From 4d9af438bc41ed20da1963a438249fee43213121 Mon Sep 17 00:00:00 2001 From: Grazvydas Ignotas Date: Thu, 3 Sep 2015 01:54:29 +0300 Subject: [PATCH] r600g: make all viewport states use single atom MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Similarly to scissor states, we can use single atom to track all viewport states. This will allow to simplify dirty atom handling later. Signed-off-by: Marek Olšák --- src/gallium/drivers/r600/evergreen_state.c | 7 +--- src/gallium/drivers/r600/r600_blit.c | 2 +- src/gallium/drivers/r600/r600_hw_context.c | 8 ++-- src/gallium/drivers/r600/r600_pipe.h | 8 ++-- src/gallium/drivers/r600/r600_state.c | 6 +-- src/gallium/drivers/r600/r600_state_common.c | 41 +++++++++++++------- 6 files changed, 38 insertions(+), 34 deletions(-) diff --git a/src/gallium/drivers/r600/evergreen_state.c b/src/gallium/drivers/r600/evergreen_state.c index 3a7f583ecfa..5212db3c1a7 100644 --- a/src/gallium/drivers/r600/evergreen_state.c +++ b/src/gallium/drivers/r600/evergreen_state.c @@ -3450,7 +3450,7 @@ fallback: void evergreen_init_state_functions(struct r600_context *rctx) { unsigned id = 4; - int i; + /* !!! * To avoid GPU lockup registers must be emited in a specific order * (no kidding ...). The order below is important and have been @@ -3503,10 +3503,7 @@ void evergreen_init_state_functions(struct r600_context *rctx) r600_init_atom(rctx, &rctx->poly_offset_state.atom, id++, evergreen_emit_polygon_offset, 6); r600_init_atom(rctx, &rctx->rasterizer_state.atom, id++, r600_emit_cso_state, 0); r600_init_atom(rctx, &rctx->scissor.atom, id++, evergreen_emit_scissor_state, 0); - for (i = 0; i < R600_MAX_VIEWPORTS; i++) { - r600_init_atom(rctx, &rctx->viewport[i].atom, id++, r600_emit_viewport_state, 8); - rctx->viewport[i].idx = i; - } + r600_init_atom(rctx, &rctx->viewport.atom, id++, r600_emit_viewport_state, 0); r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4); r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, evergreen_emit_vertex_fetch_shader, 5); r600_add_atom(rctx, &rctx->b.streamout.begin_atom, id++); diff --git a/src/gallium/drivers/r600/r600_blit.c b/src/gallium/drivers/r600/r600_blit.c index 8774cb5d655..d1370cd8f26 100644 --- a/src/gallium/drivers/r600/r600_blit.c +++ b/src/gallium/drivers/r600/r600_blit.c @@ -65,7 +65,7 @@ static void r600_blitter_begin(struct pipe_context *ctx, enum r600_blitter_op op util_blitter_save_rasterizer(rctx->blitter, rctx->rasterizer_state.cso); if (op & R600_SAVE_FRAGMENT_STATE) { - util_blitter_save_viewport(rctx->blitter, &rctx->viewport[0].state); + util_blitter_save_viewport(rctx->blitter, &rctx->viewport.state[0]); util_blitter_save_scissor(rctx->blitter, &rctx->scissor.scissor[0]); util_blitter_save_fragment_shader(rctx->blitter, rctx->ps_shader); util_blitter_save_blend(rctx->blitter, rctx->blend_state.cso); diff --git a/src/gallium/drivers/r600/r600_hw_context.c b/src/gallium/drivers/r600/r600_hw_context.c index c5403252d36..092f261700c 100644 --- a/src/gallium/drivers/r600/r600_hw_context.c +++ b/src/gallium/drivers/r600/r600_hw_context.c @@ -287,7 +287,7 @@ void r600_context_gfx_flush(void *context, unsigned flags, void r600_begin_new_cs(struct r600_context *ctx) { unsigned shader; - int i; + ctx->b.flags = 0; ctx->b.gtt = 0; ctx->b.vram = 0; @@ -311,9 +311,9 @@ void r600_begin_new_cs(struct r600_context *ctx) ctx->scissor.dirty_mask = (1 << R600_MAX_VIEWPORTS) - 1; ctx->scissor.atom.num_dw = R600_MAX_VIEWPORTS * 4; r600_mark_atom_dirty(ctx, &ctx->scissor.atom); - for (i = 0; i < R600_MAX_VIEWPORTS; i++) { - r600_mark_atom_dirty(ctx, &ctx->viewport[i].atom); - } + ctx->viewport.dirty_mask = (1 << R600_MAX_VIEWPORTS) - 1; + ctx->viewport.atom.num_dw = R600_MAX_VIEWPORTS * 8; + r600_mark_atom_dirty(ctx, &ctx->viewport.atom); if (ctx->b.chip_class < EVERGREEN) { r600_mark_atom_dirty(ctx, &ctx->config_state.atom); } diff --git a/src/gallium/drivers/r600/r600_pipe.h b/src/gallium/drivers/r600/r600_pipe.h index e09bee15d46..bf6bd766097 100644 --- a/src/gallium/drivers/r600/r600_pipe.h +++ b/src/gallium/drivers/r600/r600_pipe.h @@ -38,7 +38,7 @@ #include "tgsi/tgsi_scan.h" -#define R600_NUM_ATOMS 60 +#define R600_NUM_ATOMS 45 #define R600_MAX_VIEWPORTS 16 @@ -208,8 +208,8 @@ struct r600_stencil_ref_state { struct r600_viewport_state { struct r600_atom atom; - struct pipe_viewport_state state; - int idx; + struct pipe_viewport_state state[R600_MAX_VIEWPORTS]; + uint32_t dirty_mask; }; struct r600_shader_stages_state { @@ -463,7 +463,7 @@ struct r600_context { struct r600_config_state config_state; struct r600_stencil_ref_state stencil_ref; struct r600_vgt_state vgt_state; - struct r600_viewport_state viewport[R600_MAX_VIEWPORTS]; + struct r600_viewport_state viewport; /* Shaders and shader resources. */ struct r600_cso_state vertex_fetch_shader; struct r600_shader_state vertex_shader; diff --git a/src/gallium/drivers/r600/r600_state.c b/src/gallium/drivers/r600/r600_state.c index c3db1437200..6724cc080b5 100644 --- a/src/gallium/drivers/r600/r600_state.c +++ b/src/gallium/drivers/r600/r600_state.c @@ -3032,7 +3032,6 @@ fallback: void r600_init_state_functions(struct r600_context *rctx) { unsigned id = 4; - int i; /* !!! * To avoid GPU lockup registers must be emited in a specific order @@ -3081,10 +3080,7 @@ void r600_init_state_functions(struct r600_context *rctx) r600_init_atom(rctx, &rctx->poly_offset_state.atom, id++, r600_emit_polygon_offset, 6); r600_init_atom(rctx, &rctx->rasterizer_state.atom, id++, r600_emit_cso_state, 0); r600_init_atom(rctx, &rctx->scissor.atom, id++, r600_emit_scissor_state, 0); - for (i = 0;i < R600_MAX_VIEWPORTS; i++) { - r600_init_atom(rctx, &rctx->viewport[i].atom, id++, r600_emit_viewport_state, 8); - rctx->viewport[i].idx = i; - } + r600_init_atom(rctx, &rctx->viewport.atom, id++, r600_emit_viewport_state, 0); r600_init_atom(rctx, &rctx->config_state.atom, id++, r600_emit_config_state, 3); r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4); r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, r600_emit_vertex_fetch_shader, 5); diff --git a/src/gallium/drivers/r600/r600_state_common.c b/src/gallium/drivers/r600/r600_state_common.c index 0b3fe857591..1f96ed6e12c 100644 --- a/src/gallium/drivers/r600/r600_state_common.c +++ b/src/gallium/drivers/r600/r600_state_common.c @@ -704,28 +704,39 @@ static void r600_set_viewport_states(struct pipe_context *ctx, const struct pipe_viewport_state *state) { struct r600_context *rctx = (struct r600_context *)ctx; + struct r600_viewport_state *rstate = &rctx->viewport; int i; - for (i = start_slot; i < start_slot + num_viewports; i++) { - rctx->viewport[i].state = state[i - start_slot]; - r600_mark_atom_dirty(rctx, &rctx->viewport[i].atom); - } + for (i = start_slot; i < start_slot + num_viewports; i++) + rstate->state[i] = state[i - start_slot]; + rstate->dirty_mask |= ((1 << num_viewports) - 1) << start_slot; + rstate->atom.num_dw = util_bitcount(rstate->dirty_mask) * 8; + r600_mark_atom_dirty(rctx, &rctx->viewport.atom); } void r600_emit_viewport_state(struct r600_context *rctx, struct r600_atom *atom) { struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs; - struct r600_viewport_state *rstate = (struct r600_viewport_state *)atom; - struct pipe_viewport_state *state = &rstate->state; - int offset = rstate->idx * 6 * 4; - - radeon_set_context_reg_seq(cs, R_02843C_PA_CL_VPORT_XSCALE_0 + offset, 6); - radeon_emit(cs, fui(state->scale[0])); /* R_02843C_PA_CL_VPORT_XSCALE_0 */ - radeon_emit(cs, fui(state->translate[0])); /* R_028440_PA_CL_VPORT_XOFFSET_0 */ - radeon_emit(cs, fui(state->scale[1])); /* R_028444_PA_CL_VPORT_YSCALE_0 */ - radeon_emit(cs, fui(state->translate[1])); /* R_028448_PA_CL_VPORT_YOFFSET_0 */ - radeon_emit(cs, fui(state->scale[2])); /* R_02844C_PA_CL_VPORT_ZSCALE_0 */ - radeon_emit(cs, fui(state->translate[2])); /* R_028450_PA_CL_VPORT_ZOFFSET_0 */ + struct r600_viewport_state *rstate = &rctx->viewport; + struct pipe_viewport_state *state; + uint32_t dirty_mask; + unsigned i, offset; + + dirty_mask = rstate->dirty_mask; + while (dirty_mask != 0) { + i = u_bit_scan(&dirty_mask); + offset = i * 6 * 4; + radeon_set_context_reg_seq(cs, R_02843C_PA_CL_VPORT_XSCALE_0 + offset, 6); + state = &rstate->state[i]; + radeon_emit(cs, fui(state->scale[0])); /* R_02843C_PA_CL_VPORT_XSCALE_0 */ + radeon_emit(cs, fui(state->translate[0])); /* R_028440_PA_CL_VPORT_XOFFSET_0 */ + radeon_emit(cs, fui(state->scale[1])); /* R_028444_PA_CL_VPORT_YSCALE_0 */ + radeon_emit(cs, fui(state->translate[1])); /* R_028448_PA_CL_VPORT_YOFFSET_0 */ + radeon_emit(cs, fui(state->scale[2])); /* R_02844C_PA_CL_VPORT_ZSCALE_0 */ + radeon_emit(cs, fui(state->translate[2])); /* R_028450_PA_CL_VPORT_ZOFFSET_0 */ + } + rstate->dirty_mask = 0; + rstate->atom.num_dw = 0; } /* Compute the key for the hw shader variant */ -- 2.30.2