All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
To: Abhinav Kumar <quic_abhinavk@quicinc.com>
Cc: Kalyan Thota <quic_kalyant@quicinc.com>,
	freedreno@lists.freedesktop.org,
	Bjorn Andersson <andersson@kernel.org>,
	dri-devel@lists.freedesktop.org,
	Stephen Boyd <swboyd@chromium.org>,
	linux-arm-msm@vger.kernel.org, Sean Paul <sean@poorly.run>
Subject: Re: [Freedreno] [PATCH v3 27/27] drm/msm/dpu: add support for wide planes
Date: Fri, 10 Feb 2023 02:09:31 +0200	[thread overview]
Message-ID: <CAA8EJppuM97q2uc04_id7fEY26WHbnb=Mc0eAU1sDMTNwguAYg@mail.gmail.com> (raw)
In-Reply-To: <396c388b-de81-6794-d846-636234f1edb0@quicinc.com>

 .

On Fri, 10 Feb 2023 at 00:12, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
>
> Hi Dmitry
>
> On 2/9/2023 1:23 PM, Dmitry Baryshkov wrote:
> > Hi Abhinav,
> >
> > On Thu, 9 Feb 2023 at 21:25, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
> >>
> >>
> >>
> >> On 2/9/2023 3:45 AM, Dmitry Baryshkov wrote:
> >>> On Thu, 9 Feb 2023 at 04:19, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
> >>>>
> >>>>
> >>>>
> >>>> On 2/3/2023 10:21 AM, Dmitry Baryshkov wrote:
> >>>>> Typically SSPP can support rectangle with width up to 2560. However it's
> >>>>
> >>>> Not always 2560. Depends on the chipset.
> >>>
> >>> _typically_
> >>>
> >>
> >> Would just say maxlinewidth of SSPP instead of giving some hardcoded number.
> >
> > Ack.
> >
> >>
> >>>>
> >>>>> possible to use multirect feature and split source to use the SSPP to
> >>>>> output two consecutive rectangles. This commit brings in this capability
> >>>>> to support wider screen resolutions.
> >>>>>
> >>>>> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> >>>>> ---
> >>>>>     drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c  |   6 ++
> >>>>>     drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c | 116 +++++++++++++++++++---
> >>>>>     drivers/gpu/drm/msm/disp/dpu1/dpu_plane.h |   4 +
> >>>>>     3 files changed, 114 insertions(+), 12 deletions(-)
> >>>>>
> >>>>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> >>>>> index 0ca3bc38ff7e..867832a752b2 100644
> >>>>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> >>>>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> >>>>> @@ -485,6 +485,12 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc,
> >>>>>                                            fetch_active,
> >>>>>                                            &pstate->pipe);
> >>>>>
> >>>>> +             _dpu_crtc_blend_setup_pipe(crtc, plane,
> >>>>> +                                        mixer, cstate->num_mixers,
> >>>>> +                                        stage_cfg, pstate->stage, 1,
> >>>>> +                                        fetch_active,
> >>>>> +                                        &pstate->r_pipe);
> >>>>> +
> >>>>>                 /* blend config update */
> >>>>>                 for (lm_idx = 0; lm_idx < cstate->num_mixers; lm_idx++) {
> >>>>>                         _dpu_crtc_setup_blend_cfg(mixer + lm_idx, pstate, format);
> >>>>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
> >>>>> index e2e85688ed3c..401ead64c6bd 100644
> >>>>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
> >>>>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
> >>>>> @@ -365,6 +365,9 @@ static void _dpu_plane_set_qos_ctrl(struct drm_plane *plane,
> >>>>>         struct dpu_plane *pdpu = to_dpu_plane(plane);
> >>>>>         struct dpu_hw_pipe_qos_cfg pipe_qos_cfg;
> >>>>>
> >>>>> +     if (!pipe->sspp)
> >>>>> +             return;
> >>>>> +
> >>>>>         memset(&pipe_qos_cfg, 0, sizeof(pipe_qos_cfg));
> >>>>>
> >>>>>         if (flags & DPU_PLANE_QOS_VBLANK_CTRL) {
> >>>>> @@ -647,6 +650,9 @@ static int _dpu_plane_color_fill_pipe(struct dpu_plane_state *pstate,
> >>>>>     {
> >>>>>         struct dpu_hw_sspp_cfg pipe_cfg;
> >>>>>
> >>>>> +     if (!pipe->sspp)
> >>>>> +             return 0;
> >>>>
> >>>> instead of checking if sspp was present, is it not better for the caller
> >>>> to check if the rpipe is valid before calling this?
> >>>>
> >>>>> +
> >>>>>         /* update sspp */
> >>>>>         if (!pipe->sspp->ops.setup_solidfill)
> >>>>>                 return 0;
> >>>>> @@ -701,6 +707,8 @@ static void _dpu_plane_color_fill(struct dpu_plane *pdpu,
> >>>>>
> >>>>>         /* update sspp */
> >>>>>         _dpu_plane_color_fill_pipe(pstate, &pstate->pipe, &pstate->pipe_cfg, fill_color, fmt);
> >>>>> +
> >>>>> +     _dpu_plane_color_fill_pipe(pstate, &pstate->r_pipe, &pstate->r_pipe_cfg, fill_color, fmt);
> >>>>>     }
> >>>>
> >>>> So cant we do
> >>>>
> >>>> if (pstate->r_pipe.sspp)
> >>>>           _dpu_plane_color_fill_pipe(pstate, &pstate->r_pipe,
> >>>>                   &pstate->r_pipe_cfg, fill_color, fmt);
> >>>>
> >>>> It just seems better to me as the caller would already know if the sspp
> >>>> was assigned.
> >>>
> >>>    I think I had this kind of code earlier, but then I found it more
> >>> logical to move the check to the called function. I'll move it back.
> >>>
> >>>>
> >>>>>
> >>>>>     int dpu_plane_validate_multirect_v2(struct dpu_multirect_plane_states *plane)
> >>>>> @@ -911,6 +919,9 @@ static int dpu_plane_atomic_check_pipe(struct dpu_plane *pdpu,
> >>>>>     {
> >>>>>         uint32_t min_src_size;
> >>>>>
> >>>>> +     if (!pipe->sspp)
> >>>>> +             return 0;
> >>>>> +
> >>>>>         min_src_size = DPU_FORMAT_IS_YUV(fmt) ? 2 : 1;
> >>>>>
> >>>>>         if (DPU_FORMAT_IS_YUV(fmt) &&
> >>>>> @@ -957,9 +968,12 @@ static int dpu_plane_atomic_check(struct drm_plane *plane,
> >>>>>         int ret = 0, min_scale;
> >>>>>         struct dpu_plane *pdpu = to_dpu_plane(plane);
> >>>>>         struct dpu_plane_state *pstate = to_dpu_plane_state(new_plane_state);
> >>>>> +     struct dpu_sw_pipe *pipe = &pstate->pipe;
> >>>>> +     struct dpu_sw_pipe *r_pipe = &pstate->r_pipe;
> >>>>>         const struct drm_crtc_state *crtc_state = NULL;
> >>>>>         const struct dpu_format *fmt;
> >>>>>         struct dpu_hw_sspp_cfg *pipe_cfg = &pstate->pipe_cfg;
> >>>>> +     struct dpu_hw_sspp_cfg *r_pipe_cfg = &pstate->r_pipe_cfg;
> >>>>>         struct drm_rect fb_rect = { 0 };
> >>>>>         uint32_t max_linewidth;
> >>>>>         unsigned int rotation;
> >>>>> @@ -983,8 +997,11 @@ static int dpu_plane_atomic_check(struct drm_plane *plane,
> >>>>>         if (!new_plane_state->visible)
> >>>>>                 return 0;
> >>>>>
> >>>>> -     pstate->pipe.multirect_index = DPU_SSPP_RECT_SOLO;
> >>>>> -     pstate->pipe.multirect_mode = DPU_SSPP_MULTIRECT_NONE;
> >>>>> +     pipe->multirect_index = DPU_SSPP_RECT_SOLO;
> >>>>> +     pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE;
> >>>>> +     r_pipe->multirect_index = DPU_SSPP_RECT_SOLO;
> >>>>> +     r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE;
> >>>>> +     r_pipe->sspp = NULL;
> >>>>>
> >>>>>         pstate->stage = DPU_STAGE_0 + pstate->base.normalized_zpos;
> >>>>>         if (pstate->stage >= pdpu->catalog->caps->max_mixer_blendstages) {
> >>>>> @@ -1016,16 +1033,53 @@ static int dpu_plane_atomic_check(struct drm_plane *plane,
> >>>>>
> >>>>>         max_linewidth = pdpu->catalog->caps->max_linewidth;
> >>>>>
> >>>>> -     /* check decimated source width */
> >>>>>         if (drm_rect_width(&pipe_cfg->src_rect) > max_linewidth) {
> >>>>> -             DPU_DEBUG_PLANE(pdpu, "invalid src " DRM_RECT_FMT " line:%u\n",
> >>>>> -                             DRM_RECT_ARG(&pipe_cfg->src_rect), max_linewidth);
> >>>>> -             return -E2BIG;
> >>>>> +             /* struct dpu_crtc_state *cstate = to_dpu_crtc_state(crtc_state); */
> >>>>> +
> >>>>> +             if (drm_rect_width(&pipe_cfg->src_rect) > 2 * max_linewidth) {
> >>>>> +                     DPU_DEBUG_PLANE(pdpu, "invalid src " DRM_RECT_FMT " line:%u\n",
> >>>>> +                                     DRM_RECT_ARG(&pipe_cfg->src_rect), max_linewidth);
> >>>>> +                     return -E2BIG;
> >>>>> +             }
> >>>>
> >>>> This is where I am a bit concerned enabling it for all chipsets in one go.
> >>>
> >>> As I wrote earlier, I'd prefer the opt-out rather than opt-in here. It
> >>> is much easier to handle the reports "I have a device with sm6543,
> >>> where the display worked before 6.4, but started failing afterwards"
> >>> rather than trying to find a person with sm6543 and asking him if he
> >>> can enable this and that on his device. And even a lower chance of a
> >>> person with sm6543 coming up with a patch 'hey, I enabled this for my
> >>> phone and it works!'.
> >>>
> >>> If we find any issues during or close to the end of the development
> >>> cycle, we can add a 'don't enable wide plane here' switch and enable
> >>> it for failing platforms. But each enablement of this switch should
> >>> come with a reason (wide planes not working here because ....). In the
> >>> end this switch should be gone and transformed into proper HW
> >>> limitation checks.
> >>>
> >>
> >> As it has become clear that with this patch series 4K with UBWC cannot
> >> be supported without true virtual planes (with two SSPPs), why do you
> >> need to relax this check right now?
> >
> > Yes. It enables support for 4k @ linear formats. So my plan for this
> > series is to land 4k with all the proper applicable restrictions.
> >
> >> You can relax this when you add the support for virtual planes till then
> >> let it be this way.
> >>
> >> Its not going to break smartDMA as such. You can still use it for layers
> >> < 2560.
> >>
> >> That way we stay true to the purpose of the feature. I think originally
> >> you wanted to get this in for smartDMA and not to support wide plane and
> >> that purpose will still be achieved even with keeping this check intact.
> >
> > Actually, no. With this series I wanted to get 4k. It was developed in
> > parallel with the 4k enablement for RB3 (posted, bridge patches are
> > being merged for 6.3) and RB5 (delayed for now, I have other issues
> > there).
> >
>
> With the UBWC related checks, this wont support 4K for UBWC layers which
> is default on QC chipsets. So I am fine with respect to that. But still
> this does not address the product spec advertized modes. Like I
> mentioned before, relaxing the maxlinewidth check with the added UBWC
> checks is fine from DPU point of view but not from the product POV.
>
> As things stand today, this is the only check failing the 4K modes on
> chipsets which shouldnt support 4k (linear or UBWC doesnt matter).
>
> >> You can relax it in the virtual plane series.
> >>
> >> Regarding issues, this is where it gets tricky. We should be aligning
> >> with what the product supports. QC will not support issues arising with
> >> 4K on chipsets on which 4K is not advertized.
> >
> > So, we have several different items here:
> > - SmartDMA v2 per se, supporting two rectangles per VIG or DMA plane,
> > - Source split support,
> > - Supporting 4k modes.
> >
> > I think we should tend them one by one. This series concerns SmartDMA
> > v2. Using SmartDMA it is possible to use two rectangles side by side
> > to emulate a wide plane. This series doesn't care at all about max
> > resolutions. These two items are completely orthogonal.
> >
>
> No its not orthogonal. Relaxing the maxlinewidth check in the
> atomic_check() will allow 4K layers now even on chipsets where 4K wasnt
> advertized. Linear or UBWC doesnt matter as the spec doesnt go into that.

Please correct my answers, if I got something wrong here:

Does sc7180 support SmartDMA? Yes it does.
Can QC or CrOS validate SmartDMA separately on sc7180? I hope you can.
Should the hw-supported feature be enabled? Yes, it should.

Now limiting out 4k by not supporting SmartDMA looks like a misfeature.
I can only suggest sending a change to block 4k on sc7180.

> >>>> As you are aware,  we have an open bug today that we do not filter out
> >>>> the modes which we do not support.
> >>>>
> >>>> https://gitlab.freedesktop.org/drm/msm/-/issues/21
> >>>
> >>> I thought that with the link-frequencies in place and with the DSI
> >>> checking the OPP tables this issue is mostly handled. Isn't it?
> >>> Is a mode check in the DPU driver itself the last missing piece?
> >>>
> >>
> >> opp based checking was implemented only for DSI. That one is byte clk based.
> >>
> >> DP uses link rate for opp table.
> >>
> >> Even with a 5.4G link rate (the one in sc7180 chromebook) 4k@30 would
> >> still be possible but it was not advertized
> >>
> >> https://www.qualcomm.com/content/dam/qcomm-martech/dm-assets/documents/prod_brief_qcom_sd7c.pdf
> >>
> >> These docs are available in public domain.
> >>
> >> As we synced up last time on
> >> https://patchwork.freedesktop.org/series/107917/, even with these limits
> >> in place, its not matching the advertized limits.
> >>
> >>>>
> >>>> Due to this, on all chipsets we will end up trying to do a 4K on
> >>>> external display which we dont know what bugs it will expose.
> >>>
> >>> If we do not expose bugs, we do not have a way to fix them. And I
> >>> definitely think that all the bugs should be listed as early as
> >>> possible, while both of us still remember the code under the question.
> >>>
> >>
> >> Yes but on chipsets where 4K is supported ( and hence needed ).
> >
> > 4k, SmartDMA, src-split, split-display, etc.
> >
>
> The visual issues reported on sdm845 on the other thread are a classic
> example of what I just wrote on that patchset and thats why I was
> emphasizing a visual validation OR in other words enable the feature on
> which you are able to visually validate it.

Yes. And if we did not enable the feature, Amit would not be able to
spot that. I can repeat my suggestion:

- prevalidate these features for the accessible platforms (e.g. I only
have sdm845, sm8250 and sm8350 at hand)
- enable SmartDMA for all chipsets where SmartDMA is supported
- collect possible tested-by and broken-at reports, while the patches
sit in linux-next
- disable SmartDMA basing on the feedback from the previous step (e.g.
select from 'mostly disable', 'disable for the bugged cases', 'do not
disable at all', etc).
  I can promise that if we see a significant validation failure rate I
will not oppose disabling SmartDMA.

As a reminder: if the patchset is ready at the time of 6.3-rc1 (in
three weeks from now), it is going to be merged into linux-next first,
after that it can go into the main Linus'es tree at 6.4-rc1. So we
will have _two_ kernel cycles to collect bug reports and to disable
(or fix) broken cases.

> We can evaluate and enable smartDMA on other chipsets on a need basis.
>
> We discussed this again even today in the team discussion. Our team's
> PoV doesnt change. We would still like to enable smartDMA only on
> chipsets which can be visually validated first to limit the debugging
> effort to one chipset first and then perfect it. Otherwise its too much
> effort on QC side to debug those issues on all chipsets.

I think QC mostly debugs issues on sc7180/sc7280 and sometimes on
sdm845/sm8250 (and now on sm8350). I think we can let people
(somainline, PmOS) test features on other platforms.
And testing happens better if we can say 'please test linux-next'
rather than 'please test linux-next + this patch to enable the feature
+ that patch to enable the second patch of the feature'.

> >>>> So lets say if we test it on sc7280 fully but not on sc7180, we will
> >>>> still hit this condition on sc7180 too but on that chipset we did not
> >>>> advertise 4K as a capability in the product spec.
> >>>
> >>> Is it 'not advertised' or 'not supported by hw'?
> >>>
> >>
> >> The document
> >> https://www.qualcomm.com/content/dam/qcomm-martech/dm-assets/documents/prod_brief_qcom_sd7c.pdf
> >> is made from inputs from not just display team but overall system
> >> limits. So even though you could argue that this falls within the
> >> display capabilities, all I can say at the moment is we have to stick to
> >> the advertized limits as its compiled with inputs from all the teams
> >> (system/performance etc).
> >
> > So, there should be a limiting factor (or a combination of them).
> > Filter out 4k modes on sc7180. Or modes using fill rate higher than N.
> > Pixel clock rate higher than M. But it has nothing to do with these
> > patches enabling SmartDMA support on this platform.
> >
> > Even if we look at the vendor kernels, we don't see 'maximum external
> > resolution'. Instead I see a combination of linewidth and bandwidth
> > limitations. If we can stick to that, that would be great.
> >
>
> Can you please point me to bandwidth limitation checks? How are other
> vendors coming up with this number? It has to be based on some
> resolution too right?

Usually it is considered in the other direction. The SoC can support
this-and-that pixel clock and bandwidth, so put the NxM resolution to
the datasheet as the max supported one.

Can you point out how the vendor kernel limits DP modes? I checked out
several dtsi files. For sm8250 the DP is limited to 1920x1080 (while
PB explicitly mentions 4k@60).
sm7125 is limited to 2560x1600. sm6150 again 1920x1080. From the pile
of the DTS that I have here the rest lists only
qcom,max-pclk-frequency-khz

> My RFC https://patchwork.freedesktop.org/series/107917/ considered pixel
> clk as the limiting factor which was posted after discussions
> internally. In the absence of another way, that remains the only
> solution to tackle this.

If I remember correctly the mentioned patchset used manually crafted
pixel clocks. And for example for sm8250 this clock doesn't correspond
to verifiable source.
Your patchset used 594000 KHz as max ext pclk for sm8250. vendor dtsi
lists 187500 KHz as maximum DP pclk for RB5 and 150000 KHz in all
other cases. And, at the same time, it lists a 3840x2160@60 mode for
the DSI/lt9611uxc with the pixel clock as high as 608040 KHz.

I suggest we stop the discussion at this point, unless there is
anything else wrong with this patchset itself.

I noted the point regarding UBWC & parallel mode. I will handle it in
the next iteration.

I noted your valid point about the visual verification. I proposed a
way to ease validation, allowing it to be enabled for testers and
early adopters for nearly two cycles. I hope you'd agree to this plan.
I on the other side agree to revert to opt-in if the failure rate is
high.

Please stop bringing max resolution issues to this patchserie. It must
be handled separately. I hope to see the mode filter patch targeting
sc7180 & sc7280. With the DSI opp check in place I think we should
concentrate on the DP case. If nothing else, I think even adding the
max PCLK to the msm_dp_desc should be sufficient to your worries. I'd
prefer to be able to override it for the particular board, but I think
this can come later, as it would require an agreement from the DT
schema team.

--
With best wishes
Dmitry

WARNING: multiple messages have this Message-ID (diff)
From: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
To: Abhinav Kumar <quic_abhinavk@quicinc.com>
Cc: Kalyan Thota <quic_kalyant@quicinc.com>,
	freedreno@lists.freedesktop.org, Sean Paul <sean@poorly.run>,
	Bjorn Andersson <andersson@kernel.org>,
	dri-devel@lists.freedesktop.org,
	Stephen Boyd <swboyd@chromium.org>,
	Rob Clark <robdclark@gmail.com>, Daniel Vetter <daniel@ffwll.ch>,
	linux-arm-msm@vger.kernel.org, David Airlie <airlied@gmail.com>
Subject: Re: [Freedreno] [PATCH v3 27/27] drm/msm/dpu: add support for wide planes
Date: Fri, 10 Feb 2023 02:09:31 +0200	[thread overview]
Message-ID: <CAA8EJppuM97q2uc04_id7fEY26WHbnb=Mc0eAU1sDMTNwguAYg@mail.gmail.com> (raw)
In-Reply-To: <396c388b-de81-6794-d846-636234f1edb0@quicinc.com>

 .

On Fri, 10 Feb 2023 at 00:12, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
>
> Hi Dmitry
>
> On 2/9/2023 1:23 PM, Dmitry Baryshkov wrote:
> > Hi Abhinav,
> >
> > On Thu, 9 Feb 2023 at 21:25, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
> >>
> >>
> >>
> >> On 2/9/2023 3:45 AM, Dmitry Baryshkov wrote:
> >>> On Thu, 9 Feb 2023 at 04:19, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
> >>>>
> >>>>
> >>>>
> >>>> On 2/3/2023 10:21 AM, Dmitry Baryshkov wrote:
> >>>>> Typically SSPP can support rectangle with width up to 2560. However it's
> >>>>
> >>>> Not always 2560. Depends on the chipset.
> >>>
> >>> _typically_
> >>>
> >>
> >> Would just say maxlinewidth of SSPP instead of giving some hardcoded number.
> >
> > Ack.
> >
> >>
> >>>>
> >>>>> possible to use multirect feature and split source to use the SSPP to
> >>>>> output two consecutive rectangles. This commit brings in this capability
> >>>>> to support wider screen resolutions.
> >>>>>
> >>>>> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> >>>>> ---
> >>>>>     drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c  |   6 ++
> >>>>>     drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c | 116 +++++++++++++++++++---
> >>>>>     drivers/gpu/drm/msm/disp/dpu1/dpu_plane.h |   4 +
> >>>>>     3 files changed, 114 insertions(+), 12 deletions(-)
> >>>>>
> >>>>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> >>>>> index 0ca3bc38ff7e..867832a752b2 100644
> >>>>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> >>>>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> >>>>> @@ -485,6 +485,12 @@ static void _dpu_crtc_blend_setup_mixer(struct drm_crtc *crtc,
> >>>>>                                            fetch_active,
> >>>>>                                            &pstate->pipe);
> >>>>>
> >>>>> +             _dpu_crtc_blend_setup_pipe(crtc, plane,
> >>>>> +                                        mixer, cstate->num_mixers,
> >>>>> +                                        stage_cfg, pstate->stage, 1,
> >>>>> +                                        fetch_active,
> >>>>> +                                        &pstate->r_pipe);
> >>>>> +
> >>>>>                 /* blend config update */
> >>>>>                 for (lm_idx = 0; lm_idx < cstate->num_mixers; lm_idx++) {
> >>>>>                         _dpu_crtc_setup_blend_cfg(mixer + lm_idx, pstate, format);
> >>>>> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
> >>>>> index e2e85688ed3c..401ead64c6bd 100644
> >>>>> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
> >>>>> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_plane.c
> >>>>> @@ -365,6 +365,9 @@ static void _dpu_plane_set_qos_ctrl(struct drm_plane *plane,
> >>>>>         struct dpu_plane *pdpu = to_dpu_plane(plane);
> >>>>>         struct dpu_hw_pipe_qos_cfg pipe_qos_cfg;
> >>>>>
> >>>>> +     if (!pipe->sspp)
> >>>>> +             return;
> >>>>> +
> >>>>>         memset(&pipe_qos_cfg, 0, sizeof(pipe_qos_cfg));
> >>>>>
> >>>>>         if (flags & DPU_PLANE_QOS_VBLANK_CTRL) {
> >>>>> @@ -647,6 +650,9 @@ static int _dpu_plane_color_fill_pipe(struct dpu_plane_state *pstate,
> >>>>>     {
> >>>>>         struct dpu_hw_sspp_cfg pipe_cfg;
> >>>>>
> >>>>> +     if (!pipe->sspp)
> >>>>> +             return 0;
> >>>>
> >>>> instead of checking if sspp was present, is it not better for the caller
> >>>> to check if the rpipe is valid before calling this?
> >>>>
> >>>>> +
> >>>>>         /* update sspp */
> >>>>>         if (!pipe->sspp->ops.setup_solidfill)
> >>>>>                 return 0;
> >>>>> @@ -701,6 +707,8 @@ static void _dpu_plane_color_fill(struct dpu_plane *pdpu,
> >>>>>
> >>>>>         /* update sspp */
> >>>>>         _dpu_plane_color_fill_pipe(pstate, &pstate->pipe, &pstate->pipe_cfg, fill_color, fmt);
> >>>>> +
> >>>>> +     _dpu_plane_color_fill_pipe(pstate, &pstate->r_pipe, &pstate->r_pipe_cfg, fill_color, fmt);
> >>>>>     }
> >>>>
> >>>> So cant we do
> >>>>
> >>>> if (pstate->r_pipe.sspp)
> >>>>           _dpu_plane_color_fill_pipe(pstate, &pstate->r_pipe,
> >>>>                   &pstate->r_pipe_cfg, fill_color, fmt);
> >>>>
> >>>> It just seems better to me as the caller would already know if the sspp
> >>>> was assigned.
> >>>
> >>>    I think I had this kind of code earlier, but then I found it more
> >>> logical to move the check to the called function. I'll move it back.
> >>>
> >>>>
> >>>>>
> >>>>>     int dpu_plane_validate_multirect_v2(struct dpu_multirect_plane_states *plane)
> >>>>> @@ -911,6 +919,9 @@ static int dpu_plane_atomic_check_pipe(struct dpu_plane *pdpu,
> >>>>>     {
> >>>>>         uint32_t min_src_size;
> >>>>>
> >>>>> +     if (!pipe->sspp)
> >>>>> +             return 0;
> >>>>> +
> >>>>>         min_src_size = DPU_FORMAT_IS_YUV(fmt) ? 2 : 1;
> >>>>>
> >>>>>         if (DPU_FORMAT_IS_YUV(fmt) &&
> >>>>> @@ -957,9 +968,12 @@ static int dpu_plane_atomic_check(struct drm_plane *plane,
> >>>>>         int ret = 0, min_scale;
> >>>>>         struct dpu_plane *pdpu = to_dpu_plane(plane);
> >>>>>         struct dpu_plane_state *pstate = to_dpu_plane_state(new_plane_state);
> >>>>> +     struct dpu_sw_pipe *pipe = &pstate->pipe;
> >>>>> +     struct dpu_sw_pipe *r_pipe = &pstate->r_pipe;
> >>>>>         const struct drm_crtc_state *crtc_state = NULL;
> >>>>>         const struct dpu_format *fmt;
> >>>>>         struct dpu_hw_sspp_cfg *pipe_cfg = &pstate->pipe_cfg;
> >>>>> +     struct dpu_hw_sspp_cfg *r_pipe_cfg = &pstate->r_pipe_cfg;
> >>>>>         struct drm_rect fb_rect = { 0 };
> >>>>>         uint32_t max_linewidth;
> >>>>>         unsigned int rotation;
> >>>>> @@ -983,8 +997,11 @@ static int dpu_plane_atomic_check(struct drm_plane *plane,
> >>>>>         if (!new_plane_state->visible)
> >>>>>                 return 0;
> >>>>>
> >>>>> -     pstate->pipe.multirect_index = DPU_SSPP_RECT_SOLO;
> >>>>> -     pstate->pipe.multirect_mode = DPU_SSPP_MULTIRECT_NONE;
> >>>>> +     pipe->multirect_index = DPU_SSPP_RECT_SOLO;
> >>>>> +     pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE;
> >>>>> +     r_pipe->multirect_index = DPU_SSPP_RECT_SOLO;
> >>>>> +     r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE;
> >>>>> +     r_pipe->sspp = NULL;
> >>>>>
> >>>>>         pstate->stage = DPU_STAGE_0 + pstate->base.normalized_zpos;
> >>>>>         if (pstate->stage >= pdpu->catalog->caps->max_mixer_blendstages) {
> >>>>> @@ -1016,16 +1033,53 @@ static int dpu_plane_atomic_check(struct drm_plane *plane,
> >>>>>
> >>>>>         max_linewidth = pdpu->catalog->caps->max_linewidth;
> >>>>>
> >>>>> -     /* check decimated source width */
> >>>>>         if (drm_rect_width(&pipe_cfg->src_rect) > max_linewidth) {
> >>>>> -             DPU_DEBUG_PLANE(pdpu, "invalid src " DRM_RECT_FMT " line:%u\n",
> >>>>> -                             DRM_RECT_ARG(&pipe_cfg->src_rect), max_linewidth);
> >>>>> -             return -E2BIG;
> >>>>> +             /* struct dpu_crtc_state *cstate = to_dpu_crtc_state(crtc_state); */
> >>>>> +
> >>>>> +             if (drm_rect_width(&pipe_cfg->src_rect) > 2 * max_linewidth) {
> >>>>> +                     DPU_DEBUG_PLANE(pdpu, "invalid src " DRM_RECT_FMT " line:%u\n",
> >>>>> +                                     DRM_RECT_ARG(&pipe_cfg->src_rect), max_linewidth);
> >>>>> +                     return -E2BIG;
> >>>>> +             }
> >>>>
> >>>> This is where I am a bit concerned enabling it for all chipsets in one go.
> >>>
> >>> As I wrote earlier, I'd prefer the opt-out rather than opt-in here. It
> >>> is much easier to handle the reports "I have a device with sm6543,
> >>> where the display worked before 6.4, but started failing afterwards"
> >>> rather than trying to find a person with sm6543 and asking him if he
> >>> can enable this and that on his device. And even a lower chance of a
> >>> person with sm6543 coming up with a patch 'hey, I enabled this for my
> >>> phone and it works!'.
> >>>
> >>> If we find any issues during or close to the end of the development
> >>> cycle, we can add a 'don't enable wide plane here' switch and enable
> >>> it for failing platforms. But each enablement of this switch should
> >>> come with a reason (wide planes not working here because ....). In the
> >>> end this switch should be gone and transformed into proper HW
> >>> limitation checks.
> >>>
> >>
> >> As it has become clear that with this patch series 4K with UBWC cannot
> >> be supported without true virtual planes (with two SSPPs), why do you
> >> need to relax this check right now?
> >
> > Yes. It enables support for 4k @ linear formats. So my plan for this
> > series is to land 4k with all the proper applicable restrictions.
> >
> >> You can relax this when you add the support for virtual planes till then
> >> let it be this way.
> >>
> >> Its not going to break smartDMA as such. You can still use it for layers
> >> < 2560.
> >>
> >> That way we stay true to the purpose of the feature. I think originally
> >> you wanted to get this in for smartDMA and not to support wide plane and
> >> that purpose will still be achieved even with keeping this check intact.
> >
> > Actually, no. With this series I wanted to get 4k. It was developed in
> > parallel with the 4k enablement for RB3 (posted, bridge patches are
> > being merged for 6.3) and RB5 (delayed for now, I have other issues
> > there).
> >
>
> With the UBWC related checks, this wont support 4K for UBWC layers which
> is default on QC chipsets. So I am fine with respect to that. But still
> this does not address the product spec advertized modes. Like I
> mentioned before, relaxing the maxlinewidth check with the added UBWC
> checks is fine from DPU point of view but not from the product POV.
>
> As things stand today, this is the only check failing the 4K modes on
> chipsets which shouldnt support 4k (linear or UBWC doesnt matter).
>
> >> You can relax it in the virtual plane series.
> >>
> >> Regarding issues, this is where it gets tricky. We should be aligning
> >> with what the product supports. QC will not support issues arising with
> >> 4K on chipsets on which 4K is not advertized.
> >
> > So, we have several different items here:
> > - SmartDMA v2 per se, supporting two rectangles per VIG or DMA plane,
> > - Source split support,
> > - Supporting 4k modes.
> >
> > I think we should tend them one by one. This series concerns SmartDMA
> > v2. Using SmartDMA it is possible to use two rectangles side by side
> > to emulate a wide plane. This series doesn't care at all about max
> > resolutions. These two items are completely orthogonal.
> >
>
> No its not orthogonal. Relaxing the maxlinewidth check in the
> atomic_check() will allow 4K layers now even on chipsets where 4K wasnt
> advertized. Linear or UBWC doesnt matter as the spec doesnt go into that.

Please correct my answers, if I got something wrong here:

Does sc7180 support SmartDMA? Yes it does.
Can QC or CrOS validate SmartDMA separately on sc7180? I hope you can.
Should the hw-supported feature be enabled? Yes, it should.

Now limiting out 4k by not supporting SmartDMA looks like a misfeature.
I can only suggest sending a change to block 4k on sc7180.

> >>>> As you are aware,  we have an open bug today that we do not filter out
> >>>> the modes which we do not support.
> >>>>
> >>>> https://gitlab.freedesktop.org/drm/msm/-/issues/21
> >>>
> >>> I thought that with the link-frequencies in place and with the DSI
> >>> checking the OPP tables this issue is mostly handled. Isn't it?
> >>> Is a mode check in the DPU driver itself the last missing piece?
> >>>
> >>
> >> opp based checking was implemented only for DSI. That one is byte clk based.
> >>
> >> DP uses link rate for opp table.
> >>
> >> Even with a 5.4G link rate (the one in sc7180 chromebook) 4k@30 would
> >> still be possible but it was not advertized
> >>
> >> https://www.qualcomm.com/content/dam/qcomm-martech/dm-assets/documents/prod_brief_qcom_sd7c.pdf
> >>
> >> These docs are available in public domain.
> >>
> >> As we synced up last time on
> >> https://patchwork.freedesktop.org/series/107917/, even with these limits
> >> in place, its not matching the advertized limits.
> >>
> >>>>
> >>>> Due to this, on all chipsets we will end up trying to do a 4K on
> >>>> external display which we dont know what bugs it will expose.
> >>>
> >>> If we do not expose bugs, we do not have a way to fix them. And I
> >>> definitely think that all the bugs should be listed as early as
> >>> possible, while both of us still remember the code under the question.
> >>>
> >>
> >> Yes but on chipsets where 4K is supported ( and hence needed ).
> >
> > 4k, SmartDMA, src-split, split-display, etc.
> >
>
> The visual issues reported on sdm845 on the other thread are a classic
> example of what I just wrote on that patchset and thats why I was
> emphasizing a visual validation OR in other words enable the feature on
> which you are able to visually validate it.

Yes. And if we did not enable the feature, Amit would not be able to
spot that. I can repeat my suggestion:

- prevalidate these features for the accessible platforms (e.g. I only
have sdm845, sm8250 and sm8350 at hand)
- enable SmartDMA for all chipsets where SmartDMA is supported
- collect possible tested-by and broken-at reports, while the patches
sit in linux-next
- disable SmartDMA basing on the feedback from the previous step (e.g.
select from 'mostly disable', 'disable for the bugged cases', 'do not
disable at all', etc).
  I can promise that if we see a significant validation failure rate I
will not oppose disabling SmartDMA.

As a reminder: if the patchset is ready at the time of 6.3-rc1 (in
three weeks from now), it is going to be merged into linux-next first,
after that it can go into the main Linus'es tree at 6.4-rc1. So we
will have _two_ kernel cycles to collect bug reports and to disable
(or fix) broken cases.

> We can evaluate and enable smartDMA on other chipsets on a need basis.
>
> We discussed this again even today in the team discussion. Our team's
> PoV doesnt change. We would still like to enable smartDMA only on
> chipsets which can be visually validated first to limit the debugging
> effort to one chipset first and then perfect it. Otherwise its too much
> effort on QC side to debug those issues on all chipsets.

I think QC mostly debugs issues on sc7180/sc7280 and sometimes on
sdm845/sm8250 (and now on sm8350). I think we can let people
(somainline, PmOS) test features on other platforms.
And testing happens better if we can say 'please test linux-next'
rather than 'please test linux-next + this patch to enable the feature
+ that patch to enable the second patch of the feature'.

> >>>> So lets say if we test it on sc7280 fully but not on sc7180, we will
> >>>> still hit this condition on sc7180 too but on that chipset we did not
> >>>> advertise 4K as a capability in the product spec.
> >>>
> >>> Is it 'not advertised' or 'not supported by hw'?
> >>>
> >>
> >> The document
> >> https://www.qualcomm.com/content/dam/qcomm-martech/dm-assets/documents/prod_brief_qcom_sd7c.pdf
> >> is made from inputs from not just display team but overall system
> >> limits. So even though you could argue that this falls within the
> >> display capabilities, all I can say at the moment is we have to stick to
> >> the advertized limits as its compiled with inputs from all the teams
> >> (system/performance etc).
> >
> > So, there should be a limiting factor (or a combination of them).
> > Filter out 4k modes on sc7180. Or modes using fill rate higher than N.
> > Pixel clock rate higher than M. But it has nothing to do with these
> > patches enabling SmartDMA support on this platform.
> >
> > Even if we look at the vendor kernels, we don't see 'maximum external
> > resolution'. Instead I see a combination of linewidth and bandwidth
> > limitations. If we can stick to that, that would be great.
> >
>
> Can you please point me to bandwidth limitation checks? How are other
> vendors coming up with this number? It has to be based on some
> resolution too right?

Usually it is considered in the other direction. The SoC can support
this-and-that pixel clock and bandwidth, so put the NxM resolution to
the datasheet as the max supported one.

Can you point out how the vendor kernel limits DP modes? I checked out
several dtsi files. For sm8250 the DP is limited to 1920x1080 (while
PB explicitly mentions 4k@60).
sm7125 is limited to 2560x1600. sm6150 again 1920x1080. From the pile
of the DTS that I have here the rest lists only
qcom,max-pclk-frequency-khz

> My RFC https://patchwork.freedesktop.org/series/107917/ considered pixel
> clk as the limiting factor which was posted after discussions
> internally. In the absence of another way, that remains the only
> solution to tackle this.

If I remember correctly the mentioned patchset used manually crafted
pixel clocks. And for example for sm8250 this clock doesn't correspond
to verifiable source.
Your patchset used 594000 KHz as max ext pclk for sm8250. vendor dtsi
lists 187500 KHz as maximum DP pclk for RB5 and 150000 KHz in all
other cases. And, at the same time, it lists a 3840x2160@60 mode for
the DSI/lt9611uxc with the pixel clock as high as 608040 KHz.

I suggest we stop the discussion at this point, unless there is
anything else wrong with this patchset itself.

I noted the point regarding UBWC & parallel mode. I will handle it in
the next iteration.

I noted your valid point about the visual verification. I proposed a
way to ease validation, allowing it to be enabled for testers and
early adopters for nearly two cycles. I hope you'd agree to this plan.
I on the other side agree to revert to opt-in if the failure rate is
high.

Please stop bringing max resolution issues to this patchserie. It must
be handled separately. I hope to see the mode filter patch targeting
sc7180 & sc7280. With the DSI opp check in place I think we should
concentrate on the DP case. If nothing else, I think even adding the
max PCLK to the msm_dp_desc should be sufficient to your worries. I'd
prefer to be able to override it for the particular board, but I think
this can come later, as it would require an agreement from the DT
schema team.

--
With best wishes
Dmitry

  reply	other threads:[~2023-02-10  0:09 UTC|newest]

Thread overview: 176+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-03 18:21 [PATCH v3 00/27] drm/msm/dpu: wide planes support Dmitry Baryshkov
2023-02-03 18:21 ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 01/27] drm/msm/dpu: rename struct dpu_hw_pipe(_cfg) to dpu_hw_sspp(_cfg) Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 19:31   ` Abhinav Kumar
2023-02-03 19:31     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 02/27] drm/msm/dpu: move SSPP allocation to the RM Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 03/27] drm/msm/dpu: move SSPP debugfs creation to dpu_kms.c Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 04/27] drm/msm/dpu: drop EAGAIN check from dpu_format_populate_layout Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 05/27] drm/msm/dpu: move pipe_hw to dpu_plane_state Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 19:34   ` Abhinav Kumar
2023-02-03 19:34     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 06/27] drm/msm/dpu: drop dpu_plane_pipe function Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 07/27] drm/msm/dpu: introduce struct dpu_sw_pipe Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 08/27] drm/msm/dpu: use dpu_sw_pipe for dpu_hw_sspp callbacks Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 09/27] drm/msm/dpu: pass dpu_format to _dpu_hw_sspp_setup_scaler3() Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 19:36   ` Abhinav Kumar
2023-02-03 19:36     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 10/27] drm/msm/dpu: clean up SRC addresses when setting up SSPP for solid fill Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 20:36   ` Abhinav Kumar
2023-02-03 20:36     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 11/27] drm/msm/dpu: move stride programming to dpu_hw_sspp_setup_sourceaddress Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 20:37   ` Abhinav Kumar
2023-02-03 20:37     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 12/27] drm/msm/dpu: remove dpu_hw_fmt_layout from struct dpu_hw_sspp_cfg Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 13/27] drm/msm/dpu: drop src_split and multirect check from dpu_crtc_atomic_check Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 14/27] drm/msm/dpu: don't use unsupported blend stages Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 20:42   ` Abhinav Kumar
2023-02-03 20:42     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 15/27] drm/msm/dpu: move the rest of plane checks to dpu_plane_atomic_check() Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 22:44   ` Abhinav Kumar
2023-02-03 22:44     ` Abhinav Kumar
2023-02-03 22:55     ` Dmitry Baryshkov
2023-02-03 22:55       ` Dmitry Baryshkov
2023-02-14 23:25       ` Abhinav Kumar
2023-02-14 23:25         ` Abhinav Kumar
2023-02-15  0:08         ` Dmitry Baryshkov
2023-02-15  0:08           ` Dmitry Baryshkov
2023-03-03 11:43           ` Dmitry Baryshkov
2023-03-03 11:43             ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 16/27] drm/msm/dpu: drop redundant plane dst check from dpu_crtc_atomic_check() Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 22:57   ` Abhinav Kumar
2023-02-03 22:57     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 17/27] drm/msm/dpu: rewrite plane's QoS-related functions to take dpu_sw_pipe and dpu_format Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 23:07   ` Abhinav Kumar
2023-02-03 23:07     ` Abhinav Kumar
2023-02-03 23:20     ` Dmitry Baryshkov
2023-02-03 23:20       ` Dmitry Baryshkov
2023-02-08 23:09       ` Abhinav Kumar
2023-02-08 23:09         ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 18/27] drm/msm/dpu: populate SmartDMA features in hw catalog Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-03 23:35   ` Abhinav Kumar
2023-02-03 23:35     ` Abhinav Kumar
2023-02-04  2:29     ` Dmitry Baryshkov
2023-02-04  2:29       ` Dmitry Baryshkov
2023-02-04  2:43       ` Abhinav Kumar
2023-02-04  2:43         ` Abhinav Kumar
2023-02-04  4:10         ` Dmitry Baryshkov
2023-02-04  4:10           ` Dmitry Baryshkov
2023-02-04  5:10           ` Abhinav Kumar
2023-02-04  5:10             ` Abhinav Kumar
2023-02-04 10:43             ` Dmitry Baryshkov
2023-02-04 10:43               ` Dmitry Baryshkov
2023-02-04 18:35               ` Abhinav Kumar
2023-02-04 18:35                 ` Abhinav Kumar
2023-02-04 21:08                 ` Dmitry Baryshkov
2023-02-04 21:08                   ` Dmitry Baryshkov
2023-02-04 23:20                   ` Abhinav Kumar
2023-02-04 23:20                     ` Abhinav Kumar
2023-02-05  0:29                     ` Dmitry Baryshkov
2023-02-05  0:29                       ` Dmitry Baryshkov
2023-02-05  0:36                       ` Abhinav Kumar
2023-02-05  0:36                         ` Abhinav Kumar
2023-02-08 23:53                         ` Dmitry Baryshkov
2023-02-08 23:53                           ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 19/27] drm/msm/dpu: make _dpu_plane_calc_clk accept mode directly Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-06 18:48   ` Abhinav Kumar
2023-02-06 18:48     ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 20/27] drm/msm/dpu: add dpu_hw_pipe_cfg to dpu_plane_state Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-06 19:07   ` Abhinav Kumar
2023-02-06 19:07     ` Abhinav Kumar
2023-02-16 16:49     ` Dmitry Baryshkov
2023-02-16 16:49       ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 21/27] drm/msm/dpu: simplify dpu_plane_validate_src() Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-06 22:40   ` Abhinav Kumar
2023-02-06 22:40     ` Abhinav Kumar
2023-02-07  0:27     ` Dmitry Baryshkov
2023-02-07  0:27       ` Dmitry Baryshkov
2023-02-07  0:42       ` Abhinav Kumar
2023-02-07  0:42         ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 22/27] drm/msm/dpu: rework dpu_plane_sspp_atomic_update() Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-07  0:22   ` Abhinav Kumar
2023-02-07  0:22     ` Abhinav Kumar
2023-02-09  0:49     ` Dmitry Baryshkov
2023-02-09  0:49       ` Dmitry Baryshkov
2023-02-09  0:51     ` Dmitry Baryshkov
2023-02-09  0:51       ` Dmitry Baryshkov
2023-02-09 11:46     ` Dmitry Baryshkov
2023-02-09 11:46       ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 23/27] drm/msm/dpu: rework dpu_plane_atomic_check() Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-07 17:49   ` Abhinav Kumar
2023-02-07 17:49     ` Abhinav Kumar
2023-02-07 17:51     ` Dmitry Baryshkov
2023-02-07 17:51       ` Dmitry Baryshkov
2023-02-07 17:57       ` Abhinav Kumar
2023-02-07 17:57         ` Abhinav Kumar
2023-02-07 17:59         ` Dmitry Baryshkov
2023-02-07 17:59           ` Dmitry Baryshkov
2023-02-07 18:08           ` Abhinav Kumar
2023-02-07 18:08             ` Abhinav Kumar
2023-02-07 19:51             ` Dmitry Baryshkov
2023-02-07 19:51               ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 24/27] drm/msm/dpu: rework plane CSC setting Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-07 20:05   ` Abhinav Kumar
2023-02-07 20:05     ` Abhinav Kumar
2023-02-07 20:44     ` Dmitry Baryshkov
2023-02-07 20:44       ` Dmitry Baryshkov
2023-02-07 20:57       ` Abhinav Kumar
2023-02-07 20:57         ` Abhinav Kumar
2023-02-03 18:21 ` [PATCH v3 25/27] drm/msm/dpu: rework static color fill code Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-08 22:34   ` Abhinav Kumar
2023-02-08 22:34     ` Abhinav Kumar
2023-02-09  0:53     ` Dmitry Baryshkov
2023-02-09  0:53       ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 26/27] drm/msm/dpu: split pipe handling from _dpu_crtc_blend_setup_mixer Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-08 23:44   ` Abhinav Kumar
2023-02-08 23:44     ` Abhinav Kumar
2023-02-08 23:47     ` Dmitry Baryshkov
2023-02-08 23:47       ` Dmitry Baryshkov
2023-03-03 12:18     ` Dmitry Baryshkov
2023-03-03 12:18       ` Dmitry Baryshkov
2023-02-03 18:21 ` [PATCH v3 27/27] drm/msm/dpu: add support for wide planes Dmitry Baryshkov
2023-02-03 18:21   ` Dmitry Baryshkov
2023-02-09  2:19   ` Abhinav Kumar
2023-02-09  2:19     ` Abhinav Kumar
2023-02-09 11:45     ` Dmitry Baryshkov
2023-02-09 11:45       ` Dmitry Baryshkov
2023-02-09 19:25       ` Abhinav Kumar
2023-02-09 19:25         ` Abhinav Kumar
2023-02-09 21:23         ` Dmitry Baryshkov
2023-02-09 21:23           ` Dmitry Baryshkov
2023-02-09 22:12           ` [Freedreno] " Abhinav Kumar
2023-02-09 22:12             ` Abhinav Kumar
2023-02-10  0:09             ` Dmitry Baryshkov [this message]
2023-02-10  0:09               ` Dmitry Baryshkov
2023-02-10  1:12               ` Abhinav Kumar
2023-02-10  1:12                 ` Abhinav Kumar
2023-02-10  2:46                 ` Dmitry Baryshkov
2023-02-10  2:46                   ` Dmitry Baryshkov
2023-02-03 18:24 ` [PATCH v3 00/27] drm/msm/dpu: wide planes support Dmitry Baryshkov
2023-02-03 18:24   ` Dmitry Baryshkov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAA8EJppuM97q2uc04_id7fEY26WHbnb=Mc0eAU1sDMTNwguAYg@mail.gmail.com' \
    --to=dmitry.baryshkov@linaro.org \
    --cc=andersson@kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=freedreno@lists.freedesktop.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=quic_abhinavk@quicinc.com \
    --cc=quic_kalyant@quicinc.com \
    --cc=sean@poorly.run \
    --cc=swboyd@chromium.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.