mbox series

[v4,00/22] btrfs: async discard support

Message ID cover.1574709825.git.dennis@kernel.org (mailing list archive)
Headers show
Series btrfs: async discard support | expand

Message

Dennis Zhou Nov. 25, 2019, 7:46 p.m. UTC 
Hello,

v3 [1] had 2 minor issues which are fixed:
 - I was generically dividing u64 which made 32 bit arches unhappy. [2]
 - Uninitialized use of trim_state local variable [3]

I've gone through and fixed the apparent u64 divisions which passes the
make.cross build provided on the end of the series.

This is based on btrfs-devel#misc-next e88411ecf657.

[1] https://lore.kernel.org/linux-btrfs/cover.1574282259.git.dennis@kernel.org/
[2] https://lore.kernel.org/linux-btrfs/201911230623.j5g9qeNZ%25lkp@intel.com/
[3] https://lore.kernel.org/linux-btrfs/CAKwvOdn5j37AYzmoOsaSqyYdBkjqevbTrSyGQypB+G_NgxX0fQ@mail.gmail.com/

diffstats below:

Dennis Zhou (22):
  bitmap: genericize percpu bitmap region iterators
  btrfs: rename DISCARD opt to DISCARD_SYNC
  btrfs: keep track of which extents have been discarded
  btrfs: keep track of cleanliness of the bitmap
  btrfs: add the beginning of async discard, discard workqueue
  btrfs: handle empty block_group removal
  btrfs: discard one region at a time in async discard
  btrfs: add removal calls for sysfs debug/
  btrfs: make UUID/debug have its own kobject
  btrfs: add discard sysfs directory
  btrfs: track discardable extents for async discard
  btrfs: keep track of discardable_bytes
  btrfs: calculate discard delay based on number of extents
  btrfs: add bps discard rate limit
  btrfs: limit max discard size for async discard
  btrfs: make max async discard size tunable
  btrfs: have multiple discard lists
  btrfs: only keep track of data extents for async discard
  btrfs: keep track of discard reuse stats
  btrfs: add async discard header
  btrfs: increase the metadata allowance for the free_space_cache
  btrfs: make smaller extents more likely to go into bitmaps

 fs/btrfs/Makefile           |   2 +-
 fs/btrfs/block-group.c      |  56 ++-
 fs/btrfs/block-group.h      |  30 ++
 fs/btrfs/ctree.h            |  52 ++-
 fs/btrfs/discard.c          | 680 ++++++++++++++++++++++++++++++++++++
 fs/btrfs/discard.h          |  46 +++
 fs/btrfs/disk-io.c          |  15 +-
 fs/btrfs/extent-tree.c      |  23 +-
 fs/btrfs/free-space-cache.c | 587 ++++++++++++++++++++++++++-----
 fs/btrfs/free-space-cache.h |  39 ++-
 fs/btrfs/inode-map.c        |  13 +-
 fs/btrfs/scrub.c            |   7 +-
 fs/btrfs/super.c            |  39 ++-
 fs/btrfs/sysfs.c            | 205 ++++++++++-
 include/linux/bitmap.h      |  35 ++
 mm/percpu.c                 |  61 +---
 16 files changed, 1737 insertions(+), 153 deletions(-)
 create mode 100644 fs/btrfs/discard.c
 create mode 100644 fs/btrfs/discard.h

Thanks,
Dennis

Comments

David Sterba Nov. 26, 2019, 9:52 p.m. UTC | #1 
On Mon, Nov 25, 2019 at 02:46:40PM -0500, Dennis Zhou wrote:
> Hello,
> 
> v3 [1] had 2 minor issues which are fixed:
>  - I was generically dividing u64 which made 32 bit arches unhappy. [2]
>  - Uninitialized use of trim_state local variable [3]

I've started fstests (misc-5.5 + v4) with some debugging options
compiled in and MOUNT_OPTIONS -o discard=async:

btrfs/003:

[   57.263321] BTRFS info (device vdb): relocating block group 30408704 flags metadata|dup
[   57.321441] BTRFS info (device vdb): found 8 extents
[   57.352217] list_add corruption. prev->next should be next (ffff89b6b157f1e0), but was 6b6b6b6b6b6b6b6b. (prev=ffff89b6b3cc59f0).
[   57.355426] ------------[ cut here ]------------
[   57.356895] kernel BUG at lib/list_debug.c:26!
[   57.358303] invalid opcode: 0000 [#1] SMP
[   57.359522] CPU: 1 PID: 19415 Comm: btrfs Not tainted 5.4.0-rc8-default+ #878
[   57.361334] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-rebuilt.opensuse.org 04/01/2014
[   57.364468] RIP: 0010:__list_add_valid+0x5e/0x70
[   57.369959] RSP: 0018:ffff90cbc4c7f8d0 EFLAGS: 00010246
[   57.371475] RAX: 0000000000000075 RBX: ffff89b6b3d04000 RCX: 0000000000000006
[   57.372930] RDX: 0000000000000000 RSI: ffff89b6b2314030 RDI: ffff89b6b2313740
[   57.374893] RBP: ffff89b6b157f0d0 R08: 0000000d5aa62d0d R09: 0000000000000000
[   57.376970] R10: 0000000000000000 R11: 0000000000000000 R12: ffff89b6b3d041f0
[   57.378961] R13: ffff89b6b3cc59f0 R14: ffff89b6b157f190 R15: ffff89b6b157f1e0
[   57.381012] FS:  00007f551b41d8c0(0000) GS:ffff89b6bd800000(0000) knlGS:0000000000000000
[   57.383531] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   57.385264] CR2: 0000563d526212d8 CR3: 0000000073dd3004 CR4: 0000000000160ee0
[   57.387223] Call Trace:
[   57.388250]  btrfs_add_to_discard_unused_list+0xa2/0xe0 [btrfs]
[   57.389804]  btrfs_discard_queue_work+0x3a/0x60 [btrfs]
[   57.390961]  __btrfs_add_free_space+0xe2/0x400 [btrfs]
[   57.401053]  ? do_raw_spin_unlock+0x4b/0xc0
[   57.402419]  ? _raw_spin_unlock+0x24/0x30
[   57.403791]  ? find_first_extent_bit+0xdc/0x120 [btrfs]
[   57.405312]  add_new_free_space+0xd1/0xf0 [btrfs]
[   57.406723]  btrfs_make_block_group+0x9f/0x2e0 [btrfs]
[   57.408222]  __btrfs_alloc_chunk+0x705/0xf00 [btrfs]
[   57.409786]  btrfs_chunk_alloc+0x1a9/0x3f0 [btrfs]
[   57.411354]  btrfs_inc_block_group_ro+0x16d/0x190 [btrfs]
[   57.412897]  btrfs_relocate_block_group+0x60/0x300 [btrfs]
[   57.414430]  btrfs_relocate_chunk+0x32/0xd0 [btrfs]
[   57.415747]  __btrfs_balance+0x41c/0xcc0 [btrfs]
[   57.417084]  btrfs_balance+0x56e/0xa40 [btrfs]
[   57.418361]  ? kmem_cache_alloc_trace+0x281/0x2c0
[   57.419448]  btrfs_ioctl_balance+0x298/0x350 [btrfs]
[   57.420568]  ? mem_cgroup_throttle_swaprate+0xcb/0x2a8
[   57.422306]  btrfs_ioctl+0x466/0x2570 [btrfs]
[   57.423648]  ? __handle_mm_fault+0x499/0x740
[   57.425008]  ? kvm_sched_clock_read+0x14/0x30
[   57.426311]  ? sched_clock+0x5/0x10
[   57.427480]  ? sched_clock_cpu+0x10/0x120
[   57.428856]  ? __handle_mm_fault+0x499/0x740
[   57.430314]  ? do_raw_spin_unlock+0x4b/0xc0
[   57.431695]  ? _raw_spin_unlock+0x24/0x30
[   57.432997]  ? __handle_mm_fault+0x499/0x740
[   57.434370]  ? do_vfs_ioctl+0x405/0x730
[   57.435920]  do_vfs_ioctl+0x405/0x730
[   57.437439]  ? do_user_addr_fault+0x1f1/0x3c0
[   57.439096]  ksys_ioctl+0x3a/0x70
[   57.440518]  ? trace_hardirqs_off_thunk+0x1a/0x20
[   57.442159]  __x64_sys_ioctl+0x16/0x20
[   57.443315]  do_syscall_64+0x50/0x1f0
[   57.444252]  entry_SYSCALL_64_after_hwframe+0x49/0xbek
Dennis Zhou Nov. 27, 2019, 6:26 p.m. UTC | #2 
On Tue, Nov 26, 2019 at 10:52:04PM +0100, David Sterba wrote:
> On Mon, Nov 25, 2019 at 02:46:40PM -0500, Dennis Zhou wrote:
> > Hello,
> > 
> > v3 [1] had 2 minor issues which are fixed:
> >  - I was generically dividing u64 which made 32 bit arches unhappy. [2]
> >  - Uninitialized use of trim_state local variable [3]
> 
> I've started fstests (misc-5.5 + v4) with some debugging options
> compiled in and MOUNT_OPTIONS -o discard=async:
> 
> btrfs/003:
> 
> [   57.263321] BTRFS info (device vdb): relocating block group 30408704 flags metadata|dup
> [   57.321441] BTRFS info (device vdb): found 8 extents
> [   57.352217] list_add corruption. prev->next should be next (ffff89b6b157f1e0), but was 6b6b6b6b6b6b6b6b. (prev=ffff89b6b3cc59f0).
> [   57.355426] ------------[ cut here ]------------
> [   57.356895] kernel BUG at lib/list_debug.c:26!
> [   57.358303] invalid opcode: 0000 [#1] SMP
> [   57.359522] CPU: 1 PID: 19415 Comm: btrfs Not tainted 5.4.0-rc8-default+ #878
> [   57.361334] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-rebuilt.opensuse.org 04/01/2014
> [   57.364468] RIP: 0010:__list_add_valid+0x5e/0x70
> [   57.369959] RSP: 0018:ffff90cbc4c7f8d0 EFLAGS: 00010246
> [   57.371475] RAX: 0000000000000075 RBX: ffff89b6b3d04000 RCX: 0000000000000006
> [   57.372930] RDX: 0000000000000000 RSI: ffff89b6b2314030 RDI: ffff89b6b2313740
> [   57.374893] RBP: ffff89b6b157f0d0 R08: 0000000d5aa62d0d R09: 0000000000000000
> [   57.376970] R10: 0000000000000000 R11: 0000000000000000 R12: ffff89b6b3d041f0
> [   57.378961] R13: ffff89b6b3cc59f0 R14: ffff89b6b157f190 R15: ffff89b6b157f1e0
> [   57.381012] FS:  00007f551b41d8c0(0000) GS:ffff89b6bd800000(0000) knlGS:0000000000000000
> [   57.383531] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   57.385264] CR2: 0000563d526212d8 CR3: 0000000073dd3004 CR4: 0000000000160ee0
> [   57.387223] Call Trace:
> [   57.388250]  btrfs_add_to_discard_unused_list+0xa2/0xe0 [btrfs]
> [   57.389804]  btrfs_discard_queue_work+0x3a/0x60 [btrfs]
> [   57.390961]  __btrfs_add_free_space+0xe2/0x400 [btrfs]
> [   57.401053]  ? do_raw_spin_unlock+0x4b/0xc0
> [   57.402419]  ? _raw_spin_unlock+0x24/0x30
> [   57.403791]  ? find_first_extent_bit+0xdc/0x120 [btrfs]
> [   57.405312]  add_new_free_space+0xd1/0xf0 [btrfs]
> [   57.406723]  btrfs_make_block_group+0x9f/0x2e0 [btrfs]
> [   57.408222]  __btrfs_alloc_chunk+0x705/0xf00 [btrfs]
> [   57.409786]  btrfs_chunk_alloc+0x1a9/0x3f0 [btrfs]
> [   57.411354]  btrfs_inc_block_group_ro+0x16d/0x190 [btrfs]
> [   57.412897]  btrfs_relocate_block_group+0x60/0x300 [btrfs]
> [   57.414430]  btrfs_relocate_chunk+0x32/0xd0 [btrfs]
> [   57.415747]  __btrfs_balance+0x41c/0xcc0 [btrfs]
> [   57.417084]  btrfs_balance+0x56e/0xa40 [btrfs]
> [   57.418361]  ? kmem_cache_alloc_trace+0x281/0x2c0
> [   57.419448]  btrfs_ioctl_balance+0x298/0x350 [btrfs]
> [   57.420568]  ? mem_cgroup_throttle_swaprate+0xcb/0x2a8
> [   57.422306]  btrfs_ioctl+0x466/0x2570 [btrfs]
> [   57.423648]  ? __handle_mm_fault+0x499/0x740
> [   57.425008]  ? kvm_sched_clock_read+0x14/0x30
> [   57.426311]  ? sched_clock+0x5/0x10
> [   57.427480]  ? sched_clock_cpu+0x10/0x120
> [   57.428856]  ? __handle_mm_fault+0x499/0x740
> [   57.430314]  ? do_raw_spin_unlock+0x4b/0xc0
> [   57.431695]  ? _raw_spin_unlock+0x24/0x30
> [   57.432997]  ? __handle_mm_fault+0x499/0x740
> [   57.434370]  ? do_vfs_ioctl+0x405/0x730
> [   57.435920]  do_vfs_ioctl+0x405/0x730
> [   57.437439]  ? do_user_addr_fault+0x1f1/0x3c0
> [   57.439096]  ksys_ioctl+0x3a/0x70
> [   57.440518]  ? trace_hardirqs_off_thunk+0x1a/0x20
> [   57.442159]  __x64_sys_ioctl+0x16/0x20
> [   57.443315]  do_syscall_64+0x50/0x1f0
> [   57.444252]  entry_SYSCALL_64_after_hwframe+0x49/0xbek

Sigh, I'm sorry about these issues with the series. I think this is
being racy in btrfs_delete_unused_bgs(). I'm working on setting up the
reproducer and fixing it.

Thanks,
Dennis