| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Fri, 4 Mar 2022 15:26:19 +0800
From: Qu Wenruo <quwenruo.btrfs@....com>
To: Oliver Sang <oliver.sang@...el.com>
Cc: Qu Wenruo <wqu@...e.com>, David Sterba <dsterba@...e.com>,
LKML <linux-kernel@...r.kernel.org>,
Linux Memory Management List <linux-mm@...ck.org>,
lkp@...ts.01.org, lkp@...el.com, linux-btrfs@...r.kernel.org
Subject: Re: [btrfs] 3626a285f8: divide_error:#[##]
On 2022/3/2 16:44, Oliver Sang wrote:
> Hi Qu,
>
> On Tue, Mar 01, 2022 at 03:47:38PM +0800, Qu Wenruo wrote:
>>
>>
>> On 2022/3/1 14:30, kernel test robot wrote:
>>>
>>>
>>> Greeting,
>>>
>>> FYI, we noticed the following commit (built with gcc-9):
>>>
>>> commit: 3626a285f87dceb4ca649d0ef015d7b295206cdf ("btrfs: introduce dedicated helper to scrub simple-stripe based range")
>>> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
>>>
>>> in testcase: xfstests
>>> version: xfstests-x86_64-1de1db8-1_20220217
>>> with following parameters:
>>>
>>> disk: 6HDD
>>> fs: btrfs
>>> test: btrfs-group-07
>>> ucode: 0x28
>>>
>>> test-description: xfstests is a regression test suite for xfs and other files ystems.
>>> test-url: git://git.kernel.org/pub/scm/fs/xfs/xfstests-dev.git
>>>
>>>
>>> on test machine: 8 threads 1 sockets Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz with 8G memory
>>>
>>> caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
>>>
>>>
>>>
>>> If you fix the issue, kindly add following tag
>>> Reported-by: kernel test robot <oliver.sang@...el.com>
>>>
>>>
>>> [ 65.408303][ T3224] BTRFS info (device sdb2): flagging fs with big metadata feature
>>> [ 65.415944][ T3224] BTRFS info (device sdb2): disk space caching is enabled
>>> [ 65.422842][ T3224] BTRFS info (device sdb2): has skinny extents
>>> [ 65.436656][ T3224] BTRFS info (device sdb2): checking UUID tree
>>> [ 66.134430][ T3293] BTRFS info (device sdb2): dev_replace from /dev/sdb3 (devid 2) to /dev/sdb6 started
>>> [ 67.823326][ T3293] divide error: 0000 [#1] SMP KASAN PTI
>>> [ 67.828668][ T3293] CPU: 3 PID: 3293 Comm: btrfs Not tainted 5.17.0-rc5-00101-g3626a285f87d #1
>>> [ 67.837169][ T3293] Hardware name: Dell Inc. OptiPlex 9020/0DNKMN, BIOS A05 12/05/2013
>>> [ 67.844982][ T3293] RIP: 0010:scrub_stripe (kbuild/src/consumer/fs/btrfs/scrub.c:3448 kbuild/src/consumer/fs/btrfs/scrub.c:3486 kbuild/src/consumer/fs/btrfs/scrub.c:3644) btrfs
>>> [ 67.850976][ T3293] Code: 00 00 fc ff df 48 89 f9 48 c1 e9 03 0f b6 0c 11 48 89 fa 83 e2 07 83 c2 03 38 ca 7c 08 84 c9 0f 85 27 09 00 00 41 8b 5d 1c 99 <f7> fb 48 8b 54 24 30 48 c1 ea 03 48 63 e8 48 b8 00 00 00 00 00 fc
>>> All code
>>
>> This is weird, the code is from simple_stripe_full_stripe_len(), which
>> means the chunk map must be RAID0 or RAID10.
>>
>> In that case, their sub_stripes should be either 1 or 2, why we got 0 there?
>>
>> In fact, from volumes.c, all sub_stripes is from btrfs_raid_array[],
>> which all have either 1 or 2 sub_stripes.
>>
>>
>> Although the code is old, not the latest version, it should still not
>> cause such problem.
>>
>> Mind to retest with my branch to see if it can be reproduced?
>> https://github.com/adam900710/linux/tree/refactor_scrub
>
> we tested head of this branch:
> d6e3a8c42f2fad btrfs: scrub: rename scrub_bio::pagev and related members
> and:
> fdad4a9615f180 btrfs: introduce dedicated helper to scrub simple-stripe based range
> on this branch.
>
> by attached config.
>
> still reproduce the same issue.
>
> attached dmesgs FYI.
Still failed to reproduce here.
Those btrfs/07[0123] tests are already in scrub/replace group, thus I
ran them almost hourly during the development.
Although there are some ASSERT()s doing extra sanity checks, they should
not affect the result anyway.
Thus I pushed a branch with more explicit BUG_ON()s to catch the
possible divide by zero bugs.
(https://github.com/adam900710/linux/tree/refactor_scrub_testing)
Mind to give it a try?
Thanks,
Qu
>
>
>>
>> Thanks,
>> Qu
>>
>>> ========
>>> 0: 00 00 add %al,(%rax)
>>> 2: fc cld
>>> 3: ff (bad)
>>> 4: df 48 89 fisttps -0x77(%rax)
>>> 7: f9 stc
>>> 8: 48 c1 e9 03 shr $0x3,%rcx
>>> c: 0f b6 0c 11 movzbl (%rcx,%rdx,1),%ecx
>>> 10: 48 89 fa mov %rdi,%rdx
>>> 13: 83 e2 07 and $0x7,%edx
>>> 16: 83 c2 03 add $0x3,%edx
>>> 19: 38 ca cmp %cl,%dl
>>> 1b: 7c 08 jl 0x25
>>> 1d: 84 c9 test %cl,%cl
>>> 1f: 0f 85 27 09 00 00 jne 0x94c
>>> 25: 41 8b 5d 1c mov 0x1c(%r13),%ebx
>>> 29: 99 cltd
>>> 2a:* f7 fb idiv %ebx <-- trapping instruction
>>> 2c: 48 8b 54 24 30 mov 0x30(%rsp),%rdx
>>> 31: 48 c1 ea 03 shr $0x3,%rdx
>>> 35: 48 63 e8 movslq %eax,%rbp
>>> 38: 48 rex.W
>>> 39: b8 00 00 00 00 mov $0x0,%eax
>>> 3e: 00 fc add %bh,%ah
>>>
>>> Code starting with the faulting instruction
>>> ===========================================
>>> 0: f7 fb idiv %ebx
>>> 2: 48 8b 54 24 30 mov 0x30(%rsp),%rdx
>>> 7: 48 c1 ea 03 shr $0x3,%rdx
>>> b: 48 63 e8 movslq %eax,%rbp
>>> e: 48 rex.W
>>> f: b8 00 00 00 00 mov $0x0,%eax
>>> 14: 00 fc add %bh,%ah
>>> [ 67.870187][ T3293] RSP: 0018:ffffc9000a71f450 EFLAGS: 00010246
>>> [ 67.876028][ T3293] RAX: 0000000000000004 RBX: 0000000000000000 RCX: 0000000000000000
>>> [ 67.883756][ T3293] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffff888129ec6d1c
>>> [ 67.891491][ T3293] RBP: ffff8881453682a0 R08: 0000000000000001 R09: 0000000000000000
>>> [ 67.899230][ T3293] R10: ffff88821534a063 R11: ffffed1042a6940c R12: ffff888121238000
>>> [ 67.906955][ T3293] R13: ffff888129ec6d00 R14: ffff888145368000 R15: 0000000000000008
>>> [ 67.914680][ T3293] FS: 00007f2851eb08c0(0000) GS:ffff8881a6d80000(0000) knlGS:0000000000000000
>>> [ 67.923351][ T3293] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [ 67.929709][ T3293] CR2: 00007ffea4ff07f8 CR3: 000000010a0fc005 CR4: 00000000001706e0
>>> [ 67.937437][ T3293] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>> [ 67.945163][ T3293] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>> [ 67.952891][ T3293] Call Trace:
>>> [ 67.955992][ T3293] <TASK>
>>> [ 67.958749][ T3293] ? kasan_save_stack (kbuild/src/consumer/mm/kasan/common.c:39)
>>> [ 67.963395][ T3293] ? kasan_set_track (kbuild/src/consumer/mm/kasan/common.c:45)
>>> [ 67.967951][ T3293] ? kasan_set_free_info (kbuild/src/consumer/mm/kasan/generic.c:372)
>>> [ 67.972851][ T3293] ? mutex_unlock (kbuild/src/consumer/arch/x86/include/asm/atomic64_64.h:190 kbuild/src/consumer/include/linux/atomic/atomic-long.h:449 kbuild/src/consumer/include/linux/atomic/atomic-instrumented.h:1790 kbuild/src/consumer/kernel/locking/mutex.c:178 kbuild/src/consumer/kernel/locking/mutex.c:537)
>>>
>>>
>>> To reproduce:
>>>
>>> git clone https://github.com/intel/lkp-tests.git
>>> cd lkp-tests
>>> sudo bin/lkp install job.yaml # job file is attached in this email
>>> bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
>>> sudo bin/lkp run generated-yaml-file
>>>
>>> # if come across any failure that blocks the test,
>>> # please remove ~/.lkp and /lkp dir to run from a clean state.
>>>
>>>
>>>
>>> ---
>>> 0DAY/LKP+ Test Infrastructure Open Source Technology Center
>>> https://lists.01.org/hyperkitty/list/lkp@lists.01.org Intel Corporation
>>>
>>> Thanks,
>>> Oliver Sang
>>>
Powered by blists - more mailing lists