[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240924082612.GF4029621@kernel.org>
Date: Tue, 24 Sep 2024 09:26:12 +0100
From: Simon Horman <horms@...nel.org>
To: Kory Maincent <kory.maincent@...tlin.com>
Cc: Jakub Kicinski <kuba@...nel.org>, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, Kyle Swenson <kyle.swenson@....tech>,
thomas.petazzoni@...tlin.com,
Oleksij Rempel <o.rempel@...gutronix.de>,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>, Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net] net: pse-pd: tps23881: Fix boolean evaluation for
bitmask checks
On Tue, Sep 24, 2024 at 10:15:29AM +0200, Kory Maincent wrote:
> On Tue, 24 Sep 2024 08:18:39 +0100
> Simon Horman <horms@...nel.org> wrote:
>
> > On Mon, Sep 23, 2024 at 05:34:26PM +0200, Kory Maincent wrote:
> > [...]
> >
> > Thanks Kory,
> >
> > I agree that these changes are correct.
> > But are they fixes; can this manifest in a bug?
>
> I didn't face it but I think yes.
> In case of a 4 pairs PoE ports without the fix:
>
> chan = priv->port[id].chan[0];
> if (chan < 4) {
> enabled = ret & BIT(chan);
> delivering = ret & BIT(chan + 4);
> ...
> }
>
> if (priv->port[id].is_4p) {
> chan = priv->port[id].chan[1];
> if (chan < 4) {
> enabled &= !!(ret & BIT(chan));
> delivering &= !!(ret & BIT(chan + 4));
>
> If enabled = 0x2 here, enabled would be assigned to 0 instead of 1.
> ...
>
> }
> }
>
> I have an issue using 4pairs PoE port with my board so I can't test it.
>
>
> > (If so, I suspect the Kernel is riddled with such bugs.)
>
> Don't know about it but if I can remove it from my driver it would be nice. :)
Right, no question from my side that this change is a good one.
I'm just wondering if it is best for net or net-next.
Powered by blists - more mailing lists