aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorDave Airlie <airlied@redhat.com>2009-09-01 01:25:57 -0400
committerDave Airlie <airlied@redhat.com>2009-09-07 18:54:31 -0400
commit551ebd837c75fc75df81811a18b7136c39cab487 (patch)
tree9703fd46cf9ad170012754f984375db37d2bf818
parent11670d3c93210793562748d83502ecbef4034765 (diff)
drm/radeon/kms: add rn50/r100/r200 CS tracker.
This adds the command stream checker for the RN50, R100 and R200 cards. It stops any access to 3D registers on RN50, and does checks on buffer sizes on the r100/r200 cards. It also fixes some texture sizing checks on r300. Signed-off-by: Dave Airlie <airlied@redhat.com>
Diffstat
-rw-r--r--drivers/gpu/drm/radeon/Makefile15
-rw-r--r--drivers/gpu/drm/radeon/r100.c811
-rw-r--r--drivers/gpu/drm/radeon/r100_track.h124
-rw-r--r--drivers/gpu/drm/radeon/r200.c456
-rw-r--r--drivers/gpu/drm/radeon/r300.c363
-rw-r--r--drivers/gpu/drm/radeon/radeon.h6
-rw-r--r--drivers/gpu/drm/radeon/radeon_asic.h1
-rw-r--r--drivers/gpu/drm/radeon/radeon_reg.h57
-rw-r--r--drivers/gpu/drm/radeon/reg_srcs/r100105
-rw-r--r--drivers/gpu/drm/radeon/reg_srcs/r200184
-rw-r--r--drivers/gpu/drm/radeon/reg_srcs/rn5030
11 files changed, 1712 insertions, 440 deletions
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index b2213a576a85..6fb842962490 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -11,6 +11,15 @@ hostprogs-y := mkregtable
11quiet_cmd_mkregtable = MKREGTABLE $@ 11quiet_cmd_mkregtable = MKREGTABLE $@
12 cmd_mkregtable = $(obj)/mkregtable $< > $@ 12 cmd_mkregtable = $(obj)/mkregtable $< > $@
13 13
14$(obj)/rn50_reg_safe.h: $(src)/reg_srcs/rn50 $(obj)/mkregtable
15 $(call if_changed,mkregtable)
16
17$(obj)/r100_reg_safe.h: $(src)/reg_srcs/r100 $(obj)/mkregtable
18 $(call if_changed,mkregtable)
19
20$(obj)/r200_reg_safe.h: $(src)/reg_srcs/r200 $(obj)/mkregtable
21 $(call if_changed,mkregtable)
22
14$(obj)/rv515_reg_safe.h: $(src)/reg_srcs/rv515 $(obj)/mkregtable 23$(obj)/rv515_reg_safe.h: $(src)/reg_srcs/rv515 $(obj)/mkregtable
15 $(call if_changed,mkregtable) 24 $(call if_changed,mkregtable)
16 25
@@ -20,6 +29,10 @@ $(obj)/r300_reg_safe.h: $(src)/reg_srcs/r300 $(obj)/mkregtable
20$(obj)/rs600_reg_safe.h: $(src)/reg_srcs/rs600 $(obj)/mkregtable 29$(obj)/rs600_reg_safe.h: $(src)/reg_srcs/rs600 $(obj)/mkregtable
21 $(call if_changed,mkregtable) 30 $(call if_changed,mkregtable)
22 31
32$(obj)/r100.o: $(obj)/r100_reg_safe.h $(obj)/rn50_reg_safe.h
33
34$(obj)/r200.o: $(obj)/r200_reg_safe.h
35
23$(obj)/rv515.o: $(obj)/rv515_reg_safe.h 36$(obj)/rv515.o: $(obj)/rv515_reg_safe.h
24 37
25$(obj)/r300.o: $(obj)/r300_reg_safe.h 38$(obj)/r300.o: $(obj)/r300_reg_safe.h
@@ -34,7 +47,7 @@ radeon-$(CONFIG_DRM_RADEON_KMS) += radeon_device.o radeon_kms.o \
34 radeon_clocks.o radeon_fb.o radeon_gem.o radeon_ring.o radeon_irq_kms.o \ 47 radeon_clocks.o radeon_fb.o radeon_gem.o radeon_ring.o radeon_irq_kms.o \
35 radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \ 48 radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \
36 rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o \ 49 rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o \
37 radeon_test.o 50 radeon_test.o r200.o
38 51
39radeon-$(CONFIG_COMPAT) += radeon_ioc32.o 52radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
40 53
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 44f34f8e2b32..ee3ab62417e2 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -34,6 +34,9 @@
34#include <linux/firmware.h> 34#include <linux/firmware.h>
35#include <linux/platform_device.h> 35#include <linux/platform_device.h>
36 36
37#include "r100_reg_safe.h"
38#include "rn50_reg_safe.h"
39
37/* Firmware Names */ 40/* Firmware Names */
38#define FIRMWARE_R100 "radeon/R100_cp.bin" 41#define FIRMWARE_R100 "radeon/R100_cp.bin"
39#define FIRMWARE_R200 "radeon/R200_cp.bin" 42#define FIRMWARE_R200 "radeon/R200_cp.bin"
@@ -51,11 +54,14 @@ MODULE_FIRMWARE(FIRMWARE_RS690);
51MODULE_FIRMWARE(FIRMWARE_RS600); 54MODULE_FIRMWARE(FIRMWARE_RS600);
52MODULE_FIRMWARE(FIRMWARE_R520); 55MODULE_FIRMWARE(FIRMWARE_R520);
53 56
57#include "r100_track.h"
58
54/* This files gather functions specifics to: 59/* This files gather functions specifics to:
55 * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 60 * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280
56 * 61 *
57 * Some of these functions might be used by newer ASICs. 62 * Some of these functions might be used by newer ASICs.
58 */ 63 */
64int r200_init(struct radeon_device *rdev);
59void r100_hdp_reset(struct radeon_device *rdev); 65void r100_hdp_reset(struct radeon_device *rdev);
60void r100_gpu_init(struct radeon_device *rdev); 66void r100_gpu_init(struct radeon_device *rdev);
61int r100_gui_wait_for_idle(struct radeon_device *rdev); 67int r100_gui_wait_for_idle(struct radeon_device *rdev);
@@ -1017,147 +1023,356 @@ int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
1017 return 0; 1023 return 0;
1018} 1024}
1019 1025
1026static int r100_get_vtx_size(uint32_t vtx_fmt)
1027{
1028 int vtx_size;
1029 vtx_size = 2;
1030 /* ordered according to bits in spec */
1031 if (vtx_fmt & RADEON_SE_VTX_FMT_W0)
1032 vtx_size++;
1033 if (vtx_fmt & RADEON_SE_VTX_FMT_FPCOLOR)
1034 vtx_size += 3;
1035 if (vtx_fmt & RADEON_SE_VTX_FMT_FPALPHA)
1036 vtx_size++;
1037 if (vtx_fmt & RADEON_SE_VTX_FMT_PKCOLOR)
1038 vtx_size++;
1039 if (vtx_fmt & RADEON_SE_VTX_FMT_FPSPEC)
1040 vtx_size += 3;
1041 if (vtx_fmt & RADEON_SE_VTX_FMT_FPFOG)
1042 vtx_size++;
1043 if (vtx_fmt & RADEON_SE_VTX_FMT_PKSPEC)
1044 vtx_size++;
1045 if (vtx_fmt & RADEON_SE_VTX_FMT_ST0)
1046 vtx_size += 2;
1047 if (vtx_fmt & RADEON_SE_VTX_FMT_ST1)
1048 vtx_size += 2;
1049 if (vtx_fmt & RADEON_SE_VTX_FMT_Q1)
1050 vtx_size++;
1051 if (vtx_fmt & RADEON_SE_VTX_FMT_ST2)
1052 vtx_size += 2;
1053 if (vtx_fmt & RADEON_SE_VTX_FMT_Q2)
1054 vtx_size++;
1055 if (vtx_fmt & RADEON_SE_VTX_FMT_ST3)
1056 vtx_size += 2;
1057 if (vtx_fmt & RADEON_SE_VTX_FMT_Q3)
1058 vtx_size++;
1059 if (vtx_fmt & RADEON_SE_VTX_FMT_Q0)
1060 vtx_size++;
1061 /* blend weight */
1062 if (vtx_fmt & (0x7 << 15))
1063 vtx_size += (vtx_fmt >> 15) & 0x7;
1064 if (vtx_fmt & RADEON_SE_VTX_FMT_N0)
1065 vtx_size += 3;
1066 if (vtx_fmt & RADEON_SE_VTX_FMT_XY1)
1067 vtx_size += 2;
1068 if (vtx_fmt & RADEON_SE_VTX_FMT_Z1)
1069 vtx_size++;
1070 if (vtx_fmt & RADEON_SE_VTX_FMT_W1)
1071 vtx_size++;
1072 if (vtx_fmt & RADEON_SE_VTX_FMT_N1)
1073 vtx_size++;
1074 if (vtx_fmt & RADEON_SE_VTX_FMT_Z)
1075 vtx_size++;
1076 return vtx_size;
1077}
1078
1020static int r100_packet0_check(struct radeon_cs_parser *p, 1079static int r100_packet0_check(struct radeon_cs_parser *p,
1021 struct radeon_cs_packet *pkt) 1080 struct radeon_cs_packet *pkt,
1081 unsigned idx, unsigned reg)
1022{ 1082{
1023 struct radeon_cs_chunk *ib_chunk; 1083 struct radeon_cs_chunk *ib_chunk;
1024 struct radeon_cs_reloc *reloc; 1084 struct radeon_cs_reloc *reloc;
1085 struct r100_cs_track *track;
1025 volatile uint32_t *ib; 1086 volatile uint32_t *ib;
1026 uint32_t tmp; 1087 uint32_t tmp;
1027 unsigned reg;
1028 unsigned i;
1029 unsigned idx;
1030 bool onereg;
1031 int r; 1088 int r;
1089 int i, face;
1032 u32 tile_flags = 0; 1090 u32 tile_flags = 0;
1033 1091
1034 ib = p->ib->ptr; 1092 ib = p->ib->ptr;
1035 ib_chunk = &p->chunks[p->chunk_ib_idx]; 1093 ib_chunk = &p->chunks[p->chunk_ib_idx];
1036 idx = pkt->idx + 1; 1094 track = (struct r100_cs_track *)p->track;
1037 reg = pkt->reg; 1095
1038 onereg = false; 1096 switch (reg) {
1039 if (CP_PACKET0_GET_ONE_REG_WR(ib_chunk->kdata[pkt->idx])) { 1097 case RADEON_CRTC_GUI_TRIG_VLINE:
1040 onereg = true; 1098 r = r100_cs_packet_parse_vline(p);
1041 } 1099 if (r) {
1042 for (i = 0; i <= pkt->count; i++, idx++, reg += 4) { 1100 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1043 switch (reg) { 1101 idx, reg);
1044 case RADEON_CRTC_GUI_TRIG_VLINE: 1102 r100_cs_dump_packet(p, pkt);
1045 r = r100_cs_packet_parse_vline(p); 1103 return r;
1046 if (r) { 1104 }
1047 DRM_ERROR("No reloc for ib[%d]=0x%04X\n", 1105 break;
1048 idx, reg);
1049 r100_cs_dump_packet(p, pkt);
1050 return r;
1051 }
1052 break;
1053 /* FIXME: only allow PACKET3 blit? easier to check for out of 1106 /* FIXME: only allow PACKET3 blit? easier to check for out of
1054 * range access */ 1107 * range access */
1055 case RADEON_DST_PITCH_OFFSET: 1108 case RADEON_DST_PITCH_OFFSET:
1056 case RADEON_SRC_PITCH_OFFSET: 1109 case RADEON_SRC_PITCH_OFFSET:
1057 r = r100_cs_packet_next_reloc(p, &reloc); 1110 r = r100_reloc_pitch_offset(p, pkt, idx, reg);
1058 if (r) { 1111 if (r)
1059 DRM_ERROR("No reloc for ib[%d]=0x%04X\n", 1112 return r;
1060 idx, reg); 1113 break;
1061 r100_cs_dump_packet(p, pkt); 1114 case RADEON_RB3D_DEPTHOFFSET:
1062 return r; 1115 r = r100_cs_packet_next_reloc(p, &reloc);
1063 } 1116 if (r) {
1064 tmp = ib_chunk->kdata[idx] & 0x003fffff; 1117 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1065 tmp += (((u32)reloc->lobj.gpu_offset) >> 10); 1118 idx, reg);
1066 1119 r100_cs_dump_packet(p, pkt);
1067 if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) 1120 return r;
1068 tile_flags |= RADEON_DST_TILE_MACRO; 1121 }
1069 if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { 1122 track->zb.robj = reloc->robj;
1070 if (reg == RADEON_SRC_PITCH_OFFSET) { 1123 track->zb.offset = ib_chunk->kdata[idx];
1071 DRM_ERROR("Cannot src blit from microtiled surface\n"); 1124 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1072 r100_cs_dump_packet(p, pkt); 1125 break;
1073 return -EINVAL; 1126 case RADEON_RB3D_COLOROFFSET:
1074 } 1127 r = r100_cs_packet_next_reloc(p, &reloc);
1075 tile_flags |= RADEON_DST_TILE_MICRO; 1128 if (r) {
1076 } 1129 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1130 idx, reg);
1131 r100_cs_dump_packet(p, pkt);
1132 return r;
1133 }
1134 track->cb[0].robj = reloc->robj;
1135 track->cb[0].offset = ib_chunk->kdata[idx];
1136 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1137 break;
1138 case RADEON_PP_TXOFFSET_0:
1139 case RADEON_PP_TXOFFSET_1:
1140 case RADEON_PP_TXOFFSET_2:
1141 i = (reg - RADEON_PP_TXOFFSET_0) / 24;
1142 r = r100_cs_packet_next_reloc(p, &reloc);
1143 if (r) {
1144 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1145 idx, reg);
1146 r100_cs_dump_packet(p, pkt);
1147 return r;
1148 }
1149 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1150 track->textures[i].robj = reloc->robj;
1151 break;
1152 case RADEON_PP_CUBIC_OFFSET_T0_0:
1153 case RADEON_PP_CUBIC_OFFSET_T0_1:
1154 case RADEON_PP_CUBIC_OFFSET_T0_2:
1155 case RADEON_PP_CUBIC_OFFSET_T0_3:
1156 case RADEON_PP_CUBIC_OFFSET_T0_4:
1157 i = (reg - RADEON_PP_CUBIC_OFFSET_T0_0) / 4;
1158 r = r100_cs_packet_next_reloc(p, &reloc);
1159 if (r) {
1160 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1161 idx, reg);
1162 r100_cs_dump_packet(p, pkt);
1163 return r;
1164 }
1165 track->textures[0].cube_info[i].offset = ib_chunk->kdata[idx];
1166 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1167 track->textures[0].cube_info[i].robj = reloc->robj;
1168 break;
1169 case RADEON_PP_CUBIC_OFFSET_T1_0:
1170 case RADEON_PP_CUBIC_OFFSET_T1_1:
1171 case RADEON_PP_CUBIC_OFFSET_T1_2:
1172 case RADEON_PP_CUBIC_OFFSET_T1_3:
1173 case RADEON_PP_CUBIC_OFFSET_T1_4:
1174 i = (reg - RADEON_PP_CUBIC_OFFSET_T1_0) / 4;
1175 r = r100_cs_packet_next_reloc(p, &reloc);
1176 if (r) {
1177 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1178 idx, reg);
1179 r100_cs_dump_packet(p, pkt);
1180 return r;
1181 }
1182 track->textures[1].cube_info[i].offset = ib_chunk->kdata[idx];
1183 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1184 track->textures[1].cube_info[i].robj = reloc->robj;
1185 break;
1186 case RADEON_PP_CUBIC_OFFSET_T2_0:
1187 case RADEON_PP_CUBIC_OFFSET_T2_1:
1188 case RADEON_PP_CUBIC_OFFSET_T2_2:
1189 case RADEON_PP_CUBIC_OFFSET_T2_3:
1190 case RADEON_PP_CUBIC_OFFSET_T2_4:
1191 i = (reg - RADEON_PP_CUBIC_OFFSET_T2_0) / 4;
1192 r = r100_cs_packet_next_reloc(p, &reloc);
1193 if (r) {
1194 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1195 idx, reg);
1196 r100_cs_dump_packet(p, pkt);
1197 return r;
1198 }
1199 track->textures[2].cube_info[i].offset = ib_chunk->kdata[idx];
1200 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1201 track->textures[2].cube_info[i].robj = reloc->robj;
1202 break;
1203 case RADEON_RE_WIDTH_HEIGHT:
1204 track->maxy = ((ib_chunk->kdata[idx] >> 16) & 0x7FF);
1205 break;
1206 case RADEON_RB3D_COLORPITCH:
1207 r = r100_cs_packet_next_reloc(p, &reloc);
1208 if (r) {
1209 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1210 idx, reg);
1211 r100_cs_dump_packet(p, pkt);
1212 return r;
1213 }
1077 1214
1078 tmp |= tile_flags; 1215 if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
1079 ib[idx] = (ib_chunk->kdata[idx] & 0x3fc00000) | tmp; 1216 tile_flags |= RADEON_COLOR_TILE_ENABLE;
1080 break; 1217 if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
1081 case RADEON_RB3D_DEPTHOFFSET: 1218 tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
1082 case RADEON_RB3D_COLOROFFSET:
1083 case R300_RB3D_COLOROFFSET0:
1084 case R300_ZB_DEPTHOFFSET:
1085 case R200_PP_TXOFFSET_0:
1086 case R200_PP_TXOFFSET_1:
1087 case R200_PP_TXOFFSET_2:
1088 case R200_PP_TXOFFSET_3:
1089 case R200_PP_TXOFFSET_4:
1090 case R200_PP_TXOFFSET_5:
1091 case RADEON_PP_TXOFFSET_0:
1092 case RADEON_PP_TXOFFSET_1:
1093 case RADEON_PP_TXOFFSET_2:
1094 case R300_TX_OFFSET_0:
1095 case R300_TX_OFFSET_0+4:
1096 case R300_TX_OFFSET_0+8:
1097 case R300_TX_OFFSET_0+12:
1098 case R300_TX_OFFSET_0+16:
1099 case R300_TX_OFFSET_0+20:
1100 case R300_TX_OFFSET_0+24:
1101 case R300_TX_OFFSET_0+28:
1102 case R300_TX_OFFSET_0+32:
1103 case R300_TX_OFFSET_0+36:
1104 case R300_TX_OFFSET_0+40:
1105 case R300_TX_OFFSET_0+44:
1106 case R300_TX_OFFSET_0+48:
1107 case R300_TX_OFFSET_0+52:
1108 case R300_TX_OFFSET_0+56:
1109 case R300_TX_OFFSET_0+60:
1110 /* rn50 has no 3D engine so fail on any 3d setup */
1111 if (ASIC_IS_RN50(p->rdev)) {
1112 DRM_ERROR("attempt to use RN50 3D engine failed\n");
1113 return -EINVAL;
1114 }
1115 r = r100_cs_packet_next_reloc(p, &reloc);
1116 if (r) {
1117 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1118 idx, reg);
1119 r100_cs_dump_packet(p, pkt);
1120 return r;
1121 }
1122 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1123 break;
1124 case R300_RB3D_COLORPITCH0:
1125 case RADEON_RB3D_COLORPITCH:
1126 r = r100_cs_packet_next_reloc(p, &reloc);
1127 if (r) {
1128 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1129 idx, reg);
1130 r100_cs_dump_packet(p, pkt);
1131 return r;
1132 }
1133 1219
1134 if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) 1220 tmp = ib_chunk->kdata[idx] & ~(0x7 << 16);
1135 tile_flags |= RADEON_COLOR_TILE_ENABLE; 1221 tmp |= tile_flags;
1136 if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) 1222 ib[idx] = tmp;
1137 tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
1138 1223
1139 tmp = ib_chunk->kdata[idx] & ~(0x7 << 16); 1224 track->cb[0].pitch = ib_chunk->kdata[idx] & RADEON_COLORPITCH_MASK;
1140 tmp |= tile_flags; 1225 break;
1141 ib[idx] = tmp; 1226 case RADEON_RB3D_DEPTHPITCH:
1227 track->zb.pitch = ib_chunk->kdata[idx] & RADEON_DEPTHPITCH_MASK;
1228 break;
1229 case RADEON_RB3D_CNTL:
1230 switch ((ib_chunk->kdata[idx] >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) {
1231 case 7:
1232 case 8:
1233 case 9:
1234 case 11:
1235 case 12:
1236 track->cb[0].cpp = 1;
1142 break; 1237 break;
1143 case RADEON_RB3D_ZPASS_ADDR: 1238 case 3:
1144 r = r100_cs_packet_next_reloc(p, &reloc); 1239 case 4:
1145 if (r) { 1240 case 15:
1146 DRM_ERROR("No reloc for ib[%d]=0x%04X\n", 1241 track->cb[0].cpp = 2;
1147 idx, reg); 1242 break;
1148 r100_cs_dump_packet(p, pkt); 1243 case 6:
1149 return r; 1244 track->cb[0].cpp = 4;
1150 } 1245 break;
1151 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); 1246 default:
1247 DRM_ERROR("Invalid color buffer format (%d) !\n",
1248 ((ib_chunk->kdata[idx] >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f));
1249 return -EINVAL;
1250 }
1251 track->z_enabled = !!(ib_chunk->kdata[idx] & RADEON_Z_ENABLE);
1252 break;
1253 case RADEON_RB3D_ZSTENCILCNTL:
1254 switch (ib_chunk->kdata[idx] & 0xf) {
1255 case 0:
1256 track->zb.cpp = 2;
1257 break;
1258 case 2:
1259 case 3:
1260 case 4:
1261 case 5:
1262 case 9:
1263 case 11:
1264 track->zb.cpp = 4;
1152 break; 1265 break;
1153 default: 1266 default:
1154 /* FIXME: we don't want to allow anyothers packet */
1155 break; 1267 break;
1156 } 1268 }
1157 if (onereg) { 1269 break;
1158 /* FIXME: forbid onereg write to register on relocate */ 1270 case RADEON_RB3D_ZPASS_ADDR:
1271 r = r100_cs_packet_next_reloc(p, &reloc);
1272 if (r) {
1273 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
1274 idx, reg);
1275 r100_cs_dump_packet(p, pkt);
1276 return r;
1277 }
1278 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1279 break;
1280 case RADEON_PP_CNTL:
1281 {
1282 uint32_t temp = ib_chunk->kdata[idx] >> 4;
1283 for (i = 0; i < track->num_texture; i++)
1284 track->textures[i].enabled = !!(temp & (1 << i));
1285 }
1286 break;
1287 case RADEON_SE_VF_CNTL:
1288 track->vap_vf_cntl = ib_chunk->kdata[idx];
1289 break;
1290 case RADEON_SE_VTX_FMT:
1291 track->vtx_size = r100_get_vtx_size(ib_chunk->kdata[idx]);
1292 break;
1293 case RADEON_PP_TEX_SIZE_0:
1294 case RADEON_PP_TEX_SIZE_1:
1295 case RADEON_PP_TEX_SIZE_2:
1296 i = (reg - RADEON_PP_TEX_SIZE_0) / 8;
1297 track->textures[i].width = (ib_chunk->kdata[idx] & RADEON_TEX_USIZE_MASK) + 1;
1298 track->textures[i].height = ((ib_chunk->kdata[idx] & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1;
1299 break;
1300 case RADEON_PP_TEX_PITCH_0:
1301 case RADEON_PP_TEX_PITCH_1:
1302 case RADEON_PP_TEX_PITCH_2:
1303 i = (reg - RADEON_PP_TEX_PITCH_0) / 8;
1304 track->textures[i].pitch = ib_chunk->kdata[idx] + 32;
1305 break;
1306 case RADEON_PP_TXFILTER_0:
1307 case RADEON_PP_TXFILTER_1:
1308 case RADEON_PP_TXFILTER_2:
1309 i = (reg - RADEON_PP_TXFILTER_0) / 24;
1310 track->textures[i].num_levels = ((ib_chunk->kdata[idx] & RADEON_MAX_MIP_LEVEL_MASK)
1311 >> RADEON_MAX_MIP_LEVEL_SHIFT);
1312 tmp = (ib_chunk->kdata[idx] >> 23) & 0x7;
1313 if (tmp == 2 || tmp == 6)
1314 track->textures[i].roundup_w = false;
1315 tmp = (ib_chunk->kdata[idx] >> 27) & 0x7;
1316 if (tmp == 2 || tmp == 6)
1317 track->textures[i].roundup_h = false;
1318 break;
1319 case RADEON_PP_TXFORMAT_0:
1320 case RADEON_PP_TXFORMAT_1:
1321 case RADEON_PP_TXFORMAT_2:
1322 i = (reg - RADEON_PP_TXFORMAT_0) / 24;
1323 if (ib_chunk->kdata[idx] & RADEON_TXFORMAT_NON_POWER2) {
1324 track->textures[i].use_pitch = 1;
1325 } else {
1326 track->textures[i].use_pitch = 0;
1327 track->textures[i].width = 1 << ((ib_chunk->kdata[idx] >> RADEON_TXFORMAT_WIDTH_SHIFT) & RADEON_TXFORMAT_WIDTH_MASK);
1328 track->textures[i].height = 1 << ((ib_chunk->kdata[idx] >> RADEON_TXFORMAT_HEIGHT_SHIFT) & RADEON_TXFORMAT_HEIGHT_MASK);
1329 }
1330 if (ib_chunk->kdata[idx] & RADEON_TXFORMAT_CUBIC_MAP_ENABLE)
1331 track->textures[i].tex_coord_type = 2;
1332 switch ((ib_chunk->kdata[idx] & RADEON_TXFORMAT_FORMAT_MASK)) {
1333 case RADEON_TXFORMAT_I8:
1334 case RADEON_TXFORMAT_RGB332:
1335 case RADEON_TXFORMAT_Y8:
1336 track->textures[i].cpp = 1;
1337 break;
1338 case RADEON_TXFORMAT_AI88:
1339 case RADEON_TXFORMAT_ARGB1555:
1340 case RADEON_TXFORMAT_RGB565:
1341 case RADEON_TXFORMAT_ARGB4444:
1342 case RADEON_TXFORMAT_VYUY422:
1343 case RADEON_TXFORMAT_YVYU422:
1344 case RADEON_TXFORMAT_DXT1:
1345 case RADEON_TXFORMAT_SHADOW16:
1346 case RADEON_TXFORMAT_LDUDV655:
1347 case RADEON_TXFORMAT_DUDV88:
1348 track->textures[i].cpp = 2;
1159 break; 1349 break;
1350 case RADEON_TXFORMAT_ARGB8888:
1351 case RADEON_TXFORMAT_RGBA8888:
1352 case RADEON_TXFORMAT_DXT23:
1353 case RADEON_TXFORMAT_DXT45:
1354 case RADEON_TXFORMAT_SHADOW32:
1355 case RADEON_TXFORMAT_LDUDUV8888:
1356 track->textures[i].cpp = 4;
1357 break;
1358 }
1359 track->textures[i].cube_info[4].width = 1 << ((ib_chunk->kdata[idx] >> 16) & 0xf);
1360 track->textures[i].cube_info[4].height = 1 << ((ib_chunk->kdata[idx] >> 20) & 0xf);
1361 break;
1362 case RADEON_PP_CUBIC_FACES_0:
1363 case RADEON_PP_CUBIC_FACES_1:
1364 case RADEON_PP_CUBIC_FACES_2:
1365 tmp = ib_chunk->kdata[idx];
1366 i = (reg - RADEON_PP_CUBIC_FACES_0) / 4;
1367 for (face = 0; face < 4; face++) {
1368 track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf);
1369 track->textures[i].cube_info[face].height = 1 << ((tmp >> ((face * 8) + 4)) & 0xf);
1160 } 1370 }
1371 break;
1372 default:
1373 printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n",
1374 reg, idx);
1375 return -EINVAL;
1161 } 1376 }
1162 return 0; 1377 return 0;
1163} 1378}
@@ -1186,6 +1401,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1186{ 1401{
1187 struct radeon_cs_chunk *ib_chunk; 1402 struct radeon_cs_chunk *ib_chunk;
1188 struct radeon_cs_reloc *reloc; 1403 struct radeon_cs_reloc *reloc;
1404 struct r100_cs_track *track;
1189 unsigned idx; 1405 unsigned idx;
1190 unsigned i, c; 1406 unsigned i, c;
1191 volatile uint32_t *ib; 1407 volatile uint32_t *ib;
@@ -1194,9 +1410,11 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1194 ib = p->ib->ptr; 1410 ib = p->ib->ptr;
1195 ib_chunk = &p->chunks[p->chunk_ib_idx]; 1411 ib_chunk = &p->chunks[p->chunk_ib_idx];
1196 idx = pkt->idx + 1; 1412 idx = pkt->idx + 1;
1413 track = (struct r100_cs_track *)p->track;
1197 switch (pkt->opcode) { 1414 switch (pkt->opcode) {
1198 case PACKET3_3D_LOAD_VBPNTR: 1415 case PACKET3_3D_LOAD_VBPNTR:
1199 c = ib_chunk->kdata[idx++]; 1416 c = ib_chunk->kdata[idx++];
1417 track->num_arrays = c;
1200 for (i = 0; i < (c - 1); i += 2, idx += 3) { 1418 for (i = 0; i < (c - 1); i += 2, idx += 3) {
1201 r = r100_cs_packet_next_reloc(p, &reloc); 1419 r = r100_cs_packet_next_reloc(p, &reloc);
1202 if (r) { 1420 if (r) {
@@ -1206,6 +1424,9 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1206 return r; 1424 return r;
1207 } 1425 }
1208 ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); 1426 ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset);
1427 track->arrays[i + 0].robj = reloc->robj;
1428 track->arrays[i + 0].esize = ib_chunk->kdata[idx] >> 8;
1429 track->arrays[i + 0].esize &= 0x7F;
1209 r = r100_cs_packet_next_reloc(p, &reloc); 1430 r = r100_cs_packet_next_reloc(p, &reloc);
1210 if (r) { 1431 if (r) {
1211 DRM_ERROR("No reloc for packet3 %d\n", 1432 DRM_ERROR("No reloc for packet3 %d\n",
@@ -1214,6 +1435,9 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1214 return r; 1435 return r;
1215 } 1436 }
1216 ib[idx+2] = ib_chunk->kdata[idx+2] + ((u32)reloc->lobj.gpu_offset); 1437 ib[idx+2] = ib_chunk->kdata[idx+2] + ((u32)reloc->lobj.gpu_offset);
1438 track->arrays[i + 1].robj = reloc->robj;
1439 track->arrays[i + 1].esize = ib_chunk->kdata[idx] >> 24;
1440 track->arrays[i + 1].esize &= 0x7F;
1217 } 1441 }
1218 if (c & 1) { 1442 if (c & 1) {
1219 r = r100_cs_packet_next_reloc(p, &reloc); 1443 r = r100_cs_packet_next_reloc(p, &reloc);
@@ -1224,6 +1448,9 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1224 return r; 1448 return r;
1225 } 1449 }
1226 ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); 1450 ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset);
1451 track->arrays[i + 0].robj = reloc->robj;
1452 track->arrays[i + 0].esize = ib_chunk->kdata[idx] >> 8;
1453 track->arrays[i + 0].esize &= 0x7F;
1227 } 1454 }
1228 break; 1455 break;
1229 case PACKET3_INDX_BUFFER: 1456 case PACKET3_INDX_BUFFER:
@@ -1240,7 +1467,6 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1240 } 1467 }
1241 break; 1468 break;
1242 case 0x23: 1469 case 0x23:
1243 /* FIXME: cleanup */
1244 /* 3D_RNDR_GEN_INDX_PRIM on r100/r200 */ 1470 /* 3D_RNDR_GEN_INDX_PRIM on r100/r200 */
1245 r = r100_cs_packet_next_reloc(p, &reloc); 1471 r = r100_cs_packet_next_reloc(p, &reloc);
1246 if (r) { 1472 if (r) {
@@ -1249,18 +1475,71 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1249 return r; 1475 return r;
1250 } 1476 }
1251 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); 1477 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
1478 track->num_arrays = 1;
1479 track->vtx_size = r100_get_vtx_size(ib_chunk->kdata[idx+2]);
1480
1481 track->arrays[0].robj = reloc->robj;
1482 track->arrays[0].esize = track->vtx_size;
1483
1484 track->max_indx = ib_chunk->kdata[idx+1];
1485
1486 track->vap_vf_cntl = ib_chunk->kdata[idx+3];
1487 track->immd_dwords = pkt->count - 1;
1488 r = r100_cs_track_check(p->rdev, track);
1489 if (r)
1490 return r;
1252 break; 1491 break;
1253 case PACKET3_3D_DRAW_IMMD: 1492 case PACKET3_3D_DRAW_IMMD:
1493 if (((ib_chunk->kdata[idx+1] >> 4) & 0x3) != 3) {
1494 DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n");
1495 return -EINVAL;
1496 }
1497 track->vap_vf_cntl = ib_chunk->kdata[idx+1];
1498 track->immd_dwords = pkt->count - 1;
1499 r = r100_cs_track_check(p->rdev, track);
1500 if (r)
1501 return r;
1502 break;
1254 /* triggers drawing using in-packet vertex data */ 1503 /* triggers drawing using in-packet vertex data */
1255 case PACKET3_3D_DRAW_IMMD_2: 1504 case PACKET3_3D_DRAW_IMMD_2:
1505 if (((ib_chunk->kdata[idx] >> 4) & 0x3) != 3) {
1506 DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n");
1507 return -EINVAL;
1508 }
1509 track->vap_vf_cntl = ib_chunk->kdata[idx];
1510 track->immd_dwords = pkt->count;
1511 r = r100_cs_track_check(p->rdev, track);
1512 if (r)
1513 return r;
1514 break;
1256 /* triggers drawing using in-packet vertex data */ 1515 /* triggers drawing using in-packet vertex data */
1257 case PACKET3_3D_DRAW_VBUF_2: 1516 case PACKET3_3D_DRAW_VBUF_2:
1517 track->vap_vf_cntl = ib_chunk->kdata[idx];
1518 r = r100_cs_track_check(p->rdev, track);
1519 if (r)
1520 return r;
1521 break;
1258 /* triggers drawing of vertex buffers setup elsewhere */ 1522 /* triggers drawing of vertex buffers setup elsewhere */
1259 case PACKET3_3D_DRAW_INDX_2: 1523 case PACKET3_3D_DRAW_INDX_2:
1524 track->vap_vf_cntl = ib_chunk->kdata[idx];
1525 r = r100_cs_track_check(p->rdev, track);
1526 if (r)
1527 return r;
1528 break;
1260 /* triggers drawing using indices to vertex buffer */ 1529 /* triggers drawing using indices to vertex buffer */
1261 case PACKET3_3D_DRAW_VBUF: 1530 case PACKET3_3D_DRAW_VBUF:
1531 track->vap_vf_cntl = ib_chunk->kdata[idx + 1];
1532 r = r100_cs_track_check(p->rdev, track);
1533 if (r)
1534 return r;
1535 break;
1262 /* triggers drawing of vertex buffers setup elsewhere */ 1536 /* triggers drawing of vertex buffers setup elsewhere */
1263 case PACKET3_3D_DRAW_INDX: 1537 case PACKET3_3D_DRAW_INDX:
1538 track->vap_vf_cntl = ib_chunk->kdata[idx + 1];
1539 r = r100_cs_track_check(p->rdev, track);
1540 if (r)
1541 return r;
1542 break;
1264 /* triggers drawing using indices to vertex buffer */ 1543 /* triggers drawing using indices to vertex buffer */
1265 case PACKET3_NOP: 1544 case PACKET3_NOP:
1266 break; 1545 break;
@@ -1274,8 +1553,11 @@ static int r100_packet3_check(struct radeon_cs_parser *p,
1274int r100_cs_parse(struct radeon_cs_parser *p) 1553int r100_cs_parse(struct radeon_cs_parser *p)
1275{ 1554{
1276 struct radeon_cs_packet pkt; 1555 struct radeon_cs_packet pkt;
1556 struct r100_cs_track track;
1277 int r; 1557 int r;
1278 1558
1559 r100_cs_track_clear(p->rdev, &track);
1560 p->track = &track;
1279 do { 1561 do {
1280 r = r100_cs_packet_parse(p, &pkt, p->idx); 1562 r = r100_cs_packet_parse(p, &pkt, p->idx);
1281 if (r) { 1563 if (r) {
@@ -1284,7 +1566,16 @@ int r100_cs_parse(struct radeon_cs_parser *p)
1284 p->idx += pkt.count + 2; 1566 p->idx += pkt.count + 2;
1285 switch (pkt.type) { 1567 switch (pkt.type) {
1286 case PACKET_TYPE0: 1568 case PACKET_TYPE0:
1287 r = r100_packet0_check(p, &pkt); 1569 if (p->rdev->family >= CHIP_R200)
1570 r = r100_cs_parse_packet0(p, &pkt,
1571 p->rdev->config.r100.reg_safe_bm,
1572 p->rdev->config.r100.reg_safe_bm_size,
1573 &r200_packet0_check);
1574 else
1575 r = r100_cs_parse_packet0(p, &pkt,
1576 p->rdev->config.r100.reg_safe_bm,
1577 p->rdev->config.r100.reg_safe_bm_size,
1578 &r100_packet0_check);
1288 break; 1579 break;
1289 case PACKET_TYPE2: 1580 case PACKET_TYPE2:
1290 break; 1581 break;
@@ -1683,6 +1974,15 @@ void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v)
1683 1974
1684int r100_init(struct radeon_device *rdev) 1975int r100_init(struct radeon_device *rdev)
1685{ 1976{
1977 if (ASIC_IS_RN50(rdev)) {
1978 rdev->config.r100.reg_safe_bm = rn50_reg_safe_bm;
1979 rdev->config.r100.reg_safe_bm_size = ARRAY_SIZE(rn50_reg_safe_bm);
1980 } else if (rdev->family < CHIP_R200) {
1981 rdev->config.r100.reg_safe_bm = r100_reg_safe_bm;
1982 rdev->config.r100.reg_safe_bm_size = ARRAY_SIZE(r100_reg_safe_bm);
1983 } else {
1984 return r200_init(rdev);
1985 }
1686 return 0; 1986 return 0;
1687} 1987}
1688 1988
@@ -2383,3 +2683,274 @@ void r100_bandwidth_update(struct radeon_device *rdev)
2383 (unsigned int)RREG32(RADEON_GRPH2_BUFFER_CNTL)); 2683 (unsigned int)RREG32(RADEON_GRPH2_BUFFER_CNTL));
2384 } 2684 }
2385} 2685}
2686
2687static inline void r100_cs_track_texture_print(struct r100_cs_track_texture *t)
2688{
2689 DRM_ERROR("pitch %d\n", t->pitch);
2690 DRM_ERROR("width %d\n", t->width);
2691 DRM_ERROR("height %d\n", t->height);
2692 DRM_ERROR("num levels %d\n", t->num_levels);
2693 DRM_ERROR("depth %d\n", t->txdepth);
2694 DRM_ERROR("bpp %d\n", t->cpp);
2695 DRM_ERROR("coordinate type %d\n", t->tex_coord_type);
2696 DRM_ERROR("width round to power of 2 %d\n", t->roundup_w);
2697 DRM_ERROR("height round to power of 2 %d\n", t->roundup_h);
2698}
2699
2700static int r100_cs_track_cube(struct radeon_device *rdev,
2701 struct r100_cs_track *track, unsigned idx)
2702{
2703 unsigned face, w, h;
2704 struct radeon_object *cube_robj;
2705 unsigned long size;
2706
2707 for (face = 0; face < 5; face++) {
2708 cube_robj = track->textures[idx].cube_info[face].robj;
2709 w = track->textures[idx].cube_info[face].width;
2710 h = track->textures[idx].cube_info[face].height;
2711
2712 size = w * h;
2713 size *= track->textures[idx].cpp;
2714
2715 size += track->textures[idx].cube_info[face].offset;
2716
2717 if (size > radeon_object_size(cube_robj)) {
2718 DRM_ERROR("Cube texture offset greater than object size %lu %lu\n",
2719 size, radeon_object_size(cube_robj));
2720 r100_cs_track_texture_print(&track->textures[idx]);
2721 return -1;
2722 }
2723 }
2724 return 0;
2725}
2726
2727static int r100_cs_track_texture_check(struct radeon_device *rdev,
2728 struct r100_cs_track *track)
2729{
2730 struct radeon_object *robj;
2731 unsigned long size;
2732 unsigned u, i, w, h;
2733 int ret;
2734
2735 for (u = 0; u < track->num_texture; u++) {
2736 if (!track->textures[u].enabled)
2737 continue;
2738 robj = track->textures[u].robj;
2739 if (robj == NULL) {
2740 DRM_ERROR("No texture bound to unit %u\n", u);
2741 return -EINVAL;
2742 }
2743 size = 0;
2744 for (i = 0; i <= track->textures[u].num_levels; i++) {
2745 if (track->textures[u].use_pitch) {
2746 if (rdev->family < CHIP_R300)
2747 w = (track->textures[u].pitch / track->textures[u].cpp) / (1 << i);
2748 else
2749 w = track->textures[u].pitch / (1 << i);
2750 } else {
2751 w = track->textures[u].width / (1 << i);
2752 if (rdev->family >= CHIP_RV515)
2753 w |= track->textures[u].width_11;
2754 if (track->textures[u].roundup_w)
2755 w = roundup_pow_of_two(w);
2756 }
2757 h = track->textures[u].height / (1 << i);
2758 if (rdev->family >= CHIP_RV515)
2759 h |= track->textures[u].height_11;
2760 if (track->textures[u].roundup_h)
2761 h = roundup_pow_of_two(h);
2762 size += w * h;
2763 }
2764 size *= track->textures[u].cpp;
2765 switch (track->textures[u].tex_coord_type) {
2766 case 0:
2767 break;
2768 case 1:
2769 size *= (1 << track->textures[u].txdepth);
2770 break;
2771 case 2:
2772 if (track->separate_cube) {
2773 ret = r100_cs_track_cube(rdev, track, u);
2774 if (ret)
2775 return ret;
2776 } else
2777 size *= 6;
2778 break;
2779 default:
2780 DRM_ERROR("Invalid texture coordinate type %u for unit "
2781 "%u\n", track->textures[u].tex_coord_type, u);
2782 return -EINVAL;
2783 }
2784 if (size > radeon_object_size(robj)) {
2785 DRM_ERROR("Texture of unit %u needs %lu bytes but is "
2786 "%lu\n", u, size, radeon_object_size(robj));
2787 r100_cs_track_texture_print(&track->textures[u]);
2788 return -EINVAL;
2789 }
2790 }
2791 return 0;
2792}
2793
2794int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track)
2795{
2796 unsigned i;
2797 unsigned long size;
2798 unsigned prim_walk;
2799 unsigned nverts;
2800
2801 for (i = 0; i < track->num_cb; i++) {
2802 if (track->cb[i].robj == NULL) {
2803 DRM_ERROR("[drm] No buffer for color buffer %d !\n", i);
2804 return -EINVAL;
2805 }
2806 size = track->cb[i].pitch * track->cb[i].cpp * track->maxy;
2807 size += track->cb[i].offset;
2808 if (size > radeon_object_size(track->cb[i].robj)) {
2809 DRM_ERROR("[drm] Buffer too small for color buffer %d "
2810 "(need %lu have %lu) !\n", i, size,
2811 radeon_object_size(track->cb[i].robj));
2812 DRM_ERROR("[drm] color buffer %d (%u %u %u %u)\n",
2813 i, track->cb[i].pitch, track->cb[i].cpp,
2814 track->cb[i].offset, track->maxy);
2815 return -EINVAL;
2816 }
2817 }
2818 if (track->z_enabled) {
2819 if (track->zb.robj == NULL) {
2820 DRM_ERROR("[drm] No buffer for z buffer !\n");
2821 return -EINVAL;
2822 }
2823 size = track->zb.pitch * track->zb.cpp * track->maxy;
2824 size += track->zb.offset;
2825 if (size > radeon_object_size(track->zb.robj)) {
2826 DRM_ERROR("[drm] Buffer too small for z buffer "
2827 "(need %lu have %lu) !\n", size,
2828 radeon_object_size(track->zb.robj));
2829 DRM_ERROR("[drm] zbuffer (%u %u %u %u)\n",
2830 track->zb.pitch, track->zb.cpp,
2831 track->zb.offset, track->maxy);
2832 return -EINVAL;
2833 }
2834 }
2835 prim_walk = (track->vap_vf_cntl >> 4) & 0x3;
2836 nverts = (track->vap_vf_cntl >> 16) & 0xFFFF;
2837 switch (prim_walk) {
2838 case 1:
2839 for (i = 0; i < track->num_arrays; i++) {
2840 size = track->arrays[i].esize * track->max_indx * 4;
2841 if (track->arrays[i].robj == NULL) {
2842 DRM_ERROR("(PW %u) Vertex array %u no buffer "
2843 "bound\n", prim_walk, i);
2844 return -EINVAL;
2845 }
2846 if (size > radeon_object_size(track->arrays[i].robj)) {
2847 DRM_ERROR("(PW %u) Vertex array %u need %lu dwords "
2848 "have %lu dwords\n", prim_walk, i,
2849 size >> 2,
2850 radeon_object_size(track->arrays[i].robj) >> 2);
2851 DRM_ERROR("Max indices %u\n", track->max_indx);
2852 return -EINVAL;
2853 }
2854 }
2855 break;
2856 case 2:
2857 for (i = 0; i < track->num_arrays; i++) {
2858 size = track->arrays[i].esize * (nverts - 1) * 4;
2859 if (track->arrays[i].robj == NULL) {
2860 DRM_ERROR("(PW %u) Vertex array %u no buffer "
2861 "bound\n", prim_walk, i);
2862 return -EINVAL;
2863 }
2864 if (size > radeon_object_size(track->arrays[i].robj)) {
2865 DRM_ERROR("(PW %u) Vertex array %u need %lu dwords "
2866 "have %lu dwords\n", prim_walk, i, size >> 2,
2867 radeon_object_size(track->arrays[i].robj) >> 2);
2868 return -EINVAL;
2869 }
2870 }
2871 break;
2872 case 3:
2873 size = track->vtx_size * nverts;
2874 if (size != track->immd_dwords) {
2875 DRM_ERROR("IMMD draw %u dwors but needs %lu dwords\n",
2876 track->immd_dwords, size);
2877 DRM_ERROR("VAP_VF_CNTL.NUM_VERTICES %u, VTX_SIZE %u\n",
2878 nverts, track->vtx_size);
2879 return -EINVAL;
2880 }
2881 break;
2882 default:
2883 DRM_ERROR("[drm] Invalid primitive walk %d for VAP_VF_CNTL\n",
2884 prim_walk);
2885 return -EINVAL;
2886 }
2887 return r100_cs_track_texture_check(rdev, track);
2888}
2889
2890void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track)
2891{
2892 unsigned i, face;
2893
2894 if (rdev->family < CHIP_R300) {
2895 track->num_cb = 1;
2896 if (rdev->family <= CHIP_RS200)
2897 track->num_texture = 3;
2898 else
2899 track->num_texture = 6;
2900 track->maxy = 2048;
2901 track->separate_cube = 1;
2902 } else {
2903 track->num_cb = 4;
2904 track->num_texture = 16;
2905 track->maxy = 4096;
2906 track->separate_cube = 0;
2907 }
2908
2909 for (i = 0; i < track->num_cb; i++) {
2910 track->cb[i].robj = NULL;
2911 track->cb[i].pitch = 8192;
2912 track->cb[i].cpp = 16;
2913 track->cb[i].offset = 0;
2914 }
2915 track->z_enabled = true;
2916 track->zb.robj = NULL;
2917 track->zb.pitch = 8192;
2918 track->zb.cpp = 4;
2919 track->zb.offset = 0;
2920 track->vtx_size = 0x7F;
2921 track->immd_dwords = 0xFFFFFFFFUL;
2922 track->num_arrays = 11;
2923 track->max_indx = 0x00FFFFFFUL;
2924 for (i = 0; i < track->num_arrays; i++) {
2925 track->arrays[i].robj = NULL;
2926 track->arrays[i].esize = 0x7F;
2927 }
2928 for (i = 0; i < track->num_texture; i++) {
2929 track->textures[i].pitch = 16536;
2930 track->textures[i].width = 16536;
2931 track->textures[i].height = 16536;
2932 track->textures[i].width_11 = 1 << 11;
2933 track->textures[i].height_11 = 1 << 11;
2934 track->textures[i].num_levels = 12;
2935 if (rdev->family <= CHIP_RS200) {
2936 track->textures[i].tex_coord_type = 0;
2937 track->textures[i].txdepth = 0;
2938 } else {
2939 track->textures[i].txdepth = 16;
2940 track->textures[i].tex_coord_type = 1;
2941 }
2942 track->textures[i].cpp = 64;
2943 track->textures[i].robj = NULL;
2944 /* CS IB emission code makes sure texture unit are disabled */
2945 track->textures[i].enabled = false;
2946 track->textures[i].roundup_w = true;
2947 track->textures[i].roundup_h = true;
2948 if (track->separate_cube)
2949 for (face = 0; face < 5; face++) {
2950 track->textures[i].cube_info[face].robj = NULL;
2951 track->textures[i].cube_info[face].width = 16536;
2952 track->textures[i].cube_info[face].height = 16536;
2953 track->textures[i].cube_info[face].offset = 0;
2954 }
2955 }
2956}
diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h
new file mode 100644
index 000000000000..70a82eda394a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/r100_track.h
@@ -0,0 +1,124 @@
1
2#define R100_TRACK_MAX_TEXTURE 3
3#define R200_TRACK_MAX_TEXTURE 6
4#define R300_TRACK_MAX_TEXTURE 16
5
6#define R100_MAX_CB 1
7#define R300_MAX_CB 4
8
9/*
10 * CS functions
11 */
12struct r100_cs_track_cb {
13 struct radeon_object *robj;
14 unsigned pitch;
15 unsigned cpp;
16 unsigned offset;
17};
18
19struct r100_cs_track_array {
20 struct radeon_object *robj;
21 unsigned esize;
22};
23
24struct r100_cs_cube_info {
25 struct radeon_object *robj;
26 unsigned offset;
27 unsigned width;
28 unsigned height;
29};
30
31struct r100_cs_track_texture {
32 struct radeon_object *robj;
33 struct r100_cs_cube_info cube_info[5]; /* info for 5 non-primary faces */
34 unsigned pitch;
35 unsigned width;
36 unsigned height;
37 unsigned num_levels;
38 unsigned cpp;
39 unsigned tex_coord_type;
40 unsigned txdepth;
41 unsigned width_11;
42 unsigned height_11;
43 bool use_pitch;
44 bool enabled;
45 bool roundup_w;
46 bool roundup_h;
47};
48
49struct r100_cs_track_limits {
50 unsigned num_cb;
51 unsigned num_texture;
52 unsigned max_levels;
53};
54
55struct r100_cs_track {
56 struct radeon_device *rdev;
57 unsigned num_cb;
58 unsigned num_texture;
59 unsigned maxy;
60 unsigned vtx_size;
61 unsigned vap_vf_cntl;
62 unsigned immd_dwords;
63 unsigned num_arrays;
64 unsigned max_indx;
65 struct r100_cs_track_array arrays[11];
66 struct r100_cs_track_cb cb[R300_MAX_CB];
67 struct r100_cs_track_cb zb;
68 struct r100_cs_track_texture textures[R300_TRACK_MAX_TEXTURE];
69 bool z_enabled;
70 bool separate_cube;
71
72};
73
74int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track);
75void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track);
76int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
77 struct radeon_cs_reloc **cs_reloc);
78void r100_cs_dump_packet(struct radeon_cs_parser *p,
79 struct radeon_cs_packet *pkt);
80
81int r100_cs_packet_parse_vline(struct radeon_cs_parser *p);
82
83int r200_packet0_check(struct radeon_cs_parser *p,
84 struct radeon_cs_packet *pkt,
85 unsigned idx, unsigned reg);
86
87static inline int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
88 struct radeon_cs_packet *pkt,
89 unsigned idx,
90 unsigned reg)
91{
92 int r;
93 u32 tile_flags = 0;
94 u32 tmp;
95 struct radeon_cs_reloc *reloc;
96 struct radeon_cs_chunk *ib_chunk;
97
98 ib_chunk = &p->chunks[p->chunk_ib_idx];
99
100 r = r100_cs_packet_next_reloc(p, &reloc);
101 if (r) {
102 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
103 idx, reg);
104 r100_cs_dump_packet(p, pkt);
105 return r;
106 }
107 tmp = ib_chunk->kdata[idx] & 0x003fffff;
108 tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
109
110 if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
111 tile_flags |= RADEON_DST_TILE_MACRO;
112 if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
113 if (reg == RADEON_SRC_PITCH_OFFSET) {
114 DRM_ERROR("Cannot src blit from microtiled surface\n");
115 r100_cs_dump_packet(p, pkt);
116 return -EINVAL;
117 }
118 tile_flags |= RADEON_DST_TILE_MICRO;
119 }
120
121 tmp |= tile_flags;
122 p->ib->ptr[idx] = (ib_chunk->kdata[idx] & 0x3fc00000) | tmp;
123 return 0;
124}
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
new file mode 100644
index 000000000000..568c74bfba3d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -0,0 +1,456 @@
1/*
2 * Copyright 2008 Advanced Micro Devices, Inc.
3 * Copyright 2008 Red Hat Inc.
4 * Copyright 2009 Jerome Glisse.
5 *
6 * Permission is hereby granted, free of charge, to any person obtaining a
7 * copy of this software and associated documentation files (the "Software"),
8 * to deal in the Software without restriction, including without limitation
9 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
10 * and/or sell copies of the Software, and to permit persons to whom the
11 * Software is furnished to do so, subject to the following conditions:
12 *
13 * The above copyright notice and this permission notice shall be included in
14 * all copies or substantial portions of the Software.
15 *
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
19 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
20 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
21 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
22 * OTHER DEALINGS IN THE SOFTWARE.
23 *
24 * Authors: Dave Airlie
25 * Alex Deucher
26 * Jerome Glisse
27 */
28#include "drmP.h"
29#include "drm.h"
30#include "radeon_drm.h"
31#include "radeon_reg.h"
32#include "radeon.h"
33
34#include "r200_reg_safe.h"
35
36#include "r100_track.h"
37
38static int r200_get_vtx_size_0(uint32_t vtx_fmt_0)
39{
40 int vtx_size, i;
41 vtx_size = 2;
42
43 if (vtx_fmt_0 & R200_VTX_Z0)
44 vtx_size++;
45 if (vtx_fmt_0 & R200_VTX_W0)
46 vtx_size++;
47 /* blend weight */
48 if (vtx_fmt_0 & (0x7 << R200_VTX_WEIGHT_COUNT_SHIFT))
49 vtx_size += (vtx_fmt_0 >> R200_VTX_WEIGHT_COUNT_SHIFT) & 0x7;
50 if (vtx_fmt_0 & R200_VTX_PV_MATRIX_SEL)
51 vtx_size++;
52 if (vtx_fmt_0 & R200_VTX_N0)
53 vtx_size += 3;
54 if (vtx_fmt_0 & R200_VTX_POINT_SIZE)
55 vtx_size++;
56 if (vtx_fmt_0 & R200_VTX_DISCRETE_FOG)
57 vtx_size++;
58 if (vtx_fmt_0 & R200_VTX_SHININESS_0)
59 vtx_size++;
60 if (vtx_fmt_0 & R200_VTX_SHININESS_1)
61 vtx_size++;
62 for (i = 0; i < 8; i++) {
63 int color_size = (vtx_fmt_0 >> (11 + 2*i)) & 0x3;
64 switch (color_size) {
65 case 0: break;
66 case 1: vtx_size++; break;
67 case 2: vtx_size += 3; break;
68 case 3: vtx_size += 4; break;
69 }
70 }
71 if (vtx_fmt_0 & R200_VTX_XY1)
72 vtx_size += 2;
73 if (vtx_fmt_0 & R200_VTX_Z1)
74 vtx_size++;
75 if (vtx_fmt_0 & R200_VTX_W1)
76 vtx_size++;
77 if (vtx_fmt_0 & R200_VTX_N1)
78 vtx_size += 3;
79 return vtx_size;
80}
81
82static int r200_get_vtx_size_1(uint32_t vtx_fmt_1)
83{
84 int vtx_size, i, tex_size;
85 vtx_size = 0;
86 for (i = 0; i < 6; i++) {
87 tex_size = (vtx_fmt_1 >> (i * 3)) & 0x7;
88 if (tex_size > 4)
89 continue;
90 vtx_size += tex_size;
91 }
92 return vtx_size;
93}
94
95int r200_packet0_check(struct radeon_cs_parser *p,
96 struct radeon_cs_packet *pkt,
97 unsigned idx, unsigned reg)
98{
99 struct radeon_cs_chunk *ib_chunk;
100 struct radeon_cs_reloc *reloc;
101 struct r100_cs_track *track;
102 volatile uint32_t *ib;
103 uint32_t tmp;
104 int r;
105 int i;
106 int face;
107 u32 tile_flags = 0;
108
109 ib = p->ib->ptr;
110 ib_chunk = &p->chunks[p->chunk_ib_idx];
111 track = (struct r100_cs_track *)p->track;
112
113 switch (reg) {
114 case RADEON_CRTC_GUI_TRIG_VLINE:
115 r = r100_cs_packet_parse_vline(p);
116 if (r) {
117 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
118 idx, reg);
119 r100_cs_dump_packet(p, pkt);
120 return r;
121 }
122 break;
123 /* FIXME: only allow PACKET3 blit? easier to check for out of
124 * range access */
125 case RADEON_DST_PITCH_OFFSET:
126 case RADEON_SRC_PITCH_OFFSET:
127 r = r100_reloc_pitch_offset(p, pkt, idx, reg);
128 if (r)
129 return r;
130 break;
131 case RADEON_RB3D_DEPTHOFFSET:
132 r = r100_cs_packet_next_reloc(p, &reloc);
133 if (r) {
134 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
135 idx, reg);
136 r100_cs_dump_packet(p, pkt);
137 return r;
138 }
139 track->zb.robj = reloc->robj;
140 track->zb.offset = ib_chunk->kdata[idx];
141 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
142 break;
143 case RADEON_RB3D_COLOROFFSET:
144 r = r100_cs_packet_next_reloc(p, &reloc);
145 if (r) {
146 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
147 idx, reg);
148 r100_cs_dump_packet(p, pkt);
149 return r;
150 }
151 track->cb[0].robj = reloc->robj;
152 track->cb[0].offset = ib_chunk->kdata[idx];
153 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
154 break;
155 case R200_PP_TXOFFSET_0:
156 case R200_PP_TXOFFSET_1:
157 case R200_PP_TXOFFSET_2:
158 case R200_PP_TXOFFSET_3:
159 case R200_PP_TXOFFSET_4:
160 case R200_PP_TXOFFSET_5:
161 i = (reg - R200_PP_TXOFFSET_0) / 24;
162 r = r100_cs_packet_next_reloc(p, &reloc);
163 if (r) {
164 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
165 idx, reg);
166 r100_cs_dump_packet(p, pkt);
167 return r;
168 }
169 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
170 track->textures[i].robj = reloc->robj;
171 break;
172 case R200_PP_CUBIC_OFFSET_F1_0:
173 case R200_PP_CUBIC_OFFSET_F2_0:
174 case R200_PP_CUBIC_OFFSET_F3_0:
175 case R200_PP_CUBIC_OFFSET_F4_0:
176 case R200_PP_CUBIC_OFFSET_F5_0:
177 case R200_PP_CUBIC_OFFSET_F1_1:
178 case R200_PP_CUBIC_OFFSET_F2_1:
179 case R200_PP_CUBIC_OFFSET_F3_1:
180 case R200_PP_CUBIC_OFFSET_F4_1:
181 case R200_PP_CUBIC_OFFSET_F5_1:
182 case R200_PP_CUBIC_OFFSET_F1_2:
183 case R200_PP_CUBIC_OFFSET_F2_2:
184 case R200_PP_CUBIC_OFFSET_F3_2:
185 case R200_PP_CUBIC_OFFSET_F4_2:
186 case R200_PP_CUBIC_OFFSET_F5_2:
187 case R200_PP_CUBIC_OFFSET_F1_3:
188 case R200_PP_CUBIC_OFFSET_F2_3:
189 case R200_PP_CUBIC_OFFSET_F3_3:
190 case R200_PP_CUBIC_OFFSET_F4_3:
191 case R200_PP_CUBIC_OFFSET_F5_3:
192 case R200_PP_CUBIC_OFFSET_F1_4:
193 case R200_PP_CUBIC_OFFSET_F2_4:
194 case R200_PP_CUBIC_OFFSET_F3_4:
195 case R200_PP_CUBIC_OFFSET_F4_4:
196 case R200_PP_CUBIC_OFFSET_F5_4:
197 case R200_PP_CUBIC_OFFSET_F1_5:
198 case R200_PP_CUBIC_OFFSET_F2_5:
199 case R200_PP_CUBIC_OFFSET_F3_5:
200 case R200_PP_CUBIC_OFFSET_F4_5:
201 case R200_PP_CUBIC_OFFSET_F5_5:
202 i = (reg - R200_PP_TXOFFSET_0) / 24;
203 face = (reg - ((i * 24) + R200_PP_TXOFFSET_0)) / 4;
204 r = r100_cs_packet_next_reloc(p, &reloc);
205 if (r) {
206 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
207 idx, reg);
208 r100_cs_dump_packet(p, pkt);
209 return r;
210 }
211 track->textures[i].cube_info[face - 1].offset = ib_chunk->kdata[idx];
212 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
213 track->textures[i].cube_info[face - 1].robj = reloc->robj;
214 break;
215 case RADEON_RE_WIDTH_HEIGHT:
216 track->maxy = ((ib_chunk->kdata[idx] >> 16) & 0x7FF);
217 break;
218 case RADEON_RB3D_COLORPITCH:
219 r = r100_cs_packet_next_reloc(p, &reloc);
220 if (r) {
221 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
222 idx, reg);
223 r100_cs_dump_packet(p, pkt);
224 return r;
225 }
226
227 if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
228 tile_flags |= RADEON_COLOR_TILE_ENABLE;
229 if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
230 tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
231
232 tmp = ib_chunk->kdata[idx] & ~(0x7 << 16);
233 tmp |= tile_flags;
234 ib[idx] = tmp;
235
236 track->cb[0].pitch = ib_chunk->kdata[idx] & RADEON_COLORPITCH_MASK;
237 break;
238 case RADEON_RB3D_DEPTHPITCH:
239 track->zb.pitch = ib_chunk->kdata[idx] & RADEON_DEPTHPITCH_MASK;
240 break;
241 case RADEON_RB3D_CNTL:
242 switch ((ib_chunk->kdata[idx] >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) {
243 case 7:
244 case 8:
245 case 9:
246 case 11:
247 case 12:
248 track->cb[0].cpp = 1;
249 break;
250 case 3:
251 case 4:
252 case 15:
253 track->cb[0].cpp = 2;
254 break;
255 case 6:
256 track->cb[0].cpp = 4;
257 break;
258 default:
259 DRM_ERROR("Invalid color buffer format (%d) !\n",
260 ((ib_chunk->kdata[idx] >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f));
261 return -EINVAL;
262 }
263 if (ib_chunk->kdata[idx] & RADEON_DEPTHXY_OFFSET_ENABLE) {
264 DRM_ERROR("No support for depth xy offset in kms\n");
265 return -EINVAL;
266 }
267
268 track->z_enabled = !!(ib_chunk->kdata[idx] & RADEON_Z_ENABLE);
269 break;
270 case RADEON_RB3D_ZSTENCILCNTL:
271 switch (ib_chunk->kdata[idx] & 0xf) {
272 case 0:
273 track->zb.cpp = 2;
274 break;
275 case 2:
276 case 3:
277 case 4:
278 case 5:
279 case 9:
280 case 11:
281 track->zb.cpp = 4;
282 break;
283 default:
284 break;
285 }
286 break;
287 case RADEON_RB3D_ZPASS_ADDR:
288 r = r100_cs_packet_next_reloc(p, &reloc);
289 if (r) {
290 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
291 idx, reg);
292 r100_cs_dump_packet(p, pkt);
293 return r;
294 }
295 ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset);
296 break;
297 case RADEON_PP_CNTL:
298 {
299 uint32_t temp = ib_chunk->kdata[idx] >> 4;
300 for (i = 0; i < track->num_texture; i++)
301 track->textures[i].enabled = !!(temp & (1 << i));
302 }
303 break;
304 case RADEON_SE_VF_CNTL:
305 track->vap_vf_cntl = ib_chunk->kdata[idx];
306 break;
307 case 0x210c:
308 /* VAP_VF_MAX_VTX_INDX */
309 track->max_indx = ib_chunk->kdata[idx] & 0x00FFFFFFUL;
310 break;
311 case R200_SE_VTX_FMT_0:
312 track->vtx_size = r200_get_vtx_size_0(ib_chunk->kdata[idx]);
313 break;
314 case R200_SE_VTX_FMT_1:
315 track->vtx_size += r200_get_vtx_size_1(ib_chunk->kdata[idx]);
316 break;
317 case R200_PP_TXSIZE_0:
318 case R200_PP_TXSIZE_1:
319 case R200_PP_TXSIZE_2:
320 case R200_PP_TXSIZE_3:
321 case R200_PP_TXSIZE_4:
322 case R200_PP_TXSIZE_5:
323 i = (reg - R200_PP_TXSIZE_0) / 32;
324 track->textures[i].width = (ib_chunk->kdata[idx] & RADEON_TEX_USIZE_MASK) + 1;
325 track->textures[i].height = ((ib_chunk->kdata[idx] & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1;
326 break;
327 case R200_PP_TXPITCH_0:
328 case R200_PP_TXPITCH_1:
329 case R200_PP_TXPITCH_2:
330 case R200_PP_TXPITCH_3:
331 case R200_PP_TXPITCH_4:
332 case R200_PP_TXPITCH_5:
333 i = (reg - R200_PP_TXPITCH_0) / 32;
334 track->textures[i].pitch = ib_chunk->kdata[idx] + 32;
335 break;
336 case R200_PP_TXFILTER_0:
337 case R200_PP_TXFILTER_1:
338 case R200_PP_TXFILTER_2:
339 case R200_PP_TXFILTER_3:
340 case R200_PP_TXFILTER_4:
341 case R200_PP_TXFILTER_5:
342 i = (reg - R200_PP_TXFILTER_0) / 32;
343 track->textures[i].num_levels = ((ib_chunk->kdata[idx] & R200_MAX_MIP_LEVEL_MASK)
344 >> R200_MAX_MIP_LEVEL_SHIFT);
345 tmp = (ib_chunk->kdata[idx] >> 23) & 0x7;
346 if (tmp == 2 || tmp == 6)
347 track->textures[i].roundup_w = false;
348 tmp = (ib_chunk->kdata[idx] >> 27) & 0x7;
349 if (tmp == 2 || tmp == 6)
350 track->textures[i].roundup_h = false;
351 break;
352 case R200_PP_TXMULTI_CTL_0:
353 case R200_PP_TXMULTI_CTL_1:
354 case R200_PP_TXMULTI_CTL_2:
355 case R200_PP_TXMULTI_CTL_3:
356 case R200_PP_TXMULTI_CTL_4:
357 case R200_PP_TXMULTI_CTL_5:
358 i = (reg - R200_PP_TXMULTI_CTL_0) / 32;
359 break;
360 case R200_PP_TXFORMAT_X_0:
361 case R200_PP_TXFORMAT_X_1:
362 case R200_PP_TXFORMAT_X_2:
363 case R200_PP_TXFORMAT_X_3:
364 case R200_PP_TXFORMAT_X_4:
365 case R200_PP_TXFORMAT_X_5:
366 i = (reg - R200_PP_TXFORMAT_X_0) / 32;
367 track->textures[i].txdepth = ib_chunk->kdata[idx] & 0x7;
368 tmp = (ib_chunk->kdata[idx] >> 16) & 0x3;
369 /* 2D, 3D, CUBE */
370 switch (tmp) {
371 case 0:
372 case 5:
373 case 6:
374 case 7:
375 track->textures[i].tex_coord_type = 0;
376 break;
377 case 1:
378 track->textures[i].tex_coord_type = 1;
379 break;
380 case 2:
381 track->textures[i].tex_coord_type = 2;
382 break;
383 }
384 break;
385 case R200_PP_TXFORMAT_0:
386 case R200_PP_TXFORMAT_1:
387 case R200_PP_TXFORMAT_2:
388 case R200_PP_TXFORMAT_3:
389 case R200_PP_TXFORMAT_4:
390 case R200_PP_TXFORMAT_5:
391 i = (reg - R200_PP_TXFORMAT_0) / 32;
392 if (ib_chunk->kdata[idx] & R200_TXFORMAT_NON_POWER2) {
393 track->textures[i].use_pitch = 1;
394 } else {
395 track->textures[i].use_pitch = 0;
396 track->textures[i].width = 1 << ((ib_chunk->kdata[idx] >> RADEON_TXFORMAT_WIDTH_SHIFT) & RADEON_TXFORMAT_WIDTH_MASK);
397 track->textures[i].height = 1 << ((ib_chunk->kdata[idx] >> RADEON_TXFORMAT_HEIGHT_SHIFT) & RADEON_TXFORMAT_HEIGHT_MASK);
398 }
399 switch ((ib_chunk->kdata[idx] & RADEON_TXFORMAT_FORMAT_MASK)) {
400 case R200_TXFORMAT_I8:
401 case R200_TXFORMAT_RGB332:
402 case R200_TXFORMAT_Y8:
403 track->textures[i].cpp = 1;
404 break;
405 case R200_TXFORMAT_DXT1:
406 case R200_TXFORMAT_AI88:
407 case R200_TXFORMAT_ARGB1555:
408 case R200_TXFORMAT_RGB565:
409 case R200_TXFORMAT_ARGB4444:
410 case R200_TXFORMAT_VYUY422:
411 case R200_TXFORMAT_YVYU422:
412 case R200_TXFORMAT_LDVDU655:
413 case R200_TXFORMAT_DVDU88:
414 case R200_TXFORMAT_AVYU4444:
415 track->textures[i].cpp = 2;
416 break;
417 case R200_TXFORMAT_ARGB8888:
418 case R200_TXFORMAT_RGBA8888:
419 case R200_TXFORMAT_ABGR8888:
420 case R200_TXFORMAT_BGR111110:
421 case R200_TXFORMAT_LDVDU8888:
422 case R200_TXFORMAT_DXT23:
423 case R200_TXFORMAT_DXT45:
424 track->textures[i].cpp = 4;
425 break;
426 }
427 track->textures[i].cube_info[4].width = 1 << ((ib_chunk->kdata[idx] >> 16) & 0xf);
428 track->textures[i].cube_info[4].height = 1 << ((ib_chunk->kdata[idx] >> 20) & 0xf);
429 break;
430 case R200_PP_CUBIC_FACES_0:
431 case R200_PP_CUBIC_FACES_1:
432 case R200_PP_CUBIC_FACES_2:
433 case R200_PP_CUBIC_FACES_3:
434 case R200_PP_CUBIC_FACES_4:
435 case R200_PP_CUBIC_FACES_5:
436 tmp = ib_chunk->kdata[idx];
437 i = (reg - R200_PP_CUBIC_FACES_0) / 32;
438 for (face = 0; face < 4; face++) {
439 track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf);
440 track->textures[i].cube_info[face].height = 1 << ((tmp >> ((face * 8) + 4)) & 0xf);
441 }
442 break;
443 default:
444 printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n",
445 reg, idx);
446 return -EINVAL;
447 }
448 return 0;
449}
450
451int r200_init(struct radeon_device *rdev)
452{
453 rdev->config.r100.reg_safe_bm = r200_reg_safe_bm;
454 rdev->config.r100.reg_safe_bm_size = ARRAY_SIZE(r200_reg_safe_bm);
455 return 0;
456}
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 9f2460cf9dbc..33a2c557eac4 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -32,6 +32,7 @@
32#include "radeon.h" 32#include "radeon.h"
33#include "radeon_drm.h" 33#include "radeon_drm.h"
34#include "radeon_share.h" 34#include "radeon_share.h"
35#include "r100_track.h"
35 36
36#include "r300_reg_safe.h" 37#include "r300_reg_safe.h"
37 38
@@ -49,14 +50,10 @@ int r100_cs_packet_parse(struct radeon_cs_parser *p,
49 struct radeon_cs_packet *pkt, 50 struct radeon_cs_packet *pkt,
50 unsigned idx); 51 unsigned idx);
51int r100_cs_packet_parse_vline(struct radeon_cs_parser *p); 52int r100_cs_packet_parse_vline(struct radeon_cs_parser *p);
52int r100_cs_packet_next_reloc(struct radeon_cs_parser *p,
53 struct radeon_cs_reloc **cs_reloc);
54int r100_cs_parse_packet0(struct radeon_cs_parser *p, 53int r100_cs_parse_packet0(struct radeon_cs_parser *p,
55 struct radeon_cs_packet *pkt, 54 struct radeon_cs_packet *pkt,
56 const unsigned *auth, unsigned n, 55 const unsigned *auth, unsigned n,
57 radeon_packet0_check_t check); 56 radeon_packet0_check_t check);
58void r100_cs_dump_packet(struct radeon_cs_parser *p,
59 struct radeon_cs_packet *pkt);
60int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p, 57int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p,
61 struct radeon_cs_packet *pkt, 58 struct radeon_cs_packet *pkt,
62 struct radeon_object *robj); 59 struct radeon_object *robj);
@@ -706,264 +703,13 @@ int rv370_debugfs_pcie_gart_info_init(struct radeon_device *rdev)
706/* 703/*
707 * CS functions 704 * CS functions
708 */ 705 */
709struct r300_cs_track_cb {
710 struct radeon_object *robj;
711 unsigned pitch;
712 unsigned cpp;
713 unsigned offset;
714};
715
716struct r300_cs_track_array {
717 struct radeon_object *robj;
718 unsigned esize;
719};
720
721struct r300_cs_track_texture {
722 struct radeon_object *robj;
723 unsigned pitch;
724 unsigned width;
725 unsigned height;
726 unsigned num_levels;
727 unsigned cpp;
728 unsigned tex_coord_type;
729 unsigned txdepth;
730 unsigned width_11;
731 unsigned height_11;
732 bool use_pitch;
733 bool enabled;
734 bool roundup_w;
735 bool roundup_h;
736};
737
738struct r300_cs_track {
739 unsigned num_cb;
740 unsigned maxy;
741 unsigned vtx_size;
742 unsigned vap_vf_cntl;
743 unsigned immd_dwords;
744 unsigned num_arrays;
745 unsigned max_indx;
746 struct r300_cs_track_array arrays[11];
747 struct r300_cs_track_cb cb[4];
748 struct r300_cs_track_cb zb;
749 struct r300_cs_track_texture textures[16];
750 bool z_enabled;
751};
752
753static inline void r300_cs_track_texture_print(struct r300_cs_track_texture *t)
754{
755 DRM_ERROR("pitch %d\n", t->pitch);
756 DRM_ERROR("width %d\n", t->width);
757 DRM_ERROR("height %d\n", t->height);
758 DRM_ERROR("num levels %d\n", t->num_levels);
759 DRM_ERROR("depth %d\n", t->txdepth);
760 DRM_ERROR("bpp %d\n", t->cpp);
761 DRM_ERROR("coordinate type %d\n", t->tex_coord_type);
762 DRM_ERROR("width round to power of 2 %d\n", t->roundup_w);
763 DRM_ERROR("height round to power of 2 %d\n", t->roundup_h);
764}
765
766static inline int r300_cs_track_texture_check(struct radeon_device *rdev,
767 struct r300_cs_track *track)
768{
769 struct radeon_object *robj;
770 unsigned long size;
771 unsigned u, i, w, h;
772
773 for (u = 0; u < 16; u++) {
774 if (!track->textures[u].enabled)
775 continue;
776 robj = track->textures[u].robj;
777 if (robj == NULL) {
778 DRM_ERROR("No texture bound to unit %u\n", u);
779 return -EINVAL;
780 }
781 size = 0;
782 for (i = 0; i <= track->textures[u].num_levels; i++) {
783 if (track->textures[u].use_pitch) {
784 w = track->textures[u].pitch / (1 << i);
785 } else {
786 w = track->textures[u].width / (1 << i);
787 if (rdev->family >= CHIP_RV515)
788 w |= track->textures[u].width_11;
789 if (track->textures[u].roundup_w)
790 w = roundup_pow_of_two(w);
791 }
792 h = track->textures[u].height / (1 << i);
793 if (rdev->family >= CHIP_RV515)
794 h |= track->textures[u].height_11;
795 if (track->textures[u].roundup_h)
796 h = roundup_pow_of_two(h);
797 size += w * h;
798 }
799 size *= track->textures[u].cpp;
800 switch (track->textures[u].tex_coord_type) {
801 case 0:
802 break;
803 case 1:
804 size *= (1 << track->textures[u].txdepth);
805 break;
806 case 2:
807 size *= 6;
808 break;
809 default:
810 DRM_ERROR("Invalid texture coordinate type %u for unit "
811 "%u\n", track->textures[u].tex_coord_type, u);
812 return -EINVAL;
813 }
814 if (size > radeon_object_size(robj)) {
815 DRM_ERROR("Texture of unit %u needs %lu bytes but is "
816 "%lu\n", u, size, radeon_object_size(robj));
817 r300_cs_track_texture_print(&track->textures[u]);
818 return -EINVAL;
819 }
820 }
821 return 0;
822}
823
824int r300_cs_track_check(struct radeon_device *rdev, struct r300_cs_track *track)
825{
826 unsigned i;
827 unsigned long size;
828 unsigned prim_walk;
829 unsigned nverts;
830
831 for (i = 0; i < track->num_cb; i++) {
832 if (track->cb[i].robj == NULL) {
833 DRM_ERROR("[drm] No buffer for color buffer %d !\n", i);
834 return -EINVAL;
835 }
836 size = track->cb[i].pitch * track->cb[i].cpp * track->maxy;
837 size += track->cb[i].offset;
838 if (size > radeon_object_size(track->cb[i].robj)) {
839 DRM_ERROR("[drm] Buffer too small for color buffer %d "
840 "(need %lu have %lu) !\n", i, size,
841 radeon_object_size(track->cb[i].robj));
842 DRM_ERROR("[drm] color buffer %d (%u %u %u %u)\n",
843 i, track->cb[i].pitch, track->cb[i].cpp,
844 track->cb[i].offset, track->maxy);
845 return -EINVAL;
846 }
847 }
848 if (track->z_enabled) {
849 if (track->zb.robj == NULL) {
850 DRM_ERROR("[drm] No buffer for z buffer !\n");
851 return -EINVAL;
852 }
853 size = track->zb.pitch * track->zb.cpp * track->maxy;
854 size += track->zb.offset;
855 if (size > radeon_object_size(track->zb.robj)) {
856 DRM_ERROR("[drm] Buffer too small for z buffer "
857 "(need %lu have %lu) !\n", size,
858 radeon_object_size(track->zb.robj));
859 return -EINVAL;
860 }
861 }
862 prim_walk = (track->vap_vf_cntl >> 4) & 0x3;
863 nverts = (track->vap_vf_cntl >> 16) & 0xFFFF;
864 switch (prim_walk) {
865 case 1:
866 for (i = 0; i < track->num_arrays; i++) {
867 size = track->arrays[i].esize * track->max_indx * 4;
868 if (track->arrays[i].robj == NULL) {
869 DRM_ERROR("(PW %u) Vertex array %u no buffer "
870 "bound\n", prim_walk, i);
871 return -EINVAL;
872 }
873 if (size > radeon_object_size(track->arrays[i].robj)) {
874 DRM_ERROR("(PW %u) Vertex array %u need %lu dwords "
875 "have %lu dwords\n", prim_walk, i,
876 size >> 2,
877 radeon_object_size(track->arrays[i].robj) >> 2);
878 DRM_ERROR("Max indices %u\n", track->max_indx);
879 return -EINVAL;
880 }
881 }
882 break;
883 case 2:
884 for (i = 0; i < track->num_arrays; i++) {
885 size = track->arrays[i].esize * (nverts - 1) * 4;
886 if (track->arrays[i].robj == NULL) {
887 DRM_ERROR("(PW %u) Vertex array %u no buffer "
888 "bound\n", prim_walk, i);
889 return -EINVAL;
890 }
891 if (size > radeon_object_size(track->arrays[i].robj)) {
892 DRM_ERROR("(PW %u) Vertex array %u need %lu dwords "
893 "have %lu dwords\n", prim_walk, i, size >> 2,
894 radeon_object_size(track->arrays[i].robj) >> 2);
895 return -EINVAL;
896 }
897 }
898 break;
899 case 3:
900 size = track->vtx_size * nverts;
901 if (size != track->immd_dwords) {
902 DRM_ERROR("IMMD draw %u dwors but needs %lu dwords\n",
903 track->immd_dwords, size);
904 DRM_ERROR("VAP_VF_CNTL.NUM_VERTICES %u, VTX_SIZE %u\n",
905 nverts, track->vtx_size);
906 return -EINVAL;
907 }
908 break;
909 default:
910 DRM_ERROR("[drm] Invalid primitive walk %d for VAP_VF_CNTL\n",
911 prim_walk);
912 return -EINVAL;
913 }
914 return r300_cs_track_texture_check(rdev, track);
915}
916
917static inline void r300_cs_track_clear(struct r300_cs_track *track)
918{
919 unsigned i;
920
921 track->num_cb = 4;
922 track->maxy = 4096;
923 for (i = 0; i < track->num_cb; i++) {
924 track->cb[i].robj = NULL;
925 track->cb[i].pitch = 8192;
926 track->cb[i].cpp = 16;
927 track->cb[i].offset = 0;
928 }
929 track->z_enabled = true;
930 track->zb.robj = NULL;
931 track->zb.pitch = 8192;
932 track->zb.cpp = 4;
933 track->zb.offset = 0;
934 track->vtx_size = 0x7F;
935 track->immd_dwords = 0xFFFFFFFFUL;
936 track->num_arrays = 11;
937 track->max_indx = 0x00FFFFFFUL;
938 for (i = 0; i < track->num_arrays; i++) {
939 track->arrays[i].robj = NULL;
940 track->arrays[i].esize = 0x7F;
941 }
942 for (i = 0; i < 16; i++) {
943 track->textures[i].pitch = 16536;
944 track->textures[i].width = 16536;
945 track->textures[i].height = 16536;
946 track->textures[i].width_11 = 1 << 11;
947 track->textures[i].height_11 = 1 << 11;
948 track->textures[i].num_levels = 12;
949 track->textures[i].txdepth = 16;
950 track->textures[i].cpp = 64;
951 track->textures[i].tex_coord_type = 1;
952 track->textures[i].robj = NULL;
953 /* CS IB emission code makes sure texture unit are disabled */
954 track->textures[i].enabled = false;
955 track->textures[i].roundup_w = true;
956 track->textures[i].roundup_h = true;
957 }
958}
959
960static int r300_packet0_check(struct radeon_cs_parser *p, 706static int r300_packet0_check(struct radeon_cs_parser *p,
961 struct radeon_cs_packet *pkt, 707 struct radeon_cs_packet *pkt,
962 unsigned idx, unsigned reg) 708 unsigned idx, unsigned reg)
963{ 709{
964 struct radeon_cs_chunk *ib_chunk; 710 struct radeon_cs_chunk *ib_chunk;
965 struct radeon_cs_reloc *reloc; 711 struct radeon_cs_reloc *reloc;
966 struct r300_cs_track *track; 712 struct r100_cs_track *track;
967 volatile uint32_t *ib; 713 volatile uint32_t *ib;
968 uint32_t tmp, tile_flags = 0; 714 uint32_t tmp, tile_flags = 0;
969 unsigned i; 715 unsigned i;
@@ -971,7 +717,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
971 717
972 ib = p->ib->ptr; 718 ib = p->ib->ptr;
973 ib_chunk = &p->chunks[p->chunk_ib_idx]; 719 ib_chunk = &p->chunks[p->chunk_ib_idx];
974 track = (struct r300_cs_track*)p->track; 720 track = (struct r100_cs_track *)p->track;
975 switch(reg) { 721 switch(reg) {
976 case AVIVO_D1MODE_VLINE_START_END: 722 case AVIVO_D1MODE_VLINE_START_END:
977 case RADEON_CRTC_GUI_TRIG_VLINE: 723 case RADEON_CRTC_GUI_TRIG_VLINE:
@@ -985,28 +731,9 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
985 break; 731 break;
986 case RADEON_DST_PITCH_OFFSET: 732 case RADEON_DST_PITCH_OFFSET:
987 case RADEON_SRC_PITCH_OFFSET: 733 case RADEON_SRC_PITCH_OFFSET:
988 r = r100_cs_packet_next_reloc(p, &reloc); 734 r = r100_reloc_pitch_offset(p, pkt, idx, reg);
989 if (r) { 735 if (r)
990 DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
991 idx, reg);
992 r100_cs_dump_packet(p, pkt);
993 return r; 736 return r;
994 }
995 tmp = ib_chunk->kdata[idx] & 0x003fffff;
996 tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
997
998 if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
999 tile_flags |= RADEON_DST_TILE_MACRO;
1000 if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
1001 if (reg == RADEON_SRC_PITCH_OFFSET) {
1002 DRM_ERROR("Cannot src blit from microtiled surface\n");
1003 r100_cs_dump_packet(p, pkt);
1004 return -EINVAL;
1005 }
1006 tile_flags |= RADEON_DST_TILE_MICRO;
1007 }
1008 tmp |= tile_flags;
1009 ib[idx] = (ib_chunk->kdata[idx] & 0x3fc00000) | tmp;
1010 break; 737 break;
1011 case R300_RB3D_COLOROFFSET0: 738 case R300_RB3D_COLOROFFSET0:
1012 case R300_RB3D_COLOROFFSET1: 739 case R300_RB3D_COLOROFFSET1:
@@ -1215,42 +942,41 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
1215 tmp = (ib_chunk->kdata[idx] >> 25) & 0x3; 942 tmp = (ib_chunk->kdata[idx] >> 25) & 0x3;
1216 track->textures[i].tex_coord_type = tmp; 943 track->textures[i].tex_coord_type = tmp;
1217 switch ((ib_chunk->kdata[idx] & 0x1F)) { 944 switch ((ib_chunk->kdata[idx] & 0x1F)) {
1218 case 0: 945 case R300_TX_FORMAT_X8:
1219 case 2: 946 case R300_TX_FORMAT_Y4X4:
1220 case 5: 947 case R300_TX_FORMAT_Z3Y3X2:
1221 case 18:
1222 case 20:
1223 case 21:
1224 track->textures[i].cpp = 1; 948 track->textures[i].cpp = 1;
1225 break; 949 break;
1226 case 1: 950 case R300_TX_FORMAT_X16:
1227 case 3: 951 case R300_TX_FORMAT_Y8X8:
1228 case 6: 952 case R300_TX_FORMAT_Z5Y6X5:
1229 case 7: 953 case R300_TX_FORMAT_Z6Y5X5:
1230 case 10: 954 case R300_TX_FORMAT_W4Z4Y4X4:
1231 case 11: 955 case R300_TX_FORMAT_W1Z5Y5X5:
1232 case 19: 956 case R300_TX_FORMAT_DXT1:
1233 case 22: 957 case R300_TX_FORMAT_D3DMFT_CxV8U8:
1234 case 24: 958 case R300_TX_FORMAT_B8G8_B8G8:
959 case R300_TX_FORMAT_G8R8_G8B8:
1235 track->textures[i].cpp = 2; 960 track->textures[i].cpp = 2;
1236 break; 961 break;
1237 case 4: 962 case R300_TX_FORMAT_Y16X16:
1238 case 8: 963 case R300_TX_FORMAT_Z11Y11X10:
1239 case 9: 964 case R300_TX_FORMAT_Z10Y11X11:
1240 case 12: 965 case R300_TX_FORMAT_W8Z8Y8X8:
1241 case 13: 966 case R300_TX_FORMAT_W2Z10Y10X10:
1242 case 23: 967 case 0x17:
1243 case 25: 968 case R300_TX_FORMAT_FL_I32:
1244 case 27: 969 case 0x1e:
1245 case 30: 970 case R300_TX_FORMAT_DXT3:
971 case R300_TX_FORMAT_DXT5:
1246 track->textures[i].cpp = 4; 972 track->textures[i].cpp = 4;
1247 break; 973 break;
1248 case 14: 974 case R300_TX_FORMAT_W16Z16Y16X16:
1249 case 26: 975 case R300_TX_FORMAT_FL_R16G16B16A16:
1250 case 28: 976 case R300_TX_FORMAT_FL_I32A32:
1251 track->textures[i].cpp = 8; 977 track->textures[i].cpp = 8;
1252 break; 978 break;
1253 case 29: 979 case R300_TX_FORMAT_FL_R32G32B32A32:
1254 track->textures[i].cpp = 16; 980 track->textures[i].cpp = 16;
1255 break; 981 break;
1256 default: 982 default:
@@ -1278,11 +1004,11 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
1278 case 0x443C: 1004 case 0x443C:
1279 /* TX_FILTER0_[0-15] */ 1005 /* TX_FILTER0_[0-15] */
1280 i = (reg - 0x4400) >> 2; 1006 i = (reg - 0x4400) >> 2;
1281 tmp = ib_chunk->kdata[idx] & 0x7;; 1007 tmp = ib_chunk->kdata[idx] & 0x7;
1282 if (tmp == 2 || tmp == 4 || tmp == 6) { 1008 if (tmp == 2 || tmp == 4 || tmp == 6) {
1283 track->textures[i].roundup_w = false; 1009 track->textures[i].roundup_w = false;
1284 } 1010 }
1285 tmp = (ib_chunk->kdata[idx] >> 3) & 0x7;; 1011 tmp = (ib_chunk->kdata[idx] >> 3) & 0x7;
1286 if (tmp == 2 || tmp == 4 || tmp == 6) { 1012 if (tmp == 2 || tmp == 4 || tmp == 6) {
1287 track->textures[i].roundup_h = false; 1013 track->textures[i].roundup_h = false;
1288 } 1014 }
@@ -1370,8 +1096,9 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
1370 struct radeon_cs_packet *pkt) 1096 struct radeon_cs_packet *pkt)
1371{ 1097{
1372 struct radeon_cs_chunk *ib_chunk; 1098 struct radeon_cs_chunk *ib_chunk;
1099
1373 struct radeon_cs_reloc *reloc; 1100 struct radeon_cs_reloc *reloc;
1374 struct r300_cs_track *track; 1101 struct r100_cs_track *track;
1375 volatile uint32_t *ib; 1102 volatile uint32_t *ib;
1376 unsigned idx; 1103 unsigned idx;
1377 unsigned i, c; 1104 unsigned i, c;
@@ -1380,7 +1107,7 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
1380 ib = p->ib->ptr; 1107 ib = p->ib->ptr;
1381 ib_chunk = &p->chunks[p->chunk_ib_idx]; 1108 ib_chunk = &p->chunks[p->chunk_ib_idx];
1382 idx = pkt->idx + 1; 1109 idx = pkt->idx + 1;
1383 track = (struct r300_cs_track*)p->track; 1110 track = (struct r100_cs_track *)p->track;
1384 switch(pkt->opcode) { 1111 switch(pkt->opcode) {
1385 case PACKET3_3D_LOAD_VBPNTR: 1112 case PACKET3_3D_LOAD_VBPNTR:
1386 c = ib_chunk->kdata[idx++] & 0x1F; 1113 c = ib_chunk->kdata[idx++] & 0x1F;
@@ -1447,7 +1174,7 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
1447 } 1174 }
1448 track->vap_vf_cntl = ib_chunk->kdata[idx+1]; 1175 track->vap_vf_cntl = ib_chunk->kdata[idx+1];
1449 track->immd_dwords = pkt->count - 1; 1176 track->immd_dwords = pkt->count - 1;
1450 r = r300_cs_track_check(p->rdev, track); 1177 r = r100_cs_track_check(p->rdev, track);
1451 if (r) { 1178 if (r) {
1452 return r; 1179 return r;
1453 } 1180 }
@@ -1462,35 +1189,35 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
1462 } 1189 }
1463 track->vap_vf_cntl = ib_chunk->kdata[idx]; 1190 track->vap_vf_cntl = ib_chunk->kdata[idx];
1464 track->immd_dwords = pkt->count; 1191 track->immd_dwords = pkt->count;
1465 r = r300_cs_track_check(p->rdev, track); 1192 r = r100_cs_track_check(p->rdev, track);
1466 if (r) { 1193 if (r) {
1467 return r; 1194 return r;
1468 } 1195 }
1469 break; 1196 break;
1470 case PACKET3_3D_DRAW_VBUF: 1197 case PACKET3_3D_DRAW_VBUF:
1471 track->vap_vf_cntl = ib_chunk->kdata[idx + 1]; 1198 track->vap_vf_cntl = ib_chunk->kdata[idx + 1];
1472 r = r300_cs_track_check(p->rdev, track); 1199 r = r100_cs_track_check(p->rdev, track);
1473 if (r) { 1200 if (r) {
1474 return r; 1201 return r;
1475 } 1202 }
1476 break; 1203 break;
1477 case PACKET3_3D_DRAW_VBUF_2: 1204 case PACKET3_3D_DRAW_VBUF_2:
1478 track->vap_vf_cntl = ib_chunk->kdata[idx]; 1205 track->vap_vf_cntl = ib_chunk->kdata[idx];
1479 r = r300_cs_track_check(p->rdev, track); 1206 r = r100_cs_track_check(p->rdev, track);
1480 if (r) { 1207 if (r) {
1481 return r; 1208 return r;
1482 } 1209 }
1483 break; 1210 break;
1484 case PACKET3_3D_DRAW_INDX: 1211 case PACKET3_3D_DRAW_INDX:
1485 track->vap_vf_cntl = ib_chunk->kdata[idx + 1]; 1212 track->vap_vf_cntl = ib_chunk->kdata[idx + 1];
1486 r = r300_cs_track_check(p->rdev, track); 1213 r = r100_cs_track_check(p->rdev, track);
1487 if (r) { 1214 if (r) {
1488 return r; 1215 return r;
1489 } 1216 }
1490 break; 1217 break;
1491 case PACKET3_3D_DRAW_INDX_2: 1218 case PACKET3_3D_DRAW_INDX_2:
1492 track->vap_vf_cntl = ib_chunk->kdata[idx]; 1219 track->vap_vf_cntl = ib_chunk->kdata[idx];
1493 r = r300_cs_track_check(p->rdev, track); 1220 r = r100_cs_track_check(p->rdev, track);
1494 if (r) { 1221 if (r) {
1495 return r; 1222 return r;
1496 } 1223 }
@@ -1507,10 +1234,10 @@ static int r300_packet3_check(struct radeon_cs_parser *p,
1507int r300_cs_parse(struct radeon_cs_parser *p) 1234int r300_cs_parse(struct radeon_cs_parser *p)
1508{ 1235{
1509 struct radeon_cs_packet pkt; 1236 struct radeon_cs_packet pkt;
1510 struct r300_cs_track track; 1237 struct r100_cs_track track;
1511 int r; 1238 int r;
1512 1239
1513 r300_cs_track_clear(&track); 1240 r100_cs_track_clear(p->rdev, &track);
1514 p->track = &track; 1241 p->track = &track;
1515 do { 1242 do {
1516 r = r100_cs_packet_parse(p, &pkt, p->idx); 1243 r = r100_cs_packet_parse(p, &pkt, p->idx);
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 149974d13aa2..6c35c3c29191 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -604,8 +604,14 @@ struct radeon_asic {
604 void (*bandwidth_update)(struct radeon_device *rdev); 604 void (*bandwidth_update)(struct radeon_device *rdev);
605}; 605};
606 606
607struct r100_asic {
608 const unsigned *reg_safe_bm;
609 unsigned reg_safe_bm_size;
610};
611
607union radeon_asic_config { 612union radeon_asic_config {
608 struct r300_asic r300; 613 struct r300_asic r300;
614 struct r100_asic r100;
609}; 615};
610 616
611 617
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 8ace15156c47..c9cbd8ae1f95 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -42,6 +42,7 @@ void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable);
42 * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 42 * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280
43 */ 43 */
44int r100_init(struct radeon_device *rdev); 44int r100_init(struct radeon_device *rdev);
45int r200_init(struct radeon_device *rdev);
45uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg); 46uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg);
46void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); 47void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
47void r100_errata(struct radeon_device *rdev); 48void r100_errata(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index 4df43f62c678..404b39bf3430 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -1945,6 +1945,11 @@
1945# define RADEON_TXFORMAT_DXT1 (12 << 0) 1945# define RADEON_TXFORMAT_DXT1 (12 << 0)
1946# define RADEON_TXFORMAT_DXT23 (14 << 0) 1946# define RADEON_TXFORMAT_DXT23 (14 << 0)
1947# define RADEON_TXFORMAT_DXT45 (15 << 0) 1947# define RADEON_TXFORMAT_DXT45 (15 << 0)
1948# define RADEON_TXFORMAT_SHADOW16 (16 << 0)
1949# define RADEON_TXFORMAT_SHADOW32 (17 << 0)
1950# define RADEON_TXFORMAT_DUDV88 (18 << 0)
1951# define RADEON_TXFORMAT_LDUDV655 (19 << 0)
1952# define RADEON_TXFORMAT_LDUDUV8888 (20 << 0)
1948# define RADEON_TXFORMAT_FORMAT_MASK (31 << 0) 1953# define RADEON_TXFORMAT_FORMAT_MASK (31 << 0)
1949# define RADEON_TXFORMAT_FORMAT_SHIFT 0 1954# define RADEON_TXFORMAT_FORMAT_SHIFT 0
1950# define RADEON_TXFORMAT_APPLE_YUV_MODE (1 << 5) 1955# define RADEON_TXFORMAT_APPLE_YUV_MODE (1 << 5)
@@ -2203,7 +2208,7 @@
2203# define RADEON_ROP_ENABLE (1 << 6) 2208# define RADEON_ROP_ENABLE (1 << 6)
2204# define RADEON_STENCIL_ENABLE (1 << 7) 2209# define RADEON_STENCIL_ENABLE (1 << 7)
2205# define RADEON_Z_ENABLE (1 << 8) 2210# define RADEON_Z_ENABLE (1 << 8)
2206# define RADEON_DEPTH_XZ_OFFEST_ENABLE (1 << 9) 2211# define RADEON_DEPTHXY_OFFSET_ENABLE (1 << 9)
2207# define RADEON_RB3D_COLOR_FORMAT_SHIFT 10 2212# define RADEON_RB3D_COLOR_FORMAT_SHIFT 10
2208 2213
2209# define RADEON_COLOR_FORMAT_ARGB1555 3 2214# define RADEON_COLOR_FORMAT_ARGB1555 3
@@ -2773,7 +2778,12 @@
2773# define R200_TXFORMAT_DXT1 (12 << 0) 2778# define R200_TXFORMAT_DXT1 (12 << 0)
2774# define R200_TXFORMAT_DXT23 (14 << 0) 2779# define R200_TXFORMAT_DXT23 (14 << 0)
2775# define R200_TXFORMAT_DXT45 (15 << 0) 2780# define R200_TXFORMAT_DXT45 (15 << 0)
2781# define R200_TXFORMAT_DVDU88 (18 << 0)
2782# define R200_TXFORMAT_LDVDU655 (19 << 0)
2783# define R200_TXFORMAT_LDVDU8888 (20 << 0)
2784# define R200_TXFORMAT_GR1616 (21 << 0)
2776# define R200_TXFORMAT_ABGR8888 (22 << 0) 2785# define R200_TXFORMAT_ABGR8888 (22 << 0)
2786# define R200_TXFORMAT_BGR111110 (23 << 0)
2777# define R200_TXFORMAT_FORMAT_MASK (31 << 0) 2787# define R200_TXFORMAT_FORMAT_MASK (31 << 0)
2778# define R200_TXFORMAT_FORMAT_SHIFT 0 2788# define R200_TXFORMAT_FORMAT_SHIFT 0
2779# define R200_TXFORMAT_ALPHA_IN_MAP (1 << 6) 2789# define R200_TXFORMAT_ALPHA_IN_MAP (1 << 6)
@@ -2818,6 +2828,13 @@
2818#define R200_PP_TXPITCH_4 0x2c90 /* NPOT only */ 2828#define R200_PP_TXPITCH_4 0x2c90 /* NPOT only */
2819#define R200_PP_TXPITCH_5 0x2cb0 /* NPOT only */ 2829#define R200_PP_TXPITCH_5 0x2cb0 /* NPOT only */
2820 2830
2831#define R200_PP_CUBIC_FACES_0 0x2c18
2832#define R200_PP_CUBIC_FACES_1 0x2c38
2833#define R200_PP_CUBIC_FACES_2 0x2c58
2834#define R200_PP_CUBIC_FACES_3 0x2c78
2835#define R200_PP_CUBIC_FACES_4 0x2c98
2836#define R200_PP_CUBIC_FACES_5 0x2cb8
2837
2821#define R200_PP_TXOFFSET_0 0x2d00 2838#define R200_PP_TXOFFSET_0 0x2d00
2822# define R200_TXO_ENDIAN_NO_SWAP (0 << 0) 2839# define R200_TXO_ENDIAN_NO_SWAP (0 << 0)
2823# define R200_TXO_ENDIAN_BYTE_SWAP (1 << 0) 2840# define R200_TXO_ENDIAN_BYTE_SWAP (1 << 0)
@@ -2829,11 +2846,44 @@
2829# define R200_TXO_MICRO_TILE (1 << 3) 2846# define R200_TXO_MICRO_TILE (1 << 3)
2830# define R200_TXO_OFFSET_MASK 0xffffffe0 2847# define R200_TXO_OFFSET_MASK 0xffffffe0
2831# define R200_TXO_OFFSET_SHIFT 5 2848# define R200_TXO_OFFSET_SHIFT 5
2849#define R200_PP_CUBIC_OFFSET_F1_0 0x2d04
2850#define R200_PP_CUBIC_OFFSET_F2_0 0x2d08
2851#define R200_PP_CUBIC_OFFSET_F3_0 0x2d0c
2852#define R200_PP_CUBIC_OFFSET_F4_0 0x2d10
2853#define R200_PP_CUBIC_OFFSET_F5_0 0x2d14
2854
2832#define R200_PP_TXOFFSET_1 0x2d18 2855#define R200_PP_TXOFFSET_1 0x2d18
2856#define R200_PP_CUBIC_OFFSET_F1_1 0x2d1c
2857#define R200_PP_CUBIC_OFFSET_F2_1 0x2d20
2858#define R200_PP_CUBIC_OFFSET_F3_1 0x2d24
2859#define R200_PP_CUBIC_OFFSET_F4_1 0x2d28
2860#define R200_PP_CUBIC_OFFSET_F5_1 0x2d2c
2861
2833#define R200_PP_TXOFFSET_2 0x2d30 2862#define R200_PP_TXOFFSET_2 0x2d30
2863#define R200_PP_CUBIC_OFFSET_F1_2 0x2d34
2864#define R200_PP_CUBIC_OFFSET_F2_2 0x2d38
2865#define R200_PP_CUBIC_OFFSET_F3_2 0x2d3c
2866#define R200_PP_CUBIC_OFFSET_F4_2 0x2d40
2867#define R200_PP_CUBIC_OFFSET_F5_2 0x2d44
2868
2834#define R200_PP_TXOFFSET_3 0x2d48 2869#define R200_PP_TXOFFSET_3 0x2d48
2870#define R200_PP_CUBIC_OFFSET_F1_3 0x2d4c
2871#define R200_PP_CUBIC_OFFSET_F2_3 0x2d50
2872#define R200_PP_CUBIC_OFFSET_F3_3 0x2d54
2873#define R200_PP_CUBIC_OFFSET_F4_3 0x2d58
2874#define R200_PP_CUBIC_OFFSET_F5_3 0x2d5c
2835#define R200_PP_TXOFFSET_4 0x2d60 2875#define R200_PP_TXOFFSET_4 0x2d60
2876#define R200_PP_CUBIC_OFFSET_F1_4 0x2d64
2877#define R200_PP_CUBIC_OFFSET_F2_4 0x2d68
2878#define R200_PP_CUBIC_OFFSET_F3_4 0x2d6c
2879#define R200_PP_CUBIC_OFFSET_F4_4 0x2d70
2880#define R200_PP_CUBIC_OFFSET_F5_4 0x2d74
2836#define R200_PP_TXOFFSET_5 0x2d78 2881#define R200_PP_TXOFFSET_5 0x2d78
2882#define R200_PP_CUBIC_OFFSET_F1_5 0x2d7c
2883#define R200_PP_CUBIC_OFFSET_F2_5 0x2d80
2884#define R200_PP_CUBIC_OFFSET_F3_5 0x2d84
2885#define R200_PP_CUBIC_OFFSET_F4_5 0x2d88
2886#define R200_PP_CUBIC_OFFSET_F5_5 0x2d8c
2837 2887
2838#define R200_PP_TFACTOR_0 0x2ee0 2888#define R200_PP_TFACTOR_0 0x2ee0
2839#define R200_PP_TFACTOR_1 0x2ee4 2889#define R200_PP_TFACTOR_1 0x2ee4
@@ -3175,6 +3225,11 @@
3175# define R200_FORCE_INORDER_PROC (1<<31) 3225# define R200_FORCE_INORDER_PROC (1<<31)
3176#define R200_PP_CNTL_X 0x2cc4 3226#define R200_PP_CNTL_X 0x2cc4
3177#define R200_PP_TXMULTI_CTL_0 0x2c1c 3227#define R200_PP_TXMULTI_CTL_0 0x2c1c
3228#define R200_PP_TXMULTI_CTL_1 0x2c3c
3229#define R200_PP_TXMULTI_CTL_2 0x2c5c
3230#define R200_PP_TXMULTI_CTL_3 0x2c7c
3231#define R200_PP_TXMULTI_CTL_4 0x2c9c
3232#define R200_PP_TXMULTI_CTL_5 0x2cbc
3178#define R200_SE_VTX_STATE_CNTL 0x2180 3233#define R200_SE_VTX_STATE_CNTL 0x2180
3179# define R200_UPDATE_USER_COLOR_0_ENA_MASK (1<<16) 3234# define R200_UPDATE_USER_COLOR_0_ENA_MASK (1<<16)
3180 3235
diff --git a/drivers/gpu/drm/radeon/reg_srcs/r100 b/drivers/gpu/drm/radeon/reg_srcs/r100
new file mode 100644
index 000000000000..f7ee062f1184
--- /dev/null
+++ b/drivers/gpu/drm/radeon/reg_srcs/r100
@@ -0,0 +1,105 @@
1r100 0x3294
20x1434 SRC_Y_X
30x1438 DST_Y_X
40x143C DST_HEIGHT_WIDTH
50x146C DP_GUI_MASTER_CNTL
60x1474 BRUSH_Y_X
70x1478 DP_BRUSH_BKGD_CLR
80x147C DP_BRUSH_FRGD_CLR
90x1480 BRUSH_DATA0
100x1484 BRUSH_DATA1
110x1598 DST_WIDTH_HEIGHT
120x15C0 CLR_CMP_CNTL
130x15C4 CLR_CMP_CLR_SRC
140x15C8 CLR_CMP_CLR_DST
150x15CC CLR_CMP_MSK
160x15D8 DP_SRC_FRGD_CLR
170x15DC DP_SRC_BKGD_CLR
180x1600 DST_LINE_START
190x1604 DST_LINE_END
200x1608 DST_LINE_PATCOUNT
210x16C0 DP_CNTL
220x16CC DP_WRITE_MSK
230x16D0 DP_CNTL_XDIR_YDIR_YMAJOR
240x16E8 DEFAULT_SC_BOTTOM_RIGHT
250x16EC SC_TOP_LEFT
260x16F0 SC_BOTTOM_RIGHT
270x16F4 SRC_SC_BOTTOM_RIGHT
280x1714 DSTCACHE_CTLSTAT
290x1720 WAIT_UNTIL
300x172C RBBM_GUICNTL
310x1810 FOG_3D_TABLE_START
320x1814 FOG_3D_TABLE_END
330x1a14 FOG_TABLE_INDEX
340x1a18 FOG_TABLE_DATA
350x1c14 PP_MISC
360x1c18 PP_FOG_COLOR
370x1c1c RE_SOLID_COLOR
380x1c20 RB3D_BLENDCNTL
390x1c4c SE_CNTL
400x1c50 SE_COORD_FMT
410x1c60 PP_TXCBLEND_0
420x1c64 PP_TXABLEND_0
430x1c68 PP_TFACTOR_0
440x1c78 PP_TXCBLEND_1
450x1c7c PP_TXABLEND_1
460x1c80 PP_TFACTOR_1
470x1c90 PP_TXCBLEND_2
480x1c94 PP_TXABLEND_2
490x1c98 PP_TFACTOR_2
500x1cc8 RE_STIPPLE_ADDR
510x1ccc RE_STIPPLE_DATA
520x1cd0 RE_LINE_PATTERN
530x1cd4 RE_LINE_STATE
540x1d40 PP_BORDER_COLOR0
550x1d44 PP_BORDER_COLOR1
560x1d48 PP_BORDER_COLOR2
570x1d7c RB3D_STENCILREFMASK
580x1d80 RB3D_ROPCNTL
590x1d84 RB3D_PLANEMASK
600x1d98 VAP_VPORT_XSCALE
610x1d9C VAP_VPORT_XOFFSET
620x1da0 VAP_VPORT_YSCALE
630x1da4 VAP_VPORT_YOFFSET
640x1da8 VAP_VPORT_ZSCALE
650x1dac VAP_VPORT_ZOFFSET
660x1db0 SE_ZBIAS_FACTOR
670x1db4 SE_ZBIAS_CONSTANT
680x1db8 SE_LINE_WIDTH
690x2140 SE_CNTL_STATUS
700x2200 SE_TCL_VECTOR_INDX_REG
710x2204 SE_TCL_VECTOR_DATA_REG
720x2208 SE_TCL_SCALAR_INDX_REG
730x220c SE_TCL_SCALAR_DATA_REG
740x2210 SE_TCL_MATERIAL_EMISSIVE_RED
750x2214 SE_TCL_MATERIAL_EMISSIVE_GREEN
760x2218 SE_TCL_MATERIAL_EMISSIVE_BLUE
770x221c SE_TCL_MATERIAL_EMISSIVE_ALPHA
780x2220 SE_TCL_MATERIAL_AMBIENT_RED
790x2224 SE_TCL_MATERIAL_AMBIENT_GREEN
800x2228 SE_TCL_MATERIAL_AMBIENT_BLUE
810x222c SE_TCL_MATERIAL_AMBIENT_ALPHA
820x2230 SE_TCL_MATERIAL_DIFFUSE_RED
830x2234 SE_TCL_MATERIAL_DIFFUSE_GREEN
840x2238 SE_TCL_MATERIAL_DIFFUSE_BLUE
850x223c SE_TCL_MATERIAL_DIFFUSE_ALPHA
860x2240 SE_TCL_MATERIAL_SPECULAR_RED
870x2244 SE_TCL_MATERIAL_SPECULAR_GREEN
880x2248 SE_TCL_MATERIAL_SPECULAR_BLUE
890x224c SE_TCL_MATERIAL_SPECULAR_ALPHA
900x2250 SE_TCL_SHININESS
910x2254 SE_TCL_OUTPUT_VTX_FMT
920x2258 SE_TCL_OUTPUT_VTX_SEL
930x225c SE_TCL_MATRIX_SELECT_0
940x2260 SE_TCL_MATRIX_SELECT_1
950x2264 SE_TCL_UCP_VERT_BLEND_CNTL
960x2268 SE_TCL_TEXTURE_PROC_CTL
970x226c SE_TCL_LIGHT_MODEL_CTL
980x2270 SE_TCL_PER_LIGHT_CTL_0
990x2274 SE_TCL_PER_LIGHT_CTL_1
1000x2278 SE_TCL_PER_LIGHT_CTL_2
1010x227c SE_TCL_PER_LIGHT_CTL_3
1020x2284 SE_TCL_STATE_FLUSH
1030x26c0 RE_TOP_LEFT
1040x26c4 RE_MISC
1050x3290 RB3D_ZPASS_DATA
diff --git a/drivers/gpu/drm/radeon/reg_srcs/r200 b/drivers/gpu/drm/radeon/reg_srcs/r200
new file mode 100644
index 000000000000..6021c8849a16
--- /dev/null
+++ b/drivers/gpu/drm/radeon/reg_srcs/r200
@@ -0,0 +1,184 @@
1r200 0x3294
20x1434 SRC_Y_X
30x1438 DST_Y_X
40x143C DST_HEIGHT_WIDTH
50x146C DP_GUI_MASTER_CNTL
60x1474 BRUSH_Y_X
70x1478 DP_BRUSH_BKGD_CLR
80x147C DP_BRUSH_FRGD_CLR
90x1480 BRUSH_DATA0
100x1484 BRUSH_DATA1
110x1598 DST_WIDTH_HEIGHT
120x15C0 CLR_CMP_CNTL
130x15C4 CLR_CMP_CLR_SRC
140x15C8 CLR_CMP_CLR_DST
150x15CC CLR_CMP_MSK
160x15D8 DP_SRC_FRGD_CLR
170x15DC DP_SRC_BKGD_CLR
180x1600 DST_LINE_START
190x1604 DST_LINE_END
200x1608 DST_LINE_PATCOUNT
210x16C0 DP_CNTL
220x16CC DP_WRITE_MSK
230x16D0 DP_CNTL_XDIR_YDIR_YMAJOR
240x16E8 DEFAULT_SC_BOTTOM_RIGHT
250x16EC SC_TOP_LEFT
260x16F0 SC_BOTTOM_RIGHT
270x16F4 SRC_SC_BOTTOM_RIGHT
280x1714 DSTCACHE_CTLSTAT
290x1720 WAIT_UNTIL
300x172C RBBM_GUICNTL
310x1c14 PP_MISC
320x1c18 PP_FOG_COLOR
330x1c1c RE_SOLID_COLOR
340x1c20 RB3D_BLENDCNTL
350x1c4c SE_CNTL
360x1c50 RE_CNTL
370x1cc8 RE_STIPPLE_ADDR
380x1ccc RE_STIPPLE_DATA
390x1cd0 RE_LINE_PATTERN
400x1cd4 RE_LINE_STATE
410x1cd8 RE_SCISSOR_TL_0
420x1cdc RE_SCISSOR_BR_0
430x1ce0 RE_SCISSOR_TL_1
440x1ce4 RE_SCISSOR_BR_1
450x1ce8 RE_SCISSOR_TL_2
460x1cec RE_SCISSOR_BR_2
470x1d60 RB3D_DEPTHXY_OFFSET
480x1d7c RB3D_STENCILREFMASK
490x1d80 RB3D_ROPCNTL
500x1d84 RB3D_PLANEMASK
510x1d98 VAP_VPORT_XSCALE
520x1d9c VAP_VPORT_XOFFSET
530x1da0 VAP_VPORT_YSCALE
540x1da4 VAP_VPORT_YOFFSET
550x1da8 VAP_VPORT_ZSCALE
560x1dac VAP_VPORT_ZOFFSET
570x1db0 SE_ZBIAS_FACTOR
580x1db4 SE_ZBIAS_CONSTANT
590x1db8 SE_LINE_WIDTH
600x2080 SE_VAP_CNTL
610x2090 SE_TCL_OUTPUT_VTX_FMT_0
620x2094 SE_TCL_OUTPUT_VTX_FMT_1
630x20b0 SE_VTE_CNTL
640x2140 SE_CNTL_STATUS
650x2180 SE_VTX_STATE_CNTL
660x2200 SE_TCL_VECTOR_INDX_REG
670x2204 SE_TCL_VECTOR_DATA_REG
680x2208 SE_TCL_SCALAR_INDX_REG
690x220c SE_TCL_SCALAR_DATA_REG
700x2230 SE_TCL_MATRIX_SEL_0
710x2234 SE_TCL_MATRIX_SEL_1
720x2238 SE_TCL_MATRIX_SEL_2
730x223c SE_TCL_MATRIX_SEL_3
740x2240 SE_TCL_MATRIX_SEL_4
750x2250 SE_TCL_OUTPUT_VTX_COMP_SEL
760x2254 SE_TCL_INPUT_VTX_VECTOR_ADDR_0
770x2258 SE_TCL_INPUT_VTX_VECTOR_ADDR_1
780x225c SE_TCL_INPUT_VTX_VECTOR_ADDR_2
790x2260 SE_TCL_INPUT_VTX_VECTOR_ADDR_3
800x2268 SE_TCL_LIGHT_MODEL_CTL_0
810x226c SE_TCL_LIGHT_MODEL_CTL_1
820x2270 SE_TCL_PER_LIGHT_CTL_0
830x2274 SE_TCL_PER_LIGHT_CTL_1
840x2278 SE_TCL_PER_LIGHT_CTL_2
850x227c SE_TCL_PER_LIGHT_CTL_3
860x2284 VAP_PVS_STATE_FLUSH_REG
870x22a8 SE_TCL_TEX_PROC_CTL_2
880x22ac SE_TCL_TEX_PROC_CTL_3
890x22b0 SE_TCL_TEX_PROC_CTL_0
900x22b4 SE_TCL_TEX_PROC_CTL_1
910x22b8 SE_TCL_TEX_CYL_WRAP_CTL
920x22c0 SE_TCL_UCP_VERT_BLEND_CNTL
930x22c4 SE_TCL_POINT_SPRITE_CNTL
940x2648 RE_POINTSIZE
950x26c0 RE_TOP_LEFT
960x26c4 RE_MISC
970x26f0 RE_AUX_SCISSOR_CNTL
980x2c14 PP_BORDER_COLOR_0
990x2c34 PP_BORDER_COLOR_1
1000x2c54 PP_BORDER_COLOR_2
1010x2c74 PP_BORDER_COLOR_3
1020x2c94 PP_BORDER_COLOR_4
1030x2cb4 PP_BORDER_COLOR_5
1040x2cc4 PP_CNTL_X
1050x2cf8 PP_TRI_PERF
1060x2cfc PP_PERF_CNTL
1070x2d9c PP_TAM_DEBUG3
1080x2ee0 PP_TFACTOR_0
1090x2ee4 PP_TFACTOR_1
1100x2ee8 PP_TFACTOR_2
1110x2eec PP_TFACTOR_3
1120x2ef0 PP_TFACTOR_4
1130x2ef4 PP_TFACTOR_5
1140x2ef8 PP_TFACTOR_6
1150x2efc PP_TFACTOR_7
1160x2f00 PP_TXCBLEND_0
1170x2f04 PP_TXCBLEND2_0
1180x2f08 PP_TXABLEND_0
1190x2f0c PP_TXABLEND2_0
1200x2f10 PP_TXCBLEND_1
1210x2f14 PP_TXCBLEND2_1
1220x2f18 PP_TXABLEND_1
1230x2f1c PP_TXABLEND2_1
1240x2f20 PP_TXCBLEND_2
1250x2f24 PP_TXCBLEND2_2
1260x2f28 PP_TXABLEND_2
1270x2f2c PP_TXABLEND2_2
1280x2f30 PP_TXCBLEND_3
1290x2f34 PP_TXCBLEND2_3
1300x2f38 PP_TXABLEND_3
1310x2f3c PP_TXABLEND2_3
1320x2f40 PP_TXCBLEND_4
1330x2f44 PP_TXCBLEND2_4
1340x2f48 PP_TXABLEND_4
1350x2f4c PP_TXABLEND2_4
1360x2f50 PP_TXCBLEND_5
1370x2f54 PP_TXCBLEND2_5
1380x2f58 PP_TXABLEND_5
1390x2f5c PP_TXABLEND2_5
1400x2f60 PP_TXCBLEND_6
1410x2f64 PP_TXCBLEND2_6
1420x2f68 PP_TXABLEND_6
1430x2f6c PP_TXABLEND2_6
1440x2f70 PP_TXCBLEND_7
1450x2f74 PP_TXCBLEND2_7
1460x2f78 PP_TXABLEND_7
1470x2f7c PP_TXABLEND2_7
1480x2f80 PP_TXCBLEND_8
1490x2f84 PP_TXCBLEND2_8
1500x2f88 PP_TXABLEND_8
1510x2f8c PP_TXABLEND2_8
1520x2f90 PP_TXCBLEND_9
1530x2f94 PP_TXCBLEND2_9
1540x2f98 PP_TXABLEND_9
1550x2f9c PP_TXABLEND2_9
1560x2fa0 PP_TXCBLEND_10
1570x2fa4 PP_TXCBLEND2_10
1580x2fa8 PP_TXABLEND_10
1590x2fac PP_TXABLEND2_10
1600x2fb0 PP_TXCBLEND_11
1610x2fb4 PP_TXCBLEND2_11
1620x2fb8 PP_TXABLEND_11
1630x2fbc PP_TXABLEND2_11
1640x2fc0 PP_TXCBLEND_12
1650x2fc4 PP_TXCBLEND2_12
1660x2fc8 PP_TXABLEND_12
1670x2fcc PP_TXABLEND2_12
1680x2fd0 PP_TXCBLEND_13
1690x2fd4 PP_TXCBLEND2_13
1700x2fd8 PP_TXABLEND_13
1710x2fdc PP_TXABLEND2_13
1720x2fe0 PP_TXCBLEND_14
1730x2fe4 PP_TXCBLEND2_14
1740x2fe8 PP_TXABLEND_14
1750x2fec PP_TXABLEND2_14
1760x2ff0 PP_TXCBLEND_15
1770x2ff4 PP_TXCBLEND2_15
1780x2ff8 PP_TXABLEND_15
1790x2ffc PP_TXABLEND2_15
1800x3218 RB3D_BLENCOLOR
1810x321c RB3D_ABLENDCNTL
1820x3220 RB3D_CBLENDCNTL
1830x3290 RB3D_ZPASS_DATA
184
diff --git a/drivers/gpu/drm/radeon/reg_srcs/rn50 b/drivers/gpu/drm/radeon/reg_srcs/rn50
new file mode 100644
index 000000000000..2687b6307260
--- /dev/null
+++ b/drivers/gpu/drm/radeon/reg_srcs/rn50
@@ -0,0 +1,30 @@
1rn50 0x3294
20x1434 SRC_Y_X
30x1438 DST_Y_X
40x143C DST_HEIGHT_WIDTH
50x146C DP_GUI_MASTER_CNTL
60x1474 BRUSH_Y_X
70x1478 DP_BRUSH_BKGD_CLR
80x147C DP_BRUSH_FRGD_CLR
90x1480 BRUSH_DATA0
100x1484 BRUSH_DATA1
110x1598 DST_WIDTH_HEIGHT
120x15C0 CLR_CMP_CNTL
130x15C4 CLR_CMP_CLR_SRC
140x15C8 CLR_CMP_CLR_DST
150x15CC CLR_CMP_MSK
160x15D8 DP_SRC_FRGD_CLR
170x15DC DP_SRC_BKGD_CLR
180x1600 DST_LINE_START
190x1604 DST_LINE_END
200x1608 DST_LINE_PATCOUNT
210x16C0 DP_CNTL
220x16CC DP_WRITE_MSK
230x16D0 DP_CNTL_XDIR_YDIR_YMAJOR
240x16E8 DEFAULT_SC_BOTTOM_RIGHT
250x16EC SC_TOP_LEFT
260x16F0 SC_BOTTOM_RIGHT
270x16F4 SRC_SC_BOTTOM_RIGHT
280x1714 DSTCACHE_CTLSTAT
290x1720 WAIT_UNTIL
300x172C RBBM_GUICNTL
generated by cgit v1.2.2 (git 2.25.0) at 2024-12-25 15:51:41 -0500