Re: [PATCH 6.1.y] virtio_net: clamp rss_max_key_size to NETDEV_RSS_KEY_LEN

["Michael S. Tsirkin" <mst@redhat.com>]

On Wed, Apr 08, 2026 at 10:24:14AM -0400, Sasha Levin wrote:
> On Wed, Apr 08, 2026 at 10:20:50AM -0400, Michael S. Tsirkin wrote:
> > On Wed, Apr 08, 2026 at 10:15:48AM -0400, Sasha Levin wrote:
> > > On Wed, Apr 08, 2026 at 09:54:33AM -0400, Michael S. Tsirkin wrote:
> > > > On Wed, Apr 08, 2026 at 09:43:51AM -0400, Sasha Levin wrote:
> > > > > From: Srujana Challa <schalla@marvell.com>
> > > > >
> > > > > [ Upstream commit b4e5f04c58a29c499faa85d12952ca9a4faf1cb9 ]
> > > > >
> > > > > rss_max_key_size in the virtio spec is the maximum key size supported by
> > > > > the device, not a mandatory size the driver must use. Also the value 40
> > > > > is a spec minimum, not a spec maximum.
> > > > >
> > > > > The current code rejects RSS and can fail probe when the device reports a
> > > > > larger rss_max_key_size than the driver buffer limit. Instead, clamp the
> > > > > effective key length to min(device rss_max_key_size, NETDEV_RSS_KEY_LEN)
> > > > > and keep RSS enabled.
> > > > >
> > > > > This keeps probe working on devices that advertise larger maximum key sizes
> > > > > while respecting the netdev RSS key buffer size limit.
> > > > >
> > > > > Fixes: 3f7d9c1964fc ("virtio_net: Add hash_key_length check")
> > > > > Cc: stable@vger.kernel.org
> > > > > Signed-off-by: Srujana Challa <schalla@marvell.com>
> > > > > Acked-by: Michael S. Tsirkin <mst@redhat.com>
> > > > > Link: https://patch.msgid.link/20260326142344.1171317-1-schalla@marvell.com
> > > > > Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> > > > > [ changed clamp target from NETDEV_RSS_KEY_LEN to VIRTIO_NET_RSS_MAX_KEY_SIZE ]
> > > >
> > > > Does this not make the subject and the commit log misleading?
> > > 
> > > Probably, but changing the commit subject will just create more confusion.
> > > 
> > > --
> > > Thanks,
> > > Sasha
> > 
> > It's not just the subject. The commit log says:
> > 
> > 	Also the value 40 is a spec minimum, not a spec maximum.
> > 
> > but the changed patch seems to treat it as a maximum:
> > 
> > +               vi->rss_key_size = min_t(u16, key_sz, VIRTIO_NET_RSS_MAX_KEY_SIZE);
> > 
> > 
> > so unless I misread the code, the value is never > 40.
> 
> I tried to explain it here: https://lore.kernel.org/all/adZitVex9UGVyH-V@laps/
> 
> -- 
> Thanks,
> Sasha


OK let us take the discussion there.