off-by: "Theodore Ts'o" --- fs/ext4/ext4.h | 1 + fs/ext4/inode.c | 3 --- fs/ext4/page-io.c | 18 +++++++++++++++++- 3 files changed, 18 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index e717dfd2f2b4..b7d7bd0f066e 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -175,6 +175,7 @@ struct mpage_da_data { */ #define EXT4_IO_END_UNWRITTEN 0x0001 #define EXT4_IO_END_ERROR 0x0002 +#define EXT4_IO_END_QUEUED 0x0004 struct ext4_io_page { struct page *p_page; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index c4da98a959ae..18d2558b7624 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -121,9 +121,6 @@ void ext4_evict_inode(struct inode *inode) trace_ext4_evict_inode(inode); - mutex_lock(&inode->i_mutex); - ext4_flush_completed_IO(inode); - mutex_unlock(&inode->i_mutex); ext4_ioend_wait(inode); if (inode->i_nlink) { diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c index 78839af7ce29..92f38ee13f8a 100644 --- a/fs/ext4/page-io.c +++ b/fs/ext4/page-io.c @@ -142,7 +142,23 @@ static void ext4_end_io_work(struct work_struct *work) unsigned long flags; int ret; - mutex_lock(&inode->i_mutex); + if (!mutex_trylock(&inode->i_mutex)) { + /* + * Requeue the work instead of waiting so that the work + * items queued after this can be processed. + */ + queue_work(EXT4_SB(inode->i_sb)->dio_unwritten_wq, &io->work); + /* + * To prevent the ext4-dio-unwritten thread from keeping + * requeueing end_io requests and occupying cpu for too long, + * yield the cpu if it sees an end_io request that has already + * been requeued. + */ + if (io->flag & EXT4_IO_END_QUEUED) + yield(); + io->flag |= EXT4_IO_END_QUEUED; + return; + } ret = ext4_end_io_nolock(io); if (ret < 0) { mutex_unlock(&inode->i_mutex); -- cgit v1.2.3 From 866e4ed77448a0c311e1b055eb72ea05423fd799 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sat, 27 Aug 2011 05:57:44 +0000 Subject: xfs: fix xfs_mark_inode_dirty during umount During umount we do not add a dirty inode to the lru and wait for it to become clean first, but force writeback of data and metadata with I_WILL_FREE set. Currently there is no way for XFS to detect that the inode has been redirtied for metadata operations, as we skip the mark_inode_dirty call during teardown. Fix this by setting i_update_core nanually in that case, so that the inode gets flushed during inode reclaim. Alternatively we could enable calling mark_inode_dirty for inodes in I_WILL_FREE state, and let the VFS dirty tracking handle this. I decided against this as we will get better I/O patterns from reclaim compared to the synchronous writeout in write_inode_now, and always marking the inode dirty in some way from xfs_mark_inode_dirty is a better safetly net in either case. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder (cherry picked from commit da6742a5a4cc844a9982fdd936ddb537c0747856) Signed-off-by: Alex Elder --- fs/xfs/xfs_iops.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index b9c172b3fbbe..673704fab748 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -70,9 +70,8 @@ xfs_synchronize_times( } /* - * If the linux inode is valid, mark it dirty. - * Used when committing a dirty inode into a transaction so that - * the inode will get written back by the linux code + * If the linux inode is valid, mark it dirty, else mark the dirty state + * in the XFS inode to make sure we pick it up when reclaiming the inode. */ void xfs_mark_inode_dirty_sync( @@ -82,6 +81,10 @@ xfs_mark_inode_dirty_sync( if (!(inode->i_state & (I_WILL_FREE|I_FREEING))) mark_inode_dirty_sync(inode); + else { + barrier(); + ip->i_update_core = 1; + } } void @@ -92,6 +95,11 @@ xfs_mark_inode_dirty( if (!(inode->i_state & (I_WILL_FREE|I_FREEING))) mark_inode_dirty(inode); + else { + barrier(); + ip->i_update_core = 1; + } + } /* -- cgit v1.2.3 From 58d84c4ee0389ddeb86238d5d8359a982c9f7a5b Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sat, 27 Aug 2011 05:57:55 +0000 Subject: xfs: fix ->write_inode return values Currently we always redirty an inode that was attempted to be written out synchronously but has been cleaned by an AIL pushed internall, which is rather bogus. Fix that by doing the i_update_core check early on and return 0 for it. Also include async calls for it, as doing any work for those is just as pointless. While we're at it also fix the sign for the EIO return in case of a filesystem shutdown, and fix the completely non-sensical locking around xfs_log_inode. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder (cherry picked from commit 297db93bb74cf687510313eb235a7aec14d67e97) Signed-off-by: Alex Elder --- fs/xfs/xfs_super.c | 34 +++++++++------------------------- 1 file changed, 9 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index c1b022f20d35..2366c54cc4fa 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -879,33 +879,17 @@ xfs_log_inode( struct xfs_trans *tp; int error; - xfs_iunlock(ip, XFS_ILOCK_SHARED); tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); error = xfs_trans_reserve(tp, 0, XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0); - if (error) { xfs_trans_cancel(tp, 0); - /* we need to return with the lock hold shared */ - xfs_ilock(ip, XFS_ILOCK_SHARED); return error; } xfs_ilock(ip, XFS_ILOCK_EXCL); - - /* - * Note - it's possible that we might have pushed ourselves out of the - * way during trans_reserve which would flush the inode. But there's - * no guarantee that the inode buffer has actually gone out yet (it's - * delwri). Plus the buffer could be pinned anyway if it's part of - * an inode in another recent transaction. So we play it safe and - * fire off the transaction anyway. - */ - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - error = xfs_trans_commit(tp, 0); - xfs_ilock_demote(ip, XFS_ILOCK_EXCL); - - return error; + return xfs_trans_commit(tp, 0); } STATIC int @@ -920,7 +904,9 @@ xfs_fs_write_inode( trace_xfs_write_inode(ip); if (XFS_FORCED_SHUTDOWN(mp)) - return XFS_ERROR(EIO); + return -XFS_ERROR(EIO); + if (!ip->i_update_core) + return 0; if (wbc->sync_mode == WB_SYNC_ALL) { /* @@ -931,12 +917,10 @@ xfs_fs_write_inode( * of synchronous log foces dramatically. */ xfs_ioend_wait(ip); - xfs_ilock(ip, XFS_ILOCK_SHARED); - if (ip->i_update_core) { - error = xfs_log_inode(ip); - if (error) - goto out_unlock; - } + error = xfs_log_inode(ip); + if (error) + goto out; + return 0; } else { /* * We make this non-blocking if the inode is contended, return -- cgit v1.2.3 From 5441ae5eb3614d3c28f77073370738a2820c88e4 Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Mon, 25 Jul 2011 18:06:32 +0000 Subject: fs/9p: Add fid before dentry instantiation d_instantiate marks the dentry positive. So a parallel lookup and mkdir of the directory can find dentry that doesn't have fid attached. This can result in both the code path doing v9fs_fid_add which results in v9fs_dentry leak. Signed-off-by: Aneesh Kumar K.V Signed-off-by: Eric Van Hensbergen --- fs/9p/vfs_inode.c | 4 +--- fs/9p/vfs_inode_dotl.c | 8 ++++---- 2 files changed, 5 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 8bb5507e822f..43dd540663af 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -645,13 +645,11 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir, P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n", err); goto error; } - d_instantiate(dentry, inode); err = v9fs_fid_add(dentry, fid); if (err < 0) goto error; - + d_instantiate(dentry, inode); return ofid; - error: if (ofid) p9_client_clunk(ofid); diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index b6c8ed205192..0ca224c8bb60 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -281,10 +281,10 @@ v9fs_vfs_create_dotl(struct inode *dir, struct dentry *dentry, int omode, P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n", err); goto error; } - d_instantiate(dentry, inode); err = v9fs_fid_add(dentry, fid); if (err < 0) goto error; + d_instantiate(dentry, inode); /* Now set the ACL based on the default value */ v9fs_set_create_acl(dentry, &dacl, &pacl); @@ -403,10 +403,10 @@ static int v9fs_vfs_mkdir_dotl(struct inode *dir, err); goto error; } - d_instantiate(dentry, inode); err = v9fs_fid_add(dentry, fid); if (err < 0) goto error; + d_instantiate(dentry, inode); fid = NULL; } else { /* @@ -657,10 +657,10 @@ v9fs_vfs_symlink_dotl(struct inode *dir, struct dentry *dentry, err); goto error; } - d_instantiate(dentry, inode); err = v9fs_fid_add(dentry, fid); if (err < 0) goto error; + d_instantiate(dentry, inode); fid = NULL; } else { /* Not in cached mode. No need to populate inode with stat */ @@ -810,10 +810,10 @@ v9fs_vfs_mknod_dotl(struct inode *dir, struct dentry *dentry, int omode, err); goto error; } - d_instantiate(dentry, inode); err = v9fs_fid_add(dentry, fid); if (err < 0) goto error; + d_instantiate(dentry, inode); fid = NULL; } else { /* -- cgit v1.2.3 From 45089142b1497dab2327d60f6c71c40766fc3ea4 Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Mon, 25 Jul 2011 18:06:33 +0000 Subject: fs/9p: Don't update file type when updating file attributes We should only update attributes that we can change on stat2inode. Also do file type initialization in v9fs_init_inode. Signed-off-by: Aneesh Kumar K.V Signed-off-by: Eric Van Hensbergen --- fs/9p/v9fs_vfs.h | 4 +-- fs/9p/vfs_inode.c | 91 +++++++++++++++++++++++++++----------------------- fs/9p/vfs_inode_dotl.c | 23 ++++++++----- fs/9p/vfs_super.c | 2 +- 4 files changed, 68 insertions(+), 52 deletions(-) (limited to 'fs') diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h index 46ce357ca1ab..7ac1faec2bde 100644 --- a/fs/9p/v9fs_vfs.h +++ b/fs/9p/v9fs_vfs.h @@ -54,9 +54,9 @@ extern struct kmem_cache *v9fs_inode_cache; struct inode *v9fs_alloc_inode(struct super_block *sb); void v9fs_destroy_inode(struct inode *inode); -struct inode *v9fs_get_inode(struct super_block *sb, int mode); +struct inode *v9fs_get_inode(struct super_block *sb, int mode, dev_t); int v9fs_init_inode(struct v9fs_session_info *v9ses, - struct inode *inode, int mode); + struct inode *inode, int mode, dev_t); void v9fs_evict_inode(struct inode *inode); ino_t v9fs_qid2ino(struct p9_qid *qid); void v9fs_stat2inode(struct p9_wstat *, struct inode *, struct super_block *); diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 43dd540663af..3563cace0a2e 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -95,15 +95,18 @@ static int unixmode2p9mode(struct v9fs_session_info *v9ses, int mode) /** * p9mode2unixmode- convert plan9 mode bits to unix mode bits * @v9ses: v9fs session information - * @mode: mode to convert + * @stat: p9_wstat from which mode need to be derived + * @rdev: major number, minor number in case of device files. * */ - -static int p9mode2unixmode(struct v9fs_session_info *v9ses, int mode) +static int p9mode2unixmode(struct v9fs_session_info *v9ses, + struct p9_wstat *stat, dev_t *rdev) { int res; + int mode = stat->mode; - res = mode & 0777; + res = mode & S_IALLUGO; + *rdev = 0; if ((mode & P9_DMDIR) == P9_DMDIR) res |= S_IFDIR; @@ -116,9 +119,26 @@ static int p9mode2unixmode(struct v9fs_session_info *v9ses, int mode) && (v9ses->nodev == 0)) res |= S_IFIFO; else if ((mode & P9_DMDEVICE) && (v9fs_proto_dotu(v9ses)) - && (v9ses->nodev == 0)) - res |= S_IFBLK; - else + && (v9ses->nodev == 0)) { + char type = 0, ext[32]; + int major = -1, minor = -1; + + strncpy(ext, stat->extension, sizeof(ext)); + sscanf(ext, "%c %u %u", &type, &major, &minor); + switch (type) { + case 'c': + res |= S_IFCHR; + break; + case 'b': + res |= S_IFBLK; + break; + default: + P9_DPRINTK(P9_DEBUG_ERROR, + "Unknown special type %c %s\n", type, + stat->extension); + }; + *rdev = MKDEV(major, minor); + } else res |= S_IFREG; if (v9fs_proto_dotu(v9ses)) { @@ -131,7 +151,6 @@ static int p9mode2unixmode(struct v9fs_session_info *v9ses, int mode) if ((mode & P9_DMSETVTX) == P9_DMSETVTX) res |= S_ISVTX; } - return res; } @@ -242,13 +261,13 @@ void v9fs_destroy_inode(struct inode *inode) } int v9fs_init_inode(struct v9fs_session_info *v9ses, - struct inode *inode, int mode) + struct inode *inode, int mode, dev_t rdev) { int err = 0; inode_init_owner(inode, NULL, mode); inode->i_blocks = 0; - inode->i_rdev = 0; + inode->i_rdev = rdev; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; inode->i_mapping->a_ops = &v9fs_addr_operations; @@ -335,7 +354,7 @@ error: * */ -struct inode *v9fs_get_inode(struct super_block *sb, int mode) +struct inode *v9fs_get_inode(struct super_block *sb, int mode, dev_t rdev) { int err; struct inode *inode; @@ -348,7 +367,7 @@ struct inode *v9fs_get_inode(struct super_block *sb, int mode) P9_EPRINTK(KERN_WARNING, "Problem allocating inode\n"); return ERR_PTR(-ENOMEM); } - err = v9fs_init_inode(v9ses, inode, mode); + err = v9fs_init_inode(v9ses, inode, mode, rdev); if (err) { iput(inode); return ERR_PTR(err); @@ -435,11 +454,12 @@ void v9fs_evict_inode(struct inode *inode) static int v9fs_test_inode(struct inode *inode, void *data) { int umode; + dev_t rdev; struct v9fs_inode *v9inode = V9FS_I(inode); struct p9_wstat *st = (struct p9_wstat *)data; struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode); - umode = p9mode2unixmode(v9ses, st->mode); + umode = p9mode2unixmode(v9ses, st, &rdev); /* don't match inode of different type */ if ((inode->i_mode & S_IFMT) != (umode & S_IFMT)) return 0; @@ -473,6 +493,7 @@ static struct inode *v9fs_qid_iget(struct super_block *sb, struct p9_wstat *st, int new) { + dev_t rdev; int retval, umode; unsigned long i_ino; struct inode *inode; @@ -496,8 +517,8 @@ static struct inode *v9fs_qid_iget(struct super_block *sb, * later. */ inode->i_ino = i_ino; - umode = p9mode2unixmode(v9ses, st->mode); - retval = v9fs_init_inode(v9ses, inode, umode); + umode = p9mode2unixmode(v9ses, st, &rdev); + retval = v9fs_init_inode(v9ses, inode, umode, rdev); if (retval) goto error; @@ -1000,7 +1021,7 @@ v9fs_vfs_getattr(struct vfsmount *mnt, struct dentry *dentry, return PTR_ERR(st); v9fs_stat2inode(st, dentry->d_inode, dentry->d_inode->i_sb); - generic_fillattr(dentry->d_inode, stat); + generic_fillattr(dentry->d_inode, stat); p9stat_free(st); kfree(st); @@ -1084,6 +1105,7 @@ void v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode, struct super_block *sb) { + mode_t mode; char ext[32]; char tag_name[14]; unsigned int i_nlink; @@ -1119,31 +1141,9 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode, inode->i_nlink = i_nlink; } } - inode->i_mode = p9mode2unixmode(v9ses, stat->mode); - if ((S_ISBLK(inode->i_mode)) || (S_ISCHR(inode->i_mode))) { - char type = 0; - int major = -1; - int minor = -1; - - strncpy(ext, stat->extension, sizeof(ext)); - sscanf(ext, "%c %u %u", &type, &major, &minor); - switch (type) { - case 'c': - inode->i_mode &= ~S_IFBLK; - inode->i_mode |= S_IFCHR; - break; - case 'b': - break; - default: - P9_DPRINTK(P9_DEBUG_ERROR, - "Unknown special type %c %s\n", type, - stat->extension); - }; - inode->i_rdev = MKDEV(major, minor); - init_special_inode(inode, inode->i_mode, inode->i_rdev); - } else - inode->i_rdev = 0; - + mode = stat->mode & S_IALLUGO; + mode |= inode->i_mode & ~S_IALLUGO; + inode->i_mode = mode; i_size_write(inode, stat->length); /* not real number of blocks, but 512 byte ones ... */ @@ -1409,6 +1409,8 @@ v9fs_vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev) int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode) { + int umode; + dev_t rdev; loff_t i_size; struct p9_wstat *st; struct v9fs_session_info *v9ses; @@ -1417,6 +1419,12 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode) st = p9_client_stat(fid); if (IS_ERR(st)) return PTR_ERR(st); + /* + * Don't update inode if the file type is different + */ + umode = p9mode2unixmode(v9ses, st, &rdev); + if ((inode->i_mode & S_IFMT) != (umode & S_IFMT)) + goto out; spin_lock(&inode->i_lock); /* @@ -1428,6 +1436,7 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode) if (v9ses->cache) inode->i_size = i_size; spin_unlock(&inode->i_lock); +out: p9stat_free(st); kfree(st); return 0; diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index 0ca224c8bb60..a3f2540cc4b2 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -153,7 +153,8 @@ static struct inode *v9fs_qid_iget_dotl(struct super_block *sb, * later. */ inode->i_ino = i_ino; - retval = v9fs_init_inode(v9ses, inode, st->st_mode); + retval = v9fs_init_inode(v9ses, inode, + st->st_mode, new_decode_dev(st->st_rdev)); if (retval) goto error; @@ -414,7 +415,7 @@ static int v9fs_vfs_mkdir_dotl(struct inode *dir, * inode with stat. We need to get an inode * so that we can set the acl with dentry */ - inode = v9fs_get_inode(dir->i_sb, mode); + inode = v9fs_get_inode(dir->i_sb, mode, 0); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto error; @@ -540,6 +541,7 @@ int v9fs_vfs_setattr_dotl(struct dentry *dentry, struct iattr *iattr) void v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode) { + mode_t mode; struct v9fs_inode *v9inode = V9FS_I(inode); if ((stat->st_result_mask & P9_STATS_BASIC) == P9_STATS_BASIC) { @@ -552,11 +554,10 @@ v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode) inode->i_uid = stat->st_uid; inode->i_gid = stat->st_gid; inode->i_nlink = stat->st_nlink; - inode->i_mode = stat->st_mode; - inode->i_rdev = new_decode_dev(stat->st_rdev); - if ((S_ISBLK(inode->i_mode)) || (S_ISCHR(inode->i_mode))) - init_special_inode(inode, inode->i_mode, inode->i_rdev); + mode = stat->st_mode & S_IALLUGO; + mode |= inode->i_mode & ~S_IALLUGO; + inode->i_mode = mode; i_size_write(inode, stat->st_size); inode->i_blocks = stat->st_blocks; @@ -664,7 +665,7 @@ v9fs_vfs_symlink_dotl(struct inode *dir, struct dentry *dentry, fid = NULL; } else { /* Not in cached mode. No need to populate inode with stat */ - inode = v9fs_get_inode(dir->i_sb, S_IFLNK); + inode = v9fs_get_inode(dir->i_sb, S_IFLNK, 0); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto error; @@ -820,7 +821,7 @@ v9fs_vfs_mknod_dotl(struct inode *dir, struct dentry *dentry, int omode, * Not in cached mode. No need to populate inode with stat. * socket syscall returns a fd, so we need instantiate */ - inode = v9fs_get_inode(dir->i_sb, mode); + inode = v9fs_get_inode(dir->i_sb, mode, rdev); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto error; @@ -886,6 +887,11 @@ int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode) st = p9_client_getattr_dotl(fid, P9_STATS_ALL); if (IS_ERR(st)) return PTR_ERR(st); + /* + * Don't update inode if the file type is different + */ + if ((inode->i_mode & S_IFMT) != (st->st_mode & S_IFMT)) + goto out; spin_lock(&inode->i_lock); /* @@ -897,6 +903,7 @@ int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode) if (v9ses->cache) inode->i_size = i_size; spin_unlock(&inode->i_lock); +out: kfree(st); return 0; } diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index feef6cdc1fd2..c70251d47ed1 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -149,7 +149,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags, else sb->s_d_op = &v9fs_dentry_operations; - inode = v9fs_get_inode(sb, S_IFDIR | mode); + inode = v9fs_get_inode(sb, S_IFDIR | mode, 0); if (IS_ERR(inode)) { retval = PTR_ERR(inode); goto release_sb; -- cgit v1.2.3 From f88657ce3f9713a0c62101dffb0e972a979e77b9 Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Wed, 3 Aug 2011 19:55:32 +0530 Subject: fs/9p: Add OS dependent open flags in 9p protocol Some of the flags are OS/arch dependent we add a 9p protocol value which maps to asm-generic/fcntl.h values in Linux Based on the original patch from Venkateswararao Jujjuri Signed-off-by: Aneesh Kumar K.V --- fs/9p/v9fs_vfs.h | 2 ++ fs/9p/vfs_file.c | 2 +- fs/9p/vfs_inode.c | 16 ++++++++++++++- fs/9p/vfs_inode_dotl.c | 55 +++++++++++++++++++++++++++++++++++++++++++++++++- 4 files changed, 72 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h index 7ac1faec2bde..410ffd6ceb5f 100644 --- a/fs/9p/v9fs_vfs.h +++ b/fs/9p/v9fs_vfs.h @@ -83,4 +83,6 @@ static inline void v9fs_invalidate_inode_attr(struct inode *inode) v9inode->cache_validity |= V9FS_INO_INVALID_ATTR; return; } + +int v9fs_open_to_dotl_flags(int flags); #endif diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c index 3c173fcc2c5a..c2f107583125 100644 --- a/fs/9p/vfs_file.c +++ b/fs/9p/vfs_file.c @@ -65,7 +65,7 @@ int v9fs_file_open(struct inode *inode, struct file *file) v9inode = V9FS_I(inode); v9ses = v9fs_inode2v9ses(inode); if (v9fs_proto_dotl(v9ses)) - omode = file->f_flags; + omode = v9fs_open_to_dotl_flags(file->f_flags); else omode = v9fs_uflags2omode(file->f_flags, v9fs_proto_dotu(v9ses)); diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 3563cace0a2e..9e3ea6ce6951 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -552,6 +552,19 @@ v9fs_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid, return inode; } +/** + * v9fs_at_to_dotl_flags- convert Linux specific AT flags to + * plan 9 AT flag. + * @flags: flags to convert + */ +static int v9fs_at_to_dotl_flags(int flags) +{ + int rflags = 0; + if (flags & AT_REMOVEDIR) + rflags |= P9_DOTL_AT_REMOVEDIR; + return rflags; +} + /** * v9fs_remove - helper function to remove files and directories * @dir: directory inode that is being deleted @@ -579,7 +592,8 @@ static int v9fs_remove(struct inode *dir, struct dentry *dentry, int flags) return retval; } if (v9fs_proto_dotl(v9ses)) - retval = p9_client_unlinkat(dfid, dentry->d_name.name, flags); + retval = p9_client_unlinkat(dfid, dentry->d_name.name, + v9fs_at_to_dotl_flags(flags)); if (retval == -EOPNOTSUPP) { /* Try the one based on path */ v9fid = v9fs_fid_clone(dentry); diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index a3f2540cc4b2..aded79fcd5cf 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -191,6 +191,58 @@ v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses, struct p9_fid *fid, return inode; } +struct dotl_openflag_map { + int open_flag; + int dotl_flag; +}; + +static int v9fs_mapped_dotl_flags(int flags) +{ + int i; + int rflags = 0; + struct dotl_openflag_map dotl_oflag_map[] = { + { O_CREAT, P9_DOTL_CREATE }, + { O_EXCL, P9_DOTL_EXCL }, + { O_NOCTTY, P9_DOTL_NOCTTY }, + { O_TRUNC, P9_DOTL_TRUNC }, + { O_APPEND, P9_DOTL_APPEND }, + { O_NONBLOCK, P9_DOTL_NONBLOCK }, + { O_DSYNC, P9_DOTL_DSYNC }, + { FASYNC, P9_DOTL_FASYNC }, + { O_DIRECT, P9_DOTL_DIRECT }, + { O_LARGEFILE, P9_DOTL_LARGEFILE }, + { O_DIRECTORY, P9_DOTL_DIRECTORY }, + { O_NOFOLLOW, P9_DOTL_NOFOLLOW }, + { O_NOATIME, P9_DOTL_NOATIME }, + { O_CLOEXEC, P9_DOTL_CLOEXEC }, + { O_SYNC, P9_DOTL_SYNC}, + }; + for (i = 0; i < ARRAY_SIZE(dotl_oflag_map); i++) { + if (flags & dotl_oflag_map[i].open_flag) + rflags |= dotl_oflag_map[i].dotl_flag; + } + return rflags; +} + +/** + * v9fs_open_to_dotl_flags- convert Linux specific open flags to + * plan 9 open flag. + * @flags: flags to convert + */ +int v9fs_open_to_dotl_flags(int flags) +{ + int rflags = 0; + + /* + * We have same bits for P9_DOTL_READONLY, P9_DOTL_WRONLY + * and P9_DOTL_NOACCESS + */ + rflags |= flags & O_ACCMODE; + rflags |= v9fs_mapped_dotl_flags(flags); + + return rflags; +} + /** * v9fs_vfs_create_dotl - VFS hook to create files for 9P2000.L protocol. * @dir: directory inode that is being created @@ -259,7 +311,8 @@ v9fs_vfs_create_dotl(struct inode *dir, struct dentry *dentry, int omode, "Failed to get acl values in creat %d\n", err); goto error; } - err = p9_client_create_dotl(ofid, name, flags, mode, gid, &qid); + err = p9_client_create_dotl(ofid, name, v9fs_open_to_dotl_flags(flags), + mode, gid, &qid); if (err < 0) { P9_DPRINTK(P9_DEBUG_VFS, "p9_client_open_dotl failed in creat %d\n", -- cgit v1.2.3 From 73f507171cfa407b19f254aef95cbb058c8180cf Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Tue, 16 Aug 2011 22:19:28 +0530 Subject: fs/9p: Always ask new inode in lookup for cache mode disabled This make sure we don't end up reusing the unlinked inode object. The ideal way is to use inode i_generation. But i_generation is not available in userspace always. Signed-off-by: Aneesh Kumar K.V --- fs/9p/vfs_inode.c | 28 +++++++++++++++++++++------- 1 file changed, 21 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 9e3ea6ce6951..e3c03db3c788 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -825,6 +825,7 @@ static int v9fs_vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nameidata) { + struct dentry *res; struct super_block *sb; struct v9fs_session_info *v9ses; struct p9_fid *dfid, *fid; @@ -856,22 +857,35 @@ struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry, return ERR_PTR(result); } - - inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb); + /* + * Make sure we don't use a wrong inode due to parallel + * unlink. For cached mode create calls request for new + * inode. But with cache disabled, lookup should do this. + */ + if (v9ses->cache) + inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb); + else + inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); if (IS_ERR(inode)) { result = PTR_ERR(inode); inode = NULL; goto error; } - result = v9fs_fid_add(dentry, fid); if (result < 0) goto error_iput; - inst_out: - d_add(dentry, inode); - return NULL; - + /* + * If we had a rename on the server and a parallel lookup + * for the new name, then make sure we instantiate with + * the new name. ie look up for a/b, while on server somebody + * moved b under k and client parallely did a lookup for + * k/b. + */ + res = d_materialise_unique(dentry, inode); + if (!IS_ERR(res)) + return res; + result = PTR_ERR(res); error_iput: iput(inode); error: -- cgit v1.2.3 From 51b8b4fb32271d39fbdd760397406177b2b0fd36 Mon Sep 17 00:00:00 2001 From: Jim Garlick Date: Sun, 21 Aug 2011 00:21:18 +0530 Subject: fs/9p: Use protocol-defined value for lock/getlock 'type' field. Signed-off-by: Jim Garlick Signed-off-by: Aneesh Kumar K.V --- fs/9p/vfs_file.c | 34 +++++++++++++++++++++++++++------- 1 file changed, 27 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c index c2f107583125..62857a810a79 100644 --- a/fs/9p/vfs_file.c +++ b/fs/9p/vfs_file.c @@ -169,7 +169,18 @@ static int v9fs_file_do_lock(struct file *filp, int cmd, struct file_lock *fl) /* convert posix lock to p9 tlock args */ memset(&flock, 0, sizeof(flock)); - flock.type = fl->fl_type; + /* map the lock type */ + switch (fl->fl_type) { + case F_RDLCK: + flock.type = P9_LOCK_TYPE_RDLCK; + break; + case F_WRLCK: + flock.type = P9_LOCK_TYPE_WRLCK; + break; + case F_UNLCK: + flock.type = P9_LOCK_TYPE_UNLCK; + break; + } flock.start = fl->fl_start; if (fl->fl_end == OFFSET_MAX) flock.length = 0; @@ -245,7 +256,7 @@ static int v9fs_file_getlock(struct file *filp, struct file_lock *fl) /* convert posix lock to p9 tgetlock args */ memset(&glock, 0, sizeof(glock)); - glock.type = fl->fl_type; + glock.type = P9_LOCK_TYPE_UNLCK; glock.start = fl->fl_start; if (fl->fl_end == OFFSET_MAX) glock.length = 0; @@ -257,17 +268,26 @@ static int v9fs_file_getlock(struct file *filp, struct file_lock *fl) res = p9_client_getlock_dotl(fid, &glock); if (res < 0) return res; - if (glock.type != F_UNLCK) { - fl->fl_type = glock.type; + /* map 9p lock type to os lock type */ + switch (glock.type) { + case P9_LOCK_TYPE_RDLCK: + fl->fl_type = F_RDLCK; + break; + case P9_LOCK_TYPE_WRLCK: + fl->fl_type = F_WRLCK; + break; + case P9_LOCK_TYPE_UNLCK: + fl->fl_type = F_UNLCK; + break; + } + if (glock.type != P9_LOCK_TYPE_UNLCK) { fl->fl_start = glock.start; if (glock.length == 0) fl->fl_end = OFFSET_MAX; else fl->fl_end = glock.start + glock.length - 1; fl->fl_pid = glock.proc_id; - } else - fl->fl_type = F_UNLCK; - + } return res; } -- cgit v1.2.3 From 0ec26fd0698285b31248e34bf1abb022c00f23d6 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Mon, 5 Sep 2011 18:06:26 +0200 Subject: vfs: automount should ignore LOOKUP_FOLLOW Prior to 2.6.38 automount would not trigger on either stat(2) or lstat(2) on the automount point. After 2.6.38, with the introduction of the ->d_automount() infrastructure, stat(2) and others would start triggering automount while lstat(2), etc. still would not. This is a regression and a userspace ABI change. Problem originally reported here: http://thread.gmane.org/gmane.linux.kernel.autofs/6098 It appears that there was an attempt at fixing various userspace tools to not trigger the automount. But since the stat system call is rather common it is impossible to "fix" all userspace. This patch reverts the original behavior, which is to not trigger on stat(2) and other symlink following syscalls. [ It's not really clear what the right behavior is. Apparently Solaris does the "automount on stat, leave alone on lstat". And some programs can get unhappy when "stat+open+fstat" ends up giving a different result from the fstat than from the initial stat. But the change in 2.6.38 resulted in problems for some people, so we're going back to old behavior. Maybe we can re-visit this discussion at some future date - Linus ] Reported-by: Leonardo Chiquitto Signed-off-by: Miklos Szeredi Acked-by: Ian Kent Cc: David Howells Cc: stable@kernel.org Signed-off-by: Linus Torvalds --- fs/namei.c | 33 +++++++++++++++------------------ 1 file changed, 15 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 2826db35dc25..b52bc685465f 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -727,25 +727,22 @@ static int follow_automount(struct path *path, unsigned flags, if ((flags & LOOKUP_NO_AUTOMOUNT) && !(flags & LOOKUP_PARENT)) return -EISDIR; /* we actually want to stop here */ - /* - * We don't want to mount if someone's just doing a stat and they've - * set AT_SYMLINK_NOFOLLOW - unless they're stat'ing a directory and - * appended a '/' to the name. + /* We don't want to mount if someone's just doing a stat - + * unless they're stat'ing a directory and appended a '/' to + * the name. + * + * We do, however, want to mount if someone wants to open or + * create a file of any type under the mountpoint, wants to + * traverse through the mountpoint or wants to open the + * mounted directory. Also, autofs may mark negative dentries + * as being automount points. These will need the attentions + * of the daemon to instantiate them before they can be used. */ - if (!(flags & LOOKUP_FOLLOW)) { - /* We do, however, want to mount if someone wants to open or - * create a file of any type under the mountpoint, wants to - * traverse through the mountpoint or wants to open the mounted - * directory. - * Also, autofs may mark negative dentries as being automount - * points. These will need the attentions of the daemon to - * instantiate them before they can be used. - */ - if (!(flags & (LOOKUP_PARENT | LOOKUP_DIRECTORY | - LOOKUP_OPEN | LOOKUP_CREATE)) && - path->dentry->d_inode) - return -EISDIR; - } + if (!(flags & (LOOKUP_PARENT | LOOKUP_DIRECTORY | + LOOKUP_OPEN | LOOKUP_CREATE)) && + path->dentry->d_inode) + return -EISDIR; + current->total_link_count++; if (current->total_link_count >= 40) return -ELOOP; -- cgit v1.2.3 From 94007751bb02797ba87bac7aacee2731ac2039a3 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Sat, 10 Sep 2011 17:20:21 +1000 Subject: Avoid dereferencing a 'request_queue' after last close. On the last close of an 'md' device which as been stopped, the device is destroyed and in particular the request_queue is freed. The free is done in a separate thread so it might happen a short time later. __blkdev_put calls bdev_inode_switch_bdi *after* ->release has been called. Since commit f758eeabeb96f878c860e8f110f94ec8820822a9 bdev_inode_switch_bdi will dereference the 'old' bdi, which lives inside a request_queue, to get a spin lock. This causes the last close on an md device to sometime take a spin_lock which lives in freed memory - which results in an oops. So move the called to bdev_inode_switch_bdi before the call to ->release. Cc: Christoph Hellwig Cc: Hugh Dickins Cc: Andrew Morton Cc: Wu Fengguang Acked-by: Wu Fengguang Cc: stable@kernel.org Signed-off-by: NeilBrown --- fs/block_dev.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/block_dev.c b/fs/block_dev.c index ff77262e887c..95f786ec7f08 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1429,6 +1429,11 @@ static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part) WARN_ON_ONCE(bdev->bd_holders); sync_blockdev(bdev); kill_bdev(bdev); + /* ->release can cause the old bdi to disappear, + * so must switch it out first + */ + bdev_inode_switch_bdi(bdev->bd_inode, + &default_backing_dev_info); } if (bdev->bd_contains == bdev) { if (disk->fops->release) @@ -1442,8 +1447,6 @@ static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part) disk_put_part(bdev->bd_part); bdev->bd_part = NULL; bdev->bd_disk = NULL; - bdev_inode_switch_bdi(bdev->bd_inode, - &default_backing_dev_info); if (bdev != bdev->bd_contains) victim = bdev->bd_contains; bdev->bd_contains = NULL; -- cgit v1.2.3 From 14c7cca780bd210564ae964f57a8bb807d0b3dbf Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: Btrfs: fix an oops when deleting snapshots We can reproduce this oops via the following steps: $ mkfs.btrfs /dev/sdb7 $ mount /dev/sdb7 /mnt/btrfs $ for ((i=0; i<3; i++)); do btrfs sub snap /mnt/btrfs /mnt/btrfs/s_$i; done $ rm -fr /mnt/btrfs/* $ rm -fr /mnt/btrfs/* then we'll get ------------[ cut here ]------------ kernel BUG at fs/btrfs/inode.c:2264! [...] Call Trace: [] btrfs_rmdir+0xf7/0x1b0 [btrfs] [] vfs_rmdir+0xa5/0xf0 [] do_rmdir+0x123/0x140 [] ? fput+0x197/0x260 [] ? audit_syscall_entry+0x1bf/0x1f0 [] sys_unlinkat+0x2d/0x40 [] system_call_fastpath+0x16/0x1b RIP [] btrfs_orphan_add+0x179/0x1a0 [btrfs] When it comes to btrfs_lookup_dentry, we may set a snapshot's inode->i_ino to BTRFS_EMPTY_SUBVOL_DIR_OBJECTID instead of BTRFS_FIRST_FREE_OBJECTID, while the snapshot's location.objectid remains unchanged. However, btrfs_ino() does not take this into account, and returns a wrong ino, and causes the oops. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/btrfs_inode.h | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index 502b9e988679..d9f99a16edd6 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -176,7 +176,11 @@ static inline u64 btrfs_ino(struct inode *inode) { u64 ino = BTRFS_I(inode)->location.objectid; - if (ino <= BTRFS_FIRST_FREE_OBJECTID) + /* + * !ino: btree_inode + * type == BTRFS_ROOT_ITEM_KEY: subvol dir + */ + if (!ino || BTRFS_I(inode)->location.type == BTRFS_ROOT_ITEM_KEY) ino = inode->i_ino; return ino; } -- cgit v1.2.3 From e0b6d65be57fb37ca67b04ce8964546a74d2125c Mon Sep 17 00:00:00 2001 From: Sergei Trofimovich Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: btrfs: fix warning in iput for bad-inode iput() shouldn't be called for inodes in I_NEW state. We need to mark inode as constructed first. WARNING: at fs/inode.c:1309 iput+0x20b/0x210() Call Trace: [] warn_slowpath_common+0x7a/0xb0 [] warn_slowpath_null+0x15/0x20 [] iput+0x20b/0x210 [] btrfs_iget+0x1eb/0x4a0 [] btrfs_run_defrag_inodes+0x136/0x210 [] cleaner_kthread+0x17f/0x1a0 [] ? sub_preempt_count+0x9d/0xd0 [] ? transaction_kthread+0x280/0x280 [] kthread+0x96/0xa0 [] kernel_thread_helper+0x4/0x10 [] ? kthread_worker_fn+0x190/0x190 [] ? gs_change+0xb/0xb Signed-off-by: Sergei Trofimovich CC: Konstantin Khlebnikov Tested-by: David Sterba CC: Josef Bacik CC: Chris Mason Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 34195f9fc6bb..edd45f709989 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3952,7 +3952,6 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, struct btrfs_root *root, int *new) { struct inode *inode; - int bad_inode = 0; inode = btrfs_iget_locked(s, location->objectid, root); if (!inode) @@ -3968,15 +3967,12 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, if (new) *new = 1; } else { - bad_inode = 1; + unlock_new_inode(inode); + iput(inode); + inode = ERR_PTR(-ESTALE); } } - if (bad_inode) { - iput(inode); - inode = ERR_PTR(-ESTALE); - } - return inode; } -- cgit v1.2.3 From ddf23b3fc6850bd4654d51ec9457fe7c77cde51e Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: Btrfs: skip locking if searching the commit root in csum lookup It's not enough to just search the commit root, since we could be cow'ing the very block we need to search through, which would mean that its locked and we'll still deadlock. So use path->skip_locking as well. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/file-item.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index b910694f61ed..a1cb7821becd 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -183,8 +183,10 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, * read from the commit root and sidestep a nasty deadlock * between reading the free space cache and updating the csum tree. */ - if (btrfs_is_free_space_inode(root, inode)) + if (btrfs_is_free_space_inode(root, inode)) { path->search_commit_root = 1; + path->skip_locking = 1; + } disk_bytenr = (u64)bio->bi_sector << 9; if (dio) -- cgit v1.2.3 From 65450aa645b1ef7ed74e41c34b28d53333744978 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: Btrfs: reset to appropriate block rsv after orphan operations While truncating free space cache, we forget to change trans->block_rsv back to the original one, but leave it with the orphan_block_rsv, and then with option inode_cache enable, it leads to countless warnings of btrfs_alloc_free_block and btrfs_orphan_commit_root: WARNING: at fs/btrfs/extent-tree.c:5711 btrfs_alloc_free_block+0x180/0x350 [btrfs]() ... WARNING: at fs/btrfs/inode.c:2193 btrfs_orphan_commit_root+0xb0/0xc0 [btrfs]() Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/free-space-cache.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 6a265b9f85f2..41ac927401d0 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -190,9 +190,11 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, struct btrfs_path *path, struct inode *inode) { + struct btrfs_block_rsv *rsv; loff_t oldsize; int ret = 0; + rsv = trans->block_rsv; trans->block_rsv = root->orphan_block_rsv; ret = btrfs_block_rsv_check(trans, root, root->orphan_block_rsv, @@ -210,6 +212,8 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, */ ret = btrfs_truncate_inode_items(trans, root, inode, 0, BTRFS_EXTENT_DATA_KEY); + + trans->block_rsv = rsv; if (ret) { WARN_ON(1); return ret; -- cgit v1.2.3 From 98c9942aca05fff198cd5ca629599cd193444809 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: Btrfs: fix misuse of trans block rsv At the beginning of create_pending_snapshot, trans->block_rsv is set to pending->block_rsv and is used for snapshot things, however, when it is done, we do not recover it as will. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/transaction.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 7dc36fab4afc..e24b7964a155 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -884,6 +884,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, struct btrfs_root *tree_root = fs_info->tree_root; struct btrfs_root *root = pending->root; struct btrfs_root *parent_root; + struct btrfs_block_rsv *rsv; struct inode *parent_inode; struct dentry *parent; struct dentry *dentry; @@ -895,6 +896,8 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, u64 objectid; u64 root_flags; + rsv = trans->block_rsv; + new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS); if (!new_root_item) { pending->error = -ENOMEM; @@ -1002,6 +1005,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_orphan_post_snapshot(trans, pending); fail: kfree(new_root_item); + trans->block_rsv = rsv; btrfs_block_rsv_release(root, &pending->block_rsv, (u64)-1); return 0; } -- cgit v1.2.3 From 5b397377e97d436fc2ed872fc53f85395bb984e0 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: Btrfs: fix unclosed transaction handle in btrfs_cont_expand The function - btrfs_cont_expand() forgot to close the transaction handle before it jump out the while loop. Fix it. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index edd45f709989..c257af2ce9cb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3510,15 +3510,19 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) err = btrfs_drop_extents(trans, inode, cur_offset, cur_offset + hole_size, &hint_byte, 1); - if (err) + if (err) { + btrfs_end_transaction(trans, root); break; + } err = btrfs_insert_file_extent(trans, root, btrfs_ino(inode), cur_offset, 0, 0, hole_size, 0, hole_size, 0, 0, 0); - if (err) + if (err) { + btrfs_end_transaction(trans, root); break; + } btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); -- cgit v1.2.3 From 0c1a98c81413e00a6c379d898e06a09350d31926 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Sun, 11 Sep 2011 10:52:24 -0400 Subject: Btrfs: fix the file extent gap when doing direct IO When we write some data to the place that is beyond the end of the file in direct I/O mode, a data hole will be created. And Btrfs should insert a file extent item that point to this hole into the fs tree. But unfortunately Btrfs forgets doing it. The following is a simple way to reproduce it: # mkfs.btrfs /dev/sdc2 # mount /dev/sdc2 /test4 # touch /test4/a # dd if=/dev/zero of=/test4/a seek=8 count=1 bs=4K oflag=direct conv=nocreat,notrunc # umount /test4 # btrfsck /dev/sdc2 root 5 inode 257 errors 100 Reported-by: Tsutomu Itoh Signed-off-by: Miao Xie Tested-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/file.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 15e5a1cd8764..98d95bb5f253 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1075,12 +1075,6 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, start_pos = pos & ~((u64)root->sectorsize - 1); last_pos = ((u64)index + num_pages) << PAGE_CACHE_SHIFT; - if (start_pos > inode->i_size) { - err = btrfs_cont_expand(inode, i_size_read(inode), start_pos); - if (err) - return err; - } - again: for (i = 0; i < num_pages; i++) { pages[i] = find_or_create_page(inode->i_mapping, index + i, @@ -1338,6 +1332,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; loff_t *ppos = &iocb->ki_pos; + u64 start_pos; ssize_t num_written = 0; ssize_t err = 0; size_t count, ocount; @@ -1386,6 +1381,15 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, file_update_time(file); BTRFS_I(inode)->sequence++; + start_pos = round_down(pos, root->sectorsize); + if (start_pos > i_size_read(inode)) { + err = btrfs_cont_expand(inode, i_size_read(inode), start_pos); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + } + if (unlikely(file->f_flags & O_DIRECT)) { num_written = __btrfs_direct_write(iocb, iov, nr_segs, pos, ppos, count, ocount); -- cgit v1.2.3 From a39f75214358d715efa21e2bccf5a709d8649144 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Sun, 11 Sep 2011 10:52:25 -0400 Subject: Btrfs: fix wrong nbytes information of the inode If we write some data into the data hole of the file(no preallocation for this hole), Btrfs will allocate some disk space, and update nbytes of the inode, but the other element--disk_i_size needn't be updated. At this condition, we must update inode metadata though disk_i_size is not changed(btrfs_ordered_update_i_size() return 1). # mkfs.btrfs /dev/sdb1 # mount /dev/sdb1 /mnt # touch /mnt/a # truncate -s 856002 /mnt/a # dd if=/dev/zero of=/mnt/a bs=4K count=1 conv=nocreat,notrunc # umount /mnt # btrfsck /dev/sdb1 root 5 inode 257 errors 400 found 32768 bytes used err is 1 Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index c257af2ce9cb..b94c0da3b43f 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1786,7 +1786,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) &ordered_extent->list); ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); - if (!ret) { + if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) { ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); } @@ -5788,7 +5788,7 @@ again: add_pending_csums(trans, inode, ordered->file_offset, &ordered->list); ret = btrfs_ordered_update_i_size(inode, 0, ordered); - if (!ret) + if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) btrfs_update_inode(trans, root, inode); ret = 0; out_unlock: -- cgit v1.2.3 From 4815053aba7f2304055745df820cd74a39fdaab2 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Sun, 11 Sep 2011 10:52:25 -0400 Subject: btrfs: xattr: fix attribute removal An attribute is not removed by 'setfattr -x attr file' and remains visible in attr list. This makes xfstests/062 pass again. Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/xattr.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index d733b9cfea34..69565e5fc6a0 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -116,6 +116,12 @@ static int do_setxattr(struct btrfs_trans_handle *trans, if (ret) goto out; btrfs_release_path(path); + + /* + * remove the attribute + */ + if (!value) + goto out; } again: @@ -158,6 +164,9 @@ out: return ret; } +/* + * @value: "" makes the attribute to empty, NULL removes it + */ int __btrfs_setxattr(struct btrfs_trans_handle *trans, struct inode *inode, const char *name, const void *value, size_t size, int flags) -- cgit v1.2.3 From d72c0842ff0e71342857723bb65f35b71f57b264 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 11 Sep 2011 10:52:25 -0400 Subject: Btrfs: calc file extent num_bytes correctly in file clone num_bytes should be 4096 not 12288. Signed-off-by: Li Zefan Signed-off-by: Chris Mason --- fs/btrfs/ioctl.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index b3d249d6eba7..028a4b8c12cd 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2333,14 +2333,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, if (type == BTRFS_FILE_EXTENT_REG || type == BTRFS_FILE_EXTENT_PREALLOC) { + /* + * a | --- range to clone ---| b + * | ------------- extent ------------- | + */ + + /* substract range b */ + if (key.offset + datal > off + len) + datal = off + len - key.offset; + + /* substract range a */ if (off > key.offset) { datao += off - key.offset; datal -= off - key.offset; } - if (key.offset + datal > off + len) - datal = off + len - key.offset; - ret = btrfs_drop_extents(trans, inode, new_key.offset, new_key.offset + datal, -- cgit v1.2.3 From d525e8ab022cb000e6e31a515ba8c3cf0d9c6130 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Sun, 11 Sep 2011 10:52:25 -0400 Subject: Btrfs: add dummy extent if dst offset excceeds file end in You can see there's no file extent with range [0, 4096]. Check this by btrfsck: # btrfsck /dev/sda7 root 5 inode 258 errors 100 ... Signed-off-by: Li Zefan Signed-off-by: Chris Mason --- fs/btrfs/ioctl.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 028a4b8c12cd..63b4de1626d2 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2228,6 +2228,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, !IS_ALIGNED(destoff, bs)) goto out_unlock; + if (destoff > inode->i_size) { + ret = btrfs_cont_expand(inode, inode->i_size, destoff); + if (ret) + goto out_unlock; + } + /* do any pending delalloc/csum calc on src, one way or another, and lock file content */ while (1) { -- cgit v1.2.3 From 24114504c4d585ec4aae7a2b2acb81bf741f8c8a Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Mon, 12 Sep 2011 09:31:49 +0200 Subject: fuse: fix flock breakage Commit 37fb3a30b4 ("fuse: fix flock") added in 3.1-rc4 caused flock() to fail with ENOSYS with the kernel ABI version 7.16 or earlier. Fix by falling back to testing FUSE_POSIX_LOCKS for ABI versions 7.16 and earlier. Reported-by: Martin Ziegler Signed-off-by: Miklos Szeredi Tested-by: Martin Ziegler Signed-off-by: Linus Torvalds --- fs/fuse/inode.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 12b502929da9..add96f6ffda5 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -812,6 +812,9 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req) if (arg->minor >= 17) { if (!(arg->flags & FUSE_FLOCK_LOCKS)) fc->no_flock = 1; + } else { + if (!(arg->flags & FUSE_POSIX_LOCKS)) + fc->no_flock = 1; } if (arg->flags & FUSE_ATOMIC_O_TRUNC) fc->atomic_o_trunc = 1; -- cgit v1.2.3 From 5dfcc87fd79dfb96ed155b524337dbd0da4f5993 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Mon, 12 Sep 2011 09:38:03 +0200 Subject: fuse: fix memory leak kmemleak is reporting that 32 bytes are being leaked by FUSE: unreferenced object 0xe373b270 (size 32): comm "fusermount", pid 1207, jiffies 4294707026 (age 2675.187s) hex dump (first 32 bytes): 01 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [] kmemleak_alloc+0x27/0x50 [] kmem_cache_alloc+0xc5/0x180 [] fuse_alloc_forget+0x1e/0x20 [] fuse_alloc_inode+0xb0/0xd0 [] alloc_inode+0x1c/0x80 [] iget5_locked+0x8f/0x1a0 [] fuse_iget+0x72/0x1a0 [] fuse_get_root_inode+0x8a/0x90 [] fuse_fill_super+0x3ef/0x590 [] mount_nodev+0x3f/0x90 [] fuse_mount+0x15/0x20 [] mount_fs+0x1c/0xc0 [] vfs_kern_mount+0x41/0x90 [] do_kern_mount+0x39/0xd0 [] do_mount+0x2e5/0x660 [] sys_mount+0x66/0xa0 This leak report is consistent and happens once per boot on 3.1.0-rc5-dirty. This happens if a FORGET request is queued after the fuse device was released. Reported-by: Sitsofe Wheeler Signed-off-by: Miklos Szeredi Tested-by: Sitsofe Wheeler Signed-off-by: Linus Torvalds --- fs/fuse/dev.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 168a80f7f12b..5cb8614508c3 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -258,10 +258,14 @@ void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forget, forget->forget_one.nlookup = nlookup; spin_lock(&fc->lock); - fc->forget_list_tail->next = forget; - fc->forget_list_tail = forget; - wake_up(&fc->waitq); - kill_fasync(&fc->fasync, SIGIO, POLL_IN); + if (fc->connected) { + fc->forget_list_tail->next = forget; + fc->forget_list_tail = forget; + wake_up(&fc->waitq); + kill_fasync(&fc->fasync, SIGIO, POLL_IN); + } else { + kfree(forget); + } spin_unlock(&fc->lock); } -- cgit v1.2.3