Bug 11414
Summary: | Random crashes with 2.6.27-rc3 on PPC | ||
---|---|---|---|
Product: | Platform Specific/Hardware | Reporter: | Rafael J. Wysocki (rjw) |
Component: | PPC-32 | Assignee: | platform_ppc-32 |
Status: | CLOSED CODE_FIX | ||
Severity: | normal | CC: | bunk, mb |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 2.6.27-rc3 | Subsystem: | |
Regression: | Yes | Bisected commit-id: | |
Bug Depends on: | |||
Bug Blocks: | 10492 |
Description
Rafael J. Wysocki
2008-08-23 12:05:25 UTC
Please don't flag this as a regression from 2.6.26 too soon. I just reproduced it with 2.6.26. It could also be related to some kconfig options that I never enabled/disabled in the past. So the kernel where it not happens (2.6.25) does actually have a slightly different config than the crashing ones. For example, it has highmem enabled. Which the crashing kernels don't. There are more differences in the config... I'm still running tests. Unfortunately testing is very time consuming... OK, let's mark it as a regression from 2.6.25, then. It turns out that this is possibly related to a -fno-omit-frame-pointer gcc miscompilation that already hit other people. I'm still testing... So it's not strictly a linux regression at all. However, it affects 2.6.26 and 2.6.27 at least. I didn't try 2.6.25 with -fno-omit-frame-pointer. whoops, didn't want to post this one :( Sorry for the noise. Workaround for PowerPC (possibly breaks other archs): http://marc.info/?l=linux-kernel&m=121958694215835&w=4 fixed by commit 7563dc64585324f443f5ac107eb6d89ee813a2d2 |