Process "perfwatch" crashes a Xen Dom0
Issue
Server crashes without any trace. The captured VMcore shows a process perfwatch
that seems to have caused the crash:
crash> bt
PID: 11494 TASK: ffff88001a8337e0 CPU: 0 COMMAND: "perfwatch"
#0 [ffff880016a87e08] schedule at ffffffff80262d67
#1 [ffff880016a87e20] xen_pgd_unpin at ffffffff8027f4ea
#2 [ffff880016a87e50] mm_unpin at ffffffff8027db99
#3 [ffff880016a87e60] _arch_exit_mmap at ffffffff8027df8b
#4 [ffff880016a87ed0] exit_mmap at ffffffff8023b2c1
#5 [ffff880016a87f00] mmput at ffffffff8023d9d1
#6 [ffff880016a87f20] do_exit at ffffffff80215f6c
#7 [ffff880016a87f80] tracesys at ffffffff802602f9 (via system_call)
RIP: 000000368de9a4bf RSP: 00007fff8a8a5a68 RFLAGS: 00000246
RAX: ffffffffffffffda RBX: ffffffff802602f9 RCX: ffffffffffffffff
RDX: 0000000000000000 RSI: 000000000000003c RDI: 0000000000000000
RBP: 000000368e14d870 R8: 00000000000000e7 R9: ffffffffffffffb0
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 000000368e14d870 R14: ffffffff8024af8b R15: 0000000000000070
ORIG_RAX: 00000000000000e7 CS: 0033 SS: e02b
Environment
- Red Hat Enterprise Linux 5.5
- Xen Virtualization
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.