Included is the debug log from 20050328 xen-devel build. SMP dom0 booted fine,
but being
used as a distcc server, gcc oopsed the dom0 kernel.
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253 T/L: 678-9253
ryanh@us.ibm.com
---
(XEN) DOM0: (file=mm.c, line=1179) Bad type (saw e0000000 != exp 40000000) for
pfn 00027747
(XEN)
(XEN) DOM0: (file=mm.c, line=322) Bad page type for pfn 00027747 (f0000001)
(XEN) DOM0: (file=mm.c, line=1286) Error while pinning pfn 00027747
Failed to execute MMU updates.
------------[ cut here ]------------
kernel BUG at <bad filename>:54763!
invalid operand: 0000 [#1]
PREEMPT SMP DEBUG_PAGEALLOC
Modules linked in: sworks_agp agpgart
CPU: 0
EIP: 0061:[<c0116303>] Not tainted VLI
EFLAGS: 00010292 (2.6.11-xen0-smp)
EIP is at _flush_page_update_queue+0x93/0xa0
eax: 00000022 ebx: c1407548 ecx: fbffc000 edx: 00000001
esi: 00ef4f60 edi: 00000000 ebp: 00000000 esp: d8193e38
ds: 007b es: 007b ss: 0069
Process gcc (pid: 2509, threadinfo=d8192000 task=d9456af0)
Stack: c03df4a0 d8192000 d88b8000 c05d2e40 c0145400 d88b8000 c05d2e40 00000001
c0117ace d8192000 d88b8000 c05d2e40 c0145829 c05d2e40 000000d0 d88b8000
c0114eb9 da2f7d8c d8193f64 da2f7ff0 d9a0e000 c0114eb9 c05d2e40 000000d0
Call Trace:
[<c0145400>] cache_alloc_debugcheck_after+0xa0/0x1b0
[<c0117ace>] __change_page_attr+0x2e/0x230
[<c0145829>] kmem_cache_alloc+0xa9/0xf0
[<c0114eb9>] pgd_alloc+0x19/0x20
[<c0114eb9>] pgd_alloc+0x19/0x20
[<c011c2e0>] mm_init+0xf0/0x120
[<c011c346>] mm_alloc+0x36/0x50
[<c0167f6c>] do_execve+0x7c/0x220
[<c01085b6>] sys_execve+0x46/0xa0
[<c0109b3f>] syscall_call+0x7/0xb
Code: 85 c0 78 1b 89 fa 8b 5c 24 04 b8 18 bb 43 c0 8b 74 24 08 8b 7c 24 0c 83 c4
10 e9 89 c5 2a 00 c7 04 24 a0 f4 3d c0 e8 ed 86 00 00 <0f> 0b eb d5 89 f6
8d bc 27 00 00 00 00 55 bd e0 29 51 c0 57 56
<6>note: gcc[2509] exited with preempt_count 1
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
> > Included is the debug log from 20050328 xen-devel build. SMP dom0 booted fine, but being > used as a distcc server, gcc oopsed the dom0 kernel.Ryan, TLB flushing in SMP builds has so far never been 100% correct, so it''s not surprising that there have been some crashes. I''ve just checked in a patch that fixes all the flush problems that I know about, so it would be useful to know whether you can still provoke the above bug (or any other bugs!). -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
* Keir Fraser <Keir.Fraser@cl.cam.ac.uk> [2005-03-31 11:08]:> > > > Included is the debug log from 20050328 xen-devel build. SMP dom0 booted fine, but being > > used as a distcc server, gcc oopsed the dom0 kernel. > > Ryan, > > TLB flushing in SMP builds has so far never been 100% correct, so it''s > not surprising that there have been some crashes. I''ve just checked in > a patch that fixes all the flush problems that I know about, so it > would be useful to know whether you can still provoke the above bug > (or any other bugs!).I run builds of the nightly snapshot as a distcc host for compiling Xen and the two kernels. I''ve not been able to reproduce any of the flushing errors I''ve seen previously since the 20050328 build, both 29, 30, and 31 builds have been stable. I''ll be sure to post any oops I get. Thanks Keir. Ryan Harper Software Engineer; Linux Technology Center IBM Corp., Austin, Tx (512) 838-9253 T/L: 678-9253 ryanh@us.ibm.com _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 31 Mar 2005, Keir Fraser wrote:> TLB flushing in SMP builds has so far never been 100% correct, so it''s > not surprising that there have been some crashes. I''ve just checked in > a patch that fixes all the flush problems that I know about, so it > would be useful to know whether you can still provoke the above bug > (or any other bugs!).I''ll try to upgrade the Fedora rawhide xen kernels to the latest and see if the bug has gone away. -- "Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." - Brian W. Kernighan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel