There are some problems in TCP/IP in the 2.2.10 kernel which have been
causing kernel panics on several of our SMP nodes at the rate of one
or two per day. These bugs have been apparently been fixed in
2.2.12.
The symptom is this:
Aug 28 19:04:27 rip7 kernel: Unable to handle kernel NULL pointer dereference at
virtual address 00000000
Aug 28 19:04:27 rip7 kernel: current->tss.cr3 = 0076a000, %cr3 = 0076a000
Aug 28 19:04:27 rip7 kernel: *pde = 00000000
Aug 28 19:04:27 rip7 kernel: Oops: 0000
Aug 28 19:04:27 rip7 kernel: CPU: 0
Aug 28 19:04:27 rip7 kernel: EIP: 0010:[tcp_v4_rehash+203/328]
Aug 28 19:04:27 rip7 kernel: EFLAGS: 00010202
Aug 28 19:04:27 rip7 kernel: eax: 000001b0 ebx: c24639c0 ecx: c1c51db0 edx
: 00000000
Aug 28 19:04:27 rip7 kernel: esi: 00000004 edi: 00000004 ebp: bfffe98c esp
: c1c59f6c
Aug 28 19:04:27 rip7 kernel: ds: 0018 es: 0018 ss: 0018
Aug 28 19:04:27 rip7 kernel: Process python (pid: 6528, process nr: 63, stackpag
e=c1c59000)
Aug 28 19:04:27 rip7 kernel: Stack: c018122d c24639c0 c1d023cc c1c58000 c0164f6b
c1d023cc 00000004 00000003
Aug 28 19:04:27 rip7 kernel: 080c5760 00000000 c0165c0f c0165cbf 0000000e
00000004 c1c58000 0000000e