Re: INFO: task hung in vfs_setxattr (3)

[David Sterba]

On Tue, Sep 15, 2020 at 10:26:26AM +0300, Nikolay Borisov wrote:
> 
> 
> On 15.09.20 г. 7:59 ч., syzbot wrote:
> > Hello,
> > 
> > syzbot found the following issue on:
> > 
> > HEAD commit:    7fe10096 Merge branch 'linus' of git://git.kernel.org/pub/..
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=140b0853900000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=a9075b36a6ae26c9
> > dashboard link: https://syzkaller.appspot.com/bug?extid=738cca7d7d9754493513
> > compiler:       gcc (GCC) 10.1.0-syz 20200507
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=108d45a5900000
> > 
> > The issue was bisected to:
> > 
> > commit 6a3c7f5c87854e948c3c234e5f5e745c7c553722
> > Author: Nikolay Borisov <nborisov@xxxxxxxx>
> > Date:   Thu May 28 08:05:13 2020 +0000
> > 
> >     btrfs: don't balance btree inode pages from buffered write path
> > 
> > bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=14884d21900000
> > final oops:     https://syzkaller.appspot.com/x/report.txt?x=16884d21900000
> > console output: https://syzkaller.appspot.com/x/log.txt?x=12884d21900000
> 
> 
> Unlikely, that patch does change the performance characteristics because
> now we don't balance metadata when doing buffered writes but I don't see
> how it can lead to a hang.

Yeah, this patch once got pointed to by syzbot bisection supposedly
causing bug in a completely unrelated subsystem, but it's more likely
that it's making the bisection unreliable.

https://lore.kernel.org/lkml/20200811065013.GI2026@xxxxxxxxxxxxx/