Re: [PATCH 1/3] spi: spi-geni-qcom: Avoid clock setting if not needed

[Doug Anderson <dianders@xxxxxxxxxxxx>]

Hi,

On Wed, Jul 8, 2020 at 3:01 AM Mark Brown <broonie@xxxxxxxxxx> wrote:
>
> On Tue, Jul 07, 2020 at 05:53:01AM -0700, Doug Anderson wrote:
> > On Tue, Jul 7, 2020 at 5:08 AM Mark Brown <broonie@xxxxxxxxxx> wrote:
>
> > > This doesn't apply against current code, please check and resend.
>
> > As mentioned in the cover letter, I posted this series against the
> > Qualcomm tree.  The commit that it is fixing landed there with your
> > Ack so I was hoping this series could land in the Qualcomm tree with
> > your Ack as well.  Would that be OK?
>
> So I didn't see this until after the patch I applied was queued...  it's
> looking like it would be good to have a cross-tree merge with the
> Qualcomm tree if there's stuff like this - is this on a branch which
> makes that practical?  Otherwise I guess...

It's not too bad.  Of the 5 patches I've sent out (3 for geni SPI, 2
for quad SPI) you've landed just one.  Here's the summary:

a) geni SPI 1/3 (Avoid clock setting): Has your Ack.
b) geni SPI 2/3 (autosuspend delay): Landed in SPI tree
c) geni SPI 3/3 (overhead in prepare_message): Has your Ack.

d) quad SPI 1/2 (Avoid clock setting): Needs your Ack.
e) quad SPI 2/2 (autosuspend delay): Needs your Ack.

Since b) has already landed in your tree, let's just leave it there.
There'll be a bit of a performance hit in the Qualcomm tree, but it'll
still be usable.

Since the rest haven't landed, it would be nice to just land them in
the Qualcomm tree.


I think there's still more work to make the Geni SPI driver more
optimized, but I don't think it'll be as urgent as those patches and I
feel like any more major work could wait a cycle.


-Doug