intel: Bump driver date

docs: more bug fixes for Mesa 7.4.1
glx: added null pointer check in glXGetFBConfigs()
2009-04-15 15:12:48 -07:00 · 2009-04-15 15:10:33 -07:00 · 2009-04-15 08:02:38 -06:00 · 2009-04-11 20:24:15 -07:00 · 2009-04-09 15:15:33 -07:00 · 2009-04-09 14:44:53 -07:00
34 changed files with 585 additions and 194 deletions
--- a/2
+++ b/2
@@ -174,7 +174,7 @@ ultrix-gcc:

 # Rules for making release tarballs

-VERSION=7.4-rc2
+VERSION=7.4
 DIRECTORY = Mesa-$(VERSION)
 LIB_NAME = MesaLib-$(VERSION)
 DEMO_NAME = MesaDemos-$(VERSION)
--- a/configs/darwin
+++ b/configs/darwin
@@ -6,6 +6,8 @@ CONFIG_NAME = darwin

 INSTALL_DIR = /usr/X11

+X11_DIR = $(INSTALL_DIR)
+
 # Compiler and flags
 CC = gcc
 CXX = gcc
@@ -21,9 +23,9 @@ DEFINES =  -D_DARWIN_C_SOURCE -D_POSIX_SOURCE -D_POSIX_C_SOURCE=199309L \
 ARCH_FLAGS += $(RC_CFLAGS)

 CFLAGS =  -ggdb3 -Os -Wall -Wmissing-prototypes -std=c99 -ffast-math -fno-strict-aliasing \
-	-I$(INSTALL_DIR)/include $(OPT_FLAGS) $(PIC_FLAGS) $(ARCH_FLAGS) $(ASM_FLAGS) $(DEFINES)
+	-I$(INSTALL_DIR)/include -I$(X11_DIR)/include $(OPT_FLAGS) $(PIC_FLAGS) $(ARCH_FLAGS) $(ASM_FLAGS) $(DEFINES)
 CXXFLAGS =  -ggdb3 -Os -Wall -fno-strict-aliasing \
-	-I$(INSTALL_DIR)/include $(OPT_FLAGS) $(PIC_FLAGS) $(ARCH_FLAGS) $(ASM_FLAGS) $(DEFINES)
+	-I$(INSTALL_DIR)/include -I$(X11_DIR)/include $(OPT_FLAGS) $(PIC_FLAGS) $(ARCH_FLAGS) $(ASM_FLAGS) $(DEFINES)

 # Library names (actual file names)
 GL_LIB_NAME = libGL.dylib
@@ -39,12 +41,12 @@ GLUT_LIB_GLOB = libglut.*dylib
 GLW_LIB_GLOB = libGLw.*dylib
 OSMESA_LIB_GLOB = libOSMesa.*dylib

-GL_LIB_DEPS = -L$(INSTALL_DIR)/$(LIB_DIR) -lX11 -lXext -lm -lpthread
+GL_LIB_DEPS = -L$(INSTALL_DIR)/$(LIB_DIR) -L$(X11_DIR)/$(LIB_DIR) -lX11 -lXext -lm -lpthread
 OSMESA_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GL_LIB)
 GLU_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GL_LIB)
-GLUT_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GLU_LIB) -l$(GL_LIB)  -L$(INSTALL_DIR)/$(LIB_DIR) -lX11 -lXmu -lXi -lXext
-GLW_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GL_LIB) -L$(INSTALL_DIR)/$(LIB_DIR) -lX11 -lXt
-APP_LIB_DEPS = -L$(TOP)/lib -l$(GLUT_LIB) -l$(GLU_LIB) -l$(GL_LIB) -L$(INSTALL_DIR)/$(LIB_DIR) -lX11 -lXmu -lXt -lXi -lm
+GLUT_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GLU_LIB) -l$(GL_LIB)  -L$(INSTALL_DIR)/$(LIB_DIR) -L$(X11_DIR)/$(LIB_DIR) -lX11 -lXmu -lXi -lXext
+GLW_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GL_LIB) -L$(INSTALL_DIR)/$(LIB_DIR) -L$(X11_DIR)/$(LIB_DIR) -lX11 -lXt
+APP_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLU_LIB) -l$(GL_LIB) -L$(INSTALL_DIR)/$(LIB_DIR) -L$(X11_DIR)/$(LIB_DIR) -lX11 -lXmu -lXt -lXi -lm

 # omit glw lib for now:
 SRC_DIRS = glx/x11 mesa glu glut/glx
--- a/docs/news.html
+++ b/docs/news.html
@@ -11,7 +11,7 @@
 <H1>News</H1>


-<h2>March ??, 2009</h2>
+<h2>March 27, 2009</h2>
 <p>
 <a href="relnotes-7.4.html">Mesa 7.4</a> is released.
 This is a stable release fixing bugs since the 7.3 release.
--- a/docs/relnotes-7.4.1.html
+++ b/docs/relnotes-7.4.1.html
@@ -0,0 +1,79 @@
+<HTML>
+
+<TITLE>Mesa Release Notes</TITLE>
+
+<head><link rel="stylesheet" type="text/css" href="mesa.css"></head>
+
+<BODY>
+
+<body bgcolor="#eeeeee">
+
+<H1>Mesa 7.4.1 Release Notes / (date TBD)</H1>
+
+<p>
+Mesa 7.4.1 is a stable development release fixing bugs since the 7.4 release.
+</p>
+<p>
+Mesa 7.4.1 implements the OpenGL 2.1 API, but the version reported by
+glGetString(GL_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 2.1.
+</p>
+<p>
+See the <a href="install.html">Compiling/Installing page</a> for prerequisites
+for DRI ardware acceleration.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+tbd
+</pre>
+
+
+<h2>New features</h2>
+<ul>
+</ul>
+
+
+<h2>Bug fixes</h2>
+<ul>
+<li>Fixed a two-sided lighting bug in fixed-function-to-GPU code generation
+<li>Fixed some Darwin issues (Jeremy Huddleston)
+<li>Indexing the GLSL gl_EyePlane[] or gl_ObjectPlane[] arrays with a variable
+    was broken, bug 20986
+<li>Fixed incorrect texture unit bias in TXB instruction
+<li>glTexParameter settings weren't always propogated to drivers
+<li>Assorted vertex/fragment program bug fixes
+<li>Fixed point rendering in software rasterizer
+<li>Fixed potential deadlock in object hash functions
+<li>Fix a couple bugs surrounding front-buffer rendering with DRI2, but this
+    is not quite complete.
+</ul>
+
+<h2>Changes</h2>
+<ul>
+</ul>
+
+
+
+<h2>Driver Status</h2>
+
+<pre>
+Driver			Status
+----------------------	----------------------
+DRI drivers		varies with the driver
+XMesa/GLX (on Xlib)	implements OpenGL 2.1
+OSMesa (off-screen)	implements OpenGL 2.1
+Windows/Win32		implements OpenGL 2.1
+Glide (3dfx Voodoo1/2)	implements OpenGL 1.3
+SVGA			unsupported
+Wind River UGL		unsupported
+DJGPP			unsupported
+GGI			unsupported
+BeOS			unsupported
+Allegro			unsupported
+D3D			unsupported
+</pre>
+
+</body>
+</html>
--- a/docs/relnotes-7.4.html
+++ b/docs/relnotes-7.4.html
@@ -8,7 +8,7 @@

 <body bgcolor="#eeeeee">

-<H1>Mesa 7.4 Release Notes / date TBD</H1>
+<H1>Mesa 7.4 Release Notes / 27 March 2009</H1>

 <p>
 Mesa 7.4 is a stable development release fixing bugs since the 7.3 release.
@@ -26,7 +26,15 @@ for DRI ardware acceleration.

 <h2>MD5 checksums</h2>
 <pre>
-tbd
+ed6bd7437177307e51e16d0c7c381dfa  MesaLib-7.4.tar.gz
+7ecddb341a2691e0dfdb02f697109834  MesaLib-7.4.tar.bz2
+433e823f8245f9fd5f397e7b719a8e47  MesaLib-7.4.zip
+656eee6128016fb237e01aa8dabbc703  MesaDemos-7.4.tar.gz
+02816f10f30b1dc5e069e0f68c177c98  MesaDemos-7.4.tar.bz2
+44a70d6db4aa4c64ecc47871b6aceee8  MesaDemos-7.4.zip
+25f80db4f8645cd3e58e2c9af53ec341  MesaGLUT-7.4.tar.gz
+04ec01caebde44f5b0d619f00716b368  MesaGLUT-7.4.tar.bz2
+019dc213baecaa3cb1278847d41b8591  MesaGLUT-7.4.zip
 </pre>


--- a/docs/relnotes.html
+++ b/docs/relnotes.html
@@ -20,6 +20,7 @@ The release notes summarize what's new or changed in each Mesa release.
 </p>

 <UL>
+<LI><A HREF="relnotes-7.4.1.html">7.4.1 release notes</A>
 <LI><A HREF="relnotes-7.4.html">7.4 release notes</A>
 <LI><A HREF="relnotes-7.3.html">7.3 release notes</A>
 <LI><A HREF="relnotes-7.2.html">7.2 release notes</A>
--- a/include/GL/internal/dri_interface.h
+++ b/include/GL/internal/dri_interface.h
@@ -636,7 +636,7 @@ struct __DRIbufferRec {
 };

 #define __DRI_DRI2_LOADER "DRI_DRI2Loader"
-#define __DRI_DRI2_LOADER_VERSION 1
+#define __DRI_DRI2_LOADER_VERSION 2
 struct __DRIdri2LoaderExtensionRec {
    __DRIextension base;

@@ -644,6 +644,19 @@ struct __DRIdri2LoaderExtensionRec {
 			       int *width, int *height,
 			       unsigned int *attachments, int count,
 			       int *out_count, void *loaderPrivate);
+
+    /**
+     * Flush pending front-buffer rendering
+     *
+     * Any rendering that has been performed to the
+     * \c __DRI_BUFFER_FAKE_FRONT_LEFT will be flushed to the
+     * \c __DRI_BUFFER_FRONT_LEFT.
+     *
+     * \param driDrawable    Drawable whose front-buffer is to be flushed
+     * \param loaderPrivate  Loader's private data that was previously passed
+     *                       into __DRIdri2ExtensionRec::createNewDrawable
+     */
+    void (*flushFrontBuffer)(__DRIdrawable *driDrawable, void *loaderPrivate);
 };

 /**
--- a/src/glx/x11/dri2.c
+++ b/src/glx/x11/dri2.c
@@ -253,7 +253,7 @@ DRI2Buffer *DRI2GetBuffers(Display *dpy, XID drawable,
    *height = rep.height;
    *outCount = rep.count;

-    buffers = Xmalloc(count * sizeof buffers[0]);
+    buffers = Xmalloc(rep.count * sizeof buffers[0]);
    if (buffers == NULL) {
 	_XEatData(dpy, rep.count * sizeof repBuffer);
 	UnlockDisplay(dpy);
--- a/src/glx/x11/dri2_glx.c
+++ b/src/glx/x11/dri2_glx.c
@@ -74,7 +74,6 @@ struct __GLXDRIdrawablePrivateRec {
    int bufferCount;
    int width, height;
    int have_back;
-    int have_front;
    int have_fake_front;
 };

@@ -195,7 +194,7 @@ static void dri2CopySubBuffer(__GLXDRIdrawable *pdraw,
    XserverRegion region;

    /* Check we have the right attachments */
-    if (!(priv->have_front && priv->have_back))
+    if (!priv->have_back)
    	return;

    xrect.x = x;
@@ -223,7 +222,7 @@ static void dri2WaitX(__GLXDRIdrawable *pdraw)
    XserverRegion region;

    /* Check we have the right attachments */
-    if (!(priv->have_fake_front && priv->have_front))
+    if (!priv->have_fake_front)
    	return;

    xrect.x = 0;
@@ -243,7 +242,7 @@ static void dri2WaitGL(__GLXDRIdrawable *pdraw)
    XRectangle xrect;
    XserverRegion region;

-    if (!(priv->have_fake_front && priv->have_front))
+    if (!priv->have_fake_front)
    	return;

    xrect.x = 0;
@@ -257,6 +256,15 @@ static void dri2WaitGL(__GLXDRIdrawable *pdraw)
    XFixesDestroyRegion(pdraw->psc->dpy, region);
 }

+
+static void dri2FlushFrontBuffer(__DRIdrawable *driDrawable,
+				 void *loaderPrivate)
+{
+    (void) driDrawable;
+    dri2WaitGL((__GLXDRIdrawable *) loaderPrivate);
+}
+
+
 static void dri2DestroyScreen(__GLXscreenConfigs *psc)
 {
    /* Free the direct rendering per screen data */
@@ -283,7 +291,6 @@ dri2GetBuffers(__DRIdrawable *driDrawable,
    pdraw->width = *width;
    pdraw->height = *height;
    pdraw->bufferCount = *out_count;
-    pdraw->have_front = 0;
    pdraw->have_fake_front = 0;
    pdraw->have_back = 0;

@@ -295,8 +302,6 @@ dri2GetBuffers(__DRIdrawable *driDrawable,
 	pdraw->buffers[i].pitch = buffers[i].pitch;
 	pdraw->buffers[i].cpp = buffers[i].cpp;
 	pdraw->buffers[i].flags = buffers[i].flags;
-	if (pdraw->buffers[i].attachment == __DRI_BUFFER_FRONT_LEFT)
-	    pdraw->have_front = 1;
 	if (pdraw->buffers[i].attachment == __DRI_BUFFER_FAKE_FRONT_LEFT)
 	    pdraw->have_fake_front = 1;
 	if (pdraw->buffers[i].attachment == __DRI_BUFFER_BACK_LEFT)
@@ -311,6 +316,7 @@ dri2GetBuffers(__DRIdrawable *driDrawable,
 static const __DRIdri2LoaderExtension dri2LoaderExtension = {
    { __DRI_DRI2_LOADER, __DRI_DRI2_LOADER_VERSION },
    dri2GetBuffers,
+    dri2FlushFrontBuffer
 };

 static const __DRIextension *loader_extensions[] = {
--- a/src/glx/x11/glxcmds.c
+++ b/src/glx/x11/glxcmds.c
@@ -1702,7 +1702,8 @@ PUBLIC GLXFBConfig *glXGetFBConfigs(Display *dpy, int screen, int *nelements)
    int   i;

    *nelements = 0;
-    if ( (priv->screenConfigs != NULL)
+    if ( priv
+         && (priv->screenConfigs != NULL)
 	 && (screen >= 0) && (screen <= ScreenCount(dpy))
 	 && (priv->screenConfigs[screen].configs != NULL)
 	 && (priv->screenConfigs[screen].configs->fbconfigID != GLX_DONT_CARE) ) {
--- a/src/glx/x11/indirect.c
+++ b/src/glx/x11/indirect.c
@@ -5198,9 +5198,13 @@ glDeleteTexturesEXT(GLsizei n, const GLuint * textures)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING
    if (gc->driContext) {
        CALL_DeleteTextures(GET_DISPATCH(), (n, textures));
    } else {
+#else
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 4 + __GLX_PAD((n * 4));
@@ -5266,9 +5270,13 @@ glGenTexturesEXT(GLsizei n, GLuint * textures)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GenTextures(GET_DISPATCH(), (n, textures));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 4;
@@ -5328,9 +5336,13 @@ glIsTextureEXT(GLuint texture)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        return CALL_IsTexture(GET_DISPATCH(), (texture));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        GLboolean retval = (GLboolean) 0;
@@ -5641,9 +5653,13 @@ glGetColorTableEXT(GLenum target, GLenum format, GLenum type, GLvoid * table)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetColorTable(GET_DISPATCH(), (target, format, type, table));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        const __GLXattribute *const state = gc->client_state_private;
        Display *const dpy = gc->currentDpy;
@@ -5714,10 +5730,14 @@ glGetColorTableParameterfvEXT(GLenum target, GLenum pname, GLfloat * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetColorTableParameterfv(GET_DISPATCH(),
                                      (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -5784,10 +5804,14 @@ glGetColorTableParameterivEXT(GLenum target, GLenum pname, GLint * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetColorTableParameteriv(GET_DISPATCH(),
                                      (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -6107,10 +6131,14 @@ gl_dispatch_stub_356(GLenum target, GLenum format, GLenum type,
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetConvolutionFilter(GET_DISPATCH(),
                                  (target, format, type, image));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        const __GLXattribute *const state = gc->client_state_private;
        Display *const dpy = gc->currentDpy;
@@ -6182,10 +6210,14 @@ gl_dispatch_stub_357(GLenum target, GLenum pname, GLfloat * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetConvolutionParameterfv(GET_DISPATCH(),
                                       (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -6252,10 +6284,14 @@ gl_dispatch_stub_358(GLenum target, GLenum pname, GLint * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetConvolutionParameteriv(GET_DISPATCH(),
                                       (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -6329,10 +6365,14 @@ gl_dispatch_stub_361(GLenum target, GLboolean reset, GLenum format,
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetHistogram(GET_DISPATCH(),
                          (target, reset, format, type, values));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        const __GLXattribute *const state = gc->client_state_private;
        Display *const dpy = gc->currentDpy;
@@ -6403,9 +6443,13 @@ gl_dispatch_stub_362(GLenum target, GLenum pname, GLfloat * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetHistogramParameterfv(GET_DISPATCH(), (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -6471,9 +6515,13 @@ gl_dispatch_stub_363(GLenum target, GLenum pname, GLint * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetHistogramParameteriv(GET_DISPATCH(), (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -6543,9 +6591,13 @@ gl_dispatch_stub_364(GLenum target, GLboolean reset, GLenum format,
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetMinmax(GET_DISPATCH(), (target, reset, format, type, values));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        const __GLXattribute *const state = gc->client_state_private;
        Display *const dpy = gc->currentDpy;
@@ -6614,9 +6666,13 @@ gl_dispatch_stub_365(GLenum target, GLenum pname, GLfloat * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetMinmaxParameterfv(GET_DISPATCH(), (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
@@ -6679,9 +6735,13 @@ gl_dispatch_stub_366(GLenum target, GLenum pname, GLint * params)
 {
    __GLXcontext *const gc = __glXGetCurrentContext();

+#ifdef GLX_DIRECT_RENDERING    
    if (gc->driContext) {
        CALL_GetMinmaxParameteriv(GET_DISPATCH(), (target, pname, params));
    } else {
+#else 
+    {
+#endif
        __GLXcontext *const gc = __glXGetCurrentContext();
        Display *const dpy = gc->currentDpy;
        const GLuint cmdlen = 8;
--- a/src/mesa/drivers/dri/intel/intel_buffers.c
+++ b/src/mesa/drivers/dri/intel/intel_buffers.c
@@ -204,6 +204,8 @@ intel_draw_buffer(GLcontext * ctx, struct gl_framebuffer *fb)
 	       intel_batchbuffer_flush(intel->batch);
 	    intel->front_cliprects = GL_TRUE;
 	    colorRegions[0] = intel_get_rb_region(fb, BUFFER_FRONT_LEFT);
+
+	    intel->front_buffer_dirty = GL_TRUE;
 	 }
 	 else {
 	    if (!intel->constant_cliprect && intel->front_cliprects)
@@ -328,6 +330,12 @@ intel_draw_buffer(GLcontext * ctx, struct gl_framebuffer *fb)
 static void
 intelDrawBuffer(GLcontext * ctx, GLenum mode)
 {
+   if ((ctx->DrawBuffer != NULL) && (ctx->DrawBuffer->Name == 0)) {
+      struct intel_context *const intel = intel_context(ctx);
+
+      intel->is_front_buffer_rendering = (mode == GL_FRONT_LEFT);
+   }
+
   intel_draw_buffer(ctx, ctx->DrawBuffer);
 }

--- a/src/mesa/drivers/dri/intel/intel_context.c
+++ b/src/mesa/drivers/dri/intel/intel_context.c
@@ -97,7 +97,7 @@ int INTEL_DEBUG = (0);

 #include "extension_helper.h"

-#define DRIVER_DATE                     "20090326 2009Q1 RC2"
+#define DRIVER_DATE                     "20090415 2009Q1 RC3"
 #define DRIVER_DATE_GEM                 "GEM " DRIVER_DATE

 static const GLubyte *
@@ -265,6 +265,11 @@ intel_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
 	   region_name = "dri2 front buffer";
 	   break;

+       case __DRI_BUFFER_FAKE_FRONT_LEFT:
+	   rb = intel_fb->color_rb[0];
+	   region_name = "dri2 fake front buffer";
+	   break;
+
       case __DRI_BUFFER_BACK_LEFT:
 	   rb = intel_fb->color_rb[1];
 	   region_name = "dri2 back buffer";
@@ -525,6 +530,26 @@ intel_flush(GLcontext *ctx, GLboolean needs_mi_flush)

   if (intel->batch->map != intel->batch->ptr)
      intel_batchbuffer_flush(intel->batch);
+
+   if ((ctx->DrawBuffer->Name == 0) && intel->front_buffer_dirty) {
+      __DRIscreen *const screen = intel->intelScreen->driScrnPriv;
+
+      if ((screen->dri2.loader->base.version >= 2)
+	  && (screen->dri2.loader->flushFrontBuffer != NULL)) {
+	 (*screen->dri2.loader->flushFrontBuffer)(intel->driDrawable,
+						  intel->driDrawable->loaderPrivate);
+
+	 /* Only clear the dirty bit if front-buffer rendering is no longer
+	  * enabled.  This is done so that the dirty bit can only be set in
+	  * glDrawBuffer.  Otherwise the dirty bit would have to be set at
+	  * each of N places that do rendering.  This has worse performances,
+	  * but it is much easier to get correct.
+	  */
+	 if (intel->is_front_buffer_rendering) {
+	    intel->front_buffer_dirty = GL_FALSE;
+	 }
+      }
+   }
 }

 void
@@ -533,7 +558,7 @@ intelFlush(GLcontext * ctx)
   intel_flush(ctx, GL_FALSE);
 }

-static void
+void
 intel_glFlush(GLcontext *ctx)
 {
   intel_flush(ctx, GL_TRUE);
--- a/src/mesa/drivers/dri/intel/intel_context.h
+++ b/src/mesa/drivers/dri/intel/intel_context.h
@@ -260,11 +260,29 @@ struct intel_context
    * flush time while the lock is held.
    */
   GLboolean constant_cliprect;
+
   /**
    * In !constant_cliprect mode, set to true if the front cliprects should be
    * used instead of back.
    */
   GLboolean front_cliprects;
+
+   /**
+    * Set if rendering has occured to the drawable's front buffer.
+    *
+    * This is used in the DRI2 case to detect that glFlush should also copy
+    * the contents of the fake front buffer to the real front buffer.
+    */
+   GLboolean front_buffer_dirty;
+
+   /**
+    * Track whether front-buffer rendering is currently enabled
+    *
+    * A separate flag is used to track this in order to support MRT more
+    * easily.
+    */
+   GLboolean is_front_buffer_rendering;
+
   drm_clip_rect_t fboRect;     /**< cliprect for FBO rendering */

   int perf_boxes;
@@ -440,6 +458,7 @@ extern void intelGetLock(struct intel_context *intel, GLuint flags);

 extern void intelFinish(GLcontext * ctx);
 extern void intelFlush(GLcontext * ctx);
+extern void intel_glFlush(GLcontext *ctx);

 extern void intelInitDriverFunctions(struct dd_function_table *functions);
 extern void intelInitExtensions(GLcontext *ctx, GLboolean enable_imaging);
--- a/src/mesa/drivers/dri/intel/intel_tex_copy.c
+++ b/src/mesa/drivers/dri/intel/intel_tex_copy.c
@@ -104,7 +104,7 @@ do_copy_texsubimage(struct intel_context *intel,
      return GL_FALSE;
   }

-   intelFlush(ctx);
+   intel_glFlush(ctx);
   LOCK_HARDWARE(intel);
   {
      GLuint image_offset = intel_miptree_image_offset(intelImage->mt,
@@ -155,6 +155,7 @@ do_copy_texsubimage(struct intel_context *intel,
   }

   UNLOCK_HARDWARE(intel);
+   intel_glFlush(ctx);

   /* GL_SGIS_generate_mipmap */
   if (intelImage->level == texObj->BaseLevel && texObj->GenerateMipmap) {
--- a/src/mesa/drivers/dri/radeon/radeon_screen.c
+++ b/src/mesa/drivers/dri/radeon/radeon_screen.c
@@ -259,8 +259,6 @@ radeonFillInModes( __DRIscreenPrivate *psp,
    __GLcontextModes *m;
    unsigned depth_buffer_factor;
    unsigned back_buffer_factor;
-    GLenum fb_format;
-    GLenum fb_type;
    int i;

    /* Right now GLX_SWAP_COPY_OML isn't supported, but it would be easy
@@ -289,19 +287,24 @@ radeonFillInModes( __DRIscreenPrivate *psp,
    depth_buffer_factor = ((depth_bits != 0) || (stencil_bits != 0)) ? 2 : 1;
    back_buffer_factor  = (have_back_buffer) ? 2 : 1;

-    if ( pixel_bits == 16 ) {
-        fb_format = GL_RGB;
-        fb_type = GL_UNSIGNED_SHORT_5_6_5;
-    }
-    else {
-        fb_format = GL_BGRA;
-        fb_type = GL_UNSIGNED_INT_8_8_8_8_REV;
-    }
+    if (pixel_bits == 16) {
+	__DRIconfig **configs_a8r8g8b8;
+	__DRIconfig **configs_r5g6b5;
+
+	configs_r5g6b5 = driCreateConfigs(GL_RGB, GL_UNSIGNED_SHORT_5_6_5,
+					  depth_bits_array, stencil_bits_array,
+					  depth_buffer_factor, back_buffer_modes,
+					  back_buffer_factor);
+	configs_a8r8g8b8 = driCreateConfigs(GL_BGRA, GL_UNSIGNED_INT_8_8_8_8_REV,
+					    depth_bits_array, stencil_bits_array,
+					    1, back_buffer_modes, 1);
+	configs = driConcatConfigs(configs_r5g6b5, configs_a8r8g8b8);
+   } else
+	configs = driCreateConfigs(GL_BGRA, GL_UNSIGNED_INT_8_8_8_8_REV,
+				   depth_bits_array, stencil_bits_array,
+				   depth_buffer_factor,
+				   back_buffer_modes, back_buffer_factor);

-    configs = driCreateConfigs(fb_format, fb_type,
-			       depth_bits_array, stencil_bits_array,
-			       depth_buffer_factor,
-			       back_buffer_modes, back_buffer_factor);
    if (configs == NULL) {
 	fprintf( stderr, "[%s:%u] Error creating FBConfig!\n",
 		 __func__, __LINE__ );
--- a/src/mesa/drivers/dri/swrast/swrast.c
+++ b/src/mesa/drivers/dri/swrast/swrast.c
@@ -166,17 +166,27 @@ swrastFillInModes(__DRIscreen *psp,
    depth_buffer_factor = 4;
    back_buffer_factor = 2;

-    if (pixel_bits == 8) {
+    switch (pixel_bits) {
+    case 8:
 	fb_format = GL_RGB;
 	fb_type = GL_UNSIGNED_BYTE_2_3_3_REV;
-    }
-    else if (pixel_bits == 16) {
+	break;
+    case 16:
 	fb_format = GL_RGB;
 	fb_type = GL_UNSIGNED_SHORT_5_6_5;
-    }
-    else {
+	break;
+    case 24:
+	fb_format = GL_BGR;
+	fb_type = GL_UNSIGNED_INT_8_8_8_8_REV;
+	break;
+    case 32:
 	fb_format = GL_BGRA;
 	fb_type = GL_UNSIGNED_INT_8_8_8_8_REV;
+	break;
+    default:
+	fprintf(stderr, "[%s:%u] bad depth %d\n", __func__, __LINE__,
+		pixel_bits);
+	return NULL;
    }

    configs = driCreateConfigs(fb_format, fb_type,
@@ -198,7 +208,7 @@ driCreateNewScreen(int scrn, const __DRIextension **extensions,
 {
    static const __DRIextension *emptyExtensionList[] = { NULL };
    __DRIscreen *psp;
-    __DRIconfig **configs8, **configs16, **configs32;
+    const __DRIconfig **configs8, **configs16, **configs24, **configs32;

    (void) data;

@@ -215,11 +225,12 @@ driCreateNewScreen(int scrn, const __DRIextension **extensions,

    configs8  = swrastFillInModes(psp,  8,  8, 0, 1);
    configs16 = swrastFillInModes(psp, 16, 16, 0, 1);
+    configs24 = swrastFillInModes(psp, 24, 24, 8, 1);
    configs32 = swrastFillInModes(psp, 32, 24, 8, 1);

-    configs16 = (__DRIconfig **)driConcatConfigs(configs8, configs16);
-
-    *driver_configs = driConcatConfigs(configs16, configs32);
+    configs16 = driConcatConfigs(configs8, configs16);
+    configs24 = driConcatConfigs(configs16, configs24);
+    *driver_configs = driConcatConfigs(configs24, configs32);

    driInitExtensions( NULL, card_extensions, GL_FALSE );

@@ -251,19 +262,24 @@ static GLuint
 choose_pixel_format(const GLvisual *v)
 {
    if (v->rgbMode) {
-	int bpp = v->rgbBits;
+	int depth = v->rgbBits;

-	if (bpp == 32
+	if (depth == 32
 	    && v->redMask   == 0xff0000
 	    && v->greenMask == 0x00ff00
 	    && v->blueMask  == 0x0000ff)
 	    return PF_A8R8G8B8;
-	else if (bpp == 16
+	else if (depth == 24
+	    && v->redMask   == 0xff0000
+	    && v->greenMask == 0x00ff00
+	    && v->blueMask  == 0x0000ff)
+	    return PF_X8R8G8B8;
+	else if (depth == 16
 	    && v->redMask   == 0xf800
 	    && v->greenMask == 0x07e0
 	    && v->blueMask  == 0x001f)
 	    return PF_R5G6B5;
-	else if (bpp == 8
+	else if (depth == 8
 	    && v->redMask   == 0x07
 	    && v->greenMask == 0x38
 	    && v->blueMask  == 0xc0)
@@ -292,7 +308,6 @@ swrast_alloc_front_storage(GLcontext *ctx, struct gl_renderbuffer *rb,
 			   GLenum internalFormat, GLuint width, GLuint height)
 {
    struct swrast_renderbuffer *xrb = swrast_renderbuffer(rb);
-    int bpp;
    unsigned mask = PITCH_ALIGN_BITS - 1;

    TRACE;
@@ -301,23 +316,8 @@ swrast_alloc_front_storage(GLcontext *ctx, struct gl_renderbuffer *rb,
    rb->Width = width;
    rb->Height = height;

-    switch (internalFormat) {
-    case GL_RGB:
-	bpp = rb->RedBits + rb->GreenBits + rb->BlueBits;
-	break;
-    case GL_RGBA:
-	bpp = rb->RedBits + rb->GreenBits + rb->BlueBits + rb->AlphaBits;
-	break;
-    case GL_COLOR_INDEX8_EXT:
-	bpp = rb->IndexBits;
-	break;
-    default:
-	_mesa_problem( NULL, "unexpected format in %s", __FUNCTION__ );
-	return GL_FALSE;
-    }
-
    /* always pad to PITCH_ALIGN_BITS */
-    xrb->pitch = ((width * bpp + mask) & ~mask) / 8;
+    xrb->pitch = ((width * xrb->bpp + mask) & ~mask) / 8;

    return GL_TRUE;
 }
@@ -373,6 +373,17 @@ swrast_new_renderbuffer(const GLvisual *visual, GLboolean front)
 	xrb->Base.GreenBits = 8 * sizeof(GLubyte);
 	xrb->Base.BlueBits  = 8 * sizeof(GLubyte);
 	xrb->Base.AlphaBits = 8 * sizeof(GLubyte);
+	xrb->bpp = 32;
+	break;
+    case PF_X8R8G8B8:
+	xrb->Base.InternalFormat = GL_RGB;
+	xrb->Base._BaseFormat = GL_RGB;
+	xrb->Base.DataType = GL_UNSIGNED_BYTE;
+	xrb->Base.RedBits   = 8 * sizeof(GLubyte);
+	xrb->Base.GreenBits = 8 * sizeof(GLubyte);
+	xrb->Base.BlueBits  = 8 * sizeof(GLubyte);
+	xrb->Base.AlphaBits = 0;
+	xrb->bpp = 32;
 	break;
    case PF_R5G6B5:
 	xrb->Base.InternalFormat = GL_RGB;
@@ -382,6 +393,7 @@ swrast_new_renderbuffer(const GLvisual *visual, GLboolean front)
 	xrb->Base.GreenBits = 6 * sizeof(GLubyte);
 	xrb->Base.BlueBits  = 5 * sizeof(GLubyte);
 	xrb->Base.AlphaBits = 0;
+	xrb->bpp = 16;
 	break;
    case PF_R3G3B2:
 	xrb->Base.InternalFormat = GL_RGB;
@@ -391,12 +403,14 @@ swrast_new_renderbuffer(const GLvisual *visual, GLboolean front)
 	xrb->Base.GreenBits = 3 * sizeof(GLubyte);
 	xrb->Base.BlueBits  = 2 * sizeof(GLubyte);
 	xrb->Base.AlphaBits = 0;
+	xrb->bpp = 8;
 	break;
    case PF_CI8:
 	xrb->Base.InternalFormat = GL_COLOR_INDEX8_EXT;
 	xrb->Base._BaseFormat = GL_COLOR_INDEX;
 	xrb->Base.DataType = GL_UNSIGNED_BYTE;
 	xrb->Base.IndexBits = 8 * sizeof(GLubyte);
+	xrb->bpp = 8;
 	break;
    default:
 	return NULL;
--- a/src/mesa/drivers/dri/swrast/swrast_priv.h
+++ b/src/mesa/drivers/dri/swrast/swrast_priv.h
@@ -90,6 +90,8 @@ struct swrast_renderbuffer {

    /* renderbuffer pitch (in bytes) */
    GLuint pitch;
+   /* bits per pixel of storage */
+    GLuint bpp;
 };

 static INLINE __DRIcontext *
@@ -115,10 +117,10 @@ swrast_renderbuffer(struct gl_renderbuffer *rb)
 * Pixel formats we support
 */
 #define PF_CI8        1		/**< Color Index mode */
-#define PF_A8R8G8B8   2		/**< 32-bit TrueColor:  8-A, 8-R, 8-G, 8-B bits */
-#define PF_R5G6B5     3		/**< 16-bit TrueColor:  5-R, 6-G, 5-B bits */
-#define PF_R3G3B2     4		/**<  8-bit TrueColor:  3-R, 3-G, 2-B bits */
-
+#define PF_A8R8G8B8   2		/**< 32bpp TrueColor:  8-A, 8-R, 8-G, 8-B bits */
+#define PF_R5G6B5     3		/**< 16bpp TrueColor:  5-R, 6-G, 5-B bits */
+#define PF_R3G3B2     4		/**<  8bpp TrueColor:  3-R, 3-G, 2-B bits */
+#define PF_X8R8G8B8   5		/**< 32bpp TrueColor:  8-R, 8-G, 8-B bits */

 /**
 * Renderbuffer pitch alignment (in bits).
--- a/src/mesa/drivers/dri/swrast/swrast_span.c
+++ b/src/mesa/drivers/dri/swrast/swrast_span.c
@@ -79,6 +79,24 @@ static const GLubyte kernel[16] = {
   DST[BCOMP] = SRC[0]


+/* 32-bit BGRX */
+#define STORE_PIXEL_X8R8G8B8(DST, X, Y, VALUE) \
+   DST[3] = 0xff; \
+   DST[2] = VALUE[RCOMP]; \
+   DST[1] = VALUE[GCOMP]; \
+   DST[0] = VALUE[BCOMP]
+#define STORE_PIXEL_RGB_X8R8G8B8(DST, X, Y, VALUE) \
+   DST[3] = 0xff; \
+   DST[2] = VALUE[RCOMP]; \
+   DST[1] = VALUE[GCOMP]; \
+   DST[0] = VALUE[BCOMP]
+#define FETCH_PIXEL_X8R8G8B8(DST, SRC) \
+   DST[ACOMP] = 0xff; \
+   DST[RCOMP] = SRC[2]; \
+   DST[GCOMP] = SRC[1]; \
+   DST[BCOMP] = SRC[0]
+
+
 /* 16-bit BGR */
 #define STORE_PIXEL_R5G6B5(DST, X, Y, VALUE) \
   do { \
@@ -139,6 +157,24 @@ static const GLubyte kernel[16] = {
 #include "swrast/s_spantemp.h"


+/* 32-bit BGRX */
+#define NAME(FUNC) FUNC##_X8R8G8B8
+#define RB_TYPE GLubyte
+#define SPAN_VARS \
+   struct swrast_renderbuffer *xrb = swrast_renderbuffer(rb);
+#define INIT_PIXEL_PTR(P, X, Y) \
+   GLubyte *P = (GLubyte *)xrb->Base.Data + YFLIP(xrb, Y) * xrb->pitch + (X) * 4;
+#define INC_PIXEL_PTR(P) P += 4
+#define STORE_PIXEL(DST, X, Y, VALUE) \
+   STORE_PIXEL_X8R8G8B8(DST, X, Y, VALUE)
+#define STORE_PIXEL_RGB(DST, X, Y, VALUE) \
+   STORE_PIXEL_RGB_X8R8G8B8(DST, X, Y, VALUE)
+#define FETCH_PIXEL(DST, SRC) \
+   FETCH_PIXEL_X8R8G8B8(DST, SRC)
+
+#include "swrast/s_spantemp.h"
+
+
 /* 16-bit BGR */
 #define NAME(FUNC) FUNC##_R5G6B5
 #define RB_TYPE GLubyte
@@ -210,6 +246,24 @@ static const GLubyte kernel[16] = {
 #include "swrast_spantemp.h"


+/* 32-bit BGRX */
+#define NAME(FUNC) FUNC##_X8R8G8B8_front
+#define RB_TYPE GLubyte
+#define SPAN_VARS \
+   struct swrast_renderbuffer *xrb = swrast_renderbuffer(rb);
+#define INIT_PIXEL_PTR(P, X, Y) \
+   GLubyte *P = (GLubyte *)row;
+#define INC_PIXEL_PTR(P) P += 4
+#define STORE_PIXEL(DST, X, Y, VALUE) \
+   STORE_PIXEL_X8R8G8B8(DST, X, Y, VALUE)
+#define STORE_PIXEL_RGB(DST, X, Y, VALUE) \
+   STORE_PIXEL_RGB_X8R8G8B8(DST, X, Y, VALUE)
+#define FETCH_PIXEL(DST, SRC) \
+   FETCH_PIXEL_X8R8G8B8(DST, SRC)
+
+#include "swrast_spantemp.h"
+
+
 /* 16-bit BGR */
 #define NAME(FUNC) FUNC##_R5G6B5_front
 #define RB_TYPE GLubyte
@@ -279,6 +333,15 @@ swrast_set_span_funcs_back(struct swrast_renderbuffer *xrb,
 	xrb->Base.PutValues = put_values_A8R8G8B8;
 	xrb->Base.PutMonoValues = put_mono_values_A8R8G8B8;
 	break;
+    case PF_X8R8G8B8:
+	xrb->Base.GetRow = get_row_X8R8G8B8;
+	xrb->Base.GetValues = get_values_X8R8G8B8;
+	xrb->Base.PutRow = put_row_X8R8G8B8;
+	xrb->Base.PutRowRGB = put_row_rgb_X8R8G8B8;
+	xrb->Base.PutMonoRow = put_mono_row_X8R8G8B8;
+	xrb->Base.PutValues = put_values_X8R8G8B8;
+	xrb->Base.PutMonoValues = put_mono_values_X8R8G8B8;
+	break;
    case PF_R5G6B5:
 	xrb->Base.GetRow = get_row_R5G6B5;
 	xrb->Base.GetValues = get_values_R5G6B5;
@@ -334,6 +397,15 @@ swrast_set_span_funcs_front(struct swrast_renderbuffer *xrb,
 	xrb->Base.PutValues = put_values_A8R8G8B8_front;
 	xrb->Base.PutMonoValues = put_mono_values_A8R8G8B8_front;
 	break;
+    case PF_X8R8G8B8:
+	xrb->Base.GetRow = get_row_X8R8G8B8_front;
+	xrb->Base.GetValues = get_values_X8R8G8B8_front;
+	xrb->Base.PutRow = put_row_X8R8G8B8_front;
+	xrb->Base.PutRowRGB = put_row_rgb_X8R8G8B8_front;
+	xrb->Base.PutMonoRow = put_mono_row_X8R8G8B8_front;
+	xrb->Base.PutValues = put_values_X8R8G8B8_front;
+	xrb->Base.PutMonoValues = put_mono_values_X8R8G8B8_front;
+	break;
    case PF_R5G6B5:
 	xrb->Base.GetRow = get_row_R5G6B5_front;
 	xrb->Base.GetValues = get_values_R5G6B5_front;
--- a/src/mesa/main/drawpix.c
+++ b/src/mesa/main/drawpix.c
@@ -77,28 +77,30 @@ _mesa_DrawPixels( GLsizei width, GLsizei height,
   }

   if (ctx->RenderMode == GL_RENDER) {
-      /* Round, to satisfy conformance tests (matches SGI's OpenGL) */
-      GLint x = IROUND(ctx->Current.RasterPos[0]);
-      GLint y = IROUND(ctx->Current.RasterPos[1]);
+      if (width > 0 && height > 0) {
+         /* Round, to satisfy conformance tests (matches SGI's OpenGL) */
+         GLint x = IROUND(ctx->Current.RasterPos[0]);
+         GLint y = IROUND(ctx->Current.RasterPos[1]);

-      if (ctx->Unpack.BufferObj->Name) {
-         /* unpack from PBO */
-         if (!_mesa_validate_pbo_access(2, &ctx->Unpack, width, height, 1,
-                                        format, type, pixels)) {
-            _mesa_error(ctx, GL_INVALID_OPERATION,
-                        "glDrawPixels(invalid PBO access)");
-            return;
-         }
-         if (ctx->Unpack.BufferObj->Pointer) {
-            /* buffer is mapped - that's an error */
-            _mesa_error(ctx, GL_INVALID_OPERATION,
-                        "glDrawPixels(PBO is mapped)");
-            return;
+         if (ctx->Unpack.BufferObj->Name) {
+            /* unpack from PBO */
+            if (!_mesa_validate_pbo_access(2, &ctx->Unpack, width, height, 1,
+                                           format, type, pixels)) {
+               _mesa_error(ctx, GL_INVALID_OPERATION,
+                           "glDrawPixels(invalid PBO access)");
+               return;
+            }
+            if (ctx->Unpack.BufferObj->Pointer) {
+               /* buffer is mapped - that's an error */
+               _mesa_error(ctx, GL_INVALID_OPERATION,
+                           "glDrawPixels(PBO is mapped)");
+               return;
+            }
         }
+
+         ctx->Driver.DrawPixels(ctx, x, y, width, height, format, type,
+                                &ctx->Unpack, pixels);
      }
-
-      ctx->Driver.DrawPixels(ctx, x, y, width, height, format, type,
-			     &ctx->Unpack, pixels);
   }
   else if (ctx->RenderMode == GL_FEEDBACK) {
      /* Feedback the current raster pos info */
@@ -159,10 +161,12 @@ _mesa_CopyPixels( GLint srcx, GLint srcy, GLsizei width, GLsizei height,

   if (ctx->RenderMode == GL_RENDER) {
      /* Round to satisfy conformance tests (matches SGI's OpenGL) */
-      GLint destx = IROUND(ctx->Current.RasterPos[0]);
-      GLint desty = IROUND(ctx->Current.RasterPos[1]);
-      ctx->Driver.CopyPixels( ctx, srcx, srcy, width, height, destx, desty,
-			      type );
+      if (width > 0 && height > 0) {
+         GLint destx = IROUND(ctx->Current.RasterPos[0]);
+         GLint desty = IROUND(ctx->Current.RasterPos[1]);
+         ctx->Driver.CopyPixels( ctx, srcx, srcy, width, height, destx, desty,
+                                 type );
+      }
   }
   else if (ctx->RenderMode == GL_FEEDBACK) {
      FLUSH_CURRENT( ctx, 0 );
--- a/src/mesa/main/ffvertex_prog.c
+++ b/src/mesa/main/ffvertex_prog.c
@@ -1104,6 +1104,9 @@ static void build_lighting( struct tnl_program *p )

   if (twoside) {
      if (!p->state->material_shininess_is_zero) {
+         /* Note that we negate the back-face specular exponent here.
+          * The negation will be un-done later in the back-face code below.
+          */
         struct ureg shininess = get_material(p, 1, STATE_SHININESS);
         emit_op1(p, OPCODE_MOV, dots, WRITEMASK_Z, 
                  negate(swizzle1(shininess,X)));
@@ -1309,6 +1312,11 @@ static void build_lighting( struct tnl_program *p )
 	       mask1 = 0;
 	    }

+            /* For the back face we need to negate the X and Y component
+             * dot products.  dots.Z has the negated back-face specular
+             * exponent.  We swizzle that into the W position.  This
+             * negation makes the back-face specular term positive again.
+             */
            dots = negate(swizzle(dots,X,Y,W,Z));

 	    if (!is_undef(att)) {
@@ -1327,8 +1335,10 @@ static void build_lighting( struct tnl_program *p )

 	    emit_op3(p, OPCODE_MAD, res0, mask0, swizzle1(lit,Y), diffuse, _bfc0);
 	    emit_op3(p, OPCODE_MAD, res1, mask1, swizzle1(lit,Z), specular, _bfc1);
-            /* restore negate flag for next lighting */
-            dots = negate(dots);
+            /* restore dots to its original state for subsequent lights
+             * by negating and swizzling again.
+             */
+            dots = negate(swizzle(dots,X,Y,W,Z));

 	    release_temp(p, ambient);
 	    release_temp(p, diffuse);
--- a/src/mesa/main/glheader.h
+++ b/src/mesa/main/glheader.h
@@ -145,7 +145,10 @@
 #if defined(__linux__)
 #include <byteswap.h>
 #define CPU_TO_LE32( x )	bswap_32( x )
-#else /*__linux__*/
+#elif defined(__APPLE__)
+#include <CoreFoundation/CFByteOrder.h>
+#define CPU_TO_LE32( x )	CFSwapInt32HostToLittle( x )
+#else /*__linux__ __APPLE__*/
 #include <sys/endian.h>
 #define CPU_TO_LE32( x )	bswap32( x )
 #endif /*__linux__*/
--- a/src/mesa/main/hash.c
+++ b/src/mesa/main/hash.c
@@ -63,6 +63,7 @@ struct _mesa_HashTable {
   struct HashEntry *Table[TABLE_SIZE];  /**< the lookup table */
   GLuint MaxKey;                        /**< highest key inserted so far */
   _glthread_Mutex Mutex;                /**< mutual exclusion lock */
+   _glthread_Mutex WalkMutex;            /**< for _mesa_HashWalk() */
   GLboolean InDeleteAll;                /**< Debug check */
 };

@@ -79,6 +80,7 @@ _mesa_NewHashTable(void)
   struct _mesa_HashTable *table = CALLOC_STRUCT(_mesa_HashTable);
   if (table) {
      _glthread_INIT_MUTEX(table->Mutex);
+      _glthread_INIT_MUTEX(table->WalkMutex);
   }
   return table;
 }
@@ -111,6 +113,7 @@ _mesa_DeleteHashTable(struct _mesa_HashTable *table)
      }
   }
   _glthread_DESTROY_MUTEX(table->Mutex);
+   _glthread_DESTROY_MUTEX(table->WalkMutex);
   _mesa_free(table);
 }

@@ -285,6 +288,11 @@ _mesa_HashDeleteAll(struct _mesa_HashTable *table,

 /**
 * Walk over all entries in a hash table, calling callback function for each.
+ * Note: we use a separate mutex in this function to avoid a recursive
+ * locking deadlock (in case the callback calls _mesa_HashRemove()) and to
+ * prevent multiple threads/contexts from getting tangled up.
+ * A lock-less version of this function could be used when the table will
+ * not be modified.
 * \param table  the hash table to walk
 * \param callback  the callback function
 * \param userData  arbitrary pointer to pass along to the callback
@@ -300,14 +308,16 @@ _mesa_HashWalk(const struct _mesa_HashTable *table,
   GLuint pos;
   ASSERT(table);
   ASSERT(callback);
-   _glthread_LOCK_MUTEX(table2->Mutex);
+   _glthread_LOCK_MUTEX(table2->WalkMutex);
   for (pos = 0; pos < TABLE_SIZE; pos++) {
-      struct HashEntry *entry;
-      for (entry = table->Table[pos]; entry; entry = entry->Next) {
+      struct HashEntry *entry, *next;
+      for (entry = table->Table[pos]; entry; entry = next) {
+         /* save 'next' pointer now in case the callback deletes the entry */
+         next = entry->Next;
         callback(entry->Key, entry->Data, userData);
      }
   }
-   _glthread_UNLOCK_MUTEX(table2->Mutex);
+   _glthread_UNLOCK_MUTEX(table2->WalkMutex);
 }


--- a/src/mesa/main/light.c
+++ b/src/mesa/main/light.c
@@ -209,7 +209,6 @@ _mesa_Lightfv( GLenum light, GLenum pname, const GLfloat *params )
 	 _math_matrix_analyse(ctx->ModelviewMatrixStack.Top);
      }
      TRANSFORM_DIRECTION(temp, params, ctx->ModelviewMatrixStack.Top->m);
-      NORMALIZE_3FV(temp);
      params = temp;
      break;
   case GL_SPOT_EXPONENT:
@@ -1137,12 +1136,18 @@ compute_light_positions( GLcontext *ctx )
      }

      if (light->_Flags & LIGHT_SPOT) {
+         /* Note: we normalize the spot direction now */
+
 	 if (ctx->_NeedEyeCoords) {
 	    COPY_3V( light->_NormDirection, light->EyeDirection );
+            NORMALIZE_3FV( light->_NormDirection );
 	 }
         else {
+            GLfloat spotDir[3];
+            COPY_3V(spotDir, light->EyeDirection);
+            NORMALIZE_3FV(spotDir);
 	    TRANSFORM_NORMAL( light->_NormDirection,
-			      light->EyeDirection,
+			      spotDir,
 			      ctx->ModelviewMatrixStack.Top->m);
 	 }

--- a/src/mesa/main/readpix.c
+++ b/src/mesa/main/readpix.c
@@ -170,6 +170,9 @@ _mesa_ReadPixels( GLint x, GLint y, GLsizei width, GLsizei height,
      return;
   }

+   if (width == 0 || height == 0)
+      return; /* nothing to do */
+
   if (ctx->Pack.BufferObj->Name) {
      if (!_mesa_validate_pbo_access(2, &ctx->Pack, width, height, 1,
                                     format, type, pixels)) {
--- a/src/mesa/main/texparam.c
+++ b/src/mesa/main/texparam.c
@@ -139,8 +139,11 @@ flush(GLcontext *ctx, struct gl_texture_object *texObj)
 }


-/** Set an integer-valued texture parameter */
-static void
+/**
+ * Set an integer-valued texture parameter
+ * \return GL_TRUE if legal AND the value changed, GL_FALSE otherwise
+ */
+static GLboolean
 set_tex_parameteri(GLcontext *ctx,
                   struct gl_texture_object *texObj,
                   GLenum pname, const GLint *params)
@@ -148,13 +151,13 @@ set_tex_parameteri(GLcontext *ctx,
   switch (pname) {
   case GL_TEXTURE_MIN_FILTER:
      if (texObj->MinFilter == params[0])
-         return;
+         return GL_FALSE;
      switch (params[0]) {
      case GL_NEAREST:
      case GL_LINEAR:
         flush(ctx, texObj);
         texObj->MinFilter = params[0];
-         return;
+         return GL_TRUE;
      case GL_NEAREST_MIPMAP_NEAREST:
      case GL_LINEAR_MIPMAP_NEAREST:
      case GL_NEAREST_MIPMAP_LINEAR:
@@ -162,77 +165,80 @@ set_tex_parameteri(GLcontext *ctx,
         if (texObj->Target != GL_TEXTURE_RECTANGLE_NV) {
            flush(ctx, texObj);
            texObj->MinFilter = params[0];
-            return;
+            return GL_TRUE;
         }
         /* fall-through */
      default:
         _mesa_error( ctx, GL_INVALID_VALUE, "glTexParameter(param)" );
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_MAG_FILTER:
      if (texObj->MagFilter == params[0])
-         return;
+         return GL_FALSE;
      switch (params[0]) {
      case GL_NEAREST:
      case GL_LINEAR:
         flush(ctx, texObj);
         texObj->MagFilter = params[0];
-         return;
+         return GL_TRUE;
      default:
         _mesa_error( ctx, GL_INVALID_VALUE, "glTexParameter(param)" );
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_WRAP_S:
      if (texObj->WrapS == params[0])
-         return;
+         return GL_FALSE;
      if (validate_texture_wrap_mode(ctx, texObj->Target, params[0])) {
         flush(ctx, texObj);
         texObj->WrapS = params[0];
+         return GL_TRUE;
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_WRAP_T:
      if (texObj->WrapT == params[0])
-         return;
+         return GL_FALSE;
      if (validate_texture_wrap_mode(ctx, texObj->Target, params[0])) {
         flush(ctx, texObj);
         texObj->WrapT = params[0];
+         return GL_TRUE;
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_WRAP_R:
      if (texObj->WrapR == params[0])
-         return;
+         return GL_FALSE;
      if (validate_texture_wrap_mode(ctx, texObj->Target, params[0])) {
         flush(ctx, texObj);
         texObj->WrapR = params[0];
+         return GL_TRUE;
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_BASE_LEVEL:
      if (texObj->BaseLevel == params[0])
-         return;
+         return GL_FALSE;
      if (params[0] < 0 ||
          (texObj->Target == GL_TEXTURE_RECTANGLE_ARB && params[0] != 0)) {
         _mesa_error(ctx, GL_INVALID_VALUE, "glTexParameter(param)");
-         return;
+         return GL_FALSE;
      }
      flush(ctx, texObj);
      texObj->BaseLevel = params[0];
-      return;
+      return GL_TRUE;

   case GL_TEXTURE_MAX_LEVEL:
      if (texObj->MaxLevel == params[0])
-         return;
+         return GL_FALSE;
      if (params[0] < 0 || texObj->Target == GL_TEXTURE_RECTANGLE_ARB) {
         _mesa_error(ctx, GL_INVALID_OPERATION, "glTexParameter(param)");
-         return;
+         return GL_FALSE;
      }
      flush(ctx, texObj);
      texObj->MaxLevel = params[0];
-      return;
+      return GL_TRUE;

   case GL_TEXTURE_COMPARE_SGIX:
      if (ctx->Extensions.SGIX_shadow) {
@@ -243,7 +249,7 @@ set_tex_parameteri(GLcontext *ctx,
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(pname=GL_TEXTURE_COMPARE_SGIX)");
      }
-      return;
+      return GL_TRUE;

   case GL_TEXTURE_COMPARE_OPERATOR_SGIX:
      if (ctx->Extensions.SGIX_shadow &&
@@ -256,20 +262,22 @@ set_tex_parameteri(GLcontext *ctx,
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(GL_TEXTURE_COMPARE_OPERATOR_SGIX)");
      }
-      return;
+      return GL_TRUE;

   case GL_GENERATE_MIPMAP_SGIS:
      if (ctx->Extensions.SGIS_generate_mipmap) {
         if (texObj->GenerateMipmap != params[0]) {
            flush(ctx, texObj);
            texObj->GenerateMipmap = params[0] ? GL_TRUE : GL_FALSE;
+            return GL_TRUE;
         }
+         return GL_FALSE;
      }
      else {
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(pname=GL_GENERATE_MIPMAP_SGIS)");
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_COMPARE_MODE_ARB:
      if (ctx->Extensions.ARB_shadow &&
@@ -278,24 +286,26 @@ set_tex_parameteri(GLcontext *ctx,
         if (texObj->CompareMode != params[0]) {
            flush(ctx, texObj);
            texObj->CompareMode = params[0];
+            return GL_TRUE;
         }
+         return GL_FALSE;
      }
      else {
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(GL_TEXTURE_COMPARE_MODE_ARB)");
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_COMPARE_FUNC_ARB:
      if (ctx->Extensions.ARB_shadow) {
         if (texObj->CompareFunc == params[0])
-            return;
+            return GL_FALSE;
         switch (params[0]) {
         case GL_LEQUAL:
         case GL_GEQUAL:
            flush(ctx, texObj);
            texObj->CompareFunc = params[0];
-            return;
+            return GL_TRUE;
         case GL_EQUAL:
         case GL_NOTEQUAL:
         case GL_LESS:
@@ -305,7 +315,7 @@ set_tex_parameteri(GLcontext *ctx,
            if (ctx->Extensions.EXT_shadow_funcs) {
               flush(ctx, texObj);
               texObj->CompareFunc = params[0];
-               return;
+               return GL_TRUE;
            }
            /* fall-through */
         default:
@@ -316,7 +326,7 @@ set_tex_parameteri(GLcontext *ctx,
      else {
         _mesa_error(ctx, GL_INVALID_ENUM, "glTexParameter(param)");
      }
-      return;
+      return GL_FALSE;

   case GL_DEPTH_TEXTURE_MODE_ARB:
      if (ctx->Extensions.ARB_depth_texture &&
@@ -326,13 +336,14 @@ set_tex_parameteri(GLcontext *ctx,
         if (texObj->DepthMode != params[0]) {
            flush(ctx, texObj);
            texObj->DepthMode = params[0];
+            return GL_TRUE;
         }
      }
      else {
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(GL_DEPTH_TEXTURE_MODE_ARB)");
      }
-      return;
+      return GL_FALSE;

 #ifdef FEATURE_OES_draw_texture
   case GL_TEXTURE_CROP_RECT_OES:
@@ -340,17 +351,21 @@ set_tex_parameteri(GLcontext *ctx,
      texObj->CropRect[1] = params[1];
      texObj->CropRect[2] = params[2];
      texObj->CropRect[3] = params[3];
-      break;
+      return GL_TRUE;
 #endif

   default:
      _mesa_error(ctx, GL_INVALID_ENUM, "glTexParameter(pname=0x%x)", pname);
   }
+   return GL_FALSE;
 }


-/** Set a float-valued texture parameter */
-static void
+/**
+ * Set a float-valued texture parameter
+ * \return GL_TRUE if legal AND the value changed, GL_FALSE otherwise
+ */
+static GLboolean
 set_tex_parameterf(GLcontext *ctx,
                   struct gl_texture_object *texObj,
                   GLenum pname, const GLfloat *params)
@@ -358,54 +373,56 @@ set_tex_parameterf(GLcontext *ctx,
   switch (pname) {
   case GL_TEXTURE_MIN_LOD:
      if (texObj->MinLod == params[0])
-         return;
+         return GL_FALSE;
      flush(ctx, texObj);
      texObj->MinLod = params[0];
-      return;
+      return GL_TRUE;

   case GL_TEXTURE_MAX_LOD:
      if (texObj->MaxLod == params[0])
-         return;
+         return GL_FALSE;
      flush(ctx, texObj);
      texObj->MaxLod = params[0];
-      return;
+      return GL_TRUE;

   case GL_TEXTURE_PRIORITY:
      flush(ctx, texObj);
      texObj->Priority = CLAMP(params[0], 0.0F, 1.0F);
-      return;
+      return GL_TRUE;

   case GL_TEXTURE_MAX_ANISOTROPY_EXT:
      if (ctx->Extensions.EXT_texture_filter_anisotropic) {
         if (texObj->MaxAnisotropy == params[0])
-            return;
+            return GL_FALSE;
         if (params[0] < 1.0) {
            _mesa_error(ctx, GL_INVALID_VALUE, "glTexParameter(param)" );
-            return;
+            return GL_FALSE;
         }
         flush(ctx, texObj);
         /* clamp to max, that's what NVIDIA does */
         texObj->MaxAnisotropy = MIN2(params[0],
                                      ctx->Const.MaxTextureMaxAnisotropy);
+         return GL_TRUE;
      }
      else {
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(pname=GL_TEXTURE_MAX_ANISOTROPY_EXT)");
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_COMPARE_FAIL_VALUE_ARB:
      if (ctx->Extensions.SGIX_shadow_ambient) {
         if (texObj->ShadowAmbient != params[0]) {
            flush(ctx, texObj);
            texObj->ShadowAmbient = CLAMP(params[0], 0.0F, 1.0F);
+            return GL_TRUE;
         }
      }
      else {
         _mesa_error(ctx, GL_INVALID_ENUM,
                     "glTexParameter(pname=GL_SHADOW_AMBIENT_SGIX)");
      }
-      return;
+      return GL_FALSE;

   case GL_TEXTURE_LOD_BIAS:
      /* NOTE: this is really part of OpenGL 1.4, not EXT_texture_lod_bias */
@@ -413,7 +430,9 @@ set_tex_parameterf(GLcontext *ctx,
         if (texObj->LodBias != params[0]) {
            flush(ctx, texObj);
            texObj->LodBias = params[0];
+            return GL_TRUE;
         }
+         return GL_FALSE;
      }
      break;

@@ -427,17 +446,19 @@ set_tex_parameterf(GLcontext *ctx,
      UNCLAMPED_FLOAT_TO_CHAN(texObj->_BorderChan[GCOMP], params[1]);
      UNCLAMPED_FLOAT_TO_CHAN(texObj->_BorderChan[BCOMP], params[2]);
      UNCLAMPED_FLOAT_TO_CHAN(texObj->_BorderChan[ACOMP], params[3]);
-      return;
+      return GL_TRUE;

   default:
      _mesa_error(ctx, GL_INVALID_ENUM, "glTexParameter(pname=0x%x)", pname);
   }
+   return GL_FALSE;
 }


 void GLAPIENTRY
 _mesa_TexParameterf(GLenum target, GLenum pname, GLfloat param)
 {
+   GLboolean need_update;
   struct gl_texture_object *texObj;
   GET_CURRENT_CONTEXT(ctx);
   ASSERT_OUTSIDE_BEGIN_END(ctx);
@@ -463,15 +484,15 @@ _mesa_TexParameterf(GLenum target, GLenum pname, GLfloat param)
      {
         /* convert float param to int */
         GLint p = (GLint) param;
-         set_tex_parameteri(ctx, texObj, pname, &p);
+         need_update = set_tex_parameteri(ctx, texObj, pname, &p);
      }
-      return;
+      break;
   default:
      /* this will generate an error if pname is illegal */
-      set_tex_parameterf(ctx, texObj, pname, &param);
+      need_update = set_tex_parameterf(ctx, texObj, pname, &param);
   }

-   if (ctx->Driver.TexParameter && ctx->ErrorValue == GL_NO_ERROR) {
+   if (ctx->Driver.TexParameter && need_update) {
      ctx->Driver.TexParameter(ctx, target, texObj, pname, &param);
   }
 }
@@ -480,6 +501,7 @@ _mesa_TexParameterf(GLenum target, GLenum pname, GLfloat param)
 void GLAPIENTRY
 _mesa_TexParameterfv(GLenum target, GLenum pname, const GLfloat *params)
 {
+   GLboolean need_update;
   struct gl_texture_object *texObj;
   GET_CURRENT_CONTEXT(ctx);
   ASSERT_OUTSIDE_BEGIN_END(ctx);
@@ -505,7 +527,7 @@ _mesa_TexParameterfv(GLenum target, GLenum pname, const GLfloat *params)
      {
         /* convert float param to int */
         GLint p = (GLint) params[0];
-         set_tex_parameteri(ctx, texObj, pname, &p);
+         need_update = set_tex_parameteri(ctx, texObj, pname, &p);
      }
      break;

@@ -518,17 +540,17 @@ _mesa_TexParameterfv(GLenum target, GLenum pname, const GLfloat *params)
         iparams[1] = (GLint) params[1];
         iparams[2] = (GLint) params[2];
         iparams[3] = (GLint) params[3];
-         set_tex_parameteri(ctx, target, iparams);
+         need_update = set_tex_parameteri(ctx, target, iparams);
      }
      break;
 #endif

   default:
      /* this will generate an error if pname is illegal */
-      set_tex_parameterf(ctx, texObj, pname, params);
+      need_update = set_tex_parameterf(ctx, texObj, pname, params);
   }

-   if (ctx->Driver.TexParameter && ctx->ErrorValue == GL_NO_ERROR) {
+   if (ctx->Driver.TexParameter && need_update) {
      ctx->Driver.TexParameter(ctx, target, texObj, pname, params);
   }
 }
@@ -537,6 +559,7 @@ _mesa_TexParameterfv(GLenum target, GLenum pname, const GLfloat *params)
 void GLAPIENTRY
 _mesa_TexParameteri(GLenum target, GLenum pname, GLint param)
 {
+   GLboolean need_update;
   struct gl_texture_object *texObj;
   GET_CURRENT_CONTEXT(ctx);
   ASSERT_OUTSIDE_BEGIN_END(ctx);
@@ -555,15 +578,15 @@ _mesa_TexParameteri(GLenum target, GLenum pname, GLint param)
      {
         GLfloat fparam = (GLfloat) param;
         /* convert int param to float */
-         set_tex_parameterf(ctx, texObj, pname, &fparam);
+         need_update = set_tex_parameterf(ctx, texObj, pname, &fparam);
      }
      break;
   default:
      /* this will generate an error if pname is illegal */
-      set_tex_parameteri(ctx, texObj, pname, &param);
+      need_update = set_tex_parameteri(ctx, texObj, pname, &param);
   }

-   if (ctx->Driver.TexParameter && ctx->ErrorValue == GL_NO_ERROR) {
+   if (ctx->Driver.TexParameter && need_update) {
      GLfloat fparam = (GLfloat) param;
      ctx->Driver.TexParameter(ctx, target, texObj, pname, &fparam);
   }
@@ -573,6 +596,7 @@ _mesa_TexParameteri(GLenum target, GLenum pname, GLint param)
 void GLAPIENTRY
 _mesa_TexParameteriv(GLenum target, GLenum pname, const GLint *params)
 {
+   GLboolean need_update;
   struct gl_texture_object *texObj;
   GET_CURRENT_CONTEXT(ctx);
   ASSERT_OUTSIDE_BEGIN_END(ctx);
@@ -590,7 +614,7 @@ _mesa_TexParameteriv(GLenum target, GLenum pname, const GLint *params)
         fparams[1] = INT_TO_FLOAT(params[1]);
         fparams[2] = INT_TO_FLOAT(params[2]);
         fparams[3] = INT_TO_FLOAT(params[3]);
-         set_tex_parameterf(ctx, texObj, pname, fparams);
+         need_update = set_tex_parameterf(ctx, texObj, pname, fparams);
      }
      break;
   case GL_TEXTURE_MIN_LOD:
@@ -602,15 +626,15 @@ _mesa_TexParameteriv(GLenum target, GLenum pname, const GLint *params)
      {
         /* convert int param to float */
         GLfloat fparam = (GLfloat) params[0];
-         set_tex_parameterf(ctx, texObj, pname, &fparam);
+         need_update = set_tex_parameterf(ctx, texObj, pname, &fparam);
      }
      break;
   default:
      /* this will generate an error if pname is illegal */
-      set_tex_parameteri(ctx, texObj, pname, params);
+      need_update = set_tex_parameteri(ctx, texObj, pname, params);
   }

-   if (ctx->Driver.TexParameter && ctx->ErrorValue == GL_NO_ERROR) {
+   if (ctx->Driver.TexParameter && need_update) {
      GLfloat fparams[4];
      fparams[0] = INT_TO_FLOAT(params[0]);
      if (pname == GL_TEXTURE_BORDER_COLOR ||
--- a/src/mesa/main/version.h
+++ b/src/mesa/main/version.h
@@ -31,7 +31,7 @@
 #define MESA_MAJOR 7
 #define MESA_MINOR 4
 #define MESA_PATCH 0
-#define MESA_VERSION_STRING "7.4-rc2"
+#define MESA_VERSION_STRING "7.4"

 /* To make version comparison easy */
 #define MESA_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c))
--- a/src/mesa/math/m_matrix.c
+++ b/src/mesa/math/m_matrix.c
@@ -1379,11 +1379,10 @@ _math_matrix_analyse( GLmatrix *mat )

   if (mat->inv && (mat->flags & MAT_DIRTY_INVERSE)) {
      matrix_invert( mat );
+      mat->flags &= ~MAT_DIRTY_INVERSE;
   }

-   mat->flags &= ~(MAT_DIRTY_FLAGS|
-		   MAT_DIRTY_TYPE|
-		   MAT_DIRTY_INVERSE);
+   mat->flags &= ~(MAT_DIRTY_FLAGS | MAT_DIRTY_TYPE);
 }

 /*@}*/
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -1823,7 +1823,6 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
               if (param_var->param_binding_begin == ~0U)
                  param_var->param_binding_begin = idx;
               param_var->param_binding_length++;
-               Program->Base.NumParameters++;
            }
         }
         else {
@@ -1832,7 +1831,6 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
            if (param_var->param_binding_begin == ~0U)
               param_var->param_binding_begin = idx;
            param_var->param_binding_length++;
-            Program->Base.NumParameters++;
         }
         break;

@@ -1843,7 +1841,6 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
         if (param_var->param_binding_begin == ~0U)
            param_var->param_binding_begin = idx;
         param_var->param_binding_length++;
-         Program->Base.NumParameters++;

         /* Check if there is more: 0 -> we're done, else its an integer */
         if (**inst) {
@@ -1879,7 +1876,6 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
               idx = _mesa_add_state_reference(Program->Base.Parameters,
                                               state_tokens);
               param_var->param_binding_length++;
-               Program->Base.NumParameters++;
            }
         }
         else {
@@ -1897,7 +1893,6 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
            param_var->param_binding_begin = idx;
         param_var->param_binding_type = PROGRAM_CONSTANT;
         param_var->param_binding_length++;
-         Program->Base.NumParameters++;
         break;

      default:
@@ -1906,12 +1901,14 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
         return 1;
   }

+   Program->Base.NumParameters = Program->Base.Parameters->NumParameters;
+
   /* Make sure we haven't blown past our parameter limits */
   if (((Program->Base.Target == GL_VERTEX_PROGRAM_ARB) &&
-        (Program->Base.NumParameters >=
+        (Program->Base.NumParameters >
         ctx->Const.VertexProgram.MaxLocalParams))
       || ((Program->Base.Target == GL_FRAGMENT_PROGRAM_ARB)
-           && (Program->Base.NumParameters >=
+           && (Program->Base.NumParameters >
               ctx->Const.FragmentProgram.MaxLocalParams))) {
      program_error(ctx, Program->Position, "Too many parameter variables");
      return 1;
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -873,7 +873,7 @@ _mesa_execute_program(GLcontext * ctx,
                * result.z = result.x * APPX(result.y)
                * We do what the ARB extension says.
                */
-               q[2] = (GLfloat) pow(2.0, t[0]);
+               q[2] = (GLfloat) _mesa_pow(2.0, t[0]);
            }
            q[1] = t[0] - floor_t0;
            q[3] = 1.0F;
@@ -986,7 +986,7 @@ _mesa_execute_program(GLcontext * ctx,
               if (a[1] == 0.0 && a[3] == 0.0)
                  result[2] = 1.0;
               else
-                  result[2] = EXPF(a[3] * LOGF(a[1]));
+                  result[2] = (GLfloat) _mesa_pow(a[1], a[3]);
            }
            else {
               result[2] = 0.0;
@@ -1577,8 +1577,8 @@ _mesa_execute_program(GLcontext * ctx,
      case OPCODE_TXB:         /* GL_ARB_fragment_program only */
         /* Texel lookup with LOD bias */
         {
-            const struct gl_texture_unit *texUnit
-               = &ctx->Texture.Unit[inst->TexSrcUnit];
+            const GLuint unit = machine->Samplers[inst->TexSrcUnit];
+            const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
            GLfloat texcoord[4], color[4], lodBias;

            fetch_vector4(&inst->SrcReg[0], machine, texcoord);
--- a/src/mesa/shader/prog_parameter.c
+++ b/src/mesa/shader/prog_parameter.c
@@ -178,15 +178,20 @@ _mesa_add_named_constant(struct gl_program_parameter_list *paramList,
                         const char *name, const GLfloat values[4],
                         GLuint size)
 {
-#if 0 /* disable this for now -- we need to save the name! */
+   /* first check if this is a duplicate constant */
   GLint pos;
-   GLuint swizzle;
-   ASSERT(size == 4); /* XXX future feature */
-   /* check if we already have this constant */
-   if (_mesa_lookup_parameter_constant(paramList, values, 4, &pos, &swizzle)) {
-      return pos;
+   for (pos = 0; pos < paramList->NumParameters; pos++) {
+      const GLfloat *pvals = paramList->ParameterValues[pos];
+      if (pvals[0] == values[0] &&
+          pvals[1] == values[1] &&
+          pvals[2] == values[2] &&
+          pvals[3] == values[3] &&
+          _mesa_strcmp(paramList->Parameters[pos].Name, name) == 0) {
+         /* Same name and value is already in the param list - reuse it */
+         return pos;
+      }
   }
-#endif
+   /* not found, add new parameter */
   return _mesa_add_parameter(paramList, PROGRAM_CONSTANT, name,
                              size, GL_NONE, values, NULL, 0x0);
 }
--- a/src/mesa/shader/slang/slang_builtin.c
+++ b/src/mesa/shader/slang/slang_builtin.c
@@ -457,28 +457,36 @@ emit_statevars(const char *name, int array_len,
         tokens[0] = STATE_TEXENV_COLOR;
      }
      else if (strcmp(name, "gl_EyePlaneS") == 0) {
-         tokens[0] = STATE_TEXGEN_EYE_S;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_EYE_S;
      }
      else if (strcmp(name, "gl_EyePlaneT") == 0) {
-         tokens[0] = STATE_TEXGEN_EYE_T;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_EYE_T;
      }
      else if (strcmp(name, "gl_EyePlaneR") == 0) {
-         tokens[0] = STATE_TEXGEN_EYE_R;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_EYE_R;
      }
      else if (strcmp(name, "gl_EyePlaneQ") == 0) {
-         tokens[0] = STATE_TEXGEN_EYE_Q;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_EYE_Q;
      }
      else if (strcmp(name, "gl_ObjectPlaneS") == 0) {
-         tokens[0] = STATE_TEXGEN_OBJECT_S;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_OBJECT_S;
      }
      else if (strcmp(name, "gl_ObjectPlaneT") == 0) {
-         tokens[0] = STATE_TEXGEN_OBJECT_T;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_OBJECT_T;
      }
      else if (strcmp(name, "gl_ObjectPlaneR") == 0) {
-         tokens[0] = STATE_TEXGEN_OBJECT_R;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_OBJECT_R;
      }
      else if (strcmp(name, "gl_ObjectPlaneQ") == 0) {
-         tokens[0] = STATE_TEXGEN_OBJECT_Q;
+         tokens[0] = STATE_TEXGEN;
+         tokens[2] = STATE_TEXGEN_OBJECT_Q;
      }
      else {
         return -1; /* invalid array name */
--- a/src/mesa/swrast/s_points.c
+++ b/src/mesa/swrast/s_points.c
@@ -570,6 +570,9 @@ void
 _swrast_choose_point(GLcontext *ctx)
 {
   SWcontext *swrast = SWRAST_CONTEXT(ctx);
+   const GLfloat size = CLAMP(ctx->Point.Size,
+                              ctx->Point.MinSize,
+                              ctx->Point.MaxSize);

   if (ctx->RenderMode == GL_RENDER) {
      if (ctx->Point.PointSprite) {
@@ -578,7 +581,7 @@ _swrast_choose_point(GLcontext *ctx)
      else if (ctx->Point.SmoothFlag) {
         swrast->Point = smooth_point;
      }
-      else if (ctx->Point.Size > 1.0 ||
+      else if (size > 1.0 ||
               ctx->Point._Attenuated ||
               ctx->VertexProgram.PointSizeEnabled) {
         swrast->Point = large_point;
--- a/src/mesa/swrast/s_span.c
+++ b/src/mesa/swrast/s_span.c
@@ -1399,11 +1399,17 @@ _swrast_write_rgba_span( GLcontext *ctx, SWspan *span)
   }

   /*
-    * Write to renderbuffers
+    * Write to renderbuffers.
+    * Depending on glDrawBuffer() state and the which color outputs are
+    * written by the fragment shader, we may either replicate one color to
+    * all renderbuffers or write a different color to each renderbuffer.
+    * multiFragOutputs=TRUE for the later case.
    */
   {
      const GLuint numBuffers = fb->_NumColorDrawBuffers;
-      const GLboolean multiFragOutputs = numBuffers > 1;
+      const struct gl_fragment_program *fp = ctx->FragmentProgram._Current;
+      const GLboolean multiFragOutputs = 
+         (fp && fp->Base.OutputsWritten >= (1 << FRAG_RESULT_DATA0));
      GLuint buf;

      for (buf = 0; buf < numBuffers; buf++) {
Author	SHA1	Message	Date
Ian Romanick	85396d8ef4	intel: Bump driver date	2009-04-15 15:12:48 -07:00
Ian Romanick	f11b84998d	docs: more bug fixes for Mesa 7.4.1	2009-04-15 15:10:33 -07:00
Brian Paul	d2f6791062	glx: added null pointer check in glXGetFBConfigs() Fixes segfault seen with glxinfo with NVIDIA OpenGL. (cherry picked from master, commit `05471828dc`)	2009-04-15 08:02:38 -06:00
Ian Romanick	49e0c74ddd	DRI2: Don't fault on NULL DrawBuffer It is possible for ctx->DrawBuffer to be NULL, so don't fault when that happens. This change is not being committed to master because it doesn't appear to be necessary there. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>	2009-04-11 20:24:15 -07:00
Eric Anholt	83e14773c5	dri2: Don't crash if the server returns more buffers than expected. (cherry picked from commit `f967e8b507`)	2009-04-09 15:15:33 -07:00
Ian Romanick	4605937843	intel / DRI2: Accept fake front-buffer from loader Handle the loader returning a fake front-buffer. Since the driver never specifically requests a fake front-buffer, the driver assumes that it will never receive both a fake and a real front-buffer. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Kristian Høgsberg <krh@redhat.com> (cherry picked from commit `6a6e478e55`)	2009-04-09 14:44:53 -07:00
Ian Romanick	b8ef30c45c	DRI2: Assume that there is always a front buffer Assume that the front-buffer exists even if the server didn't tell the client that it exists. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Kristian Høgsberg <krh@redhat.com> (cherry picked from commit `e6386e0d9a`)	2009-04-09 14:44:44 -07:00
Ian Romanick	4f17040cb1	intel / DRI2: Track and flush front-buffer rendering Track two flags: whether or not front-buffer rendering is currently enabled and whether or not front-buffer rendering has been enabled since the last glFlush. If the second flag is set, the front-buffer is flushed via a loader call back. If the first flag is cleared, the second flag is cleared at this time. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Kristian Høgsberg <krh@redhat.com> (cherry picked from commit `43cf0d1eeb`)	2009-04-09 14:44:34 -07:00
Ian Romanick	98c2c9ef8c	DRI2: Provide an interface for drivers to flush front-buffer rendering Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Kristian Høgsberg <krh@redhat.com> (cherry picked from commit `82634ee8df`)	2009-04-09 14:44:17 -07:00
Brian Paul	d805c82068	mesa: fix potential recursive locking deadlock in _mesa_HashWalk() If the walk callback called _mesa_HashRemove() we'd deadlock. (cherry picked from master, commit `deff099215`)	2009-04-09 10:55:06 -06:00
Brian Paul	6c3bec4387	swrast: fix point rendering function selection Need to clamp default point size to min/max range before checking if it's one. Fixes glean pointAtten test. (cherry picked from master, commit `228f20e324`)	2009-04-07 13:44:50 -06:00
Brian Paul	7d3561c871	docs: more bug fixes for Mesa 7.4.1	2009-04-07 08:37:22 -06:00
Brian Paul	5c47d53924	mesa: in mesa_add_named_constant(), avoid adding duplicate constants (cherry picked from master, commit `80197a0c1b`)	2009-04-07 08:35:01 -06:00
Brian Paul	a18216308a	mesa: fix parameter counting in ARB vertex/fragment program parsing Duplicated unnamed constants were getting counted more than once. (cherry picked from master, commit `866bdd0509`)	2009-04-07 08:34:32 -06:00
Brian Paul	14f13fbcfa	mesa: replace >= with > when testing if we've exceeded max local params Now a program that uses 256 locals works as it should. (cherry picked from master, commit `a4173956eb`)	2009-04-07 08:34:04 -06:00
Brian Paul	ca24095c1e	mesa: only clear matrix MAT_DIRTY_INVERSE flag when we actually compute the inverse If _math_matrix_analyse() got called before we allocated the inverse matrix array we could lose the flag indicating that we needed to compute the inverse. This could happen with certain vertex shader cases. (cherry picked from master, commit `ce461ffc5a`)	2009-04-07 08:33:43 -06:00
Brian Paul	d5e0e03d43	mesa: for OPCODE_LIT, use _mesa_pow() instead of exp() and log() Also, s/pow/_mesa_pow/ (cherry picked from master, commit `b8a200ac9d`)	2009-04-07 08:32:42 -06:00
Michel Dänzer	775ca8e3fa	radeon: Expose a 32 bit RGBA fbconfig even when the screen depth is 16. Otherwise current xserver / libGL no longer expose a 32 bit RGBA GLX visual, and compiz fails. Fixes http://bugs.freedesktop.org/show_bug.cgi?id=20479 . (Cherry picked from commit `e798bf8053`)	2009-04-06 09:27:17 +02:00
Brian Paul	39345a4928	mesa: don't normalize spot light direction until validation time In glLight() we're only supposed to transform the direction by the modelview matrix, not normalized it too. (cherry picked from commit `650d147289`)	2009-04-03 10:16:31 -06:00
Roland Scheidegger	7be149cfd1	mesa: fix TexParameter functions premature return in TexParameterf caused mesa to never call Driver.TexParameter breaking drivers relying on this (fix bug #20966). While here, also fix using ctx->ErrorValue when deciding to call Driver.TexParameter. Errors are sticky and uncleared errors thus would cause this to no longer get called. Since we thus need return value of set_tex_parameter[if] can also optimize this to only call when value changed.	2009-04-03 00:00:40 +02:00
Brian Paul	8c9d7b2474	docs: fix TXB texture unit bias error	2009-04-02 13:11:56 -06:00
Brian Paul	6ef0951b17	mesa: don't call ctx->Driver.ReadPixels() if width or height is zero (cherry picked from commit master, `7b9bf39543`)	2009-04-02 13:11:33 -06:00
Brian Paul	2b33b77337	mesa: don't call ctx->Driver.Draw/CopyPixels() if width or height is zero (cherry picked from master, commit `f6a3f1f52a`)	2009-04-02 13:10:41 -06:00
Brian Paul	0590edeea0	mesa: use correct tex unit lod bias for TXB instruction (cherry picked from master, commit `1ab225017e`)	2009-04-02 13:09:32 -06:00
Brian Paul	7511d76bd7	swrast: fix glDrawBuffer(GL_FRONT_AND_BACK) We weren't putting the right colors into the back buffer in this mode. (cherry picked from master, commit `9cc79fc2dc`)	2009-04-02 13:07:53 -06:00
Brian Paul	7f30aba920	docs: fix bug 20986	2009-04-01 07:39:43 -06:00
Brian Paul	4fa700260e	glsl: fix texgen state variable tokens in emit_statevars() This fixes broken variable indexing into the gl_Eye/ObjectPlaneS/T/R/Q arrays. See bug 20986. (cherry picked from master, commit `f8dd6594bf`)	2009-04-01 07:39:26 -06:00
Brian Paul	781fb79c59	docs: first 7.4.1 bug fixes	2009-03-31 16:29:23 -06:00
Brian Paul	5b7e9f2f3a	mesa: fix bug in GPU codegen for fixed-function two-sided lighting The 'dots' register wasn't getting properly un-negated and un-swizzled after emitting the code for back-face lighting. So, if more than one light source was enabled, the specular exponent for the next light source was wrong. During execution we were evaluating pow(x, y) where y was negative instead of positive. This led to the outcome being zero or NaN. This fixes the occasional black triangles seen in isosurf when hacked to enable two-sided lighting. (cherry picked from master, commit `919f57078b`)	2009-03-31 16:28:17 -06:00
Brian Paul	d4d4b63d7f	docs: prep for Mesa 7.4.1 release notes	2009-03-31 16:26:24 -06:00
Jeremy Huddleston	9e9fe51acd	Updated CPU_TO_LE32 to work on darwin	2009-03-31 14:59:28 -07:00
Jeremy Huddleston	b65bc1b6cb	Fix compiling indirect.c when GLX_DIRECT_RENDERING is not defined	2009-03-31 14:48:19 -07:00
Jeremy Huddleston	6cb796f6fc	Updated darwin config for when X11 is not in the same location as we're installing to (cherry picked from commit `7817fea0b6`)	2009-03-31 14:37:27 -07:00
Brian Paul	de197cf991	docs: 7.4 release md5 sums	2009-03-27 19:12:16 -06:00
Brian Paul	bf1e6581a5	docs: set 7.4 release date	2009-03-27 18:57:39 -06:00
Brian Paul	80ec7843bc	mesa: set VERSION=7.4	2009-03-27 18:57:23 -06:00
Brian Paul	366d6e8cd1	mesa: set version to 7.4	2009-03-27 18:57:07 -06:00
Ian Romanick	63ed288396	intel: Add extra, stronger flushes around CopyTexSubImage blits	2009-03-27 15:22:27 -07:00
Eric Anholt	88ce34f400	swrast: Add support for x8r8g8b8 fbconfig. This lets swrast produce an fbconfig suitable for the root visual now that the server's not allowing mismatched fbconfigs. (cherry picked from commit `529d1d720e`)	2009-03-27 15:20:09 -07:00