19 files changed, 346 insertions, 315 deletions

diff --git a/src/mesa/drivers/dri/r600/Makefile b/src/mesa/drivers/dri/r600/Makefile
index 3c3100ab91..d925a2dfe3 100644
--- a/src/mesa/drivers/dri/r600/Makefile
+++ b/src/mesa/drivers/dri/r600/Makefile
@@ -27,16 +27,17 @@ COMMON_SOURCES = \
 	../common/dri_util.c
 
 RADEON_COMMON_SOURCES = \
-	radeon_texture.c \
+	radeon_bo_legacy.c \
 	radeon_common_context.c \
 	radeon_common.c \
+	radeon_cs_legacy.c \
 	radeon_dma.c \
+	radeon_debug.c \
+	radeon_fbo.c \
 	radeon_lock.c \
-	radeon_bo_legacy.c \
-	radeon_cs_legacy.c \
 	radeon_mipmap_tree.c \
 	radeon_span.c \
-	radeon_fbo.c \
+	radeon_texture.c \
 	radeon_queryobj.c
 
 DRIVER_SOURCES = \
diff --git a/src/mesa/drivers/dri/r600/r600_cmdbuf.c b/src/mesa/drivers/dri/r600/r600_cmdbuf.c
index 65930ff589..3cfe03a45f 100644
--- a/src/mesa/drivers/dri/r600/r600_cmdbuf.c
+++ b/src/mesa/drivers/dri/r600/r600_cmdbuf.c
@@ -192,9 +192,9 @@ static int r600_cs_begin(struct radeon_cs *cs,
 
     if (cs->cdw + ndw > cs->ndw) {
         uint32_t tmp, *ptr;
-	int num = (ndw > 0x3FF) ? ndw : 0x3FF;
+	int num = (ndw > 0x400) ? ndw : 0x400;
 
-        tmp = (cs->cdw + 1 + num) & (~num);
+        tmp = (cs->cdw + num + 0x3FF) & (~0x3FF);
         ptr = (uint32_t*)realloc(cs->packets, 4 * tmp);
         if (ptr == NULL) {
             return -ENOMEM;
@@ -229,6 +229,14 @@ static int r600_cs_end(struct radeon_cs *cs,
         return -EPIPE;
     }
 
+    if (cs->cdw > cs->ndw) {
+	    fprintf(stderr, "CS section overflow at (%s,%s,%d) cdw %d ndw %d\n",
+		    cs->section_file, cs->section_func, cs->section_line,cs->cdw,cs->ndw);
+	    fprintf(stderr, "CS section end at (%s,%s,%d)\n",
+		    file, func, line);
+	    assert(0);
+    }
+
     return 0;
 }
 
diff --git a/src/mesa/drivers/dri/r600/r600_context.c b/src/mesa/drivers/dri/r600/r600_context.c
index 6a066f3510..07a7bcf11f 100644
--- a/src/mesa/drivers/dri/r600/r600_context.c
+++ b/src/mesa/drivers/dri/r600/r600_context.c
@@ -56,6 +56,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "drivers/common/driverfuncs.h"
 
+#include "radeon_debug.h"
 #include "r600_context.h"
 #include "radeon_common_context.h"
 #include "radeon_span.h"
@@ -84,6 +85,7 @@ int hw_tcl_on = 1;
 #define need_GL_EXT_framebuffer_object
 #define need_GL_EXT_fog_coord
 #define need_GL_EXT_gpu_program_parameters
+#define need_GL_EXT_provoking_vertex
 #define need_GL_EXT_secondary_color
 #define need_GL_EXT_stencil_two_side
 #define need_GL_ATI_separate_stencil
@@ -116,6 +118,7 @@ const struct dri_extension card_extensions[] = {
   {"GL_EXT_packed_depth_stencil",	NULL},
   {"GL_EXT_fog_coord",			GL_EXT_fog_coord_functions },
   {"GL_EXT_gpu_program_parameters",     GL_EXT_gpu_program_parameters_functions},
+  {"GL_EXT_provoking_vertex",           GL_EXT_provoking_vertex_functions },
   {"GL_EXT_secondary_color", 		GL_EXT_secondary_color_functions},
   {"GL_EXT_shadow_funcs",		NULL},
   {"GL_EXT_stencil_two_side",		GL_EXT_stencil_two_side_functions},
@@ -127,6 +130,7 @@ const struct dri_extension card_extensions[] = {
   {"GL_EXT_texture_lod_bias",		NULL},
   {"GL_EXT_texture_mirror_clamp",	NULL},
   {"GL_EXT_texture_rectangle",		NULL},
+  {"GL_EXT_texture_sRGB",               NULL},
   {"GL_ATI_separate_stencil",		GL_ATI_separate_stencil_functions},
   {"GL_ATI_texture_env_combine3",	NULL},
   {"GL_ATI_texture_mirror_once",	NULL},
@@ -225,8 +229,10 @@ GLboolean r600CreateContext(const __GLcontextModes * glVisual,
 
 	/* Allocate the R600 context */
 	r600 = (context_t*) CALLOC(sizeof(*r600));
-	if (!r600)
+	if (!r600) {
+		radeon_error("Failed to allocate memory for context.\n");
 		return GL_FALSE;
+	}
 
 	if (!(screen->chip_flags & RADEON_CHIPSET_TCL))
 		hw_tcl_on = future_hw_tcl_on = 0;
@@ -255,6 +261,7 @@ GLboolean r600CreateContext(const __GLcontextModes * glVisual,
 	if (!radeonInitContext(&r600->radeon, &functions,
 			       glVisual, driContextPriv,
 			       sharedContextPrivate)) {
+		radeon_error("Initializing context failed.\n");
 		FREE(r600);
 		return GL_FALSE;
 	}
@@ -347,6 +354,8 @@ GLboolean r600CreateContext(const __GLcontextModes * glVisual,
 	ctx->VertexProgram._MaintainTnlProgram = GL_TRUE;
 	ctx->FragmentProgram._MaintainTexEnvProgram = GL_TRUE;
 
+	radeon_init_debug();
+
 	driInitExtensions(ctx, card_extensions, GL_TRUE);
 	if (r600->radeon.radeonScreen->kernel_mm)
 	  driInitExtensions(ctx, mm_extensions, GL_FALSE);
@@ -375,7 +384,7 @@ GLboolean r600CreateContext(const __GLcontextModes * glVisual,
 	TNL_CONTEXT(ctx)->Driver.RunPipeline = r600RunPipeline;
 
 	if (driQueryOptionb(&r600->radeon.optionCache, "no_rast")) {
-		fprintf(stderr, "disabling 3D acceleration\n");
+		radeon_warning("disabling 3D acceleration\n");
 #if R200_MERGED
 		FALLBACK(&r600->radeon, RADEON_FALLBACK_DISABLE, 1);
 #endif
diff --git a/src/mesa/drivers/dri/r600/r600_emit.c b/src/mesa/drivers/dri/r600/r600_emit.c
index f8c56e7774..5c250c2418 100644
--- a/src/mesa/drivers/dri/r600/r600_emit.c
+++ b/src/mesa/drivers/dri/r600/r600_emit.c
@@ -60,28 +60,30 @@ GLboolean r600EmitShader(GLcontext * ctx,
 	radeonContextPtr radeonctx = RADEON_CONTEXT(ctx);
 	struct radeon_bo * pbo;
 	uint32_t *out;
-
 shader_again_alloc:
 	pbo = radeon_bo_open(radeonctx->radeonScreen->bom,
-			     0,
-			     sizeinDWORD * 4,
-			     256,
-			     RADEON_GEM_DOMAIN_GTT,
-			     0);
+			0,
+			sizeinDWORD * 4,
+			256,
+			RADEON_GEM_DOMAIN_GTT,
+			0);
+
+	radeon_print(RADEON_SHADER, RADEON_NORMAL, "%s %p size %d: %s\n", __func__, pbo, sizeinDWORD, szShaderUsage);
 
 	if (!pbo) {
+		radeon_print(RADEON_MEMORY | RADEON_CS, RADEON_IMPORTANT, "No memory for buffer object. Flushing command buffer.\n");
 		rcommonFlushCmdBuf(radeonctx, __FUNCTION__);
 		goto shader_again_alloc;
 	}
 
 	radeon_cs_space_add_persistent_bo(radeonctx->cmdbuf.cs,
-					  pbo,
-					  RADEON_GEM_DOMAIN_GTT, 0);
+			pbo,
+			RADEON_GEM_DOMAIN_GTT, 0);
 
-        if (radeon_cs_space_check_with_bo(radeonctx->cmdbuf.cs,
-                                          pbo,
-                                          RADEON_GEM_DOMAIN_GTT, 0)) {
-                fprintf(stderr,"failure to revalidate BOs - badness\n");
+	if (radeon_cs_space_check_with_bo(radeonctx->cmdbuf.cs,
+				pbo,
+				RADEON_GEM_DOMAIN_GTT, 0)) {
+		radeon_error("failure to revalidate BOs - badness\n");
 		return GL_FALSE;
 	}
 
@@ -103,6 +105,8 @@ GLboolean r600DeleteShader(GLcontext * ctx,
 {
     struct radeon_bo * pbo = (struct radeon_bo *)shaderbo;
 
+    radeon_print(RADEON_SHADER, RADEON_NORMAL, "%s: %p\n", __func__, pbo);
+
     if (pbo) {
 	    if (pbo->ptr)
 		radeon_bo_unmap(pbo);
diff --git a/src/mesa/drivers/dri/r600/r600_tex.c b/src/mesa/drivers/dri/r600/r600_tex.c
index a06a1fa921..d105b90cd1 100644
--- a/src/mesa/drivers/dri/r600/r600_tex.c
+++ b/src/mesa/drivers/dri/r600/r600_tex.c
@@ -67,7 +67,7 @@ static unsigned int translate_wrap_mode(GLenum wrapmode)
 	case GL_MIRROR_CLAMP_TO_EDGE_EXT: return SQ_TEX_MIRROR_ONCE_LAST_TEXEL;
 	case GL_MIRROR_CLAMP_TO_BORDER_EXT: return SQ_TEX_MIRROR_ONCE_BORDER;
 	default:
-		_mesa_problem(NULL, "bad wrap mode in %s", __FUNCTION__);
+		radeon_error("bad wrap mode in %s", __FUNCTION__);
 		return 0;
 	}
 }
@@ -208,8 +208,7 @@ static void r600SetTexFilter(radeonTexObjPtr t, GLenum minf, GLenum magf, GLfloa
 			| R300_TX_MIN_FILTER_ANISO
 			| R300_TX_MIN_FILTER_MIP_LINEAR
 			| aniso_filter(anisotropy);*/
-		if (RADEON_DEBUG & DEBUG_TEXTURE)
-			fprintf(stderr, "Using maximum anisotropy of %f\n", anisotropy);
+		radeon_print(RADEON_TEXTURE, RADEON_NORMAL, "Using maximum anisotropy of %f\n", anisotropy);
 		return;
 	}
 
@@ -288,10 +287,9 @@ static void r600TexParameter(GLcontext * ctx, GLenum target,
 {
 	radeonTexObj* t = radeon_tex_obj(texObj);
 
-	if (RADEON_DEBUG & (DEBUG_STATE | DEBUG_TEXTURE)) {
-		fprintf(stderr, "%s( %s )\n", __FUNCTION__,
+	radeon_print(RADEON_STATE | RADEON_TEXTURE, RADEON_VERBOSE,
+			"%s( %s )\n", __FUNCTION__,
 			_mesa_lookup_enum_by_nr(pname));
-	}
 
 	switch (pname) {
 	case GL_TEXTURE_MIN_FILTER:
@@ -351,11 +349,10 @@ static void r600DeleteTexture(GLcontext * ctx, struct gl_texture_object *texObj)
 	context_t* rmesa = R700_CONTEXT(ctx);
 	radeonTexObj* t = radeon_tex_obj(texObj);
 
-	if (RADEON_DEBUG & (DEBUG_STATE | DEBUG_TEXTURE)) {
-		fprintf(stderr, "%s( %p (target = %s) )\n", __FUNCTION__,
+	radeon_print(RADEON_STATE | RADEON_TEXTURE, RADEON_NORMAL,
+		"%s( %p (target = %s) )\n", __FUNCTION__,
 			(void *)texObj,
 			_mesa_lookup_enum_by_nr(texObj->Target));
-	}
 
 	if (rmesa) {
 		int i;
@@ -393,10 +390,9 @@ static struct gl_texture_object *r600NewTextureObject(GLcontext * ctx,
 	radeonTexObj* t = CALLOC_STRUCT(radeon_tex_obj);
 
 
-	if (RADEON_DEBUG & (DEBUG_STATE | DEBUG_TEXTURE)) {
-		fprintf(stderr, "%s( %p (target = %s) )\n", __FUNCTION__,
+	radeon_print(RADEON_STATE | RADEON_TEXTURE, RADEON_NORMAL,
+		"%s( %p (target = %s) )\n", __FUNCTION__,
 			t, _mesa_lookup_enum_by_nr(target));
-	}
 
 	_mesa_initialize_texture_object(&t->base, name, target);
 	t->base.MaxAnisotropy = rmesa->radeon.initialMaxAnisotropy;
diff --git a/src/mesa/drivers/dri/r600/r600_texstate.c b/src/mesa/drivers/dri/r600/r600_texstate.c
index 9f1bf45246..fff6e74386 100644
--- a/src/mesa/drivers/dri/r600/r600_texstate.c
+++ b/src/mesa/drivers/dri/r600/r600_texstate.c
@@ -49,6 +49,8 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "r700_state.h"
 #include "radeon_mipmap_tree.h"
 #include "r600_tex.h"
+#include "r700_fragprog.h"
+#include "r700_vertprog.h"
 
 void r600UpdateTextureState(GLcontext * ctx);
 
@@ -529,6 +531,49 @@ static GLboolean r600GetTexFormat(struct gl_texture_object *tObj, GLuint mesa_fo
 			return GL_FALSE;
 		}
 		break;
+	/* EXT_texture_sRGB */
+	case MESA_FORMAT_SRGBA8:
+		SETfield(t->SQ_TEX_RESOURCE1, FMT_8_8_8_8,
+			 SQ_TEX_RESOURCE_WORD1_0__DATA_FORMAT_shift, SQ_TEX_RESOURCE_WORD1_0__DATA_FORMAT_mask);
+
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_W,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_X_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_X_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_Z,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Y_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Y_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_Y,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Z_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Z_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_W_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_W_mask);
+		SETbit(t->SQ_TEX_RESOURCE4, SQ_TEX_RESOURCE_WORD4_0__FORCE_DEGAMMA_bit);
+		break;
+	case MESA_FORMAT_SLA8:
+		SETfield(t->SQ_TEX_RESOURCE1, FMT_8_8,
+			 SQ_TEX_RESOURCE_WORD1_0__DATA_FORMAT_shift, SQ_TEX_RESOURCE_WORD1_0__DATA_FORMAT_mask);
+
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_X_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_X_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Y_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Y_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Z_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Z_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_Y,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_W_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_W_mask);
+		SETbit(t->SQ_TEX_RESOURCE4, SQ_TEX_RESOURCE_WORD4_0__FORCE_DEGAMMA_bit);
+		break;
+	case MESA_FORMAT_SL8: /* X, X, X, ONE */
+		SETfield(t->SQ_TEX_RESOURCE1, FMT_8,
+			 SQ_TEX_RESOURCE_WORD1_0__DATA_FORMAT_shift, SQ_TEX_RESOURCE_WORD1_0__DATA_FORMAT_mask);
+
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_X_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_X_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Y_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Y_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_X,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Z_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_Z_mask);
+		SETfield(t->SQ_TEX_RESOURCE4, SQ_SEL_1,
+			 SQ_TEX_RESOURCE_WORD4_0__DST_SEL_W_shift, SQ_TEX_RESOURCE_WORD4_0__DST_SEL_W_mask);
+		SETbit(t->SQ_TEX_RESOURCE4, SQ_TEX_RESOURCE_WORD4_0__FORCE_DEGAMMA_bit);
+		break;
 	default:
 		/* Not supported format */
 		return GL_FALSE;
@@ -567,7 +612,7 @@ static void setup_hardware_state(context_t *rmesa, struct gl_texture_object *tex
 
 	if (!t->image_override) {
 		if (!r600GetTexFormat(texObj, firstImage->TexFormat->MesaFormat)) {
-			_mesa_problem(NULL, "unexpected texture format in %s",
+			radeon_error("unexpected texture format in %s\n",
 				      __FUNCTION__);
 			return;
 		}
@@ -593,7 +638,7 @@ static void setup_hardware_state(context_t *rmesa, struct gl_texture_object *tex
 		SETfield(t->SQ_TEX_RESOURCE1, 0, TEX_DEPTH_shift, TEX_DEPTH_mask);
 		break;
         default:
-		_mesa_problem(NULL, "unexpected texture target type in %s", __FUNCTION__);
+		radeon_error("unexpected texture target type in %s\n", __FUNCTION__);
 		return;
 	}
 
@@ -647,6 +692,7 @@ GLboolean r600ValidateBuffers(GLcontext * ctx)
 {
 	context_t *rmesa = R700_CONTEXT(ctx);
 	struct radeon_renderbuffer *rrb;
+	struct radeon_bo *pbo;
 	int i;
 	int ret;
 
@@ -675,9 +721,7 @@ GLboolean r600ValidateBuffers(GLcontext * ctx)
 			continue;
 
 		if (!r600_validate_texture(ctx, ctx->Texture.Unit[i]._Current)) {
-			_mesa_warning(ctx,
-				      "failed to validate texture for unit %d.\n",
-				      i);
+			radeon_warning("failed to validate texture for unit %d.\n", i);
 		}
 		t = radeon_tex_obj(ctx->Texture.Unit[i]._Current);
 		if (t->image_override && t->bo)
@@ -690,6 +734,18 @@ GLboolean r600ValidateBuffers(GLcontext * ctx)
 							  RADEON_GEM_DOMAIN_GTT | RADEON_GEM_DOMAIN_VRAM, 0);
 	}
 
+	pbo = (struct radeon_bo *)r700GetActiveFpShaderBo(ctx);
+	if (pbo) {
+		radeon_cs_space_add_persistent_bo(rmesa->radeon.cmdbuf.cs, pbo,
+						  RADEON_GEM_DOMAIN_GTT, 0);
+	}
+
+	pbo = (struct radeon_bo *)r700GetActiveVpShaderBo(ctx);
+	if (pbo) {
+		radeon_cs_space_add_persistent_bo(rmesa->radeon.cmdbuf.cs, pbo,
+						  RADEON_GEM_DOMAIN_GTT, 0);
+	}
+
 	ret = radeon_cs_space_check_with_bo(rmesa->radeon.cmdbuf.cs, first_elem(&rmesa->radeon.dma.reserved)->bo, RADEON_GEM_DOMAIN_GTT, 0);
 	if (ret)
 		return GL_FALSE;
diff --git a/src/mesa/drivers/dri/r600/r700_assembler.c b/src/mesa/drivers/dri/r600/r700_assembler.c
index 834c23593b..2d8480daaf 100644
--- a/src/mesa/drivers/dri/r600/r700_assembler.c
+++ b/src/mesa/drivers/dri/r600/r700_assembler.c
@@ -33,8 +33,8 @@
 #include "main/mtypes.h"
 #include "main/imports.h"
 
+#include "radeon_debug.h"
 #include "r600_context.h"
-#include "r700_debug.h"
 
 #include "r700_assembler.h"
 
@@ -366,8 +366,8 @@ unsigned int r700GetNumOperands(r700_AssemblerBase* pAsm)
     case SQ_OP2_INST_COS:
         return 1;
         
-    default: r700_error(TODO_ASM_NEEDIMPINST, 
-                        "Need instruction operand number. \n");;
+    default: radeon_error(
+		    "Need instruction operand number for %x.\n", pAsm->D.dst.opcode);
     };
 
     return 3;
@@ -531,7 +531,7 @@ int check_current_clause(r700_AssemblerBase* pAsm,
 		case CF_EMPTY_CLAUSE:
 			break;
 		default:
-            r700_error(ERROR_ASM_VTX_CLAUSE,
+            radeon_error(
                        "Unknown CF_CLAUSE_TYPE (%d) in check_current_clause. \n", (int) new_clause_type);
 			return GL_FALSE;
 		}
@@ -565,7 +565,7 @@ int check_current_clause(r700_AssemblerBase* pAsm,
                 }
                 else 
                 {
-                    r700_error(ERROR_ASM_ALLOCEXPORTCF,
+                    radeon_error(
                                "Error allocating new EXPORT CF instruction in check_current_clause. \n");
                     return GL_FALSE;
                 }
@@ -578,7 +578,7 @@ int check_current_clause(r700_AssemblerBase* pAsm,
             pAsm->cf_current_clause_type = CF_OTHER_CLAUSE;
             break;
         default:
-            r700_error(ERROR_ASM_UNKOWNCLAUSE,
+            radeon_error(
                        "Unknown CF_CLAUSE_TYPE (%d) in check_current_clause. \n", (int) new_clause_type);
             return GL_FALSE;
         }
@@ -611,7 +611,7 @@ GLboolean add_vfetch_instruction(r700_AssemblerBase*     pAsm,
 		}
 		else 
 		{
-            r700_error(ERROR_ASM_ALLOCVTXCF, "Could not allocate a new VFetch CF instruction.");
+            radeon_error("Could not allocate a new VFetch CF instruction.\n");
 			return GL_FALSE;
 		}
 
@@ -661,7 +661,7 @@ GLboolean add_tex_instruction(r700_AssemblerBase*     pAsm,
 		}
 		else 
 		{
-            r700_error(ERROR_ASM_ALLOCTEXCF, "Could not allocate a new TEX CF instruction.");
+            radeon_error("Could not allocate a new TEX CF instruction.\n");
 			return GL_FALSE;
 		}
         
@@ -1047,7 +1047,7 @@ GLboolean assemble_src(r700_AssemblerBase *pAsm,
             }
             break;      
         default:
-            r700_error(ERROR_ASM_SRCARGUMENT, "Invalid source argument type");          
+            radeon_error("Invalid source argument type\n");
             return GL_FALSE;
         }
     } 
@@ -1094,7 +1094,7 @@ GLboolean assemble_dst(r700_AssemblerBase *pAsm)
         }
         break;   
     default:
-        r700_error(ERROR_ASM_DSTARGUMENT, "Invalid destination output argument type");
+        radeon_error("Invalid destination output argument type\n");
         return GL_FALSE;
     }
 
@@ -1134,7 +1134,7 @@ GLboolean tex_dst(r700_AssemblerBase *pAsm)
     }
     else 
     {
-        r700_error(ERROR_ASM_DSTARGUMENT, "Invalid destination output argument type");
+        radeon_error("Invalid destination output argument type\n");
         return GL_FALSE;
     }
 
@@ -1188,7 +1188,7 @@ GLboolean tex_src(r700_AssemblerBase *pAsm)
     }
     else
     {
-        r700_error(ERROR_ASM_BADTEXSRC, "Invalid source texcoord for TEX instruction");
+        radeon_error("Invalid source texcoord for TEX instruction\n");
         return GL_FALSE;
     }
 
@@ -1269,7 +1269,7 @@ GLboolean assemble_tex_instruction(r700_AssemblerBase *pAsm, GLboolean normalize
     }
     else 
     {
-        r700_error(ERROR_ASM_TEXDSTBADTYPE, "Only temp destination registers supported for TEX dest regs.");
+        radeon_error("Only temp destination registers supported for TEX dest regs.\n");
         return GL_FALSE;
     }
 
@@ -1362,7 +1362,7 @@ GLboolean assemble_alu_src(R700ALUInstruction*  alu_instruction_ptr,
         }
         else
         {
-            r700_error(ERROR_ASM_ALUSRCBADTYPE, "Source (%d) register type (%d) not one of TEMP, INPUT, or CONSTANT.", 
+            radeon_error("Source (%d) register type (%d) not one of TEMP, INPUT, or CONSTANT.\n",
                      source_index, pSource->rtype);
             return GL_FALSE;
         }
@@ -1397,7 +1397,7 @@ GLboolean assemble_alu_src(R700ALUInstruction*  alu_instruction_ptr,
             src_chan = SQ_CHAN_X; 
             break;
         default:
-            r700_error(ERROR_ASM_ALUSRCSELECT, "Unknown source select value (%d) in assemble_alu_src().");
+            radeon_error("Unknown source select value (%d) in assemble_alu_src().\n", channel_swizzle);
             return GL_FALSE;
             break;
     }
@@ -1432,7 +1432,7 @@ GLboolean assemble_alu_src(R700ALUInstruction*  alu_instruction_ptr,
             alu_instruction_ptr->m_Word1_OP3.f.src2_neg  = src_neg;
             break;
         default:
-            r700_error(ERROR_ASM_ALUSRCNUMBER, "Only three sources allowed in ALU opcodes.");
+            radeon_error("Only three sources allowed in ALU opcodes.\n");
           return GL_FALSE;
           break;
     }
@@ -1467,7 +1467,7 @@ GLboolean add_alu_instruction(r700_AssemblerBase* pAsm,
         }
         else 
         {
-            r700_error(ERROR_ASM_ALLOCALUCF, "Could not allocate a new ALU CF instruction.");
+            radeon_error("Could not allocate a new ALU CF instruction.\n");
             return GL_FALSE;
         }
 
@@ -1635,7 +1635,7 @@ GLboolean reserve_cfile(r700_AssemblerBase* pAsm,
     }
     else 
     {
-        r700_error(ERROR_ASM_CONSTCHANNEL, "All cfile read ports are used, cannot reference C$sel, channel $chan.");
+        radeon_error("All cfile read ports are used, cannot reference C$sel, channel $chan.\n");
         return GL_FALSE;
     }
     return GL_TRUE;
@@ -1649,7 +1649,7 @@ GLboolean reserve_gpr(r700_AssemblerBase* pAsm, GLuint sel, GLuint chan, GLuint
     }
     else if(pAsm->hw_gpr[cycle][chan] != (int)sel) 
     {
-        r700_error(ERROR_ASM_BADGPRRESERVE, "Another scalar operation has already used GPR read port for given channel");
+        radeon_error("Another scalar operation has already used GPR read port for given channel\n");
         return GL_FALSE;
     }
 
@@ -1689,7 +1689,7 @@ GLboolean cycle_for_scalar_bank_swizzle(const int swiz, const int sel, GLuint* p
             }
             break;
         default:
-            r700_error(ERROR_ASM_BADSCALARBZ, "Bad Scalar bank swizzle value");
+            radeon_error("Bad Scalar bank swizzle value\n");
             break;
     }
 
@@ -1737,7 +1737,7 @@ GLboolean cycle_for_vector_bank_swizzle(const int swiz, const int sel, GLuint* p
             }
             break;
         default:
-            r700_error(ERROR_ASM_BADVECTORBZ, "Bad Vec bank swizzle value");
+            radeon_error("Bad Vec bank swizzle value\n");
             return GL_FALSE;
             break;
     }
@@ -2056,7 +2056,7 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
         }
         else 
         {
-            r700_error(ERROR_ASM_ALUDSTBADTYPE, "Only temp destination registers supported for ALU dest regs.");
+            radeon_error("Only temp destination registers supported for ALU dest regs.\n");
             return GL_FALSE;
         }
 
@@ -2207,13 +2207,13 @@ GLboolean next_ins(r700_AssemblerBase *pAsm)
 	    if (pILInst->TexSrcTarget == TEXTURE_RECT_INDEX) {
 		    if( GL_FALSE == assemble_tex_instruction(pAsm, GL_FALSE) ) 
 		    {
-			    r700_error(ERROR_ASM_TEXINSTRUCTION, "Error assembling TEX instruction");
+			    radeon_error("Error assembling TEX instruction\n");
 			    return GL_FALSE;
 		    }
 	    } else {
 		    if( GL_FALSE == assemble_tex_instruction(pAsm, GL_TRUE) ) 
 		    {
-			    r700_error(ERROR_ASM_TEXINSTRUCTION, "Error assembling TEX instruction");
+			    radeon_error("Error assembling TEX instruction\n");
 			    return GL_FALSE;
 		    }
 	    }
@@ -2222,7 +2222,7 @@ GLboolean next_ins(r700_AssemblerBase *pAsm)
     {   //ALU      
         if( GL_FALSE == assemble_alu_instruction(pAsm) ) 
         {
-            r700_error(ERROR_ASM_TEXINSTRUCTION, "Error assembling ALU instruction");
+            radeon_error("Error assembling ALU instruction\n");
             return GL_FALSE;
         }
     } 
@@ -2367,7 +2367,7 @@ GLboolean assemble_ADD(r700_AssemblerBase *pAsm)
 
 GLboolean assemble_BAD(char *opcode_str) 
 {
-    r700_error(TODO_ASM_NEEDIMPINST, "Not yet implemented instruction (%s)", opcode_str);
+    radeon_error("Not yet implemented instruction (%s)\n", opcode_str);
     return GL_FALSE;
 }
 
@@ -3381,7 +3381,7 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
 
     if (GL_TRUE == src_const) 
     {
-        r700_error(TODO_ASM_CONSTTEXADDR, "TODO: Texture coordinates from a constant register not supported.");
+        radeon_error("TODO: Texture coordinates from a constant register not supported.\n");
         return GL_FALSE;
     }
 
@@ -3391,7 +3391,7 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
             pAsm->D.dst.opcode = SQ_TEX_INST_SAMPLE;            
             break;
         case OPCODE_TXB:            
-            r700_error(TODO_ASM_TXB, "do not support TXB yet");
+            radeon_error("do not support TXB yet\n");
             return GL_FALSE;
             break;
         case OPCODE_TXP:            
@@ -3399,7 +3399,7 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
             pAsm->D.dst.opcode = SQ_TEX_INST_SAMPLE;
             break;
         default:
-            r700_error(ERROR_ASM_BADTEXINST, "Internal error: bad texture op (not TEX)");
+            radeon_error("Internal error: bad texture op (not TEX)\n");
             return GL_FALSE;
             break;
     }
@@ -3581,12 +3581,12 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
             break;  
 
         case OPCODE_ARL: 
-            r700_error(TODO_ASM_NEEDIMPINST, "Not yet implemented instruction OPCODE_ARL ");
+            radeon_error("Not yet implemented instruction OPCODE_ARL \n");
             //if ( GL_FALSE == assemble_BAD("ARL") ) 
                 return GL_FALSE;
             break;
         case OPCODE_ARR: 
-            r700_error(TODO_ASM_NEEDIMPINST, "Not yet implemented instruction OPCODE_ARR ");
+            radeon_error("Not yet implemented instruction OPCODE_ARR \n");
             //if ( GL_FALSE == assemble_BAD("ARR") ) 
                 return GL_FALSE;
             break;
@@ -3617,7 +3617,7 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
                 return GL_FALSE;
             break;  
         case OPCODE_EXP: 
-            r700_error(TODO_ASM_NEEDIMPINST, "Not yet implemented instruction OPCODE_EXP ");
+            radeon_error("Not yet implemented instruction OPCODE_EXP \n");
             //if ( GL_FALSE == assemble_BAD("EXP") ) 
                 return GL_FALSE;
             break; // approx of EX2
@@ -3653,7 +3653,7 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
                 return GL_FALSE;
             break;  
         case OPCODE_LOG: 
-            r700_error(TODO_ASM_NEEDIMPINST, "Not yet implemented instruction OPCODE_LOG ");
+            radeon_error("Not yet implemented instruction OPCODE_LOG \n");
             //if ( GL_FALSE == assemble_BAD("LOG") ) 
                 return GL_FALSE;
             break; // approx of LG2
@@ -3752,7 +3752,7 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
                 return GL_FALSE;
             break;
         case OPCODE_ELSE : 
-            r700_error(TODO_ASM_NEEDIMPINST, "Not yet implemented instruction OPCODE_ELSE ");
+            radeon_error("Not yet implemented instruction OPCODE_ELSE \n");
             //if ( GL_FALSE == assemble_BAD("ELSE") ) 
                 return GL_FALSE;
             break;
@@ -3774,7 +3774,7 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
             return GL_TRUE;
 
         default:
-            r700_error(ERROR_ASM_UNKNOWNILINST, "internal: unknown instruction");
+            radeon_error("internal: unknown instruction\n");
             return GL_FALSE;
         }
     }
@@ -3818,7 +3818,7 @@ GLboolean Process_Export(r700_AssemblerBase* pAsm,
             break;
 
         default:
-            r700_error(ERROR_ASM_BADEXPORTTYPE, "Unknown export type: %d", type);
+            radeon_error("Unknown export type: %d\n", type);
             return GL_FALSE;
             break;
     }
diff --git a/src/mesa/drivers/dri/r600/r700_chip.c b/src/mesa/drivers/dri/r600/r700_chip.c
index 16f1a3df36..37bff56f5a 100644
--- a/src/mesa/drivers/dri/r600/r700_chip.c
+++ b/src/mesa/drivers/dri/r600/r700_chip.c
@@ -49,6 +49,8 @@ static void r700SendTexState(GLcontext *ctx, struct radeon_state_atom *atom)
 	unsigned int i;
 	BATCH_LOCALS(&context->radeon);
 
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
+
 	for (i = 0; i < R700_TEXTURE_NUMBERUNITS; i++) {
 		radeonTexObj *t = r700->textures[i];
 		if (t) {
@@ -93,6 +95,7 @@ static void r700SendTexSamplerState(GLcontext *ctx, struct radeon_state_atom *at
 	R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
 	unsigned int i;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	for (i = 0; i < R700_TEXTURE_NUMBERUNITS; i++) {
 		radeonTexObj *t = r700->textures[i];
@@ -115,6 +118,7 @@ static void r700SendTexBorderColorState(GLcontext *ctx, struct radeon_state_atom
 	R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
 	unsigned int i;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	for (i = 0; i < R700_TEXTURE_NUMBERUNITS; i++) {
 		radeonTexObj *t = r700->textures[i];
@@ -141,6 +145,7 @@ static void r700SetupVTXConstants(GLcontext  * ctx,
     context_t *context = R700_CONTEXT(ctx);
     struct radeon_aos * paos = (struct radeon_aos *)pAos;
     BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
     unsigned int uSQ_VTX_CONSTANT_WORD0_0;
     unsigned int uSQ_VTX_CONSTANT_WORD1_0;
@@ -205,6 +210,7 @@ void r700SetupStreams(GLcontext *ctx)
     TNLcontext *tnl = TNL_CONTEXT(ctx);
     struct vertex_buffer *vb = &tnl->vb;
     unsigned int i, j = 0;
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
     R600_STATECHANGE(context, vtx);
 
@@ -229,6 +235,7 @@ static void r700SendVTXState(GLcontext *ctx, struct radeon_state_atom *atom)
              = (struct r700_vertex_program *)ctx->VertexProgram._Current;
     unsigned int i, j = 0;
     BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
     if (context->radeon.tcl.aos_count == 0)
 	    return;
@@ -258,12 +265,100 @@ static void r700SendVTXState(GLcontext *ctx, struct radeon_state_atom *atom)
     }
 }
 
+static void r700SetRenderTarget(context_t *context, int id)
+{
+    R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
+
+    struct radeon_renderbuffer *rrb;
+    unsigned int nPitchInPixel;
+
+    rrb = radeon_get_colorbuffer(&context->radeon);
+    if (!rrb || !rrb->bo) {
+	    return;
+    }
+
+    R600_STATECHANGE(context, cb_target);
+
+    /* color buffer */
+    r700->render_target[id].CB_COLOR0_BASE.u32All = context->radeon.state.color.draw_offset;
+
+    nPitchInPixel = rrb->pitch/rrb->cpp;
+    SETfield(r700->render_target[id].CB_COLOR0_SIZE.u32All, (nPitchInPixel/8)-1,
+             PITCH_TILE_MAX_shift, PITCH_TILE_MAX_mask);
+    SETfield(r700->render_target[id].CB_COLOR0_SIZE.u32All, ( (nPitchInPixel * context->radeon.radeonScreen->driScreen->fbHeight)/64 )-1,
+             SLICE_TILE_MAX_shift, SLICE_TILE_MAX_mask);
+    r700->render_target[id].CB_COLOR0_BASE.u32All = 0;
+    SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, ENDIAN_NONE, ENDIAN_shift, ENDIAN_mask);
+    SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, ARRAY_LINEAR_GENERAL,
+             CB_COLOR0_INFO__ARRAY_MODE_shift, CB_COLOR0_INFO__ARRAY_MODE_mask);
+    if(4 == rrb->cpp)
+    {
+        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, COLOR_8_8_8_8,
+                 CB_COLOR0_INFO__FORMAT_shift, CB_COLOR0_INFO__FORMAT_mask);
+        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, SWAP_ALT, COMP_SWAP_shift, COMP_SWAP_mask);
+    }
+    else
+    {
+        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, COLOR_5_6_5,
+                 CB_COLOR0_INFO__FORMAT_shift, CB_COLOR0_INFO__FORMAT_mask);
+        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, SWAP_ALT_REV,
+                 COMP_SWAP_shift, COMP_SWAP_mask);
+    }
+    SETbit(r700->render_target[id].CB_COLOR0_INFO.u32All, SOURCE_FORMAT_bit);
+    SETbit(r700->render_target[id].CB_COLOR0_INFO.u32All, BLEND_CLAMP_bit);
+    SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, NUMBER_UNORM, NUMBER_TYPE_shift, NUMBER_TYPE_mask);
+
+    r700->render_target[id].enabled = GL_TRUE;
+}
+
+static void r700SetDepthTarget(context_t *context)
+{
+    R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
+
+    struct radeon_renderbuffer *rrb;
+    unsigned int nPitchInPixel;
+
+    rrb = radeon_get_depthbuffer(&context->radeon);
+    if (!rrb)
+	    return;
+
+    R600_STATECHANGE(context, db_target);
+
+    /* depth buf */
+    r700->DB_DEPTH_SIZE.u32All = 0;
+    r700->DB_DEPTH_BASE.u32All = 0;
+    r700->DB_DEPTH_INFO.u32All = 0;
+    r700->DB_DEPTH_VIEW.u32All = 0;
+
+    nPitchInPixel = rrb->pitch/rrb->cpp;
+
+    SETfield(r700->DB_DEPTH_SIZE.u32All, (nPitchInPixel/8)-1,
+             PITCH_TILE_MAX_shift, PITCH_TILE_MAX_mask);
+    SETfield(r700->DB_DEPTH_SIZE.u32All, ( (nPitchInPixel * context->radeon.radeonScreen->driScreen->fbHeight)/64 )-1,
+             SLICE_TILE_MAX_shift, SLICE_TILE_MAX_mask); /* size in pixel / 64 - 1 */
+
+    if(4 == rrb->cpp)
+    {
+        SETfield(r700->DB_DEPTH_INFO.u32All, DEPTH_8_24,
+                 DB_DEPTH_INFO__FORMAT_shift, DB_DEPTH_INFO__FORMAT_mask);
+    }
+    else
+    {
+        SETfield(r700->DB_DEPTH_INFO.u32All, DEPTH_16,
+                     DB_DEPTH_INFO__FORMAT_shift, DB_DEPTH_INFO__FORMAT_mask);
+    }
+    SETfield(r700->DB_DEPTH_INFO.u32All, ARRAY_2D_TILED_THIN1,
+             DB_DEPTH_INFO__ARRAY_MODE_shift, DB_DEPTH_INFO__ARRAY_MODE_mask);
+    /* r700->DB_PREFETCH_LIMIT.bits.DEPTH_HEIGHT_TILE_MAX = (context->currentDraw->h >> 3) - 1; */ /* z buffer sie may much bigger than what need, so use actual used h. */
+}
+
 static void r700SendDepthTargetState(GLcontext *ctx, struct radeon_state_atom *atom)
 {
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	struct radeon_renderbuffer *rrb;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	rrb = radeon_get_depthbuffer(&context->radeon);
 	if (!rrb || !rrb->bo) {
@@ -271,6 +366,8 @@ static void r700SendDepthTargetState(GLcontext *ctx, struct radeon_state_atom *a
 		return;
 	}
 
+	r700SetDepthTarget(context);
+
         BEGIN_BATCH_NO_AUTOSTATE(8 + 2);
 	R600_OUT_BATCH_REGSEQ(DB_DEPTH_SIZE, 2);
 	R600_OUT_BATCH(r700->DB_DEPTH_SIZE.u32All);
@@ -303,6 +400,7 @@ static void r700SendRenderTargetState(GLcontext *ctx, struct radeon_state_atom *
 	struct radeon_renderbuffer *rrb;
 	BATCH_LOCALS(&context->radeon);
 	int id = 0;
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	rrb = radeon_get_colorbuffer(&context->radeon);
 	if (!rrb || !rrb->bo) {
@@ -310,6 +408,8 @@ static void r700SendRenderTargetState(GLcontext *ctx, struct radeon_state_atom *
 		return;
 	}
 
+	r700SetRenderTarget(context, 0);
+
 	if (id > R700_MAX_RENDER_TARGETS)
 		return;
 
@@ -352,6 +452,7 @@ static void r700SendPSState(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	struct radeon_bo * pbo;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	pbo = (struct radeon_bo *)r700GetActiveFpShaderBo(GL_CONTEXT(context));
 
@@ -385,6 +486,7 @@ static void r700SendVSState(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	struct radeon_bo * pbo;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	pbo = (struct radeon_bo *)r700GetActiveVpShaderBo(GL_CONTEXT(context));
 
@@ -416,6 +518,7 @@ static void r700SendFSState(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	struct radeon_bo * pbo;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	/* XXX fixme
 	 * R6xx chips require a FS be emitted, even if it's not used.
@@ -457,6 +560,7 @@ static void r700SendViewportState(GLcontext *ctx, struct radeon_state_atom *atom
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
 	int id = 0;
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	if (id > R700_MAX_VIEWPORTS)
 		return;
@@ -489,6 +593,7 @@ static void r700SendSQConfig(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
         BEGIN_BATCH_NO_AUTOSTATE(34);
 	R600_OUT_BATCH_REGSEQ(SQ_CONFIG, 6);
@@ -526,6 +631,7 @@ static void r700SendUCPState(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
 	int i;
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	for (i = 0; i < R700_MAX_UCP; i++) {
 		if (r700->ucp[i].enabled) {
@@ -547,6 +653,7 @@ static void r700SendSPIState(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
 	unsigned int ui;
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	BEGIN_BATCH_NO_AUTOSTATE(59 + R700_MAX_SHADER_EXPORTS);
 
@@ -620,6 +727,7 @@ static void r700SendVGTState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
         BEGIN_BATCH_NO_AUTOSTATE(41);
 
@@ -665,6 +773,7 @@ static void r700SendSXState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
         BEGIN_BATCH_NO_AUTOSTATE(9);
 	R600_OUT_BATCH_REGVAL(SX_MISC, r700->SX_MISC.u32All);
@@ -679,6 +788,7 @@ static void r700SendDBState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
         BEGIN_BATCH_NO_AUTOSTATE(23);
 	R600_OUT_BATCH_REGVAL(DB_HTILE_DATA_BASE, r700->DB_HTILE_DATA_BASE.u32All);
@@ -720,6 +830,7 @@ static void r700SendCBState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	if (context->radeon.radeonScreen->chip_family < CHIP_FAMILY_RV770) {
 		BEGIN_BATCH_NO_AUTOSTATE(11);
@@ -766,6 +877,7 @@ static void r700SendCBBlendState(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
 	unsigned int ui;
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	if (context->radeon.radeonScreen->chip_family < CHIP_FAMILY_RV770) {
 		BEGIN_BATCH_NO_AUTOSTATE(3);
@@ -796,6 +908,7 @@ static void r700SendCBBlendColorState(GLcontext *ctx, struct radeon_state_atom *
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	BEGIN_BATCH_NO_AUTOSTATE(6);
 	R600_OUT_BATCH_REGSEQ(CB_BLEND_RED, 4);
@@ -850,6 +963,7 @@ static void r700SendCLState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	BEGIN_BATCH_NO_AUTOSTATE(12);
 	R600_OUT_BATCH_REGVAL(PA_CL_CLIP_CNTL, r700->PA_CL_CLIP_CNTL.u32All);
@@ -881,6 +995,7 @@ static void r700SendScissorState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	BEGIN_BATCH_NO_AUTOSTATE(22);
 	R600_OUT_BATCH_REGSEQ(PA_SC_SCREEN_SCISSOR_TL, 2);
@@ -913,6 +1028,7 @@ static void r700SendSCState(GLcontext *ctx, struct radeon_state_atom *atom)
 	context_t *context = R700_CONTEXT(ctx);
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	BEGIN_BATCH_NO_AUTOSTATE(15);
 	R600_OUT_BATCH_REGVAL(R7xx_PA_SC_EDGERULE, r700->PA_SC_EDGERULE.u32All);
@@ -969,6 +1085,7 @@ static void r700SendVSConsts(GLcontext *ctx, struct radeon_state_atom *atom)
 	R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
 	int i;
 	BATCH_LOCALS(&context->radeon);
+	radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
 	if (r700->vs.num_consts == 0)
 		return;
@@ -999,6 +1116,7 @@ static int check_cb(GLcontext *ctx, struct radeon_state_atom *atom)
 
 	if (context->radeon.radeonScreen->chip_family < CHIP_FAMILY_RV770)
 		count += 11;
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 
 	return count;
 }
@@ -1019,6 +1137,7 @@ static int check_blnd(GLcontext *ctx, struct radeon_state_atom *atom)
 				count += 3;
 		}
 	}
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 
 	return count;
 }
@@ -1034,6 +1153,7 @@ static int check_ucp(GLcontext *ctx, struct radeon_state_atom *atom)
 		if (r700->ucp[i].enabled)
 			count += 6;
 	}
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 	return count;
 }
 
@@ -1045,6 +1165,7 @@ static int check_vtx(GLcontext *ctx, struct radeon_state_atom *atom)
 	if (count)
 		count += 6;
 
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 	return count;
 }
 
@@ -1059,6 +1180,7 @@ static int check_tx(GLcontext *ctx, struct radeon_state_atom *atom)
 		if (t)
 			count++;
 	}
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 	return count * 31;
 }
 
@@ -1070,6 +1192,7 @@ static int check_ps_consts(GLcontext *ctx, struct radeon_state_atom *atom)
 
 	if (count)
 		count += 2;
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 
 	return count;
 }
@@ -1082,6 +1205,7 @@ static int check_vs_consts(GLcontext *ctx, struct radeon_state_atom *atom)
 
 	if (count)
 		count += 2;
+	radeon_print(RADEON_STATE, RADEON_TRACE, "%s %d\n", __func__, count);
 
 	return count;
 }
@@ -1101,6 +1225,7 @@ do {									\
 
 void r600InitAtoms(context_t *context)
 {
+	radeon_print(RADEON_STATE, RADEON_NORMAL, "%s %p\n", __func__, context);
 	context->radeon.hw.max_state_size = 10 + 5 + 14; /* start 3d, idle, cb/db flush */
 
 	/* Setup the atom linked list */
diff --git a/src/mesa/drivers/dri/r600/r700_clear.c b/src/mesa/drivers/dri/r600/r700_clear.c
index 6d4ea90ccc..c6546ab00c 100644
--- a/src/mesa/drivers/dri/r600/r700_clear.c
+++ b/src/mesa/drivers/dri/r600/r700_clear.c
@@ -55,6 +55,8 @@ void r700Clear(GLcontext * ctx, GLbitfield mask)
     int i;
     struct gl_framebuffer *fb = ctx->DrawBuffer;
 
+    radeon_print(RADEON_RENDER, RADEON_VERBOSE, "%s %x\n", __func__, mask);
+
     if( GL_TRUE == r700ClearFast(context, mask) )
     {
         return;
@@ -106,8 +108,7 @@ void r700Clear(GLcontext * ctx, GLbitfield mask)
 	}
 
 	if (swrast_mask) {
-		if (RADEON_DEBUG & DEBUG_FALLBACKS)
-			fprintf(stderr, "%s: swrast clear, mask: %x\n",
+		radeon_print(RADEON_FALLBACKS, RADEON_IMPORTANT, "%s: swrast clear, mask: %x\n",
 				__FUNCTION__, swrast_mask);
 		_swrast_Clear(ctx, swrast_mask);
 	}
diff --git a/src/mesa/drivers/dri/r600/r700_debug.c b/src/mesa/drivers/dri/r600/r700_debug.c
index dabd0d2c25..cd1ba9eca3 100644
--- a/src/mesa/drivers/dri/r600/r700_debug.c
+++ b/src/mesa/drivers/dri/r600/r700_debug.c
@@ -25,31 +25,8 @@
  *   CooperYuan <cooper.yuan@amd.com>, <cooperyuan@gmail.com>
  */
 
-#include <stdio.h>
-#include <stdarg.h>
-#include <stdlib.h>
-#include <string.h>
-
-#include "main/glheader.h"
-
 #include "r700_debug.h"
-#include "r600_context.h"
-
-void NormalizeLogErrorCode(int nError)
-{
-    //TODO
-}
-
-void r700_error(int nLocalError, char* fmt, ...)
-{
-    va_list args;
-
-    NormalizeLogErrorCode(nLocalError);
-
-	va_start(args, fmt);
-    fprintf(stderr, fmt, args);
-    va_end(args);
-}
+#include "radeon_debug.h"
 
 void DumpHwBinary(int type, void *addr, int size)
 {
@@ -61,21 +38,21 @@ void DumpHwBinary(int type, void *addr, int size)
     switch (type)
     {
         case DUMP_PIXEL_SHADER:
-            DEBUGF("Pixel Shader\n");
+            radeon_print(RADEON_SHADER, RADEON_TRACE, "Pixel Shader\n");
         break;
         case DUMP_VERTEX_SHADER:
-            DEBUGF("Vertex Shader\n");
+            radeon_print(RADEON_SHADER, RADEON_TRACE, "Vertex Shader\n");
         break;
         case DUMP_FETCH_SHADER:
-            DEBUGF("Fetch Shader\n");
+            radeon_print(RADEON_SHADER, RADEON_TRACE, "Fetch Shader\n");
         break;
     }
 
     for (i = 0; i < size; i++)
     {
-        DEBUGP("0x%08x,\t", *pHw);
+        radeon_print(RADEON_SHADER, RADEON_TRACE, "0x%08x,\t", *pHw);
         if (i%4 == 3)
-            DEBUGP("0x%08x\n", *pHw);
+            radeon_print(RADEON_SHADER, RADEON_TRACE, "0x%08x\n", *pHw);
         pHw++;
 
     }
diff --git a/src/mesa/drivers/dri/r600/r700_debug.h b/src/mesa/drivers/dri/r600/r700_debug.h
index e810e6da08..c0921bf610 100644
--- a/src/mesa/drivers/dri/r600/r700_debug.h
+++ b/src/mesa/drivers/dri/r600/r700_debug.h
@@ -27,67 +27,13 @@
 
 #ifndef _R700_DEBUG_H_
 #define _R700_DEBUG_H_
-
-enum R700_ERROR
-{
-    ERROR_ASM_VTX_CLAUSE     = 0x1000,
-    ERROR_ASM_UNKOWNCLAUSE   = 0x1001,
-    ERROR_ASM_ALLOCEXPORTCF  = 0x1002,
-    ERROR_ASM_ALLOCVTXCF     = 0x1003,
-    ERROR_ASM_ALLOCTEXCF     = 0x1004,
-    ERROR_ASM_ALLOCALUCF     = 0x1005,
-    ERROR_ASM_UNKNOWNILINST  = 0x1006,
-    ERROR_ASM_SRCARGUMENT    = 0x1007,
-    ERROR_ASM_DSTARGUMENT    = 0x1008,
-    ERROR_ASM_TEXINSTRUCTION = 0x1009,
-    ERROR_ASM_ALUINSTRUCTION = 0x100A,
-    ERROR_ASM_INSTDSTTRACK   = 0x100B,
-    ERROR_ASM_TEXDSTBADTYPE  = 0x100C,
-    ERROR_ASM_ALUSRCBADTYPE  = 0x100D,
-    ERROR_ASM_ALUSRCSELECT   = 0x100E,
-    ERROR_ASM_ALUSRCNUMBER   = 0x100F,
-    ERROR_ASM_ALUDSTBADTYPE  = 0x1010,
-    ERROR_ASM_CONSTCHANNEL   = 0x1011,
-    ERROR_ASM_BADSCALARBZ    = 0x1012,
-    ERROR_ASM_BADGPRRESERVE  = 0x1013,
-    ERROR_ASM_BADVECTORBZ    = 0x1014,
-    ERROR_ASM_BADTEXINST     = 0x1015,
-    ERROR_ASM_BADTEXSRC      = 0x1016,
-    ERROR_ASM_BADEXPORTTYPE  = 0x1017,
-
-
-    TODO_ASM_CONSTTEXADDR   = 0x8000,
-    TODO_ASM_NEEDIMPINST    = 0x8001,
-    TODO_ASM_TXB            = 0x8002,
-    TODO_ASM_TXP            = 0x8003
-};
-
 enum R700_DUMP_TYPE
 {
-    DUMP_VERTEX_SHADER      = 0x1,
-    DUMP_PIXEL_SHADER       = 0x2,
-    DUMP_FETCH_SHADER       = 0x4,
+	DUMP_VERTEX_SHADER      = 0x1,
+	DUMP_PIXEL_SHADER       = 0x2,
+	DUMP_FETCH_SHADER       = 0x4,
 };
 
-#define DEBUGF printf
-#define DEBUGP printf
-
-void NormalizeLogErrorCode(int nError);
-/*NormalizeLogErrorCode(nLocalError); */
-void r700_error(int nLocalError, char *fmt, ...);      
 extern void DumpHwBinary(int, void *, int);
 
-#ifdef STANDALONE_COMPILER
-#ifdef __cplusplus
-extern "C"
-{
-#endif //__cplusplus
-
-void LogString(char* szStr);
-
-#ifdef __cplusplus
-}
-#endif //__cplusplus
-#endif /*STANDALONE_COMPILER*/
-
 #endif /*_R700_DEBUG_H_*/
diff --git a/src/mesa/drivers/dri/r600/r700_oglprog.c b/src/mesa/drivers/dri/r600/r700_oglprog.c
index c49b90c1cc..3c8c1fd7a3 100644
--- a/src/mesa/drivers/dri/r600/r700_oglprog.c
+++ b/src/mesa/drivers/dri/r600/r700_oglprog.c
@@ -49,6 +49,9 @@ static struct gl_program *r700NewProgram(GLcontext * ctx,
     struct r700_vertex_program *vp;
 	struct r700_fragment_program *fp;
 
+	radeon_print(RADEON_SHADER, RADEON_VERBOSE,
+			"%s %u, %u\n", __func__, target, id);
+
     switch (target) 
     {
     case GL_VERTEX_STATE_PROGRAM_NV:
@@ -89,6 +92,9 @@ static void r700DeleteProgram(GLcontext * ctx, struct gl_program *prog)
     struct r700_vertex_program   * vp;
     struct r700_fragment_program * fp;
 
+	radeon_print(RADEON_SHADER, RADEON_VERBOSE,
+			"%s %p\n", __func__, prog);
+
     switch (prog->Target) 
     {
     case GL_VERTEX_STATE_PROGRAM_NV:
diff --git a/src/mesa/drivers/dri/r600/r700_render.c b/src/mesa/drivers/dri/r600/r700_render.c
index 0b3ceb01ba..3566bf3ca7 100644
--- a/src/mesa/drivers/dri/r600/r700_render.c
+++ b/src/mesa/drivers/dri/r600/r700_render.c
@@ -69,6 +69,7 @@ GLboolean r700SyncSurf(context_t *context,
 void r700WaitForIdle(context_t *context)
 {
     BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_RENDER | RADEON_STATE, RADEON_TRACE, "%s\n", __func__);
     BEGIN_BATCH_NO_AUTOSTATE(3);
 
     R600_OUT_BATCH(CP_PACKET3(R600_IT_SET_CONFIG_REG, 1));
@@ -82,6 +83,7 @@ void r700WaitForIdle(context_t *context)
 void r700WaitForIdleClean(context_t *context)
 {
     BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_RENDER | RADEON_STATE, RADEON_TRACE, "%s\n", __func__);
     BEGIN_BATCH_NO_AUTOSTATE(5);
 
     R600_OUT_BATCH(CP_PACKET3(R600_IT_EVENT_WRITE, 0));
@@ -98,6 +100,7 @@ void r700WaitForIdleClean(context_t *context)
 void r700Start3D(context_t *context)
 {
     BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_RENDER | RADEON_STATE, RADEON_TRACE, "%s\n", __func__);
     if (context->radeon.radeonScreen->chip_family < CHIP_FAMILY_RV770)
     {
         BEGIN_BATCH_NO_AUTOSTATE(2);
@@ -124,6 +127,7 @@ GLboolean r700SyncSurf(context_t *context,
 		       uint32_t sync_type)
 {
     BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_RENDER | RADEON_STATE, RADEON_TRACE, "%s\n", __func__);
     uint32_t cp_coher_size;
 
     if (!pbo)
@@ -253,10 +257,16 @@ static void r700RunRenderPrimitive(GLcontext * ctx, int start, int end, int prim
 	uint32_t vgt_index_type     = 0;
 	uint32_t vgt_primitive_type = 0;
 	uint32_t vgt_num_indices    = 0;
+	TNLcontext *tnl = TNL_CONTEXT(ctx);
+	struct vertex_buffer *vb = &tnl->vb;
 
 	type = r700PrimitiveType(prim);
 	num_indices = r700NumVerts(end - start, prim);
 
+	radeon_print(RADEON_RENDER, RADEON_TRACE,
+		"%s type %x num_indices %d\n",
+		__func__, type, num_indices);
+
 	if (type < 0 || num_indices <= 0)
 		return;
 
@@ -292,7 +302,10 @@ static void r700RunRenderPrimitive(GLcontext * ctx, int start, int end, int prim
         R600_OUT_BATCH(vgt_draw_initiator);
 
         for (i = start; i < (start + num_indices); i++) {
-            R600_OUT_BATCH(i);
+		if(vb->Elts)
+			R600_OUT_BATCH(vb->Elts[i]);
+		else
+			R600_OUT_BATCH(i);
         }
         END_BATCH();
         COMMIT_BATCH();
@@ -327,8 +340,8 @@ static GLuint r700PredictRenderSize(GLcontext* ctx)
     else
         dwords += state_size;
 
-    if (RADEON_DEBUG & DEBUG_PRIMS)
-        fprintf(stderr, "%s: total prediction size is %d.\n", __FUNCTION__, dwords);
+    radeon_print(RADEON_RENDER, RADEON_VERBOSE,
+	"%s: total prediction size is %d.\n", __FUNCTION__, dwords);
     return dwords;
 }
 
@@ -342,8 +355,7 @@ static GLboolean r700RunRender(GLcontext * ctx,
     struct vertex_buffer *vb = &tnl->vb;
     struct radeon_renderbuffer *rrb;
 
-    if (RADEON_DEBUG & DEBUG_PRIMS)
-        fprintf(stderr, "%s: cs begin at %d\n",
+    radeon_print(RADEON_RENDER, RADEON_NORMAL, "%s: cs begin at %d\n",
                 __func__, context->radeon.cmdbuf.cs->cdw);
 
     /* always emit CB base to prevent
@@ -365,6 +377,7 @@ static GLboolean r700RunRender(GLcontext * ctx,
 
     radeonEmitState(radeon);
 
+    radeon_debug_add_indent();
     /* richard test code */
     for (i = 0; i < vb->PrimitiveCount; i++) {
         GLuint prim = _tnl_translate_prim(&vb->Primitive[i]);
@@ -372,6 +385,7 @@ static GLboolean r700RunRender(GLcontext * ctx,
         GLuint end = vb->Primitive[i].start + vb->Primitive[i].count;
         r700RunRenderPrimitive(ctx, start, end, prim);
     }
+    radeon_debug_remove_indent();
 
     /* Flush render op cached for last several quads. */
     r700WaitForIdleClean(context);
@@ -388,8 +402,7 @@ static GLboolean r700RunRender(GLcontext * ctx,
 
     radeonReleaseArrays(ctx, ~0);
 
-    if (RADEON_DEBUG & DEBUG_PRIMS)
-        fprintf(stderr, "%s: cs end at %d\n",
+    radeon_print(RADEON_RENDER, RADEON_TRACE, "%s: cs end at %d\n",
                 __func__, context->radeon.cmdbuf.cs->cdw);
 
     if ( emit_end < context->radeon.cmdbuf.cs->cdw )
diff --git a/src/mesa/drivers/dri/r600/r700_shaderinst.c b/src/mesa/drivers/dri/r600/r700_shaderinst.c
index c1bffee91f..f120d9f941 100644
--- a/src/mesa/drivers/dri/r600/r700_shaderinst.c
+++ b/src/mesa/drivers/dri/r600/r700_shaderinst.c
@@ -27,7 +27,7 @@
 
 #include "main/mtypes.h"
 
-#include "r700_debug.h"
+#include "radeon_debug.h"
 #include "r700_shaderinst.h"
 
 void Init_R700ControlFlowGenericClause(R700ControlFlowGenericClause* pInst)
@@ -208,16 +208,16 @@ unsigned int GetCFMaxInstructions(ShaderInstType instType)
 GLboolean LinkVertexInstruction(R700ControlFlowGenericClause *pCFGeneric,
 								R700VertexInstruction *pVTXInstruction)
 {
-	if (pCFGeneric->m_pLinkedTEXInstruction != 0) 
-	{
-		r700_error(ERROR_ASM_VTX_CLAUSE, "This instruction is already linked to a texture instruction");
-		return GL_FALSE;
+    if (pCFGeneric->m_pLinkedTEXInstruction != 0)
+    {
+	radeon_error("This instruction is already linked to a texture instruction.\n");
+	return GL_FALSE;
     }
 
     pCFGeneric->m_pLinkedVTXInstruction     = pVTXInstruction;
-	pVTXInstruction->m_pLinkedGenericClause = pCFGeneric;
+    pVTXInstruction->m_pLinkedGenericClause = pCFGeneric;
 
-	return GL_TRUE;
+    return GL_TRUE;
 }
 
 
diff --git a/src/mesa/drivers/dri/r600/r700_state.c b/src/mesa/drivers/dri/r600/r700_state.c
index 91cb492ce6..93b4ebfdb9 100644
--- a/src/mesa/drivers/dri/r600/r700_state.c
+++ b/src/mesa/drivers/dri/r600/r700_state.c
@@ -60,13 +60,6 @@ static void r700SetClipPlaneState(GLcontext * ctx, GLenum cap, GLboolean state);
 static void r700UpdatePolygonMode(GLcontext * ctx);
 static void r700SetPolygonOffsetState(GLcontext * ctx, GLboolean state);
 static void r700SetStencilState(GLcontext * ctx, GLboolean state);
-static void r700SetRenderTarget(context_t *context, int id);
-static void r700SetDepthTarget(context_t *context);
-
-void r700SetDefaultStates(context_t *context) //--------------------
-{
-    
-}
 
 void r700UpdateShaders (GLcontext * ctx)  //----------------------------------
 {
@@ -133,21 +126,6 @@ void r700UpdateViewportOffset(GLcontext * ctx) //------------------
 	radeonUpdateScissor(ctx);
 }
 
-/**
- * Tell the card where to render (offset, pitch).
- * Effected by glDrawBuffer, etc
- */
-void r700UpdateDrawBuffer(GLcontext * ctx) /* TODO */ //---------------------
-{
-	context_t *context = R700_CONTEXT(ctx);
-
-	R600_STATECHANGE(context, cb_target);
-	R600_STATECHANGE(context, db_target);
-
-	r700SetRenderTarget(context, 0);
-	r700SetDepthTarget(context);
-}
-
 void r700UpdateStateParameters(GLcontext * ctx, GLuint new_state) //--------------------
 {
 	struct r700_fragment_program *fp =
@@ -179,21 +157,29 @@ static void r700InvalidateState(GLcontext * ctx, GLuint new_state) //-----------
     R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
 
     _swrast_InvalidateState(ctx, new_state);
-	_swsetup_InvalidateState(ctx, new_state);
-	_vbo_InvalidateState(ctx, new_state);
-	_tnl_InvalidateState(ctx, new_state);
-	_ae_invalidate_state(ctx, new_state);
-
-	if (new_state & (_NEW_BUFFERS | _NEW_COLOR | _NEW_PIXEL)) 
-    {
-        _mesa_update_framebuffer(ctx);
-		/* this updates the DrawBuffer's Width/Height if it's a FBO */
-		_mesa_update_draw_buffer_bounds(ctx);
+    _swsetup_InvalidateState(ctx, new_state);
+    _vbo_InvalidateState(ctx, new_state);
+    _tnl_InvalidateState(ctx, new_state);
+    _ae_invalidate_state(ctx, new_state);
+
+    if (new_state & _NEW_BUFFERS) {
+	    _mesa_update_framebuffer(ctx);
+	    /* this updates the DrawBuffer's Width/Height if it's a FBO */
+	    _mesa_update_draw_buffer_bounds(ctx);
+
+	    R600_STATECHANGE(context, cb_target);
+	    R600_STATECHANGE(context, db_target);
+    }
 
-		r700UpdateDrawBuffer(ctx);
-	}
+    if (new_state & (_NEW_LIGHT)) {
+	    R600_STATECHANGE(context, su);
+	    if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION)
+		    SETbit(r700->PA_SU_SC_MODE_CNTL.u32All, PROVOKING_VTX_LAST_bit);
+	    else
+		    CLEARbit(r700->PA_SU_SC_MODE_CNTL.u32All, PROVOKING_VTX_LAST_bit);
+    }
 
-	r700UpdateStateParameters(ctx, new_state);
+    r700UpdateStateParameters(ctx, new_state);
 
     R600_STATECHANGE(context, cl);
     R600_STATECHANGE(context, spi);
@@ -1373,107 +1359,6 @@ void r700SetScissor(context_t *context) //---------------
 	r700->viewport[id].enabled = GL_TRUE;
 }
 
-static void r700SetRenderTarget(context_t *context, int id)
-{
-    R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
-
-    struct radeon_renderbuffer *rrb;
-    unsigned int nPitchInPixel;
-
-    rrb = radeon_get_colorbuffer(&context->radeon);
-    if (!rrb || !rrb->bo) {
-	    return;
-    }
-
-    R600_STATECHANGE(context, cb_target);
-    R600_STATECHANGE(context, cb);
-
-    /* screen/window/view */
-    SETfield(r700->CB_TARGET_MASK.u32All, 0xF, (4 * id), TARGET0_ENABLE_mask);
-
-    /* color buffer */
-    r700->render_target[id].CB_COLOR0_BASE.u32All = context->radeon.state.color.draw_offset;
-
-    nPitchInPixel = rrb->pitch/rrb->cpp;
-    SETfield(r700->render_target[id].CB_COLOR0_SIZE.u32All, (nPitchInPixel/8)-1,
-             PITCH_TILE_MAX_shift, PITCH_TILE_MAX_mask);
-    SETfield(r700->render_target[id].CB_COLOR0_SIZE.u32All, ( (nPitchInPixel * context->radeon.radeonScreen->driScreen->fbHeight)/64 )-1,
-             SLICE_TILE_MAX_shift, SLICE_TILE_MAX_mask);
-    r700->render_target[id].CB_COLOR0_BASE.u32All = 0;
-    SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, ENDIAN_NONE, ENDIAN_shift, ENDIAN_mask);
-    SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, ARRAY_LINEAR_GENERAL,
-             CB_COLOR0_INFO__ARRAY_MODE_shift, CB_COLOR0_INFO__ARRAY_MODE_mask);
-    if(4 == rrb->cpp)
-    {
-        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, COLOR_8_8_8_8,
-                 CB_COLOR0_INFO__FORMAT_shift, CB_COLOR0_INFO__FORMAT_mask);
-        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, SWAP_ALT, COMP_SWAP_shift, COMP_SWAP_mask);
-    }
-    else
-    {
-        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, COLOR_5_6_5,
-                 CB_COLOR0_INFO__FORMAT_shift, CB_COLOR0_INFO__FORMAT_mask);
-        SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, SWAP_ALT_REV,
-                 COMP_SWAP_shift, COMP_SWAP_mask);
-    }
-    SETbit(r700->render_target[id].CB_COLOR0_INFO.u32All, SOURCE_FORMAT_bit);
-    SETbit(r700->render_target[id].CB_COLOR0_INFO.u32All, BLEND_CLAMP_bit);
-    SETfield(r700->render_target[id].CB_COLOR0_INFO.u32All, NUMBER_UNORM, NUMBER_TYPE_shift, NUMBER_TYPE_mask);
-
-    r700->render_target[id].enabled = GL_TRUE;
-}
-
-static void r700SetDepthTarget(context_t *context)
-{
-    R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
-
-    struct radeon_renderbuffer *rrb;
-    unsigned int nPitchInPixel;
-
-    rrb = radeon_get_depthbuffer(&context->radeon);
-    if (!rrb)
-	    return;
-
-    R600_STATECHANGE(context, db_target);
-
-    /* depth buf */
-    r700->DB_DEPTH_SIZE.u32All = 0;
-    r700->DB_DEPTH_BASE.u32All = 0;
-    r700->DB_DEPTH_INFO.u32All = 0;
-    r700->DB_DEPTH_VIEW.u32All = 0;
-
-    nPitchInPixel = rrb->pitch/rrb->cpp;
-
-    SETfield(r700->DB_DEPTH_SIZE.u32All, (nPitchInPixel/8)-1,
-             PITCH_TILE_MAX_shift, PITCH_TILE_MAX_mask);
-    SETfield(r700->DB_DEPTH_SIZE.u32All, ( (nPitchInPixel * context->radeon.radeonScreen->driScreen->fbHeight)/64 )-1,
-             SLICE_TILE_MAX_shift, SLICE_TILE_MAX_mask); /* size in pixel / 64 - 1 */
-
-    if(4 == rrb->cpp)
-    {
-        switch (GL_CONTEXT(context)->Visual.depthBits)
-        {
-        case 16:
-        case 24:
-            SETfield(r700->DB_DEPTH_INFO.u32All, DEPTH_8_24,
-                     DB_DEPTH_INFO__FORMAT_shift, DB_DEPTH_INFO__FORMAT_mask);
-            break;
-        default:
-            fprintf(stderr, "Error: Unsupported depth %d... exiting\n",
-                GL_CONTEXT(context)->Visual.depthBits);
-            _mesa_exit(-1);
-        }
-    }
-    else
-    {
-        SETfield(r700->DB_DEPTH_INFO.u32All, DEPTH_16,
-                     DB_DEPTH_INFO__FORMAT_shift, DB_DEPTH_INFO__FORMAT_mask);
-    }
-    SETfield(r700->DB_DEPTH_INFO.u32All, ARRAY_2D_TILED_THIN1,
-             DB_DEPTH_INFO__ARRAY_MODE_shift, DB_DEPTH_INFO__ARRAY_MODE_mask);
-    /* r700->DB_PREFETCH_LIMIT.bits.DEPTH_HEIGHT_TILE_MAX = (context->currentDraw->h >> 3) - 1; */ /* z buffer sie may much bigger than what need, so use actual used h. */
-}
-
 static void r700InitSQConfig(GLcontext * ctx)
 {
     context_t *context = R700_CONTEXT(ctx);
@@ -1676,6 +1561,7 @@ void r700InitState(GLcontext * ctx) //-------------------
 {
     context_t *context = R700_CONTEXT(ctx);
     R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
+    int id = 0;
 
     radeon_firevertices(&context->radeon);
 
@@ -1869,6 +1755,9 @@ void r700InitState(GLcontext * ctx) //-------------------
     /* Set up color compare mask */
     r700->CB_CLRCMP_MSK.u32All = 0xFFFFFFFF;
 
+    /* screen/window/view */
+    SETfield(r700->CB_TARGET_MASK.u32All, 0xF, (4 * id), TARGET0_ENABLE_mask);
+
     context->radeon.hw.all_dirty = GL_TRUE;
 
 }
diff --git a/src/mesa/drivers/dri/r600/r700_state.h b/src/mesa/drivers/dri/r600/r700_state.h
index 30eb54e8b0..0f53d5b4c5 100644
--- a/src/mesa/drivers/dri/r600/r700_state.h
+++ b/src/mesa/drivers/dri/r600/r700_state.h
@@ -37,13 +37,10 @@ extern void r700UpdateStateParameters(GLcontext * ctx, GLuint new_state);
 extern void r700UpdateShaders (GLcontext * ctx);
 
 extern void r700UpdateViewportOffset(GLcontext * ctx);
-extern void r700UpdateDrawBuffer (GLcontext * ctx);
 
 extern void r700InitState (GLcontext * ctx);
 extern void r700InitStateFuncs (struct dd_function_table *functions);
 
-extern void r700SetDefaultStates(context_t * context);
-
 extern void r700SetScissor(context_t *context);
 
 #endif	/* _R600_SCREEN_H */
diff --git a/src/mesa/drivers/dri/r600/r700_vertprog.c b/src/mesa/drivers/dri/r600/r700_vertprog.c
index 86a67ab766..d107f99e7b 100644
--- a/src/mesa/drivers/dri/r600/r700_vertprog.c
+++ b/src/mesa/drivers/dri/r600/r700_vertprog.c
@@ -38,6 +38,7 @@
 #include "shader/prog_parameter.h"
 #include "shader/prog_statevars.h"
 
+#include "radeon_debug.h"
 #include "r600_context.h"
 #include "r600_cmdbuf.h"
 
@@ -174,7 +175,7 @@ void Map_Vertex_Program(struct r700_vertex_program *vp,
 	// Create VFETCH instructions for inputs
 	if (GL_TRUE != Process_Vertex_Program_Vfetch_Instructions(vp, mesa_vp) ) 
 	{
-		r700_error(ERROR_ASM_VTX_CLAUSE, "Calling Process_Vertex_Program_Vfetch_Instructions return error. \n");
+		radeon_error("Calling Process_Vertex_Program_Vfetch_Instructions return error. \n");
 		return; //error
 	}
 
diff --git a/src/mesa/drivers/dri/r600/radeon_debug.c b/src/mesa/drivers/dri/r600/radeon_debug.c
new file mode 120000
index 0000000000..c98c2e074c
--- /dev/null
+++ b/src/mesa/drivers/dri/r600/radeon_debug.c
@@ -0,0 +1 @@
+../radeon/radeon_debug.c
\ No newline at end of file
diff --git a/src/mesa/drivers/dri/r600/radeon_debug.h b/src/mesa/drivers/dri/r600/radeon_debug.h
new file mode 120000
index 0000000000..bd8aa28e89
--- /dev/null
+++ b/src/mesa/drivers/dri/r600/radeon_debug.h
@@ -0,0 +1 @@
+../radeon/radeon_debug.h
\ No newline at end of file