Re: Bug#394392: mm/msync in recent kernel broken?
reassign 401006 linux-2.6
severity 401006 grave
retitle 401006 mm/msync leading to filesystem corruption
thanks
Seems other people see problems with mm/msync too.
* bugsgrid@gmail.com <bugsgrid@gmail.com> [2006-12-16 13:32]:
> I think something got wrong with this change, regarding mmap/msync on
> somewhat-slow? filesystem, resulting in sporadic data loss.
> For my case it occures during running rtorrent on cifs (ok I know it's
> too bad idea anyway), sometime with tons of strange log (see below),
> sometime just gets corrupted silently.
> Please see #401006 and its upstream ml, reportedly regression on
> linux-image-2.6.18-3 and 2.6.19. This is not cifs specific (same
> problem reported on local reiserfs and so on).
>
> kernel at that time is home-built one from linux-source-2.6.18-6.
> Dec 3 18:26:16 x kernel: Linux version 2.6.18 (Version: m0) (x@x)
> (gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-20)) #1 Sat Dec
> 2 14:22:41 JST 2006
> Dec 11 09:15:16 x kernel: printk: 536 messages suppressed.
> Dec 11 09:15:16 x kernel: swapper: page allocation failure. order:0,
> mode:0x20
> Dec 11 09:15:16 x kernel: [<c012b8d5>] [<c013e7b9>] [<c013ea03>]
> [<c0230223>] [<c01ddc39>] [<c01de97d>] [<c010c6e4>] [<c0116151>]
> [<c0236068>] [<c0114372>] [<c01143e6>] [<c01041d3>] [<c01041d3>]
> [<c0102976>] [<c0102976>] [<c0101561>] [<c01015c9>] [<c02c66d9>]
> [<c02c6230>] <6>Mem-info:
> Dec 11 09:15:16 x kernel: DMA per-cpu:
> Dec 11 09:15:16 x kernel: cpu 0 hot: high 0, batch 1 used:0
> Dec 11 09:15:16 x kernel: cpu 0 cold: high 0, batch 1 used:0
> Dec 11 09:15:16 x kernel: DMA32 per-cpu: empty
> Dec 11 09:15:16 x kernel: Normal per-cpu:
> Dec 11 09:15:16 x kernel: cpu 0 hot: high 90, batch 15 used:18
> Dec 11 09:15:16 x kernel: cpu 0 cold: high 30, batch 7 used:23
> Dec 11 09:15:16 x kernel: HighMem per-cpu: empty
> Dec 11 09:15:16 x kernel: Free pages: 1668kB (0kB HighMem)
> Dec 11 09:15:16 x kernel: Active:37320 inactive:24391 dirty:6568
> writeback:0 unstable:0 free:417 slab:1872 mapped:11210 pagetables:86
> Dec 11 09:15:16 x kernel: DMA free:1004kB min:124kB low:152kB
> high:184kB active:8016kB inactive:4404kB present:16384kB
> pages_scanned:0 all_unreclaimable? no
> Dec 11 09:15:16 x kernel: lowmem_reserve[]: 0 0 239 239
> Dec 11 09:15:16 x kernel: DMA32 free:0kB min:0kB low:0kB high:0kB
> active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
> no
> Dec 11 09:15:16 x kernel: lowmem_reserve[]: 0 0 239 239
> Dec 11 09:15:16 x kernel: Normal free:664kB min:1916kB low:2392kB
> high:2872kB active:141264kB inactive:93160kB present:245696kB
> pages_scanned:192 all_unreclaimable? no
> Dec 11 09:15:16 x kernel: lowmem_reserve[]: 0 0 0 0
> Dec 11 09:15:16 x kernel: HighMem free:0kB min:128kB low:128kB
> high:128kB active:0kB inactive:0kB present:0kB pages_scanned:0
> all_unreclaimable? no
> Dec 11 09:15:16 x kernel: lowmem_reserve[]: 0 0 0 0
> Dec 11 09:15:16 x kernel: DMA: 3*4kB 0*8kB 0*16kB 1*32kB 1*64kB
> 1*128kB 1*256kB 1*512kB 0*1024kB 0*2048kB 0*4096kB = 1004kB
> Dec 11 09:15:16 x kernel: DMA32: empty
> Dec 11 09:15:16 x kernel: Normal: 0*4kB 1*8kB 1*16kB 0*32kB 0*64kB
> 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB 0*4096kB = 664kB
> Dec 11 09:15:16 x kernel: HighMem: empty
> Dec 11 09:15:16 x kernel: Swap cache: add 14520, delete 14520, find
> 218/354, race 0+0
> Dec 11 09:15:16 x kernel: Free swap = 249712kB
> Dec 11 09:15:16 x kernel: Total swap = 249944kB
> Dec 11 09:15:16 x kernel: Free swap: 249712kB
> Dec 11 09:15:16 x kernel: 65520 pages of RAM
> Dec 11 09:15:16 x kernel: 0 pages of HIGHMEM
> Dec 11 09:15:16 x kernel: 1147 reserved pages
> Dec 11 09:15:16 x kernel: 14274 pages shared
> Dec 11 09:15:16 x kernel: 0 pages swap cached
> Dec 11 09:15:16 x kernel: 6568 pages dirty
> Dec 11 09:15:16 x kernel: 0 pages writeback
> Dec 11 09:15:16 x kernel: 11210 pages mapped
> Dec 11 09:15:16 x kernel: 1872 pages slab
> Dec 11 09:15:16 x kernel: 86 pages pagetables
>
> On my box swap is local dedicated partition. I can't test with
> official linux-image because my box won't boot due to buggy usb-legacy
> (my box is something crap not installable with d-i).
>
> Sorry if i should have opened new bug...
--
Martin Michlmayr
http://www.cyrius.com/
Reply to: