RE: [v1 2/2] drm/msm/disp/dpu1: enable crtc color management based on encoder topology

From: Kalyan Thota
Date: Mon Jun 27 2022 - 08:03:23 EST


Thanks for the comments, Dmitry. I haven't noticed mode->hdisplay being used. My idea was to run thru the topology and tie up the encoders with dspp to the CRTCs.
Since mode is available only in the commit, we cannot use the dpu_encoder_get_topology during initialization sequence.

The requirement here is that when we initialize the crtc, we need to enable drm_crtc_enable_color_mgmt only for the crtcs that support it. As I understand from Rob, chrome framework will check for the enablement in order to exercise the feature.

Do you have any ideas on how to handle this requirement ? Since we will reserve the dspp's only when a commit is issued, I guess it will be too late to enable color management by then.

@robdclark@xxxxxxxxx
Is it okay, if we disable color management for all the crtcs during initialization and enable it when we have dspps available during modeset. Can we framework code query for the property before issuing a commit for the frame after modeset ?

Thanks,
Kalyan

> -----Original Message-----
> From: Dmitry Baryshkov <dmitry.baryshkov@xxxxxxxxxx>
> Sent: Tuesday, June 21, 2022 4:43 PM
> To: Kalyan Thota (QUIC) <quic_kalyant@xxxxxxxxxxx>
> Cc: y@xxxxxxxxxxxx; dri-devel@xxxxxxxxxxxxxxxxxxxxx; linux-arm-
> msm@xxxxxxxxxxxxxxx; freedreno@xxxxxxxxxxxxxxxxxxxxx;
> devicetree@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> robdclark@xxxxxxxxx; dianders@xxxxxxxxxxxx; swboyd@xxxxxxxxxxxx;
> Vinod Polimera (QUIC) <quic_vpolimer@xxxxxxxxxxx>; Abhinav Kumar (QUIC)
> <quic_abhinavk@xxxxxxxxxxx>
> Subject: Re: [v1 2/2] drm/msm/disp/dpu1: enable crtc color management based
> on encoder topology
>
> WARNING: This email originated from outside of Qualcomm. Please be wary of
> any links or attachments, and do not enable macros.
>
> Generic comment: y@xxxxxxxxxxxx address bounces. Please remove it from
> the cc list. If you need to send a patch for the internal reasons, please use Bcc.
>
> On Tue, 21 Jun 2022 at 12:06, Kalyan Thota <quic_kalyant@xxxxxxxxxxx> wrote:
> >
> > Crtc color management needs to be registered only for the crtc which
> > has the capability to handle it. Since topology decides which encoder
> > will get the dspp hw block, tie up the crtc and the encoder together
> > (encoder->possible_crtcs)
> >
> > Change-Id: If5a0f33547b6f527ca4b8fbb78424b141dbbd711
>
> No change-id's please. This is not the gerrit.
>
> > Signed-off-by: Kalyan Thota <quic_kalyant@xxxxxxxxxxx>
> > ---
> > drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c | 8 ++++++--
> > drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h | 2 +-
> > drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 20 ++++++++++++++++----
> > drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h | 5 +++++
> > drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 22 ++++++++++++++++++----
> > 5 files changed, 46 insertions(+), 11 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> > b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> > index 7763558..2913acb 100644
> > --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> > +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.c
> > @@ -1511,7 +1511,7 @@ static const struct drm_crtc_helper_funcs
> > dpu_crtc_helper_funcs = {
> >
> > /* initialize crtc */
> > struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane
> *plane,
> > - struct drm_plane *cursor)
> > + struct drm_plane *cursor, unsigned int
> > + enc_mask)
> > {
> > struct drm_crtc *crtc = NULL;
> > struct dpu_crtc *dpu_crtc = NULL; @@ -1544,7 +1544,11 @@
> > struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct
> > drm_plane *plane,
> >
> > drm_crtc_helper_add(crtc, &dpu_crtc_helper_funcs);
> >
> > - drm_crtc_enable_color_mgmt(crtc, 0, true, 0);
> > + /* Register crtc color management if the encoder has dspp, use the
> > + * crtc to mark it as possible_crtcs for that encoder.
> > + */
> > + if(BIT(crtc->index) & enc_mask)
>
> So, we are checking CRTC's index against the encoder's mask? This is
> counterintuitive.
>
> > + drm_crtc_enable_color_mgmt(crtc, 0, true, 0);
> >
> > /* save user friendly CRTC name for later */
> > snprintf(dpu_crtc->name, DPU_CRTC_NAME_SIZE, "crtc%u",
> > crtc->base.id); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h
> > b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h
> > index b8785c3..0a6458e 100644
> > --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h
> > +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_crtc.h
> > @@ -269,7 +269,7 @@ void dpu_crtc_complete_commit(struct drm_crtc
> *crtc);
> > * @Return: new crtc object or error
> > */
> > struct drm_crtc *dpu_crtc_init(struct drm_device *dev, struct drm_plane
> *plane,
> > - struct drm_plane *cursor);
> > + struct drm_plane *cursor, unsigned int
> > + enc_mask);
> >
> > /**
> > * dpu_crtc_register_custom_event - api for enabling/disabling crtc
> > event diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> > b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> > index f2cb497..893ce68 100644
> > --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> > +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c
> > @@ -13,6 +13,8 @@
> > #include <drm/drm_crtc.h>
> > #include <drm/drm_file.h>
> > #include <drm/drm_probe_helper.h>
> > +#include <drm/drm_bridge.h>
> > +#include <drm/drm_bridge_connector.h>
> >
> > #include "msm_drv.h"
> > #include "dpu_kms.h"
> > @@ -511,13 +513,18 @@ void dpu_encoder_helper_split_config(
> > }
> > }
> >
> > -static struct msm_display_topology dpu_encoder_get_topology(
> > - struct dpu_encoder_virt *dpu_enc,
> > +struct msm_display_topology dpu_encoder_get_topology(
> > + struct drm_encoder *drm_enc,
> > struct dpu_kms *dpu_kms,
> > struct drm_display_mode *mode) {
> > struct msm_display_topology topology = {0};
> > + struct dpu_encoder_virt *dpu_enc;
> > + struct drm_bridge *bridge;
> > int i, intf_count = 0;
> > + bool primary_display = false;
> > +
> > + dpu_enc = to_dpu_encoder_virt(drm_enc);
> >
> > for (i = 0; i < MAX_PHYS_ENCODERS_PER_VIRTUAL; i++)
> > if (dpu_enc->phys_encs[i]) @@ -542,7 +549,12 @@ static
> > struct msm_display_topology dpu_encoder_get_topology(
> > else
> > topology.num_lm = (mode->hdisplay >
> > MAX_HDISPLAY_SPLIT) ? 2 : 1;
> >
> > - if (dpu_enc->disp_info.intf_type == DRM_MODE_ENCODER_DSI) {
> > + drm_for_each_bridge_in_chain(drm_enc, bridge) {
> > + if (bridge->type != DRM_MODE_CONNECTOR_DisplayPort)
> > + primary_display = true;
> > + }
>
> I must admit, I never actually liked the original (intf_type == DSI) check. However
> the new one is even worse. We are checking the whole bridge chain in an
> attempt to rule out the DP ports for whatever reason. What about the HDMI
> ports? Should they be also frowned upon?
> The ugly part is that we are making the decision for the user, which displays are
> "primary" for him. Can we let the user make this setting?
>
> > +
> > + if (primary_display) {
> > if (dpu_kms->catalog->dspp &&
> > (dpu_kms->catalog->dspp_count >= topology.num_lm))
> > topology.num_dspp = topology.num_lm; @@ -601,7
> > +613,7 @@ static int dpu_encoder_virt_atomic_check(
> > }
> > }
> >
> > - topology = dpu_encoder_get_topology(dpu_enc, dpu_kms, adj_mode);
> > + topology = dpu_encoder_get_topology(drm_enc, dpu_kms,
> > + adj_mode);
>
> extra whitespace change. Please drop.
>
> >
> > /* Reserve dynamic resources now. */
> > if (!ret) {
> > diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> > b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> > index 1f39327..c4daf7c 100644
> > --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> > +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.h
> > @@ -172,4 +172,9 @@ int dpu_encoder_get_vsync_count(struct
> drm_encoder
> > *drm_enc);
> >
> > bool dpu_encoder_is_widebus_enabled(const struct drm_encoder
> > *drm_enc);
> >
> > +struct msm_display_topology dpu_encoder_get_topology(
> > + struct drm_encoder *drm_enc,
> > + struct dpu_kms *dpu_kms,
> > + struct drm_display_mode *mode);
> > +
> > #endif /* __DPU_ENCODER_H__ */
> > diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> > b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> > index 3a4da0d..486ff9d 100644
> > --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> > +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c
> > @@ -687,9 +687,12 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms
> *dpu_kms)
> > unsigned cursor_idx = 0;
> > unsigned primary_idx = 0;
> > bool pin_overlays;
> > + unsigned int max_dspp_count = 0;
> > + unsigned int enc_mask = 0;
> >
> > struct msm_drm_private *priv;
> > struct dpu_mdss_cfg *catalog;
> > + struct msm_display_topology topology = {0};
> >
> > int primary_planes_idx = 0, cursor_planes_idx = 0, i, ret;
> > int max_crtc_count;
> > @@ -754,10 +757,19 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms
> *dpu_kms)
> > }
> >
> > max_crtc_count = min(max_crtc_count, primary_planes_idx);
> > + max_dspp_count = catalog->dspp_count;
> > +
> > + drm_for_each_encoder(encoder, dev) {
> > + topology = dpu_encoder_get_topology(encoder, dpu_kms,
> > + NULL);
>
> This can crash dpu_encoder_get_topology() because it checks mode->hdisplay.
> And the check anyway is futile here. We do not know if the encoder is going to
> use 1 or 2 LMs (since we do not know the resolution), so we do not know
> whether it will use 1 or 2 DSPP blocks.
>
> > + if (topology.num_dspp > 0 && (topology.num_dspp <=
> max_dspp_count)) {
> > + enc_mask |= BIT(encoder->index);
> > + max_dspp_count -= topology.num_dspp;
> > + }
> > + }
> >
> > /* Create one CRTC per encoder */
> > for (i = 0; i < max_crtc_count; i++) {
> > - crtc = dpu_crtc_init(dev, primary_planes[i], cursor_planes[i]);
> > + crtc = dpu_crtc_init(dev, primary_planes[i],
> > + cursor_planes[i], enc_mask);
> > if (IS_ERR(crtc)) {
> > ret = PTR_ERR(crtc);
> > return ret;
> > @@ -765,9 +777,11 @@ static int _dpu_kms_drm_obj_init(struct dpu_kms
> *dpu_kms)
> > priv->crtcs[priv->num_crtcs++] = crtc;
> > }
> >
> > - /* All CRTCs are compatible with all encoders */
> > - drm_for_each_encoder(encoder, dev)
> > - encoder->possible_crtcs = (1 << priv->num_crtcs) - 1;
> > + /* Attach CRTC's to compatiable encoders */
>
> compatible
>
>
> > + drm_for_each_encoder(encoder, dev) {
> > + encoder->possible_crtcs = (enc_mask & BIT(encoder->index)) ?
> > + BIT(encoder->index) : (((1 << priv->num_crtcs) - 1) &
> ~enc_mask);
> > + }
> >
> > return 0;
> > }
> > --
> > 2.7.4
> >
>
>
> --
> With best wishes
> Dmitry