reiserfs locking (v2)
Hello,
I searched lkml and found the following report (titled reiserfs locking):
http://lkml.org/lkml/2010/4/15/389 (Fri, 16 Apr 2010)
I can see this backtrace while configuring emacs:
[ 6136.579013]
[ 6136.579015] =======================================================
[ 6136.579021] [ INFO: possible circular locking dependency detected ]
[ 6136.579025] 2.6.35-rc3-dbg-git3-00350-g94d119f #61
[ 6136.579027] -------------------------------------------------------
[ 6136.579031] conftest/13950 is trying to acquire lock:
[ 6136.579034] (&sb->s_type->i_mutex_key#10){+.+.+.}, at: [<c10f1cf6>] reiserfs_file_release+0x11d/0x344
[ 6136.579050]
[ 6136.579051] but task is already holding lock:
[ 6136.579054] (&mm->mmap_sem){++++++}, at: [<c1091b87>] sys_munmap+0x26/0x42
[ 6136.579064]
[ 6136.579065] which lock already depends on the new lock.
[ 6136.579066]
[ 6136.579069]
[ 6136.579070] the existing dependency chain (in reverse order) is:
[ 6136.579073]
[ 6136.579074] -> #1 (&mm->mmap_sem){++++++}:
[ 6136.579081] [<c104f53a>] lock_acquire+0x59/0x70
[ 6136.579089] [<c108cf78>] might_fault+0x53/0x70
[ 6136.579094] [<c11853b8>] copy_to_user+0x30/0x48
[ 6136.579101] [<c10afaf9>] filldir64+0x95/0xc9
[ 6136.579106] [<c10f2504>] reiserfs_readdir_dentry+0x35d/0x4d9
[ 6136.579112] [<c10f2692>] reiserfs_readdir+0x12/0x17
[ 6136.579117] [<c10afd17>] vfs_readdir+0x6d/0x92
[ 6136.579122] [<c10afe91>] sys_getdents64+0x63/0xa2
[ 6136.579127] [<c10027d3>] sysenter_do_call+0x12/0x32
[ 6136.579134]
[ 6136.579135] -> #0 (&sb->s_type->i_mutex_key#10){+.+.+.}:
[ 6136.579142] [<c104ef30>] __lock_acquire+0x96d/0xbe1
[ 6136.579148] [<c104f53a>] lock_acquire+0x59/0x70
[ 6136.579153] [<c12c5614>] __mutex_lock_common+0x39/0x36b
[ 6136.579160] [<c12c5980>] mutex_lock_nested+0x12/0x15
[ 6136.579165] [<c10f1cf6>] reiserfs_file_release+0x11d/0x344
[ 6136.579171] [<c10a580d>] fput+0xe0/0x16a
[ 6136.579177] [<c1090ca6>] remove_vma+0x28/0x47
[ 6136.579182] [<c1091a68>] do_munmap+0x1e8/0x200
[ 6136.579188] [<c1091b93>] sys_munmap+0x32/0x42
[ 6136.579193] [<c10027d3>] sysenter_do_call+0x12/0x32
[ 6136.579198]
[ 6136.579199] other info that might help us debug this:
[ 6136.579200]
[ 6136.579204] 1 lock held by conftest/13950:
[ 6136.579207] #0: (&mm->mmap_sem){++++++}, at: [<c1091b87>] sys_munmap+0x26/0x42
[ 6136.579216]
[ 6136.579217] stack backtrace:
[ 6136.579221] Pid: 13950, comm: conftest Not tainted 2.6.35-rc3-dbg-git3-00350-g94d119f #61
[ 6136.579225] Call Trace:
[ 6136.579230] [<c12c4893>] ? printk+0xf/0x11
[ 6136.579235] [<c104dbdd>] print_circular_bug+0x8a/0x96
[ 6136.579241] [<c104ef30>] __lock_acquire+0x96d/0xbe1
[ 6136.579247] [<c104e436>] ? mark_lock+0x26/0x1b3
[ 6136.579254] [<c104f53a>] lock_acquire+0x59/0x70
[ 6136.579259] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344
[ 6136.579265] [<c12c5614>] __mutex_lock_common+0x39/0x36b
[ 6136.579270] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344
[ 6136.579276] [<c1084375>] ? release_pages+0x55/0x116
[ 6136.579282] [<c12c5980>] mutex_lock_nested+0x12/0x15
[ 6136.579287] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344
[ 6136.579293] [<c10f1cf6>] reiserfs_file_release+0x11d/0x344
[ 6136.579299] [<c10a57bd>] ? fput+0x90/0x16a
[ 6136.579304] [<c10a580d>] fput+0xe0/0x16a
[ 6136.579309] [<c1090ca6>] remove_vma+0x28/0x47
[ 6136.579314] [<c1091819>] ? arch_unmap_area_topdown+0x0/0x18
[ 6136.579319] [<c1091a68>] do_munmap+0x1e8/0x200
[ 6136.579325] [<c1091b93>] sys_munmap+0x32/0x42
[ 6136.579330] [<c10027d3>] sysenter_do_call+0x12/0x32So, below is the simple code which produces similar result:
[ 573.405720]
[ 573.405722] =======================================================
[ 573.405728] [ INFO: possible circular locking dependency detected ]
[ 573.405732] 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65
[ 573.405735] -------------------------------------------------------
[ 573.405739] a.out/7287 is trying to acquire lock:
[ 573.405742] (&sb->s_type->i_mutex_key#10){+.+.+.}, at: [<c10f1d4e>] reiserfs_file_release+0x11d/0x344
[ 573.405758]
[ 573.405759] but task is already holding lock:
[ 573.405762] (&mm->mmap_sem){++++++}, at: [<c1092937>] sys_mmap_pgoff+0xa4/0xe7
[ 573.405772]
[ 573.405773] which lock already depends on the new lock.
[ 573.405774]
[ 573.405777]
[ 573.405778] the existing dependency chain (in reverse order) is:
[ 573.405781]
[ 573.405782] -> #1 (&mm->mmap_sem){++++++}:
[ 573.405789] [<c104f566>] lock_acquire+0x59/0x70
[ 573.405797] [<c108cf70>] might_fault+0x53/0x70
[ 573.405803] [<c1185418>] copy_to_user+0x30/0x48
[ 573.405809] [<c10afaf9>] filldir64+0x95/0xc9
[ 573.405815] [<c10f255c>] reiserfs_readdir_dentry+0x35d/0x4d9
[ 573.405821] [<c10f26ea>] reiserfs_readdir+0x12/0x17
[ 573.405827] [<c10afd17>] vfs_readdir+0x6d/0x92
[ 573.405831] [<c10afe91>] sys_getdents64+0x63/0xa2
[ 573.405836] [<c10027d3>] sysenter_do_call+0x12/0x32
[ 573.405843]
[ 573.405843] -> #0 (&sb->s_type->i_mutex_key#10){+.+.+.}:
[ 573.405851] [<c104ef5c>] __lock_acquire+0x96d/0xbe1
[ 573.405857] [<c104f566>] lock_acquire+0x59/0x70
[ 573.405862] [<c12c5674>] __mutex_lock_common+0x39/0x36b
[ 573.405869] [<c12c59e0>] mutex_lock_nested+0x12/0x15
[ 573.405874] [<c10f1d4e>] reiserfs_file_release+0x11d/0x344
[ 573.405880] [<c10a5805>] fput+0xe0/0x16a
[ 573.405886] [<c1090c9e>] remove_vma+0x28/0x47
[ 573.405892] [<c1091a60>] do_munmap+0x1e8/0x200
[ 573.405897] [<c109230a>] mmap_region+0x6b/0x372
[ 573.405902] [<c109284d>] do_mmap_pgoff+0x23c/0x282
[ 573.405908] [<c1092950>] sys_mmap_pgoff+0xbd/0xe7
[ 573.405913] [<c10027d3>] sysenter_do_call+0x12/0x32
[ 573.405919]
[ 573.405920] other info that might help us debug this:
[ 573.405921]
[ 573.405925] 1 lock held by a.out/7287:
[ 573.405928] #0: (&mm->mmap_sem){++++++}, at: [<c1092937>] sys_mmap_pgoff+0xa4/0xe7
[ 573.405937]
[ 573.405938] stack backtrace:
[ 573.405942] Pid: 7287, comm: a.out Not tainted 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65
[ 573.405946] Call Trace:
[ 573.405951] [<c12c48f3>] ? printk+0xf/0x11
[ 573.405957] [<c104dc09>] print_circular_bug+0x8a/0x96
[ 573.405962] [<c104ef5c>] __lock_acquire+0x96d/0xbe1
[ 573.405969] [<c104e462>] ? mark_lock+0x26/0x1b3
[ 573.405975] [<c104f566>] lock_acquire+0x59/0x70
[ 573.405980] [<c10f1d4e>] ? reiserfs_file_release+0x11d/0x344
[ 573.405986] [<c12c5674>] __mutex_lock_common+0x39/0x36b
[ 573.405991] [<c10f1d4e>] ? reiserfs_file_release+0x11d/0x344
[ 573.405997] [<c12c59e0>] mutex_lock_nested+0x12/0x15
[ 573.406003] [<c10f1d4e>] ? reiserfs_file_release+0x11d/0x344
[ 573.406008] [<c10f1d4e>] reiserfs_file_release+0x11d/0x344
[ 573.406014] [<c10a57b5>] ? fput+0x90/0x16a
[ 573.406019] [<c10a5805>] fput+0xe0/0x16a
[ 573.406024] [<c1090c9e>] remove_vma+0x28/0x47
[ 573.406030] [<c1091811>] ? arch_unmap_area_topdown+0x0/0x18
[ 573.406035] [<c1091a60>] do_munmap+0x1e8/0x200
[ 573.406040] [<c109230a>] mmap_region+0x6b/0x372
[ 573.406046] [<c109284d>] do_mmap_pgoff+0x23c/0x282
[ 573.406052] [<c1092950>] sys_mmap_pgoff+0xbd/0xe7
[ 573.406058] [<c10027d3>] sysenter_do_call+0x12/0x32As well as this:
[ 202.300464] REISERFS error (device sda9): vs-2100 add_save_link: search_by_key ([-1 7812832 0x1 IND]) returned 1 [ 202.300473] REISERFS (device sda9): Remounting filesystem read-only [ 202.301603] ------------[ cut here ]------------ [ 202.301615] WARNING: at fs/reiserfs/journal.c:3436 journal_end+0x5b/0xaf() [ 202.301619] Hardware name: F3JC [ 202.301622] Modules linked in: snd_seq_dummy snd_seq_oss
snd_seq_midi_event snd_seq snd_seq_device snd_hda_codec_si3054 snd_hda_codec_realtek snd_hwdep snd_pcm_oss snd_mixer_oss asus_laptop sparse_keymap sdhci_pci sdhci snd_hda_intel mmc_core led_class snd_hda_codec rng_core snd_pcm snd_timer snd_page_alloc snd i2c_i801 soundcore psmouse sg evdev serio_raw r8169 mii usbhid hid uhci_hcd ehci_hcd sr_mod usbcore cdrom sd_mod ata_piix
[ 202.301689] Pid: 5055, comm: a.out Not tainted 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65 [ 202.301693] Call Trace: [ 202.301701] [<c102e3a2>] warn_slowpath_common+0x65/0x7a [ 202.301707] [<c1102d95>] ? journal_end+0x5b/0xaf [ 202.301712] [<c102e3c6>] warn_slowpath_null+0xf/0x13 [ 202.301718] [<c1102d95>] journal_end+0x5b/0xaf [ 202.301725] [<c10eebaa>] reiserfs_truncate_file+0x19f/0x233 [ 202.301733] [<c10f1ba1>] reiserfs_vfs_truncate_file+0xd/0xf [ 202.301738] [<c1084883>] vmtruncate+0x23/0x29 [ 202.301745] [<c10b4ad4>] inode_setattr+0x47/0x68 [ 202.301751] [<c10f1b3f>] reiserfs_setattr+0x242/0x297 [ 202.301758] [<c12c5d05>] ? down_write+0x22/0x2a [ 202.301764] [<c10b4d6f>] notify_change+0x15c/0x26b [ 202.301770] [<c10a35c7>] do_truncate+0x64/0x7d [ 202.301776] [<c12c69b0>] ? _raw_spin_unlock+0x33/0x3f [ 202.301783] [<c10ad892>] do_last+0x450/0x459 [ 202.301789] [<c10ada5b>] do_filp_open+0x1c0/0x41a [ 202.301798] [<c1029081>] ? get_parent_ip+0xb/0x31 [ 202.301804] [<c1029123>] ? sub_preempt_count+0x7c/0x89 [ 202.301810] [<c10b5746>] ? alloc_fd+0xb4/0xbf [ 202.301816] [<c10a3f46>] do_sys_open+0x48/0xdf [ 202.301821] [<c10a3ffb>] sys_open+0x1e/0x26 [ 202.301827] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 202.301833] ---[ end trace c4e3312bdadd2dc5 ]--- [ 202.301864] REISERFS warning (device sda9): clm-6006 reiserfs_dirty_inode: writing inode 7812832 on readonly FS [ 202.302011] REISERFS warning (device sda9): clm-6006 reiserfs_dirty_inode: writing inode 7812832 on readonly FS [ 202.302396] ------------[ cut here ]------------ [ 202.302403] WARNING: at fs/reiserfs/journal.c:3436 journal_end+0x5b/0xaf() [ 202.302406] Hardware name: F3JC [ 202.302409] Modules linked in: snd_seq_dummy snd_seq_oss
snd_seq_midi_event snd_seq snd_seq_device snd_hda_codec_si3054 snd_hda_codec_realtek snd_hwdep snd_pcm_oss snd_mixer_oss asus_laptop sparse_keymap sdhci_pci sdhci snd_hda_intel mmc_core led_class snd_hda_codec rng_core snd_pcm snd_timer snd_page_alloc snd i2c_i801 soundcore psmouse sg evdev serio_raw r8169 mii usbhid hid uhci_hcd ehci_hcd sr_mod usbcore cdrom sd_mod ata_piix
[ 202.302466] Pid: 4951, comm: a.out Tainted: G W 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65 [ 202.302470] Call Trace: [ 202.302476] [<c102e3a2>] warn_slowpath_common+0x65/0x7a [ 202.302481] [<c1102d95>] ? journal_end+0x5b/0xaf [ 202.302487] [<c102e3c6>] warn_slowpath_null+0xf/0x13 [ 202.302492] [<c1102d95>] journal_end+0x5b/0xaf [ 202.302497] [<c10fd687>] ? reiserfs_delete_object+0x29/0x60 [ 202.302503] [<c10ee1ee>] reiserfs_delete_inode+0xc4/0x100 [ 202.302510] [<c10b4265>] generic_delete_inode+0x72/0xc2 [ 202.302515] [<c10b42c6>] generic_drop_inode+0x11/0x4c [ 202.302520] [<c10b37ff>] iput+0x49/0x4c [ 202.302525] [<c10b165e>] dentry_iput+0x9a/0xb3 [ 202.302530] [<c10b16a4>] d_kill+0x2d/0x47 [ 202.302534] [<c10b1c94>] dput+0xdf/0xe8 [ 202.302540] [<c10a5872>] fput+0x14d/0x16a [ 202.302546] [<c1090c9e>] remove_vma+0x28/0x47 [ 202.302551] [<c1091c29>] exit_mmap+0x8e/0xa6 [ 202.302556] [<c102c4b3>] mmput+0x24/0xab [ 202.302561] [<c102fd83>] exit_mm+0xe8/0xf0 [ 202.302567] [<c1031295>] do_exit+0x1b7/0x5cb [ 202.302573] [<c10434a6>] ? up_read+0x1d/0x23 [ 202.302578] [<c10318b3>] do_group_exit+0x60/0x83 [ 202.302584] [<c10318e9>] sys_exit_group+0x13/0x17 [ 202.302589] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 202.302594] ---[ end trace c4e3312bdadd2dc6 ]---
==== code ====
conftest.c
/*
2010, Sergey Senozhatsky. GPLv2
Description:
We have several PIDs working with conftest.mmap. Actually this is (seems) what hapenning during emacs configure.
traced emacs configure:
vfork() ...
====
... PID 5446
5446 open("conftest.mmap", O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0600) = 3 write(3, "g\306isQ\377J\354)\315\272\253\362\373\343F|\302T\370\33\350\347\215vZ.c3\237\311\232"..., 4096) = 4096
close(3) = 0open("conftest.txt", O_RDWR|O_CREAT|O_TRUNC|O_LARGEFILE, 0600) = 3write(3, "\0", 1) = 1mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = 0xb78a8000
close(3) = 0munmap(0xb78a8000, 4096) = 0open("conftest.mmap", O_RDWR|O_LARGEFILE) = 3
mmap2(0xb78a8000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED,
3, 0) = 0xb78a8000
read(3,
"*****"...,
4096) = 4096
close(3) = 0
open(".", O_RDONLY|O_LARGEFILE) = 3
close(3) = 0fstatat64(AT_FDCWD, "conftest.mmap", {st_mode=S_IFREG|0600, st_size=4096, ...}, AT_SYMLINK_NOFOLLOW) = 0 unlinkat(AT_FDCWD, "conftest.mmap", 0) = 0
====
... PID 5449
5449 fstatat64(AT_FDCWD, "conftest.mmap", {st_mode=S_IFREG|0600, st_size=4096, ...}, AT_SYMLINK_NOFOLLOW) = 0
5449 unlinkat(AT_FDCWD, "conftest.mmap", 0) = 0
5449 execve("/bin/rm", ["rm", "-f", "conftest.mmap", "conftest.txt"]..
*/
/*
The code below produces:
[ 46.727489] =======================================================
[ 46.727495] [ INFO: possible circular locking dependency detected ]
[ 46.727499] 2.6.35-rc3-dbg-git5-00446-g36336bc-dirty #64
[ 46.727503] -------------------------------------------------------
[ 46.727506] a.out/5840 is trying to acquire lock:
[ 46.727510] (&sb->s_type->i_mutex_key#10){+.+.+.}, at: [<c10f1d5c>] reiserfs_file_release+0x12b/0x367
[ 46.727526]
[ 46.727527] but task is already holding lock:
[ 46.727530] (&mm->mmap_sem){++++++}, at: [<c1092937>] sys_mmap_pgoff+0xa4/0xe7
[ 46.727540]
[ 46.727541] which lock already depends on the new lock.
[ 46.727543]
[ 46.727546]
[ 46.727546] the existing dependency chain (in reverse order) is:
[ 46.727550]
[ 46.727551] -> #1 (&mm->mmap_sem){++++++}:
[ 46.727557] [<c104f566>] lock_acquire+0x59/0x70
[ 46.727565] [<c108cf70>] might_fault+0x53/0x70
[ 46.727571] [<c1185438>] copy_to_user+0x30/0x48
[ 46.727578] [<c10afaf9>] filldir64+0x95/0xc9
[ 46.727584] [<c10f257c>] reiserfs_readdir_dentry+0x35d/0x4d9
[ 46.727590] [<c10f270a>] reiserfs_readdir+0x12/0x17
[ 46.727596] [<c10afd17>] vfs_readdir+0x6d/0x92
[ 46.727600] [<c10afe91>] sys_getdents64+0x63/0xa2
[ 46.727606] [<c10027d3>] sysenter_do_call+0x12/0x32
[ 46.727612]
[ 46.727613] -> #0 (&sb->s_type->i_mutex_key#10){+.+.+.}:
[ 46.727621] [<c104ef5c>] __lock_acquire+0x96d/0xbe1
[ 46.727626] [<c104f566>] lock_acquire+0x59/0x70
[ 46.727632] [<c12c5694>] __mutex_lock_common+0x39/0x36b
[ 46.727639] [<c12c5a00>] mutex_lock_nested+0x12/0x15
[ 46.727644] [<c10f1d5c>] reiserfs_file_release+0x12b/0x367
[ 46.727650] [<c10a5805>] fput+0xe0/0x16a
[ 46.727657] [<c1090c9e>] remove_vma+0x28/0x47
[ 46.727662] [<c1091a60>] do_munmap+0x1e8/0x200
[ 46.727667] [<c109230a>] mmap_region+0x6b/0x372
[ 46.727672] [<c109284d>] do_mmap_pgoff+0x23c/0x282
[ 46.727678] [<c1092950>] sys_mmap_pgoff+0xbd/0xe7
[ 46.727683] [<c10027d3>] sysenter_do_call+0x12/0x32
[ 46.727689]
[ 46.727690] other info that might help us debug this:
[ 46.727691]
[ 46.727695] 1 lock held by a.out/5840:
[ 46.727698] #0: (&mm->mmap_sem){++++++}, at: [<c1092937>] sys_mmap_pgoff+0xa4/0xe7
[ 46.727707]
[ 46.727708] stack backtrace:
[ 46.727713] Pid: 5840, comm: a.out Not tainted 2.6.35-rc3-dbg-git5-00446-g36336bc-dirty #64
[ 46.727717] Call Trace:
[ 46.727722] [<c12c4913>] ? printk+0xf/0x11
[ 46.727728] [<c104dc09>] print_circular_bug+0x8a/0x96
[ 46.727734] [<c104ef5c>] __lock_acquire+0x96d/0xbe1
[ 46.727740] [<c104ccc8>] ? look_up_lock_class+0x6c/0x7b
[ 46.727746] [<c104e462>] ? mark_lock+0x26/0x1b3
[ 46.727752] [<c104f566>] lock_acquire+0x59/0x70
[ 46.727758] [<c10f1d5c>] ? reiserfs_file_release+0x12b/0x367
[ 46.727764] [<c12c5694>] __mutex_lock_common+0x39/0x36b
[ 46.727769] [<c10f1d5c>] ? reiserfs_file_release+0x12b/0x367
[ 46.727775] [<c12c5a00>] mutex_lock_nested+0x12/0x15
[ 46.727781] [<c10f1d5c>] ? reiserfs_file_release+0x12b/0x367
[ 46.727786] [<c10f1d5c>] reiserfs_file_release+0x12b/0x367
[ 46.727792] [<c108d77d>] ? free_pgd_range+0x96/0x12f
[ 46.727798] [<c10a57b5>] ? fput+0x90/0x16a
[ 46.727803] [<c10a5805>] fput+0xe0/0x16a
[ 46.727808] [<c1090c9e>] remove_vma+0x28/0x47
[ 46.727814] [<c1091811>] ? arch_unmap_area_topdown+0x0/0x18
[ 46.727819] [<c1091a60>] do_munmap+0x1e8/0x200
[ 46.727825] [<c109230a>] mmap_region+0x6b/0x372
[ 46.727831] [<c109284d>] do_mmap_pgoff+0x23c/0x282
[ 46.727837] [<c1092950>] sys_mmap_pgoff+0xbd/0xe7
[ 46.727842] [<c10027d3>] sysenter_do_call+0x12/0x32*/
#define _GNU_SOURCE
#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/mman.h>
int main()
{
char buf[4096];
int i = 0;
/* we don't really care */
for (; i < 4096; i++)
buf[i] = (i + 65) % 255;
for (i = 0; i < 10; i++) {
int pid = fork();
if (pid > 0 ) {
printf("parent...");
} else if (pid == 0) {
printf("child...\n");
int fd = open("conftest.mmap", O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0600);
if (fd > 0) {
printf("OPEN ok %d\n", fd);
if (write(fd, buf, 4096) < 0)
printf("WRITE error\n");
else
printf("WRITE ok\n");
close(fd);
} else {
printf("OPEN error\n");
}
fd = open("conftest.mmap", O_RDWR|O_LARGEFILE);
if (fd > 0) {
printf("OPEN conftest.mmap %d\n", fd);
void *map = mmap((void*)0xb78a8000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, fd, 0);
if (map == MAP_FAILED) {
printf("MMAP failed\n");
goto out;
} else {
printf("MMAP ok\n");
}
if (read(fd, buf, 4096) < 0)
printf("READ failed\n");
else
printf("READ ok\n");
close(fd);
} else {
printf("Error: can't open conftest.mmap\n");
}
out:
fd = open(".", O_RDONLY|O_LARGEFILE);
if (fd > 0) {
printf("OPEN . ok %d... closing\n", fd);
close(fd);
} else {
printf("OPEN error\n");
}
struct stat _stat;
if (fstatat(AT_FDCWD, "conftest.mmap", &_stat, AT_SYMLINK_NOFOLLOW) < 0)
printf("FSTATAT error\n");
else
printf("FSTATAT ok\n");
if (unlinkat(AT_FDCWD, "conftest.mmap", 0) < 0)
printf("UNLINKAT error\n");
else
printf("UNLINKAT ok\n");
} else {
printf("FORK error\n");
}
}
return 0;
}Quoted 65 lines · Sergey Senozhatsky
Hello,
I searched lkml and found the following report (titled reiserfs locking):
http://lkml.org/lkml/2010/4/15/389 (Fri, 16 Apr 2010)I can see this backtrace while configuring emacs:
[ 6136.579013] [ 6136.579015] ======================================================= [ 6136.579021] [ INFO: possible circular locking dependency detected ] [ 6136.579025] 2.6.35-rc3-dbg-git3-00350-g94d119f #61 [ 6136.579027] ------------------------------------------------------- [ 6136.579031] conftest/13950 is trying to acquire lock: [ 6136.579034] (&sb->s_type->i_mutex_key#10){+.+.+.}, at: [<c10f1cf6>] reiserfs_file_release+0x11d/0x344 [ 6136.579050] [ 6136.579051] but task is already holding lock: [ 6136.579054] (&mm->mmap_sem){++++++}, at: [<c1091b87>] sys_munmap+0x26/0x42 [ 6136.579064] [ 6136.579065] which lock already depends on the new lock. [ 6136.579066] [ 6136.579069] [ 6136.579070] the existing dependency chain (in reverse order) is: [ 6136.579073] [ 6136.579074] -> #1 (&mm->mmap_sem){++++++}: [ 6136.579081] [<c104f53a>] lock_acquire+0x59/0x70 [ 6136.579089] [<c108cf78>] might_fault+0x53/0x70 [ 6136.579094] [<c11853b8>] copy_to_user+0x30/0x48 [ 6136.579101] [<c10afaf9>] filldir64+0x95/0xc9 [ 6136.579106] [<c10f2504>] reiserfs_readdir_dentry+0x35d/0x4d9 [ 6136.579112] [<c10f2692>] reiserfs_readdir+0x12/0x17 [ 6136.579117] [<c10afd17>] vfs_readdir+0x6d/0x92 [ 6136.579122] [<c10afe91>] sys_getdents64+0x63/0xa2 [ 6136.579127] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 6136.579134] [ 6136.579135] -> #0 (&sb->s_type->i_mutex_key#10){+.+.+.}: [ 6136.579142] [<c104ef30>] __lock_acquire+0x96d/0xbe1 [ 6136.579148] [<c104f53a>] lock_acquire+0x59/0x70 [ 6136.579153] [<c12c5614>] __mutex_lock_common+0x39/0x36b [ 6136.579160] [<c12c5980>] mutex_lock_nested+0x12/0x15 [ 6136.579165] [<c10f1cf6>] reiserfs_file_release+0x11d/0x344 [ 6136.579171] [<c10a580d>] fput+0xe0/0x16a [ 6136.579177] [<c1090ca6>] remove_vma+0x28/0x47 [ 6136.579182] [<c1091a68>] do_munmap+0x1e8/0x200 [ 6136.579188] [<c1091b93>] sys_munmap+0x32/0x42 [ 6136.579193] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 6136.579198] [ 6136.579199] other info that might help us debug this: [ 6136.579200] [ 6136.579204] 1 lock held by conftest/13950: [ 6136.579207] #0: (&mm->mmap_sem){++++++}, at: [<c1091b87>] sys_munmap+0x26/0x42 [ 6136.579216] [ 6136.579217] stack backtrace: [ 6136.579221] Pid: 13950, comm: conftest Not tainted 2.6.35-rc3-dbg-git3-00350-g94d119f #61 [ 6136.579225] Call Trace: [ 6136.579230] [<c12c4893>] ? printk+0xf/0x11 [ 6136.579235] [<c104dbdd>] print_circular_bug+0x8a/0x96 [ 6136.579241] [<c104ef30>] __lock_acquire+0x96d/0xbe1 [ 6136.579247] [<c104e436>] ? mark_lock+0x26/0x1b3 [ 6136.579254] [<c104f53a>] lock_acquire+0x59/0x70 [ 6136.579259] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344 [ 6136.579265] [<c12c5614>] __mutex_lock_common+0x39/0x36b [ 6136.579270] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344 [ 6136.579276] [<c1084375>] ? release_pages+0x55/0x116 [ 6136.579282] [<c12c5980>] mutex_lock_nested+0x12/0x15 [ 6136.579287] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344
Quoted 9 lines
[ 6136.579293] [<c10f1cf6>] reiserfs_file_release+0x11d/0x344 [ 6136.579299] [<c10a57bd>] ? fput+0x90/0x16a [ 6136.579304] [<c10a580d>] fput+0xe0/0x16a [ 6136.579309] [<c1090ca6>] remove_vma+0x28/0x47 [ 6136.579314] [<c1091819>] ? arch_unmap_area_topdown+0x0/0x18 [ 6136.579319] [<c1091a68>] do_munmap+0x1e8/0x200 [ 6136.579325] [<c1091b93>] sys_munmap+0x32/0x42 [ 6136.579330] [<c10027d3>] sysenter_do_call+0x12/0x32
Right.
The problem is:
vfs_readdir() { do_munmap() {
mutex_lock(inode); read or write(don't know)_lock(mm->mmap_sem)
reiserfs_readdir() { reiserfs_file_release() {
read_lock(mm->mmap_sem) mutex_lock(inode);
} }
} }I don't think the deadlock can really happen, as we can't release the directory while we are reading it. Plus I guess we can't mmap a directory (someone correct me if I'm wrong).
But still the warning is annoying. I suspect this was there for a while. There are other known "inversions that can't happen" in reiserfs, one is on the xattrs code (reiserfs/xattr.c:xattr_unlink()) which warning you can trigger under testing pressure.
But this one looks easy to trigger, although I've never seen it, but I'll try your testcase.
Anyway, we can't change the readdir path. mutex -> mm->mmap_sem is the correct ordering, we need to call copy_to_user there anyway.
So the challenge is to look at this mutex_lock(&inode->i_mutex) in reiserfs_file_release(). I really don't know what it is protecting.
Is there someone who could give me a hint here?
Quoted 33 lines
As well as this:
[ 202.300464] REISERFS error (device sda9): vs-2100 add_save_link: search_by_key ([-1 7812832 0x1 IND]) returned 1 [ 202.300473] REISERFS (device sda9): Remounting filesystem read-only [ 202.301603] ------------[ cut here ]------------ [ 202.301615] WARNING: at fs/reiserfs/journal.c:3436 journal_end+0x5b/0xaf() [ 202.301619] Hardware name: F3JC [ 202.301622] Modules linked in: snd_seq_dummy snd_seq_oss
snd_seq_midi_eventsnd_seqsnd_seq_devicesnd_hda_codec_si3054snd_hda_codec_realteksnd_hwdepsnd_pcm_osssnd_mixer_ossasus_laptopsparse_keymapsdhci_pcisdhcisnd_hda_intelmmc_coreled_classsnd_hda_codecrng_coresnd_pcmsnd_timersnd_page_allocsndi2c_i801soundcore psmouse sg evdevserio_rawr8169 mii usbhid hiduhci_hcdehci_hcdsr_modusbcore cdromsd_modata_piix[ 202.301689] Pid: 5055, comm: a.out Not tainted 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65 [ 202.301693] Call Trace: [ 202.301701] [<c102e3a2>] warn_slowpath_common+0x65/0x7a [ 202.301707] [<c1102d95>] ? journal_end+0x5b/0xaf [ 202.301712] [<c102e3c6>] warn_slowpath_null+0xf/0x13 [ 202.301718] [<c1102d95>] journal_end+0x5b/0xaf [ 202.301725] [<c10eebaa>] reiserfs_truncate_file+0x19f/0x233 [ 202.301733] [<c10f1ba1>] reiserfs_vfs_truncate_file+0xd/0xf [ 202.301738] [<c1084883>] vmtruncate+0x23/0x29 [ 202.301745] [<c10b4ad4>] inode_setattr+0x47/0x68 [ 202.301751] [<c10f1b3f>] reiserfs_setattr+0x242/0x297 [ 202.301758] [<c12c5d05>] ? down_write+0x22/0x2a [ 202.301764] [<c10b4d6f>] notify_change+0x15c/0x26b [ 202.301770] [<c10a35c7>] do_truncate+0x64/0x7d [ 202.301776] [<c12c69b0>] ? _raw_spin_unlock+0x33/0x3f [ 202.301783] [<c10ad892>] do_last+0x450/0x459
Quoted 8 lines
[ 202.301789] [<c10ada5b>] do_filp_open+0x1c0/0x41a [ 202.301798] [<c1029081>] ? get_parent_ip+0xb/0x31 [ 202.301804] [<c1029123>] ? sub_preempt_count+0x7c/0x89 [ 202.301810] [<c10b5746>] ? alloc_fd+0xb4/0xbf [ 202.301816] [<c10a3f46>] do_sys_open+0x48/0xdf [ 202.301821] [<c10a3ffb>] sys_open+0x1e/0x26 [ 202.301827] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 202.301833] ---[ end trace c4e3312bdadd2dc5 ]---
Ah that's wicked. I'll try your testcase for that too.
Thanks a lot for providing these!
Quoted 7 lines · Frederic Weisbecker
I don't think the deadlock can really happen, as we can't release the directory while we are reading it. Plus I guess we can't mmap a directory (someone correct me if I'm wrong).
Is there someone who could give me a hint here?
If its purely directories you can try and give directory inode locks a different class.
Quoted 11 lines · Peter Zijlstra
Quoted 7 lines · Frederic Weisbecker
I don't think the deadlock can really happen, as we can't release the directory while we are reading it. Plus I guess we can't mmap a directory (someone correct me if I'm wrong).
Is there someone who could give me a hint here?
If its purely directories you can try and give directory inode locks a different class.
We have a static layout in include/linux/fs.h:
enum inode_i_mutex_lock_class
{
I_MUTEX_NORMAL,
I_MUTEX_PARENT,
I_MUTEX_CHILD,
I_MUTEX_XATTR,
I_MUTEX_QUOTA
};
I fear none of them fits in our scheme, except the normal one.
And playing with a supplementary set of classes only used in a single place would make the lockdep checks useless there, even worse it would make us missing a lot of right checks.
vfs_readdir() locks the directory before calling the fs, and none of the nested classes would be in its good right there, as the directory is the current inode to work on, not a parent nor a child or so.
And unfortunately it's the same in reiserfs_file_release() that can close about whatever inode, we are locking the current inode, not a parent, child, xattr, or so.
So changing the nesting class here would not be a good thing to do I fear.
Quoted 4 lines · Frederic Weisbecker
vfs_readdir()locks the directory before calling the fs, and none of the nested classes would be in its good right there, as the directory is the current inode to work on, not a parent nor a child or so.
Nah, what I was referring to is sticking the i_mutex into a different class on inode creation, but apparently we already do, see unlock_new_inode().
So its not a pure directory thing..
Quoted 96 lines · Frederic Weisbecker
Quoted 55 lines · Sergey Senozhatsky
Hello,
I searched lkml and found the following report (titled reiserfs locking):
http://lkml.org/lkml/2010/4/15/389 (Fri, 16 Apr 2010)I can see this backtrace while configuring emacs:
[ 6136.579013] [ 6136.579015] ======================================================= [ 6136.579021] [ INFO: possible circular locking dependency detected ] [ 6136.579025] 2.6.35-rc3-dbg-git3-00350-g94d119f #61 [ 6136.579027] ------------------------------------------------------- [ 6136.579031] conftest/13950 is trying to acquire lock: [ 6136.579034] (&sb->s_type->i_mutex_key#10){+.+.+.}, at: [<c10f1cf6>] reiserfs_file_release+0x11d/0x344 [ 6136.579050] [ 6136.579051] but task is already holding lock: [ 6136.579054] (&mm->mmap_sem){++++++}, at: [<c1091b87>] sys_munmap+0x26/0x42 [ 6136.579064] [ 6136.579065] which lock already depends on the new lock. [ 6136.579066] [ 6136.579069] [ 6136.579070] the existing dependency chain (in reverse order) is: [ 6136.579073] [ 6136.579074] -> #1 (&mm->mmap_sem){++++++}: [ 6136.579081] [<c104f53a>] lock_acquire+0x59/0x70 [ 6136.579089] [<c108cf78>] might_fault+0x53/0x70 [ 6136.579094] [<c11853b8>] copy_to_user+0x30/0x48 [ 6136.579101] [<c10afaf9>] filldir64+0x95/0xc9 [ 6136.579106] [<c10f2504>] reiserfs_readdir_dentry+0x35d/0x4d9 [ 6136.579112] [<c10f2692>] reiserfs_readdir+0x12/0x17 [ 6136.579117] [<c10afd17>] vfs_readdir+0x6d/0x92 [ 6136.579122] [<c10afe91>] sys_getdents64+0x63/0xa2 [ 6136.579127] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 6136.579134] [ 6136.579135] -> #0 (&sb->s_type->i_mutex_key#10){+.+.+.}: [ 6136.579142] [<c104ef30>] __lock_acquire+0x96d/0xbe1 [ 6136.579148] [<c104f53a>] lock_acquire+0x59/0x70 [ 6136.579153] [<c12c5614>] __mutex_lock_common+0x39/0x36b [ 6136.579160] [<c12c5980>] mutex_lock_nested+0x12/0x15 [ 6136.579165] [<c10f1cf6>] reiserfs_file_release+0x11d/0x344 [ 6136.579171] [<c10a580d>] fput+0xe0/0x16a [ 6136.579177] [<c1090ca6>] remove_vma+0x28/0x47 [ 6136.579182] [<c1091a68>] do_munmap+0x1e8/0x200 [ 6136.579188] [<c1091b93>] sys_munmap+0x32/0x42 [ 6136.579193] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 6136.579198] [ 6136.579199] other info that might help us debug this: [ 6136.579200] [ 6136.579204] 1 lock held by conftest/13950: [ 6136.579207] #0: (&mm->mmap_sem){++++++}, at: [<c1091b87>] sys_munmap+0x26/0x42 [ 6136.579216] [ 6136.579217] stack backtrace: [ 6136.579221] Pid: 13950, comm: conftest Not tainted 2.6.35-rc3-dbg-git3-00350-g94d119f #61 [ 6136.579225] Call Trace: [ 6136.579230] [<c12c4893>] ? printk+0xf/0x11 [ 6136.579235] [<c104dbdd>] print_circular_bug+0x8a/0x96 [ 6136.579241] [<c104ef30>] __lock_acquire+0x96d/0xbe1 [ 6136.579247] [<c104e436>] ? mark_lock+0x26/0x1b3 [ 6136.579254] [<c104f53a>] lock_acquire+0x59/0x70 [ 6136.579259] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344 [ 6136.579265] [<c12c5614>] __mutex_lock_common+0x39/0x36b [ 6136.579270] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344 [ 6136.579276] [<c1084375>] ? release_pages+0x55/0x116 [ 6136.579282] [<c12c5980>] mutex_lock_nested+0x12/0x15 [ 6136.579287] [<c10f1cf6>] ? reiserfs_file_release+0x11d/0x344Quoted 10 lines
[ 6136.579293] [<c10f1cf6>] reiserfs_file_release+0x11d/0x344 [ 6136.579299] [<c10a57bd>] ? fput+0x90/0x16a [ 6136.579304] [<c10a580d>] fput+0xe0/0x16a [ 6136.579309] [<c1090ca6>] remove_vma+0x28/0x47 [ 6136.579314] [<c1091819>] ? arch_unmap_area_topdown+0x0/0x18 [ 6136.579319] [<c1091a68>] do_munmap+0x1e8/0x200 [ 6136.579325] [<c1091b93>] sys_munmap+0x32/0x42 [ 6136.579330] [<c10027d3>] sysenter_do_call+0x12/0x32Right.
The problem is:
vfs_readdir() { do_munmap() { mutex_lock(inode); read or write(don't know)_lock(mm->mmap_sem) reiserfs_readdir() { reiserfs_file_release() { read_lock(mm->mmap_sem) mutex_lock(inode); } } } }I don't think the deadlock can really happen, as we can't release the directory while we are reading it. Plus I guess we can't mmap a directory (someone correct me if I'm wrong).
But still the warning is annoying. I suspect this was there for a while. There are other known "inversions that can't happen" in reiserfs, one is on the xattrs code (reiserfs/xattr.c:xattr_unlink()) which warning you can trigger under testing pressure.
But this one looks easy to trigger, although I've never seen it, but I'll try your testcase.
Quoted 8 lines
Anyway, we can't change the readdir path. mutex ->
mm->mmap_semis the correct ordering, we need to callcopy_to_userthere anyway.So the challenge is to look at this
mutex_lock(&inode->i_mutex)inreiserfs_file_release(). I really don't know what it is protecting.Is there someone who could give me a hint here?
I guess this protects file's data during so-called tail conversions. This is a comment in indirect2direct():
// we are protected by i_mutex. The tail can not disapper, not // append can be done either
// we are in truncate or packing tail in file_release
Edward.
Quoted 56 lines
Quoted 34 lines
As well as this:
[ 202.300464] REISERFS error (device sda9): vs-2100 add_save_link: search_by_key ([-1 7812832 0x1 IND]) returned 1 [ 202.300473] REISERFS (device sda9): Remounting filesystem read-only [ 202.301603] ------------[ cut here ]------------ [ 202.301615] WARNING: at fs/reiserfs/journal.c:3436 journal_end+0x5b/0xaf() [ 202.301619] Hardware name: F3JC [ 202.301622] Modules linked in: snd_seq_dummy snd_seq_oss
snd_seq_midi_eventsnd_seqsnd_seq_devicesnd_hda_codec_si3054snd_hda_codec_realteksnd_hwdepsnd_pcm_osssnd_mixer_ossasus_laptopsparse_keymapsdhci_pcisdhcisnd_hda_intelmmc_coreled_classsnd_hda_codecrng_coresnd_pcmsnd_timersnd_page_allocsndi2c_i801soundcore psmouse sg evdevserio_rawr8169 mii usbhid hiduhci_hcdehci_hcdsr_modusbcore cdromsd_modata_piix[ 202.301689] Pid: 5055, comm: a.out Not tainted 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65 [ 202.301693] Call Trace: [ 202.301701] [<c102e3a2>] warn_slowpath_common+0x65/0x7a [ 202.301707] [<c1102d95>] ? journal_end+0x5b/0xaf [ 202.301712] [<c102e3c6>] warn_slowpath_null+0xf/0x13 [ 202.301718] [<c1102d95>] journal_end+0x5b/0xaf [ 202.301725] [<c10eebaa>] reiserfs_truncate_file+0x19f/0x233 [ 202.301733] [<c10f1ba1>] reiserfs_vfs_truncate_file+0xd/0xf [ 202.301738] [<c1084883>] vmtruncate+0x23/0x29 [ 202.301745] [<c10b4ad4>] inode_setattr+0x47/0x68 [ 202.301751] [<c10f1b3f>] reiserfs_setattr+0x242/0x297 [ 202.301758] [<c12c5d05>] ? down_write+0x22/0x2a [ 202.301764] [<c10b4d6f>] notify_change+0x15c/0x26b [ 202.301770] [<c10a35c7>] do_truncate+0x64/0x7d [ 202.301776] [<c12c69b0>] ? _raw_spin_unlock+0x33/0x3f [ 202.301783] [<c10ad892>] do_last+0x450/0x459Quoted 9 lines
[ 202.301789] [<c10ada5b>] do_filp_open+0x1c0/0x41a [ 202.301798] [<c1029081>] ? get_parent_ip+0xb/0x31 [ 202.301804] [<c1029123>] ? sub_preempt_count+0x7c/0x89 [ 202.301810] [<c10b5746>] ? alloc_fd+0xb4/0xbf [ 202.301816] [<c10a3f46>] do_sys_open+0x48/0xdf [ 202.301821] [<c10a3ffb>] sys_open+0x1e/0x26 [ 202.301827] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 202.301833] ---[ end trace c4e3312bdadd2dc5 ]---Ah that's wicked. I'll try your testcase for that too.
Thanks a lot for providing these!
--
To unsubscribe from this list: send the line "unsubscribe reiserfs-devel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 44 lines · Frederic Weisbecker
Quoted 33 lines
As well as this:
[ 202.300464] REISERFS error (device sda9): vs-2100 add_save_link: search_by_key ([-1 7812832 0x1 IND]) returned 1 [ 202.300473] REISERFS (device sda9): Remounting filesystem read-only [ 202.301603] ------------[ cut here ]------------ [ 202.301615] WARNING: at fs/reiserfs/journal.c:3436 journal_end+0x5b/0xaf() [ 202.301619] Hardware name: F3JC [ 202.301622] Modules linked in: snd_seq_dummy snd_seq_oss
snd_seq_midi_eventsnd_seqsnd_seq_devicesnd_hda_codec_si3054snd_hda_codec_realteksnd_hwdepsnd_pcm_osssnd_mixer_ossasus_laptopsparse_keymapsdhci_pcisdhcisnd_hda_intelmmc_coreled_classsnd_hda_codecrng_coresnd_pcmsnd_timersnd_page_allocsndi2c_i801soundcore psmouse sg evdevserio_rawr8169 mii usbhid hiduhci_hcdehci_hcdsr_modusbcore cdromsd_modata_piix[ 202.301689] Pid: 5055, comm: a.out Not tainted 2.6.35-rc3-dbg-git6-00502-g94feaba-dirty #65 [ 202.301693] Call Trace: [ 202.301701] [<c102e3a2>] warn_slowpath_common+0x65/0x7a [ 202.301707] [<c1102d95>] ? journal_end+0x5b/0xaf [ 202.301712] [<c102e3c6>] warn_slowpath_null+0xf/0x13 [ 202.301718] [<c1102d95>] journal_end+0x5b/0xaf [ 202.301725] [<c10eebaa>] reiserfs_truncate_file+0x19f/0x233 [ 202.301733] [<c10f1ba1>] reiserfs_vfs_truncate_file+0xd/0xf [ 202.301738] [<c1084883>] vmtruncate+0x23/0x29 [ 202.301745] [<c10b4ad4>] inode_setattr+0x47/0x68 [ 202.301751] [<c10f1b3f>] reiserfs_setattr+0x242/0x297 [ 202.301758] [<c12c5d05>] ? down_write+0x22/0x2a [ 202.301764] [<c10b4d6f>] notify_change+0x15c/0x26b [ 202.301770] [<c10a35c7>] do_truncate+0x64/0x7d [ 202.301776] [<c12c69b0>] ? _raw_spin_unlock+0x33/0x3f [ 202.301783] [<c10ad892>] do_last+0x450/0x459Quoted 8 lines
[ 202.301789] [<c10ada5b>] do_filp_open+0x1c0/0x41a [ 202.301798] [<c1029081>] ? get_parent_ip+0xb/0x31 [ 202.301804] [<c1029123>] ? sub_preempt_count+0x7c/0x89 [ 202.301810] [<c10b5746>] ? alloc_fd+0xb4/0xbf [ 202.301816] [<c10a3f46>] do_sys_open+0x48/0xdf [ 202.301821] [<c10a3ffb>] sys_open+0x1e/0x26 [ 202.301827] [<c10027d3>] sysenter_do_call+0x12/0x32 [ 202.301833] ---[ end trace c4e3312bdadd2dc5 ]---
Quoted 4 lines
Ah that's wicked. I'll try your testcase for that too.
Thanks a lot for providing these!
I'm not very lucky at hitting this. Saw it several times (may be 3).
Sergey
Quoted 14 lines · Frederic Weisbecker
Right.
The problem is:
vfs_readdir() { do_munmap() { mutex_lock(inode); read or write(don't know)_lock(mm->mmap_sem) reiserfs_readdir() { reiserfs_file_release() { read_lock(mm->mmap_sem) mutex_lock(inode); } } } }
Quoted 3 lines
I don't think the deadlock can really happen, as we can't release the directory while we are reading it. Plus I guess we can't mmap a directory (someone correct me if I'm wrong).
Gyah... For the 1001st time: readdir() is far from being the only thing that nests mmap_sem inside i_mutex. In particular, write() does the same thing.
So yes, it *is* a real deadlock, TYVM, with no directories involved. Open the same file twice, mmap one fd, close it, then have munmap() hitting i_mutex in reiserfs_file_release() race with write() through another fd.
Incidentally, reiserfs_file_release() checks in the fastpath look completely bogus. Checking i_count? What the hell is that one about? And no, these checks won't stop open() coming between them and grabbing i_mutex, so they couldn't prevent the deadlock in question anyway.
Quoted 11 lines · Al Viro
Gyah... For the 1001st time:
readdir()is far from being the only thing that nestsmmap_seminsidei_mutex. In particular,write()does the same thing.So yes, it *is* a real deadlock, TYVM, with no directories involved. Open the same file twice, mmap one fd, close it, then have
munmap()hittingi_mutexinreiserfs_file_release()race withwrite()through another fd.Incidentally,
reiserfs_file_release()checks in the fastpath look completely bogus. Checkingi_count? What the hell is that one about? And no, these checks won't stopopen()coming between them and grabbingi_mutex, so they couldn't prevent the deadlock in question anyway.
... and unfortunately it's been that way since the the initial merge in 2.4.early. FWIW, it seems that i_count check was a misguided attempt to check that no other opened struct file are there, but it's
a) wrong, since way, _way_ back - open() affects d_count, not i_count
b) wrong even with such modification (consider hardlinks)
c) wrong for even more reasons since forever - i_count and d_count could
be bumped by many things at any time
d) hopelessly racy anyway, since another open() could very well have
happened just as we'd finished these checks.
Quoted 14 lines · Al Viro
Quoted 11 lines · Al Viro
Gyah... For the 1001st time:
readdir()is far from being the only thing that nestsmmap_seminsidei_mutex. In particular,write()does the same thing.So yes, it *is* a real deadlock, TYVM, with no directories involved. Open the same file twice, mmap one fd, close it, then have
munmap()hittingi_mutexinreiserfs_file_release()race withwrite()through another fd.Incidentally,
reiserfs_file_release()checks in the fastpath look completely bogus. Checkingi_count? What the hell is that one about? And no, these checks won't stopopen()coming between them and grabbingi_mutex, so they couldn't prevent the deadlock in question anyway.
Quoted 9 lines
... and unfortunately it's been that way since the the initial merge in 2.4.early. FWIW, it seems that
i_countcheck was a misguided attempt to check that no other opened struct file are there, but it's
a) wrong, since way, _way_ back -open()affectsd_count, noti_count
b) wrong even with such modification (consider hardlinks)
c) wrong for even more reasons since forever -i_countandd_countcould
be bumped by many things at any time
d) hopelessly racy anyway, since anotheropen()could very well have
happened just as we'd finished these checks.
OK... See 22093b8f3d387f77 in vfs-2.6.git for-next (should propagate to git.kernel.org shortly). That ought to deal with this crap, assuming I hadn't fucked up somewhere...
Quoted 25 lines · Al Viro
Quoted 14 lines · Al Viro
Quoted 11 lines · Al Viro
Gyah... For the 1001st time:
readdir()is far from being the only thing that nestsmmap_seminsidei_mutex. In particular,write()does the same thing.So yes, it *is* a real deadlock, TYVM, with no directories involved. Open the same file twice, mmap one fd, close it, then have
munmap()hittingi_mutexinreiserfs_file_release()race withwrite()through another fd.Incidentally,
reiserfs_file_release()checks in the fastpath look completely bogus. Checkingi_count? What the hell is that one about? And no, these checks won't stopopen()coming between them and grabbingi_mutex, so they couldn't prevent the deadlock in question anyway.Quoted 9 lines
... and unfortunately it's been that way since the the initial merge in 2.4.early. FWIW, it seems that
i_countcheck was a misguided attempt to check that no other opened struct file are there, but it's
a) wrong, since way, _way_ back -open()affectsd_count, noti_count
b) wrong even with such modification (consider hardlinks)
c) wrong for even more reasons since forever -i_countandd_countcould
be bumped by many things at any time
d) hopelessly racy anyway, since anotheropen()could very well have
happened just as we'd finished these checks.
Quoted 3 lines
OK... See
22093b8f3d387f77in vfs-2.6.git for-next (should propagate to git.kernel.org shortly). That ought to deal with this crap, assuming I hadn't fucked up somewhere...
Looks good. Thanks for fixing this!
Quoted 26 lines · Frederic Weisbecker
Quoted 25 lines · Al Viro
Quoted 23 lines · Al Viro
Quoted 11 lines · Al Viro
Gyah... For the 1001st time:
readdir()is far from being the only thing that nestsmmap_seminsidei_mutex. In particular,write()does the same thing.So yes, it *is* a real deadlock, TYVM, with no directories involved. Open the same file twice, mmap one fd, close it, then have
munmap()hittingi_mutexinreiserfs_file_release()race withwrite()through another fd.Incidentally,
reiserfs_file_release()checks in the fastpath look completely bogus. Checkingi_count? What the hell is that one about? And no, these checks won't stopopen()coming between them and grabbingi_mutex, so they couldn't prevent the deadlock in question anyway.... and unfortunately it's been that way since the the initial merge in 2.4.early. FWIW, it seems that
i_countcheck was a misguided attempt to check that no other opened struct file are there, but it's
a) wrong, since way, _way_ back -open()affectsd_count, noti_count
b) wrong even with such modification (consider hardlinks)
c) wrong for even more reasons since forever -i_countandd_countcould
be bumped by many things at any time
d) hopelessly racy anyway, since anotheropen()could very well have
happened just as we'd finished these checks.
Quoted 7 lines
Quoted 3 lines
OK... See
22093b8f3d387f77in vfs-2.6.git for-next (should propagate to git.kernel.org shortly). That ought to deal with this crap, assuming I hadn't fucked up somewhere...Looks good. Thanks for fixing this!
Seems to work fine with my test app.
Reported-by/Tested-by: Sergey Senozhatsky <[email protected]>
Sergey
Quoted 32 lines · Sergey Senozhatsky
Quoted 31 lines · Frederic Weisbecker
Quoted 28 lines · Al Viro
Quoted 14 lines · Al Viro
Quoted 11 lines · Al Viro
Gyah... For the 1001st time:
readdir()is far from being the only thing that nestsmmap_seminsidei_mutex. In particular,write()does the same thing.So yes, it *is* a real deadlock, TYVM, with no directories involved. Open the same file twice, mmap one fd, close it, then have
munmap()hittingi_mutexinreiserfs_file_release()race withwrite()through another fd.Incidentally,
reiserfs_file_release()checks in the fastpath look completely bogus. Checkingi_count? What the hell is that one about? And no, these checks won't stopopen()coming between them and grabbingi_mutex, so they couldn't prevent the deadlock in question anyway.Quoted 9 lines
... and unfortunately it's been that way since the the initial merge in 2.4.early. FWIW, it seems that
i_countcheck was a misguided attempt to check that no other opened struct file are there, but it's
a) wrong, since way, _way_ back -open()affectsd_count, noti_count
b) wrong even with such modification (consider hardlinks)
c) wrong for even more reasons since forever -i_countandd_countcould
be bumped by many things at any time
d) hopelessly racy anyway, since anotheropen()could very well have
happened just as we'd finished these checks.OK... See
22093b8f3d387f77in vfs-2.6.git for-next (should propagate to git.kernel.org shortly). That ought to deal with this crap, assuming I hadn't fucked up somewhere...
Quoted 6 lines
Quoted 2 lines
Looks good. Thanks for fixing this!
Seems to work fine with my test app.
Reported-by/Tested-by: Sergey Senozhatsky <[email protected]>
Great!