From: Marek Olšák <[email protected]>

---
 src/gallium/drivers/radeon/Makefile.sources   |   1 -
 src/gallium/drivers/radeon/cayman_msaa.c      | 269 --------------------------
 src/gallium/drivers/radeon/r600_pipe_common.c |   1 -
 src/gallium/drivers/radeon/r600_pipe_common.h |  25 ---
 src/gallium/drivers/radeonsi/Makefile.sources |   1 +
 src/gallium/drivers/radeonsi/si_pipe.c        |   1 +
 src/gallium/drivers/radeonsi/si_pipe.h        |   9 +
 src/gallium/drivers/radeonsi/si_state.c       |  79 ++++++--
 src/gallium/drivers/radeonsi/si_state.h       |   4 +
 src/gallium/drivers/radeonsi/si_state_msaa.c  | 209 ++++++++++++++++++++
 10 files changed, 292 insertions(+), 307 deletions(-)
 delete mode 100644 src/gallium/drivers/radeon/cayman_msaa.c
 create mode 100644 src/gallium/drivers/radeonsi/si_state_msaa.c

diff --git a/src/gallium/drivers/radeon/Makefile.sources 
b/src/gallium/drivers/radeon/Makefile.sources
index c32ebea..22de129 100644
--- a/src/gallium/drivers/radeon/Makefile.sources
+++ b/src/gallium/drivers/radeon/Makefile.sources
@@ -1,12 +1,11 @@
 C_SOURCES := \
-       cayman_msaa.c \
        r600_buffer_common.c \
        r600_cs.h \
        r600_gpu_load.c \
        r600_perfcounter.c \
        r600_pipe_common.c \
        r600_pipe_common.h \
        r600_query.c \
        r600_query.h \
        r600_test_dma.c \
        r600_texture.c \
diff --git a/src/gallium/drivers/radeon/cayman_msaa.c 
b/src/gallium/drivers/radeon/cayman_msaa.c
deleted file mode 100644
index 72a393e..0000000
--- a/src/gallium/drivers/radeon/cayman_msaa.c
+++ /dev/null
@@ -1,269 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING 
FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN 
THE
- * SOFTWARE.
- *
- * Authors: Marek Olšák <[email protected]>
- *
- */
-
-#include "r600_cs.h"
-
-/* 2xMSAA
- * There are two locations (4, 4), (-4, -4). */
-static const uint32_t eg_sample_locs_2x[4] = {
-       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
-       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
-       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
-       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
-};
-static const unsigned eg_max_dist_2x = 4;
-/* 4xMSAA
- * There are 4 locations: (-2, -6), (6, -2), (-6, 2), (2, 6). */
-static const uint32_t eg_sample_locs_4x[4] = {
-       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
-       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
-       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
-       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
-};
-static const unsigned eg_max_dist_4x = 6;
-
-/* Cayman 8xMSAA */
-static const uint32_t cm_sample_locs_8x[] = {
-       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
-       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
-       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
-       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
-       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
-       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
-       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
-       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
-};
-static const unsigned cm_max_dist_8x = 8;
-/* Cayman 16xMSAA */
-static const uint32_t cm_sample_locs_16x[] = {
-       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
-       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
-       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
-       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
-       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
-       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
-       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
-       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
-       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
-       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
-       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
-       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
-       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
-       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
-       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
-       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
-};
-static const unsigned cm_max_dist_16x = 8;
-
-void si_get_sample_position(struct pipe_context *ctx, unsigned sample_count,
-                               unsigned sample_index, float *out_value)
-{
-       int offset, index;
-       struct {
-               int idx:4;
-       } val;
-       switch (sample_count) {
-       case 1:
-       default:
-               out_value[0] = out_value[1] = 0.5;
-               break;
-       case 2:
-               offset = 4 * (sample_index * 2);
-               val.idx = (eg_sample_locs_2x[0] >> offset) & 0xf;
-               out_value[0] = (float)(val.idx + 8) / 16.0f;
-               val.idx = (eg_sample_locs_2x[0] >> (offset + 4)) & 0xf;
-               out_value[1] = (float)(val.idx + 8) / 16.0f;
-               break;
-       case 4:
-               offset = 4 * (sample_index * 2);
-               val.idx = (eg_sample_locs_4x[0] >> offset) & 0xf;
-               out_value[0] = (float)(val.idx + 8) / 16.0f;
-               val.idx = (eg_sample_locs_4x[0] >> (offset + 4)) & 0xf;
-               out_value[1] = (float)(val.idx + 8) / 16.0f;
-               break;
-       case 8:
-               offset = 4 * (sample_index % 4 * 2);
-               index = (sample_index / 4) * 4;
-               val.idx = (cm_sample_locs_8x[index] >> offset) & 0xf;
-               out_value[0] = (float)(val.idx + 8) / 16.0f;
-               val.idx = (cm_sample_locs_8x[index] >> (offset + 4)) & 0xf;
-               out_value[1] = (float)(val.idx + 8) / 16.0f;
-               break;
-       case 16:
-               offset = 4 * (sample_index % 4 * 2);
-               index = (sample_index / 4) * 4;
-               val.idx = (cm_sample_locs_16x[index] >> offset) & 0xf;
-               out_value[0] = (float)(val.idx + 8) / 16.0f;
-               val.idx = (cm_sample_locs_16x[index] >> (offset + 4)) & 0xf;
-               out_value[1] = (float)(val.idx + 8) / 16.0f;
-               break;
-       }
-}
-
-void si_init_msaa(struct pipe_context *ctx)
-{
-       struct r600_common_context *rctx = (struct r600_common_context*)ctx;
-       int i;
-
-       si_get_sample_position(ctx, 1, 0, rctx->sample_locations_1x[0]);
-
-       for (i = 0; i < 2; i++)
-               si_get_sample_position(ctx, 2, i, rctx->sample_locations_2x[i]);
-       for (i = 0; i < 4; i++)
-               si_get_sample_position(ctx, 4, i, rctx->sample_locations_4x[i]);
-       for (i = 0; i < 8; i++)
-               si_get_sample_position(ctx, 8, i, rctx->sample_locations_8x[i]);
-       for (i = 0; i < 16; i++)
-               si_get_sample_position(ctx, 16, i, 
rctx->sample_locations_16x[i]);
-}
-
-void si_common_emit_msaa_sample_locs(struct radeon_winsys_cs *cs, int 
nr_samples)
-{
-       switch (nr_samples) {
-       default:
-       case 1:
-               radeon_set_context_reg(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 0);
-               radeon_set_context_reg(cs, 
CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, 0);
-               radeon_set_context_reg(cs, 
CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, 0);
-               radeon_set_context_reg(cs, 
CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, 0);
-               break;
-       case 2:
-               radeon_set_context_reg(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, eg_sample_locs_2x[0]);
-               radeon_set_context_reg(cs, 
CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, eg_sample_locs_2x[1]);
-               radeon_set_context_reg(cs, 
CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, eg_sample_locs_2x[2]);
-               radeon_set_context_reg(cs, 
CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, eg_sample_locs_2x[3]);
-               break;
-       case 4:
-               radeon_set_context_reg(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, eg_sample_locs_4x[0]);
-               radeon_set_context_reg(cs, 
CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, eg_sample_locs_4x[1]);
-               radeon_set_context_reg(cs, 
CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, eg_sample_locs_4x[2]);
-               radeon_set_context_reg(cs, 
CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, eg_sample_locs_4x[3]);
-               break;
-       case 8:
-               radeon_set_context_reg_seq(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 14);
-               radeon_emit(cs, cm_sample_locs_8x[0]);
-               radeon_emit(cs, cm_sample_locs_8x[4]);
-               radeon_emit(cs, 0);
-               radeon_emit(cs, 0);
-               radeon_emit(cs, cm_sample_locs_8x[1]);
-               radeon_emit(cs, cm_sample_locs_8x[5]);
-               radeon_emit(cs, 0);
-               radeon_emit(cs, 0);
-               radeon_emit(cs, cm_sample_locs_8x[2]);
-               radeon_emit(cs, cm_sample_locs_8x[6]);
-               radeon_emit(cs, 0);
-               radeon_emit(cs, 0);
-               radeon_emit(cs, cm_sample_locs_8x[3]);
-               radeon_emit(cs, cm_sample_locs_8x[7]);
-               break;
-       case 16:
-               radeon_set_context_reg_seq(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 16);
-               radeon_emit(cs, cm_sample_locs_16x[0]);
-               radeon_emit(cs, cm_sample_locs_16x[4]);
-               radeon_emit(cs, cm_sample_locs_16x[8]);
-               radeon_emit(cs, cm_sample_locs_16x[12]);
-               radeon_emit(cs, cm_sample_locs_16x[1]);
-               radeon_emit(cs, cm_sample_locs_16x[5]);
-               radeon_emit(cs, cm_sample_locs_16x[9]);
-               radeon_emit(cs, cm_sample_locs_16x[13]);
-               radeon_emit(cs, cm_sample_locs_16x[2]);
-               radeon_emit(cs, cm_sample_locs_16x[6]);
-               radeon_emit(cs, cm_sample_locs_16x[10]);
-               radeon_emit(cs, cm_sample_locs_16x[14]);
-               radeon_emit(cs, cm_sample_locs_16x[3]);
-               radeon_emit(cs, cm_sample_locs_16x[7]);
-               radeon_emit(cs, cm_sample_locs_16x[11]);
-               radeon_emit(cs, cm_sample_locs_16x[15]);
-               break;
-       }
-}
-
-void si_common_emit_msaa_config(struct radeon_winsys_cs *cs, int nr_samples,
-                               int ps_iter_samples, int overrast_samples,
-                               unsigned sc_mode_cntl_1)
-{
-       int setup_samples = nr_samples > 1 ? nr_samples :
-                           overrast_samples > 1 ? overrast_samples : 0;
-       /* Required by OpenGL line rasterization.
-        *
-        * TODO: We should also enable perpendicular endcaps for AA lines,
-        *       but that requires implementing line stippling in the pixel
-        *       shader. SC can only do line stippling with axis-aligned
-        *       endcaps.
-        */
-       unsigned sc_line_cntl = S_028BDC_DX10_DIAMOND_TEST_ENA(1);
-
-       if (setup_samples > 1) {
-               /* indexed by log2(nr_samples) */
-               unsigned max_dist[] = {
-                       0,
-                       eg_max_dist_2x,
-                       eg_max_dist_4x,
-                       cm_max_dist_8x,
-                       cm_max_dist_16x
-               };
-               unsigned log_samples = util_logbase2(setup_samples);
-               unsigned log_ps_iter_samples =
-                       util_logbase2(util_next_power_of_two(ps_iter_samples));
-
-               radeon_set_context_reg_seq(cs, CM_R_028BDC_PA_SC_LINE_CNTL, 2);
-               radeon_emit(cs, sc_line_cntl |
-                           S_028BDC_EXPAND_LINE_WIDTH(1)); /* 
CM_R_028BDC_PA_SC_LINE_CNTL */
-               radeon_emit(cs, S_028BE0_MSAA_NUM_SAMPLES(log_samples) |
-                           S_028BE0_MAX_SAMPLE_DIST(max_dist[log_samples]) |
-                           S_028BE0_MSAA_EXPOSED_SAMPLES(log_samples)); /* 
CM_R_028BE0_PA_SC_AA_CONFIG */
-
-               if (nr_samples > 1) {
-                       radeon_set_context_reg(cs, CM_R_028804_DB_EQAA,
-                                              
S_028804_MAX_ANCHOR_SAMPLES(log_samples) |
-                                              
S_028804_PS_ITER_SAMPLES(log_ps_iter_samples) |
-                                              
S_028804_MASK_EXPORT_NUM_SAMPLES(log_samples) |
-                                              
S_028804_ALPHA_TO_MASK_NUM_SAMPLES(log_samples) |
-                                              
S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
-                                              
S_028804_STATIC_ANCHOR_ASSOCIATIONS(1));
-                       radeon_set_context_reg(cs, 
EG_R_028A4C_PA_SC_MODE_CNTL_1,
-                                              
EG_S_028A4C_PS_ITER_SAMPLE(ps_iter_samples > 1) |
-                                              sc_mode_cntl_1);
-               } else if (overrast_samples > 1) {
-                       radeon_set_context_reg(cs, CM_R_028804_DB_EQAA,
-                                              
S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
-                                              
S_028804_STATIC_ANCHOR_ASSOCIATIONS(1) |
-                                              
S_028804_OVERRASTERIZATION_AMOUNT(log_samples));
-                       radeon_set_context_reg(cs, 
EG_R_028A4C_PA_SC_MODE_CNTL_1,
-                                              sc_mode_cntl_1);
-               }
-       } else {
-               radeon_set_context_reg_seq(cs, CM_R_028BDC_PA_SC_LINE_CNTL, 2);
-               radeon_emit(cs, sc_line_cntl); /* CM_R_028BDC_PA_SC_LINE_CNTL */
-               radeon_emit(cs, 0); /* CM_R_028BE0_PA_SC_AA_CONFIG */
-
-               radeon_set_context_reg(cs, CM_R_028804_DB_EQAA,
-                                      S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
-                                      S_028804_STATIC_ANCHOR_ASSOCIATIONS(1));
-               radeon_set_context_reg(cs, EG_R_028A4C_PA_SC_MODE_CNTL_1,
-                                      sc_mode_cntl_1);
-       }
-}
diff --git a/src/gallium/drivers/radeon/r600_pipe_common.c 
b/src/gallium/drivers/radeon/r600_pipe_common.c
index 262476f..f8a826f 100644
--- a/src/gallium/drivers/radeon/r600_pipe_common.c
+++ b/src/gallium/drivers/radeon/r600_pipe_common.c
@@ -631,21 +631,20 @@ bool si_common_context_init(struct r600_common_context 
*rctx,
                rctx->b.get_device_reset_status = r600_get_reset_status;
                rctx->gpu_reset_counter =
                        rctx->ws->query_value(rctx->ws,
                                              RADEON_GPU_RESET_COUNTER);
        }
 
        rctx->b.set_device_reset_callback = r600_set_device_reset_callback;
 
        si_init_context_texture_functions(rctx);
        si_init_query_functions(rctx);
-       si_init_msaa(&rctx->b);
 
        if (rctx->chip_class == CIK ||
            rctx->chip_class == VI ||
            rctx->chip_class == GFX9) {
                rctx->eop_bug_scratch = (struct r600_resource*)
                        pipe_buffer_create(&rscreen->b, 0, PIPE_USAGE_DEFAULT,
                                           16 * 
rscreen->info.num_render_backends);
                if (!rctx->eop_bug_scratch)
                        return false;
        }
diff --git a/src/gallium/drivers/radeon/r600_pipe_common.h 
b/src/gallium/drivers/radeon/r600_pipe_common.h
index b620e6b..a7c91cb 100644
--- a/src/gallium/drivers/radeon/r600_pipe_common.h
+++ b/src/gallium/drivers/radeon/r600_pipe_common.h
@@ -571,29 +571,20 @@ struct r600_common_context {
        uint64_t                        num_alloc_tex_transfer_bytes;
        unsigned                        last_tex_ps_draw_ratio; /* for query */
 
        /* Render condition. */
        struct r600_atom                render_cond_atom;
        struct pipe_query               *render_cond;
        unsigned                        render_cond_mode;
        bool                            render_cond_invert;
        bool                            render_cond_force_off; /* for u_blitter 
*/
 
-       /* MSAA sample locations.
-        * The first index is the sample index.
-        * The second index is the coordinate: X, Y. */
-       float                           sample_locations_1x[1][2];
-       float                           sample_locations_2x[2][2];
-       float                           sample_locations_4x[4][2];
-       float                           sample_locations_8x[8][2];
-       float                           sample_locations_16x[16][2];
-
        /* Statistics gathering for the DCC enablement heuristic. It can't be
         * in r600_texture because r600_texture can be shared by multiple
         * contexts. This is for back buffers only. We shouldn't get too many
         * of those.
         *
         * X11 DRI3 rotates among a finite set of back buffers. They should
         * all fit in this array. If they don't, separate DCC might never be
         * enabled by DCC stat gathering.
         */
        struct {
@@ -799,29 +790,20 @@ void vi_dcc_clear_level(struct r600_common_context *rctx,
 void si_do_fast_color_clear(struct r600_common_context *rctx,
                            struct pipe_framebuffer_state *fb,
                            struct r600_atom *fb_state,
                            unsigned *buffers, ubyte *dirty_cbufs,
                            const union pipe_color_union *color);
 bool si_texture_disable_dcc(struct r600_common_context *rctx,
                            struct r600_texture *rtex);
 void si_init_screen_texture_functions(struct r600_common_screen *rscreen);
 void si_init_context_texture_functions(struct r600_common_context *rctx);
 
-/* cayman_msaa.c */
-void si_get_sample_position(struct pipe_context *ctx, unsigned sample_count,
-                           unsigned sample_index, float *out_value);
-void si_init_msaa(struct pipe_context *ctx);
-void si_common_emit_msaa_sample_locs(struct radeon_winsys_cs *cs, int 
nr_samples);
-void si_common_emit_msaa_config(struct radeon_winsys_cs *cs, int nr_samples,
-                               int ps_iter_samples, int overrast_samples,
-                               unsigned sc_mode_cntl_1);
-
 
 /* Inline helpers. */
 
 static inline struct r600_resource *r600_resource(struct pipe_resource *r)
 {
        return (struct r600_resource*)r;
 }
 
 static inline void
 r600_resource_reference(struct r600_resource **ptr, struct r600_resource *res)
@@ -916,23 +898,16 @@ vi_tc_compat_htile_enabled(struct r600_texture *tex, 
unsigned level)
 }
 
 #define COMPUTE_DBG(rscreen, fmt, args...) \
        do { \
                if ((rscreen->b.debug_flags & DBG(COMPUTE))) fprintf(stderr, 
fmt, ##args); \
        } while (0);
 
 #define R600_ERR(fmt, args...) \
        fprintf(stderr, "EE %s:%d %s - " fmt, __FILE__, __LINE__, __func__, 
##args)
 
-/* For MSAA sample positions. */
-#define FILL_SREG(s0x, s0y, s1x, s1y, s2x, s2y, s3x, s3y)  \
-       (((s0x) & 0xf) | (((unsigned)(s0y) & 0xf) << 4) |                  \
-       (((unsigned)(s1x) & 0xf) << 8) | (((unsigned)(s1y) & 0xf) << 12) |      
   \
-       (((unsigned)(s2x) & 0xf) << 16) | (((unsigned)(s2y) & 0xf) << 20) |     
   \
-        (((unsigned)(s3x) & 0xf) << 24) | (((unsigned)(s3y) & 0xf) << 28))
-
 static inline int S_FIXED(float value, unsigned frac_bits)
 {
        return value * (1 << frac_bits);
 }
 
 #endif
diff --git a/src/gallium/drivers/radeonsi/Makefile.sources 
b/src/gallium/drivers/radeonsi/Makefile.sources
index 63cd7a3..2d746ab 100644
--- a/src/gallium/drivers/radeonsi/Makefile.sources
+++ b/src/gallium/drivers/radeonsi/Makefile.sources
@@ -22,15 +22,16 @@ C_SOURCES := \
        si_shader.c \
        si_shader.h \
        si_shader_internal.h \
        si_shader_nir.c \
        si_shader_tgsi_alu.c \
        si_shader_tgsi_mem.c \
        si_shader_tgsi_setup.c \
        si_state.c \
        si_state_binning.c \
        si_state_draw.c \
+       si_state_msaa.c \
        si_state_shaders.c \
        si_state_streamout.c \
        si_state_viewport.c \
        si_state.h \
        si_uvd.c
diff --git a/src/gallium/drivers/radeonsi/si_pipe.c 
b/src/gallium/drivers/radeonsi/si_pipe.c
index b9840ad..4384987 100644
--- a/src/gallium/drivers/radeonsi/si_pipe.c
+++ b/src/gallium/drivers/radeonsi/si_pipe.c
@@ -198,20 +198,21 @@ static struct pipe_context *si_create_context(struct 
pipe_screen *screen,
        if (!si_common_context_init(&sctx->b, &sscreen->b, flags))
                goto fail;
 
        if (sscreen->b.info.drm_major == 3)
                sctx->b.b.get_device_reset_status = si_amdgpu_get_reset_status;
 
        si_init_blit_functions(sctx);
        si_init_compute_functions(sctx);
        si_init_cp_dma_functions(sctx);
        si_init_debug_functions(sctx);
+       si_init_msaa_functions(sctx);
        si_init_streamout_functions(sctx);
 
        if (sscreen->b.info.has_hw_decode) {
                sctx->b.b.create_video_codec = si_uvd_create_decoder;
                sctx->b.b.create_video_buffer = si_video_buffer_create;
        } else {
                sctx->b.b.create_video_codec = vl_create_decoder;
                sctx->b.b.create_video_buffer = vl_video_buffer_create;
        }
 
diff --git a/src/gallium/drivers/radeonsi/si_pipe.h 
b/src/gallium/drivers/radeonsi/si_pipe.h
index 4e54b7e..c162a0f 100644
--- a/src/gallium/drivers/radeonsi/si_pipe.h
+++ b/src/gallium/drivers/radeonsi/si_pipe.h
@@ -536,20 +536,29 @@ struct si_context {
        struct util_dynarray    resident_img_handles;
 
        /* Resident bindless handles which need decompression */
        struct util_dynarray    resident_tex_needs_color_decompress;
        struct util_dynarray    resident_img_needs_color_decompress;
        struct util_dynarray    resident_tex_needs_depth_decompress;
 
        /* Bindless state */
        bool                    uses_bindless_samplers;
        bool                    uses_bindless_images;
+
+       /* MSAA sample locations.
+        * The first index is the sample index.
+        * The second index is the coordinate: X, Y. */
+       float                   sample_locations_1x[1][2];
+       float                   sample_locations_2x[2][2];
+       float                   sample_locations_4x[4][2];
+       float                   sample_locations_8x[8][2];
+       float                   sample_locations_16x[16][2];
 };
 
 /* cik_sdma.c */
 void cik_init_sdma_functions(struct si_context *sctx);
 
 /* si_blit.c */
 void si_init_blit_functions(struct si_context *sctx);
 void si_decompress_textures(struct si_context *sctx, unsigned shader_mask);
 void si_resource_copy_region(struct pipe_context *ctx,
                             struct pipe_resource *dst,
diff --git a/src/gallium/drivers/radeonsi/si_state.c 
b/src/gallium/drivers/radeonsi/si_state.c
index 82f3962..3c6b7ca 100644
--- a/src/gallium/drivers/radeonsi/si_state.c
+++ b/src/gallium/drivers/radeonsi/si_state.c
@@ -2879,33 +2879,33 @@ static void si_set_framebuffer_state(struct 
pipe_context *ctx,
             (zstex && zstex->surface.has_stencil != old_has_stencil)))
                si_mark_atom_dirty(sctx, &sctx->msaa_config);
 
        if (sctx->framebuffer.nr_samples != old_nr_samples) {
                si_mark_atom_dirty(sctx, &sctx->msaa_config);
                si_mark_atom_dirty(sctx, &sctx->db_render_state);
 
                /* Set sample locations as fragment shader constants. */
                switch (sctx->framebuffer.nr_samples) {
                case 1:
-                       constbuf.user_buffer = sctx->b.sample_locations_1x;
+                       constbuf.user_buffer = sctx->sample_locations_1x;
                        break;
                case 2:
-                       constbuf.user_buffer = sctx->b.sample_locations_2x;
+                       constbuf.user_buffer = sctx->sample_locations_2x;
                        break;
                case 4:
-                       constbuf.user_buffer = sctx->b.sample_locations_4x;
+                       constbuf.user_buffer = sctx->sample_locations_4x;
                        break;
                case 8:
-                       constbuf.user_buffer = sctx->b.sample_locations_8x;
+                       constbuf.user_buffer = sctx->sample_locations_8x;
                        break;
                case 16:
-                       constbuf.user_buffer = sctx->b.sample_locations_16x;
+                       constbuf.user_buffer = sctx->sample_locations_16x;
                        break;
                default:
                        R600_ERR("Requested an invalid number of samples %i.\n",
                                 sctx->framebuffer.nr_samples);
                        assert(0);
                }
                constbuf.buffer_size = sctx->framebuffer.nr_samples * 2 * 4;
                si_set_rw_buffer(sctx, SI_PS_CONST_SAMPLE_POSITIONS, &constbuf);
 
                si_mark_atom_dirty(sctx, &sctx->msaa_sample_locs.atom);
@@ -3184,21 +3184,21 @@ static void si_emit_msaa_sample_locs(struct si_context 
*sctx,
                nr_samples = SI_NUM_SMOOTH_AA_SAMPLES;
 
        /* On Polaris, the small primitive filter uses the sample locations
         * even when MSAA is off, so we need to make sure they're set to 0.
         */
        if (has_msaa_sample_loc_bug)
                nr_samples = MAX2(nr_samples, 1);
 
        if (nr_samples != sctx->msaa_sample_locs.nr_samples) {
                sctx->msaa_sample_locs.nr_samples = nr_samples;
-               si_common_emit_msaa_sample_locs(cs, nr_samples);
+               si_emit_sample_locations(cs, nr_samples);
        }
 
        if (sctx->b.family >= CHIP_POLARIS10) {
                struct si_state_rasterizer *rs = sctx->queued.named.rasterizer;
                unsigned small_prim_filter_cntl =
                        S_028830_SMALL_PRIM_FILTER_ENABLE(1) |
                        /* line bug */
                        S_028830_LINE_FILTER_DISABLE(sctx->b.family <= 
CHIP_POLARIS12);
 
                /* The alternative of setting sample locations to 0 would
@@ -3296,24 +3296,82 @@ static void si_emit_msaa_config(struct si_context 
*sctx, struct r600_atom *atom)
                S_028A4C_OUT_OF_ORDER_PRIMITIVE_ENABLE(out_of_order_rast) |
                S_028A4C_OUT_OF_ORDER_WATER_MARK(0x7) |
                /* always 1: */
                S_028A4C_WALK_ALIGN8_PRIM_FITS_ST(1) |
                S_028A4C_SUPERTILE_WALK_ORDER_ENABLE(1) |
                S_028A4C_TILE_WALK_ORDER_ENABLE(1) |
                S_028A4C_MULTI_SHADER_ENGINE_PRIM_DISCARD_ENABLE(1) |
                S_028A4C_FORCE_EOV_CNTDWN_ENABLE(1) |
                S_028A4C_FORCE_EOV_REZ_ENABLE(1);
 
-       si_common_emit_msaa_config(cs, sctx->framebuffer.nr_samples,
-                               sctx->ps_iter_samples,
-                               sctx->smoothing_enabled ? 
SI_NUM_SMOOTH_AA_SAMPLES : 0,
-                               sc_mode_cntl_1);
+       int setup_samples = sctx->framebuffer.nr_samples > 1 ? 
sctx->framebuffer.nr_samples :
+                           sctx->smoothing_enabled ? SI_NUM_SMOOTH_AA_SAMPLES 
: 0;
+
+       /* Required by OpenGL line rasterization.
+        *
+        * TODO: We should also enable perpendicular endcaps for AA lines,
+        *       but that requires implementing line stippling in the pixel
+        *       shader. SC can only do line stippling with axis-aligned
+        *       endcaps.
+        */
+       unsigned sc_line_cntl = S_028BDC_DX10_DIAMOND_TEST_ENA(1);
+
+       if (setup_samples > 1) {
+               /* distance from the pixel center, indexed by log2(nr_samples) 
*/
+               static unsigned max_dist[] = {
+                       0, /* unused */
+                       4, /* 2x MSAA */
+                       6, /* 4x MSAA */
+                       7, /* 8x MSAA */
+                       8, /* 16x MSAA */
+               };
+               unsigned log_samples = util_logbase2(setup_samples);
+               unsigned log_ps_iter_samples =
+                       
util_logbase2(util_next_power_of_two(sctx->ps_iter_samples));
+
+               radeon_set_context_reg_seq(cs, CM_R_028BDC_PA_SC_LINE_CNTL, 2);
+               radeon_emit(cs, sc_line_cntl |
+                           S_028BDC_EXPAND_LINE_WIDTH(1)); /* 
CM_R_028BDC_PA_SC_LINE_CNTL */
+               radeon_emit(cs, S_028BE0_MSAA_NUM_SAMPLES(log_samples) |
+                           S_028BE0_MAX_SAMPLE_DIST(max_dist[log_samples]) |
+                           S_028BE0_MSAA_EXPOSED_SAMPLES(log_samples)); /* 
CM_R_028BE0_PA_SC_AA_CONFIG */
+
+               if (sctx->framebuffer.nr_samples > 1) {
+                       radeon_set_context_reg(cs, CM_R_028804_DB_EQAA,
+                                              
S_028804_MAX_ANCHOR_SAMPLES(log_samples) |
+                                              
S_028804_PS_ITER_SAMPLES(log_ps_iter_samples) |
+                                              
S_028804_MASK_EXPORT_NUM_SAMPLES(log_samples) |
+                                              
S_028804_ALPHA_TO_MASK_NUM_SAMPLES(log_samples) |
+                                              
S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
+                                              
S_028804_STATIC_ANCHOR_ASSOCIATIONS(1));
+                       radeon_set_context_reg(cs, 
EG_R_028A4C_PA_SC_MODE_CNTL_1,
+                                              
EG_S_028A4C_PS_ITER_SAMPLE(sctx->ps_iter_samples > 1) |
+                                              sc_mode_cntl_1);
+               } else if (sctx->smoothing_enabled) {
+                       radeon_set_context_reg(cs, CM_R_028804_DB_EQAA,
+                                              
S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
+                                              
S_028804_STATIC_ANCHOR_ASSOCIATIONS(1) |
+                                              
S_028804_OVERRASTERIZATION_AMOUNT(log_samples));
+                       radeon_set_context_reg(cs, 
EG_R_028A4C_PA_SC_MODE_CNTL_1,
+                                              sc_mode_cntl_1);
+               }
+       } else {
+               radeon_set_context_reg_seq(cs, CM_R_028BDC_PA_SC_LINE_CNTL, 2);
+               radeon_emit(cs, sc_line_cntl); /* CM_R_028BDC_PA_SC_LINE_CNTL */
+               radeon_emit(cs, 0); /* CM_R_028BE0_PA_SC_AA_CONFIG */
+
+               radeon_set_context_reg(cs, CM_R_028804_DB_EQAA,
+                                      S_028804_HIGH_QUALITY_INTERSECTIONS(1) |
+                                      S_028804_STATIC_ANCHOR_ASSOCIATIONS(1));
+               radeon_set_context_reg(cs, EG_R_028A4C_PA_SC_MODE_CNTL_1,
+                                      sc_mode_cntl_1);
+       }
 
        /* GFX9: Flush DFSM when the AA mode changes. */
        if (sctx->screen->dfsm_allowed) {
                radeon_emit(cs, PKT3(PKT3_EVENT_WRITE, 0, 0));
                radeon_emit(cs, EVENT_TYPE(V_028A90_FLUSH_DFSM) | 
EVENT_INDEX(0));
        }
 }
 
 static void si_set_min_samples(struct pipe_context *ctx, unsigned min_samples)
 {
@@ -4440,21 +4498,20 @@ void si_init_state_functions(struct si_context *sctx)
        sctx->custom_dsa_flush = si_create_db_flush_dsa(sctx);
        sctx->custom_blend_resolve = si_create_blend_custom(sctx, 
V_028808_CB_RESOLVE);
        sctx->custom_blend_fmask_decompress = si_create_blend_custom(sctx, 
V_028808_CB_FMASK_DECOMPRESS);
        sctx->custom_blend_eliminate_fastclear = si_create_blend_custom(sctx, 
V_028808_CB_ELIMINATE_FAST_CLEAR);
        sctx->custom_blend_dcc_decompress = si_create_blend_custom(sctx, 
V_028808_CB_DCC_DECOMPRESS);
 
        sctx->b.b.set_clip_state = si_set_clip_state;
        sctx->b.b.set_stencil_ref = si_set_stencil_ref;
 
        sctx->b.b.set_framebuffer_state = si_set_framebuffer_state;
-       sctx->b.b.get_sample_position = si_get_sample_position;
 
        sctx->b.b.create_sampler_state = si_create_sampler_state;
        sctx->b.b.delete_sampler_state = si_delete_sampler_state;
 
        sctx->b.b.create_sampler_view = si_create_sampler_view;
        sctx->b.b.sampler_view_destroy = si_sampler_view_destroy;
 
        sctx->b.b.set_sample_mask = si_set_sample_mask;
 
        sctx->b.b.create_vertex_elements_state = si_create_vertex_elements;
diff --git a/src/gallium/drivers/radeonsi/si_state.h 
b/src/gallium/drivers/radeonsi/si_state.h
index a686d0d..8122ddb 100644
--- a/src/gallium/drivers/radeonsi/si_state.h
+++ b/src/gallium/drivers/radeonsi/si_state.h
@@ -416,20 +416,24 @@ void si_emit_cache_flush(struct si_context *sctx);
 void si_draw_vbo(struct pipe_context *ctx, const struct pipe_draw_info *dinfo);
 void si_draw_rectangle(struct blitter_context *blitter,
                       void *vertex_elements_cso,
                       blitter_get_vs_func get_vs,
                       int x1, int y1, int x2, int y2,
                       float depth, unsigned num_instances,
                       enum blitter_attrib_type type,
                       const union blitter_attrib *attrib);
 void si_trace_emit(struct si_context *sctx);
 
+/* si_state_msaa.c */
+void si_init_msaa_functions(struct si_context *sctx);
+void si_emit_sample_locations(struct radeon_winsys_cs *cs, int nr_samples);
+
 /* si_state_streamout.c */
 void si_streamout_buffers_dirty(struct si_context *sctx);
 void si_emit_streamout_end(struct si_context *sctx);
 void si_update_prims_generated_query_state(struct si_context *sctx,
                                           unsigned type, int diff);
 void si_init_streamout_functions(struct si_context *sctx);
 
 
 static inline unsigned
 si_tile_mode_index(struct r600_texture *rtex, unsigned level, bool stencil)
diff --git a/src/gallium/drivers/radeonsi/si_state_msaa.c 
b/src/gallium/drivers/radeonsi/si_state_msaa.c
new file mode 100644
index 0000000..977f54b
--- /dev/null
+++ b/src/gallium/drivers/radeonsi/si_state_msaa.c
@@ -0,0 +1,209 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING 
FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN 
THE
+ * SOFTWARE.
+ *
+ * Authors: Marek Olšák <[email protected]>
+ *
+ */
+
+#include "si_pipe.h"
+#include "sid.h"
+#include "radeon/r600_cs.h"
+
+/* For MSAA sample positions. */
+#define FILL_SREG(s0x, s0y, s1x, s1y, s2x, s2y, s3x, s3y)  \
+       (((s0x) & 0xf) | (((unsigned)(s0y) & 0xf) << 4) |                  \
+       (((unsigned)(s1x) & 0xf) << 8) | (((unsigned)(s1y) & 0xf) << 12) |      
   \
+       (((unsigned)(s2x) & 0xf) << 16) | (((unsigned)(s2y) & 0xf) << 20) |     
   \
+        (((unsigned)(s3x) & 0xf) << 24) | (((unsigned)(s3y) & 0xf) << 28))
+
+/* 2xMSAA
+ * There are two locations (4, 4), (-4, -4). */
+static const uint32_t sample_locs_2x[4] = {
+       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
+       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
+       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
+       FILL_SREG(4, 4, -4, -4, 4, 4, -4, -4),
+};
+/* 4xMSAA
+ * There are 4 locations: (-2, -6), (6, -2), (-6, 2), (2, 6). */
+static const uint32_t sample_locs_4x[4] = {
+       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
+       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
+       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
+       FILL_SREG(-2, -6, 6, -2, -6, 2, 2, 6),
+};
+
+/* Cayman 8xMSAA */
+static const uint32_t sample_locs_8x[] = {
+       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
+       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
+       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
+       FILL_SREG( 1, -3, -1,  3, 5,  1, -3, -5),
+       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
+       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
+       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
+       FILL_SREG(-5,  5, -7, -1, 3,  7,  7, -7),
+};
+/* Cayman 16xMSAA */
+static const uint32_t sample_locs_16x[] = {
+       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
+       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
+       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
+       FILL_SREG( 1,  1, -1, -3, -3,  2,  4, -1),
+       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
+       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
+       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
+       FILL_SREG(-5, -2,  2,  5,  5,  3,  3, -5),
+       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
+       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
+       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
+       FILL_SREG(-2,  6,  0, -7, -4, -6, -6,  4),
+       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
+       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
+       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
+       FILL_SREG(-8,  0,  7, -4,  6,  7, -7, -8),
+};
+
+static void si_get_sample_position(struct pipe_context *ctx, unsigned 
sample_count,
+                                  unsigned sample_index, float *out_value)
+{
+       int offset, index;
+       struct {
+               int idx:4;
+       } val;
+
+       switch (sample_count) {
+       case 1:
+       default:
+               out_value[0] = out_value[1] = 0.5;
+               break;
+       case 2:
+               offset = 4 * (sample_index * 2);
+               val.idx = (sample_locs_2x[0] >> offset) & 0xf;
+               out_value[0] = (float)(val.idx + 8) / 16.0f;
+               val.idx = (sample_locs_2x[0] >> (offset + 4)) & 0xf;
+               out_value[1] = (float)(val.idx + 8) / 16.0f;
+               break;
+       case 4:
+               offset = 4 * (sample_index * 2);
+               val.idx = (sample_locs_4x[0] >> offset) & 0xf;
+               out_value[0] = (float)(val.idx + 8) / 16.0f;
+               val.idx = (sample_locs_4x[0] >> (offset + 4)) & 0xf;
+               out_value[1] = (float)(val.idx + 8) / 16.0f;
+               break;
+       case 8:
+               offset = 4 * (sample_index % 4 * 2);
+               index = (sample_index / 4) * 4;
+               val.idx = (sample_locs_8x[index] >> offset) & 0xf;
+               out_value[0] = (float)(val.idx + 8) / 16.0f;
+               val.idx = (sample_locs_8x[index] >> (offset + 4)) & 0xf;
+               out_value[1] = (float)(val.idx + 8) / 16.0f;
+               break;
+       case 16:
+               offset = 4 * (sample_index % 4 * 2);
+               index = (sample_index / 4) * 4;
+               val.idx = (sample_locs_16x[index] >> offset) & 0xf;
+               out_value[0] = (float)(val.idx + 8) / 16.0f;
+               val.idx = (sample_locs_16x[index] >> (offset + 4)) & 0xf;
+               out_value[1] = (float)(val.idx + 8) / 16.0f;
+               break;
+       }
+}
+
+void si_emit_sample_locations(struct radeon_winsys_cs *cs, int nr_samples)
+{
+       switch (nr_samples) {
+       default:
+       case 1:
+               radeon_set_context_reg(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 0);
+               radeon_set_context_reg(cs, 
CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, 0);
+               radeon_set_context_reg(cs, 
CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, 0);
+               radeon_set_context_reg(cs, 
CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, 0);
+               break;
+       case 2:
+               radeon_set_context_reg(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, sample_locs_2x[0]);
+               radeon_set_context_reg(cs, 
CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, sample_locs_2x[1]);
+               radeon_set_context_reg(cs, 
CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, sample_locs_2x[2]);
+               radeon_set_context_reg(cs, 
CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, sample_locs_2x[3]);
+               break;
+       case 4:
+               radeon_set_context_reg(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, sample_locs_4x[0]);
+               radeon_set_context_reg(cs, 
CM_R_028C08_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0, sample_locs_4x[1]);
+               radeon_set_context_reg(cs, 
CM_R_028C18_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0, sample_locs_4x[2]);
+               radeon_set_context_reg(cs, 
CM_R_028C28_PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0, sample_locs_4x[3]);
+               break;
+       case 8:
+               radeon_set_context_reg_seq(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 14);
+               radeon_emit(cs, sample_locs_8x[0]);
+               radeon_emit(cs, sample_locs_8x[4]);
+               radeon_emit(cs, 0);
+               radeon_emit(cs, 0);
+               radeon_emit(cs, sample_locs_8x[1]);
+               radeon_emit(cs, sample_locs_8x[5]);
+               radeon_emit(cs, 0);
+               radeon_emit(cs, 0);
+               radeon_emit(cs, sample_locs_8x[2]);
+               radeon_emit(cs, sample_locs_8x[6]);
+               radeon_emit(cs, 0);
+               radeon_emit(cs, 0);
+               radeon_emit(cs, sample_locs_8x[3]);
+               radeon_emit(cs, sample_locs_8x[7]);
+               break;
+       case 16:
+               radeon_set_context_reg_seq(cs, 
CM_R_028BF8_PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0, 16);
+               radeon_emit(cs, sample_locs_16x[0]);
+               radeon_emit(cs, sample_locs_16x[4]);
+               radeon_emit(cs, sample_locs_16x[8]);
+               radeon_emit(cs, sample_locs_16x[12]);
+               radeon_emit(cs, sample_locs_16x[1]);
+               radeon_emit(cs, sample_locs_16x[5]);
+               radeon_emit(cs, sample_locs_16x[9]);
+               radeon_emit(cs, sample_locs_16x[13]);
+               radeon_emit(cs, sample_locs_16x[2]);
+               radeon_emit(cs, sample_locs_16x[6]);
+               radeon_emit(cs, sample_locs_16x[10]);
+               radeon_emit(cs, sample_locs_16x[14]);
+               radeon_emit(cs, sample_locs_16x[3]);
+               radeon_emit(cs, sample_locs_16x[7]);
+               radeon_emit(cs, sample_locs_16x[11]);
+               radeon_emit(cs, sample_locs_16x[15]);
+               break;
+       }
+}
+
+void si_init_msaa_functions(struct si_context *sctx)
+{
+       int i;
+
+       sctx->b.b.get_sample_position = si_get_sample_position;
+
+       si_get_sample_position(&sctx->b.b, 1, 0, sctx->sample_locations_1x[0]);
+
+       for (i = 0; i < 2; i++)
+               si_get_sample_position(&sctx->b.b, 2, i, 
sctx->sample_locations_2x[i]);
+       for (i = 0; i < 4; i++)
+               si_get_sample_position(&sctx->b.b, 4, i, 
sctx->sample_locations_4x[i]);
+       for (i = 0; i < 8; i++)
+               si_get_sample_position(&sctx->b.b, 8, i, 
sctx->sample_locations_8x[i]);
+       for (i = 0; i < 16; i++)
+               si_get_sample_position(&sctx->b.b, 16, i, 
sctx->sample_locations_16x[i]);
+}
-- 
2.7.4

_______________________________________________
mesa-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Reply via email to