radeonsi: Program RASTER_CONFIG for harvested GPUs v5

Harvested GPUs have some of their render backends disabled, so
in order to prevent the hardware from trying to render things
with these disabled backends we need to correctly program
the PA_SC_RASTER_CONFIG register.

v2:
  - Write RASTER_CONFIG for all SEs.

v3:
  - Set GRBM_GFX_INDEX.INSTANCE_BROADCAST_WRITES bit.
  - Set GRBM_GFX_INFEX.SH_BROADCAST_WRITES bit when done setting
    PA_SC_RASTER_CONFIG.
  - Get num_se and num_sh_per_se from kernel.

v4:
  - Get correct value for num_se
  - Remove loop for setting PA_SC_RASTER_CONFIG
  - Only compute raster config when a backend has been disabled.

v5: Michel Dänzer
  - Fix computation for chips with multiple SEs

https://bugs.freedesktop.org/show_bug.cgi?id=60879

CC: "10.4 10.3" <[email protected]>

2 files changed, 130 insertions, 6 deletions

diff --git a/src/gallium/drivers/radeonsi/si_state.c b/src/gallium/drivers/radeonsi/si_state.c
index ea8e61a0419..f24c28e9220 100644
--- a/src/gallium/drivers/radeonsi/si_state.c
+++ b/src/gallium/drivers/radeonsi/si_state.c
@@ -3081,6 +3081,110 @@ void si_init_state_functions(struct si_context *sctx)
 	sctx->b.b.draw_vbo = si_draw_vbo;
 }
 
+static void
+si_write_harvested_raster_configs(struct si_context *sctx,
+				  struct si_pm4_state *pm4,
+				  unsigned raster_config)
+{
+	unsigned sh_per_se = MAX2(sctx->screen->b.info.max_sh_per_se, 1);
+	unsigned num_se = MAX2(sctx->screen->b.info.max_se, 1);
+	unsigned rb_mask = sctx->screen->b.info.si_backend_enabled_mask;
+	unsigned num_rb = sctx->screen->b.info.r600_num_backends;
+	unsigned rb_per_pkr = num_rb / num_se / sh_per_se;
+	unsigned rb_per_se = num_rb / num_se;
+	unsigned se0_mask = (1 << rb_per_se) - 1;
+	unsigned se1_mask = se0_mask << rb_per_se;
+	unsigned se;
+
+	assert(num_se == 1 || num_se == 2);
+	assert(sh_per_se == 1 || sh_per_se == 2);
+	assert(rb_per_pkr == 1 || rb_per_pkr == 2);
+
+	/* XXX: I can't figure out what the *_XSEL and *_YSEL
+	 * fields are for, so I'm leaving them as their default
+	 * values. */
+
+	se0_mask &= rb_mask;
+	se1_mask &= rb_mask;
+	if (num_se == 2 && (!se0_mask || !se1_mask)) {
+		raster_config &= C_028350_SE_MAP;
+
+		if (!se0_mask) {
+			raster_config |=
+				S_028350_SE_MAP(V_028350_RASTER_CONFIG_SE_MAP_3);
+		} else {
+			raster_config |=
+				S_028350_SE_MAP(V_028350_RASTER_CONFIG_SE_MAP_0);
+		}
+	}
+
+	for (se = 0; se < num_se; se++) {
+		unsigned raster_config_se = raster_config;
+		unsigned pkr0_mask = ((1 << rb_per_pkr) - 1) << (se * rb_per_se);
+		unsigned pkr1_mask = pkr0_mask << rb_per_pkr;
+
+		pkr0_mask &= rb_mask;
+		pkr1_mask &= rb_mask;
+		if (sh_per_se == 2 && (!pkr0_mask || !pkr1_mask)) {
+			raster_config_se &= C_028350_PKR_MAP;
+
+			if (!pkr0_mask) {
+				raster_config_se |=
+					S_028350_PKR_MAP(V_028350_RASTER_CONFIG_PKR_MAP_3);
+			} else {
+				raster_config_se |=
+					S_028350_PKR_MAP(V_028350_RASTER_CONFIG_PKR_MAP_0);
+			}
+		}
+
+		if (rb_per_pkr == 2) {
+			unsigned rb0_mask = 1 << (se * rb_per_se);
+			unsigned rb1_mask = rb0_mask << 1;
+
+			rb0_mask &= rb_mask;
+			rb1_mask &= rb_mask;
+			if (!rb0_mask || !rb1_mask) {
+				raster_config_se &= C_028350_RB_MAP_PKR0;
+
+				if (!rb0_mask) {
+					raster_config_se |=
+						S_028350_RB_MAP_PKR0(V_028350_RASTER_CONFIG_RB_MAP_3);
+				} else {
+					raster_config_se |=
+						S_028350_RB_MAP_PKR0(V_028350_RASTER_CONFIG_RB_MAP_0);
+				}
+			}
+
+			if (sh_per_se == 2) {
+				rb0_mask = 1 << (se * rb_per_se + rb_per_pkr);
+				rb1_mask = rb0_mask << 1;
+				rb0_mask &= rb_mask;
+				rb1_mask &= rb_mask;
+				if (!rb0_mask || !rb1_mask) {
+					raster_config_se &= C_028350_RB_MAP_PKR1;
+
+					if (!rb0_mask) {
+						raster_config_se |=
+							S_028350_RB_MAP_PKR1(V_028350_RASTER_CONFIG_RB_MAP_3);
+					} else {
+						raster_config_se |=
+							S_028350_RB_MAP_PKR1(V_028350_RASTER_CONFIG_RB_MAP_0);
+					}
+				}
+			}
+		}
+
+		si_pm4_set_reg(pm4, GRBM_GFX_INDEX,
+			       SE_INDEX(se) | SH_BROADCAST_WRITES |
+			       INSTANCE_BROADCAST_WRITES);
+		si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG, raster_config_se);
+	}
+
+	si_pm4_set_reg(pm4, GRBM_GFX_INDEX,
+		       SE_BROADCAST_WRITES | SH_BROADCAST_WRITES |
+		       INSTANCE_BROADCAST_WRITES);
+}
+
 void si_init_config(struct si_context *sctx)
 {
 	struct si_pm4_state *pm4 = CALLOC_STRUCT(si_pm4_state);
@@ -3152,24 +3256,38 @@ void si_init_config(struct si_context *sctx)
 			break;
 		}
 	} else {
+		unsigned rb_mask = sctx->screen->b.info.si_backend_enabled_mask;
+		unsigned num_rb = sctx->screen->b.info.r600_num_backends;
+		unsigned raster_config;
+
 		switch (sctx->screen->b.family) {
 		case CHIP_TAHITI:
 		case CHIP_PITCAIRN:
-			si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG, 0x2a00126a);
+			raster_config = 0x2a00126a;
 			break;
 		case CHIP_VERDE:
-			si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG, 0x0000124a);
+			raster_config = 0x0000124a;
 			break;
 		case CHIP_OLAND:
-			si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG, 0x00000082);
+			raster_config = 0x00000082;
 			break;
 		case CHIP_HAINAN:
-			si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG, 0x00000000);
+			raster_config = 0x00000000;
 			break;
 		default:
-			si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG, 0x00000000);
+			fprintf(stderr,
+				"radeonsi: Unknown GPU, using 0 for raster_config\n");
+			raster_config = 0x00000000;
 			break;
 		}
+
+		/* Always use the default config when all backends are enabled. */
+		if (rb_mask && util_bitcount(rb_mask) >= num_rb) {
+			si_pm4_set_reg(pm4, R_028350_PA_SC_RASTER_CONFIG,
+				       raster_config);
+		} else {
+			si_write_harvested_raster_configs(sctx, pm4, raster_config);
+		}
 	}
 
 	si_pm4_set_reg(pm4, R_028204_PA_SC_WINDOW_SCISSOR_TL, S_028204_WINDOW_OFFSET_DISABLE(1));
diff --git a/src/gallium/drivers/radeonsi/sid.h b/src/gallium/drivers/radeonsi/sid.h
index c4d4afd6f02..c78ba7c2991 100644
--- a/src/gallium/drivers/radeonsi/sid.h
+++ b/src/gallium/drivers/radeonsi/sid.h
@@ -204,7 +204,13 @@
  * 6. COMMAND [29:22] | BYTE_COUNT [20:0]
  */
 
-
+#define GRBM_GFX_INDEX                                                  0x802C
+#define         INSTANCE_INDEX(x)                                     ((x) << 0)
+#define         SH_INDEX(x)                                           ((x) << 8)
+#define         SE_INDEX(x)                                           ((x) << 16)
+#define         SH_BROADCAST_WRITES                                   (1 << 29)
+#define         INSTANCE_BROADCAST_WRITES                             (1 << 30)
+#define         SE_BROADCAST_WRITES                                   (1 << 31)
 #define R_0084FC_CP_STRMOUT_CNTL		                        0x0084FC
 #define   S_0084FC_OFFSET_UPDATE_DONE(x)		              (((x) & 0x1) << 0)
 #define R_0085F0_CP_COHER_CNTL                                          0x0085F0