Re: [PATCH] Revert "spi: gpio: Don't request CS GPIO in DT use-case"

[Linus Walleij <linus.walleij@xxxxxxxxxx>]

On Tue, Jul 2, 2019 at 1:39 PM Mark Brown <broonie@xxxxxxxxxx> wrote:
> On Mon, Jul 01, 2019 at 07:25:17PM +0200, Linus Walleij wrote:
> > This reverts commit 249e2632dcd0509b8f8f296f5aabf4d48dfd6da8.
> >
> > After this commit drivers/net/dsa/vitesse-vsc73xx.c stopped
> > working. Apparently CS is not working because the reads
> > from the device is just returning just 1:s or just 0:s at
> > all bisection points, so it is a complete regression and
> > I think spi-gpio CS is essentially broken.
>
> Also I'm a bit concerned that nobody else noticed this - do we
> have any understanding of what the actual problem is?

I am trying to rootcause it, I suspect I was just the first to try this
on real hardware actually.

The users in the device trees in the kernel
mostly point to consumers like LEDs, displays and switches
that wouldn't cause boot regressions
instead you must observe more elaborate aspects of the system
so that is why the automatic test farms cannot see it.

Yours,
Linus Walleij