[PATCH 14/21] drm/msm/dpu: Reserve resources for CWB
Dmitry Baryshkov
dmitry.baryshkov at linaro.org
Fri Aug 30 22:27:12 UTC 2024
On Fri, 30 Aug 2024 at 23:28, Jessica Zhang <quic_jesszhan at quicinc.com> wrote:
> On 8/30/2024 10:25 AM, Dmitry Baryshkov wrote:
> > On Thu, Aug 29, 2024 at 01:48:35PM GMT, Jessica Zhang wrote:
> >> Reserve dedicated pingpong blocks for CWB
> >
> > Please explain design ideas. Having just a single phrase is usually not
> > enough.
> >
> >>
> >> Signed-off-by: Jessica Zhang <quic_jesszhan at quicinc.com>
> >> ---
> >> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 17 ++++++--
> >> drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h | 1 +
> >> drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c | 65 +++++++++++++++++++++++++++++
> >> 3 files changed, 79 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> >> index 36b677cf9c7a..f1bd14d1f89e 100644
> >> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> >> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> >> @@ -2,7 +2,7 @@
> >> /*
> >> * Copyright (C) 2013 Red Hat
> >> * Copyright (c) 2014-2018, 2020-2021 The Linux Foundation. All rights reserved.
> >> - * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
> >> + * Copyright (c) 2022-2024 Qualcomm Innovation Center, Inc. All rights reserved.
> >> *
> >> * Author: Rob Clark <robdclark at gmail.com>
> >> */
> >> @@ -1054,6 +1054,7 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc,
> >> struct dpu_hw_blk *hw_ctl[MAX_CHANNELS_PER_ENC];
> >> struct dpu_hw_blk *hw_dsc[MAX_CHANNELS_PER_ENC];
> >> int num_pp, num_dsc;
> >> + bool is_cwb_encoder;
> >> unsigned int dsc_mask = 0;
> >> int i;
> >>
> >> @@ -1067,6 +1068,8 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc,
> >>
> >> priv = drm_enc->dev->dev_private;
> >> dpu_kms = to_dpu_kms(priv->kms);
> >> + is_cwb_encoder = drm_crtc_in_clone_mode(crtc_state) &&
> >> + dpu_enc->disp_info.intf_type == INTF_WB;
> >>
> >> global_state = dpu_kms_get_existing_global_state(dpu_kms);
> >> if (IS_ERR_OR_NULL(global_state)) {
> >> @@ -1077,9 +1080,15 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc,
> >> trace_dpu_enc_mode_set(DRMID(drm_enc));
> >>
> >> /* Query resource that have been reserved in atomic check step. */
> >> - num_pp = dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> - drm_enc->crtc, DPU_HW_BLK_PINGPONG, hw_pp,
> >> - ARRAY_SIZE(hw_pp));
> >> + if (is_cwb_encoder)
> >> + num_pp = dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> + drm_enc->crtc, DPU_HW_BLK_DCWB_PINGPONG, hw_pp,
> >> + ARRAY_SIZE(hw_pp));
> >> + else
> >> + num_pp = dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> + drm_enc->crtc, DPU_HW_BLK_PINGPONG, hw_pp,
> >> + ARRAY_SIZE(hw_pp));
> >
> > Why is this necessary? Can we still use DPU_HW_BLK_PINGPONG?
>
> Hi Dmitry,
>
> Since both regular non-cwb pingpongs and pingpongs dedicated for cwb
> both share the same pingpong_to_crtc_id map, I need a way to distinguish
> between the real-time and dedicated cwb pingpongs.
>
> If I don't do that, get_assigned_resources would populate the hw_pp
> array for both the real-time and WB encoders with *only* the real-time
> display pingpongs.
I see. It looks like PINGPONG blocks should be allocated / filled
per-encoder, not per-CRTC. Likewise CDM is encoder-specific, not
CRTC-specific.
Does that solve the issue?
>
> >
> >> +
> >> dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> drm_enc->crtc, DPU_HW_BLK_CTL, hw_ctl, ARRAY_SIZE(hw_ctl));
> >>
> >> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h
> >> index c43cb55fe1d2..c87790a1b940 100644
> >> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h
> >> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h
> >> @@ -77,6 +77,7 @@ enum dpu_hw_blk_type {
> >> DPU_HW_BLK_LM,
> >> DPU_HW_BLK_CTL,
> >> DPU_HW_BLK_PINGPONG,
> >> + DPU_HW_BLK_DCWB_PINGPONG,
> >> DPU_HW_BLK_INTF,
> >> DPU_HW_BLK_WB,
> >> DPU_HW_BLK_DSPP,
> >> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> >> index 13f84375e15d..afad26556cd5 100644
> >> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> >> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> >> @@ -236,6 +236,48 @@ static int _dpu_rm_get_lm_peer(struct dpu_rm *rm, int primary_idx)
> >> return -EINVAL;
> >> }
> >>
> >> +static int _dpu_rm_reserve_cwb_pingpong(struct dpu_rm *rm,
> >> + struct dpu_global_state *global_state, uint32_t crtc_id,
> >> + struct msm_display_topology *topology)
> >> +{
> >> + int num_cwb_pp = topology->num_lm, cwb_pp_count = 0;
> >> + int cwb_pp_idx[MAX_BLOCKS];
> >> +
> >> + /*
> >> + * Reserve additional dedicated CWB pingpong blocks for each mixer
> >> + *
> >> + * TODO: add support for reserving non-dedicated CWB pingpong blocks
> >> + */
> >> + for (int i = 0; i < ARRAY_SIZE(rm->mixer_blks) &&
> >> + cwb_pp_count < num_cwb_pp; i++) {
> >> + for (int j = PINGPONG_CWB_0 - PINGPONG_0;
> >> + j < ARRAY_SIZE(rm->pingpong_blks); j++) {
> >
> > Alignment...
>
> Ack -- for future reference, should I always align the second line with
> the bracket of the previous line?
>
> I thought the general guideline was 2 tabs, but I might be misremembering.
I can't find the textual reference, however .clang-format in the root
dir has 'AlignAfterOpenBracket: Align'. I'm more or less ignorant for
the function argument alignment (especially if an old function is
being modified), but aligning the conditions and loops makes it easier
to read the code.
>
> Thanks,
>
> Jessica Zhang
>
> >
> >> + /*
> >> + * Odd LMs must be assigned to odd pingpongs and even
> >> + * LMs with even pingpongs
> >> + */
> >> + if (reserved_by_other(global_state->pingpong_to_crtc_id,
> >> + j, crtc_id) ||
> >> + i % 2 != j % 2)
> >> + continue;
> >> +
> >> + cwb_pp_idx[cwb_pp_count] = j;
> >> + cwb_pp_count++;
> >> + break;
> >> + }
> >> + }
> >> +
> >> + if (cwb_pp_count != num_cwb_pp) {
> >> + DPU_ERROR("Unable to reserve all cwb pingpongs\n");
> >> + return -ENAVAIL;
> >> + }
> >> +
> >> + for (int i = 0; i < cwb_pp_count; i++)
> >> + global_state->pingpong_to_crtc_id[cwb_pp_idx[i]] = crtc_id;
> >> +
> >> + return 0;
> >> +}
> >> +
> >> /**
> >> * _dpu_rm_check_lm_and_get_connected_blks - check if proposed layer mixer meets
> >> * proposed use case requirements, incl. hardwired dependent blocks like
> >> @@ -617,6 +659,14 @@ static int _dpu_rm_make_reservation(
> >> return ret;
> >> }
> >>
> >> + if (topology->cwb_enabled) {
> >> + ret = _dpu_rm_reserve_cwb_pingpong(rm, global_state,
> >> + crtc_id, topology);
> >> + if (ret) {
> >> + DPU_ERROR("unable to find appropriate dcwb pingpongs\n");
> >> + return ret;
> >> + }
> >> + }
> >>
> >> ret = _dpu_rm_reserve_ctls(rm, global_state, crtc_id,
> >> topology);
> >> @@ -706,6 +756,7 @@ int dpu_rm_get_assigned_resources(struct dpu_rm *rm,
> >>
> >> switch (type) {
> >> case DPU_HW_BLK_PINGPONG:
> >> + case DPU_HW_BLK_DCWB_PINGPONG:
> >> hw_blks = rm->pingpong_blks;
> >> hw_to_crtc_id = global_state->pingpong_to_crtc_id;
> >> max_blks = ARRAY_SIZE(rm->pingpong_blks);
> >> @@ -745,6 +796,20 @@ int dpu_rm_get_assigned_resources(struct dpu_rm *rm,
> >> if (hw_to_crtc_id[i] != crtc_id)
> >> continue;
> >>
> >> + if (type == DPU_HW_BLK_PINGPONG) {
> >> + struct dpu_hw_pingpong *pp = to_dpu_hw_pingpong(hw_blks[i]);
> >> +
> >> + if (pp->idx >= PINGPONG_CWB_0)
> >> + continue;
> >> + }
> >> +
> >> + if (type == DPU_HW_BLK_DCWB_PINGPONG) {
> >> + struct dpu_hw_pingpong *pp = to_dpu_hw_pingpong(hw_blks[i]);
> >> +
> >> + if (pp->idx < PINGPONG_CWB_0)
> >> + continue;
> >> + }
> >> +
> >> if (num_blks == blks_size) {
> >> DPU_ERROR("More than %d resources assigned to crtc %d\n",
> >> blks_size, crtc_id);
> >>
> >> --
> >> 2.34.1
> >>
> >
> > --
> > With best wishes
> > Dmitry
--
With best wishes
Dmitry
On Fri, 30 Aug 2024 at 23:28, Jessica Zhang <quic_jesszhan at quicinc.com> wrote:
>
>
>
> On 8/30/2024 10:25 AM, Dmitry Baryshkov wrote:
> > On Thu, Aug 29, 2024 at 01:48:35PM GMT, Jessica Zhang wrote:
> >> Reserve dedicated pingpong blocks for CWB
> >
> > Please explain design ideas. Having just a single phrase is usually not
> > enough.
> >
> >>
> >> Signed-off-by: Jessica Zhang <quic_jesszhan at quicinc.com>
> >> ---
> >> drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 17 ++++++--
> >> drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h | 1 +
> >> drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c | 65 +++++++++++++++++++++++++++++
> >> 3 files changed, 79 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> >> index 36b677cf9c7a..f1bd14d1f89e 100644
> >> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> >> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> >> @@ -2,7 +2,7 @@
> >> /*
> >> * Copyright (C) 2013 Red Hat
> >> * Copyright (c) 2014-2018, 2020-2021 The Linux Foundation. All rights reserved.
> >> - * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
> >> + * Copyright (c) 2022-2024 Qualcomm Innovation Center, Inc. All rights reserved.
> >> *
> >> * Author: Rob Clark <robdclark at gmail.com>
> >> */
> >> @@ -1054,6 +1054,7 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc,
> >> struct dpu_hw_blk *hw_ctl[MAX_CHANNELS_PER_ENC];
> >> struct dpu_hw_blk *hw_dsc[MAX_CHANNELS_PER_ENC];
> >> int num_pp, num_dsc;
> >> + bool is_cwb_encoder;
> >> unsigned int dsc_mask = 0;
> >> int i;
> >>
> >> @@ -1067,6 +1068,8 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc,
> >>
> >> priv = drm_enc->dev->dev_private;
> >> dpu_kms = to_dpu_kms(priv->kms);
> >> + is_cwb_encoder = drm_crtc_in_clone_mode(crtc_state) &&
> >> + dpu_enc->disp_info.intf_type == INTF_WB;
> >>
> >> global_state = dpu_kms_get_existing_global_state(dpu_kms);
> >> if (IS_ERR_OR_NULL(global_state)) {
> >> @@ -1077,9 +1080,15 @@ static void dpu_encoder_virt_atomic_mode_set(struct drm_encoder *drm_enc,
> >> trace_dpu_enc_mode_set(DRMID(drm_enc));
> >>
> >> /* Query resource that have been reserved in atomic check step. */
> >> - num_pp = dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> - drm_enc->crtc, DPU_HW_BLK_PINGPONG, hw_pp,
> >> - ARRAY_SIZE(hw_pp));
> >> + if (is_cwb_encoder)
> >> + num_pp = dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> + drm_enc->crtc, DPU_HW_BLK_DCWB_PINGPONG, hw_pp,
> >> + ARRAY_SIZE(hw_pp));
> >> + else
> >> + num_pp = dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> + drm_enc->crtc, DPU_HW_BLK_PINGPONG, hw_pp,
> >> + ARRAY_SIZE(hw_pp));
> >
> > Why is this necessary? Can we still use DPU_HW_BLK_PINGPONG?
>
> Hi Dmitry,
>
> Since both regular non-cwb pingpongs and pingpongs dedicated for cwb
> both share the same pingpong_to_crtc_id map, I need a way to distinguish
> between the real-time and dedicated cwb pingpongs.
>
> If I don't do that, get_assigned_resources would populate the hw_pp
> array for both the real-time and WB encoders with *only* the real-time
> display pingpongs.
>
> >
> >> +
> >> dpu_rm_get_assigned_resources(&dpu_kms->rm, global_state,
> >> drm_enc->crtc, DPU_HW_BLK_CTL, hw_ctl, ARRAY_SIZE(hw_ctl));
> >>
> >> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h
> >> index c43cb55fe1d2..c87790a1b940 100644
> >> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h
> >> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_mdss.h
> >> @@ -77,6 +77,7 @@ enum dpu_hw_blk_type {
> >> DPU_HW_BLK_LM,
> >> DPU_HW_BLK_CTL,
> >> DPU_HW_BLK_PINGPONG,
> >> + DPU_HW_BLK_DCWB_PINGPONG,
> >> DPU_HW_BLK_INTF,
> >> DPU_HW_BLK_WB,
> >> DPU_HW_BLK_DSPP,
> >> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> >> index 13f84375e15d..afad26556cd5 100644
> >> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> >> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c
> >> @@ -236,6 +236,48 @@ static int _dpu_rm_get_lm_peer(struct dpu_rm *rm, int primary_idx)
> >> return -EINVAL;
> >> }
> >>
> >> +static int _dpu_rm_reserve_cwb_pingpong(struct dpu_rm *rm,
> >> + struct dpu_global_state *global_state, uint32_t crtc_id,
> >> + struct msm_display_topology *topology)
> >> +{
> >> + int num_cwb_pp = topology->num_lm, cwb_pp_count = 0;
> >> + int cwb_pp_idx[MAX_BLOCKS];
> >> +
> >> + /*
> >> + * Reserve additional dedicated CWB pingpong blocks for each mixer
> >> + *
> >> + * TODO: add support for reserving non-dedicated CWB pingpong blocks
> >> + */
> >> + for (int i = 0; i < ARRAY_SIZE(rm->mixer_blks) &&
> >> + cwb_pp_count < num_cwb_pp; i++) {
> >> + for (int j = PINGPONG_CWB_0 - PINGPONG_0;
> >> + j < ARRAY_SIZE(rm->pingpong_blks); j++) {
> >
> > Alignment...
>
> Ack -- for future reference, should I always align the second line with
> the bracket of the previous line?
>
> I thought the general guideline was 2 tabs, but I might be misremembering.
>
> Thanks,
>
> Jessica Zhang
>
> >
> >> + /*
> >> + * Odd LMs must be assigned to odd pingpongs and even
> >> + * LMs with even pingpongs
> >> + */
> >> + if (reserved_by_other(global_state->pingpong_to_crtc_id,
> >> + j, crtc_id) ||
> >> + i % 2 != j % 2)
> >> + continue;
> >> +
> >> + cwb_pp_idx[cwb_pp_count] = j;
> >> + cwb_pp_count++;
> >> + break;
> >> + }
> >> + }
> >> +
> >> + if (cwb_pp_count != num_cwb_pp) {
> >> + DPU_ERROR("Unable to reserve all cwb pingpongs\n");
> >> + return -ENAVAIL;
> >> + }
> >> +
> >> + for (int i = 0; i < cwb_pp_count; i++)
> >> + global_state->pingpong_to_crtc_id[cwb_pp_idx[i]] = crtc_id;
> >> +
> >> + return 0;
> >> +}
> >> +
> >> /**
> >> * _dpu_rm_check_lm_and_get_connected_blks - check if proposed layer mixer meets
> >> * proposed use case requirements, incl. hardwired dependent blocks like
> >> @@ -617,6 +659,14 @@ static int _dpu_rm_make_reservation(
> >> return ret;
> >> }
> >>
> >> + if (topology->cwb_enabled) {
> >> + ret = _dpu_rm_reserve_cwb_pingpong(rm, global_state,
> >> + crtc_id, topology);
> >> + if (ret) {
> >> + DPU_ERROR("unable to find appropriate dcwb pingpongs\n");
> >> + return ret;
> >> + }
> >> + }
> >>
> >> ret = _dpu_rm_reserve_ctls(rm, global_state, crtc_id,
> >> topology);
> >> @@ -706,6 +756,7 @@ int dpu_rm_get_assigned_resources(struct dpu_rm *rm,
> >>
> >> switch (type) {
> >> case DPU_HW_BLK_PINGPONG:
> >> + case DPU_HW_BLK_DCWB_PINGPONG:
> >> hw_blks = rm->pingpong_blks;
> >> hw_to_crtc_id = global_state->pingpong_to_crtc_id;
> >> max_blks = ARRAY_SIZE(rm->pingpong_blks);
> >> @@ -745,6 +796,20 @@ int dpu_rm_get_assigned_resources(struct dpu_rm *rm,
> >> if (hw_to_crtc_id[i] != crtc_id)
> >> continue;
> >>
> >> + if (type == DPU_HW_BLK_PINGPONG) {
> >> + struct dpu_hw_pingpong *pp = to_dpu_hw_pingpong(hw_blks[i]);
> >> +
> >> + if (pp->idx >= PINGPONG_CWB_0)
> >> + continue;
> >> + }
> >> +
> >> + if (type == DPU_HW_BLK_DCWB_PINGPONG) {
> >> + struct dpu_hw_pingpong *pp = to_dpu_hw_pingpong(hw_blks[i]);
> >> +
> >> + if (pp->idx < PINGPONG_CWB_0)
> >> + continue;
> >> + }
> >> +
> >> if (num_blks == blks_size) {
> >> DPU_ERROR("More than %d resources assigned to crtc %d\n",
> >> blks_size, crtc_id);
> >>
> >> --
> >> 2.34.1
> >>
> >
> > --
> > With best wishes
> > Dmitry
--
With best wishes
Dmitry
More information about the Freedreno
mailing list