[BUG] softirq.c:229 - LTP nightly bk testing

[Paul Larson]

In the nightly testing that Linux Test Project does on bitkeeper trees,
I noticed that one of the two machines failed to reboot to run last
nights test.  The UP box came up and worked fine, but the 2-way PIII-550
had this panic on boot:

kernel BUG at softirq.c:229!
invalid operand: 0000
CPU:    1
EIP:    0060:[<c011cbad>]       Not tainted
EFLAGS: 00010246
eax: 00000000   ebx: c0454034   ecx: c265a000   edx: c03f20cc
esi: c0453f80   edi: c265a000   ebp: c0436560   esp: c265bfb0
ds: 0068   es: 0068   ss: 0068
Process ksoftirqd_CPU1 (pid: 5, threadinfo=c265a000 task=c2658000)
Stack: 00000001 c0432960 fffffffe 00000020 c011c8fa c0432960 c265a000
00000020
       c265a000 c0453b80 00000246 c011ce56
10 50 8b 43 0c ff d0 83 c4 04
 <0>Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing

Changesets between last test and this one were 1.519.3.1 through 1.558.