[PATCH 2/2] x86/config/efi: drop DRIVER_SERIAL_NS16550

[Sascha Hauer]

On Mon, Oct 26, 2020 at 10:07:19PM +0100, Ahmad Fatoum wrote:
> 
> 
> On 10/26/20 7:06 PM, Michael Olbrich wrote:
> > On Tue, Oct 20, 2020 at 09:35:30AM +0200, Ahmad Fatoum wrote:
> >> On 10/20/20 9:26 AM, Uwe Kleine-König wrote:
> >>> On Tue, Oct 20, 2020 at 07:08:47AM +0200, Ahmad Fatoum wrote:
> >>>> Hello Uwe,
> >>>>
> >>>> On 9/18/20 1:04 PM, Uwe Kleine-König wrote:
> >>>>> For most machines having DRIVER_SERIAL_NS16550 and DRIVER_SERIAL_EFI_STDIO
> >>>>> both enabled results in the output appearing twice on the serial output.
> >>>>> So remove disable the lowlevel driver and rely on the EFI bios for serial.
> >>>
> >>> How embarrassing, s/remove //.
> >>>
> >>>> On the flip-side, some systems now won't output anything over serial with
> >>>> the defconfig, like the one I just ran into. Maybe we should keep both
> >>>> enabled, but have a quirk table, which boards to disable the 16550 for?
> >>>
> >>> Expecting that most systems are happy with EFI_STDIO I would suggest a
> >>> quirk table for those systems which are not?!
> >>
> >> Doubled output is better than no output, so a defconfig that
> >> by default displays anything if the BIOS is not in the quirk
> >> table sounds like it's the better choice?
> > 
> > Maybe disable CONSOLE_ACTIVATE_ALL instead? Then you should only get the
> > output from one driver.
> 
> And which one should be default? You'll get more compile time coverage
> with the defconfig but the user experience problem of having a defconfig
> that does print nothing at all on some devices is still there.

Maybe there is some way to find out where DRIVER_SERIAL_EFI_STDIO
directs it output to?

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |