| Message ID | 1343852708-24009-1-git-send-email-jbacik@fusionio.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, Aug 1, 2012 at 3:25 PM, Josef Bacik <jbacik@fusionio.com> wrote: > We need an smb_mb() before waitqueue_active to avoid missing wakeups. > Before Mitch was hitting a deadlock between the ordered flushers and the > transaction commit because the ordered flushers were waiting for more refs > and were never woken up, so those smp_mb()'s are the most important. > Everything else I added for correctness sake and to avoid getting bitten by > this again somewhere else. Thanks, > This patch seems to make it tougher to hit a deadlock, but I'm still encountering intermittent deadlocks using this patch when running multiple rsync threads. I've also tested "Patch 2", and that has me hitting a deadlock even quicker (when starting several copying threads). I also found a slight performance hit using this patch. On a 3.4.6 kernel (merged with the 3.5_rc for-linus branch), I would typically complete my rsync test in ~265 seconds. Also, I can't recall hitting a deadlock on the 3.4.6 kernel (with 3.5_rc for-linus). When using this patch, the test would take ~310 seconds (when it didn't hit a deadlock). Here's the Delayed Tasks (Ctrl-SysRq-W) when using JUST this patch: [ 1568.794030] SysRq : Show Blocked State [ 1568.794101] task PC stack pid father [ 1568.794123] btrfs-endio-wri D ffff88012579c000 0 3845 2 0x00000000 [ 1568.794128] ffff8801254f3c20 0000000000000046 ffff8801254f2000 ffff8801241b5a80 [ 1568.794132] 0000000000012280 ffff8801254f3fd8 0000000000012280 0000000000004000 [ 1568.794136] ffff8801254f3fd8 0000000000012280 ffff880129af16a0 ffff8801241b5a80 [ 1568.794140] Call Trace: [ 1568.794179] [<ffffffffa0068785>] ? memcpy_extent_buffer+0x159/0x17a [btrfs] [ 1568.794200] [<ffffffffa0082ab7>] ? find_ref_head+0xa3/0xc6 [btrfs] [ 1568.794220] [<ffffffffa008343c>] ? btrfs_find_ref_cluster+0xdd/0x117 [btrfs] [ 1568.794225] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.794241] [<ffffffffa003fc86>] btrfs_run_delayed_refs+0x269/0x3f0 [btrfs] [ 1568.794246] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.794265] [<ffffffffa004fdc4>] __btrfs_end_transaction+0xca/0x283 [btrfs] [ 1568.794283] [<ffffffffa004ffda>] btrfs_end_transaction+0x15/0x17 [btrfs] [ 1568.794302] [<ffffffffa00555da>] btrfs_finish_ordered_io+0x2e4/0x334 [btrfs] [ 1568.794306] [<ffffffff8103b980>] ? run_timer_softirq+0x2d4/0x2d4 [ 1568.794325] [<ffffffffa005563f>] finish_ordered_fn+0x15/0x17 [btrfs] [ 1568.794344] [<ffffffffa0070ef8>] worker_loop+0x188/0x4e0 [btrfs] [ 1568.794365] [<ffffffffa0070d70>] ? btrfs_queue_worker+0x275/0x275 [btrfs] [ 1568.794384] [<ffffffffa0070d70>] ? btrfs_queue_worker+0x275/0x275 [btrfs] [ 1568.794387] [<ffffffff8104ac37>] kthread+0x89/0x91 [ 1568.794391] [<ffffffff8162fd74>] kernel_thread_helper+0x4/0x10 [ 1568.794395] [<ffffffff8104abae>] ? kthread_freezable_should_stop+0x57/0x57 [ 1568.794398] [<ffffffff8162fd70>] ? gs_change+0xb/0xb [ 1568.794400] btrfs-transacti D ffff88009912ba50 0 3851 2 0x00000000 [ 1568.794403] ffff8801241cfc70 0000000000000046 ffff8801241ce000 ffff8801248cda80 [ 1568.794407] 0000000000012280 ffff8801241cffd8 0000000000012280 0000000000004000 [ 1568.794411] ffff8801241cffd8 0000000000012280 ffff8801254b8000 ffff8801248cda80 [ 1568.794415] Call Trace: [ 1568.794436] [<ffffffffa0066646>] ? extent_writepages+0x53/0x5d [btrfs] [ 1568.794455] [<ffffffffa005357b>] ? uncompress_inline.clone.33+0x15f/0x15f [btrfs] [ 1568.794459] [<ffffffff810c9ada>] ? pagevec_lookup_tag+0x24/0x2e [ 1568.794478] [<ffffffffa0052e0e>] ? btrfs_writepages+0x27/0x29 [btrfs] [ 1568.794481] [<ffffffff810c90b1>] ? do_writepages+0x20/0x29 [ 1568.794485] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.794505] [<ffffffffa0061547>] btrfs_start_ordered_extent+0xde/0xfa [btrfs] [ 1568.794508] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.794529] [<ffffffffa0061984>] ? btrfs_lookup_first_ordered_extent+0x65/0x99 [btrfs] [ 1568.794549] [<ffffffffa0061a6a>] btrfs_wait_ordered_range+0xb2/0xda [btrfs] [ 1568.794569] [<ffffffffa0061bcc>] btrfs_run_ordered_operations+0x13a/0x1c1 [btrfs] [ 1568.794587] [<ffffffffa004f5f5>] btrfs_commit_transaction+0x287/0x960 [btrfs] [ 1568.794606] [<ffffffffa00502b1>] ? start_transaction+0x2d5/0x310 [btrfs] [ 1568.794609] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.794627] [<ffffffffa004913b>] transaction_kthread+0x187/0x258 [btrfs] [ 1568.794644] [<ffffffffa0048fb4>] ? btrfs_alloc_root+0x42/0x42 [btrfs] [ 1568.794661] [<ffffffffa0048fb4>] ? btrfs_alloc_root+0x42/0x42 [btrfs] [ 1568.794664] [<ffffffff8104ac37>] kthread+0x89/0x91 [ 1568.794668] [<ffffffff8162fd74>] kernel_thread_helper+0x4/0x10 [ 1568.794671] [<ffffffff8104abae>] ? kthread_freezable_should_stop+0x57/0x57 [ 1568.794674] [<ffffffff8162fd70>] ? gs_change+0xb/0xb [ 1568.794676] flush-btrfs-1 D ffff88012579c000 0 3857 2 0x00000000 [ 1568.794680] ffff880037125670 0000000000000046 ffff880037124000 ffff8801254b8000 [ 1568.794684] 0000000000012280 ffff880037125fd8 0000000000012280 0000000000004000 [ 1568.794687] ffff880037125fd8 0000000000012280 ffffffff81c13410 ffff8801254b8000 [ 1568.794691] Call Trace: [ 1568.794711] [<ffffffffa0082ab7>] ? find_ref_head+0xa3/0xc6 [btrfs] [ 1568.794731] [<ffffffffa008343c>] ? btrfs_find_ref_cluster+0xdd/0x117 [btrfs] [ 1568.794735] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.794750] [<ffffffffa003fc86>] btrfs_run_delayed_refs+0x269/0x3f0 [btrfs] [ 1568.794754] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.794772] [<ffffffffa004fdc4>] __btrfs_end_transaction+0xca/0x283 [btrfs] [ 1568.794790] [<ffffffffa004ffda>] btrfs_end_transaction+0x15/0x17 [btrfs] [ 1568.794809] [<ffffffffa005607e>] cow_file_range+0x3fa/0x453 [btrfs] [ 1568.794829] [<ffffffffa00635a6>] ? __set_extent_bit+0x3ce/0x403 [btrfs] [ 1568.794848] [<ffffffffa00569d7>] run_delalloc_range+0xbc/0x38f [btrfs] [ 1568.794868] [<ffffffffa006485f>] ? find_lock_delalloc_range.clone.26+0x18b/0x1b0 [btrfs] [ 1568.794889] [<ffffffffa0065d1d>] __extent_writepage+0x200/0x600 [btrfs] [ 1568.794909] [<ffffffffa006528d>] ? end_extent_writepage+0x5b/0x5b [btrfs] [ 1568.794913] [<ffffffff810bf853>] ? find_get_pages_tag+0xf8/0x134 [ 1568.794934] [<ffffffffa00662e4>] extent_write_cache_pages.clone.16.clone.29+0x1c7/0x30c [btrfs] [ 1568.794952] [<ffffffffa0051c2f>] ? __btrfs_submit_bio_done+0x1d/0x1d [btrfs] [ 1568.794971] [<ffffffffa0051c12>] ? btrfs_submit_bio_hook+0x122/0x122 [btrfs] [ 1568.794991] [<ffffffffa0062284>] ? submit_one_bio+0x8d/0x97 [btrfs] [ 1568.795011] [<ffffffffa006663b>] extent_writepages+0x48/0x5d [btrfs] [ 1568.795023] [<ffffffffa005357b>] ? uncompress_inline.clone.33+0x15f/0x15f [btrfs] [ 1568.795023] [<ffffffffa0052e0e>] btrfs_writepages+0x27/0x29 [btrfs] [ 1568.795023] [<ffffffff810c90b1>] do_writepages+0x20/0x29 [ 1568.795023] [<ffffffff8112e343>] __writeback_single_inode.clone.22+0x48/0x11c [ 1568.795023] [<ffffffff8112e8ab>] writeback_sb_inodes+0x1f0/0x332 [ 1568.795023] [<ffffffff8112ea65>] __writeback_inodes_wb+0x78/0xb9 [ 1568.795023] [<ffffffff8112ebec>] wb_writeback+0x146/0x23e [ 1568.795023] [<ffffffff81121cfd>] ? get_nr_inodes+0x48/0x5f [ 1568.795023] [<ffffffff8112f48a>] wb_do_writeback+0x154/0x1b0 [ 1568.795023] [<ffffffff8112f574>] bdi_writeback_thread+0x8e/0x1f1 [ 1568.795023] [<ffffffff8112f4e6>] ? wb_do_writeback+0x1b0/0x1b0 [ 1568.795023] [<ffffffff8112f4e6>] ? wb_do_writeback+0x1b0/0x1b0 [ 1568.795023] [<ffffffff8104ac37>] kthread+0x89/0x91 [ 1568.795023] [<ffffffff8162fd74>] kernel_thread_helper+0x4/0x10 [ 1568.795023] [<ffffffff8104abae>] ? kthread_freezable_should_stop+0x57/0x57 [ 1568.795023] [<ffffffff8162fd70>] ? gs_change+0xb/0xb [ 1568.795023] btrfs-endio-wri D ffff88012579c000 0 3899 2 0x00000000 [ 1568.795023] ffff880124ea5c20 0000000000000046 ffff880124ea4000 ffff8800a3015a80 [ 1568.795023] 0000000000012280 ffff880124ea5fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff880124ea5fd8 0000000000012280 ffff8800a30143e0 ffff8800a3015a80 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa0068785>] ? memcpy_extent_buffer+0x159/0x17a [btrfs] [ 1568.795023] [<ffffffffa0082ab7>] ? find_ref_head+0xa3/0xc6 [btrfs] [ 1568.795023] [<ffffffffa008343c>] ? btrfs_find_ref_cluster+0xdd/0x117 [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa003fc86>] btrfs_run_delayed_refs+0x269/0x3f0 [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa004fdc4>] __btrfs_end_transaction+0xca/0x283 [btrfs] [ 1568.795023] [<ffffffffa004ffda>] btrfs_end_transaction+0x15/0x17 [btrfs] [ 1568.795023] [<ffffffffa00555da>] btrfs_finish_ordered_io+0x2e4/0x334 [btrfs] [ 1568.795023] [<ffffffff8103b980>] ? run_timer_softirq+0x2d4/0x2d4 [ 1568.795023] [<ffffffffa005563f>] finish_ordered_fn+0x15/0x17 [btrfs] [ 1568.795023] [<ffffffffa0070ef8>] worker_loop+0x188/0x4e0 [btrfs] [ 1568.795023] [<ffffffffa0070d70>] ? btrfs_queue_worker+0x275/0x275 [btrfs] [ 1568.795023] [<ffffffff8104ac37>] kthread+0x89/0x91 [ 1568.795023] [<ffffffff8162fd74>] kernel_thread_helper+0x4/0x10 [ 1568.795023] [<ffffffff8104abae>] ? kthread_freezable_should_stop+0x57/0x57 [ 1568.795023] [<ffffffff8162fd70>] ? gs_change+0xb/0xb [ 1568.795023] rsync D 0000000000000000 0 3996 3989 0x00000000 [ 1568.795023] ffff88010cd05ca8 0000000000000082 ffff88010cd04000 ffff88011368ad40 [ 1568.795023] 0000000000012280 ffff88010cd05fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88010cd05fd8 0000000000012280 ffffffff81c13410 ffff88011368ad40 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 3997 3993 0x00000000 [ 1568.795023] ffff88010cd07ca8 0000000000000082 ffff88010cd06000 ffff880109a816a0 [ 1568.795023] 0000000000012280 ffff88010cd07fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88010cd07fd8 0000000000012280 ffff880129af16a0 ffff880109a816a0 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 3998 3991 0x00000000 [ 1568.795023] ffff88011f6f1ca8 0000000000000082 ffff88011f6f0000 ffff880125d55a80 [ 1568.795023] 0000000000012280 ffff88011f6f1fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011f6f1fd8 0000000000012280 ffff880129af16a0 ffff880125d55a80 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 3999 3994 0x00000000 [ 1568.795023] ffff88011f6f3ca8 0000000000000086 ffff88011f6f2000 ffff880109a843e0 [ 1568.795023] 0000000000012280 ffff88011f6f3fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011f6f3fd8 0000000000012280 ffff880129af16a0 ffff880109a843e0 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4000 3990 0x00000000 [ 1568.795023] ffff88011cbb9ca8 0000000000000086 ffff88011cbb8000 ffff880109a80000 [ 1568.795023] 0000000000012280 ffff88011cbb9fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011cbb9fd8 0000000000012280 ffff88011368ad40 ffff880109a80000 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4001 3992 0x00000000 [ 1568.795023] ffff88011cbbbca8 0000000000000086 ffff88011cbba000 ffff8801254b96a0 [ 1568.795023] 0000000000012280 ffff88011cbbbfd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011cbbbfd8 0000000000012280 ffff880129af16a0 ffff8801254b96a0 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa005012e>] ? start_transaction+0x152/0x310 [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4002 3995 0x00000000 [ 1568.795023] ffff88011dd67ca8 0000000000000082 ffff88011dd66000 ffff88012885da80 [ 1568.795023] 0000000000012280 ffff88011dd67fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011dd67fd8 0000000000012280 ffffffff81c13410 ffff88012885da80 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005126d>] __unlink_start_trans+0x94/0x404 [btrfs] [ 1568.795023] [<ffffffff810384a3>] ? inode_capable+0x15/0x1a [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffffa00577bb>] btrfs_unlink+0x2c/0xac [btrfs] [ 1568.795023] [<ffffffff81116c8f>] vfs_unlink+0x78/0xdd [ 1568.795023] [<ffffffff811189e8>] do_unlinkat+0xe6/0x178 [ 1568.795023] [<ffffffff8110de68>] ? fput+0x1e8/0x1f7 [ 1568.795023] [<ffffffff8110ac43>] ? filp_close+0x70/0x7b [ 1568.795023] [<ffffffff81119c7c>] sys_unlink+0x16/0x18 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4004 3998 0x00000000 [ 1568.795023] ffff88011a9adb68 0000000000000086 ffff88011a9ac000 ffff8801248cad40 [ 1568.795023] 0000000000012280 ffff88011a9adfd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011a9adfd8 0000000000012280 ffff880109a816a0 ffff8801248cad40 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffff8105136f>] ? in_group_p+0x31/0x33 [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005b47d>] btrfs_create+0x3a/0x1e0 [btrfs] [ 1568.795023] [<ffffffff8111ff3d>] ? d_splice_alias+0xcc/0xd8 [ 1568.795023] [<ffffffff81116a86>] vfs_create+0x9c/0xf5 [ 1568.795023] [<ffffffff81118e62>] do_last+0x2b9/0x807 [ 1568.795023] [<ffffffff811194af>] path_openat+0xcc/0x37f [ 1568.795023] [<ffffffff81119864>] do_filp_open+0x3d/0x89 [ 1568.795023] [<ffffffff810fe9b0>] ? kmem_cache_alloc+0x31/0x104 [ 1568.795023] [<ffffffff81123bb4>] ? alloc_fd+0x74/0x103 [ 1568.795023] [<ffffffff8110c09d>] do_sys_open+0x10f/0x1a1 [ 1568.795023] [<ffffffff8110c150>] sys_open+0x21/0x23 [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4006 3997 0x00000000 [ 1568.795023] ffff88011475dbb8 0000000000000086 ffff88011475c000 ffff880128995a80 [ 1568.795023] 0000000000012280 ffff88011475dfd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011475dfd8 0000000000012280 ffffffff81c13410 ffff880128995a80 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffff81116008>] ? generic_permission+0x86/0x209 [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005a8d0>] btrfs_rename+0x1bd/0x5af [btrfs] [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffff81116e9f>] ? vfs_rename+0xbe/0x3df [ 1568.795023] [<ffffffff8111705d>] vfs_rename+0x27c/0x3df [ 1568.795023] [<ffffffff8111a072>] sys_renameat+0x1ac/0x259 [ 1568.795023] [<ffffffff81122f26>] ? notify_change+0x2a2/0x2b8 [ 1568.795023] [<ffffffff811197e3>] ? user_path_at_empty+0x61/0x92 [ 1568.795023] [<ffffffff81124e45>] ? mntput_no_expire+0x3f/0x138 [ 1568.795023] [<ffffffff81124f68>] ? mntput+0x2a/0x2c [ 1568.795023] [<ffffffff811156a6>] ? path_put+0x22/0x26 [ 1568.795023] [<ffffffff8111a13a>] sys_rename+0x1b/0x1d [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4007 3996 0x00000000 [ 1568.795023] ffff88010c7cdbb8 0000000000000086 ffff88010c7cc000 ffff8801289916a0 [ 1568.795023] 0000000000012280 ffff88010c7cdfd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88010c7cdfd8 0000000000012280 ffff880129af16a0 ffff8801289916a0 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005a8d0>] btrfs_rename+0x1bd/0x5af [btrfs] [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffff81116e9f>] ? vfs_rename+0xbe/0x3df [ 1568.795023] [<ffffffff8111705d>] vfs_rename+0x27c/0x3df [ 1568.795023] [<ffffffff8111a072>] sys_renameat+0x1ac/0x259 [ 1568.795023] [<ffffffff81122f26>] ? notify_change+0x2a2/0x2b8 [ 1568.795023] [<ffffffff811197e3>] ? user_path_at_empty+0x61/0x92 [ 1568.795023] [<ffffffff81124e45>] ? mntput_no_expire+0x3f/0x138 [ 1568.795023] [<ffffffff81124f68>] ? mntput+0x2a/0x2c [ 1568.795023] [<ffffffff811156a6>] ? path_put+0x22/0x26 [ 1568.795023] [<ffffffff8111a13a>] sys_rename+0x1b/0x1d [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] rsync D 0000000000000000 0 4008 4001 0x00000000 [ 1568.795023] ffff88011ae33bb8 0000000000000086 ffff88011ae32000 ffff8801296a2d40 [ 1568.795023] 0000000000012280 ffff88011ae33fd8 0000000000012280 0000000000004000 [ 1568.795023] ffff88011ae33fd8 0000000000012280 ffff880128995a80 ffff8801296a2d40 [ 1568.795023] Call Trace: [ 1568.795023] [<ffffffffa003cb51>] ? get_alloc_profile+0x4e/0x50 [btrfs] [ 1568.795023] [<ffffffffa003cb82>] ? btrfs_get_alloc_profile+0x2f/0x31 [btrfs] [ 1568.795023] [<ffffffffa003db7f>] ? reserve_metadata_bytes.clone.57+0x339/0x66c [btrfs] [ 1568.795023] [<ffffffff8162d58c>] schedule+0x64/0x66 [ 1568.795023] [<ffffffffa004ee57>] wait_current_trans.clone.26+0xac/0xdd [btrfs] [ 1568.795023] [<ffffffff8104b10e>] ? wake_up_bit+0x2a/0x2a [ 1568.795023] [<ffffffffa0050157>] start_transaction+0x17b/0x310 [btrfs] [ 1568.795023] [<ffffffff81116008>] ? generic_permission+0x86/0x209 [ 1568.795023] [<ffffffffa0050553>] btrfs_start_transaction+0x13/0x15 [btrfs] [ 1568.795023] [<ffffffffa005a8d0>] btrfs_rename+0x1bd/0x5af [btrfs] [ 1568.795023] [<ffffffff81116127>] ? generic_permission+0x1a5/0x209 [ 1568.795023] [<ffffffff81116e9f>] ? vfs_rename+0xbe/0x3df [ 1568.795023] [<ffffffff8111705d>] vfs_rename+0x27c/0x3df [ 1568.795023] [<ffffffff8111a072>] sys_renameat+0x1ac/0x259 [ 1568.795023] [<ffffffff81122f26>] ? notify_change+0x2a2/0x2b8 [ 1568.795023] [<ffffffff811197e3>] ? user_path_at_empty+0x61/0x92 [ 1568.795023] [<ffffffff81124e45>] ? mntput_no_expire+0x3f/0x138 [ 1568.795023] [<ffffffff81124f68>] ? mntput+0x2a/0x2c [ 1568.795023] [<ffffffff811156a6>] ? path_put+0x22/0x26 [ 1568.795023] [<ffffffff8111a13a>] sys_rename+0x1b/0x1d [ 1568.795023] [<ffffffff8162ec92>] system_call_fastpath+0x16/0x1b [ 1568.795023] Sched Debug Version: v0.10, 3.5.0-baseline+ #1 [ 1568.795023] ktime : 1568797.341765 [ 1568.795023] sched_clk : 1568905.405572 [ 1568.795023] cpu_clk : 1568795.023448 [ 1568.795023] jiffies : 4296236090 [ 1568.795023] sched_clock_stable : 0 [ 1568.795023] [ 1568.795023] sysctl_sched [ 1568.795023] .sysctl_sched_latency : 12.000000 [ 1568.795023] .sysctl_sched_min_granularity : 1.500000 [ 1568.795023] .sysctl_sched_wakeup_granularity : 2.000000 [ 1568.795023] .sysctl_sched_child_runs_first : 0 [ 1568.795023] .sysctl_sched_features : 24119 [ 1568.795023] .sysctl_sched_tunable_scaling : 1 (logaritmic) [ 1568.795023] [ 1568.795023] cpu#0, 2666.627 MHz [ 1568.795023] .nr_running : 0 [ 1568.795023] .load : 0 [ 1568.795023] .nr_switches : 2092841 [ 1568.795023] .nr_load_updates : 319578 [ 1568.795023] .nr_uninterruptible : -79 [ 1568.795023] .next_balance : 4296.235984 [ 1568.795023] .curr->pid : 0 [ 1568.795023] .clock : 1568792.396031 [ 1568.795023] .cpu_load[0] : 0 [ 1568.795023] .cpu_load[1] : 0 [ 1568.795023] .cpu_load[2] : 0 [ 1568.795023] .cpu_load[3] : 0 [ 1568.795023] .cpu_load[4] : 0 [ 1568.795023] .yld_count : 21 [ 1568.795023] .sched_count : 2345950 [ 1568.795023] .sched_goidle : 683438 [ 1568.795023] .avg_idle : 1000000 [ 1568.795023] .ttwu_count : 1179205 [ 1568.795023] .ttwu_local : 432796 [ 1568.795023] [ 1568.795023] cfs_rq[0]:/ [ 1568.795023] .exec_clock : 122611.778679 [ 1568.795023] .MIN_vruntime : 0.000001 [ 1568.795023] .min_vruntime : 155427.205936 [ 1568.795023] .max_vruntime : 0.000001 [ 1568.795023] .spread : 0.000000 [ 1568.795023] .spread0 : 0.000000 [ 1568.795023] .nr_spread_over : 817 [ 1568.795023] .nr_running : 0 [ 1568.795023] .load : 0 [ 1568.795023] .load_avg : 0.000000 [ 1568.795023] .load_period : 0.000000 [ 1568.795023] .load_contrib : 0 [ 1568.795023] .load_tg : 0 [ 1568.795023] [ 1568.795023] rt_rq[0]: [ 1568.795023] .rt_nr_running : 0 [ 1568.795023] .rt_throttled : 0 [ 1568.795023] .rt_time : 0.000000 [ 1568.795023] .rt_runtime : 950.000000 [ 1568.795023] [ 1568.795023] runnable tasks: [ 1568.795023] task PID tree-key switches prio exec-runtime sum-exec sum-sleep [ 1568.795023] ---------------------------------------------------------------------------------------------------------- [ 1568.795023] [ 1568.795023] cpu#1, 2666.627 MHz [ 1568.795023] .nr_running : 0 [ 1568.795023] .load : 0 [ 1568.795023] .nr_switches : 2079916 [ 1568.795023] .nr_load_updates : 319982 [ 1568.795023] .nr_uninterruptible : 94 [ 1568.795023] .next_balance : 4296.235484 [ 1568.795023] .curr->pid : 0 [ 1568.795023] .clock : 1568784.396001 [ 1568.795023] .cpu_load[0] : 0 [ 1568.795023] .cpu_load[1] : 0 [ 1568.795023] .cpu_load[2] : 0 [ 1568.795023] .cpu_load[3] : 0 [ 1568.795023] .cpu_load[4] : 0 [ 1568.795023] .yld_count : 4 [ 1568.795023] .sched_count : 2308965 [ 1568.795023] .sched_goidle : 690586 [ 1568.795023] .avg_idle : 1000000 [ 1568.795023] .ttwu_count : 1159413 [ 1568.795023] .ttwu_local : 419527 [ 1568.795023] [ 1568.795023] cfs_rq[1]:/autogroup-12 [ 1568.795023] .exec_clock : 320.130894 [ 1568.795023] .MIN_vruntime : 0.000001 [ 1568.795023] .min_vruntime : 1163.780537 [ 1568.795023] .max_vruntime : 0.000001 [ 1568.795023] .spread : 0.000000 [ 1568.795023] .spread0 : -154263.425399 [ 1568.795023] .nr_spread_over : 178 [ 1568.795023] .nr_running : 0 [ 1568.795023] .load : 0 [ 1568.795023] .load_avg : 2559.999744 [ 1568.795023] .load_period : 5.241463 [ 1568.795023] .load_contrib : 488 [ 1568.795023] .load_tg : 488 [ 1568.795023] .se->exec_start : 1568778.912838 [ 1568.795023] .se->vruntime : 183270.984759 [ 1568.795023] .se->sum_exec_runtime : 320.130894 [ 1568.795023] .se->statistics.wait_start : 0.000000 [ 1568.795023] .se->statistics.sleep_start : 0.000000 [ 1568.795023] .se->statistics.block_start : 0.000000 [ 1568.795023] .se->statistics.sleep_max : 0.000000 [ 1568.795023] .se->statistics.block_max : 0.000000 [ 1568.795023] .se->statistics.exec_max : 1.000806 [ 1568.795023] .se->statistics.slice_max : 0.030516 [ 1568.795023] .se->statistics.wait_max : 2.745282 [ 1568.795023] .se->statistics.wait_sum : 8.576517 [ 1568.795023] .se->statistics.wait_count : 1446 [ 1568.795023] .se->load.weight : 2 [ 1568.795023] [ 1568.795023] cfs_rq[1]:/ [ 1568.795023] .exec_clock : 120018.151379 [ 1568.795023] .MIN_vruntime : 0.000001 [ 1568.795023] .min_vruntime : 183276.982041 [ 1568.795023] .max_vruntime : 0.000001 [ 1568.795023] .spread : 0.000000 [ 1568.795023] .spread0 : 27849.776105 [ 1568.795023] .nr_spread_over : 1197 [ 1568.795023] .nr_running : 0 [ 1568.795023] .load : 0 [ 1568.795023] .load_avg : 0.000000 [ 1568.795023] .load_period : 0.000000 [ 1568.795023] .load_contrib : 0 [ 1568.795023] .load_tg : 0 [ 1568.795023] [ 1568.795023] rt_rq[1]: [ 1568.795023] .rt_nr_running : 0 [ 1568.795023] .rt_throttled : 0 [ 1568.795023] .rt_time : 0.000000 [ 1568.795023] .rt_runtime : 950.000000 [ 1568.795023] [ 1568.795023] runnable tasks: [ 1568.795023] task PID tree-key switches prio exec-runtime sum-exec sum-sleep [ 1568.795023] ---------------------------------------------------------------------------------------------------------- [ 1568.795023] -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08/02/2012 04:25 AM, Josef Bacik wrote: > We need an smb_mb() before waitqueue_active to avoid missing wakeups. > Before Mitch was hitting a deadlock between the ordered flushers and the > transaction commit because the ordered flushers were waiting for more refs > and were never woken up, so those smp_mb()'s are the most important. > Everything else I added for correctness sake and to avoid getting bitten by > this again somewhere else. Thanks, > Hi Josef, I'll appreciate a lot if you can add some comments for each memory barrier, because not everyone knows why it is used here and there. :) thanks, liubo > Signed-off-by: Josef Bacik <jbacik@fusionio.com> > --- > fs/btrfs/compression.c | 1 + > fs/btrfs/delayed-inode.c | 16 ++++++++++------ > fs/btrfs/delayed-ref.c | 18 ++++++++++++------ > fs/btrfs/disk-io.c | 11 ++++++++--- > fs/btrfs/inode.c | 8 +++++--- > fs/btrfs/volumes.c | 8 +++++--- > 6 files changed, 41 insertions(+), 21 deletions(-) > > diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c > index 86eff48..43d1c5a 100644 > --- a/fs/btrfs/compression.c > +++ b/fs/btrfs/compression.c > @@ -818,6 +818,7 @@ static void free_workspace(int type, struct list_head *workspace) > btrfs_compress_op[idx]->free_workspace(workspace); > atomic_dec(alloc_workspace); > wake: > + smp_mb(); > if (waitqueue_active(workspace_wait)) > wake_up(workspace_wait); > } > diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c > index 335605c..8cc9b19 100644 > --- a/fs/btrfs/delayed-inode.c > +++ b/fs/btrfs/delayed-inode.c > @@ -513,9 +513,11 @@ static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) > rb_erase(&delayed_item->rb_node, root); > delayed_item->delayed_node->count--; > atomic_dec(&delayed_root->items); > - if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND && > - waitqueue_active(&delayed_root->wait)) > - wake_up(&delayed_root->wait); > + if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) { > + smp_mb(); > + if (waitqueue_active(&delayed_root->wait)) > + wake_up(&delayed_root->wait); > + } > } > > static void btrfs_release_delayed_item(struct btrfs_delayed_item *item) > @@ -1057,9 +1059,11 @@ static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) > delayed_root = delayed_node->root->fs_info->delayed_root; > atomic_dec(&delayed_root->items); > if (atomic_read(&delayed_root->items) < > - BTRFS_DELAYED_BACKGROUND && > - waitqueue_active(&delayed_root->wait)) > - wake_up(&delayed_root->wait); > + BTRFS_DELAYED_BACKGROUND) { > + smp_mb(); > + if (waitqueue_active(&delayed_root->wait)) > + wake_up(&delayed_root->wait); > + } > } > } > > diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c > index da7419e..858ef02 100644 > --- a/fs/btrfs/delayed-ref.c > +++ b/fs/btrfs/delayed-ref.c > @@ -662,9 +662,12 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info, > add_delayed_tree_ref(fs_info, trans, &ref->node, bytenr, > num_bytes, parent, ref_root, level, action, > for_cow); > - if (!need_ref_seq(for_cow, ref_root) && > - waitqueue_active(&fs_info->tree_mod_seq_wait)) > - wake_up(&fs_info->tree_mod_seq_wait); > + if (!need_ref_seq(for_cow, ref_root)) { > + smp_mb(); > + if (waitqueue_active(&fs_info->tree_mod_seq_wait)) > + wake_up(&fs_info->tree_mod_seq_wait); > + } > + > spin_unlock(&delayed_refs->lock); > if (need_ref_seq(for_cow, ref_root)) > btrfs_qgroup_record_ref(trans, &ref->node, extent_op); > @@ -713,9 +716,11 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info, > add_delayed_data_ref(fs_info, trans, &ref->node, bytenr, > num_bytes, parent, ref_root, owner, offset, > action, for_cow); > - if (!need_ref_seq(for_cow, ref_root) && > - waitqueue_active(&fs_info->tree_mod_seq_wait)) > - wake_up(&fs_info->tree_mod_seq_wait); > + if (!need_ref_seq(for_cow, ref_root)) { > + smp_mb(); > + if (waitqueue_active(&fs_info->tree_mod_seq_wait)) > + wake_up(&fs_info->tree_mod_seq_wait); > + } > spin_unlock(&delayed_refs->lock); > if (need_ref_seq(for_cow, ref_root)) > btrfs_qgroup_record_ref(trans, &ref->node, extent_op); > @@ -744,6 +749,7 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info, > num_bytes, BTRFS_UPDATE_DELAYED_HEAD, > extent_op->is_data); > > + smp_mb(); > if (waitqueue_active(&fs_info->tree_mod_seq_wait)) > wake_up(&fs_info->tree_mod_seq_wait); > spin_unlock(&delayed_refs->lock); > diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c > index 502b20c..a355c89 100644 > --- a/fs/btrfs/disk-io.c > +++ b/fs/btrfs/disk-io.c > @@ -756,9 +756,11 @@ static void run_one_async_done(struct btrfs_work *work) > > atomic_dec(&fs_info->nr_async_submits); > > - if (atomic_read(&fs_info->nr_async_submits) < limit && > - waitqueue_active(&fs_info->async_submit_wait)) > - wake_up(&fs_info->async_submit_wait); > + if (atomic_read(&fs_info->nr_async_submits) < limit) { > + smp_mb(); > + if (waitqueue_active(&fs_info->async_submit_wait)) > + wake_up(&fs_info->async_submit_wait); > + } > > /* If an error occured we just want to clean up the bio and move on */ > if (async->error) { > @@ -3785,14 +3787,17 @@ int btrfs_cleanup_transaction(struct btrfs_root *root) > /* FIXME: cleanup wait for commit */ > t->in_commit = 1; > t->blocked = 1; > + smp_mb(); > if (waitqueue_active(&root->fs_info->transaction_blocked_wait)) > wake_up(&root->fs_info->transaction_blocked_wait); > > t->blocked = 0; > + smp_mb(); > if (waitqueue_active(&root->fs_info->transaction_wait)) > wake_up(&root->fs_info->transaction_wait); > > t->commit_done = 1; > + smp_mb(); > if (waitqueue_active(&t->commit_wait)) > wake_up(&t->commit_wait); > > diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c > index 4b82ae2..acea7d9 100644 > --- a/fs/btrfs/inode.c > +++ b/fs/btrfs/inode.c > @@ -1010,9 +1010,11 @@ static noinline void async_cow_submit(struct btrfs_work *work) > atomic_sub(nr_pages, &root->fs_info->async_delalloc_pages); > > if (atomic_read(&root->fs_info->async_delalloc_pages) < > - 5 * 1024 * 1024 && > - waitqueue_active(&root->fs_info->async_submit_wait)) > - wake_up(&root->fs_info->async_submit_wait); > + 5 * 1024 * 1024) { > + smp_mb(); > + if (waitqueue_active(&root->fs_info->async_submit_wait)) > + wake_up(&root->fs_info->async_submit_wait); > + } > > if (async_cow->inode) > submit_compressed_extents(async_cow->inode, async_cow); > diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c > index b8708f9..871f43f 100644 > --- a/fs/btrfs/volumes.c > +++ b/fs/btrfs/volumes.c > @@ -229,9 +229,11 @@ loop_lock: > cur->bi_next = NULL; > atomic_dec(&fs_info->nr_async_bios); > > - if (atomic_read(&fs_info->nr_async_bios) < limit && > - waitqueue_active(&fs_info->async_submit_wait)) > - wake_up(&fs_info->async_submit_wait); > + if (atomic_read(&fs_info->nr_async_bios) < limit) { > + smp_mb(); > + if (waitqueue_active(&fs_info->async_submit_wait)) > + wake_up(&fs_info->async_submit_wait); > + } > > BUG_ON(atomic_read(&cur->bi_cnt) == 0); > > -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Aug 02, 2012 at 04:46:44AM -0600, Liu Bo wrote: > On 08/02/2012 04:25 AM, Josef Bacik wrote: > > We need an smb_mb() before waitqueue_active to avoid missing wakeups. > > Before Mitch was hitting a deadlock between the ordered flushers and the > > transaction commit because the ordered flushers were waiting for more refs > > and were never woken up, so those smp_mb()'s are the most important. > > Everything else I added for correctness sake and to avoid getting bitten by > > this again somewhere else. Thanks, > > > > Hi Josef, > > I'll appreciate a lot if you can add some comments for each memory > barrier, because not everyone knows why it is used here and there. :) > I'm not going to add comments to all those places, you need a memory barrier in places you don't have an implicit barrier before you do waitqueue_active because you could miss somebody being added to the waitqueue, it's just basic correctness. Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Aug 2, 2012 at 4:46 AM, Liu Bo <liub.liubo@gmail.com> wrote: > On 08/02/2012 04:25 AM, Josef Bacik wrote: >> We need an smb_mb() before waitqueue_active to avoid missing wakeups. >> Before Mitch was hitting a deadlock between the ordered flushers and the >> transaction commit because the ordered flushers were waiting for more refs >> and were never woken up, so those smp_mb()'s are the most important. >> Everything else I added for correctness sake and to avoid getting bitten by >> this again somewhere else. Thanks, >> > > Hi Josef, > > I'll appreciate a lot if you can add some comments for each memory > barrier, because not everyone knows why it is used here and there. :) Everyone who wants to know should read the memory-barriers.txt file that's hiding in the oddly named "Documentation" folder of their kernel tree. :) -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Aug 02, 2012 at 08:11:58AM -0400, Josef Bacik wrote: > On Thu, Aug 02, 2012 at 04:46:44AM -0600, Liu Bo wrote: > > On 08/02/2012 04:25 AM, Josef Bacik wrote: > > > We need an smb_mb() before waitqueue_active to avoid missing wakeups. > > > Before Mitch was hitting a deadlock between the ordered flushers and the > > > transaction commit because the ordered flushers were waiting for more refs > > > and were never woken up, so those smp_mb()'s are the most important. > > > Everything else I added for correctness sake and to avoid getting bitten by > > > this again somewhere else. Thanks, > > > > I'll appreciate a lot if you can add some comments for each memory > > barrier, because not everyone knows why it is used here and there. :) > > I'm not going to add comments to all those places, you need a memory barrier in > places you don't have an implicit barrier before you do waitqueue_active because > you could miss somebody being added to the waitqueue, it's just basic > correctness. Thanks, This asks for a helper: + smp_mb(); + if (waitqueue_active(&fs_info->async_submit_wait)) + wake_up(&fs_info->async_submit_wait); -> void wake_up_if_active(wait) { /* * the comment */ smp_mb(); if(waitqueue_active(wait) wake_up(wait); } -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Aug 1, 2012 at 7:21 PM, Mitch Harder <mitch.harder@sabayonlinux.org> wrote: > On Wed, Aug 1, 2012 at 3:25 PM, Josef Bacik <jbacik@fusionio.com> wrote: >> We need an smb_mb() before waitqueue_active to avoid missing wakeups. >> Before Mitch was hitting a deadlock between the ordered flushers and the >> transaction commit because the ordered flushers were waiting for more refs >> and were never woken up, so those smp_mb()'s are the most important. >> Everything else I added for correctness sake and to avoid getting bitten by >> this again somewhere else. Thanks, >> > > This patch seems to make it tougher to hit a deadlock, but I'm still > encountering intermittent deadlocks using this patch when running > multiple rsync threads. > > I've also tested "Patch 2", and that has me hitting a deadlock even > quicker (when starting several copying threads). > > I also found a slight performance hit using this patch. On a 3.4.6 > kernel (merged with the 3.5_rc for-linus branch), I would typically > complete my rsync test in ~265 seconds. Also, I can't recall hitting > a deadlock on the 3.4.6 kernel (with 3.5_rc for-linus). When using > this patch, the test would take ~310 seconds (when it didn't hit a > deadlock). > I've bisected my deadlock back to: Btrfs: hooks for qgroup to record delayed refs (commit 546adb0d). This issue may be the same problem Alexander Block is discussing in another thread on the Btrfs Mailing List: http://article.gmane.org/gmane.comp.file-systems.btrfs/19028 I'm using multiple rsync threads instead of the new send/receive function. But we're both hitting deadlocks that bisect back to the same commit. -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08/03/2012 04:43 PM, Mitch Harder wrote: > On Wed, Aug 1, 2012 at 7:21 PM, Mitch Harder > <mitch.harder@sabayonlinux.org> wrote: >> On Wed, Aug 1, 2012 at 3:25 PM, Josef Bacik <jbacik@fusionio.com> wrote: >>> We need an smb_mb() before waitqueue_active to avoid missing wakeups. >>> Before Mitch was hitting a deadlock between the ordered flushers and the >>> transaction commit because the ordered flushers were waiting for more refs >>> and were never woken up, so those smp_mb()'s are the most important. >>> Everything else I added for correctness sake and to avoid getting bitten by >>> this again somewhere else. Thanks, >>> >> >> This patch seems to make it tougher to hit a deadlock, but I'm still >> encountering intermittent deadlocks using this patch when running >> multiple rsync threads. >> >> I've also tested "Patch 2", and that has me hitting a deadlock even >> quicker (when starting several copying threads). >> >> I also found a slight performance hit using this patch. On a 3.4.6 >> kernel (merged with the 3.5_rc for-linus branch), I would typically >> complete my rsync test in ~265 seconds. Also, I can't recall hitting >> a deadlock on the 3.4.6 kernel (with 3.5_rc for-linus). When using >> this patch, the test would take ~310 seconds (when it didn't hit a >> deadlock). >> > > I've bisected my deadlock back to: > Btrfs: hooks for qgroup to record delayed refs (commit 546adb0d). > I've got it reproduced here and, I think, nailed it down. I'll send a patch tomorrow after discussing it with Jan. -Arne > This issue may be the same problem Alexander Block is discussing in > another thread on the Btrfs Mailing List: > http://article.gmane.org/gmane.comp.file-systems.btrfs/19028 > > I'm using multiple rsync threads instead of the new send/receive > function. But we're both hitting deadlocks that bisect back to the > same commit. > -- > To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 86eff48..43d1c5a 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -818,6 +818,7 @@ static void free_workspace(int type, struct list_head *workspace) btrfs_compress_op[idx]->free_workspace(workspace); atomic_dec(alloc_workspace); wake: + smp_mb(); if (waitqueue_active(workspace_wait)) wake_up(workspace_wait); } diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 335605c..8cc9b19 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -513,9 +513,11 @@ static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) rb_erase(&delayed_item->rb_node, root); delayed_item->delayed_node->count--; atomic_dec(&delayed_root->items); - if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND && - waitqueue_active(&delayed_root->wait)) - wake_up(&delayed_root->wait); + if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) { + smp_mb(); + if (waitqueue_active(&delayed_root->wait)) + wake_up(&delayed_root->wait); + } } static void btrfs_release_delayed_item(struct btrfs_delayed_item *item) @@ -1057,9 +1059,11 @@ static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) delayed_root = delayed_node->root->fs_info->delayed_root; atomic_dec(&delayed_root->items); if (atomic_read(&delayed_root->items) < - BTRFS_DELAYED_BACKGROUND && - waitqueue_active(&delayed_root->wait)) - wake_up(&delayed_root->wait); + BTRFS_DELAYED_BACKGROUND) { + smp_mb(); + if (waitqueue_active(&delayed_root->wait)) + wake_up(&delayed_root->wait); + } } } diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index da7419e..858ef02 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -662,9 +662,12 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info, add_delayed_tree_ref(fs_info, trans, &ref->node, bytenr, num_bytes, parent, ref_root, level, action, for_cow); - if (!need_ref_seq(for_cow, ref_root) && - waitqueue_active(&fs_info->tree_mod_seq_wait)) - wake_up(&fs_info->tree_mod_seq_wait); + if (!need_ref_seq(for_cow, ref_root)) { + smp_mb(); + if (waitqueue_active(&fs_info->tree_mod_seq_wait)) + wake_up(&fs_info->tree_mod_seq_wait); + } + spin_unlock(&delayed_refs->lock); if (need_ref_seq(for_cow, ref_root)) btrfs_qgroup_record_ref(trans, &ref->node, extent_op); @@ -713,9 +716,11 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info, add_delayed_data_ref(fs_info, trans, &ref->node, bytenr, num_bytes, parent, ref_root, owner, offset, action, for_cow); - if (!need_ref_seq(for_cow, ref_root) && - waitqueue_active(&fs_info->tree_mod_seq_wait)) - wake_up(&fs_info->tree_mod_seq_wait); + if (!need_ref_seq(for_cow, ref_root)) { + smp_mb(); + if (waitqueue_active(&fs_info->tree_mod_seq_wait)) + wake_up(&fs_info->tree_mod_seq_wait); + } spin_unlock(&delayed_refs->lock); if (need_ref_seq(for_cow, ref_root)) btrfs_qgroup_record_ref(trans, &ref->node, extent_op); @@ -744,6 +749,7 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info, num_bytes, BTRFS_UPDATE_DELAYED_HEAD, extent_op->is_data); + smp_mb(); if (waitqueue_active(&fs_info->tree_mod_seq_wait)) wake_up(&fs_info->tree_mod_seq_wait); spin_unlock(&delayed_refs->lock); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 502b20c..a355c89 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -756,9 +756,11 @@ static void run_one_async_done(struct btrfs_work *work) atomic_dec(&fs_info->nr_async_submits); - if (atomic_read(&fs_info->nr_async_submits) < limit && - waitqueue_active(&fs_info->async_submit_wait)) - wake_up(&fs_info->async_submit_wait); + if (atomic_read(&fs_info->nr_async_submits) < limit) { + smp_mb(); + if (waitqueue_active(&fs_info->async_submit_wait)) + wake_up(&fs_info->async_submit_wait); + } /* If an error occured we just want to clean up the bio and move on */ if (async->error) { @@ -3785,14 +3787,17 @@ int btrfs_cleanup_transaction(struct btrfs_root *root) /* FIXME: cleanup wait for commit */ t->in_commit = 1; t->blocked = 1; + smp_mb(); if (waitqueue_active(&root->fs_info->transaction_blocked_wait)) wake_up(&root->fs_info->transaction_blocked_wait); t->blocked = 0; + smp_mb(); if (waitqueue_active(&root->fs_info->transaction_wait)) wake_up(&root->fs_info->transaction_wait); t->commit_done = 1; + smp_mb(); if (waitqueue_active(&t->commit_wait)) wake_up(&t->commit_wait); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4b82ae2..acea7d9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1010,9 +1010,11 @@ static noinline void async_cow_submit(struct btrfs_work *work) atomic_sub(nr_pages, &root->fs_info->async_delalloc_pages); if (atomic_read(&root->fs_info->async_delalloc_pages) < - 5 * 1024 * 1024 && - waitqueue_active(&root->fs_info->async_submit_wait)) - wake_up(&root->fs_info->async_submit_wait); + 5 * 1024 * 1024) { + smp_mb(); + if (waitqueue_active(&root->fs_info->async_submit_wait)) + wake_up(&root->fs_info->async_submit_wait); + } if (async_cow->inode) submit_compressed_extents(async_cow->inode, async_cow); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index b8708f9..871f43f 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -229,9 +229,11 @@ loop_lock: cur->bi_next = NULL; atomic_dec(&fs_info->nr_async_bios); - if (atomic_read(&fs_info->nr_async_bios) < limit && - waitqueue_active(&fs_info->async_submit_wait)) - wake_up(&fs_info->async_submit_wait); + if (atomic_read(&fs_info->nr_async_bios) < limit) { + smp_mb(); + if (waitqueue_active(&fs_info->async_submit_wait)) + wake_up(&fs_info->async_submit_wait); + } BUG_ON(atomic_read(&cur->bi_cnt) == 0);
We need an smb_mb() before waitqueue_active to avoid missing wakeups. Before Mitch was hitting a deadlock between the ordered flushers and the transaction commit because the ordered flushers were waiting for more refs and were never woken up, so those smp_mb()'s are the most important. Everything else I added for correctness sake and to avoid getting bitten by this again somewhere else. Thanks, Signed-off-by: Josef Bacik <jbacik@fusionio.com> --- fs/btrfs/compression.c | 1 + fs/btrfs/delayed-inode.c | 16 ++++++++++------ fs/btrfs/delayed-ref.c | 18 ++++++++++++------ fs/btrfs/disk-io.c | 11 ++++++++--- fs/btrfs/inode.c | 8 +++++--- fs/btrfs/volumes.c | 8 +++++--- 6 files changed, 41 insertions(+), 21 deletions(-)