r600g: add support for multiple viewports.

tested on rv635 and barts.

Signed-off-by: Dave Airlie <airlied@redhat.com>

9 files changed, 98 insertions, 40 deletions

diff --git a/src/gallium/drivers/r600/evergreen_state.c b/src/gallium/drivers/r600/evergreen_state.c
index c86e812182b..41607b19d4c 100644
--- a/src/gallium/drivers/r600/evergreen_state.c
+++ b/src/gallium/drivers/r600/evergreen_state.c
@@ -1320,20 +1320,25 @@ static void evergreen_set_scissor_states(struct pipe_context *ctx,
 					const struct pipe_scissor_state *state)
 {
 	struct r600_context *rctx = (struct r600_context *)ctx;
+	int i;
 
-	rctx->scissor.scissor = *state;
-	rctx->scissor.atom.dirty = true;
+	for (i = start_slot; i < start_slot + num_scissors; i++) {
+		rctx->scissor[i].scissor = state[i - start_slot];
+		rctx->scissor[i].atom.dirty = true;
+	}
 }
 
 static void evergreen_emit_scissor_state(struct r600_context *rctx, struct r600_atom *atom)
 {
 	struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
-	struct pipe_scissor_state *state = &rctx->scissor.scissor;
+	struct r600_scissor_state *rstate = (struct r600_scissor_state *)atom;
+	struct pipe_scissor_state *state = &rstate->scissor;
+	unsigned offset = rstate->idx * 4 * 2;
 	uint32_t tl, br;
 
 	evergreen_get_scissor_rect(rctx, state->minx, state->miny, state->maxx, state->maxy, &tl, &br);
 
-	r600_write_context_reg_seq(cs, R_028250_PA_SC_VPORT_SCISSOR_0_TL, 2);
+	r600_write_context_reg_seq(cs, R_028250_PA_SC_VPORT_SCISSOR_0_TL + offset, 2);
 	radeon_emit(cs, tl);
 	radeon_emit(cs, br);
 }
@@ -2817,6 +2822,7 @@ void cayman_init_common_regs(struct r600_command_buffer *cb,
 static void cayman_init_atom_start_cs(struct r600_context *rctx)
 {
 	struct r600_command_buffer *cb = &rctx->start_cs_cmd;
+	int tmp;
 
 	r600_init_command_buffer(cb, 256);
 
@@ -2904,9 +2910,11 @@ static void cayman_init_atom_start_cs(struct r600_context *rctx)
 	r600_store_context_reg(cb, R_028200_PA_SC_WINDOW_OFFSET, 0);
 	r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
 
-	r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2);
-	r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
-	r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+	r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * 16);
+	for (tmp = 0; tmp < 16; tmp++) {
+		r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
+		r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+	}
 
 	r600_store_context_reg(cb, R_028230_PA_SC_EDGERULE, 0xAAAAAAAA);
 	r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL, 0x0000043F);
@@ -3353,9 +3361,11 @@ void evergreen_init_atom_start_cs(struct r600_context *rctx)
 	r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
 	r600_store_context_reg(cb, R_028230_PA_SC_EDGERULE, 0xAAAAAAAA);
 
-	r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2);
-	r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
-	r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+	r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * 16);
+	for (tmp = 0; tmp < 16; tmp++) {
+		r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
+		r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+	}
 
 	r600_store_context_reg(cb, R_0286DC_SPI_FOG_CNTL, 0);
 	r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL, 0x0000043F);
@@ -3727,6 +3737,7 @@ void evergreen_update_vs_state(struct pipe_context *ctx, struct r600_pipe_shader
 		S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->clip_dist_write & 0xF0) != 0) |
 		S_02881C_VS_OUT_MISC_VEC_ENA(rshader->vs_out_misc_write) |
 		S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size) |
+		S_02881C_USE_VTX_VIEWPORT_INDX(rshader->vs_out_viewport) |
 		S_02881C_USE_VTX_RENDER_TARGET_INDX(rshader->vs_out_layer);
 }
 
@@ -4037,7 +4048,7 @@ static boolean evergreen_dma_blit(struct pipe_context *ctx,
 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
@@ -4087,9 +4098,13 @@ void evergreen_init_state_functions(struct r600_context *rctx)
 	r600_init_atom(rctx, &rctx->dsa_state.atom, id++, r600_emit_cso_state, 0);
 	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, 4);
+	for (i = 0; i < 16; i++) {
+		r600_init_atom(rctx, &rctx->viewport[i].atom, id++, r600_emit_viewport_state, 8);
+		r600_init_atom(rctx, &rctx->scissor[i].atom, id++, evergreen_emit_scissor_state, 4);
+		rctx->viewport[i].idx = i;
+		rctx->scissor[i].idx = i;
+	}
 	r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4);
-	r600_init_atom(rctx, &rctx->viewport.atom, id++, r600_emit_viewport_state, 8);
 	r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, evergreen_emit_vertex_fetch_shader, 5);
 	rctx->atoms[id++] = &rctx->b.streamout.begin_atom;
 	r600_init_atom(rctx, &rctx->vertex_shader.atom, id++, r600_emit_shader, 23);
diff --git a/src/gallium/drivers/r600/r600_blit.c b/src/gallium/drivers/r600/r600_blit.c
index 6bb7cfe9b9e..770eb38e5d9 100644
--- a/src/gallium/drivers/r600/r600_blit.c
+++ b/src/gallium/drivers/r600/r600_blit.c
@@ -65,8 +65,8 @@ 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.state);
-		util_blitter_save_scissor(rctx->blitter, &rctx->scissor.scissor);
+		util_blitter_save_viewport(rctx->blitter, &rctx->viewport[0].state);
+		util_blitter_save_scissor(rctx->blitter, &rctx->scissor[0].scissor);
 		util_blitter_save_fragment_shader(rctx->blitter, rctx->ps_shader);
 		util_blitter_save_blend(rctx->blitter, rctx->blend_state.cso);
 		util_blitter_save_depth_stencil_alpha(rctx->blitter, rctx->dsa_state.cso);
diff --git a/src/gallium/drivers/r600/r600_hw_context.c b/src/gallium/drivers/r600/r600_hw_context.c
index fc81e95accd..dc3c22156ff 100644
--- a/src/gallium/drivers/r600/r600_hw_context.c
+++ b/src/gallium/drivers/r600/r600_hw_context.c
@@ -276,7 +276,7 @@ void r600_context_flush(struct r600_context *ctx, 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;
@@ -297,7 +297,10 @@ void r600_begin_new_cs(struct r600_context *ctx)
 	ctx->poly_offset_state.atom.dirty = true;
 	ctx->vgt_state.atom.dirty = true;
 	ctx->sample_mask.atom.dirty = true;
-	ctx->scissor.atom.dirty = true;
+	for (i = 0; i < 16; i++) {
+		ctx->scissor[i].atom.dirty = true;
+		ctx->viewport[i].atom.dirty = true;
+	}
 	ctx->config_state.atom.dirty = true;
 	ctx->stencil_ref.atom.dirty = true;
 	ctx->vertex_fetch_shader.atom.dirty = true;
@@ -308,7 +311,7 @@ void r600_begin_new_cs(struct r600_context *ctx)
 		ctx->gs_rings.atom.dirty = true;
 	}
 	ctx->vertex_shader.atom.dirty = true;
-	ctx->viewport.atom.dirty = true;
+
 
 	if (ctx->blend_state.cso)
 		ctx->blend_state.atom.dirty = true;
diff --git a/src/gallium/drivers/r600/r600_pipe.c b/src/gallium/drivers/r600/r600_pipe.c
index d9b4509e078..796f0f5066a 100644
--- a/src/gallium/drivers/r600/r600_pipe.c
+++ b/src/gallium/drivers/r600/r600_pipe.c
@@ -434,7 +434,7 @@ static int r600_get_param(struct pipe_screen* pscreen, enum pipe_cap param)
 		return 8;
 
 	case PIPE_CAP_MAX_VIEWPORTS:
-		return 1;
+		return 16;
 
 	/* Timer queries, present when the clock frequency is non zero. */
 	case PIPE_CAP_QUERY_TIME_ELAPSED:
diff --git a/src/gallium/drivers/r600/r600_pipe.h b/src/gallium/drivers/r600/r600_pipe.h
index 9f27a17cd53..15052da8d22 100644
--- a/src/gallium/drivers/r600/r600_pipe.h
+++ b/src/gallium/drivers/r600/r600_pipe.h
@@ -38,7 +38,7 @@
 #include "util/u_double_list.h"
 #include "util/u_transfer.h"
 
-#define R600_NUM_ATOMS 42
+#define R600_NUM_ATOMS 72
 
 /* the number of CS dwords for flushing and drawing */
 #define R600_MAX_FLUSH_CS_DWORDS	16
@@ -179,6 +179,7 @@ struct r600_stencil_ref_state {
 struct r600_viewport_state {
 	struct r600_atom atom;
 	struct pipe_viewport_state state;
+	int idx;
 };
 
 struct r600_shader_stages_state {
@@ -358,6 +359,7 @@ struct r600_scissor_state
 	struct r600_atom		atom;
 	struct pipe_scissor_state	scissor;
 	bool				enable; /* r6xx only */
+	int idx;
 };
 
 struct r600_fetch_shader {
@@ -419,12 +421,12 @@ struct r600_context {
 	struct r600_poly_offset_state	poly_offset_state;
 	struct r600_cso_state		rasterizer_state;
 	struct r600_sample_mask		sample_mask;
-	struct r600_scissor_state	scissor;
+	struct r600_scissor_state	scissor[16];
 	struct r600_seamless_cube_map	seamless_cube_map;
 	struct r600_config_state	config_state;
 	struct r600_stencil_ref_state	stencil_ref;
 	struct r600_vgt_state		vgt_state;
-	struct r600_viewport_state	viewport;
+	struct r600_viewport_state	viewport[16];
 	/* 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_shader.c b/src/gallium/drivers/r600/r600_shader.c
index 758abd76bf4..baeef57b22d 100644
--- a/src/gallium/drivers/r600/r600_shader.c
+++ b/src/gallium/drivers/r600/r600_shader.c
@@ -492,6 +492,7 @@ static int r600_spi_sid(struct r600_shader_io * io)
 	if (name == TGSI_SEMANTIC_POSITION ||
 		name == TGSI_SEMANTIC_PSIZE ||
 		name == TGSI_SEMANTIC_LAYER ||
+		name == TGSI_SEMANTIC_VIEWPORT_INDEX ||
 		name == TGSI_SEMANTIC_FACE)
 		index = 0;
 	else {
@@ -623,6 +624,10 @@ static int tgsi_declaration(struct r600_shader_ctx *ctx)
 				ctx->shader->vs_out_misc_write = 1;
 				ctx->shader->vs_out_point_size = 1;
 				break;
+			case TGSI_SEMANTIC_VIEWPORT_INDEX:
+				ctx->shader->vs_out_misc_write = 1;
+				ctx->shader->vs_out_viewport = 1;
+				break;
 			case TGSI_SEMANTIC_LAYER:
 				ctx->shader->vs_out_misc_write = 1;
 				ctx->shader->vs_out_layer = 1;
@@ -1296,6 +1301,18 @@ static int generate_gs_copy_shader(struct r600_context *rctx,
 			ctx.shader->vs_out_misc_write = 1;
 			ctx.shader->vs_out_layer = 1;
 			break;
+		case TGSI_SEMANTIC_VIEWPORT_INDEX:
+			output.array_base = 61;
+			if (next_clip_pos == 61)
+				next_clip_pos = 62;
+			output.type = V_SQ_CF_ALLOC_EXPORT_WORD0_SQ_EXPORT_POS;
+			ctx.shader->vs_out_misc_write = 1;
+			ctx.shader->vs_out_viewport = 1;
+			output.swizzle_x = 7;
+			output.swizzle_y = 7;
+			output.swizzle_z = 7;
+			output.swizzle_w = 0;
+			break;
 		case TGSI_SEMANTIC_CLIPDIST:
 			/* spi_sid is 0 for clipdistance outputs that were generated
 			 * for clipvertex - we don't need to pass them to PS */
diff --git a/src/gallium/drivers/r600/r600_shader.h b/src/gallium/drivers/r600/r600_shader.h
index 89602c7013f..519d0aaf0bf 100644
--- a/src/gallium/drivers/r600/r600_shader.h
+++ b/src/gallium/drivers/r600/r600_shader.h
@@ -63,6 +63,7 @@ struct r600_shader {
 	boolean			vs_out_misc_write;
 	boolean			vs_out_point_size;
 	boolean			vs_out_layer;
+	boolean			vs_out_viewport;
 	boolean			has_txq_cube_array_z_comp;
 	boolean			uses_tex_buffers;
 	boolean                 gs_prim_id_input;
diff --git a/src/gallium/drivers/r600/r600_state.c b/src/gallium/drivers/r600/r600_state.c
index 0dfd4260a82..5538898d2fe 100644
--- a/src/gallium/drivers/r600/r600_state.c
+++ b/src/gallium/drivers/r600/r600_state.c
@@ -1205,10 +1205,12 @@ static void r600_set_polygon_stipple(struct pipe_context *ctx,
 static void r600_emit_scissor_state(struct r600_context *rctx, struct r600_atom *atom)
 {
 	struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
-	struct pipe_scissor_state *state = &rctx->scissor.scissor;
+	struct r600_scissor_state *rstate = (struct r600_scissor_state *)atom;
+	struct pipe_scissor_state *state = &rstate->scissor;
+	unsigned offset = rstate->idx * 4 * 2;
 
-	if (rctx->b.chip_class != R600 || rctx->scissor.enable) {
-		r600_write_context_reg_seq(cs, R_028250_PA_SC_VPORT_SCISSOR_0_TL, 2);
+	if (rctx->b.chip_class != R600 || rctx->scissor[0].enable) {
+		r600_write_context_reg_seq(cs, R_028250_PA_SC_VPORT_SCISSOR_0_TL + offset, 2);
 		radeon_emit(cs, S_028240_TL_X(state->minx) | S_028240_TL_Y(state->miny) |
 				     S_028240_WINDOW_OFFSET_DISABLE(1));
 		radeon_emit(cs, S_028244_BR_X(state->maxx) | S_028244_BR_Y(state->maxy));
@@ -1226,13 +1228,18 @@ static void r600_set_scissor_states(struct pipe_context *ctx,
                                     const struct pipe_scissor_state *state)
 {
 	struct r600_context *rctx = (struct r600_context *)ctx;
+	int i;
 
-	rctx->scissor.scissor = *state;
+	for (i = start_slot ; i < start_slot + num_scissors; i++) {
+		rctx->scissor[i].scissor = state[i - start_slot];
+	}
 
-	if (rctx->b.chip_class == R600 && !rctx->scissor.enable)
+	if (rctx->b.chip_class == R600 && !rctx->scissor[0].enable)
 		return;
 
-	rctx->scissor.atom.dirty = true;
+	for (i = start_slot ; i < start_slot + num_scissors; i++) {
+		rctx->scissor[i].atom.dirty = true;
+	}
 }
 
 static struct r600_resource *r600_buffer_create_helper(struct r600_screen *rscreen,
@@ -2775,9 +2782,11 @@ void r600_init_atom_start_cs(struct r600_context *rctx)
 	r600_store_value(cb, 0x3F800000); /* R_028C14_PA_CL_GB_HORZ_CLIP_ADJ */
 	r600_store_value(cb, 0x3F800000); /* R_028C18_PA_CL_GB_HORZ_DISC_ADJ */
 
-	r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2);
-	r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
-	r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+	r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * 16);
+	for (tmp = 0; tmp < 16; tmp++) {
+		r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
+		r600_store_value(cb, 0x3F800000); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
+	}
 
 	r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL, 0x43F);
 
@@ -2996,6 +3005,7 @@ void r600_update_vs_state(struct pipe_context *ctx, struct r600_pipe_shader *sha
 		S_02881C_VS_OUT_CCDIST0_VEC_ENA((rshader->clip_dist_write & 0x0F) != 0) |
 		S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->clip_dist_write & 0xF0) != 0) |
 		S_02881C_VS_OUT_MISC_VEC_ENA(rshader->vs_out_misc_write) |
+		S_02881C_USE_VTX_VIEWPORT_INDX(rshader->vs_out_viewport) |
 		S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size);
 }
 
@@ -3393,6 +3403,7 @@ static boolean r600_dma_blit(struct pipe_context *ctx,
 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
@@ -3440,10 +3451,14 @@ void r600_init_state_functions(struct r600_context *rctx)
 	r600_init_atom(rctx, &rctx->dsa_state.atom, id++, r600_emit_cso_state, 0);
 	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, 4);
+	for (i = 0;i < 16; i++) {
+		r600_init_atom(rctx, &rctx->scissor[i].atom, id++, r600_emit_scissor_state, 4);
+		r600_init_atom(rctx, &rctx->viewport[i].atom, id++, r600_emit_viewport_state, 8);
+		rctx->scissor[i].idx = i;
+		rctx->viewport[i].idx = i;
+	}
 	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->viewport.atom, id++, r600_emit_viewport_state, 8);
 	r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, r600_emit_vertex_fetch_shader, 5);
 	rctx->atoms[id++] = &rctx->b.streamout.begin_atom;
 	r600_init_atom(rctx, &rctx->vertex_shader.atom, id++, r600_emit_shader, 23);
diff --git a/src/gallium/drivers/r600/r600_state_common.c b/src/gallium/drivers/r600/r600_state_common.c
index 22f2b062a96..da4578a3781 100644
--- a/src/gallium/drivers/r600/r600_state_common.c
+++ b/src/gallium/drivers/r600/r600_state_common.c
@@ -334,9 +334,9 @@ static void r600_bind_rs_state(struct pipe_context *ctx, void *state)
 
 	/* Workaround for a missing scissor enable on r600. */
 	if (rctx->b.chip_class == R600 &&
-	    rs->scissor_enable != rctx->scissor.enable) {
-		rctx->scissor.enable = rs->scissor_enable;
-		rctx->scissor.atom.dirty = true;
+	    rs->scissor_enable != rctx->scissor[0].enable) {
+		rctx->scissor[0].enable = rs->scissor_enable;
+		rctx->scissor[0].atom.dirty = true;
 	}
 
 	/* Re-emit PA_SC_LINE_STIPPLE. */
@@ -657,17 +657,22 @@ static void r600_set_viewport_states(struct pipe_context *ctx,
                                      const struct pipe_viewport_state *state)
 {
 	struct r600_context *rctx = (struct r600_context *)ctx;
+	int i;
 
-	rctx->viewport.state = *state;
-	rctx->viewport.atom.dirty = true;
+	for (i = start_slot; i < start_slot + num_viewports; i++) {
+		rctx->viewport[i].state = state[i - start_slot];
+		rctx->viewport[i].atom.dirty = true;
+	}
 }
 
 void r600_emit_viewport_state(struct r600_context *rctx, struct r600_atom *atom)
 {
 	struct radeon_winsys_cs *cs = rctx->b.rings.gfx.cs;
-	struct pipe_viewport_state *state = &rctx->viewport.state;
+	struct r600_viewport_state *rstate = (struct r600_viewport_state *)atom;
+	struct pipe_viewport_state *state = &rstate->state;
+	int offset = rstate->idx * 6 * 4;
 
-	r600_write_context_reg_seq(cs, R_02843C_PA_CL_VPORT_XSCALE_0, 6);
+	r600_write_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  */