linux-kernel - Re: page_cache_tree_insert WARN

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20161004173425.GA1223@cmpxchg.org>
Date:   Tue, 4 Oct 2016 19:34:25 +0200
From:   Johannes Weiner <hannes@...xchg.org>
To:     Dave Jones <davej@...emonkey.org.uk>
Cc:     Linux Kernel <linux-kernel@...r.kernel.org>, linux-mm@...ck.org
Subject: Re: page_cache_tree_insert WARN_ON hit on 4.8+

Hi Dave,

On Tue, Oct 04, 2016 at 01:09:55PM -0400, Dave Jones wrote:
> Hit this during a trinity run.
> Kernel built from v4.8-1558-g21f54ddae449
> 
> WARNING: CPU: 0 PID: 5670 at ./include/linux/swap.h:276 page_cache_tree_insert+0x198/0x1b0
> CPU: 0 PID: 5670 Comm: trinity-c6 Not tainted 4.8.0-think+ #2
>  ffffc900003a3ab8 ffffffffb03dc311 0000000000000000 0000000000000000
>  ffffffffb0c063d6 ffffffffb018d898 ffffc900003a3af8 ffffffffb008b550
>  00000114003a3b30 ffffffffb0c063d6 0000000000000114 0000000000000000
> Call Trace:
>  [<ffffffffb03dc311>] dump_stack+0x6c/0x9b
>  [<ffffffffb018d898>] ? page_cache_tree_insert+0x198/0x1b0
>  [<ffffffffb008b550>] __warn+0x110/0x130
>  [<ffffffffb008b6dc>] warn_slowpath_null+0x2c/0x40
>  [<ffffffffb018d898>] page_cache_tree_insert+0x198/0x1b0
>  [<ffffffffb01900c4>] __add_to_page_cache_locked+0x1a4/0x3a0
>  [<ffffffffb0190379>] add_to_page_cache_lru+0x79/0x1c0
>  [<ffffffffb0193456>] generic_file_read_iter+0x916/0xce0
>  [<ffffffffb0232090>] do_iter_readv_writev+0x120/0x1c0
>  [<ffffffffb0192b40>] ? wait_on_page_bit_killable+0x100/0x100
>  [<ffffffffb0192b40>] ? wait_on_page_bit_killable+0x100/0x100
>  [<ffffffffb0232e29>] do_readv_writev+0x1f9/0x2e0
>  [<ffffffffb025d44e>] ? __fdget_pos+0x5e/0x70
>  [<ffffffffb025d44e>] ? __fdget_pos+0x5e/0x70
>  [<ffffffffb025d44e>] ? __fdget_pos+0x5e/0x70
>  [<ffffffffb0232f74>] vfs_readv+0x64/0x90
>  [<ffffffffb023300d>] do_readv+0x6d/0x120
>  [<ffffffffb0234907>] SyS_readv+0x27/0x30
>  [<ffffffffb000276f>] do_syscall_64+0x7f/0x200
>  [<ffffffffb09cfa8b>] entry_SYSCALL64_slow_path+0x25/0x25

Thanks for the report.

I've been trying to reproduce this too after Linus got hit by it. Is
there any way to trace back the steps what trinity was doing exactly?
What kind of file(system) this was operating on, file size, what the
/proc/vmstat delta before the operation until the trigger looks like?

The call to WARN() is new, the actual underlying bug could have been
around for a while, making it hard to narrow down.