[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160502220745.44c4bd91@canb.auug.org.au>
Date: Mon, 2 May 2016 22:07:45 +1000
From: Stephen Rothwell <sfr@...b.auug.org.au>
To: Mateusz Guzik <mguzik@...hat.com>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
linux-fsdevel@...r.kernel.org,
Alexander Viro <viro@...iv.linux.org.uk>,
linux-next@...r.kernel.org, linux-kernel@...r.kernel.org,
Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
Theodore Ts'o <tytso@....edu>,
Miklos Szeredi <mszeredi@...hat.com>
Subject: Re: linux-next: Tree for May 2 [WARNING: at fs/dcache.c]
Hi Mateusz,
On Mon, 2 May 2016 12:33:54 +0200 Mateusz Guzik <mguzik@...hat.com> wrote:
>
> On Mon, May 02, 2016 at 07:15:24PM +0900, Sergey Senozhatsky wrote:
> > On (05/02/16 18:40), Stephen Rothwell wrote:
> > > Hi all,
> > >
> > > Changes since 20160429
> >
> > Hello,
> >
> > [ 0.368791] ------------[ cut here ]------------
> > [ 0.368850] WARNING: CPU: 0 PID: 1 at fs/dcache.c:1688 d_set_d_op+0x5e/0xcc
> > [ 0.368911] Modules linked in:
> > [ 0.369002] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.6.0-rc6-next-20160502-dbg-00005-gf58c9da-dirty #404
> > [ 0.369161] 0000000000000000 ffff880133067908 ffffffff811b8202 0000000000000000
> > [ 0.369371] 0000000000000000 ffff880133067948 ffffffff81039365 00000698e5dffe26
> > [ 0.369580] ffff880132c090c0 ffffffff81613680 ffff880132c040a0 ffff880132c08000
> > [ 0.369791] Call Trace:
> > [ 0.369846] [<ffffffff811b8202>] dump_stack+0x4d/0x63
> > [ 0.369904] [<ffffffff81039365>] __warn+0xb8/0xd3
> > [ 0.369962] [<ffffffff810393e6>] warn_slowpath_null+0x18/0x1a
> > [ 0.370021] [<ffffffff811012f6>] d_set_d_op+0x5e/0xcc
> > [ 0.370079] [<ffffffff8110d887>] simple_lookup+0x2e/0x45
>
> The issue is that 2 macros have the same value:
>
> #define DCACHE_OP_REAL 0x08000000
>
> #define DCACHE_PAR_LOOKUP 0x08000000 /* being looked up
> (with parent locked shared) */
>
> Verified with switching one to 0x10000000 and the warning went away.
This comes from a bad automatic merge resolution between commit
d101a125954e ("fs: add file_dentry()")
from Linus' tree (introduced before v4.5-rc3) and commit
340450a54573 ("beginning of transition to parallel lookups - marking in-lookup dentries")
from the vfs tree.
I will fix this up in tomorrow's tree (unless Al beats me to it).
--
Cheers,
Stephen Rothwell
Powered by blists - more mailing lists