please pull powerpc-merge.git
Linus Torvalds
torvalds at osdl.org
Tue Nov 15 15:20:12 EST 2005
On Tue, 15 Nov 2005, Paul Mackerras wrote:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc-merge.git
Hey, I just got an oops.
This was with a plain 2.6.15-rc1 kernel (so before the merge):
Unable to handle kernel paging request for data at address 0xc0000000ff000000
Faulting instruction address: 0xc000000000031200
Oops: Kernel access of bad area, sig: 11 [#1]
SMP NR_CPUS=2 NUMA POWERMAC
Modules linked in: autofs
NIP: C000000000031200 LR: C000000000097B90 CTR: 0000000000000020
REGS: c0000001561f38d0 TRAP: 0300 Not tainted (2.6.15-rc1-g508862e4)
MSR: 9000000000009032 <EE,ME,IR,DR> CR: 88004448 XER: 00000000
DAR: C0000000FF000000, DSISR: 0000000042010000
TASK = c000000105185040[6398] 'git-fsck-object' THREAD: c0000001561f0000
CPU: 0
GPR00: 0000000000000080 C0000001561F3B50 C0000000006CD7E0 C0000000FF000000
GPR04: 00000000F2113000 C0000000040E6000 C0000000005BA400 9000000000009032
GPR08: C00000000FFFA800 C0000000006CED38 C0000000006D1CA8 0000000000000020
GPR12: 0000000048004424 C0000000005BA400 0000000010020000 0000000010010000
GPR16: 0000000000044836 00000000004B1EC2 0000000000000000 0000000000000000
GPR20: C0000001791DCC80 C00000017BE63700 C000000145D66700 00000000F2113000
GPR24: 0000000002000000 0000000000000000 0000000000000898 C0000000040E6000
GPR28: C00000013914E898 C00000013914E000 C0000000005D7830 0000000000000000
NIP [C000000000031200] .clear_user_page+0x10/0x60
LR [C000000000097B90] .__handle_mm_fault+0xda0/0xf10
Call Trace:
[C0000001561F3B50] [C000000000097B24] .__handle_mm_fault+0xd34/0xf10 (unreliable)
[C0000001561F3C60] [C00000000049729C] .do_page_fault+0x4ec/0x7f0
[C0000001561F3E30] [C000000000004760] .handle_page_fault+0x20/0x54
Instruction dump:
4d820020 7c0018a8 7c004878 7c0019ad 40c2fff4 4e800020 60000000 60000000
e922a3e0 8169000c 80090004 7d6903a6 <7c001fec> 7c630214 4320fff8 e922a3d8
the worrisome thing is that it happened while running git-fsck-objects..
Any ideas? I'm really bad at decoding ppc64 oopses, so ...
(There are other reports of VM-induced problems on -rc1, this is probably
not ppc64-related).
Linus
More information about the Linuxppc64-dev
mailing list