Re: PROBLEM: kernel BUG at mm/page_alloc.c:775

[Mel Gorman]

On Tue, Jan 12, 2010 at 03:25:23PM -0600, Christoph Lameter wrote:
> On Sat, 9 Jan 2010, Michail Bachmann wrote:
> 
> > [   48.505381] kernel BUG at mm/page_alloc.c:775!
> 
> Somehow nodes got mixed up or the lookup tables for pages / zones are not
> giving the right node numbers.
> 

Agreed. On this type of machine, I'm not sure how that could happen
short of struct page information being corrupted. The range should
always be aligned to a pageblock boundary and I cannot see how that
would cross a zone boundary on this machine.

Does this machine pass memtest?
Is there any chance the problem can be bisected?

> > [   48.505467] invalid opcode: 0000 [#1]
> > [   48.505589] last sysfs file:
> > [   48.505672] Modules linked in:
> > [   48.505788]
> > [   48.505870] Pid: 343, comm: fsck.ext3 Not tainted (2.6.32.2-200912310108
> > #1) System Name
> > [   48.505994] EIP: 0060:[<c01641c6>] EFLAGS: 00010093 CPU: 0
> > [   48.506094] EIP is at move_freepages_block+0x86/0x130
> > [   48.506178] EAX: 000002fc EBX: 00000040 ECX: 00000000 EDX: 00000001
> > [   48.506264] ESI: 000041ed EDI: c1368000 EBP: e7267c70 ESP: e7267c50
> > [   48.506350]  DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
> > [   48.506438] Process fsck.ext3 (pid: 343, ti=e7266000 task=e7b78720
> > task.ti=e7266000)
> > [   48.506558] Stack:
> > [   48.506634]  00000000 00000000 c042595c c136ffe0 e7267c78 c1368018 00000002
> > 000001b8
> > [   48.506974] <0> e7267cc0 c0164751 00000000 c03fd58c 00000001
> > c0361780e7267ca8 00000206
> > [   48.507379] <0> 00000000 00000000 c042595c c1368000 00011210
> > c0425b50c0425b50 c0425998
> > [   48.507855] Call Trace:
> > [   48.507938]  [<c0164751>] ? __rmqueue+0x1a1/0x350
> > [   48.508027]  [<c016636b>] ? get_page_from_freelist+0x35b/0x420
> > [   48.508115]  [<c01664f8>] ? __alloc_pages_nodemask+0xc8/0x510
> > [   48.508215]  [<c011db73>] ? dequeue_task+0x63/0xb0
> > [   48.508304]  [<c0167ea8>] ? __do_page_cache_readahead+0xb8/0x1b0
> > [   48.508396]  [<c0167fc8>] ? ra_submit+0x28/0x30
> > [   48.508480]  [<c01681cd>] ? ondemand_readahead+0xfd/0x1e0
> > [   48.508567]  [<c0168320>] ? page_cache_async_readahead+0x70/0x90
> > [   48.508653]  [<c0162c8c>] ? generic_file_aio_read+0x2fc/0x620
> > [   48.508741]  [<c018b101>] ? do_sync_read+0xd1/0x110
> > [   48.508834]  [<c01369f0>] ? autoremove_wake_function+0x0/0x40
> > [   48.508928]  [<c0284c20>] ? n_tty_write+0x0/0x3d0
> > [   48.509018]  [<c02402ff>] ? security_file_permission+0xf/0x20
> > [   48.509103]  [<c018b194>] ? rw_verify_area+0x54/0xd0
> > [   48.509188]  [<c018bd39>] ? vfs_read+0x99/0x160
> > [   48.509269]  [<c018b030>] ? do_sync_read+0x0/0x110
> > [   48.509351]  [<c018bebd>] ? sys_read+0x3d/0x70
> > [   48.509434]  [<c0102c44>] ? sysenter_do_call+0x12/0x22
> > [   48.509517] Code: c1 e2 06 c1 e1 08 29 d1 8b 93 e0 7f 00 00 29 c1 c1 e1 02
> > c1 ea 1e 89 d3 89 d0 c1 e3 06 c1 e0 08 29 d8 29 d0 c1 e0 02 39 c1 74 1c <0f>
> > 0b eb fe 8d b6 00 00 00 00 c7 45 f0 00 00 00 00 8b 45 f0 83
> > [   48.511798] EIP: [<c01641c6>] move_freepages_block+0x86/0x130 SS:ESP
> > 0068:e7267c50
> > [   48.511990] ---[ end trace 45c7d49cba718751 ]---
> >
> > My memory layout on this box is (from dmesg):
> >
> > ----
> > Zone PFN ranges:
> >   DMA      0x00000000 -> 0x00001000
> >   Normal   0x00001000 -> 0x00027fec
> > Movable zone start PFN for each node
> > early_node_map[3] active PFN ranges
> >     0: 0x00000000 -> 0x00000001
> >     0: 0x00000010 -> 0x000000a0
> >     0: 0x00000100 -> 0x00027fec
> > On node 0 totalpages: 163709
> > free_area_init_node: node 0, pgdat c0425660, node_mem_map c1000000
> >   DMA zone: 32 pages used for memmap
> >   DMA zone: 0 pages reserved
> >   DMA zone: 3953 pages, LIFO batch:0
> >   Normal zone: 1248 pages used for memmap
> >   Normal zone: 158476 pages, LIFO batch:31
> > ----
> >
> > The last kernel version without this problem seems to be the 2.6.30.x (I am
> > running 2.6.30.10 right now without any problems).
> >
> > If you need any more information from my system don't hesitate to ask.
> >
> > CU Micha
> >
> 

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/