diff options
Diffstat (limited to 'drivers/gpu/drm/amd')
36 files changed, 615 insertions, 266 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index fa8ac9d19a7a..6326c1792270 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -1304,7 +1304,7 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( } /* Free the BO*/ - amdgpu_bo_unref(&mem->bo); + drm_gem_object_put_unlocked(&mem->bo->tbo.base); mutex_destroy(&mem->lock); kfree(mem); @@ -1647,7 +1647,8 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd, ALLOC_MEM_FLAGS_VRAM : ALLOC_MEM_FLAGS_GTT) | ALLOC_MEM_FLAGS_WRITABLE | ALLOC_MEM_FLAGS_EXECUTABLE; - (*mem)->bo = amdgpu_bo_ref(bo); + drm_gem_object_get(&bo->tbo.base); + (*mem)->bo = bo; (*mem)->va = va; (*mem)->domain = (bo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index f184cdca938d..5fcbacddb9b0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3325,15 +3325,12 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) } } - amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE); - amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE); - - amdgpu_amdkfd_suspend(adev); - amdgpu_ras_suspend(adev); r = amdgpu_device_ip_suspend_phase1(adev); + amdgpu_amdkfd_suspend(adev); + /* evict vram memory */ amdgpu_bo_evict_vram(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c index 2672dc64a310..6a76ab16500f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c @@ -133,8 +133,7 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev, u32 cpp; u64 flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED | AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | - AMDGPU_GEM_CREATE_VRAM_CLEARED | - AMDGPU_GEM_CREATE_CPU_GTT_USWC; + AMDGPU_GEM_CREATE_VRAM_CLEARED; info = drm_get_format_info(adev->ddev, mode_cmd); cpp = info->cpp[0]; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index b03b1eb7ba04..1ae174c3d160 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -91,7 +91,8 @@ void amdgpu_pm_acpi_event_handler(struct amdgpu_device *adev) adev->pm.ac_power = true; else adev->pm.ac_power = false; - if (adev->powerplay.pp_funcs->enable_bapm) + if (adev->powerplay.pp_funcs && + adev->powerplay.pp_funcs->enable_bapm) amdgpu_dpm_enable_bapm(adev, adev->pm.ac_power); mutex_unlock(&adev->pm.mutex); } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index 02702597ddeb..012df3d574bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -4241,11 +4241,7 @@ static int gfx_v10_0_set_powergating_state(void *handle, switch (adev->asic_type) { case CHIP_NAVI10: case CHIP_NAVI14: - if (!enable) { - amdgpu_gfx_off_ctrl(adev, false); - cancel_delayed_work_sync(&adev->gfx.gfx_off_delay_work); - } else - amdgpu_gfx_off_ctrl(adev, true); + amdgpu_gfx_off_ctrl(adev, enable); break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index 73337e658aff..914dbd901b98 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -1177,6 +1177,8 @@ static const struct amdgpu_gfxoff_quirk amdgpu_gfxoff_quirk_list[] = { { 0x1002, 0x15dd, 0x1002, 0x15dd, 0xc8 }, /* https://bugzilla.kernel.org/show_bug.cgi?id=207171 */ { 0x1002, 0x15dd, 0x103c, 0x83e7, 0xd3 }, + /* GFXOFF is unstable on C6 parts with a VBIOS 113-RAVEN-114 */ + { 0x1002, 0x15dd, 0x1002, 0x15dd, 0xc6 }, { 0, 0, 0, 0, 0 }, }; @@ -4732,10 +4734,9 @@ static int gfx_v9_0_set_powergating_state(void *handle, switch (adev->asic_type) { case CHIP_RAVEN: case CHIP_RENOIR: - if (!enable) { + if (!enable) amdgpu_gfx_off_ctrl(adev, false); - cancel_delayed_work_sync(&adev->gfx.gfx_off_delay_work); - } + if (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS) { gfx_v9_0_enable_sck_slow_down_on_power_up(adev, true); gfx_v9_0_enable_sck_slow_down_on_power_down(adev, true); @@ -4759,12 +4760,7 @@ static int gfx_v9_0_set_powergating_state(void *handle, amdgpu_gfx_off_ctrl(adev, true); break; case CHIP_VEGA12: - if (!enable) { - amdgpu_gfx_off_ctrl(adev, false); - cancel_delayed_work_sync(&adev->gfx.gfx_off_delay_work); - } else { - amdgpu_gfx_off_ctrl(adev, true); - } + amdgpu_gfx_off_ctrl(adev, enable); break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h b/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h index 074a9a09c0a7..a5b60c9a2418 100644 --- a/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h +++ b/drivers/gpu/drm/amd/amdgpu/navi10_sdma_pkt_open.h @@ -73,6 +73,22 @@ #define SDMA_OP_AQL_COPY 0 #define SDMA_OP_AQL_BARRIER_OR 0 +#define SDMA_GCR_RANGE_IS_PA (1 << 18) +#define SDMA_GCR_SEQ(x) (((x) & 0x3) << 16) +#define SDMA_GCR_GL2_WB (1 << 15) +#define SDMA_GCR_GL2_INV (1 << 14) +#define SDMA_GCR_GL2_DISCARD (1 << 13) +#define SDMA_GCR_GL2_RANGE(x) (((x) & 0x3) << 11) +#define SDMA_GCR_GL2_US (1 << 10) +#define SDMA_GCR_GL1_INV (1 << 9) +#define SDMA_GCR_GLV_INV (1 << 8) +#define SDMA_GCR_GLK_INV (1 << 7) +#define SDMA_GCR_GLK_WB (1 << 6) +#define SDMA_GCR_GLM_INV (1 << 5) +#define SDMA_GCR_GLM_WB (1 << 4) +#define SDMA_GCR_GL1_RANGE(x) (((x) & 0x3) << 2) +#define SDMA_GCR_GLI_INV(x) (((x) & 0x3) << 0) + /*define for op field*/ #define SDMA_PKT_HEADER_op_offset 0 #define SDMA_PKT_HEADER_op_mask 0x000000FF diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c index 67b9830b7c7e..ddc8b217e8c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c @@ -382,6 +382,18 @@ static void sdma_v5_0_ring_emit_ib(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); uint64_t csa_mc_addr = amdgpu_sdma_get_csa_mc_addr(ring, vmid); + /* Invalidate L2, because if we don't do it, we might get stale cache + * lines from previous IBs. + */ + amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_GCR_REQ)); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, (SDMA_GCR_GL2_INV | + SDMA_GCR_GL2_WB | + SDMA_GCR_GLM_INV | + SDMA_GCR_GLM_WB) << 16); + amdgpu_ring_write(ring, 0xffffff80); + amdgpu_ring_write(ring, 0xffff); + /* An IB packet must end on a 8 DW boundary--the next dword * must be on a 8-dword boundary. Our IB packet below is 6 * dwords long, thus add x number of NOPs, such that, in @@ -1597,7 +1609,7 @@ static const struct amdgpu_ring_funcs sdma_v5_0_ring_funcs = { SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 * 2 + 10 + 10 + 10, /* sdma_v5_0_ring_emit_fence x3 for user fence, vm fence */ - .emit_ib_size = 7 + 6, /* sdma_v5_0_ring_emit_ib */ + .emit_ib_size = 5 + 7 + 6, /* sdma_v5_0_ring_emit_ib */ .emit_ib = sdma_v5_0_ring_emit_ib, .emit_fence = sdma_v5_0_ring_emit_fence, .emit_pipeline_sync = sdma_v5_0_ring_emit_pipeline_sync, 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 6240259b3a93..8e7cffe10cc5 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -1941,17 +1941,22 @@ amdgpu_dm_update_connector_after_detect(struct amdgpu_dm_connector *aconnector) dc_sink_retain(aconnector->dc_sink); if (sink->dc_edid.length == 0) { aconnector->edid = NULL; - drm_dp_cec_unset_edid(&aconnector->dm_dp_aux.aux); + if (aconnector->dc_link->aux_mode) { + drm_dp_cec_unset_edid( + &aconnector->dm_dp_aux.aux); + } } else { aconnector->edid = - (struct edid *) sink->dc_edid.raw_edid; - + (struct edid *)sink->dc_edid.raw_edid; drm_connector_update_edid_property(connector, - aconnector->edid); - drm_dp_cec_set_edid(&aconnector->dm_dp_aux.aux, - aconnector->edid); + aconnector->edid); + + if (aconnector->dc_link->aux_mode) + drm_dp_cec_set_edid(&aconnector->dm_dp_aux.aux, + aconnector->edid); } + amdgpu_dm_update_freesync_caps(connector, aconnector->edid); } else { @@ -3212,7 +3217,8 @@ fill_plane_dcc_attributes(struct amdgpu_device *adev, const union dc_tiling_info *tiling_info, const uint64_t info, struct dc_plane_dcc_param *dcc, - struct dc_plane_address *address) + struct dc_plane_address *address, + bool force_disable_dcc) { struct dc *dc = adev->dm.dc; struct dc_dcc_surface_param input; @@ -3224,6 +3230,9 @@ fill_plane_dcc_attributes(struct amdgpu_device *adev, memset(&input, 0, sizeof(input)); memset(&output, 0, sizeof(output)); + if (force_disable_dcc) + return 0; + if (!offset) return 0; @@ -3273,7 +3282,8 @@ fill_plane_buffer_attributes(struct amdgpu_device *adev, union dc_tiling_info *tiling_info, struct plane_size *plane_size, struct dc_plane_dcc_param *dcc, - struct dc_plane_address *address) + struct dc_plane_address *address, + bool force_disable_dcc) { const struct drm_framebuffer *fb = &afb->base; int ret; @@ -3379,7 +3389,8 @@ fill_plane_buffer_attributes(struct amdgpu_device *adev, ret = fill_plane_dcc_attributes(adev, afb, format, rotation, plane_size, tiling_info, - tiling_flags, dcc, address); + tiling_flags, dcc, address, + force_disable_dcc); if (ret) return ret; } @@ -3471,7 +3482,8 @@ fill_dc_plane_info_and_addr(struct amdgpu_device *adev, const struct drm_plane_state *plane_state, const uint64_t tiling_flags, struct dc_plane_info *plane_info, - struct dc_plane_address *address) + struct dc_plane_address *address, + bool force_disable_dcc) { const struct drm_framebuffer *fb = plane_state->fb; const struct amdgpu_framebuffer *afb = @@ -3550,7 +3562,8 @@ fill_dc_plane_info_and_addr(struct amdgpu_device *adev, plane_info->rotation, tiling_flags, &plane_info->tiling_info, &plane_info->plane_size, - &plane_info->dcc, address); + &plane_info->dcc, address, + force_disable_dcc); if (ret) return ret; @@ -3573,6 +3586,7 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev, struct dc_plane_info plane_info; uint64_t tiling_flags; int ret; + bool force_disable_dcc = false; ret = fill_dc_scaling_info(plane_state, &scaling_info); if (ret) @@ -3587,9 +3601,11 @@ static int fill_dc_plane_attributes(struct amdgpu_device *adev, if (ret) return ret; + force_disable_dcc = adev->asic_type == CHIP_RAVEN && adev->in_suspend; ret = fill_dc_plane_info_and_addr(adev, plane_state, tiling_flags, &plane_info, - &dc_plane_state->address); + &dc_plane_state->address, + force_disable_dcc); if (ret) return ret; @@ -5171,6 +5187,7 @@ static int dm_plane_helper_prepare_fb(struct drm_plane *plane, uint64_t tiling_flags; uint32_t domain; int r; + bool force_disable_dcc = false; dm_plane_state_old = to_dm_plane_state(plane->state); dm_plane_state_new = to_dm_plane_state(new_state); @@ -5229,11 +5246,13 @@ static int dm_plane_helper_prepare_fb(struct drm_plane *plane, dm_plane_state_old->dc_state != dm_plane_state_new->dc_state) { struct dc_plane_state *plane_state = dm_plane_state_new->dc_state; + force_disable_dcc = adev->asic_type == CHIP_RAVEN && adev->in_suspend; fill_plane_buffer_attributes( adev, afb, plane_state->format, plane_state->rotation, tiling_flags, &plane_state->tiling_info, &plane_state->plane_size, &plane_state->dcc, - &plane_state->address); + &plane_state->address, + force_disable_dcc); } return 0; @@ -6514,7 +6533,12 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, fill_dc_plane_info_and_addr( dm->adev, new_plane_state, tiling_flags, &bundle->plane_infos[planes_count], - &bundle->flip_addrs[planes_count].address); + &bundle->flip_addrs[planes_count].address, + false); + + DRM_DEBUG_DRIVER("plane: id=%d dcc_en=%d\n", + new_plane_state->plane->index, + bundle->plane_infos[planes_count].dcc.enable); bundle->surface_updates[planes_count].plane_info = &bundle->plane_infos[planes_count]; @@ -7697,6 +7721,7 @@ static int dm_update_plane_state(struct dc *dc, struct drm_crtc_state *old_crtc_state, *new_crtc_state; struct dm_crtc_state *dm_new_crtc_state, *dm_old_crtc_state; struct dm_plane_state *dm_new_plane_state, *dm_old_plane_state; + struct amdgpu_crtc *new_acrtc; bool needs_reset; int ret = 0; @@ -7706,9 +7731,23 @@ static int dm_update_plane_state(struct dc *dc, dm_new_plane_state = to_dm_plane_state(new_plane_state); dm_old_plane_state = to_dm_plane_state(old_plane_state); - /*TODO Implement atomic check for cursor plane */ - if (plane->type == DRM_PLANE_TYPE_CURSOR) + /*TODO Implement better atomic check for cursor plane */ + if (plane->type == DRM_PLANE_TYPE_CURSOR) { + if (!enable || !new_plane_crtc || + drm_atomic_plane_disabling(plane->state, new_plane_state)) + return 0; + + new_acrtc = to_amdgpu_crtc(new_plane_crtc); + + if ((new_plane_state->crtc_w > new_acrtc->max_cursor_width) || + (new_plane_state->crtc_h > new_acrtc->max_cursor_height)) { + DRM_DEBUG_ATOMIC("Bad cursor size %d x %d\n", + new_plane_state->crtc_w, new_plane_state->crtc_h); + return -EINVAL; + } + return 0; + } needs_reset = should_reset_plane(state, plane, old_plane_state, new_plane_state); @@ -7935,7 +7974,8 @@ dm_determine_update_type_for_commit(struct amdgpu_display_manager *dm, ret = fill_dc_plane_info_and_addr( dm->adev, new_plane_state, tiling_flags, plane_info, - &flip_addr->address); + &flip_addr->address, + false); if (ret) goto cleanup; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c index 3abeff7722e3..e80371542622 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c @@ -316,15 +316,15 @@ static void update_config(void *handle, struct cp_psp_stream_config *config) struct mod_hdcp_display *display = &hdcp_work[link_index].display; struct mod_hdcp_link *link = &hdcp_work[link_index].link; - memset(display, 0, sizeof(*display)); - memset(link, 0, sizeof(*link)); - - display->index = aconnector->base.index; - if (config->dpms_off) { hdcp_remove_display(hdcp_work, link_index, aconnector); return; } + + memset(display, 0, sizeof(*display)); + memset(link, 0, sizeof(*link)); + + display->index = aconnector->base.index; display->state = MOD_HDCP_DISPLAY_ACTIVE; if (aconnector->dc_sink != NULL) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 188e51600070..fff95e6b46c7 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -763,6 +763,29 @@ static bool disable_all_writeback_pipes_for_stream( return true; } +void apply_ctx_interdependent_lock(struct dc *dc, struct dc_state *context, struct dc_stream_state *stream, bool lock) +{ + int i = 0; + + /* Checks if interdependent update function pointer is NULL or not, takes care of DCE110 case */ + if (dc->hwss.interdependent_update_lock) + dc->hwss.interdependent_update_lock(dc, context, lock); + else { + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; + struct pipe_ctx *old_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i]; + + // Copied conditions that were previously in dce110_apply_ctx_for_surface + if (stream == pipe_ctx->stream) { + if (!pipe_ctx->top_pipe && + (pipe_ctx->plane_state || old_pipe_ctx->plane_state)) + dc->hwss.pipe_control_lock(dc, pipe_ctx, lock); + break; + } + } + } +} + static void disable_dangling_plane(struct dc *dc, struct dc_state *context) { int i, j; @@ -788,11 +811,20 @@ static void disable_dangling_plane(struct dc *dc, struct dc_state *context) if (should_disable && old_stream) { dc_rem_all_planes_for_stream(dc, old_stream, dangling_context); disable_all_writeback_pipes_for_stream(dc, old_stream, dangling_context); - if (dc->hwss.apply_ctx_for_surface) + + if (dc->hwss.apply_ctx_for_surface) { + apply_ctx_interdependent_lock(dc, dc->current_state, old_stream, true); dc->hwss.apply_ctx_for_surface(dc, old_stream, 0, dangling_context); + apply_ctx_interdependent_lock(dc, dc->current_state, old_stream, false); + dc->hwss.post_unlock_program_front_end(dc, dangling_context); + } + if (dc->hwss.program_front_end_for_ctx) { + dc->hwss.interdependent_update_lock(dc, dc->current_state, true); + dc->hwss.program_front_end_for_ctx(dc, dangling_context); + dc->hwss.interdependent_update_lock(dc, dc->current_state, false); + dc->hwss.post_unlock_program_front_end(dc, dangling_context); + } } - if (dc->hwss.program_front_end_for_ctx) - dc->hwss.program_front_end_for_ctx(dc, dangling_context); } current_ctx = dc->current_state; @@ -803,11 +835,10 @@ static void disable_dangling_plane(struct dc *dc, struct dc_state *context) static void wait_for_no_pipes_pending(struct dc *dc, struct dc_state *context) { int i; - int count = 0; - struct pipe_ctx *pipe; PERF_TRACE(); for (i = 0; i < MAX_PIPES; i++) { - pipe = &context->res_ctx.pipe_ctx[i]; + int count = 0; + struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; if (!pipe->plane_state) continue; @@ -1212,16 +1243,19 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c /* re-program planes for existing stream, in case we need to * free up plane resource for later use */ - if (dc->hwss.apply_ctx_for_surface) + if (dc->hwss.apply_ctx_for_surface) { for (i = 0; i < context->stream_count; i++) { if (context->streams[i]->mode_changed) continue; - + apply_ctx_interdependent_lock(dc, context, context->streams[i], true); dc->hwss.apply_ctx_for_surface( dc, context->streams[i], context->stream_status[i].plane_count, context); /* use new pipe config in new context */ + apply_ctx_interdependent_lock(dc, context, context->streams[i], false); + dc->hwss.post_unlock_program_front_end(dc, context); } + } /* Program hardware */ for (i = 0; i < dc->res_pool->pipe_count; i++) { @@ -1240,19 +1274,27 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c } /* Program all planes within new context*/ - if (dc->hwss.program_front_end_for_ctx) + if (dc->hwss.program_front_end_for_ctx) { + dc->hwss.interdependent_update_lock(dc, context, true); dc->hwss.program_front_end_for_ctx(dc, context); + dc->hwss.interdependent_update_lock(dc, context, false); + dc->hwss.post_unlock_program_front_end(dc, context); + } for (i = 0; i < context->stream_count; i++) { const struct dc_link *link = context->streams[i]->link; if (!context->streams[i]->mode_changed) continue; - if (dc->hwss.apply_ctx_for_surface) + if (dc->hwss.apply_ctx_for_surface) { + apply_ctx_interdependent_lock(dc, context, context->streams[i], true); dc->hwss.apply_ctx_for_surface( dc, context->streams[i], context->stream_status[i].plane_count, context); + apply_ctx_interdependent_lock(dc, context, context->streams[i], false); + dc->hwss.post_unlock_program_front_end(dc, context); + } /* * enable stereo @@ -1320,10 +1362,24 @@ bool dc_commit_state(struct dc *dc, struct dc_state *context) return (result == DC_OK); } -bool dc_is_hw_initialized(struct dc *dc) +static bool is_flip_pending_in_pipes(struct dc *dc, struct dc_state *context) { - struct dc_bios *dcb = dc->ctx->dc_bios; - return dcb->funcs->is_accelerated_mode(dcb); + int i; + struct pipe_ctx *pipe; + + for (i = 0; i < MAX_PIPES; i++) { + pipe = &context->res_ctx.pipe_ctx[i]; + + if (!pipe->plane_state) + continue; + + /* Must set to false to start with, due to OR in update function */ + pipe->plane_state->status.is_flip_pending = false; + dc->hwss.update_pending_status(pipe); + if (pipe->plane_state->status.is_flip_pending) + return true; + } + return false; } bool dc_post_update_surfaces_to_stream(struct dc *dc) @@ -1336,6 +1392,9 @@ bool dc_post_update_surfaces_to_stream(struct dc *dc) post_surface_trace(dc); + if (is_flip_pending_in_pipes(dc, context)) + return true; + for (i = 0; i < dc->res_pool->pipe_count; i++) if (context->res_ctx.pipe_ctx[i].stream == NULL || context->res_ctx.pipe_ctx[i].plane_state == NULL) { @@ -1736,14 +1795,15 @@ static enum surface_update_type check_update_surfaces_for_stream( if (stream_update->wb_update) su_flags->bits.wb_update = 1; + + if (stream_update->dsc_config) + su_flags->bits.dsc_changed = 1; + if (su_flags->raw != 0) overall_type = UPDATE_TYPE_FULL; if (stream_update->output_csc_transform || stream_update->output_color_space) su_flags->bits.out_csc = 1; - - if (stream_update->dsc_config) - overall_type = UPDATE_TYPE_FULL; } for (i = 0 ; i < surface_count; i++) { @@ -1778,8 +1838,11 @@ enum surface_update_type dc_check_update_surfaces_for_stream( type = check_update_surfaces_for_stream(dc, updates, surface_count, stream_update, stream_status); if (type == UPDATE_TYPE_FULL) { - if (stream_update) + if (stream_update) { + uint32_t dsc_changed = stream_update->stream->update_flags.bits.dsc_changed; stream_update->stream->update_flags.raw = 0xFFFFFFFF; + stream_update->stream->update_flags.bits.dsc_changed = dsc_changed; + } for (i = 0; i < surface_count; i++) updates[i].surface->update_flags.raw = 0xFFFFFFFF; } @@ -2095,18 +2158,14 @@ static void commit_planes_do_stream_update(struct dc *dc, } } - if (stream_update->dsc_config && dc->hwss.pipe_control_lock_global) { - dc->hwss.pipe_control_lock_global(dc, pipe_ctx, true); - dp_update_dsc_config(pipe_ctx); - dc->hwss.pipe_control_lock_global(dc, pipe_ctx, false); - } /* Full fe update*/ if (update_type == UPDATE_TYPE_FAST) continue; - if (stream_update->dpms_off) { - dc->hwss.pipe_control_lock(dc, pipe_ctx, true); + if (stream_update->dsc_config) + dp_update_dsc_config(pipe_ctx); + if (stream_update->dpms_off) { if (*stream_update->dpms_off) { core_link_disable_stream(pipe_ctx); /* for dpms, keep acquired resources*/ @@ -2120,8 +2179,6 @@ static void commit_planes_do_stream_update(struct dc *dc, core_link_enable_stream(dc->current_state, pipe_ctx); } - - dc->hwss.pipe_control_lock(dc, pipe_ctx, false); } if (stream_update->abm_level && pipe_ctx->stream_res.abm) { @@ -2177,6 +2234,27 @@ static void commit_planes_for_stream(struct dc *dc, context_clock_trace(dc, context); } + for (j = 0; j < dc->res_pool->pipe_count; j++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; + + if (!pipe_ctx->top_pipe && + !pipe_ctx->prev_odm_pipe && + pipe_ctx->stream && + pipe_ctx->stream == stream) { + top_pipe_to_program = pipe_ctx; + } + } + + if ((update_type != UPDATE_TYPE_FAST) && dc->hwss.interdependent_update_lock) + dc->hwss.interdependent_update_lock(dc, context, true); + else + /* Lock the top pipe while updating plane addrs, since freesync requires + * plane addr update event triggers to be synchronized. + * top_pipe_to_program is expected to never be NULL + */ + dc->hwss.pipe_control_lock(dc, top_pipe_to_program, true); + + // Stream updates if (stream_update) commit_planes_do_stream_update(dc, stream, stream_update, update_type, context); @@ -2191,6 +2269,12 @@ static void commit_planes_for_stream(struct dc *dc, if (dc->hwss.program_front_end_for_ctx) dc->hwss.program_front_end_for_ctx(dc, context); + if ((update_type != UPDATE_TYPE_FAST) && dc->hwss.interdependent_update_lock) + dc->hwss.interdependent_update_lock(dc, context, false); + else + dc->hwss.pipe_control_lock(dc, top_pipe_to_program, false); + + dc->hwss.post_unlock_program_front_end(dc, context); return; } @@ -2226,8 +2310,6 @@ static void commit_planes_for_stream(struct dc *dc, pipe_ctx->stream == stream) { struct dc_stream_status *stream_status = NULL; - top_pipe_to_program = pipe_ctx; - if (!pipe_ctx->plane_state) continue; @@ -2272,12 +2354,6 @@ static void commit_planes_for_stream(struct dc *dc, // Update Type FAST, Surface updates if (update_type == UPDATE_TYPE_FAST) { - /* Lock the top pipe while updating plane addrs, since freesync requires - * plane addr update event triggers to be synchronized. - * top_pipe_to_program is expected to never be NULL - */ - dc->hwss.pipe_control_lock(dc, top_pipe_to_program, true); - if (dc->hwss.set_flip_control_gsl) for (i = 0; i < surface_count; i++) { struct dc_plane_state *plane_state = srf_updates[i].surface; @@ -2319,9 +2395,15 @@ static void commit_planes_for_stream(struct dc *dc, dc->hwss.update_plane_addr(dc, pipe_ctx); } } + } + if ((update_type != UPDATE_TYPE_FAST) && dc->hwss.interdependent_update_lock) + dc->hwss.interdependent_update_lock(dc, context, false); + else dc->hwss.pipe_control_lock(dc, top_pipe_to_program, false); - } + + if (update_type != UPDATE_TYPE_FAST) + dc->hwss.post_unlock_program_front_end(dc, context); // Fire manual trigger only when bottom plane is flipped for (j = 0; j < dc->res_pool->pipe_count; j++) { diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c index fd9e69634c50..fbcd979438e2 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c @@ -220,6 +220,30 @@ static enum dpcd_training_patterns return dpcd_tr_pattern; } +static uint8_t dc_dp_initialize_scrambling_data_symbols( + struct dc_link *link, + enum dc_dp_training_pattern pattern) +{ + uint8_t disable_scrabled_data_symbols = 0; + + switch (pattern) { + case DP_TRAINING_PATTERN_SEQUENCE_1: + case DP_TRAINING_PATTERN_SEQUENCE_2: + case DP_TRAINING_PATTERN_SEQUENCE_3: + disable_scrabled_data_symbols = 1; + break; + case DP_TRAINING_PATTERN_SEQUENCE_4: + disable_scrabled_data_symbols = 0; + break; + default: + ASSERT(0); + DC_LOG_HW_LINK_TRAINING("%s: Invalid HW Training pattern: %d\n", + __func__, pattern); + break; + } + return disable_scrabled_data_symbols; +} + static inline bool is_repeater(struct dc_link *link, uint32_t offset) { return (!link->is_lttpr_mode_transparent && offset != 0); @@ -252,6 +276,9 @@ static void dpcd_set_lt_pattern_and_lane_settings( dpcd_pattern.v1_4.TRAINING_PATTERN_SET = dc_dp_training_pattern_to_dpcd_training_pattern(link, pattern); + dpcd_pattern.v1_4.SCRAMBLING_DISABLE = + dc_dp_initialize_scrambling_data_symbols(link, pattern); + dpcd_lt_buffer[DP_TRAINING_PATTERN_SET - DP_TRAINING_PATTERN_SET] = dpcd_pattern.raw; @@ -2888,6 +2915,12 @@ bool dc_link_handle_hpd_rx_irq(struct dc_link *link, union hpd_irq_data *out_hpd for (i = 0; i < MAX_PIPES; i++) { pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i]; if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link) + link->dc->hwss.blank_stream(pipe_ctx); + } + + for (i = 0; i < MAX_PIPES; i++) { + pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i]; + if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link) break; } @@ -2904,6 +2937,12 @@ bool dc_link_handle_hpd_rx_irq(struct dc_link *link, union hpd_irq_data *out_hpd if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST) dc_link_reallocate_mst_payload(link); + for (i = 0; i < MAX_PIPES; i++) { + pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i]; + if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link) + link->dc->hwss.unblank_stream(pipe_ctx, &previous_link_settings); + } + status = false; if (out_link_loss) *out_link_loss = true; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c index 6ddbb00ed37a..4f0e7203dba4 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c @@ -231,34 +231,6 @@ struct dc_stream_status *dc_stream_get_status( return dc_stream_get_status_from_state(dc->current_state, stream); } -static void delay_cursor_until_vupdate(struct pipe_ctx *pipe_ctx, struct dc *dc) -{ -#if defined(CONFIG_DRM_AMD_DC_DCN) - unsigned int vupdate_line; - unsigned int lines_to_vupdate, us_to_vupdate, vpos, nvpos; - struct dc_stream_state *stream = pipe_ctx->stream; - unsigned int us_per_line; - - if (stream->ctx->asic_id.chip_family == FAMILY_RV && - ASICREV_IS_RAVEN(stream->ctx->asic_id.hw_internal_rev)) { - - vupdate_line = dc->hwss.get_vupdate_offset_from_vsync(pipe_ctx); - if (!dc_stream_get_crtc_position(dc, &stream, 1, &vpos, &nvpos)) - return; - - if (vpos >= vupdate_line) - return; - - us_per_line = stream->timing.h_total * 10000 / stream->timing.pix_clk_100hz; - lines_to_vupdate = vupdate_line - vpos; - us_to_vupdate = lines_to_vupdate * us_per_line; - - /* 70 us is a conservative estimate of cursor update time*/ - if (us_to_vupdate < 70) - udelay(us_to_vupdate); - } -#endif -} /** * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address @@ -298,9 +270,7 @@ bool dc_stream_set_cursor_attributes( if (!pipe_to_program) { pipe_to_program = pipe_ctx; - - delay_cursor_until_vupdate(pipe_ctx, dc); - dc->hwss.pipe_control_lock(dc, pipe_to_program, true); + dc->hwss.cursor_lock(dc, pipe_to_program, true); } dc->hwss.set_cursor_attribute(pipe_ctx); @@ -309,7 +279,7 @@ bool dc_stream_set_cursor_attributes( } if (pipe_to_program) - dc->hwss.pipe_control_lock(dc, pipe_to_program, false); + dc->hwss.cursor_lock(dc, pipe_to_program, false); return true; } @@ -349,16 +319,14 @@ bool dc_stream_set_cursor_position( if (!pipe_to_program) { pipe_to_program = pipe_ctx; - - delay_cursor_until_vupdate(pipe_ctx, dc); - dc->hwss.pipe_control_lock(dc, pipe_to_program, true); + dc->hwss.cursor_lock(dc, pipe_to_program, true); } dc->hwss.set_cursor_position(pipe_ctx); } if (pipe_to_program) - dc->hwss.pipe_control_lock(dc, pipe_to_program, false); + dc->hwss.cursor_lock(dc, pipe_to_program, false); return true; } diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 8ff25b5dd2f6..e8d126890d7e 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -1075,7 +1075,6 @@ unsigned int dc_get_current_backlight_pwm(struct dc *dc); unsigned int dc_get_target_backlight_pwm(struct dc *dc); bool dc_is_dmcu_initialized(struct dc *dc); -bool dc_is_hw_initialized(struct dc *dc); enum dc_status dc_set_clock(struct dc *dc, enum dc_clock_type clock_type, uint32_t clk_khz, uint32_t stepping); void dc_get_clock(struct dc *dc, enum dc_clock_type clock_type, struct dc_clock_config *clock_cfg); diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index 92096de79dec..a5c7ef47b8d3 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -118,6 +118,7 @@ union stream_update_flags { uint32_t dpms_off:1; uint32_t gamut_remap:1; uint32_t wb_update:1; + uint32_t dsc_changed : 1; } bits; uint32_t raw; diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c index 5b689273ff44..454a123b92fc 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c @@ -2576,17 +2576,6 @@ static void dce110_apply_ctx_for_surface( for (i = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - struct pipe_ctx *old_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i]; - - if (stream == pipe_ctx->stream) { - if (!pipe_ctx->top_pipe && - (pipe_ctx->plane_state || old_pipe_ctx->plane_state)) - dc->hwss.pipe_control_lock(dc, pipe_ctx, true); - } - } - - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; if (pipe_ctx->stream != stream) continue; @@ -2607,20 +2596,16 @@ static void dce110_apply_ctx_for_surface( } - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - struct pipe_ctx *old_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i]; - - if ((stream == pipe_ctx->stream) && - (!pipe_ctx->top_pipe) && - (pipe_ctx->plane_state || old_pipe_ctx->plane_state)) - dc->hwss.pipe_control_lock(dc, pipe_ctx, false); - } - if (dc->fbc_compressor) enable_fbc(dc, context); } +static void dce110_post_unlock_program_front_end( + struct dc *dc, + struct dc_state *context) +{ +} + static void dce110_power_down_fe(struct dc *dc, struct pipe_ctx *pipe_ctx) { struct dce_hwseq *hws = dc->hwseq; @@ -2722,6 +2707,7 @@ static const struct hw_sequencer_funcs dce110_funcs = { .init_hw = init_hw, .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = dce110_apply_ctx_for_surface, + .post_unlock_program_front_end = dce110_post_unlock_program_front_end, .update_plane_addr = update_plane_addr, .update_pending_status = dce110_update_pending_status, .enable_accelerated_mode = dce110_enable_accelerated_mode, @@ -2736,6 +2722,8 @@ static const struct hw_sequencer_funcs dce110_funcs = { .disable_audio_stream = dce110_disable_audio_stream, .disable_plane = dce110_power_down_fe, .pipe_control_lock = dce_pipe_control_lock, + .interdependent_update_lock = NULL, + .cursor_lock = dce_pipe_control_lock, .prepare_bandwidth = dce110_prepare_bandwidth, .optimize_bandwidth = dce110_optimize_bandwidth, .set_drr = set_drr, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c index 1008ac8a0f2a..0c987b5d68e2 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c @@ -82,7 +82,7 @@ void print_microsec(struct dc_context *dc_ctx, us_x10 % frac); } -static void dcn10_lock_all_pipes(struct dc *dc, +void dcn10_lock_all_pipes(struct dc *dc, struct dc_state *context, bool lock) { @@ -93,6 +93,7 @@ static void dcn10_lock_all_pipes(struct dc *dc, for (i = 0; i < dc->res_pool->pipe_count; i++) { pipe_ctx = &context->res_ctx.pipe_ctx[i]; tg = pipe_ctx->stream_res.tg; + /* * Only lock the top pipe's tg to prevent redundant * (un)locking. Also skip if pipe is disabled. @@ -103,9 +104,9 @@ static void dcn10_lock_all_pipes(struct dc *dc, continue; if (lock) - tg->funcs->lock(tg); + dc->hwss.pipe_control_lock(dc, pipe_ctx, true); else - tg->funcs->unlock(tg); + dc->hwss.pipe_control_lock(dc, pipe_ctx, false); } } @@ -1576,7 +1577,7 @@ void dcn10_pipe_control_lock( /* use TG master update lock to lock everything on the TG * therefore only top pipe need to lock */ - if (pipe->top_pipe) + if (!pipe || pipe->top_pipe) return; if (dc->debug.sanity_checks) @@ -1591,6 +1592,85 @@ void dcn10_pipe_control_lock( hws->funcs.verify_allow_pstate_change_high(dc); } +/** + * delay_cursor_until_vupdate() - Delay cursor update if too close to VUPDATE. + * + * Software keepout workaround to prevent cursor update locking from stalling + * out cursor updates indefinitely or from old values from being retained in + * the case where the viewport changes in the same frame as the cursor. + * + * The idea is to calculate the remaining time from VPOS to VUPDATE. If it's + * too close to VUPDATE, then stall out until VUPDATE finishes. + * + * TODO: Optimize cursor programming to be once per frame before VUPDATE + * to avoid the need for this workaround. + */ +static void delay_cursor_until_vupdate(struct dc *dc, struct pipe_ctx *pipe_ctx) +{ + struct dc_stream_state *stream = pipe_ctx->stream; + struct crtc_position position; + uint32_t vupdate_start, vupdate_end; + unsigned int lines_to_vupdate, us_to_vupdate, vpos; + unsigned int us_per_line, us_vupdate; + + if (!dc->hwss.calc_vupdate_position || !dc->hwss.get_position) + return; + + if (!pipe_ctx->stream_res.stream_enc || !pipe_ctx->stream_res.tg) + return; + + dc->hwss.calc_vupdate_position(dc, pipe_ctx, &vupdate_start, + &vupdate_end); + + dc->hwss.get_position(&pipe_ctx, 1, &position); + vpos = position.vertical_count; + + /* Avoid wraparound calculation issues */ + vupdate_start += stream->timing.v_total; + vupdate_end += stream->timing.v_total; + vpos += stream->timing.v_total; + + if (vpos <= vupdate_start) { + /* VPOS is in VACTIVE or back porch. */ + lines_to_vupdate = vupdate_start - vpos; + } else if (vpos > vupdate_end) { + /* VPOS is in the front porch. */ + return; + } else { + /* VPOS is in VUPDATE. */ + lines_to_vupdate = 0; + } + + /* Calculate time until VUPDATE in microseconds. */ + us_per_line = + stream->timing.h_total * 10000u / stream->timing.pix_clk_100hz; + us_to_vupdate = lines_to_vupdate * us_per_line; + + /* 70 us is a conservative estimate of cursor update time*/ + if (us_to_vupdate > 70) + return; + + /* Stall out until the cursor update completes. */ + if (vupdate_end < vupdate_start) + vupdate_end += stream->timing.v_total; + us_vupdate = (vupdate_end - vupdate_start + 1) * us_per_line; + udelay(us_to_vupdate + us_vupdate); +} + +void dcn10_cursor_lock(struct dc *dc, struct pipe_ctx *pipe, bool lock) +{ + /* cursor lock is per MPCC tree, so only need to lock one pipe per stream */ + if (!pipe || pipe->top_pipe) + return; + + /* Prevent cursor lock from stalling out cursor updates. */ + if (lock) + delay_cursor_until_vupdate(dc, pipe); + + dc->res_pool->mpc->funcs->cursor_lock(dc->res_pool->mpc, + pipe->stream_res.opp->inst, lock); +} + static bool wait_for_reset_trigger_to_occur( struct dc_context *dc_ctx, struct timing_generator *tg) @@ -2512,7 +2592,6 @@ void dcn10_apply_ctx_for_surface( int i; struct timing_generator *tg; uint32_t underflow_check_delay_us; - bool removed_pipe[4] = { false }; bool interdependent_update = false; struct pipe_ctx *top_pipe_to_program = dcn10_find_top_pipe_for_stream(dc, context, stream); @@ -2531,11 +2610,6 @@ void dcn10_apply_ctx_for_surface( if (underflow_check_delay_us != 0xFFFFFFFF && hws->funcs.did_underflow_occur) ASSERT(hws->funcs.did_underflow_occur(dc, top_pipe_to_program)); - if (interdependent_update) - dcn10_lock_all_pipes(dc, context, true); - else - dcn10_pipe_control_lock(dc, top_pipe_to_program, true); - if (underflow_check_delay_us != 0xFFFFFFFF) udelay(underflow_check_delay_us); @@ -2552,18 +2626,8 @@ void dcn10_apply_ctx_for_surface( struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; struct pipe_ctx *old_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i]; - /* - * Powergate reused pipes that are not powergated - * fairly hacky right now, using opp_id as indicator - * TODO: After move dc_post to dc_update, this will - * be removed. - */ - if (pipe_ctx->plane_state && !old_pipe_ctx->plane_state) { - if (old_pipe_ctx->stream_res.tg == tg && - old_pipe_ctx->plane_res.hubp && - old_pipe_ctx->plane_res.hubp->opp_id != OPP_ID_INVALID) - dc->hwss.disable_plane(dc, old_pipe_ctx); - } + + pipe_ctx->update_flags.raw = 0; if ((!pipe_ctx->plane_state || pipe_ctx->stream_res.tg != old_pipe_ctx->stream_res.tg) && @@ -2571,7 +2635,7 @@ void dcn10_apply_ctx_for_surface( old_pipe_ctx->stream_res.tg == tg) { hws->funcs.plane_atomic_disconnect(dc, old_pipe_ctx); - removed_pipe[i] = true; + pipe_ctx->update_flags.bits.disable = 1; DC_LOG_DC("Reset mpcc for pipe %d\n", old_pipe_ctx->pipe_idx); @@ -2597,21 +2661,41 @@ void dcn10_apply_ctx_for_surface( &pipe_ctx->dlg_regs, &pipe_ctx->ttu_regs); } +} - if (interdependent_update) - dcn10_lock_all_pipes(dc, context, false); - else - dcn10_pipe_control_lock(dc, top_pipe_to_program, false); +void dcn10_post_unlock_program_front_end( + struct dc *dc, + struct dc_state *context) +{ + int i, j; - if (num_planes == 0) - false_optc_underflow_wa(dc, stream, tg); + DC_LOGGER_INIT(dc->ctx->logger); + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; + + if (!pipe_ctx->top_pipe && + !pipe_ctx->prev_odm_pipe && + pipe_ctx->stream) { + struct dc_stream_status *stream_status = NULL; + struct timing_generator *tg = pipe_ctx->stream_res.tg; + + for (j = 0; j < context->stream_count; j++) { + if (pipe_ctx->stream == context->streams[j]) + stream_status = &context->stream_status[j]; + } + + if (context->stream_status[i].plane_count == 0) + false_optc_underflow_wa(dc, pipe_ctx->stream, tg); + } + } for (i = 0; i < dc->res_pool->pipe_count; i++) - if (removed_pipe[i]) + if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable) dc->hwss.disable_plane(dc, &dc->current_state->res_ctx.pipe_ctx[i]); for (i = 0; i < dc->res_pool->pipe_count; i++) - if (removed_pipe[i]) { + if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable) { dc->hwss.optimize_bandwidth(dc, context); break; } @@ -3127,7 +3211,7 @@ int dcn10_get_vupdate_offset_from_vsync(struct pipe_ctx *pipe_ctx) return vertical_line_start; } -static void dcn10_calc_vupdate_position( +void dcn10_calc_vupdate_position( struct dc *dc, struct pipe_ctx *pipe_ctx, uint32_t *start_line, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h index 4d20f6586bb5..42b6e016d71e 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.h @@ -34,6 +34,11 @@ struct dc; void dcn10_hw_sequencer_construct(struct dc *dc); int dcn10_get_vupdate_offset_from_vsync(struct pipe_ctx *pipe_ctx); +void dcn10_calc_vupdate_position( + struct dc *dc, + struct pipe_ctx *pipe_ctx, + uint32_t *start_line, + uint32_t *end_line); void dcn10_setup_vupdate_interrupt(struct dc *dc, struct pipe_ctx *pipe_ctx); enum dc_status dcn10_enable_stream_timing( struct pipe_ctx *pipe_ctx, @@ -49,6 +54,7 @@ void dcn10_pipe_control_lock( struct dc *dc, struct pipe_ctx *pipe, bool lock); +void dcn10_cursor_lock(struct dc *dc, struct pipe_ctx *pipe, bool lock); void dcn10_blank_pixel_data( struct dc *dc, struct pipe_ctx *pipe_ctx, @@ -70,11 +76,18 @@ void dcn10_reset_hw_ctx_wrap( struct dc *dc, struct dc_state *context); void dcn10_disable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx); +void dcn10_lock_all_pipes( + struct dc *dc, + struct dc_state *context, + bool lock); void dcn10_apply_ctx_for_surface( struct dc *dc, const struct dc_stream_state *stream, int num_planes, struct dc_state *context); +void dcn10_post_unlock_program_front_end( + struct dc *dc, + struct dc_state *context); void dcn10_hubp_pg_control( struct dce_hwseq *hws, unsigned int hubp_inst, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c index e7e5352ec424..0900c861204f 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_init.c @@ -32,6 +32,7 @@ static const struct hw_sequencer_funcs dcn10_funcs = { .init_hw = dcn10_init_hw, .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = dcn10_apply_ctx_for_surface, + .post_unlock_program_front_end = dcn10_post_unlock_program_front_end, .update_plane_addr = dcn10_update_plane_addr, .update_dchub = dcn10_update_dchub, .update_pending_status = dcn10_update_pending_status, @@ -49,6 +50,8 @@ static const struct hw_sequencer_funcs dcn10_funcs = { .disable_audio_stream = dce110_disable_audio_stream, .disable_plane = dcn10_disable_plane, .pipe_control_lock = dcn10_pipe_control_lock, + .cursor_lock = dcn10_cursor_lock, + .interdependent_update_lock = dcn10_lock_all_pipes, .prepare_bandwidth = dcn10_prepare_bandwidth, .optimize_bandwidth = dcn10_optimize_bandwidth, .set_drr = dcn10_set_drr, @@ -69,6 +72,7 @@ static const struct hw_sequencer_funcs dcn10_funcs = { .set_clock = dcn10_set_clock, .get_clock = dcn10_get_clock, .get_vupdate_offset_from_vsync = dcn10_get_vupdate_offset_from_vsync, + .calc_vupdate_position = dcn10_calc_vupdate_position, }; static const struct hwseq_private_funcs dcn10_private_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c index 04f863499cfb..3fcd408e9103 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c @@ -223,6 +223,9 @@ struct mpcc *mpc1_insert_plane( REG_SET(MPCC_TOP_SEL[mpcc_id], 0, MPCC_TOP_SEL, dpp_id); REG_SET(MPCC_OPP_ID[mpcc_id], 0, MPCC_OPP_ID, tree->opp_id); + /* Configure VUPDATE lock set for this MPCC to map to the OPP */ + REG_SET(MPCC_UPDATE_LOCK_SEL[mpcc_id], 0, MPCC_UPDATE_LOCK_SEL, tree->opp_id); + /* update mpc tree mux setting */ if (tree->opp_list == insert_above_mpcc) { /* insert the toppest mpcc */ @@ -318,6 +321,7 @@ void mpc1_remove_mpcc( REG_SET(MPCC_TOP_SEL[mpcc_id], 0, MPCC_TOP_SEL, 0xf); REG_SET(MPCC_BOT_SEL[mpcc_id], 0, MPCC_BOT_SEL, 0xf); REG_SET(MPCC_OPP_ID[mpcc_id], 0, MPCC_OPP_ID, 0xf); + REG_SET(MPCC_UPDATE_LOCK_SEL[mpcc_id], 0, MPCC_UPDATE_LOCK_SEL, 0xf); /* mark this mpcc as not in use */ mpc10->mpcc_in_use_mask &= ~(1 << mpcc_id); @@ -328,6 +332,7 @@ void mpc1_remove_mpcc( REG_SET(MPCC_TOP_SEL[mpcc_id], 0, MPCC_TOP_SEL, 0xf); REG_SET(MPCC_BOT_SEL[mpcc_id], 0, MPCC_BOT_SEL, 0xf); REG_SET(MPCC_OPP_ID[mpcc_id], 0, MPCC_OPP_ID, 0xf); + REG_SET(MPCC_UPDATE_LOCK_SEL[mpcc_id], 0, MPCC_UPDATE_LOCK_SEL, 0xf); } } @@ -361,6 +366,7 @@ void mpc1_mpc_init(struct mpc *mpc) REG_SET(MPCC_TOP_SEL[mpcc_id], 0, MPCC_TOP_SEL, 0xf); REG_SET(MPCC_BOT_SEL[mpcc_id], 0, MPCC_BOT_SEL, 0xf); REG_SET(MPCC_OPP_ID[mpcc_id], 0, MPCC_OPP_ID, 0xf); + REG_SET(MPCC_UPDATE_LOCK_SEL[mpcc_id], 0, MPCC_UPDATE_LOCK_SEL, 0xf); mpc1_init_mpcc(&(mpc->mpcc_array[mpcc_id]), mpcc_id); } @@ -381,6 +387,7 @@ void mpc1_mpc_init_single_inst(struct mpc *mpc, unsigned int mpcc_id) REG_SET(MPCC_TOP_SEL[mpcc_id], 0, MPCC_TOP_SEL, 0xf); REG_SET(MPCC_BOT_SEL[mpcc_id], 0, MPCC_BOT_SEL, 0xf); REG_SET(MPCC_OPP_ID[mpcc_id], 0, MPCC_OPP_ID, 0xf); + REG_SET(MPCC_UPDATE_LOCK_SEL[mpcc_id], 0, MPCC_UPDATE_LOCK_SEL, 0xf); mpc1_init_mpcc(&(mpc->mpcc_array[mpcc_id]), mpcc_id); @@ -453,6 +460,13 @@ void mpc1_read_mpcc_state( MPCC_BUSY, &s->busy); } +void mpc1_cursor_lock(struct mpc *mpc, int opp_id, bool lock) +{ + struct dcn10_mpc *mpc10 = TO_DCN10_MPC(mpc); + + REG_SET(CUR[opp_id], 0, CUR_VUPDATE_LOCK_SET, lock ? 1 : 0); +} + static const struct mpc_funcs dcn10_mpc_funcs = { .read_mpcc_state = mpc1_read_mpcc_state, .insert_plane = mpc1_insert_plane, @@ -464,6 +478,7 @@ static const struct mpc_funcs dcn10_mpc_funcs = { .assert_mpcc_idle_before_connect = mpc1_assert_mpcc_idle_before_connect, .init_mpcc_list_from_hw = mpc1_init_mpcc_list_from_hw, .update_blending = mpc1_update_blending, + .cursor_lock = mpc1_cursor_lock, .set_denorm = NULL, .set_denorm_clamp = NULL, .set_output_csc = NULL, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h index 962a68e322ee..66a4719c22a0 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h @@ -39,11 +39,12 @@ SRII(MPCC_BG_G_Y, MPCC, inst),\ SRII(MPCC_BG_R_CR, MPCC, inst),\ SRII(MPCC_BG_B_CB, MPCC, inst),\ - SRII(MPCC_BG_B_CB, MPCC, inst),\ - SRII(MPCC_SM_CONTROL, MPCC, inst) + SRII(MPCC_SM_CONTROL, MPCC, inst),\ + SRII(MPCC_UPDATE_LOCK_SEL, MPCC, inst) #define MPC_OUT_MUX_COMMON_REG_LIST_DCN1_0(inst) \ - SRII(MUX, MPC_OUT, inst) + SRII(MUX, MPC_OUT, inst),\ + VUPDATE_SRII(CUR, VUPDATE_LOCK_SET, inst) #define MPC_COMMON_REG_VARIABLE_LIST \ uint32_t MPCC_TOP_SEL[MAX_MPCC]; \ @@ -55,7 +56,9 @@ uint32_t MPCC_BG_R_CR[MAX_MPCC]; \ uint32_t MPCC_BG_B_CB[MAX_MPCC]; \ uint32_t MPCC_SM_CONTROL[MAX_MPCC]; \ - uint32_t MUX[MAX_OPP]; + uint32_t MUX[MAX_OPP]; \ + uint32_t MPCC_UPDATE_LOCK_SEL[MAX_MPCC]; \ + uint32_t CUR[MAX_OPP]; #define MPC_COMMON_MASK_SH_LIST_DCN1_0(mask_sh)\ SF(MPCC0_MPCC_TOP_SEL, MPCC_TOP_SEL, mask_sh),\ @@ -78,7 +81,8 @@ SF(MPCC0_MPCC_SM_CONTROL, MPCC_SM_FIELD_ALT, mask_sh),\ SF(MPCC0_MPCC_SM_CONTROL, MPCC_SM_FORCE_NEXT_FRAME_POL, mask_sh),\ SF(MPCC0_MPCC_SM_CONTROL, MPCC_SM_FORCE_NEXT_TOP_POL, mask_sh),\ - SF(MPC_OUT0_MUX, MPC_OUT_MUX, mask_sh) + SF(MPC_OUT0_MUX, MPC_OUT_MUX, mask_sh),\ + SF(MPCC0_MPCC_UPDATE_LOCK_SEL, MPCC_UPDATE_LOCK_SEL, mask_sh) #define MPC_REG_FIELD_LIST(type) \ type MPCC_TOP_SEL;\ @@ -101,7 +105,9 @@ type MPCC_SM_FIELD_ALT;\ type MPCC_SM_FORCE_NEXT_FRAME_POL;\ type MPCC_SM_FORCE_NEXT_TOP_POL;\ - type MPC_OUT_MUX; + type MPC_OUT_MUX;\ + type MPCC_UPDATE_LOCK_SEL;\ + type CUR_VUPDATE_LOCK_SET; struct dcn_mpc_registers { MPC_COMMON_REG_VARIABLE_LIST @@ -192,4 +198,6 @@ void mpc1_read_mpcc_state( int mpcc_inst, struct mpcc_state *s); +void mpc1_cursor_lock(struct mpc *mpc, int opp_id, bool lock); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c index 3b71898e859e..e3c4c06ac191 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c @@ -181,6 +181,14 @@ enum dcn10_clk_src_array_id { .reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \ mm ## block ## id ## _ ## reg_name +#define VUPDATE_SRII(reg_name, block, id)\ + .reg_name[id] = BASE(mm ## reg_name ## 0 ## _ ## block ## id ## _BASE_IDX) + \ + mm ## reg_name ## 0 ## _ ## block ## id + +/* set field/register/bitfield name */ +#define SFRB(field_name, reg_name, bitfield, post_fix)\ + .field_name = reg_name ## __ ## bitfield ## post_fix + /* NBIO */ #define NBIO_BASE_INNER(seg) \ NBIF_BASE__INST0_SEG ## seg @@ -419,11 +427,13 @@ static const struct dcn_mpc_registers mpc_regs = { }; static const struct dcn_mpc_shift mpc_shift = { - MPC_COMMON_MASK_SH_LIST_DCN1_0(__SHIFT) + MPC_COMMON_MASK_SH_LIST_DCN1_0(__SHIFT),\ + SFRB(CUR_VUPDATE_LOCK_SET, CUR0_VUPDATE_LOCK_SET0, CUR0_VUPDATE_LOCK_SET, __SHIFT) }; static const struct dcn_mpc_mask mpc_mask = { - MPC_COMMON_MASK_SH_LIST_DCN1_0(_MASK), + MPC_COMMON_MASK_SH_LIST_DCN1_0(_MASK),\ + SFRB(CUR_VUPDATE_LOCK_SET, CUR0_VUPDATE_LOCK_SET0, CUR0_VUPDATE_LOCK_SET, _MASK) }; #define tg_regs(id)\ diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c index a444fed94184..611dac544bfe 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c @@ -1088,40 +1088,18 @@ void dcn20_enable_plane( // } } - -void dcn20_pipe_control_lock_global( - struct dc *dc, - struct pipe_ctx *pipe, - bool lock) -{ - if (lock) { - pipe->stream_res.tg->funcs->lock_doublebuffer_enable( - pipe->stream_res.tg); - pipe->stream_res.tg->funcs->lock(pipe->stream_res.tg); - } else { - pipe->stream_res.tg->funcs->unlock(pipe->stream_res.tg); - pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, - CRTC_STATE_VACTIVE); - pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, - CRTC_STATE_VBLANK); - pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, - CRTC_STATE_VACTIVE); - pipe->stream_res.tg->funcs->lock_doublebuffer_disable( - pipe->stream_res.tg); - } -} - void dcn20_pipe_control_lock( struct dc *dc, struct pipe_ctx *pipe, bool lock) { bool flip_immediate = false; + bool dig_update_required = false; /* use TG master update lock to lock everything on the TG * therefore only top pipe need to lock */ - if (pipe->top_pipe) + if (!pipe || pipe->top_pipe) return; if (pipe->plane_state != NULL) @@ -1154,6 +1132,19 @@ void dcn20_pipe_control_lock( (!flip_immediate && pipe->stream_res.gsl_group > 0)) dcn20_setup_gsl_group_as_lock(dc, pipe, flip_immediate); + if (pipe->stream && pipe->stream->update_flags.bits.dsc_changed) + dig_update_required = true; + + /* Need double buffer lock mode in order to synchronize front end pipe + * updates with dig updates. + */ + if (dig_update_required) { + if (lock) { + pipe->stream_res.tg->funcs->lock_doublebuffer_enable( + pipe->stream_res.tg); + } + } + if (pipe->plane_state != NULL && pipe->plane_state->triplebuffer_flips) { if (lock) pipe->stream_res.tg->funcs->triplebuffer_lock(pipe->stream_res.tg); @@ -1165,6 +1156,19 @@ void dcn20_pipe_control_lock( else pipe->stream_res.tg->funcs->unlock(pipe->stream_res.tg); } + + if (dig_update_required) { + if (!lock) { + pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, + CRTC_STATE_VACTIVE); + pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, + CRTC_STATE_VBLANK); + pipe->stream_res.tg->funcs->wait_for_state(pipe->stream_res.tg, + CRTC_STATE_VACTIVE); + pipe->stream_res.tg->funcs->lock_doublebuffer_disable( + pipe->stream_res.tg); + } + } } static void dcn20_detect_pipe_changes(struct pipe_ctx *old_pipe, struct pipe_ctx *new_pipe) @@ -1536,27 +1540,28 @@ static void dcn20_program_pipe( } } -static bool does_pipe_need_lock(struct pipe_ctx *pipe) -{ - if ((pipe->plane_state && pipe->plane_state->update_flags.raw) - || pipe->update_flags.raw) - return true; - if (pipe->bottom_pipe) - return does_pipe_need_lock(pipe->bottom_pipe); - - return false; -} - void dcn20_program_front_end_for_ctx( struct dc *dc, struct dc_state *context) { - const unsigned int TIMEOUT_FOR_PIPE_ENABLE_MS = 100; int i; struct dce_hwseq *hws = dc->hwseq; - bool pipe_locked[MAX_PIPES] = {false}; DC_LOGGER_INIT(dc->ctx->logger); + for (i = 0; i < dc->res_pool->pipe_count; i++) { + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; + + if (!pipe_ctx->top_pipe && !pipe_ctx->prev_odm_pipe && pipe_ctx->plane_state) { + ASSERT(!pipe_ctx->plane_state->triplebuffer_flips); + if (dc->hwss.program_triplebuffer != NULL && + !dc->debug.disable_tri_buf) { + /*turn off triple buffer for full update*/ + dc->hwss.program_triplebuffer( + dc, pipe_ctx, pipe_ctx->plane_state->triplebuffer_flips); + } + } + } + /* Carry over GSL groups in case the context is changing. */ for (i = 0; i < dc->res_pool->pipe_count; i++) if (context->res_ctx.pipe_ctx[i].stream == dc->current_state->res_ctx.pipe_ctx[i].stream) @@ -1567,17 +1572,6 @@ void dcn20_program_front_end_for_ctx( for (i = 0; i < dc->res_pool->pipe_count; i++) dcn20_detect_pipe_changes(&dc->current_state->res_ctx.pipe_ctx[i], &context->res_ctx.pipe_ctx[i]); - for (i = 0; i < dc->res_pool->pipe_count; i++) - if (!context->res_ctx.pipe_ctx[i].top_pipe && - does_pipe_need_lock(&context->res_ctx.pipe_ctx[i])) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - - if (pipe_ctx->update_flags.bits.tg_changed || pipe_ctx->update_flags.bits.enable) - dc->hwss.pipe_control_lock(dc, pipe_ctx, true); - if (!pipe_ctx->update_flags.bits.enable) - dc->hwss.pipe_control_lock(dc, &dc->current_state->res_ctx.pipe_ctx[i], true); - pipe_locked[i] = true; - } /* OTG blank before disabling all front ends */ for (i = 0; i < dc->res_pool->pipe_count; i++) @@ -1615,17 +1609,16 @@ void dcn20_program_front_end_for_ctx( hws->funcs.program_all_writeback_pipes_in_tree(dc, pipe->stream, context); } } +} - /* Unlock all locked pipes */ - for (i = 0; i < dc->res_pool->pipe_count; i++) - if (pipe_locked[i]) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; +void dcn20_post_unlock_program_front_end( + struct dc *dc, + struct dc_state *context) +{ + int i; + const unsigned int TIMEOUT_FOR_PIPE_ENABLE_MS = 100; - if (pipe_ctx->update_flags.bits.tg_changed || pipe_ctx->update_flags.bits.enable) - dc->hwss.pipe_control_lock(dc, pipe_ctx, false); - if (!pipe_ctx->update_flags.bits.enable) - dc->hwss.pipe_control_lock(dc, &dc->current_state->res_ctx.pipe_ctx[i], false); - } + DC_LOGGER_INIT(dc->ctx->logger); for (i = 0; i < dc->res_pool->pipe_count; i++) if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable) @@ -1655,7 +1648,6 @@ void dcn20_program_front_end_for_ctx( dc->res_pool->hubbub->funcs->apply_DEDCN21_147_wa(dc->res_pool->hubbub); } - void dcn20_prepare_bandwidth( struct dc *dc, struct dc_state *context) @@ -2306,7 +2298,8 @@ void dcn20_fpga_init_hw(struct dc *dc) REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_REFDIV, 2); REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); - REG_WRITE(REFCLK_CNTL, 0); + if (REG(REFCLK_CNTL)) + REG_WRITE(REFCLK_CNTL, 0); // diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h index 02c9be5ebd47..63ce763f148e 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h @@ -35,6 +35,9 @@ bool dcn20_set_shaper_3dlut( void dcn20_program_front_end_for_ctx( struct dc *dc, struct dc_state *context); +void dcn20_post_unlock_program_front_end( + struct dc *dc, + struct dc_state *context); void dcn20_update_plane_addr(const struct dc *dc, struct pipe_ctx *pipe_ctx); void dcn20_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx); bool dcn20_set_input_transfer_func(struct dc *dc, struct pipe_ctx *pipe_ctx, @@ -58,10 +61,6 @@ void dcn20_pipe_control_lock( struct dc *dc, struct pipe_ctx *pipe, bool lock); -void dcn20_pipe_control_lock_global( - struct dc *dc, - struct pipe_ctx *pipe, - bool lock); void dcn20_prepare_bandwidth( struct dc *dc, struct dc_state *context); diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c index 5e640f17d3d4..71bfde2cf646 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c @@ -33,6 +33,7 @@ static const struct hw_sequencer_funcs dcn20_funcs = { .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = NULL, .program_front_end_for_ctx = dcn20_program_front_end_for_ctx, + .post_unlock_program_front_end = dcn20_post_unlock_program_front_end, .update_plane_addr = dcn20_update_plane_addr, .update_dchub = dcn10_update_dchub, .update_pending_status = dcn10_update_pending_status, @@ -50,7 +51,8 @@ static const struct hw_sequencer_funcs dcn20_funcs = { .disable_audio_stream = dce110_disable_audio_stream, .disable_plane = dcn20_disable_plane, .pipe_control_lock = dcn20_pipe_control_lock, - .pipe_control_lock_global = dcn20_pipe_control_lock_global, + .interdependent_update_lock = dcn10_lock_all_pipes, + .cursor_lock = dcn10_cursor_lock, .prepare_bandwidth = dcn20_prepare_bandwidth, .optimize_bandwidth = dcn20_optimize_bandwidth, .update_bandwidth = dcn20_update_bandwidth, @@ -81,6 +83,7 @@ static const struct hw_sequencer_funcs dcn20_funcs = { .init_vm_ctx = dcn20_init_vm_ctx, .set_flip_control_gsl = dcn20_set_flip_control_gsl, .get_vupdate_offset_from_vsync = dcn10_get_vupdate_offset_from_vsync, + .calc_vupdate_position = dcn10_calc_vupdate_position, }; static const struct hwseq_private_funcs dcn20_private_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c index de9c857ab3e9..570dfd9a243f 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c @@ -545,6 +545,7 @@ const struct mpc_funcs dcn20_mpc_funcs = { .mpc_init = mpc1_mpc_init, .mpc_init_single_inst = mpc1_mpc_init_single_inst, .update_blending = mpc2_update_blending, + .cursor_lock = mpc1_cursor_lock, .get_mpcc_for_dpp = mpc2_get_mpcc_for_dpp, .wait_for_idle = mpc2_assert_idle_mpcc, .assert_mpcc_idle_before_connect = mpc2_assert_mpcc_idle_before_connect, diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.h b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.h index c78fd5123497..496658f420db 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.h +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.h @@ -179,7 +179,8 @@ SF(MPC_OUT0_DENORM_CLAMP_G_Y, MPC_OUT_DENORM_CLAMP_MAX_G_Y, mask_sh),\ SF(MPC_OUT0_DENORM_CLAMP_G_Y, MPC_OUT_DENORM_CLAMP_MIN_G_Y, mask_sh),\ SF(MPC_OUT0_DENORM_CLAMP_B_CB, MPC_OUT_DENORM_CLAMP_MAX_B_CB, mask_sh),\ - SF(MPC_OUT0_DENORM_CLAMP_B_CB, MPC_OUT_DENORM_CLAMP_MIN_B_CB, mask_sh) + SF(MPC_OUT0_DENORM_CLAMP_B_CB, MPC_OUT_DENORM_CLAMP_MIN_B_CB, mask_sh),\ + SF(CUR_VUPDATE_LOCK_SET0, CUR_VUPDATE_LOCK_SET, mask_sh) /* * DCN2 MPC_OCSC debug status register: 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 e310d67c399a..1ba47f3a6857 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c @@ -506,6 +506,10 @@ enum dcn20_clk_src_array_id { .block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \ mm ## block ## id ## _ ## reg_name +#define VUPDATE_SRII(reg_name, block, id)\ + .reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \ + mm ## reg_name ## _ ## block ## id + /* NBIO */ #define NBIO_BASE_INNER(seg) \ NBIO_BASE__INST0_SEG ## seg @@ -3034,25 +3038,32 @@ validate_out: return out; } - -bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, - bool fast_validate) +/* + * This must be noinline to ensure anything that deals with FP registers + * is contained within this call; previously our compiling with hard-float + * would result in fp instructions being emitted outside of the boundaries + * of the DC_FP_START/END macros, which makes sense as the compiler has no + * idea about what is wrapped and what is not + * + * This is largely just a workaround to avoid breakage introduced with 5.6, + * ideally all fp-using code should be moved into its own file, only that + * should be compiled with hard-float, and all code exported from there + * should be strictly wrapped with DC_FP_START/END + */ +static noinline bool dcn20_validate_bandwidth_fp(struct dc *dc, + struct dc_state *context, bool fast_validate) { bool voltage_supported = false; bool full_pstate_supported = false; bool dummy_pstate_supported = false; double p_state_latency_us; - DC_FP_START(); p_state_latency_us = context->bw_ctx.dml.soc.dram_clock_change_latency_us; context->bw_ctx.dml.soc.disable_dram_clock_change_vactive_support = dc->debug.disable_dram_clock_change_vactive_support; if (fast_validate) { - voltage_supported = dcn20_validate_bandwidth_internal(dc, context, true); - - DC_FP_END(); - return voltage_supported; + return dcn20_validate_bandwidth_internal(dc, context, true); } // Best case, we support full UCLK switch latency @@ -3081,7 +3092,15 @@ bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, restore_dml_state: context->bw_ctx.dml.soc.dram_clock_change_latency_us = p_state_latency_us; + return voltage_supported; +} +bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context, + bool fast_validate) +{ + bool voltage_supported = false; + DC_FP_START(); + voltage_supported = dcn20_validate_bandwidth_fp(dc, context, fast_validate); DC_FP_END(); return voltage_supported; } diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c index fddbd59bf4f9..7f53bf724fce 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c @@ -34,6 +34,7 @@ static const struct hw_sequencer_funcs dcn21_funcs = { .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = NULL, .program_front_end_for_ctx = dcn20_program_front_end_for_ctx, + .post_unlock_program_front_end = dcn20_post_unlock_program_front_end, .update_plane_addr = dcn20_update_plane_addr, .update_dchub = dcn10_update_dchub, .update_pending_status = dcn10_update_pending_status, @@ -51,7 +52,8 @@ static const struct hw_sequencer_funcs dcn21_funcs = { .disable_audio_stream = dce110_disable_audio_stream, .disable_plane = dcn20_disable_plane, .pipe_control_lock = dcn20_pipe_control_lock, - .pipe_control_lock_global = dcn20_pipe_control_lock_global, + .interdependent_update_lock = dcn10_lock_all_pipes, + .cursor_lock = dcn10_cursor_lock, .prepare_bandwidth = dcn20_prepare_bandwidth, .optimize_bandwidth = dcn20_optimize_bandwidth, .update_bandwidth = dcn20_update_bandwidth, @@ -84,6 +86,7 @@ static const struct hw_sequencer_funcs dcn21_funcs = { .optimize_pwr_state = dcn21_optimize_pwr_state, .exit_optimized_pwr_state = dcn21_exit_optimized_pwr_state, .get_vupdate_offset_from_vsync = dcn10_get_vupdate_offset_from_vsync, + .calc_vupdate_position = dcn10_calc_vupdate_position, .set_cursor_position = dcn10_set_cursor_position, .set_cursor_attribute = dcn10_set_cursor_attribute, .set_cursor_sdr_white_level = dcn10_set_cursor_sdr_white_level, diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c index 33d0a176841a..5286cc7d1261 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c @@ -250,7 +250,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn2_1_soc = { .dram_channel_width_bytes = 4, .fabric_datapath_to_dcn_data_return_bytes = 32, .dcn_downspread_percent = 0.5, - .downspread_percent = 0.5, + .downspread_percent = 0.38, .dram_page_open_time_ns = 50.0, .dram_rw_turnaround_time_ns = 17.5, .dram_return_buffer_per_channel_bytes = 8192, @@ -306,6 +306,10 @@ struct _vcs_dpi_soc_bounding_box_st dcn2_1_soc = { .block ## _ ## reg_name[id] = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \ mm ## block ## id ## _ ## reg_name +#define VUPDATE_SRII(reg_name, block, id)\ + .reg_name[id] = BASE(mm ## reg_name ## _ ## block ## id ## _BASE_IDX) + \ + mm ## reg_name ## _ ## block ## id + /* NBIO */ #define NBIO_BASE_INNER(seg) \ NBIF0_BASE__INST0_SEG ## seg diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h b/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h index 094afc4c8173..50ee8aa7ec3b 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h @@ -210,6 +210,22 @@ struct mpc_funcs { struct mpcc_blnd_cfg *blnd_cfg, int mpcc_id); + /* + * Lock cursor updates for the specified OPP. + * OPP defines the set of MPCC that are locked together for cursor. + * + * Parameters: + * [in] mpc - MPC context. + * [in] opp_id - The OPP to lock cursor updates on + * [in] lock - lock/unlock the OPP + * + * Return: void + */ + void (*cursor_lock)( + struct mpc *mpc, + int opp_id, + bool lock); + struct mpcc* (*get_mpcc_for_dpp)( struct mpc_tree *tree, int dpp_id); diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h index 209118f9f193..08307f3796e3 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h @@ -66,6 +66,8 @@ struct hw_sequencer_funcs { int num_planes, struct dc_state *context); void (*program_front_end_for_ctx)(struct dc *dc, struct dc_state *context); + void (*post_unlock_program_front_end)(struct dc *dc, + struct dc_state *context); void (*update_plane_addr)(const struct dc *dc, struct pipe_ctx *pipe_ctx); void (*update_dchub)(struct dce_hwseq *hws, @@ -78,17 +80,23 @@ struct hw_sequencer_funcs { void (*update_pending_status)(struct pipe_ctx *pipe_ctx); /* Pipe Lock Related */ - void (*pipe_control_lock_global)(struct dc *dc, - struct pipe_ctx *pipe, bool lock); void (*pipe_control_lock)(struct dc *dc, struct pipe_ctx *pipe, bool lock); + void (*interdependent_update_lock)(struct dc *dc, + struct dc_state *context, bool lock); void (*set_flip_control_gsl)(struct pipe_ctx *pipe_ctx, bool flip_immediate); + void (*cursor_lock)(struct dc *dc, struct pipe_ctx *pipe, bool lock); /* Timing Related */ void (*get_position)(struct pipe_ctx **pipe_ctx, int num_pipes, struct crtc_position *position); int (*get_vupdate_offset_from_vsync)(struct pipe_ctx *pipe_ctx); + void (*calc_vupdate_position)( + struct dc *dc, + struct pipe_ctx *pipe_ctx, + uint32_t *start_line, + uint32_t *end_line); void (*enable_per_frame_crtc_position_reset)(struct dc *dc, int group_size, struct pipe_ctx *grouped_pipes[]); void (*enable_timing_synchronization)(struct dc *dc, diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c index c195575366a3..8e2acb4df860 100644 --- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c +++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c @@ -319,12 +319,12 @@ static void pp_dpm_en_umd_pstate(struct pp_hwmgr *hwmgr, if (*level & profile_mode_mask) { hwmgr->saved_dpm_level = hwmgr->dpm_level; hwmgr->en_umd_pstate = true; - amdgpu_device_ip_set_clockgating_state(hwmgr->adev, - AMD_IP_BLOCK_TYPE_GFX, - AMD_CG_STATE_UNGATE); amdgpu_device_ip_set_powergating_state(hwmgr->adev, AMD_IP_BLOCK_TYPE_GFX, AMD_PG_STATE_UNGATE); + amdgpu_device_ip_set_clockgating_state(hwmgr->adev, + AMD_IP_BLOCK_TYPE_GFX, + AMD_CG_STATE_UNGATE); } } else { /* exit umd pstate, restore level, enable gfx cg*/ @@ -1435,7 +1435,8 @@ static int pp_get_asic_baco_capability(void *handle, bool *cap) if (!hwmgr) return -EINVAL; - if (!hwmgr->pm_en || !hwmgr->hwmgr_func->get_asic_baco_capability) + if (!(hwmgr->not_vf && amdgpu_dpm) || + !hwmgr->hwmgr_func->get_asic_baco_capability) return 0; mutex_lock(&hwmgr->smu_lock); @@ -1469,7 +1470,8 @@ static int pp_set_asic_baco_state(void *handle, int state) if (!hwmgr) return -EINVAL; - if (!hwmgr->pm_en || !hwmgr->hwmgr_func->set_asic_baco_state) + if (!(hwmgr->not_vf && amdgpu_dpm) || + !hwmgr->hwmgr_func->set_asic_baco_state) return 0; mutex_lock(&hwmgr->smu_lock); diff --git a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c index 96e81c7bc266..e2565967db07 100644 --- a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c @@ -1675,12 +1675,12 @@ static int smu_enable_umd_pstate(void *handle, if (*level & profile_mode_mask) { smu_dpm_ctx->saved_dpm_level = smu_dpm_ctx->dpm_level; smu_dpm_ctx->enable_umd_pstate = true; - amdgpu_device_ip_set_clockgating_state(smu->adev, - AMD_IP_BLOCK_TYPE_GFX, - AMD_CG_STATE_UNGATE); amdgpu_device_ip_set_powergating_state(smu->adev, AMD_IP_BLOCK_TYPE_GFX, AMD_PG_STATE_UNGATE); + amdgpu_device_ip_set_clockgating_state(smu->adev, + AMD_IP_BLOCK_TYPE_GFX, + AMD_CG_STATE_UNGATE); } } else { /* exit umd pstate, restore level, enable gfx cg*/ diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c index 77c14671866c..719597c5d27d 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c @@ -984,6 +984,32 @@ static int init_thermal_controller( struct pp_hwmgr *hwmgr, const ATOM_PPLIB_POWERPLAYTABLE *powerplay_table) { + hwmgr->thermal_controller.ucType = + powerplay_table->sThermalController.ucType; + hwmgr->thermal_controller.ucI2cLine = + powerplay_table->sThermalController.ucI2cLine; + hwmgr->thermal_controller.ucI2cAddress = + powerplay_table->sThermalController.ucI2cAddress; + + hwmgr->thermal_controller.fanInfo.bNoFan = + (0 != (powerplay_table->sThermalController.ucFanParameters & + ATOM_PP_FANPARAMETERS_NOFAN)); + + hwmgr->thermal_controller.fanInfo.ucTachometerPulsesPerRevolution = + powerplay_table->sThermalController.ucFanParameters & + ATOM_PP_FANPARAMETERS_TACHOMETER_PULSES_PER_REVOLUTION_MASK; + + hwmgr->thermal_controller.fanInfo.ulMinRPM + = powerplay_table->sThermalController.ucFanMinRPM * 100UL; + hwmgr->thermal_controller.fanInfo.ulMaxRPM + = powerplay_table->sThermalController.ucFanMaxRPM * 100UL; + + set_hw_cap(hwmgr, + ATOM_PP_THERMALCONTROLLER_NONE != hwmgr->thermal_controller.ucType, + PHM_PlatformCaps_ThermalController); + + hwmgr->thermal_controller.use_hw_fan_control = 1; + return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/renoir_ppt.c b/drivers/gpu/drm/amd/powerplay/renoir_ppt.c index f7a1ce37227c..4a52c310058d 100644 --- a/drivers/gpu/drm/amd/powerplay/renoir_ppt.c +++ b/drivers/gpu/drm/amd/powerplay/renoir_ppt.c @@ -889,12 +889,17 @@ static int renoir_read_sensor(struct smu_context *smu, static bool renoir_is_dpm_running(struct smu_context *smu) { + struct amdgpu_device *adev = smu->adev; + /* * Util now, the pmfw hasn't exported the interface of SMU * feature mask to APU SKU so just force on all the feature * at early initial stage. */ - return true; + if (adev->in_suspend) + return false; + else + return true; } |