On Wed, Dec 14, 2022 at 01:52:17PM +0100, Krzysztof Kozlowski wrote:
> On 14/12/2022 13:30, Brian Masney wrote:
> > I triple checked that I have the QUP pins defined correctly for the 5
> > buses. I checked them against what's in the downstream kernel and I also
> > checked them against what's in upstream's
> > drivers/pinctrl/qcom/pinctrl-sc8280xp.c. This is the pin mapping that I
>
> What's the base of this kernel? Are you sure you have d21f4b7ffc22?
I'm based on top of linux-next-20221208 with no other changes. I have
that commit.
commit d21f4b7ffc22c009da925046b69b15af08de9d75
Author: Douglas Anderson <dianders@xxxxxxxxxxxx>
Date: Fri Oct 14 10:33:18 2022 -0700
pinctrl: qcom: Avoid glitching lines when we first mux to output
On Wed, Dec 14, 2022 at 01:53:38PM +0100, Konrad Dybcio wrote:
> > This is the style where i2cdetect seems to be happy for all 5 buses and
> > is fast:
> >
> > i2c0_default: i2c0-default-state {
> > mux-pins {
> > pins = "gpio135", "gpio136";
> > function = "qup0";
> > };
> >
> > config-pins {
> > pins = "gpio135", "gpio136";
> > drive-strength = <2>;
> > bias-pull-up;
> > };
> > };
> Unless you made a typo somewhere, I genuinely have no explanation for this..
I have my unpublished v2 patch set committed to my tree and a clean tree
according to git. I started with the state that I have quoted above. As I
did the various tests I described in my last email, I would do a
'git diff' just to be sure that I didn't have any typos.
I'll wait to hear from Shazad about whether or not the output that I got
from i2cdetect is supposed to be the same for those 5 buses.
Brian
