Re: [RFC net-next v2 5/5] net: phy: nxp-c45-tja11xx: implement mdo_insert_tx_tag

[Sabrina Dubroca]

2023-09-01, 17:22:49 +0300, Radu Pirea (OSS) wrote:
> 
> 
> On 01.09.2023 16:57, Sabrina Dubroca wrote:
> > 2023-09-01, 14:58:12 +0300, Radu Pirea (OSS) wrote:
> > > On 01.09.2023 13:07, Sabrina Dubroca wrote:
> > > > > (the interface was up before)
> > > > > [root@alarm ~]# ip link add link end0 macsec0 type macsec encrypt on
> > > > > offload phy
> > > > > [root@alarm ~]# ip link set end0 down
> > > > > [root@alarm ~]# ip macsec add macsec0 rx port 1 address
> > > > > 00:01:be:be:ef:33
> > > > > RTNETLINK answers: Operation not supported
> > > > 
> > > > Where does that EOPNOTSUPP come from? nxp_c45_mdo_add_rxsc from this
> > > > version of the code can't return that, and macsec_add_rxsc also
> > > > shouldn't at this point.
> > > 
> > > This is the source of -EOPNOTSUPP
> > > https://elixir.bootlin.com/linux/latest/source/drivers/net/macsec.c#L1928
> > 
> > Could you check which part of macsec_get_ops is failing? Since
> > macsec_newlink with "offload phy" worked, macsec_check_offload
> > shouldn't fail, so why does macsec_get_ops return NULL?
> > real_dev->phydev was NULL'ed?
> 
> This check logical and returns false:
> https://elixir.bootlin.com/linux/latest/source/drivers/net/macsec.c#L343
> 
> real_dev->phydev was nulled.
> The call stack is next:
> fec_enet_close -> phy_disconnect -> phy_detach ->
> https://elixir.bootlin.com/linux/latest/source/drivers/net/phy/phy_device.c#L1815

Ok, thanks for looking this up. So we can't have a consistent behavior
between SW and PHY modes unfortunately.

-- 
Sabrina