Kernel NULL pointer dereference at ipoib_cm_tx_reap+0xc9 / list_del+0x1b
Issue
- An NFS server under a medium to high load with Infiniband network cards has crashed with the kernel NULL pointer dereference at ipoib_cm_tx_reap+0xc9 / list_del+0x1b with the following call trace:
PID: 18499 TASK: ffff8824202b2080 CPU: 0 COMMAND: "ipoib"
#7 [ffff882412d67cd0] do_page_fault at ffffffff8151339e
#8 [ffff882412d67d00] page_fault at ffffffff81510755
[exception RIP: list_del+0x1b]
RIP: ffffffff81288eeb RSP: ffff882412d67db0 RFLAGS: 00010046
RAX: 0000000000000000 RBX: ffff88120ffafad0 RCX: 000000000000af76
RDX: 0000000000000246 RSI: 0000000000000000 RDI: ffff88120ffafad0
RBP: ffff882412d67dc0 R8: ffff88120ffafad0 R9: 0040000000000000
R10: ffffea003b2a59d0 R11: 0000000000000002 R12: ffff88241d564020
R13: 0000000000000246 R14: ffff88241bd7f880 R15: ffff88241d5646e0
ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018
#9 [ffff882412d67dc8] ipoib_cm_tx_reap at ffffffffa037c619 [ib_ipoib]
#10 [ffff882412d67e38] worker_thread at ffffffff81090ae0
#11 [ffff882412d67ee8] kthread at ffffffff81096936
Environment
- Red Hat Enterprise Linux 6.4, kernel-2.6.32-345.el6 and later (including 2.6.32-358.2.1)
- Red Hat Enterprise Linux 6.3, kernel-2.6.32-279.21.1.el6 and later
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.