[LU-14698] sanityn test_77b scheduling in_atomic()

• Type: Bug
• Resolution: Unresolved
• Priority: Minor
• Affects Version/s: Lustre 2.15.0
• Severity: 3
• Assignee: WC Triage
• Reporter: Andreas Dilger

Description

Seen in the following test runs on the server console, but likely has existed for a long time.

https://testing-archive.whamcloud.com/gerrit-janitor/16001/testresults/sanityn-ldiskfs-DNE-centos7_x86_64-centos7_x86_64/
https://testing-archive.whamcloud.com/gerrit-janitor/16001/testresults/sanityn-ldiskfs-centos7_x86_64-centos7_x86_64/

May 21 17:05:54 oleg242-server kernel: BUG: sleeping function called from invalid context at kernel/workqueue.c:2925
May 21 17:05:54 oleg242-server kernel: in_atomic(): 1, irqs_disabled(): 0, pid: 21346, name: lctl
May 21 17:05:54 oleg242-server kernel: CPU: 3 PID: 21346 Comm: lctl Kdump: loaded Tainted: P           OE  ------------   3.10.0-7.9-debug #1
May 21 17:05:54 oleg242-server kernel: Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
May 21 17:05:54 oleg242-server kernel: Call Trace:
May 21 17:05:54 oleg242-server kernel: [<ffffffff817da09e>] dump_stack+0x19/0x1b
May 21 17:05:54 oleg242-server kernel: [<ffffffff810c9049>] __might_sleep+0xd9/0x100
May 21 17:05:54 oleg242-server kernel: [<ffffffff810b1218>] flush_work+0x48/0x1f0
May 21 17:05:54 oleg242-server kernel: [<ffffffff810b1d51>] __cancel_work_timer+0xa1/0x150
May 21 17:05:54 oleg242-server kernel: [<ffffffff810b1e10>] cancel_work_sync+0x10/0x20
May 21 17:05:54 oleg242-server kernel: [<ffffffff8140cc53>] rhashtable_free_and_destroy+0x23/0x150
May 21 17:05:54 oleg242-server kernel: [<ffffffffa0693437>] nrs_crrn_stop+0x57/0x2e0 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffffa06431f9>] nrs_policy_stop0+0x39/0x1c0 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffffa0643a08>] nrs_policy_stop_primary.isra.8+0x78/0x1f0 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffffa0644924>] nrs_policy_start_locked+0x624/0x670 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffffa0644a9b>] nrs_policy_ctl+0x12b/0x2b0 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffffa06472df>] ptlrpc_nrs_policy_control+0x10f/0x2d0 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffffa0629293>] ptlrpc_lprocfs_nrs_seq_write+0x423/0x630 [ptlrpc]
May 21 17:05:54 oleg242-server kernel: [<ffffffff81244876>] vfs_write+0xd6/0x230
May 21 17:05:54 oleg242-server kernel: [<ffffffff812456bf>] SyS_write+0x7f/0xf0

Issue Links

• is related to: LU-14364 Switching QoS from tbf uid to fifo caused soft lockup
• is related to: LU-8130 Migrate from libcfs hash to rhashtable

Comments

[Alex Zhuravlev]

still hitting this with master branch:

[   34.360117] BUG: sleeping function called from invalid context at kernel/workqueue.c:3092
[   34.360279] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 7706, name: lctl
..
[   34.361026]  rhashtable_free_and_destroy+0x20/0x360
[   34.361057]  ? lock_acquire+0x1ac/0x3f0
[   34.361076]  nrs_crrn_stop+0x4e/0x1e0 [ptlrpc]
[   34.361149]  nrs_policy_stop0+0x30/0x1b0 [ptlrpc]
[   34.361220]  nrs_policy_stop_primary.isra.3+0x17c/0x1c0 [ptlrpc]
[   34.361296]  nrs_policy_start_locked+0x374/0x5d0 [ptlrpc]
[   34.361367]  nrs_policy_ctl+0x183/0x280 [ptlrpc]
[   34.361468]  ptlrpc_nrs_policy_control+0x10d/0x2b0 [ptlrpc]
[   34.361550]  ptlrpc_lprocfs_nrs_seq_write+0x42c/0x570 [ptlrpc]
[   34.361636]  full_proxy_write+0x4b/0x70
[   34.361684]  vfs_write+0xb9/0x1c0
[   34.361707]  ksys_write+0x3d/0xa0
[   34.361727]  do_syscall_64+0x4b/0x1b0