[Bridge] memory leak in fdb_create

Nikolay Aleksandrov nikolay at cumulusnetworks.com
Sun Jul 28 16:51:03 UTC 2019


On 28/07/2019 17:20, syzbot wrote:
> syzbot has bisected this bug to:
> 
> commit 04cf31a759ef575f750a63777cee95500e410994
> Author: Michael Ellerman <mpe at ellerman.id.au>
> Date:   Thu Mar 24 11:04:01 2016 +0000
> 
>     ftrace: Make ftrace_location_range() global
> 
> bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=1538c778600000
> start commit:   abf02e29 Merge tag 'pm-5.2-rc6' of git://git.kernel.org/pu..
> git tree:       upstream
> final crash:    https://syzkaller.appspot.com/x/report.txt?x=1738c778600000
> console output: https://syzkaller.appspot.com/x/log.txt?x=1338c778600000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=56f1da14935c3cce
> dashboard link: https://syzkaller.appspot.com/bug?extid=88533dc8b582309bf3ee
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=16de5c06a00000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=10546026a00000
> 
> Reported-by: syzbot+88533dc8b582309bf3ee at syzkaller.appspotmail.com
> Fixes: 04cf31a759ef ("ftrace: Make ftrace_location_range() global")
> 
> For information about bisection process see: https://goo.gl/tpsmEJ#bisection

I see the problem, it'd happen if the multicast stats memory allocation fails on bridge
init then the fdb added due to the default vlan would remain and the bridge kmem cache
would be destroyed while not empty (you can even trigger a BUG because of that).
I'll post a patch shortly after running a few tests.

Thanks,
 Nik



More information about the Bridge mailing list