[Linux-stm32] [PATCH net-next 5/8] net: dpaa2-mac: use resolved link config in mac_link_up()

[Ioana Ciornei]

> Subject: Re: [PATCH net-next 5/8] net: dpaa2-mac: use resolved link config in
> mac_link_up()
> 
> On Tue, Feb 25, 2020 at 04:36:32PM +0000, Ioana Ciornei wrote:
> > > Subject: [PATCH net-next 5/8] net: dpaa2-mac: use resolved link
> > > config in
> > > mac_link_up()
> > >
> > > Convert the DPAA2 ethernet driver to use the finalised link
> > > parameters in
> > > mac_link_up() rather than the parameters in mac_config(), which are
> > > more suited to the needs of the DPAA2 MC firmware than those
> > > available via mac_config().
> > >
> > > Signed-off-by: Russell King <rmk+kernel at armlinux.org.uk>
> >
> > Tested-by: Ioana Ciornei <ioana.ciornei at nxp.com>
> 
> Thanks.
> 
> > > +
> > > +		/* This is lossy; the firmware really should take the pause
> > > +		 * enablement status rather than pause/asym pause status.
> > > +		 */
> >
> > In what sense it's lossy? I cannot see how information can be lost by
> translating the rx/tx pause state to pause/asym.
> > If it's just about the unnecessary double translation, then I agree.. this could
> have been done in an easier manner.
> 
> If you're just translating rx/tx to pause/asym and then doing the reverse, then it
> isn't lossy, but if the firmware is resolving pause/asym according to the table in
> IEEE 802.3, then it will be lossy.

The firmware is just doing the reverse translation.

> 
> If the firmware doesn't interpret the bits, then why not do the sensible thing and
> just pass the enablement status rather than trying to confusingly encode it back
> to pause/asym?

I agree. It's unnecessary and confusing. I'll add this on the list of fixups to be made by the firmware team.

Ioana

> 
> --