Hi,
On <2008...@valinux.co.jp>,
Isaku Yamahata wrote:
Hmm, I'm afraind that c/s 18090:1e5d42cf61ec may cause the issue.
Could you please test again without c/s 18090:1e5d42cf61ec.
Without 18090:1e5d42cf61ec, hypervisor still crashes.
Next, I will try your patch.
I attached the panic message on Tiger4.
The source code is:
static __inline__ int
__test_and_clear_bit(int nr, volatile void * addr)
{
__u32 *p = (__u32 *) addr + (nr >> 5);
__u32 m = 1 << (nr & 31);
f40000000405fae6: 10 01 3c 20 20 00 ld4 r17=[r15]
f40000000405faec: 00 00 00 20 nop.b 0x0;;
f40000000405faf0: 10 98 7c 22 2c 20 [MIB] and r19=31,r17
f40000000405faf6: 20 59 44 34 29 00 extr r18=r17,5,27
f40000000405fafc: 00 00 00 20 nop.b 0x0
f40000000405fb00: 02 78 00 1c 08 10 [MII] ld2 r15=[r14]
f40000000405fb06: e0 f8 03 fe 49 c0 mov r14=32767;;
f40000000405fb0c: f0 70 1c e2 cmp4.eq
p6,p7=r15,r14
Best regards,