[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Yn5TWIZPBVLFipVD@shredder>
Date: Fri, 13 May 2022 15:47:20 +0300
From: Ido Schimmel <idosch@...dia.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: Vladimir Oltean <vladimir.oltean@....com>,
netdev <netdev@...r.kernel.org>,
Nikolay Aleksandrov <razor@...ckwall.org>,
Ido Schimmel <idosch@...lanox.com>,
"bridge@...ts.linux-foundation.org"
<bridge@...ts.linux-foundation.org>
Subject: Re: [PATCH RFC] net: bridge: Clear offload_fwd_mark when passing
frame up bridge interface.
On Thu, May 12, 2022 at 10:38:03PM +0200, Andrew Lunn wrote:
> > I like Andrew's patch because it is the Rx equivalent of
> > br_switchdev_frame_unmark() in br_dev_xmit(). However, if we go with the
> > second option, it should allow us to remove the clearing of the mark in
> > the Tx path as the control block is cleared in the Tx path since commit
> > fd65e5a95d08 ("net: bridge: clear bridge's private skb space on xmit").
> >
> > I don't know how far back Nik's patch was backported and I don't know
> > how far back Andrew's patch will be backported, so it might be best to
> > submit Andrew's patch to net as-is and then in net-next change
> > nbp_switchdev_allowed_egress() and remove br_switchdev_frame_unmark()
> > from both the Rx and Tx paths.
> >
> > Anyway, I have applied this patch to our tree for testing. Will report
> > tomorrow in case there are any regressions.
>
> Hi Ido
>
> Did your testing find any issues?
No, patch is fine. Thanks!
Powered by blists - more mailing lists