This is the mail archive of the mailing list for the systemtap project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

S390x Weekly testing

Some bad luck this week. I am hitting what looks like a kprobes bug on the 2.6.21-rc3.
The system is crashing in the stap prologue test.

illegal operation: 0001 Ý#29¨
CPU: 3 Not tainted
Process find (pid: 1571, task: 000000003fc28718, ksp: 000000003a357c70)
Krnl PSW : 0704200180000000 000000000018d844 (sys_write+0x24/0x80)
Krnl GPRS: 0000000000000000 000002000018fd38 0000000000000001 0000020000020000
0000000000000048 000002000018a8a8 0000000000000048 0000000000000010
000000000018d820 000000003a356000 0000020000020000 fffffffffffffff7
0000000000000048 000000000010e408 000000003a357eb8 000000003a357e70
Krnl Code: f0 98 00 24 41 30 f0 ac c0 e5 00 00 03 94 b9 04 00 4c b9 02
Call Trace:
(Ý<040000003a357e78>¨ 0x40000003a357e78)

The crashing PSW (ip) is in the middle of the probed instruction.
Digging into it..

000000000018d820 <sys_write>:
18d820: eb af f0 68 00 24 stmg %r10,%r15,104(%r15)
18d826: a7 f1 1f 00 tml %r15,7936
18d82a: b9 04 00 ef lgr %r14,%r15
18d82e: a7 84 00 01 je 18d830 <sys_write+0x10>
18d832: a7 fb ff b8 aghi %r15,-72
18d836: b9 04 00 a3 lgr %r10,%r3
18d83a: b9 04 00 c4 lgr %r12,%r4
18d83e: a7 b9 ff f7 lghi %r11,-9
>> 18d842: e3 e0 f0 98 00 24 stg %r14,152(%r15) << this is the probe point
18d848: 41 30 f0 ac la %r3,172(%r15)
18d84c: c0 e5 00 00 03 94 brasl %r14,18df74 <fget_light>
18d852: b9 04 00 4c lgr %r4,%r12

David Wilder
IBM Linux Technology Center
Beaverton, Oregon, USA

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]