[PATCH v2 2/2] drm/komeda: Enable writeback split support
james qian wang (Arm Technology China)
james.qian.wang at arm.com
Wed Jun 12 10:20:18 UTC 2019
Writeback split is also for workaround the size limitation of d71 scaler.
Like layer_split, writeback downscaling also can use two scalers to handle
the scaling half-by-half. The only differnence is writback needs a
standalone component (splitter)'s help to split the composition result.
The data pipeline of writeback split as below:
/-> scaler-0 ->\
compiz -> splitter merger -> wb_layer -> memory
\-> scaler-1 ->/
v2: Rebase
Signed-off-by: James Qian Wang (Arm Technology China) <james.qian.wang at arm.com>
---
.../drm/arm/display/komeda/komeda_pipeline.h | 4 +
.../display/komeda/komeda_pipeline_state.c | 117 ++++++++++++++++--
.../arm/display/komeda/komeda_wb_connector.c | 17 ++-
3 files changed, 124 insertions(+), 14 deletions(-)
diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h
index b6fed54b1cf1..c0378b18f803 100644
--- a/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h
+++ b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h
@@ -519,6 +519,10 @@ int komeda_build_layer_split_data_flow(struct komeda_layer *left,
struct komeda_plane_state *kplane_st,
struct komeda_crtc_state *kcrtc_st,
struct komeda_data_flow_cfg *dflow);
+int komeda_build_wb_split_data_flow(struct komeda_layer *wb_layer,
+ struct drm_connector_state *conn_st,
+ struct komeda_crtc_state *kcrtc_st,
+ struct komeda_data_flow_cfg *dflow);
int komeda_release_unclaimed_resources(struct komeda_pipeline *pipe,
struct komeda_crtc_state *kcrtc_st);
diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_pipeline_state.c b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline_state.c
index 23182edeb09a..b58a32f1b158 100644
--- a/drivers/gpu/drm/arm/display/komeda/komeda_pipeline_state.c
+++ b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline_state.c
@@ -284,30 +284,33 @@ komeda_layer_check_cfg(struct komeda_layer *layer,
struct komeda_fb *kfb,
struct komeda_data_flow_cfg *dflow)
{
- u32 hsize_in, vsize_in;
+ u32 src_x, src_y, src_w, src_h;
if (!komeda_fb_is_layer_supported(kfb, layer->layer_type, dflow->rot))
return -EINVAL;
- if (komeda_fb_check_src_coords(kfb, dflow->in_x, dflow->in_y,
- dflow->in_w, dflow->in_h))
- return -EINVAL;
-
if (layer->base.id == KOMEDA_COMPONENT_WB_LAYER) {
- hsize_in = dflow->out_w;
- vsize_in = dflow->out_h;
+ src_x = dflow->out_x;
+ src_y = dflow->out_y;
+ src_w = dflow->out_w;
+ src_h = dflow->out_h;
} else {
- hsize_in = dflow->in_w;
- vsize_in = dflow->in_h;
+ src_x = dflow->in_x;
+ src_y = dflow->in_y;
+ src_w = dflow->in_w;
+ src_h = dflow->in_h;
}
- if (!in_range(&layer->hsize_in, hsize_in)) {
- DRM_DEBUG_ATOMIC("invalidate src_w %d.\n", hsize_in);
+ if (komeda_fb_check_src_coords(kfb, src_x, src_y, src_w, src_h))
+ return -EINVAL;
+
+ if (!in_range(&layer->hsize_in, src_w)) {
+ DRM_DEBUG_ATOMIC("invalidate src_w %d.\n", src_w);
return -EINVAL;
}
- if (!in_range(&layer->vsize_in, vsize_in)) {
- DRM_DEBUG_ATOMIC("invalidate src_h %d.\n", vsize_in);
+ if (!in_range(&layer->vsize_in, src_h)) {
+ DRM_DEBUG_ATOMIC("invalidate src_h %d.\n", src_h);
return -EINVAL;
}
@@ -534,6 +537,59 @@ komeda_scaler_validate(void *user,
return err;
}
+static void komeda_split_data_flow(struct komeda_scaler *scaler,
+ struct komeda_data_flow_cfg *dflow,
+ struct komeda_data_flow_cfg *l_dflow,
+ struct komeda_data_flow_cfg *r_dflow);
+
+static int
+komeda_splitter_validate(struct komeda_splitter *splitter,
+ struct drm_connector_state *conn_st,
+ struct komeda_data_flow_cfg *dflow,
+ struct komeda_data_flow_cfg *l_output,
+ struct komeda_data_flow_cfg *r_output)
+{
+ struct komeda_component_state *c_st;
+ struct komeda_splitter_state *st;
+
+ if (!splitter) {
+ DRM_DEBUG_ATOMIC("Current HW doesn't support splitter.\n");
+ return -EINVAL;
+ }
+
+ if (!in_range(&splitter->hsize, dflow->in_w)) {
+ DRM_DEBUG_ATOMIC("split in_w:%d is out of the acceptable range.\n",
+ dflow->in_w);
+ return -EINVAL;
+ }
+
+ if (!in_range(&splitter->vsize, dflow->in_h)) {
+ DRM_DEBUG_ATOMIC("split in_in: %d exceed the acceptable range.\n",
+ dflow->in_w);
+ return -EINVAL;
+ }
+
+ c_st = komeda_component_get_state_and_set_user(&splitter->base,
+ conn_st->state, conn_st->connector, conn_st->crtc);
+
+ if (IS_ERR(c_st))
+ return PTR_ERR(c_st);
+
+ komeda_split_data_flow(splitter->base.pipeline->scalers[0],
+ dflow, l_output, r_output);
+
+ st = to_splitter_st(c_st);
+ st->hsize = dflow->in_w;
+ st->vsize = dflow->in_h;
+ st->overlap = dflow->overlap;
+
+ komeda_component_add_input(&st->base, &dflow->input, 0);
+ komeda_component_set_output(&l_output->input, &splitter->base, 0);
+ komeda_component_set_output(&r_output->input, &splitter->base, 1);
+
+ return 0;
+}
+
static int
komeda_merger_validate(struct komeda_merger *merger,
void *user,
@@ -1025,6 +1081,41 @@ int komeda_build_wb_data_flow(struct komeda_layer *wb_layer,
return komeda_wb_layer_validate(wb_layer, conn_st, dflow);
}
+/* writeback scaling split data path:
+ * /-> scaler ->\
+ * compiz -> splitter merger -> wb_layer -> memory
+ * \-> scaler ->/
+ */
+int komeda_build_wb_split_data_flow(struct komeda_layer *wb_layer,
+ struct drm_connector_state *conn_st,
+ struct komeda_crtc_state *kcrtc_st,
+ struct komeda_data_flow_cfg *dflow)
+{
+ struct komeda_pipeline *pipe = wb_layer->base.pipeline;
+ struct drm_connector *conn = conn_st->connector;
+ struct komeda_data_flow_cfg l_dflow, r_dflow;
+ int err;
+
+ err = komeda_splitter_validate(pipe->splitter, conn_st,
+ dflow, &l_dflow, &r_dflow);
+ if (err)
+ return err;
+ err = komeda_scaler_validate(conn, kcrtc_st, &l_dflow);
+ if (err)
+ return err;
+
+ err = komeda_scaler_validate(conn, kcrtc_st, &r_dflow);
+ if (err)
+ return err;
+
+ err = komeda_merger_validate(pipe->merger, conn_st, kcrtc_st,
+ &l_dflow, &r_dflow, dflow);
+ if (err)
+ return err;
+
+ return komeda_wb_layer_validate(wb_layer, conn_st, dflow);
+}
+
/* build display output data flow, the data path is:
* compiz -> improc -> timing_ctrlr
*/
diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_wb_connector.c b/drivers/gpu/drm/arm/display/komeda/komeda_wb_connector.c
index 0d734244f662..bb8a61f6e9a4 100644
--- a/drivers/gpu/drm/arm/display/komeda/komeda_wb_connector.c
+++ b/drivers/gpu/drm/arm/display/komeda/komeda_wb_connector.c
@@ -13,6 +13,7 @@ komeda_wb_init_data_flow(struct komeda_layer *wb_layer,
struct komeda_crtc_state *kcrtc_st,
struct komeda_data_flow_cfg *dflow)
{
+ struct komeda_scaler *scaler = wb_layer->base.pipeline->scalers[0];
struct drm_framebuffer *fb = conn_st->writeback_job->fb;
memset(dflow, 0, sizeof(*dflow));
@@ -29,6 +30,13 @@ komeda_wb_init_data_flow(struct komeda_layer *wb_layer,
komeda_complete_data_flow_cfg(dflow, fb);
+ /* if scaling exceed the acceptable scaler input/output range, try to
+ * enable split.
+ */
+ if (dflow->en_scaling && scaler)
+ dflow->en_split = !in_range(&scaler->hsize, dflow->in_w) ||
+ !in_range(&scaler->hsize, dflow->out_w);
+
return 0;
}
@@ -66,7 +74,14 @@ komeda_wb_encoder_atomic_check(struct drm_encoder *encoder,
if (err)
return err;
- return komeda_build_wb_data_flow(wb_layer, conn_st, kcrtc_st, &dflow);
+ if (dflow.en_split)
+ err = komeda_build_wb_split_data_flow(wb_layer,
+ conn_st, kcrtc_st, &dflow);
+ else
+ err = komeda_build_wb_data_flow(wb_layer,
+ conn_st, kcrtc_st, &dflow);
+
+ return err;
}
static const struct drm_encoder_helper_funcs komeda_wb_encoder_helper_funcs = {
--
2.17.1
More information about the dri-devel
mailing list