Re: memory leak in rds_send_probe

[Eric Biggers <ebiggers@kernel.org>]

On Tue, Jul 23, 2019 at 03:23:36PM -0700, Andrew Morton wrote:
> On Tue, 23 Jul 2019 15:17:00 -0700 syzbot <syzbot+5134cdf021c4ed5aaa5f@syzkaller.appspotmail.com> wrote:
> 
> > syzbot has bisected this bug to:
> > 
> > commit af49a63e101eb62376cc1d6bd25b97eb8c691d54
> > Author: Matthew Wilcox <willy@linux.intel.com>
> > Date:   Sat May 21 00:03:33 2016 +0000
> > 
> >      radix-tree: change naming conventions in radix_tree_shrink
> > 
> > bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=176528c8600000
> > start commit:   c6dd78fc Merge branch 'x86-urgent-for-linus' of git://git...
> > git tree:       upstream
> > final crash:    https://syzkaller.appspot.com/x/report.txt?x=14e528c8600000
> > console output: https://syzkaller.appspot.com/x/log.txt?x=10e528c8600000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=8de7d700ea5ac607
> > dashboard link: https://syzkaller.appspot.com/bug?extid=5134cdf021c4ed5aaa5f
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=145df0c8600000
> > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=170001f4600000
> > 
> > Reported-by: syzbot+5134cdf021c4ed5aaa5f@syzkaller.appspotmail.com
> > Fixes: af49a63e101e ("radix-tree: change naming conventions in  
> > radix_tree_shrink")
> > 
> > For information about bisection process see: https://goo.gl/tpsmEJ#bisection
> 
> That's rather hard to believe.  af49a63e101eb6237 simply renames a
> couple of local variables.
> 

It's been known for months (basically ever since bisection was added) that about
50% of syzbot bisection results are obviously incorrect, often a commit selected
at random.  Unfortunately, the people actually funded to work on syzbot
apparently don't consider fixing this to be high priority issue, so we have to
live with it for now.  Or until someone volunteers to fix it themselves; source
is at https://github.com/google/syzkaller.

So for now, please don't waste much time on bisection results that look wonky.

But please do pay attention to any bisection results in reminders I've been
sending like "Reminder: 10 open syzbot bugs in foo subsystem", since I've
manually reviewed those to exclude the obviously wrong results...

- Eric