[LU-8106] kiblnd_pool_alloc_node() crashed because of NULL pointer dereference

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.9.0
• Severity: 3
• Assignee: Doug Oucharek (Inactive)
• Reporter: Doug Oucharek (Inactive)

Description

A customer experienced a crash with this traceback:

<4><d> ffff88100f5dfaa8 ffff88100f5dfab8 0000000000000012 ffff88100f5dfa98
<4><d> ffff8810449c7a80 ffffffffa0654379 ffff881045390000 ffff88100f5dfa00
<4>Call Trace:
<4> [<ffffffffa0654379>] kiblnd_get_idle_tx+0x29/0x2c0 [ko2iblnd]
<4> [<ffffffffa0657a8f>] kiblnd_check_sends_locked+0x2ef/0x4a0 [ko2iblnd]
<4> [<ffffffffa0659f42>] kiblnd_post_rx+0x162/0x540 [ko2iblnd]
<4> [<ffffffffa065a436>] kiblnd_recv+0x116/0x560 [ko2iblnd]
<4> [<ffffffffa05edefb>] lnet_ni_recv+0xbb/0x320 [lnet]
<4> [<ffffffffa05eed71>] lnet_drop_message+0x81/0xa0 [lnet]
<4> [<ffffffffa05f3712>] lnet_parse+0x1b2/0x18d0 [lnet]
<4> [<ffffffff8106faa4>] ? enqueue_task_fair+0x64/0x100
<4> [<ffffffff8106711e>] ? try_to_wake_up+0x24e/0x3e0
<4> [<ffffffffa065ab4b>] kiblnd_handle_rx+0x2cb/0x640 [ko2iblnd]
<4> [<ffffffff810672c2>] ? default_wake_function+0x12/0x20
<4> [<ffffffffa065bc43>] kiblnd_rx_complete+0x2d3/0x420 [ko2iblnd]
<4> [<ffffffffa065bdf2>] kiblnd_complete+0x62/0xe0 [ko2iblnd]
<4> [<ffffffffa065c1aa>] kiblnd_scheduler+0x33a/0x7d0 [ko2iblnd]
<4> [<ffffffff810672b0>] ? default_wake_function+0x0/0x20
<4> [<ffffffffa065be70>] ? kiblnd_scheduler+0x0/0x7d0 [ko2iblnd]
<4> [<ffffffff810a0fce>] kthread+0x9e/0xc0
<4> [<ffffffff8100c28a>] child_rip+0xa/0x20
<4> [<ffffffff810a0f30>] ? kthread+0x0/0xc0
<4> [<ffffffff8100c280>] ? child_rip+0x0/0x20
<4>Code: 8d 67 30 4c 8d 77 10 4c 8d 6d c8 48 89 fb 48 89 df e8 e8 41 ef e0 48 8b 43 30 4c 39 e0 48 89 45 c8 0f 84 87 00 00 00 48 8d 50 10 <48> 39 50 10 74 6d 83 40 30 01 bf e8 03 00 00 4c 8b 65 c8 e8 cb
<1>RIP [<ffffffffa064953d>] kiblnd_pool_alloc_node+0x4d/0x2c0 [ko2iblnd]
<4> RSP <ffff8810449c7a20>
<4>CR2: 0000000000000010

Issue Links

• is related to: LU-15618 ksock_conn ref leak on shutdown

Comments

[Joseph Gmitter (Inactive)]

Patch has landed to master for 2.9.0

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/19993/
Subject: LU-8106 lnet: Do not drop message when shutting down LNet
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 47b7b319783f27023b0cefe54a2a2eea678284f2

[Gerrit Updater]

Doug Oucharek ([e-mail]) uploaded a new patch: http://review.whamcloud.com/19993
Subject: LU-8106 lnet: Do not drop message when shutting down LNet
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 70e7feaa701a8a7192b91d029edcd1fb0dec33aa