Re: [PATCH v5 3/7] drm/msm/dpu: add DPU_PINGPONG_DSC bits into PP_BLK and PP_BLK_TE marcos

[Abhinav Kumar]

On 5/4/2023 12:59 PM, Dmitry Baryshkov wrote:
> On 04/05/2023 22:50, Abhinav Kumar wrote:
> > On 5/4/2023 12:36 PM, Marijn Suijten wrote:
> > > On 2023-05-04 11:25:44, Abhinav Kumar wrote:
> > > <snip>
> > > > > Sure, if you really prefer a split I'd go for two patches:
> > > > > 1. Add the flag to the enum and catalog;
> > > > > 2. Add the ops guard (functional change).
> > > > >
> > > > > Then don't forget to reword the commit message, following the 
> > > > > guidelines
> > > > > below and the suggestion for 2/7.
> > > > >
> > > > > - Marijn
> > > >
> > > > Plan sounds good to me.
> > > >
> > > > Marijn, we will wait for a couple of days to post the next rev but 
> > > > would
> > > > be hard more than that as we need to pick up other things which are
> > > > pending on top of this. Hence would appreciate if you can finish 
> > > > reviews
> > > > by then.
> > >
> > > It depends on how many more revisions are needed after that, and not all
> > > patches in this series have an r-b just yet.  Given the amount of review
> > > comments that are still trickling in (also on patches that already have
> > > maintainer r-b) I don't think we're quite there to start thinging about
> > > picking this up in drm-msm just yet.  I doubt anyone wants a repeat of
> > > the original DSC series, which went through many review rounds yet still
> > > required multiple series of bugfixes (some of which were pointed out and
> > > ignored in review) to be brought to a working state.  But the split
> > > across topics per series already makes this a lot less likely, many
> > > thanks for that.
> >
> > I think the outstanding comments shouldnt last more than 1-2 revs more 
> > on this one as its mostly due to multiple patches on the list touching 
> > catalog at the same time. I have been monitoring the comments closely 
> > even though I dont respond to all of them.
> >
> > One of the major reasons of the number of issues with DSC 1.1 was QC 
> > didn't really have the devices or panels to support it. Thats why I 
> > changed that this time around to take more control of validation of 
> > DSC 1.2 and ofcourse decided to break up of series into the least 
> > amount of functionality needed to keep the DPU driver intact.
> >
> > All that being said, we still value your comments and would gladly 
> > wait for a couple of days like I already wrote. But there are more 
> > incremental series on top of this:
> >
> > -> DSI changes for DSC 1.2
> > -> proper teardown for DSC
> > -> DSC pair allocation support
> > -> DSC 1.2 over DP
> >
> > We will be posting all of these within next couple of weeks on top of 
> > this.
>
> I'd say, it's fine to post them now, as we have more or less agreed on 
> the helper series. The interface between the series should be stable then.
>
> The RM series is probably the one having bigger dependencies/conflicts 
> on other pending patches (include virtual wide planes)

1 is already posted, will keep fixing review comments
2 will be posted pretty soon

DSC1.2 over DSI will be complete with this set.

I will finish up virtual planes review by early next week. Already 
underway ...

3 & 4 will be posted soon after that.

> > > In other words, let's take it slow and do things properly this time. And
> > > who knows, perhaps the rest of these patches are more straightforward.
> >
> > Ack. the intent is always to do things right the first time.
> >
> > > - Marijn
> > >
> > > <snip>