From: <mikita.lipski@amd.com> To: <amd-gfx@lists.freedesktop.org> Cc: Manasi Navare <manasi.d.navare@intel.com>, David Francis <David.Francis@amd.com>, Mikita Lipski <mikita.lipski@amd.com>, dri-devel@lists.freedesktop.org Subject: [PATCH v8 14/17] drm/amd/display: MST DSC compute fair share Date: Tue, 3 Dec 2019 09:35:27 -0500 [thread overview] Message-ID: <20191203143530.27262-15-mikita.lipski@amd.com> (raw) In-Reply-To: <20191203143530.27262-1-mikita.lipski@amd.com> From: David Francis <David.Francis@amd.com> If there is limited link bandwidth on a MST network, it must be divided fairly between the streams on that network Implement an algorithm to determine the correct DSC config for each stream The algorithm: This [ ] ( ) represents the range of bandwidths possible for a given stream. The [] area represents the range of DSC configs, and the () represents no DSC. The bandwidth used increases from left to right. First, try disabling DSC on all streams [ ] (|) [ ] (|) Check this against the bandwidth limits of the link and each branch (including each endpoint). If it passes, the job is done Second, try maximum DSC compression on all streams that support DSC [| ] ( ) [| ] ( ) If this does not pass, then enabling this combination of streams is impossible Otherwise, divide the remaining bandwidth evenly amongst the streams [ | ] ( ) [ | ] ( ) If one or more of the streams reach minimum compression, evenly divide the reamining bandwidth amongst the remaining streams [ |] ( ) [ |] ( ) [ | ] ( ) [ | ] ( ) If all streams can reach minimum compression, disable compression greedily [ |] ( ) [ |] ( ) [ ] (|) Perform this algorithm on each full update, on each MST link with at least one DSC stream on it After the configs are computed, call dcn20_add_dsc_to_stream_resource on each stream with DSC enabled. It is only after all streams are created that we can know which of them will need DSC. Do all of this at the end of amdgpu atomic check. If it fails, fail check; This combination of timings cannot be supported. v2: Use drm_dp_mst_atomic_check to validate bw for certain dsc configurations v3: Use dc_dsc_policy structure to get min and max bpp rate for DSC configuration Cc: Lyude Paul <lyude@redhat.com> Cc: Manasi Navare <manasi.d.navare@intel.com> Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com> Signed-off-by: David Francis <David.Francis@amd.com> Signed-off-by: Mikita Lipski <mikita.lipski@amd.com> --- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3 + .../display/amdgpu_dm/amdgpu_dm_mst_types.c | 364 ++++++++++++++++++ .../display/amdgpu_dm/amdgpu_dm_mst_types.h | 3 + .../drm/amd/display/dc/dcn20/dcn20_resource.c | 7 +- .../drm/amd/display/dc/dcn20/dcn20_resource.h | 1 + 5 files changed, 376 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 753a79734817..93a230d956ee 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -8095,6 +8095,9 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, if (ret) goto fail; + if (!compute_mst_dsc_configs_for_state(state, dm_state->context)) + goto fail; + if (dc_validate_global_state(dc, dm_state->context, false) != DC_OK) { ret = -EINVAL; goto fail; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index c376c8ccd391..6d13d1c33530 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -40,6 +40,10 @@ #if defined(CONFIG_DEBUG_FS) #include "amdgpu_dm_debugfs.h" #endif + + +#include "dc/dcn20/dcn20_resource.h" + /* #define TRACE_DPCD */ #ifdef TRACE_DPCD @@ -499,3 +503,363 @@ int dm_mst_get_pbn_divider(struct dc_link *link) return dc_link_bandwidth_kbps(link, dc_link_get_link_cap(link)) / (8 * 1000 * 54); } + +struct dsc_mst_fairness_params { + struct dc_crtc_timing *timing; + struct dc_sink *sink; + struct dc_dsc_bw_range bw_range; + bool compression_possible; + struct drm_dp_mst_port *port; +}; + +struct dsc_mst_fairness_vars { + int pbn; + bool dsc_enabled; + int bpp_x16; +}; + +static int kbps_to_peak_pbn(int kbps) +{ + u64 peak_kbps = kbps; + + peak_kbps *= 1006; + peak_kbps /= 1000; + return (int) DIV_ROUND_UP(peak_kbps * 64, (54 * 8 * 1000)); +} + +static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *params, + struct dsc_mst_fairness_vars *vars, + int count) +{ + int i; + + for (i = 0; i < count; i++) { + memset(¶ms[i].timing->dsc_cfg, 0, sizeof(params[i].timing->dsc_cfg)); + if (vars[i].dsc_enabled && dc_dsc_compute_config( + params[i].sink->ctx->dc->res_pool->dscs[0], + ¶ms[i].sink->sink_dsc_caps.dsc_dec_caps, + params[i].sink->ctx->dc->debug.dsc_min_slice_height_override, + 0, + params[i].timing, + ¶ms[i].timing->dsc_cfg)) { + params[i].timing->flags.DSC = 1; + params[i].timing->dsc_cfg.bits_per_pixel = vars[i].bpp_x16; + } else { + params[i].timing->flags.DSC = 0; + } + } +} + +static int bpp_x16_from_pbn(struct dsc_mst_fairness_params param, int pbn) +{ + struct dc_dsc_config dsc_config; + u64 kbps; + + kbps = (u64)pbn * 994 * 8 * 54 / 64; + dc_dsc_compute_config( + param.sink->ctx->dc->res_pool->dscs[0], + ¶m.sink->sink_dsc_caps.dsc_dec_caps, + param.sink->ctx->dc->debug.dsc_min_slice_height_override, + (int) kbps, param.timing, &dsc_config); + + return dsc_config.bits_per_pixel; +} + +static void increase_dsc_bpp(struct drm_atomic_state *state, + struct dc_link *dc_link, + struct dsc_mst_fairness_params *params, + struct dsc_mst_fairness_vars *vars, + int count) +{ + int i; + bool bpp_increased[MAX_PIPES]; + int initial_slack[MAX_PIPES]; + int min_initial_slack; + int next_index; + int remaining_to_increase = 0; + int pbn_per_timeslot; + int link_timeslots_used; + int fair_pbn_alloc; + + for (i = 0; i < count; i++) { + if (vars[i].dsc_enabled) { + initial_slack[i] = kbps_to_peak_pbn(params[i].bw_range.max_kbps) - vars[i].pbn; + bpp_increased[i] = false; + remaining_to_increase += 1; + } else { + initial_slack[i] = 0; + bpp_increased[i] = true; + } + } + + pbn_per_timeslot = dc_link_bandwidth_kbps(dc_link, + dc_link_get_link_cap(dc_link)) / (8 * 1000 * 54); + + while (remaining_to_increase) { + next_index = -1; + min_initial_slack = -1; + for (i = 0; i < count; i++) { + if (!bpp_increased[i]) { + if (min_initial_slack == -1 || min_initial_slack > initial_slack[i]) { + min_initial_slack = initial_slack[i]; + next_index = i; + } + } + } + + if (next_index == -1) + break; + + link_timeslots_used = 0; + + for (i = 0; i < count; i++) + link_timeslots_used += DIV_ROUND_UP(vars[i].pbn, pbn_per_timeslot); + + fair_pbn_alloc = (63 - link_timeslots_used) / remaining_to_increase * pbn_per_timeslot; + + if (initial_slack[next_index] > fair_pbn_alloc) { + vars[next_index].pbn += fair_pbn_alloc; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + if (!drm_dp_mst_atomic_check(state)) { + vars[next_index].bpp_x16 = bpp_x16_from_pbn(params[next_index], vars[next_index].pbn); + } else { + vars[next_index].pbn -= fair_pbn_alloc; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + } + } else { + vars[next_index].pbn += initial_slack[next_index]; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + if (!drm_dp_mst_atomic_check(state)) { + vars[next_index].bpp_x16 = params[next_index].bw_range.max_target_bpp_x16; + } else { + vars[next_index].pbn -= initial_slack[next_index]; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + } + } + + bpp_increased[next_index] = true; + remaining_to_increase--; + } +} + +static void try_disable_dsc(struct drm_atomic_state *state, + struct dc_link *dc_link, + struct dsc_mst_fairness_params *params, + struct dsc_mst_fairness_vars *vars, + int count) +{ + int i; + bool tried[MAX_PIPES]; + int kbps_increase[MAX_PIPES]; + int max_kbps_increase; + int next_index; + int remaining_to_try = 0; + + for (i = 0; i < count; i++) { + if (vars[i].dsc_enabled && vars[i].bpp_x16 == params[i].bw_range.max_target_bpp_x16) { + kbps_increase[i] = params[i].bw_range.stream_kbps - params[i].bw_range.max_kbps; + tried[i] = false; + remaining_to_try += 1; + } else { + kbps_increase[i] = 0; + tried[i] = true; + } + } + + while (remaining_to_try) { + next_index = -1; + max_kbps_increase = -1; + for (i = 0; i < count; i++) { + if (!tried[i]) { + if (max_kbps_increase == -1 || max_kbps_increase < kbps_increase[i]) { + max_kbps_increase = kbps_increase[i]; + next_index = i; + } + } + } + + if (next_index == -1) + break; + + vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps); + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + 0); + + if (!drm_dp_mst_atomic_check(state)) { + vars[next_index].dsc_enabled = false; + vars[next_index].bpp_x16 = 0; + } else { + vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.max_kbps); + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + } + + tried[next_index] = true; + remaining_to_try--; + } +} + +static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, + struct dc_state *dc_state, + struct dc_link *dc_link) +{ + int i; + struct dc_stream_state *stream; + struct dsc_mst_fairness_params params[MAX_PIPES]; + struct dsc_mst_fairness_vars vars[MAX_PIPES]; + struct amdgpu_dm_connector *aconnector; + int count = 0; + + memset(params, 0, sizeof(params)); + + /* Set up params */ + for (i = 0; i < dc_state->stream_count; i++) { + struct dc_dsc_policy dsc_policy = {0}; + + stream = dc_state->streams[i]; + + if (stream->link != dc_link) + continue; + + stream->timing.flags.DSC = 0; + + params[count].timing = &stream->timing; + params[count].sink = stream->sink; + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; + params[count].port = aconnector->port; + params[count].compression_possible = stream->sink->sink_dsc_caps.dsc_dec_caps.is_dsc_supported; + dc_dsc_get_policy_for_timing(params[count].timing, &dsc_policy); + if (!dc_dsc_compute_bandwidth_range( + stream->sink->ctx->dc->res_pool->dscs[0], + stream->sink->ctx->dc->debug.dsc_min_slice_height_override, + dsc_policy.min_target_bpp, + dsc_policy.max_target_bpp, + &stream->sink->sink_dsc_caps.dsc_dec_caps, + &stream->timing, ¶ms[count].bw_range)) + params[count].bw_range.stream_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing); + + count++; + } + /* Try no compression */ + for (i = 0; i < count; i++) { + vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps); + vars[i].dsc_enabled = false; + vars[i].bpp_x16 = 0; + drm_dp_atomic_find_vcpi_slots(state, + params[i].port->mgr, + params[i].port, + vars[i].pbn, + 0); + } + if (!drm_dp_mst_atomic_check(state)) { + set_dsc_configs_from_fairness_vars(params, vars, count); + return true; + } + + /* Try max compression */ + for (i = 0; i < count; i++) { + if (params[i].compression_possible) { + vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps); + vars[i].dsc_enabled = true; + vars[i].bpp_x16 = params[i].bw_range.min_target_bpp_x16; + drm_dp_atomic_find_vcpi_slots(state, + params[i].port->mgr, + params[i].port, + vars[i].pbn, + dm_mst_get_pbn_divider(dc_link)); + } else { + vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps); + vars[i].dsc_enabled = false; + vars[i].bpp_x16 = 0; + drm_dp_atomic_find_vcpi_slots(state, + params[i].port->mgr, + params[i].port, + vars[i].pbn, + 0); + } + } + if (drm_dp_mst_atomic_check(state)) + return false; + + /* Optimize degree of compression */ + increase_dsc_bpp(state, dc_link, params, vars, count); + + try_disable_dsc(state, dc_link, params, vars, count); + + set_dsc_configs_from_fairness_vars(params, vars, count); + + return true; +} + +bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, + struct dc_state *dc_state) +{ + int i, j; + struct dc_stream_state *stream; + bool computed_streams[MAX_PIPES]; + struct amdgpu_dm_connector *aconnector; + + for (i = 0; i < dc_state->stream_count; i++) + computed_streams[i] = false; + + for (i = 0; i < dc_state->stream_count; i++) { + stream = dc_state->streams[i]; + + if (stream->signal != SIGNAL_TYPE_DISPLAY_PORT_MST) + continue; + + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; + + if (!aconnector || !aconnector->dc_sink) + continue; + + if (!aconnector->dc_sink->sink_dsc_caps.dsc_dec_caps.is_dsc_supported) + continue; + + if (computed_streams[i]) + continue; + + mutex_lock(&aconnector->mst_mgr.lock); + if (!compute_mst_dsc_configs_for_link(state, dc_state, stream->link)) { + mutex_unlock(&aconnector->mst_mgr.lock); + return false; + } + mutex_unlock(&aconnector->mst_mgr.lock); + + for (j = 0; j < dc_state->stream_count; j++) { + if (dc_state->streams[j]->link == stream->link) + computed_streams[j] = true; + } + } + + for (i = 0; i < dc_state->stream_count; i++) { + stream = dc_state->streams[i]; + + if (stream->timing.flags.DSC == 1) + dcn20_add_dsc_to_stream_resource(stream->ctx->dc, dc_state, stream); + } + + return true; +} diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h index a553ea046185..d451ce9cecc0 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h @@ -34,4 +34,7 @@ int dm_mst_get_pbn_divider(struct dc_link *link); void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm, struct amdgpu_dm_connector *aconnector); + +bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, + struct dc_state *dc_state); #endif diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c index 2ccfd84a7da4..7f03a6b5ae1b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c @@ -1558,7 +1558,7 @@ static void release_dsc(struct resource_context *res_ctx, -static enum dc_status add_dsc_to_stream_resource(struct dc *dc, +enum dc_status dcn20_add_dsc_to_stream_resource(struct dc *dc, struct dc_state *dc_ctx, struct dc_stream_state *dc_stream) { @@ -1573,6 +1573,9 @@ static enum dc_status add_dsc_to_stream_resource(struct dc *dc, if (pipe_ctx->stream != dc_stream) continue; + if (pipe_ctx->stream_res.dsc) + continue; + acquire_dsc(&dc_ctx->res_ctx, pool, &pipe_ctx->stream_res.dsc); /* The number of DSCs can be less than the number of pipes */ @@ -1622,7 +1625,7 @@ enum dc_status dcn20_add_stream_to_ctx(struct dc *dc, struct dc_state *new_ctx, /* Get a DSC if required and available */ if (result == DC_OK && dc_stream->timing.flags.DSC) - result = add_dsc_to_stream_resource(dc, new_ctx, dc_stream); + result = dcn20_add_dsc_to_stream_resource(dc, new_ctx, dc_stream); if (result == DC_OK) result = dcn20_build_mapped_resource(dc, new_ctx, dc_stream); diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h index 840ca66c34e1..f5893840b79b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h @@ -157,6 +157,7 @@ void dcn20_calculate_dlg_params( enum dc_status dcn20_build_mapped_resource(const struct dc *dc, struct dc_state *context, struct dc_stream_state *stream); enum dc_status dcn20_add_stream_to_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream); +enum dc_status dcn20_add_dsc_to_stream_resource(struct dc *dc, struct dc_state *dc_ctx, struct dc_stream_state *dc_stream); enum dc_status dcn20_remove_stream_from_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream); enum dc_status dcn20_get_default_swizzle_mode(struct dc_plane_state *plane_state); -- 2.17.1 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
WARNING: multiple messages have this Message-ID (diff)
From: <mikita.lipski@amd.com> To: <amd-gfx@lists.freedesktop.org> Cc: Manasi Navare <manasi.d.navare@intel.com>, David Francis <David.Francis@amd.com>, Mikita Lipski <mikita.lipski@amd.com>, dri-devel@lists.freedesktop.org Subject: [PATCH v8 14/17] drm/amd/display: MST DSC compute fair share Date: Tue, 3 Dec 2019 09:35:27 -0500 [thread overview] Message-ID: <20191203143530.27262-15-mikita.lipski@amd.com> (raw) In-Reply-To: <20191203143530.27262-1-mikita.lipski@amd.com> From: David Francis <David.Francis@amd.com> If there is limited link bandwidth on a MST network, it must be divided fairly between the streams on that network Implement an algorithm to determine the correct DSC config for each stream The algorithm: This [ ] ( ) represents the range of bandwidths possible for a given stream. The [] area represents the range of DSC configs, and the () represents no DSC. The bandwidth used increases from left to right. First, try disabling DSC on all streams [ ] (|) [ ] (|) Check this against the bandwidth limits of the link and each branch (including each endpoint). If it passes, the job is done Second, try maximum DSC compression on all streams that support DSC [| ] ( ) [| ] ( ) If this does not pass, then enabling this combination of streams is impossible Otherwise, divide the remaining bandwidth evenly amongst the streams [ | ] ( ) [ | ] ( ) If one or more of the streams reach minimum compression, evenly divide the reamining bandwidth amongst the remaining streams [ |] ( ) [ |] ( ) [ | ] ( ) [ | ] ( ) If all streams can reach minimum compression, disable compression greedily [ |] ( ) [ |] ( ) [ ] (|) Perform this algorithm on each full update, on each MST link with at least one DSC stream on it After the configs are computed, call dcn20_add_dsc_to_stream_resource on each stream with DSC enabled. It is only after all streams are created that we can know which of them will need DSC. Do all of this at the end of amdgpu atomic check. If it fails, fail check; This combination of timings cannot be supported. v2: Use drm_dp_mst_atomic_check to validate bw for certain dsc configurations v3: Use dc_dsc_policy structure to get min and max bpp rate for DSC configuration Cc: Lyude Paul <lyude@redhat.com> Cc: Manasi Navare <manasi.d.navare@intel.com> Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com> Signed-off-by: David Francis <David.Francis@amd.com> Signed-off-by: Mikita Lipski <mikita.lipski@amd.com> --- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3 + .../display/amdgpu_dm/amdgpu_dm_mst_types.c | 364 ++++++++++++++++++ .../display/amdgpu_dm/amdgpu_dm_mst_types.h | 3 + .../drm/amd/display/dc/dcn20/dcn20_resource.c | 7 +- .../drm/amd/display/dc/dcn20/dcn20_resource.h | 1 + 5 files changed, 376 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 753a79734817..93a230d956ee 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -8095,6 +8095,9 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, if (ret) goto fail; + if (!compute_mst_dsc_configs_for_state(state, dm_state->context)) + goto fail; + if (dc_validate_global_state(dc, dm_state->context, false) != DC_OK) { ret = -EINVAL; goto fail; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index c376c8ccd391..6d13d1c33530 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -40,6 +40,10 @@ #if defined(CONFIG_DEBUG_FS) #include "amdgpu_dm_debugfs.h" #endif + + +#include "dc/dcn20/dcn20_resource.h" + /* #define TRACE_DPCD */ #ifdef TRACE_DPCD @@ -499,3 +503,363 @@ int dm_mst_get_pbn_divider(struct dc_link *link) return dc_link_bandwidth_kbps(link, dc_link_get_link_cap(link)) / (8 * 1000 * 54); } + +struct dsc_mst_fairness_params { + struct dc_crtc_timing *timing; + struct dc_sink *sink; + struct dc_dsc_bw_range bw_range; + bool compression_possible; + struct drm_dp_mst_port *port; +}; + +struct dsc_mst_fairness_vars { + int pbn; + bool dsc_enabled; + int bpp_x16; +}; + +static int kbps_to_peak_pbn(int kbps) +{ + u64 peak_kbps = kbps; + + peak_kbps *= 1006; + peak_kbps /= 1000; + return (int) DIV_ROUND_UP(peak_kbps * 64, (54 * 8 * 1000)); +} + +static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *params, + struct dsc_mst_fairness_vars *vars, + int count) +{ + int i; + + for (i = 0; i < count; i++) { + memset(¶ms[i].timing->dsc_cfg, 0, sizeof(params[i].timing->dsc_cfg)); + if (vars[i].dsc_enabled && dc_dsc_compute_config( + params[i].sink->ctx->dc->res_pool->dscs[0], + ¶ms[i].sink->sink_dsc_caps.dsc_dec_caps, + params[i].sink->ctx->dc->debug.dsc_min_slice_height_override, + 0, + params[i].timing, + ¶ms[i].timing->dsc_cfg)) { + params[i].timing->flags.DSC = 1; + params[i].timing->dsc_cfg.bits_per_pixel = vars[i].bpp_x16; + } else { + params[i].timing->flags.DSC = 0; + } + } +} + +static int bpp_x16_from_pbn(struct dsc_mst_fairness_params param, int pbn) +{ + struct dc_dsc_config dsc_config; + u64 kbps; + + kbps = (u64)pbn * 994 * 8 * 54 / 64; + dc_dsc_compute_config( + param.sink->ctx->dc->res_pool->dscs[0], + ¶m.sink->sink_dsc_caps.dsc_dec_caps, + param.sink->ctx->dc->debug.dsc_min_slice_height_override, + (int) kbps, param.timing, &dsc_config); + + return dsc_config.bits_per_pixel; +} + +static void increase_dsc_bpp(struct drm_atomic_state *state, + struct dc_link *dc_link, + struct dsc_mst_fairness_params *params, + struct dsc_mst_fairness_vars *vars, + int count) +{ + int i; + bool bpp_increased[MAX_PIPES]; + int initial_slack[MAX_PIPES]; + int min_initial_slack; + int next_index; + int remaining_to_increase = 0; + int pbn_per_timeslot; + int link_timeslots_used; + int fair_pbn_alloc; + + for (i = 0; i < count; i++) { + if (vars[i].dsc_enabled) { + initial_slack[i] = kbps_to_peak_pbn(params[i].bw_range.max_kbps) - vars[i].pbn; + bpp_increased[i] = false; + remaining_to_increase += 1; + } else { + initial_slack[i] = 0; + bpp_increased[i] = true; + } + } + + pbn_per_timeslot = dc_link_bandwidth_kbps(dc_link, + dc_link_get_link_cap(dc_link)) / (8 * 1000 * 54); + + while (remaining_to_increase) { + next_index = -1; + min_initial_slack = -1; + for (i = 0; i < count; i++) { + if (!bpp_increased[i]) { + if (min_initial_slack == -1 || min_initial_slack > initial_slack[i]) { + min_initial_slack = initial_slack[i]; + next_index = i; + } + } + } + + if (next_index == -1) + break; + + link_timeslots_used = 0; + + for (i = 0; i < count; i++) + link_timeslots_used += DIV_ROUND_UP(vars[i].pbn, pbn_per_timeslot); + + fair_pbn_alloc = (63 - link_timeslots_used) / remaining_to_increase * pbn_per_timeslot; + + if (initial_slack[next_index] > fair_pbn_alloc) { + vars[next_index].pbn += fair_pbn_alloc; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + if (!drm_dp_mst_atomic_check(state)) { + vars[next_index].bpp_x16 = bpp_x16_from_pbn(params[next_index], vars[next_index].pbn); + } else { + vars[next_index].pbn -= fair_pbn_alloc; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + } + } else { + vars[next_index].pbn += initial_slack[next_index]; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + if (!drm_dp_mst_atomic_check(state)) { + vars[next_index].bpp_x16 = params[next_index].bw_range.max_target_bpp_x16; + } else { + vars[next_index].pbn -= initial_slack[next_index]; + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + } + } + + bpp_increased[next_index] = true; + remaining_to_increase--; + } +} + +static void try_disable_dsc(struct drm_atomic_state *state, + struct dc_link *dc_link, + struct dsc_mst_fairness_params *params, + struct dsc_mst_fairness_vars *vars, + int count) +{ + int i; + bool tried[MAX_PIPES]; + int kbps_increase[MAX_PIPES]; + int max_kbps_increase; + int next_index; + int remaining_to_try = 0; + + for (i = 0; i < count; i++) { + if (vars[i].dsc_enabled && vars[i].bpp_x16 == params[i].bw_range.max_target_bpp_x16) { + kbps_increase[i] = params[i].bw_range.stream_kbps - params[i].bw_range.max_kbps; + tried[i] = false; + remaining_to_try += 1; + } else { + kbps_increase[i] = 0; + tried[i] = true; + } + } + + while (remaining_to_try) { + next_index = -1; + max_kbps_increase = -1; + for (i = 0; i < count; i++) { + if (!tried[i]) { + if (max_kbps_increase == -1 || max_kbps_increase < kbps_increase[i]) { + max_kbps_increase = kbps_increase[i]; + next_index = i; + } + } + } + + if (next_index == -1) + break; + + vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps); + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + 0); + + if (!drm_dp_mst_atomic_check(state)) { + vars[next_index].dsc_enabled = false; + vars[next_index].bpp_x16 = 0; + } else { + vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.max_kbps); + drm_dp_atomic_find_vcpi_slots(state, + params[next_index].port->mgr, + params[next_index].port, + vars[next_index].pbn, + dm_mst_get_pbn_divider(dc_link)); + } + + tried[next_index] = true; + remaining_to_try--; + } +} + +static bool compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, + struct dc_state *dc_state, + struct dc_link *dc_link) +{ + int i; + struct dc_stream_state *stream; + struct dsc_mst_fairness_params params[MAX_PIPES]; + struct dsc_mst_fairness_vars vars[MAX_PIPES]; + struct amdgpu_dm_connector *aconnector; + int count = 0; + + memset(params, 0, sizeof(params)); + + /* Set up params */ + for (i = 0; i < dc_state->stream_count; i++) { + struct dc_dsc_policy dsc_policy = {0}; + + stream = dc_state->streams[i]; + + if (stream->link != dc_link) + continue; + + stream->timing.flags.DSC = 0; + + params[count].timing = &stream->timing; + params[count].sink = stream->sink; + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; + params[count].port = aconnector->port; + params[count].compression_possible = stream->sink->sink_dsc_caps.dsc_dec_caps.is_dsc_supported; + dc_dsc_get_policy_for_timing(params[count].timing, &dsc_policy); + if (!dc_dsc_compute_bandwidth_range( + stream->sink->ctx->dc->res_pool->dscs[0], + stream->sink->ctx->dc->debug.dsc_min_slice_height_override, + dsc_policy.min_target_bpp, + dsc_policy.max_target_bpp, + &stream->sink->sink_dsc_caps.dsc_dec_caps, + &stream->timing, ¶ms[count].bw_range)) + params[count].bw_range.stream_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing); + + count++; + } + /* Try no compression */ + for (i = 0; i < count; i++) { + vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps); + vars[i].dsc_enabled = false; + vars[i].bpp_x16 = 0; + drm_dp_atomic_find_vcpi_slots(state, + params[i].port->mgr, + params[i].port, + vars[i].pbn, + 0); + } + if (!drm_dp_mst_atomic_check(state)) { + set_dsc_configs_from_fairness_vars(params, vars, count); + return true; + } + + /* Try max compression */ + for (i = 0; i < count; i++) { + if (params[i].compression_possible) { + vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps); + vars[i].dsc_enabled = true; + vars[i].bpp_x16 = params[i].bw_range.min_target_bpp_x16; + drm_dp_atomic_find_vcpi_slots(state, + params[i].port->mgr, + params[i].port, + vars[i].pbn, + dm_mst_get_pbn_divider(dc_link)); + } else { + vars[i].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps); + vars[i].dsc_enabled = false; + vars[i].bpp_x16 = 0; + drm_dp_atomic_find_vcpi_slots(state, + params[i].port->mgr, + params[i].port, + vars[i].pbn, + 0); + } + } + if (drm_dp_mst_atomic_check(state)) + return false; + + /* Optimize degree of compression */ + increase_dsc_bpp(state, dc_link, params, vars, count); + + try_disable_dsc(state, dc_link, params, vars, count); + + set_dsc_configs_from_fairness_vars(params, vars, count); + + return true; +} + +bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, + struct dc_state *dc_state) +{ + int i, j; + struct dc_stream_state *stream; + bool computed_streams[MAX_PIPES]; + struct amdgpu_dm_connector *aconnector; + + for (i = 0; i < dc_state->stream_count; i++) + computed_streams[i] = false; + + for (i = 0; i < dc_state->stream_count; i++) { + stream = dc_state->streams[i]; + + if (stream->signal != SIGNAL_TYPE_DISPLAY_PORT_MST) + continue; + + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; + + if (!aconnector || !aconnector->dc_sink) + continue; + + if (!aconnector->dc_sink->sink_dsc_caps.dsc_dec_caps.is_dsc_supported) + continue; + + if (computed_streams[i]) + continue; + + mutex_lock(&aconnector->mst_mgr.lock); + if (!compute_mst_dsc_configs_for_link(state, dc_state, stream->link)) { + mutex_unlock(&aconnector->mst_mgr.lock); + return false; + } + mutex_unlock(&aconnector->mst_mgr.lock); + + for (j = 0; j < dc_state->stream_count; j++) { + if (dc_state->streams[j]->link == stream->link) + computed_streams[j] = true; + } + } + + for (i = 0; i < dc_state->stream_count; i++) { + stream = dc_state->streams[i]; + + if (stream->timing.flags.DSC == 1) + dcn20_add_dsc_to_stream_resource(stream->ctx->dc, dc_state, stream); + } + + return true; +} diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h index a553ea046185..d451ce9cecc0 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.h @@ -34,4 +34,7 @@ int dm_mst_get_pbn_divider(struct dc_link *link); void amdgpu_dm_initialize_dp_connector(struct amdgpu_display_manager *dm, struct amdgpu_dm_connector *aconnector); + +bool compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, + struct dc_state *dc_state); #endif diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c index 2ccfd84a7da4..7f03a6b5ae1b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c @@ -1558,7 +1558,7 @@ static void release_dsc(struct resource_context *res_ctx, -static enum dc_status add_dsc_to_stream_resource(struct dc *dc, +enum dc_status dcn20_add_dsc_to_stream_resource(struct dc *dc, struct dc_state *dc_ctx, struct dc_stream_state *dc_stream) { @@ -1573,6 +1573,9 @@ static enum dc_status add_dsc_to_stream_resource(struct dc *dc, if (pipe_ctx->stream != dc_stream) continue; + if (pipe_ctx->stream_res.dsc) + continue; + acquire_dsc(&dc_ctx->res_ctx, pool, &pipe_ctx->stream_res.dsc); /* The number of DSCs can be less than the number of pipes */ @@ -1622,7 +1625,7 @@ enum dc_status dcn20_add_stream_to_ctx(struct dc *dc, struct dc_state *new_ctx, /* Get a DSC if required and available */ if (result == DC_OK && dc_stream->timing.flags.DSC) - result = add_dsc_to_stream_resource(dc, new_ctx, dc_stream); + result = dcn20_add_dsc_to_stream_resource(dc, new_ctx, dc_stream); if (result == DC_OK) result = dcn20_build_mapped_resource(dc, new_ctx, dc_stream); diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h index 840ca66c34e1..f5893840b79b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h @@ -157,6 +157,7 @@ void dcn20_calculate_dlg_params( enum dc_status dcn20_build_mapped_resource(const struct dc *dc, struct dc_state *context, struct dc_stream_state *stream); enum dc_status dcn20_add_stream_to_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream); +enum dc_status dcn20_add_dsc_to_stream_resource(struct dc *dc, struct dc_state *dc_ctx, struct dc_stream_state *dc_stream); enum dc_status dcn20_remove_stream_from_ctx(struct dc *dc, struct dc_state *new_ctx, struct dc_stream_state *dc_stream); enum dc_status dcn20_get_default_swizzle_mode(struct dc_plane_state *plane_state); -- 2.17.1 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx
next prev parent reply other threads:[~2019-12-03 14:36 UTC|newest] Thread overview: 70+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-12-03 14:35 [PATCH v8 00/17] DSC MST support for DRM and AMDGPU mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 01/17] drm/dp_mst: Add PBN calculation for DSC modes mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-06 0:22 ` kbuild test robot 2019-12-06 0:22 ` kbuild test robot 2019-12-06 0:22 ` kbuild test robot 2019-12-07 0:17 ` Lyude Paul 2019-12-07 0:17 ` Lyude Paul 2019-12-07 22:13 ` Lipski, Mikita 2019-12-07 22:13 ` Lipski, Mikita 2019-12-03 14:35 ` [PATCH v8 02/17] drm/dp_mst: Parse FEC capability on MST ports mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 21:09 ` kbuild test robot 2019-12-03 21:09 ` kbuild test robot 2019-12-03 21:09 ` kbuild test robot 2019-12-03 14:35 ` [PATCH v8 03/17] drm/dp_mst: Add MST support to DP DPCD R/W functions mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 04/17] drm/dp_mst: Fill branch->num_ports mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-06 4:30 ` Harry Wentland 2019-12-06 4:30 ` Harry Wentland 2019-12-03 14:35 ` [PATCH v8 05/17] drm/dp_mst: Add helpers for MST DSC and virtual DPCD aux mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 06/17] drm/dp_mst: Add new quirk for Synaptics MST hubs mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 07/17] drm/amd/display: Initialize DSC PPS variables to 0 mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 08/17] drm/amd/display: Validate DSC caps on MST endpoints mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 09/17] drm/amd/display: Write DSC enable to MST DPCD mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 10/17] drm/dp_mst: Manually overwrite PBN divider for calculating timeslots mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` [PATCH v8 11/17] drm/dp_mst: Add DSC enablement helpers to DRM mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 21:52 ` kbuild test robot 2019-12-03 21:52 ` kbuild test robot 2019-12-03 21:52 ` kbuild test robot 2019-12-07 0:24 ` Lyude Paul 2019-12-07 0:24 ` Lyude Paul 2019-12-09 14:07 ` Mikita Lipski 2019-12-09 14:07 ` Mikita Lipski 2019-12-03 14:35 ` [PATCH v8 12/17] drm/dp_mst: Add branch bandwidth validation to MST atomic check mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-05 19:28 ` Harry Wentland 2019-12-05 19:28 ` Harry Wentland 2019-12-07 0:30 ` Lyude Paul 2019-12-07 0:30 ` Lyude Paul 2019-12-03 14:35 ` [PATCH v8 13/17] drm/amd/display: Add PBN per slot calculation for DSC mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 14:35 ` mikita.lipski [this message] 2019-12-03 14:35 ` [PATCH v8 14/17] drm/amd/display: MST DSC compute fair share mikita.lipski 2019-12-07 0:53 ` Lyude Paul 2019-12-07 0:53 ` Lyude Paul 2019-12-03 14:35 ` [PATCH v8 15/17] drm/amd/display: Recalculate VCPI slots for new DSC connectors mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-07 0:52 ` Lyude Paul 2019-12-07 0:52 ` Lyude Paul 2019-12-03 14:35 ` [PATCH v8 16/17] drm/dp_mst: Add helper to trigger modeset on affected DSC MST CRTCs mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-03 22:40 ` kbuild test robot 2019-12-03 22:40 ` kbuild test robot 2019-12-03 22:40 ` kbuild test robot 2019-12-07 0:51 ` Lyude Paul 2019-12-07 0:51 ` Lyude Paul 2019-12-03 14:35 ` [PATCH v8 17/17] drm/amd/display: Trigger modesets on MST DSC connectors mikita.lipski 2019-12-03 14:35 ` mikita.lipski 2019-12-07 0:53 ` Lyude Paul 2019-12-07 0:53 ` Lyude Paul
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=20191203143530.27262-15-mikita.lipski@amd.com \ --to=mikita.lipski@amd.com \ --cc=David.Francis@amd.com \ --cc=amd-gfx@lists.freedesktop.org \ --cc=dri-devel@lists.freedesktop.org \ --cc=manasi.d.navare@intel.com \ /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: linkBe 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.