kernel/git/stable/linux.git/fs/logfs/inode.c, branch linux-3.6.y

logfs: maintain the ordering of meta-inode destruction

2012-07-23T04:05:52+00:00

LogFS does not use a specialized area to maintain the inodes. The inodes information is kept in a specialized file called inode file. Similarly, the segment information is kept in a segment file. Since the segment file also has an inode which is kept in the inode file, the inode for segment file must be evicted before the inode for inode file. The change fixes the following BUG during unmount Pid: 2057, comm: umount Not tainted 3.5.0-rc6+ #25 Bochs Bochs RIP: 0010:[] [] move_page_to_btree+0x32/0x1f0 [logfs] Process umount (pid: 2057, threadinfo ...) Call Trace: [] ? find_get_pages+0x2a/0x180 [] logfs_invalidatepage+0x85/0x90 [logfs] [] truncate_inode_page+0xb1/0xd0 [] truncate_inode_pages_range+0x15f/0x490 [] ? printk+0x78/0x7a [] truncate_inode_pages+0x15/0x20 [] logfs_evict_inode+0x6c/0x190 [logfs] [] ? _raw_spin_unlock+0x2b/0x40 [] evict+0xa7/0x1b0 [] dispose_list+0x3e/0x60 [] evict_inodes+0xf4/0x110 [] generic_shutdown_super+0x53/0xf0 [] logfs_kill_sb+0x52/0xf0 [logfs] [] deactivate_locked_super+0x45/0x80 [] deactivate_super+0x4a/0x70 [] mntput_no_expire+0xde/0x140 [] sys_umount+0x6f/0x3a0 [] system_call_fastpath+0x16/0x1b ---[ end trace 45f7752082cefafd ]--- Signed-off-by: Prasad Joshi

logfs: destroy the reserved inodes while unmounting

2012-04-02T03:50:33+00:00

We were assuming that the evict_inode() would never be called on reserved inodes. However, (after the commit 8e22c1a4e logfs: get rid of magical inodes) while unmounting the file system, in put_super, we call iput() on all of the reserved inodes. The following simple test used to cause a kernel panic on LogFS: 1. Mount a LogFS file system on /mnt 2. Create a file $ touch /mnt/a 3. Try to unmount the FS $ umount /mnt The simple fix would be to drop the assumption and properly destroy the reserved inodes. Signed-off-by: Prasad Joshi

Merge tag 'for-linus' of git://github.com/prasad-joshi/logfs_upstream

2012-01-31T17:23:59+00:00

There are few important bug fixes for LogFS * tag 'for-linus' of git://github.com/prasad-joshi/logfs_upstream: Logfs: Allow NULL block_isbad() methods logfs: Grow inode in delete path logfs: Free areas before calling generic_shutdown_super() logfs: remove useless BUG_ON MAINTAINERS: Add Prasad Joshi in LogFS maintiners logfs: Propagate page parameter to __logfs_write_inode logfs: set superblock shutdown flag after generic sb shutdown logfs: take write mutex lock during fsync and sync logfs: Prevent memory corruption logfs: update page reference count for pined pages Fix up conflict in fs/logfs/dev_mtd.c due to semantic change in what "mtd->block_isbad" means in commit f2933e86ad93: "Logfs: Allow NULL block_isbad() methods" clashing with the abstraction changes in the commits 7086c19d0742: "mtd: introduce mtd_block_isbad interface" and d58b27ed58a3: "logfs: do not use 'mtd->block_isbad' directly". This resolution takes the semantics from commit f2933e86ad93, and just makes mtd_block_isbad() return zero (false) if the 'block_isbad' function is NULL. But that also means that now "mtd_can_have_bb()" always returns 0. Now, "mtd_block_markbad()" will obviously return an error if the low-level driver doesn't support bad blocks, so this is somewhat non-symmetric, but it actually makes sense if a NULL "block_isbad" function is considered to mean "I assume that all my blocks are always good".

logfs: Propagate page parameter to __logfs_write_inode

2012-01-28T06:08:25+00:00

During GC LogFS has to rewrite each valid block to a separate segment. Rewrite operation reads data from an old segment and writes it to a newly allocated segment. Since every write operation changes data block pointers maintained in inode, inode should also be rewritten. In GC path to avoid AB-BA deadlock LogFS marks a page with PG_pre_locked in addition to locking the page (PG_locked). The page lock is ignored iff the page is pre-locked. LogFS uses a special file called segment file. The segment file maintains an 8 bytes entry for every segment. It keeps track of erase count, level etc. for every segment. Bad things happen with a segment belonging to the segment file is GCed ------------[ cut here ]------------ kernel BUG at /home/prasad/logfs/readwrite.c:297! invalid opcode: 0000 [#1] SMP Modules linked in: logfs joydev usbhid hid psmouse e1000 i2c_piix4 serio_raw [last unloaded: logfs] Pid: 20161, comm: mount Not tainted 3.1.0-rc3+ #3 innotek GmbH VirtualBox EIP: 0060:[] EFLAGS: 00010292 CPU: 0 EIP is at logfs_lock_write_page+0x6a/0x70 [logfs] EAX: 00000027 EBX: f73f5b20 ECX: c16007c8 EDX: 00000094 ESI: 00000000 EDI: e59be6e4 EBP: c7337b28 ESP: c7337b18 DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 Process mount (pid: 20161, ti=c7336000 task=eb323f70 task.ti=c7336000) Stack: f8099a3d c7337b24 f73f5b20 00001002 c7337b50 f8091f6d f8099a4d f80994e4 00000003 00000000 c7337b68 00000000 c67e4400 00001000 c7337b80 f80935e5 00000000 00000000 00000000 00000000 e1fcf000 0000000f e59be618 c70bf900 Call Trace: [] logfs_get_write_page.clone.16+0xdd/0x100 [logfs] [] logfs_mod_segment_entry+0x55/0x110 [logfs] [] logfs_get_segment_entry+0x1d/0x20 [logfs] [] ? logfs_cleanup_journal+0x50/0x50 [logfs] [] ostore_get_erase_count+0x1b/0x40 [logfs] [] logfs_open_area+0xc8/0x150 [logfs] [] ? kmemleak_alloc+0x2c/0x60 [] __logfs_segment_write.clone.16+0x4e/0x1b0 [logfs] [] ? mempool_kmalloc+0x13/0x20 [] ? mempool_kmalloc+0x13/0x20 [] logfs_segment_write+0x17f/0x1d0 [logfs] [] logfs_write_i0+0x11c/0x180 [logfs] [] logfs_write_direct+0x45/0x90 [logfs] [] __logfs_write_buf+0xbd/0xf0 [logfs] [] ? kmap_atomic_prot+0x4e/0xe0 [] logfs_write_buf+0x3b/0x60 [logfs] [] __logfs_write_inode+0xa9/0x110 [logfs] [] logfs_rewrite_block+0xc0/0x110 [logfs] [] ? get_mapping_page+0x10/0x60 [logfs] [] ? logfs_load_object_aliases+0x2e0/0x2f0 [logfs] [] logfs_gc_segment+0x2ad/0x310 [logfs] [] __logfs_gc_once+0x4a/0x80 [logfs] [] logfs_gc_pass+0x683/0x6a0 [logfs] [] logfs_mount+0x5a9/0x680 [logfs] [] mount_fs+0x21/0xd0 [] ? __alloc_percpu+0xf/0x20 [] ? alloc_vfsmnt+0xb1/0x130 [] vfs_kern_mount+0x4b/0xa0 [] do_kern_mount+0x3e/0xe0 [] do_mount+0x34d/0x670 [] ? strndup_user+0x49/0x70 [] sys_mount+0x6b/0xa0 [] syscall_call+0x7/0xb Code: f8 e8 8b 93 39 c9 8b 45 f8 3e 0f ba 28 00 19 d2 85 d2 74 ca eb d0 0f 0b 8d 45 fc 89 44 24 04 c7 04 24 3d 9a 09 f8 e8 09 92 39 c9 <0f> 0b 8d 74 26 00 55 89 e5 3e 8d 74 26 00 8b 10 80 e6 01 74 09 EIP: [] logfs_lock_write_page+0x6a/0x70 [logfs] SS:ESP 0068:c7337b18 ---[ end trace 96e67d5b3aa3d6ca ]--- The patch passes locked page to __logfs_write_inode. It calls function logfs_get_wblocks() to pre-lock the page. This ensures any further attempts to lock the page are ignored (esp from get_erase_count). Acked-by: Joern Engel Signed-off-by: Prasad Joshi

logfs: take write mutex lock during fsync and sync

2012-01-28T06:06:06+00:00

LogFS uses super->s_write_mutex while writing data to disk. Taking the same mutex lock in sync and fsync code path solves the following BUG: ------------[ cut here ]------------ kernel BUG at /home/prasad/logfs/dev_bdev.c:134! Pid: 2387, comm: flush-253:16 Not tainted 3.0.0+ #4 Bochs Bochs RIP: 0010:[] [] bdev_writeseg+0x25d/0x270 [logfs] Call Trace: [] logfs_open_area+0x91/0x150 [logfs] [] ? find_level.clone.9+0x62/0x100 [] __logfs_segment_write.clone.20+0x5c/0x190 [logfs] [] ? mempool_kmalloc+0x15/0x20 [] ? mempool_alloc+0x53/0x130 [] logfs_segment_write+0x1d4/0x230 [logfs] [] logfs_write_i0+0x12e/0x190 [logfs] [] __logfs_write_rec+0x140/0x220 [logfs] [] logfs_write_rec+0x64/0xd0 [logfs] [] __logfs_write_buf+0x106/0x110 [logfs] [] logfs_write_buf+0x4e/0x80 [logfs] [] __logfs_writepage+0x23/0x80 [logfs] [] logfs_writepage+0xdc/0x110 [logfs] [] __writepage+0x17/0x40 [] write_cache_pages+0x208/0x4f0 [] ? set_page_dirty+0x70/0x70 [] generic_writepages+0x4a/0x70 [] do_writepages+0x21/0x40 [] writeback_single_inode+0x101/0x250 [] writeback_sb_inodes+0xed/0x1c0 [] writeback_inodes_wb+0x7b/0x1e0 [] wb_writeback+0x4c3/0x530 [] ? sub_preempt_count+0x9d/0xd0 [] wb_do_writeback+0xdb/0x290 [] ? sub_preempt_count+0x9d/0xd0 [] ? _raw_spin_unlock_irqrestore+0x18/0x40 [] ? del_timer+0x8a/0x120 [] bdi_writeback_thread+0x8c/0x2e0 [] ? wb_do_writeback+0x290/0x290 [] kthread+0x96/0xa0 [] kernel_thread_helper+0x4/0x10 [] ? kthread_worker_fn+0x190/0x190 [] ? gs_change+0xb/0xb RIP [] bdev_writeseg+0x25d/0x270 [logfs] ---[ end trace 0211ad60a57657c4 ]--- Reviewed-by: Joern Engel Signed-off-by: Prasad Joshi

logfs: propagate umode_t

2012-01-04T03:55:06+00:00

Signed-off-by: Al Viro

vfs: fix the stupidity with i_dentry in inode destructors

2012-01-04T03:52:40+00:00

Seeing that just about every destructor got that INIT_LIST_HEAD() copied into it, there is no point whatsoever keeping this INIT_LIST_HEAD in inode_init_once(); the cost of taking it into inode_init_always() will be negligible for pipes and sockets and negative for everything else. Not to mention the removal of boilerplate code from ->destroy_inode() instances... Signed-off-by: Al Viro

filesystems: add missing nlink wrappers

2011-11-02T11:53:43+00:00

Replace direct i_nlink updates with the respective updater function (inc_nlink, drop_nlink, clear_nlink, inode_dec_link_count). Signed-off-by: Miklos Szeredi

logfs: remove unnecessary nlink setting

2011-11-02T11:53:42+00:00

alloc_inode() initializes i_nlink to 1. Remove unnecessary re-initialization. Signed-off-by: Miklos Szeredi CC: Joern Engel CC: Prasad Joshi Signed-off-by: Christoph Hellwig

fs: remove inode_lock from iput_final and prune_icache

2011-03-25T01:16:32+00:00

Now that inode state changes are protected by the inode->i_lock and the inode LRU manipulations by the inode_lru_lock, we can remove the inode_lock from prune_icache and the initial part of iput_final(). instead of using the inode_lock to protect the inode during iput_final, use the inode->i_lock instead. This protects the inode against new references being taken while we change the inode state to I_FREEING, as well as preventing prune_icache from grabbing the inode while we are manipulating it. Hence we no longer need the inode_lock in iput_final prior to setting I_FREEING on the inode. For prune_icache, we no longer need the inode_lock to protect the LRU list, and the inodes themselves are protected against freeing races by the inode->i_lock. Hence we can lift the inode_lock from prune_icache as well. Signed-off-by: Dave Chinner Signed-off-by: Al Viro