netdev - Re: Report mlx5

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]

Date: Thu, 8 Feb 2024 11:12:23 +0800
From: Tao Liu <taoliu828@....com>
To: Cosmin Ratiu <cratiu@...dia.com>
Cc: roid@...dia.com, paulb@...dia.com, vladbu@...dia.com,
	dchumak@...dia.com, saeedm@...dia.com, taoliu828@....com,
	netdev@...r.kernel.org
Subject: Re: Report mlx5_core crash

On 02/07  , Cosmin Ratiu wrote:
> On Tue, 2024-02-06 at 15:01 +0800, Tao Liu wrote:
> > On 01/31  , Tao Liu wrote:
> > > Hi Mellanox team,
> > > 
> > >    We hit a crash in mlx5_core which is similar with commit
> > >    de31854ece17 ("net/mlx5e: Fix nullptr on deleting mirroring rule").
> > >    But they are different cases, our case is:
> > >    in_port(...),eth(...) \
> > > actions:set(tunnel(...)),vxlan_sys_4789,set(tunnel(...)),vxlan_sys_4789,...
> > > 
> > >      BUG: kernel NULL pointer dereference, address: 0000000000000270
> > >      RIP: 0010:del_sw_hw_rule+0x29/0x190 [mlx5_core]
> 
> Hello,
> 
> I'll help you find and fix the problem.
> Your core dump analysis was very useful, but not sufficient to find the
> cause of the crash. Would you mind sharing a set of reproduction steps
> so we can debug this further?
> 
> Thank you,
> Cosmin.

Hi Cosmin,

Thanks for your reply.

It's hard to reproduce the crash directly.  In our case the rule forwards ip
broadcast traffic to 5 vxlan remotes. And driver creates 6 mlx5_flow_rule
which include 5 mlx5_pkt_reformat and 1 counter.
It triggers only when two *dr_action in struct mlx5_pkt_reformat have same
lower 32 bits, which determined by memory allocation.

Is it possible that we do some fault injection in unit test to reproduce?

Best regards,
Tao