* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
@ 2009-10-20 5:34 ` pavan dot naregundi at in dot ibm dot com
2009-11-12 20:18 ` fche at redhat dot com
` (5 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: pavan dot naregundi at in dot ibm dot com @ 2009-10-20 5:34 UTC (permalink / raw)
To: systemtap
------- Additional Comments From pavan dot naregundi at in dot ibm dot com 2009-10-20 05:33 -------
Created an attachment (id=4302)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4302&action=view)
logs collected by autotest
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
2009-10-20 5:34 ` [Bug runtime/10812] " pavan dot naregundi at in dot ibm dot com
@ 2009-11-12 20:18 ` fche at redhat dot com
2010-01-22 9:20 ` pavan dot naregundi at in dot ibm dot com
` (4 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: fche at redhat dot com @ 2009-11-12 20:18 UTC (permalink / raw)
To: systemtap
------- Additional Comments From fche at redhat dot com 2009-11-12 20:17 -------
If you still have access to a stap*.ko that exhibits this problem, could
you "objdump -drS" it to disassemble it, to see where this memcmp() call
is coming from? (We have one explicit call to it, in buildid checking,
but that doesn't seem THAT close to the write_cmd path.)
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
2009-10-20 5:34 ` [Bug runtime/10812] " pavan dot naregundi at in dot ibm dot com
2009-11-12 20:18 ` fche at redhat dot com
@ 2010-01-22 9:20 ` pavan dot naregundi at in dot ibm dot com
2010-01-22 9:21 ` pavan dot naregundi at in dot ibm dot com
` (3 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: pavan dot naregundi at in dot ibm dot com @ 2010-01-22 9:20 UTC (permalink / raw)
To: systemtap
------- Additional Comments From pavan dot naregundi at in dot ibm dot com 2010-01-22 09:20 -------
Can recreate this with git (2.6.33-rc4-git5 ..6ccf80eb..). Here is the
supporting info.
stap_07505bf65828be510a2a08652a12f45b_4303: systemtap: 1.1/0.144, base:
d0000000099f0000, memory: 29data/40text/145ctx/13net/24856alloc kb, probes: 4
Unable to handle kernel paging request for data at address 0xd000000001d02710
Faulting instruction address: 0xc00000000004460c
Oops: Kernel access of bad area, sig: 11 [#1]
SMP NR_CPUS=1024 DEBUG_PAGEALLOC NUMA pSeries
Modules linked in: stap_5e64ec831c6c1c2fa9784d7081f8e5fa_7859 ipv6 fuse loop
dm_mod sr_mod ibmveth cdrom sg sd_mod crc_t10dif ibmvscsic scsi_transport_srp
scsi_tgt scsi_mod [last unloaded: stap_07505bf65828be510a2a08652a12f45b_4303]
NIP: c00000000004460c LR: d000000009d563d4 CTR: 0000000000000014
REGS: c0000000e6513950 TRAP: 0300 Not tainted (2.6.33-rc4-git5-autotest)
MSR: 8000000000009032 <EE,ME,IR,DR> CR: 44000424 XER: 00000001
DAR: d000000001d02710, DSISR: 0000000040010000
TASK = c0000000ec9fc050[18734] 'stapio' THREAD: c0000000e6510000 CPU: 17
GPR00: 000000000000006b c0000000e6513bd0 c000000000bb1e70 000000000000009d
GPR04: d000000001d0270f 0000000000000014 d000000009d59315 0000000000000000
GPR08: 000000000000e99c d000000009d59200 c000000010001d20 c0000000000445f4
GPR12: d000000009d58160 c000000000c85080 0000000010008f48 00000fffedd0aa60
GPR16: 0000000010022670 0000000010008af0 0000000010008ad0 0000000000000008
GPR20: 0000000010022890 0000000010008ea0 0000000010003954 000000001001b018
GPR24: 00000fffedd0aa64 0000000010022668 d000000009d591f8 d000000009d59b34
GPR28: d000000001d02724 d000000001cf8638 d000000009d629b8 d000000009d5a540
NIP [c00000000004460c] .memcmp+0x18/0x30
LR [d000000009d563d4] ._stp_ctl_write_cmd+0x45c/0x98c
[stap_5e64ec831c6c1c2fa9784d7081f8e5fa_7859]
Call Trace:
[c0000000e6513bd0] [d000000009d562e0] ._stp_ctl_write_cmd+0x368/0x98c
[stap_5e64ec831c6c1c2fa9784d7081f8e5fa_7859] (unreliable)
[c0000000e6513ce0] [c000000000182cd8] .vfs_write+0xd0/0x1dc
[c0000000e6513d80] [c000000000182eec] .SyS_write+0x58/0xa0
[c0000000e6513e30] [c0000000000085b4] syscall_exit+0x0/0x40
Instruction dump:
3883ffff 8c040001 2c000000 4082fff8 7c632050 4e800020 2c050000 40c10024
7ca903a6 38c3ffff 3884ffff 8c660001 <8c040001> 7c601851 4102fff4 4e800020
---[ end trace 21b12feeeff20bee ]---
objdump of the corresponding module shows this following code
63b8: 7f 9d 2a 14 add r28,r29,r5
}
/* build-id note payload start address */
/* XXX: But see
https://bugzilla.redhat.com/show_bug.cgi?id=465872;
dwfl_module_build_id was not intended to return the end
address. */
notes_addr -= m->build_id_len;
63bc: e8 bf 00 7a lwa r5,120(r31)
63c0: 7c 85 e0 50 subf r4,r5,r28
if (notes_addr <= base_addr) /* shouldn't happen */
63c4: 7f a4 e8 40 cmpld cr7,r4,r29
63c8: 40 9d ff 48 ble+ cr7,6310 <._stp_ctl_write_cmd+0x398>
continue;
if (memcmp(m->build_id_bits, (unsigned char*) notes_addr,
m->build_id_len)) {
63cc: e8 7f 00 60 ld r3,96(r31)
63d0: 48 00 00 01 bl 63d0 <._stp_ctl_write_cmd+0x458>
63d0: R_PPC64_REL24 .memcmp
63d4: 60 00 00 00 nop
63d8: 2f a3 00 00 cmpdi cr7,r3,0
63dc: 41 9e ff 34 beq+ cr7,6310 <._stp_ctl_write_cmd+0x398>
const char *basename;
basename = strrchr(m->path, '/');
Attaching the objdump o/p.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
` (2 preceding siblings ...)
2010-01-22 9:20 ` pavan dot naregundi at in dot ibm dot com
@ 2010-01-22 9:21 ` pavan dot naregundi at in dot ibm dot com
2010-01-24 14:22 ` fche at redhat dot com
` (2 subsequent siblings)
6 siblings, 0 replies; 8+ messages in thread
From: pavan dot naregundi at in dot ibm dot com @ 2010-01-22 9:21 UTC (permalink / raw)
To: systemtap
------- Additional Comments From pavan dot naregundi at in dot ibm dot com 2010-01-22 09:21 -------
Created an attachment (id=4549)
--> (http://sourceware.org/bugzilla/attachment.cgi?id=4549&action=view)
objdump -drS output
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
` (3 preceding siblings ...)
2010-01-22 9:21 ` pavan dot naregundi at in dot ibm dot com
@ 2010-01-24 14:22 ` fche at redhat dot com
2010-03-16 18:58 ` fche at redhat dot com
2010-09-29 17:00 ` fche at redhat dot com
6 siblings, 0 replies; 8+ messages in thread
From: fche at redhat dot com @ 2010-01-24 14:22 UTC (permalink / raw)
To: systemtap
------- Additional Comments From fche at redhat dot com 2010-01-24 14:19 -------
It appears as though one of the memcmp address parameters used
for buildid verification may be wrong. Without further symbol
data, I don't have a theory as to why that would be, but we might try ...
- changing the stap code to emit a loop of kread()s to compare the
buildid bytes, so in case of an address miscalculation, it will
fail cleanly
- disabling the buildid comparison generation code in the translator
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
` (4 preceding siblings ...)
2010-01-24 14:22 ` fche at redhat dot com
@ 2010-03-16 18:58 ` fche at redhat dot com
2010-09-29 17:00 ` fche at redhat dot com
6 siblings, 0 replies; 8+ messages in thread
From: fche at redhat dot com @ 2010-03-16 18:58 UTC (permalink / raw)
To: systemtap
------- Additional Comments From fche at redhat dot com 2010-03-16 18:58 -------
Please see if commit #f4191849c improves the situation.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |WAITING
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Bug runtime/10812] 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests
2009-10-20 5:25 [Bug runtime/10812] New: 2.6.32-rc5: Vector: 300 (Data Access) with systemtap tests pavan dot naregundi at in dot ibm dot com
` (5 preceding siblings ...)
2010-03-16 18:58 ` fche at redhat dot com
@ 2010-09-29 17:00 ` fche at redhat dot com
6 siblings, 0 replies; 8+ messages in thread
From: fche at redhat dot com @ 2010-09-29 17:00 UTC (permalink / raw)
To: systemtap
------- Additional Comments From fche at redhat dot com 2010-09-29 17:00 -------
Please check whether commits 71fa1fe39 and 77123275 fix the problem.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=10812
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
^ permalink raw reply [flat|nested] 8+ messages in thread