Merge commit 'origin/gallium-0.2' into gallium-xlib-rework

20 files changed, 127 insertions, 91 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c
index a415e378ff..d7a2bd95ee 100644
--- a/src/mesa/drivers/dri/i965/brw_context.c
+++ b/src/mesa/drivers/dri/i965/brw_context.c
@@ -143,6 +143,9 @@ GLboolean brwCreateContext( const __GLcontextModes *mesaVis,
    ctx->Const.MaxCubeTextureLevels = 12;
    ctx->Const.MaxTextureRectSize = (1<<11);
    
+   /* if conformance mode is set, swrast can handle any size AA point */
+   ctx->Const.MaxPointSizeAA = 255.0;
+
 /*    ctx->Const.MaxNativeVertexProgramTemps = 32; */
 
    brw_init_attribs( brw );
diff --git a/src/mesa/drivers/dri/i965/brw_program.c b/src/mesa/drivers/dri/i965/brw_program.c
index a18dee85e8..0c86911044 100644
--- a/src/mesa/drivers/dri/i965/brw_program.c
+++ b/src/mesa/drivers/dri/i965/brw_program.c
@@ -111,9 +111,15 @@ static void brwProgramStringNotify( GLcontext *ctx,
 				    struct gl_program *prog )
 {
    if (target == GL_FRAGMENT_PROGRAM_ARB) {
+      struct gl_fragment_program *fprog = (struct gl_fragment_program *) prog;
       struct brw_context *brw = brw_context(ctx);
       struct brw_fragment_program *p = (struct brw_fragment_program *)prog;
       struct brw_fragment_program *fp = (struct brw_fragment_program *)brw->fragment_program;
+      if (fprog->FogOption) {
+         _mesa_append_fog_code(ctx, fprog);
+         fprog->FogOption = GL_NONE;
+      }
+
       if (p == fp)
 	 brw->state.dirty.brw |= BRW_NEW_FRAGMENT_PROGRAM;
       p->id = brw->program_id++;      
diff --git a/src/mesa/drivers/dri/i965/brw_sf.c b/src/mesa/drivers/dri/i965/brw_sf.c
index 9dce6cd8e6..1a11d54621 100644
--- a/src/mesa/drivers/dri/i965/brw_sf.c
+++ b/src/mesa/drivers/dri/i965/brw_sf.c
@@ -73,10 +73,12 @@ static void compile_sf_prog( struct brw_context *brw,
 	 c.attr_to_idx[i] = idx;
 	 c.idx_to_attr[idx] = i;
 	 if (i >= VERT_RESULT_TEX0 && i <= VERT_RESULT_TEX7) {
-		 c.point_attrs[i].CoordReplace = 
-			brw->attribs.Point->CoordReplace[i - VERT_RESULT_TEX0];
-	 } else
-		 c.point_attrs[i].CoordReplace = GL_FALSE;
+            c.point_attrs[i].CoordReplace = 
+               brw->attribs.Point->CoordReplace[i - VERT_RESULT_TEX0];
+	 }
+         else {
+            c.point_attrs[i].CoordReplace = GL_FALSE;
+         }
 	 idx++;
       }
    
@@ -106,7 +108,6 @@ static void compile_sf_prog( struct brw_context *brw,
       assert(0);
       return;
    }
-	 
 
    /* get the program
     */
diff --git a/src/mesa/drivers/dri/i965/brw_sf_state.c b/src/mesa/drivers/dri/i965/brw_sf_state.c
index 506126fcfb..741a7c53c4 100644
--- a/src/mesa/drivers/dri/i965/brw_sf_state.c
+++ b/src/mesa/drivers/dri/i965/brw_sf_state.c
@@ -229,7 +229,7 @@ sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
    /* XXX clamp max depends on AA vs. non-AA */
 
    sf.sf7.sprite_point = key->point_sprite;
-   sf.sf7.point_size = CLAMP(nearbyint(key->point_size), 1, 255) * (1<<3);
+   sf.sf7.point_size = CLAMP(rint(key->point_size), 1, 255) * (1<<3);
    sf.sf7.use_point_size_state = !key->point_attenuated;
    sf.sf7.aa_line_distance_mode = 0;
 
diff --git a/src/mesa/drivers/dri/i965/brw_wm_fp.c b/src/mesa/drivers/dri/i965/brw_wm_fp.c
index f4583877aa..1a00b69825 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_fp.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_fp.c
@@ -811,57 +811,6 @@ static void precalc_txp( struct brw_wm_compile *c,
 
 
 
-
-
-/***********************************************************************
- * Add instructions to perform fog blending
- */
-
-static void fog_blend( struct brw_wm_compile *c,
-			     struct prog_src_register fog_factor )
-{
-   struct prog_dst_register outcolor = dst_reg(PROGRAM_OUTPUT, FRAG_RESULT_COLR);
-   struct prog_src_register fogcolor = search_or_add_param5( c, STATE_FOG_COLOR, 0,0,0,0 );
-
-   /* color.xyz = LRP fog_factor.xxxx, output_color, fog_color */
-   
-   emit_op(c, 
-	   OPCODE_LRP,
-	   dst_mask(outcolor, WRITEMASK_XYZ),
-	   0, 0, 0,
-	   fog_factor,
-	   src_reg_from_dst(outcolor),
-	   fogcolor);
-}
-
-
-
-/* This one is simple - just take the interpolated fog coordinate and
- * use it as the fog blend factor.
- */
-static void fog_interpolated( struct brw_wm_compile *c )
-{
-   struct prog_src_register fogc = src_reg(PROGRAM_INPUT, FRAG_ATTRIB_FOGC);
-   
-   if (!(c->fp_interp_emitted & (1<<FRAG_ATTRIB_FOGC))) 
-      emit_interp(c, FRAG_ATTRIB_FOGC);
-
-   fog_blend( c, src_swizzle1(fogc, GET_SWZ(fogc.Swizzle,X)));
-}
-
-static void emit_fog( struct brw_wm_compile *c ) 
-{
-   if (!c->fp->program.FogOption)
-      return;
-
-   if (1) 
-      fog_interpolated( c );
-   else {
-      /* TODO: per-pixel fog */
-      assert(0);
-   }
-}
-
 static void emit_fb_write( struct brw_wm_compile *c )
 {
    struct prog_src_register payload_r0_depth = src_reg(PROGRAM_PAYLOAD, PAYLOAD_DEPTH);
@@ -1059,7 +1008,6 @@ void brw_wm_pass_fp( struct brw_wm_compile *c )
          emit_ddy(c, inst);
 	break;
       case OPCODE_END:
-	 emit_fog(c);
 	 emit_fb_write(c);
 	 break;
       case OPCODE_PRINT:
diff --git a/src/mesa/drivers/dri/i965/brw_wm_glsl.c b/src/mesa/drivers/dri/i965/brw_wm_glsl.c
index 942ebe1ed4..8fd776ac39 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_glsl.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_glsl.c
@@ -892,15 +892,19 @@ static void emit_lrp(struct brw_wm_compile *c,
     }
 }
 
+/**
+ * For GLSL shaders, this KIL will be unconditional.
+ * It may be contained inside an IF/ENDIF structure of course.
+ */
 static void emit_kil(struct brw_wm_compile *c)
 {
-	struct brw_compile *p = &c->func;
-	struct brw_reg depth = retype(brw_vec1_grf(0, 0), BRW_REGISTER_TYPE_UW);
-	brw_push_insn_state(p);
-	brw_set_mask_control(p, BRW_MASK_DISABLE);
-	brw_NOT(p, c->emit_mask_reg, brw_mask_reg(1)); //IMASK
-	brw_AND(p, depth, c->emit_mask_reg, depth);
-	brw_pop_insn_state(p);
+    struct brw_compile *p = &c->func;
+    struct brw_reg depth = retype(brw_vec1_grf(0, 0), BRW_REGISTER_TYPE_UW);
+    brw_push_insn_state(p);
+    brw_set_mask_control(p, BRW_MASK_DISABLE);
+    brw_NOT(p, c->emit_mask_reg, brw_mask_reg(1)); //IMASK
+    brw_AND(p, depth, c->emit_mask_reg, depth);
+    brw_pop_insn_state(p);
 }
 
 static void emit_mad(struct brw_wm_compile *c,
diff --git a/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c b/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
index f12ef47a7d..8c9cb78945 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
@@ -244,7 +244,7 @@ brw_wm_sampler_populate_key(struct brw_context *brw,
 	 entry->minfilter = texObj->MinFilter;
 	 entry->magfilter = texObj->MagFilter;
 	 entry->comparemode = texObj->CompareMode;
-    entry->comparefunc = texObj->CompareFunc;
+         entry->comparefunc = texObj->CompareFunc;
 
 	 dri_bo_unreference(brw->wm.sdc_bo[unit]);
 	 if (firstImage->_BaseFormat == GL_DEPTH_COMPONENT) {
diff --git a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
index 63e14cc390..06e71e6d69 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
@@ -192,21 +192,27 @@ brw_create_texture_surface( struct brw_context *brw,
    if (key->bo) 
       surf.ss0.surface_format = translate_tex_format(key->format, key->depthmode);
    else {
-     switch(key->depth) {
-     case 32: surf.ss0.surface_format = BRW_SURFACEFORMAT_B8G8R8A8_UNORM; break;
-     default:
-     case 24: surf.ss0.surface_format = BRW_SURFACEFORMAT_B8G8R8X8_UNORM; break;
-     case 16: surf.ss0.surface_format = BRW_SURFACEFORMAT_B5G6R5_UNORM; break;
-     }
+      switch (key->depth) {
+      case 32:
+         surf.ss0.surface_format = BRW_SURFACEFORMAT_B8G8R8A8_UNORM;
+         break;
+      default:
+      case 24:
+         surf.ss0.surface_format = BRW_SURFACEFORMAT_B8G8R8X8_UNORM;
+         break;
+      case 16:
+         surf.ss0.surface_format = BRW_SURFACEFORMAT_B5G6R5_UNORM;
+         break;
+      }
    }
 
    /* This is ok for all textures with channel width 8bit or less:
     */
 /*    surf.ss0.data_return_format = BRW_SURFACERETURNFORMAT_S1; */
    if (key->bo)
-     surf.ss1.base_addr = key->bo->offset; /* reloc */
+      surf.ss1.base_addr = key->bo->offset; /* reloc */
    else
-     surf.ss1.base_addr = key->offset;
+      surf.ss1.base_addr = key->offset;
 
    surf.ss2.mip_count = key->last_level - key->first_level;
    surf.ss2.width = key->width - 1;
diff --git a/src/mesa/drivers/dri/intel/intel_context.c b/src/mesa/drivers/dri/intel/intel_context.c
index 44b276a123..7b7f7d8c14 100644
--- a/src/mesa/drivers/dri/intel/intel_context.c
+++ b/src/mesa/drivers/dri/intel/intel_context.c
@@ -95,7 +95,7 @@ int INTEL_DEBUG = (0);
 
 #include "extension_helper.h"
 
-#define DRIVER_DATE                     "20080716"
+#define DRIVER_DATE                     "20090114"
 #define DRIVER_DATE_GEM                 "GEM " DRIVER_DATE
 
 static const GLubyte *
diff --git a/src/mesa/drivers/dri/intel/intel_depthstencil.c b/src/mesa/drivers/dri/intel/intel_depthstencil.c
index c2b4d7728b..354b3bf0d7 100644
--- a/src/mesa/drivers/dri/intel/intel_depthstencil.c
+++ b/src/mesa/drivers/dri/intel/intel_depthstencil.c
@@ -110,7 +110,10 @@ intel_unpair_depth_stencil(GLcontext *ctx, struct intel_renderbuffer *irb)
 	 ASSERT(stencilIrb->PairedDepth == rb->Name);
 	 intel_renderbuffer_map(intel, rb);
 	 intel_renderbuffer_map(intel, stencilRb);
+#if 0
+         /* disable for now */
 	 _mesa_extract_stencil(ctx, rb, stencilRb);
+#endif
 	 intel_renderbuffer_unmap(intel, stencilRb);
 	 intel_renderbuffer_unmap(intel, rb);
          stencilIrb->PairedDepth = 0;
@@ -132,7 +135,10 @@ intel_unpair_depth_stencil(GLcontext *ctx, struct intel_renderbuffer *irb)
 	 ASSERT(depthIrb->PairedStencil == rb->Name);
 	 intel_renderbuffer_map(intel, rb);
 	 intel_renderbuffer_map(intel, depthRb);
+#if 0
+         /* disable for now */
 	 _mesa_extract_stencil(ctx, depthRb, rb);
+#endif
 	 intel_renderbuffer_unmap(intel, depthRb);
 	 intel_renderbuffer_unmap(intel, rb);
          depthIrb->PairedStencil = 0;
@@ -177,8 +183,11 @@ intel_validate_paired_depth_stencil(GLcontext * ctx,
       }
       else {
          /* Separate depth/stencil buffers, need to interleave now */
-         ASSERT(depthRb->Base._BaseFormat == GL_DEPTH_COMPONENT);
-         ASSERT(stencilRb->Base._BaseFormat == GL_STENCIL_INDEX);
+         ASSERT(depthRb->Base._BaseFormat == GL_DEPTH_COMPONENT ||
+                depthRb->Base._BaseFormat == GL_DEPTH_STENCIL);
+         ASSERT(stencilRb->Base._BaseFormat == GL_STENCIL_INDEX ||
+                stencilRb->Base._BaseFormat == GL_DEPTH_STENCIL);
+
          /* may need to interleave depth/stencil now */
          if (depthRb->PairedStencil == stencilRb->Base.Name) {
             /* OK, the depth and stencil buffers are already interleaved */
diff --git a/src/mesa/drivers/dri/intel/intel_fbo.c b/src/mesa/drivers/dri/intel/intel_fbo.c
index 7453b96dc5..54f2fa5287 100644
--- a/src/mesa/drivers/dri/intel/intel_fbo.c
+++ b/src/mesa/drivers/dri/intel/intel_fbo.c
@@ -248,11 +248,18 @@ intel_alloc_renderbuffer_storage(GLcontext * ctx, struct gl_renderbuffer *rb,
       cpp = 4;
       break;
    case GL_DEPTH_COMPONENT16:
+#if 0
       rb->_ActualFormat = GL_DEPTH_COMPONENT16;
       rb->DataType = GL_UNSIGNED_SHORT;
       rb->DepthBits = 16;
       cpp = 2;
       break;
+#else
+      /* fall-through.
+       * 16bpp depth renderbuffer can't be paired with a stencil buffer so
+       * always used combined depth/stencil format.
+       */
+#endif
    case GL_DEPTH_COMPONENT:
    case GL_DEPTH_COMPONENT24:
    case GL_DEPTH_COMPONENT32:
@@ -529,20 +536,25 @@ intel_update_wrapper(GLcontext *ctx, struct intel_renderbuffer *irb,
    if (texImage->TexFormat == &_mesa_texformat_argb8888) {
       irb->Base._ActualFormat = GL_RGBA8;
       irb->Base._BaseFormat = GL_RGBA;
+      irb->Base.DataType = GL_UNSIGNED_BYTE;
       DBG("Render to RGBA8 texture OK\n");
    }
    else if (texImage->TexFormat == &_mesa_texformat_rgb565) {
       irb->Base._ActualFormat = GL_RGB5;
       irb->Base._BaseFormat = GL_RGB;
+      irb->Base.DataType = GL_UNSIGNED_SHORT;
       DBG("Render to RGB5 texture OK\n");
    }
    else if (texImage->TexFormat == &_mesa_texformat_z16) {
       irb->Base._ActualFormat = GL_DEPTH_COMPONENT16;
       irb->Base._BaseFormat = GL_DEPTH_COMPONENT;
+      irb->Base.DataType = GL_UNSIGNED_SHORT;
       DBG("Render to DEPTH16 texture OK\n");
-   } else if (texImage->TexFormat == &_mesa_texformat_s8_z24) {
+   }
+   else if (texImage->TexFormat == &_mesa_texformat_s8_z24) {
       irb->Base._ActualFormat = GL_DEPTH24_STENCIL8_EXT;
       irb->Base._BaseFormat = GL_DEPTH_STENCIL_EXT;
+      irb->Base.DataType = GL_UNSIGNED_INT_24_8_EXT;
       DBG("Render to DEPTH_STENCIL texture OK\n");
    }
    else {
@@ -554,7 +566,6 @@ intel_update_wrapper(GLcontext *ctx, struct intel_renderbuffer *irb,
    irb->Base.InternalFormat = irb->Base._ActualFormat;
    irb->Base.Width = texImage->Width;
    irb->Base.Height = texImage->Height;
-   irb->Base.DataType = GL_UNSIGNED_BYTE;       /* FBO XXX fix */
    irb->Base.RedBits = texImage->TexFormat->RedBits;
    irb->Base.GreenBits = texImage->TexFormat->GreenBits;
    irb->Base.BlueBits = texImage->TexFormat->BlueBits;
diff --git a/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c b/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
index 1d7f15f10a..3a01f63dc7 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
@@ -204,6 +204,14 @@ do_blit_bitmap( GLcontext *ctx,
    /* Update draw buffer bounds */
    _mesa_update_state(ctx);
 
+   if (ctx->Depth.Test) {
+      /* The blit path produces incorrect results when depth testing is on.
+       * It seems the blit Z coord is always 1.0 (the far plane) so fragments
+       * will likely be obscured by other, closer geometry.
+       */
+      return GL_FALSE;
+   }
+
    if (!dst)
        return GL_FALSE;
 
@@ -357,6 +365,7 @@ intel_texture_bitmap(GLcontext * ctx,
    GLubyte *unpacked_bitmap;
    GLubyte *a8_bitmap;
    int x, y;
+   GLfloat dst_z;
 
    /* We need a fragment program for the KIL effect */
    if (!ctx->Extensions.ARB_fragment_program ||
@@ -456,21 +465,24 @@ intel_texture_bitmap(GLcontext * ctx,
    intel_meta_set_passthrough_vertex_program(intel);
    intel_meta_set_passthrough_transform(intel);
 
+   /* convert rasterpos Z from [0,1] to NDC coord in [-1,1] */
+   dst_z = -1.0 + 2.0 * ctx->Current.RasterPos[2];
+
    vertices[0][0] = dst_x;
    vertices[0][1] = dst_y;
-   vertices[0][2] = ctx->Current.RasterPos[2];
+   vertices[0][2] = dst_z;
    vertices[0][3] = 1.0;
    vertices[1][0] = dst_x + width;
    vertices[1][1] = dst_y;
-   vertices[1][2] = ctx->Current.RasterPos[2];
+   vertices[1][2] = dst_z;
    vertices[1][3] = 1.0;
    vertices[2][0] = dst_x + width;
    vertices[2][1] = dst_y + height;
-   vertices[2][2] = ctx->Current.RasterPos[2];
+   vertices[2][2] = dst_z;
    vertices[2][3] = 1.0;
    vertices[3][0] = dst_x;
    vertices[3][1] = dst_y + height;
-   vertices[3][2] = ctx->Current.RasterPos[2];
+   vertices[3][2] = dst_z;
    vertices[3][3] = 1.0;
 
    texcoords[0][0] = 0.0;
diff --git a/src/mesa/drivers/dri/intel/intel_pixel_copy.c b/src/mesa/drivers/dri/intel/intel_pixel_copy.c
index 447c6494e7..7c7aa6097c 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel_copy.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel_copy.c
@@ -119,6 +119,12 @@ do_texture_copypixels(GLcontext * ctx,
    if (!src || !dst || type != GL_COLOR)
       return GL_FALSE;
 
+   if (ctx->_ImageTransferState) {
+      if (INTEL_DEBUG & DEBUG_PIXEL)
+         fprintf(stderr, "%s: check_color failed\n", __FUNCTION__);
+      return GL_FALSE;
+   }
+
    /* Can't handle overlapping regions.  Don't have sufficient control
     * over rasterization to pull it off in-place.  Punt on these for
     * now.
diff --git a/src/mesa/drivers/dri/intel/intel_pixel_draw.c b/src/mesa/drivers/dri/intel/intel_pixel_draw.c
index 2af839b803..0e83afa645 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel_draw.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel_draw.c
@@ -71,6 +71,7 @@ intel_texture_drawpixels(GLcontext * ctx,
    GLuint texname;
    GLfloat vertices[4][4];
    GLfloat texcoords[4][2];
+   GLfloat z;
 
    /* We're going to mess with texturing with no regard to existing texture
     * state, so if there is some set up we have to bail.
@@ -140,6 +141,9 @@ intel_texture_drawpixels(GLcontext * ctx,
 
    intel_meta_set_passthrough_transform(intel);
 
+   /* convert rasterpos Z from [0,1] to NDC coord in [-1,1] */
+   z = -1.0 + 2.0 * ctx->Current.RasterPos[2];
+
    /* Create the vertex buffer based on the current raster pos.  The x and y
     * we're handed are ctx->Current.RasterPos[0,1] rounded to integers.
     * We also apply the depth.  However, the W component is already multiplied
@@ -147,19 +151,19 @@ intel_texture_drawpixels(GLcontext * ctx,
     */
    vertices[0][0] = x;
    vertices[0][1] = y;
-   vertices[0][2] = ctx->Current.RasterPos[2];
+   vertices[0][2] = z;
    vertices[0][3] = 1.0;
    vertices[1][0] = x + width * ctx->Pixel.ZoomX;
    vertices[1][1] = y;
-   vertices[1][2] = ctx->Current.RasterPos[2];
+   vertices[1][2] = z;
    vertices[1][3] = 1.0;
    vertices[2][0] = x + width * ctx->Pixel.ZoomX;
    vertices[2][1] = y + height * ctx->Pixel.ZoomY;
-   vertices[2][2] = ctx->Current.RasterPos[2];
+   vertices[2][2] = z;
    vertices[2][3] = 1.0;
    vertices[3][0] = x;
    vertices[3][1] = y + height * ctx->Pixel.ZoomY;
-   vertices[3][2] = ctx->Current.RasterPos[2];
+   vertices[3][2] = z;
    vertices[3][3] = 1.0;
 
    texcoords[0][0] = 0.0;
diff --git a/src/mesa/drivers/dri/intel/intel_span.c b/src/mesa/drivers/dri/intel/intel_span.c
index 8f4e681ffe..d9315043e6 100644
--- a/src/mesa/drivers/dri/intel/intel_span.c
+++ b/src/mesa/drivers/dri/intel/intel_span.c
@@ -633,7 +633,7 @@ intelSpanRenderStart(GLcontext * ctx)
    intelFlush(&intel->ctx);
    LOCK_HARDWARE(intel);
 
-   for (i = 0; i < ctx->Const.MaxTextureCoordUnits; i++) {
+   for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
       if (ctx->Texture.Unit[i]._ReallyEnabled) {
          struct gl_texture_object *texObj = ctx->Texture.Unit[i]._Current;
          intel_tex_map_images(intel, intel_texture_object(texObj));
@@ -655,7 +655,7 @@ intelSpanRenderFinish(GLcontext * ctx)
 
    _swrast_flush(ctx);
 
-   for (i = 0; i < ctx->Const.MaxTextureCoordUnits; i++) {
+   for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
       if (ctx->Texture.Unit[i]._ReallyEnabled) {
          struct gl_texture_object *texObj = ctx->Texture.Unit[i]._Current;
          intel_tex_unmap_images(intel, intel_texture_object(texObj));
diff --git a/src/mesa/drivers/dri/intel/intel_tex_format.c b/src/mesa/drivers/dri/intel/intel_tex_format.c
index 2be060dd3e..5e418ac446 100644
--- a/src/mesa/drivers/dri/intel/intel_tex_format.c
+++ b/src/mesa/drivers/dri/intel/intel_tex_format.c
@@ -134,8 +134,14 @@ intelChooseTextureFormat(GLcontext * ctx, GLint internalFormat,
    case GL_DEPTH_COMPONENT16:
    case GL_DEPTH_COMPONENT24:
    case GL_DEPTH_COMPONENT32:
+#if 0
       return &_mesa_texformat_z16;
-
+#else
+      /* fall-through.
+       * 16bpp depth texture can't be paired with a stencil buffer so
+       * always used combined depth/stencil format.
+       */
+#endif
    case GL_DEPTH_STENCIL_EXT:
    case GL_DEPTH24_STENCIL8_EXT:
       return &_mesa_texformat_s8_z24;
@@ -158,7 +164,7 @@ intelChooseTextureFormat(GLcontext * ctx, GLint internalFormat,
    case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT:
    case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT:
    case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT:
-     return &_mesa_texformat_srgb_dxt1;
+      return &_mesa_texformat_srgb_dxt1;
 #endif
 
    default:
diff --git a/src/mesa/drivers/dri/mach64/mach64_context.h b/src/mesa/drivers/dri/mach64/mach64_context.h
index 55e0618ff8..854751626d 100644
--- a/src/mesa/drivers/dri/mach64/mach64_context.h
+++ b/src/mesa/drivers/dri/mach64/mach64_context.h
@@ -294,7 +294,13 @@ extern GLboolean mach64UnbindContext( __DRIcontextPrivate *driContextPriv );
 #define LE32_OUT( x, y )	do { *(GLuint *)(x) = (y); } while (0)
 #define LE32_OUT_FLOAT( x, y )	do { *(GLfloat *)(x) = (y); } while (0)
 #else
+#ifndef __OpenBSD__
 #include <byteswap.h>
+#else
+#include <machine/endian.h>
+#define bswap_32 bswap32
+#endif
+
 #define LE32_IN( x )		bswap_32( *(GLuint *)(x) )
 #define LE32_IN_FLOAT( x )						\
 ({									\
diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c
index 6a5c3633a2..a63dbac343 100644
--- a/src/mesa/drivers/dri/r300/r300_state.c
+++ b/src/mesa/drivers/dri/r300/r300_state.c
@@ -1675,6 +1675,13 @@ static void r300SetupRSUnit(GLcontext * ctx)
 		rs_col_count += count;
 	}
 
+	if (InputsRead & FRAG_BIT_FOGC) {
+		/* XXX FIX THIS
+		 * Just turn off the bit for now.
+		 * Need to do something similar to the color/texcoord inputs.
+		 */
+		InputsRead &= ~FRAG_BIT_FOGC;
+	}
 
 	for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
 		int swiz;
diff --git a/src/mesa/drivers/dri/r300/radeon_program_pair.c b/src/mesa/drivers/dri/r300/radeon_program_pair.c
index 5ad50d2863..58bc0d5843 100644
--- a/src/mesa/drivers/dri/r300/radeon_program_pair.c
+++ b/src/mesa/drivers/dri/r300/radeon_program_pair.c
@@ -473,6 +473,11 @@ static void allocate_input_registers(struct pair_state *s)
 		alloc_hw_reg(s, PROGRAM_INPUT, FRAG_ATTRIB_COL1, hwindex++);
 	InputsRead &= ~FRAG_BIT_COL1;
 
+	/* Fog coordinate */
+	if (InputsRead & FRAG_BIT_FOGC)
+		alloc_hw_reg(s, PROGRAM_INPUT, FRAG_ATTRIB_FOGC, hwindex++);
+	InputsRead &= ~FRAG_BIT_FOGC;
+
 	/* Anything else */
 	if (InputsRead)
 		error("Don't know how to handle inputs 0x%x\n", InputsRead);
diff --git a/src/mesa/drivers/dri/swrast/swrast.c b/src/mesa/drivers/dri/swrast/swrast.c
index 49f1b8b9ee..15b57244dc 100644
--- a/src/mesa/drivers/dri/swrast/swrast.c
+++ b/src/mesa/drivers/dri/swrast/swrast.c
@@ -66,6 +66,7 @@
 #define need_GL_ARB_vertex_program
 #define need_GL_APPLE_vertex_array_object
 #define need_GL_ATI_fragment_shader
+#define need_GL_ATI_separate_stencil
 #define need_GL_EXT_depth_bounds_test
 #define need_GL_EXT_framebuffer_object
 #define need_GL_EXT_framebuffer_blit
@@ -96,6 +97,7 @@ const struct dri_extension card_extensions[] =
     { "GL_ARB_vertex_program",		GL_ARB_vertex_program_functions },
     { "GL_APPLE_vertex_array_object",	GL_APPLE_vertex_array_object_functions },
     { "GL_ATI_fragment_shader",		GL_ATI_fragment_shader_functions },
+    { "GL_ATI_separate_stencil",	GL_ATI_separate_stencil_functions },
     { "GL_EXT_depth_bounds_test",	GL_EXT_depth_bounds_test_functions },
     { "GL_EXT_framebuffer_object",	GL_EXT_framebuffer_object_functions },
     { "GL_EXT_framebuffer_blit",	GL_EXT_framebuffer_blit_functions },