1 files changed, 48 insertions, 33 deletions

diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index cf89aa2eb28c..3970e62eaab8 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -1628,6 +1628,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
                 case RADEON_TXFORMAT_RGB332:
                 case RADEON_TXFORMAT_Y8:
                         track->textures[i].cpp = 1;
+                        track->textures[i].compress_format = R100_TRACK_COMP_NONE;
                         break;
                 case RADEON_TXFORMAT_AI88:
                 case RADEON_TXFORMAT_ARGB1555:
@@ -1639,12 +1640,14 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
                 case RADEON_TXFORMAT_LDUDV655:
                 case RADEON_TXFORMAT_DUDV88:
                         track->textures[i].cpp = 2;
+                        track->textures[i].compress_format = R100_TRACK_COMP_NONE;
                         break;
                 case RADEON_TXFORMAT_ARGB8888:
                 case RADEON_TXFORMAT_RGBA8888:
                 case RADEON_TXFORMAT_SHADOW32:
                 case RADEON_TXFORMAT_LDUDUV8888:
                         track->textures[i].cpp = 4;
+                        track->textures[i].compress_format = R100_TRACK_COMP_NONE;
                         break;
                 case RADEON_TXFORMAT_DXT1:
                         track->textures[i].cpp = 1;
@@ -2604,12 +2607,6 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
         int surf_index = reg * 16;
         int flags = 0;
 
-        /* r100/r200 divide by 16 */
-        if (rdev->family < CHIP_R300)
-                flags = pitch / 16;
-        else
-                flags = pitch / 8;
-
         if (rdev->family <= CHIP_RS200) {
                 if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO))
                                  == (RADEON_TILING_MACRO|RADEON_TILING_MICRO))
@@ -2633,6 +2630,20 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
         if (tiling_flags & RADEON_TILING_SWAP_32BIT)
                 flags |= RADEON_SURF_AP0_SWP_32BPP | RADEON_SURF_AP1_SWP_32BPP;
 
+        /* when we aren't tiling the pitch seems to needs to be furtherdivided down. - tested on power5 + rn50 server */
+        if (tiling_flags & (RADEON_TILING_SWAP_16BIT | RADEON_TILING_SWAP_32BIT)) {
+                if (!(tiling_flags & (RADEON_TILING_MACRO | RADEON_TILING_MICRO)))
+                        if (ASIC_IS_RN50(rdev))
+                                pitch /= 16;
+        }
+
+        /* r100/r200 divide by 16 */
+        if (rdev->family < CHIP_R300)
+                flags |= pitch / 16;
+        else
+                flags |= pitch / 8;
+
+
         DRM_DEBUG("writing surface %d %d %x %x\n", reg, flags, offset, offset+obj_size-1);
         WREG32(RADEON_SURFACE0_INFO + surf_index, flags);
         WREG32(RADEON_SURFACE0_LOWER_BOUND + surf_index, offset);
@@ -3147,33 +3158,6 @@ static inline void r100_cs_track_texture_print(struct r100_cs_track_texture *t)
         DRM_ERROR("compress format            %d\n", t->compress_format);
 }
 
-static int r100_cs_track_cube(struct radeon_device *rdev,
-                              struct r100_cs_track *track, unsigned idx)
-{
-        unsigned face, w, h;
-        struct radeon_bo *cube_robj;
-        unsigned long size;
-
-        for (face = 0; face < 5; face++) {
-                cube_robj = track->textures[idx].cube_info[face].robj;
-                w = track->textures[idx].cube_info[face].width;
-                h = track->textures[idx].cube_info[face].height;
-
-                size = w * h;
-                size *= track->textures[idx].cpp;
-
-                size += track->textures[idx].cube_info[face].offset;
-
-                if (size > radeon_bo_size(cube_robj)) {
-                        DRM_ERROR("Cube texture offset greater than object size %lu %lu\n",
-                                  size, radeon_bo_size(cube_robj));
-                        r100_cs_track_texture_print(&track->textures[idx]);
-                        return -1;
-                }
-        }
-        return 0;
-}
-
 static int r100_track_compress_size(int compress_format, int w, int h)
 {
         int block_width, block_height, block_bytes;
@@ -3204,6 +3188,37 @@ static int r100_track_compress_size(int compress_format, int w, int h)
         return sz;
 }
 
+static int r100_cs_track_cube(struct radeon_device *rdev,
+                              struct r100_cs_track *track, unsigned idx)
+{
+        unsigned face, w, h;
+        struct radeon_bo *cube_robj;
+        unsigned long size;
+        unsigned compress_format = track->textures[idx].compress_format;
+
+        for (face = 0; face < 5; face++) {
+                cube_robj = track->textures[idx].cube_info[face].robj;
+                w = track->textures[idx].cube_info[face].width;
+                h = track->textures[idx].cube_info[face].height;
+
+                if (compress_format) {
+                        size = r100_track_compress_size(compress_format, w, h);
+                } else
+                        size = w * h;
+                size *= track->textures[idx].cpp;
+
+                size += track->textures[idx].cube_info[face].offset;
+
+                if (size > radeon_bo_size(cube_robj)) {
+                        DRM_ERROR("Cube texture offset greater than object size %lu %lu\n",
+                                  size, radeon_bo_size(cube_robj));
+                        r100_cs_track_texture_print(&track->textures[idx]);
+                        return -1;
+                }
+        }
+        return 0;
+}
+
 static int r100_cs_track_texture_check(struct radeon_device *rdev,
                                        struct r100_cs_track *track)
 {