Re: [PATCH v4 3/7] btrfs: split RAID stripes on deletion

[Qu Wenruo]

在 2024/7/8 20:22, Johannes Thumshirn 写道:
> On 08.07.24 07:26, Johannes Thumshirn wrote:
> > On 08.07.24 07:20, Qu Wenruo wrote:
> > > Can the ASSERT() be reproduced without a zoned device? (I'm really not a
> > > fan of the existing tcmu emulated solution, meanwhile libvirt still
> > > doesn't support ZNS devices)
> > >
> > > If it can be reproduced just with RST feature, I may provide some help
> > > digging into the ASSERT().
> >
> > Let me check. It's very sporadic as well unfortunately.
>
> OK, I've managed to trigger the failure with btrfs/070 on a
> SCRATCH_DEV_POOL with 5 non-zoned devices.

I'm hitting errors like this:

[  227.898320] ------------[ cut here ]------------
[  227.898817] BTRFS: Transaction aborted (error -17)
[  227.899250] WARNING: CPU: 7 PID: 65 at
fs/btrfs/raid-stripe-tree.c:116 btrfs_insert_raid_extent+0x337/0x3d0 [btrfs]
[  227.900122] Modules linked in: btrfs blake2b_generic xor
zstd_compress vfat fat intel_rapl_msr intel_rapl_common crct10dif_pclmul
crc32_pclmul ghash_clmulni_intel iTCO_wdt iTCO_vendor_support
aesni_intel crypto_simd cryptd psmouse i2c_i801 pcspkr i2c_smbus lpc_ich
intel_agp intel_gtt joydev agpgart mousedev raid6_pq libcrc32c loop drm
fuse qemu_fw_cfg ext4 crc32c_generic crc16 mbcache jbd2 dm_mod
virtio_rng virtio_net virtio_blk virtio_balloon net_failover
virtio_console failover virtio_scsi rng_core dimlib usbhid virtio_pci
virtio_pci_legacy_dev crc32c_intel virtio_pci_modern_dev serio_raw
[  227.904452] CPU: 7 PID: 65 Comm: kworker/u40:0 Not tainted
6.10.0-rc6-custom+ #167
[  227.905220] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS
unknown 2/2/2022
[  227.905827] Workqueue: btrfs-endio-write btrfs_work_helper [btrfs]
[  227.906558] RIP: 0010:btrfs_insert_raid_extent+0x337/0x3d0 [btrfs]
[  227.907246] Code: 89 6b 08 e8 4b 18 f7 ff 49 8b 84 24 50 02 00 00 4c
39 f0 75 be 31 db e9 7d fe ff ff 89 de 48 c7 c7 f0 8d 9d a0 e8 29 a1 79
e0 <0f> 0b e9 69 ff ff ff e8 bd 95 3e e1 49 8b 46 60 48 05 48 1a 00 00
[  227.908277] BTRFS: error (device dm-3 state A) in
btrfs_insert_one_raid_extent:116: errno=-17 Object already exists
[  227.909356] RSP: 0018:ffffc9000026fca0 EFLAGS: 00010282
[  227.909361] RAX: 0000000000000000 RBX: 00000000ffffffef RCX:
0000000000000027
[  227.911934] RDX: ffff88817bda1948 RSI: 0000000000000001 RDI:
ffff88817bda1940
[  227.912722] RBP: ffff8881029dcbe0 R08: 0000000000000000 R09:
0000000000000003
[  227.913095] BTRFS info (device dm-3 state EA): forced readonly
[  227.913569] R10: ffffc9000026fb38 R11: ffffffff826d0508 R12:
0000000000000010
[  227.915182] R13: ffff8881029dcbe0 R14: ffff88812a5ff790 R15:
ffff8881488f2780
[  227.916130] FS:  0000000000000000(0000) GS:ffff88817bd80000(0000)
knlGS:0000000000000000
[  227.916912] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  227.917500] CR2: 0000561364dec000 CR3: 00000001583ca000 CR4:
0000000000750ef0
[  227.918210] PKRU: 55555554
[  227.918484] Call Trace:
[  227.918727]  <TASK>
[  227.918940]  ? __warn+0x8c/0x180
[  227.919299]  ? btrfs_insert_raid_extent+0x337/0x3d0 [btrfs]
[  227.919891]  ? report_bug+0x164/0x190
[  227.920272]  ? prb_read_valid+0x1b/0x30
[  227.920666]  ? handle_bug+0x3c/0x80
[  227.921013]  ? exc_invalid_op+0x17/0x70
[  227.921397]  ? asm_exc_invalid_op+0x1a/0x20
[  227.921835]  ? btrfs_insert_raid_extent+0x337/0x3d0 [btrfs]
[  227.922440]  btrfs_finish_one_ordered+0x3c3/0xaa0 [btrfs]
[  227.923055]  ? srso_alias_return_thunk+0x5/0xfbef5
[  227.923549]  ? srso_alias_return_thunk+0x5/0xfbef5
[  227.924062]  btrfs_work_helper+0x107/0x4c0 [btrfs]
[  227.924612]  ? lock_is_held_type+0x9a/0x110
[  227.925040]  process_one_work+0x212/0x720
[  227.925454]  ? srso_alias_return_thunk+0x5/0xfbef5
[  227.926010]  worker_thread+0x1dc/0x3b0
[  227.926411]  ? __pfx_worker_thread+0x10/0x10
[  227.926918]  kthread+0xe0/0x110
[  227.927377]  ? __pfx_kthread+0x10/0x10
[  227.927776]  ret_from_fork+0x31/0x50
[  227.928151]  ? __pfx_kthread+0x10/0x10
[  227.928564]  ret_from_fork_asm+0x1a/0x30
[  227.929035]  </TASK>
[  227.929305] irq event stamp: 11077
[  227.929710] hardirqs last  enabled at (11085): [<ffffffff8115daf5>]
console_unlock+0x135/0x160
[  227.930725] hardirqs last disabled at (11094): [<ffffffff8115dada>]
console_unlock+0x11a/0x160
[  227.931730] softirqs last  enabled at (10728): [<ffffffff810b4684>]
__irq_exit_rcu+0x84/0xa0
[  227.932568] softirqs last disabled at (10723): [<ffffffff810b4684>]
__irq_exit_rcu+0x84/0xa0
[  227.933494] ---[ end trace 0000000000000000 ]---
[  227.933992] BTRFS: error (device dm-3 state EA) in
btrfs_insert_one_raid_extent:116: errno=-17 Object already exists
[  227.935193] BTRFS warning (device dm-3 state EA): Skipping commit of
aborted transaction.
[  227.936383] BTRFS: error (device dm-3 state EA) in
cleanup_transaction:2018: errno=-17 Object already exists

But not that ASSERT() yet.

I guess I need the first patch to pass this first?

Thanks,
Qu