Closed Bug 743123 Opened 8 years ago Closed 6 years ago

Crashes on Talos on Linux64 have useless stacks

Categories

(Testing :: Talos, defect)

x86_64
Linux
defect
Not set

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: philor, Unassigned)

References

Details

https://tbpl.mozilla.org/php/getParsedLog.php?id=10675395&tree=Mozilla-Inbound
https://tbpl.mozilla.org/php/getParsedLog.php?id=10681030&tree=Mozilla-Inbound
https://tbpl.mozilla.org/php/getParsedLog.php?id=10681872&tree=Mozilla-Inbound

Stacks like

Thread 0 (crashed)
 0  libpthread-2.11.so + 0xb04c
    rbx = 0x84b54040   r12 = 0xffffffff   r13 = 0x826f9d00   r14 = 0x80070057
    r15 = 0x00000000   rip = 0xd360b04c   rsp = 0x905fea18   rbp = 0x73d38800

Thread 1
 0  libc-2.11.so + 0xda7f9
    rbx = 0x7a729000   r12 = 0x795dbc10   r13 = 0x795dbc90   r14 = 0x795dbc00
    r15 = 0x00000000   rip = 0xd2eda7f9   rsp = 0x795dbb88   rbp = 0x7a74a000

Thread 2
 0  libpthread-2.11.so + 0xb04c
    rbx = 0x84b54590   r12 = 0xffffffff   r13 = 0x00000000   r14 = 0x78522e5f
    r15 = 0x00000003   rip = 0xd360b04c   rsp = 0x78522d58   rbp = 0x7a733b00

Thread 3
 0  libpthread-2.11.so + 0xb04c
    rbx = 0x84b546a0   r12 = 0xffffffff   r13 = 0x6c2ef390   r14 = 0x6c27bc00
    r15 = 0x00000000   rip = 0xd360b04c   rsp = 0x777fe258   rbp = 0x6af5e300

Thread 4
 0  libpthread-2.11.so + 0xb04c
    rbx = 0x84b547b0   r12 = 0xffffffff   r13 = 0x00000000   r14 = 0x00000001
    r15 = 0x00000000   rip = 0xd360b04c   rsp = 0x765ccdf8   rbp = 0x77a54140

Thread 5
 0  libpthread-2.11.so + 0xb04c
    rbx = 0x84b549d0   r12 = 0xffffffff   r13 = 0x828b0af4   r14 = 0x00000000
    r15 = 0x00000003   rip = 0xd360b04c   rsp = 0x759fee18   rbp = 0x77a2ea40

Thread 6
 0  libpthread-2.11.so + 0xb04c
    rbx = 0x84b54bf0   r12 = 0xffffffff   r13 = 0x74bfe710   r14 = 0x00000000
    r15 = 0x00000003   rip = 0xd360b04c   rsp = 0x74bfde48   rbp = 0x77a2eac0

Thread 7
 0  libpthread-2.11.so + 0xb3b9
    rbx = 0x7a733a08   r12 = 0x0000012b   r13 = 0x726fecd0   r14 = 0x00000001
    r15 = 0x00000000   rip = 0xd360b3b9   rsp = 0x726fec78   rbp = 0x84be7b60

just don't get the bugs fixed.
It's hard to say anything useful without having a minidump to play with. Obviously if we had symbols for OS libraries (bug 528231) we'd be likely to get something more useful, but I'm surprised that we don't get anywhere from stack scanning. Maybe we just need a newer minidump_stackwalk binary?
Inconveniently, https://tbpl.mozilla.org/php/getParsedLog.php?id=10726905&tree=Mozilla-Inbound is a Linux64 talos crash with a useful stack, so it's not the case that they are always busted, which would probably be easier to deal with.
(In reply to Phil Ringnalda (:philor) from comment #5)
> Inconveniently,
> https://tbpl.mozilla.org/php/getParsedLog.php?id=10726905&tree=Mozilla-
> Inbound is a Linux64 talos crash with a useful stack, so it's not the case
> that they are always busted, which would probably be easier to deal with.

Okay, so, that shows a crash in Gecko code, and we get a useful stack. If you look at the full log, however, all of the other threads display the same problem, a single useless frame in libpthread. Clearly the stackwalker is having an issue walking out of those pthread frames.
Do we have a plan to make progress on this? Is bug 745193 all we need to do here?
Getting that bug fixed will at least allow us to examine these minidumps and see if we can figure out what's going on. As of right now there's not enough info in the log to do anything at all.
Depends on: 745193, 744442
This appears to affect all platforms, not just Linux64.
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.