commit c1eb78f7ff9a6d80a6db06727a521b5aa3871a3d
Author: Roland Scheidegger <sroland@tungstengraphics.com>
Date: Sat Dec 22 18:49:42 2007 +0100
fix GL_LINE_LOOP with drivers using own render pipeline stage (#12410, #13527)
primitive needs to include the begin/end flags (broken since vbo-0.2). Should
fix missing first/last line segment on gamma, i810, i915, mga, r200, radeon,
s3v, savage, unichrome (r300 already correct). Tested on r200, fixes #13527.
commit 8aa0fd6b206930ababb1bae2acc629e2a945c0c0
Author: Brian <brian.paul@tungstengraphics.com>
Date: Thu Dec 20 09:07:57 2007 -0700
glGetActiveUniform fix
commit a1b3a908f557502fc1e5bf4e6357014538778ed8
Author: Brian <brian.paul@tungstengraphics.com>
Date: Thu Dec 20 09:05:37 2007 -0700
return correct size from glGetActiveUniform (bug 13751)
commit f9e70d951a3047a2af87ee43ca6c1045a187ef08
Author: Xiang, Haihao <haihao.xiang@intel.com>
Date: Wed Dec 19 10:44:00 2007 +0800
i965: allocate GRF registers before building subroutines,
it ensures there are sufficient registers for all subroutines.
commit 41ed6be1da92fe8b48f106e7b540bd6a8ad2446a
Author: Xiang, Haihao <haihao.xiang@intel.com>
Date: Wed Dec 19 10:36:56 2007 +0800
i965: restore the flag after building the subroutine of the
GS thread. fix #13240
commit 27de28fc3ec46cd5cac1e6de52d27922aabe340e
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Dec 18 17:07:11 2007 -0700
Remove -I$(TOP)/src/mesa/transform
commit dd2f01c229e9a9692fb0c424cb3c587419d04f83
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Dec 18 16:12:34 2007 -0700
simplify update two-side lighting test (follow-on to previous front/back-face changes)
commit deb5c56c77d385786d0b4fceed01cdee82f5355d
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Dec 18 16:10:06 2007 -0700
fix NEED_SECONDARY_COLOR for vert/frag progs
commit 0107acded03c510df2093b0b98bca52b5734dd5b
Author: Michel Dänzer <michel@tungstengraphics.com>
Date: Mon Dec 17 12:20:56 2007 +0100
i915tex: Fix issues with glDrawBuffer(GL_NONE).
Don't dereference NULL renderbuffer pointer, and make sure the software
fallback sticks.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=13694 .
commit 2ac5e08d1d41377dc873b31875701f59e74242ec
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Dec 14 14:33:53 2007 -0700
fix polygon cull regression
commit 1837b8c214a362ccecc441f7b6880f623e70b8d6
Author: Dan Nicholson <dbn.lists@gmail.com>
Date: Wed Oct 31 17:12:07 2007 -0700
configs: Fix linking with static libGL and --as-needed
Linking of the programs breaks when using a static libGL and the GNU ld
option --as-needed. This is because libXext is needed for the XShm
functions.
(cherry picked from commit 6aaf9bdd418e11d5ccb76a6ab85a8df36b0c8804)
commit 03d2bc1774e3612edca4be48c5aeafaafa2a52b1
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Dec 4 14:06:26 2007 -0700
fix gl_FrontFacing
commit 385bddbde04ca3ec2a159e8c8a50d9009da25b63
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Dec 4 14:06:01 2007 -0700
fix span->facing computation and gl_FrontFacing initialization
commit c13542626729a6fe3e4dfbc71f3569119255b668
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Dec 4 14:05:11 2007 -0700
Fix gl_FrontFacing compilation problem
commit 4be0c981205b5530366a10eb96e06b93a2439391
Author: Michel Dänzer <michel@tungstengraphics.com>
Date: Tue Dec 4 10:46:45 2007 +0100
Use -Bsymbolic for linking all shared objects.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=10132 .
Also remove comment about SONAME, as SONAME only applies to shared libraries.
(cherry picked from commit fc7ddea8535f1a9c196bf30f7864414e4ac18b8a)
commit e760aebd5aaecc220c6789fe925897909a876acd
Author: Brian <brian@i915.localnet.net>
Date: Mon Dec 3 14:11:46 2007 -0700
fix two-side lighting bug/crash
commit 1f6e7d9c297aba2e936352d6ae8a4198b5c64af5
Author: Brian <brian@i915.localnet.net>
Date: Mon Dec 3 14:11:26 2007 -0700
fix DD_TRI_LIGHT_TWOSIDE bug (#13368)
commit c6d421af05ccdd377df5cade301c5c25814ddf0c
Author: Dan Nicholson <dbn.lists@gmail.com>
Date: Mon Dec 3 11:57:14 2007 -0800
Fix the library name in glw.pc
Fix a copy and paste error s/GLU/GLw/ in glw.pc.
(cherry picked from commit 2dc85e8078bed67cb5d8bc3e16484f56c61c61ca)
commit 952df5e493198de91b329983dae5366102eb0a25
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Dec 3 11:59:40 2007 -0700
fix two-sided stencil
commit 45acb8b058fc0f61a1c2687897c5d4042b534760
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Dec 3 11:58:16 2007 -0700
added missing quote char
commit e8997c0183dbdb59884cc02af81d2faab1beaba1
Author: Michel Dänzer <michel@tungstengraphics.com>
Date: Fri Nov 30 11:40:06 2007 +0100
i915tex: Fix up state changes for i8xx.
(cherry picked from commit 2af613e0b8e6f89cc7528ff5a969e18f077e61fc)
commit 169e62f563c4eb7dfa9c456a1ee03168af6b546a
Author: Brian <brian@i915.localnet.net>
Date: Fri Nov 30 15:52:41 2007 -0700
fix-build: remove ctx->_Facing assignment
commit 1b43babfb126c6318d2f2468ec81b93ee1923774
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 30 13:01:42 2007 -0700
fix broken two-sided stencil
commit 43e902f774f8c3cd58310eeb44eec9d3f8f81746
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 30 09:08:11 2007 -0700
better front-plane clip test
commit b0a800e249990de0cbc577f8b48a1a301a3b9361
Author: Xiang, Haihao <haihao.xiang@intel.com>
Date: Fri Nov 30 11:50:08 2007 +0800
i965: use uncompressed instruction to ensure only
Pixel Mask Copy is modified as the pixel shader thread
turns off pixels based on kill instructions.
commit 73d5f232ad5905e6566ddd22f2cbceaac48bd4ad
Author: Brian <brian.paul@tungstengraphics.com>
Date: Thu Nov 29 08:12:33 2007 -0700
New ctx->Driver.Map/UnmapTexture() functions for accessing textures from t_vb_program.c
commit f7209541e44c35e64bc9cd4fd8342ee7743b7774
Author: Brian <brian.paul@tungstengraphics.com>
Date: Thu Nov 29 08:17:37 2007 -0700
cleanups, comments
Conflicts:
src/mesa/tnl/t_vb_program.c
commit 397a32dca7b58b1120874b304e0f39fc314dbb3b
Author: Brian <brian.paul@tungstengraphics.com>
Date: Wed Nov 28 15:19:46 2007 -0700
Move _mesa_load_tracked_matrices() from TNL module to prog_statevars.c
commit e0719d7122e1ac0659c8195ae267a45579c4070c
Author: Roland Scheidegger <sroland@tungstengraphics.com>
Date: Thu Nov 29 03:08:18 2007 +0100
make sure state token values are fully initialized
commit 3f18c0a9f27fded8f52a5f0c5b8ad71f71c46aa2
Author: Michel Dänzer <michel@tungstengraphics.com>
Date: Wed Nov 28 10:20:04 2007 +0100
r200: Fix texture format regression on big endian systems.
See https://bugs.freedesktop.org/show_bug.cgi?id=13324 .
Also use tx_table_be for VALID_FORMAT, in case r200SetTexImages ever gets
called for MESA_FORMAT_RGB888.
(cherry picked from commit dc88a96631c21c2bc2cdffefc8947e8924d77c91)
commit b14be6193891dd7f03efcadb674a759acf3722c9
Author: Delle <delle@ciaoweb.it>
Date: Tue Nov 27 18:18:48 2007 -0700
use DEFAULT_SOFTWARE_DEPTH_BITS
commit 440d6203080650bc112c4759dad6bcb3ed9e7fa2
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 27 10:30:48 2007 -0700
minor additions to avoid FAQs
commit 8dd9df0f3e53549acc874d891388cfa34ab038c0
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 27 10:28:30 2007 -0700
document GLSL float f/F suffix bug
commit 6b9534eaf5d617fd8ce3888885913c9665b4ca7c
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 27 10:17:02 2007 -0700
set fp->UsesKill when emitting OPCODE_KIL
commit c91d374ad7736c5c98010780a5701722d50f4f30
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Nov 26 09:35:31 2007 -0700
improve 24-bit Z to 32-bit Z conversion
commit 1cab4160bc5b22264023ffde05aeab5bebf87db7
Author: Keith Whitwell <keith@tungstengraphics.com>
Date: Thu Nov 15 15:25:11 2007 +0000
i915tex: Fix some minor batchbuffer errors.
commit 718d2dfbbddbca3bcaebbd2b57f9e72ca594991d
Author: Keith Whitwell <keith@tungstengraphics.com>
Date: Thu Nov 15 09:59:33 2007 +0000
i915tex: Catch cases where not all state is emitted for a new batchbuffer.
This could lead to incorrect rendering or even lockups.
commit 0967e1270d3b3bc0917fb841314ebb8d04766746
Author: Michel Dänzer <michel@tungstengraphics.com>
Date: Fri Nov 9 18:46:55 2007 +0100
i915tex: Some additional blit fixes and assertions.
commit a08c02f1a85823d7116758c3c0996776932fc530
Author: Michel Dänzer <michel@tungstengraphics.com>
Date: Fri Nov 9 18:44:11 2007 +0100
i915tex: Actually wait for previous commands to complete for glFinish().
commit bdfd9afeadd19c0bd71d904d9d04b01b74b9b70f
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 23 12:01:39 2007 -0700
Consolidate texture fetch code and use partial derivatives when possible.
commit 1c91a590d6b574a59063a1c276b060b60459565c
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 23 10:28:33 2007 -0700
fix a few GLSL bugs
commit 65bd7968bf78e77bfa1d30fe2c459bf6b135059c
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 23 10:26:02 2007 -0700
Fix parsing of gl_FrontLightModelProduct.sceneColor, don't segfault on variable
array indexes.
commit 6ccd23b87b9d07ce01bb84d406b64fb18d37a33e
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 23 09:14:27 2007 -0700
need to check border width in sample_linear_2d() - fixes failed assertion in texwrap.c test
commit 823409b7d09cab74db3715de5f1e1685fe0fc70d
Author: Brian <brian.paul@tungstengraphics.com>
Date: Thu Nov 22 09:36:13 2007 -0700
Bump version numbers to 7.0.3 for next release
commit c85d31f4a049193afe2452b4f130671b16fbc32c
Author: Brian <brian.paul@tungstengraphics.com>
Date: Thu Nov 22 09:33:57 2007 -0700
Recompute ctx->Point._Size if GL_POINT_SIZE_MIN/MAX changes.
Note that all the point size clamping stuff has been redone in Mesa/master.
commit 24af5c44da2b57707976728c9374f44b20b4efe4
Author: Roland Scheidegger <sroland@tungstengraphics.com>
Date: Thu Nov 22 02:55:25 2007 +0100
fix z buffer read/write issue with rv100-like chips and old ddx
commit 75efacf8eb725049525ed724482ffeace76dd2fd
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 20 08:16:25 2007 -0700
Initial 7.0.3 relnotes
commit 825e810247db3c210616648f74a96a6f2c42d9d8
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 20 08:15:19 2007 -0700
bring over Fortran fixes from master
commit e20723cfc18fee93c1a5f5ff346baa8694f53d09
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 20 08:13:14 2007 -0700
Obsolete
commit 9f39a67c1d71006f58be72091dd37ebe644738fd
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 20 08:13:03 2007 -0700
Rename glut_fbc.c -> glut_fcb.c (cb=callback)
commit 2044f3f791b0caf36d875691d10c627b3fda49bd
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Nov 19 13:05:10 2007 -0700
clamp lambda to Min/MaxLod
commit 30d4e8c422b52826f7a14c25103df9f68d59b172
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Nov 19 10:37:41 2007 -0700
fix some texture format assertions, etc
commit 92d552f6a2688d7fa410a88d3ed96a0144e34a6f
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Nov 19 09:54:41 2007 -0700
fix out-of-bounds array index (ix=-1)
commit 5adfcbbc4f8ab7438654bac3a423aab9522bb5a1
Author: Brian <brian.paul@tungstengraphics.com>
Date: Fri Nov 16 10:03:12 2007 -0700
added gl_dispatch_stub_772()
commit 69969b38197b847abaec5248f67775416adb499c
Author: Roland Scheidegger <sroland@tungstengraphics.com>
Date: Thu Nov 15 00:52:38 2007 +0100
fix position invariant vertex programs for sw-tnl
do the same math as for fixed function pipe, including
user clip planes.
(mostly resurrected from the dead t_vb_arbprogram.c code)
commit edfee04fcb1431aa079c69b7c0b67e87de4063c7
Author: Roland Scheidegger <sroland@tungstengraphics.com>
Date: Thu Nov 15 01:16:27 2007 +0100
fix bogus assumption if ddx has set up surface reg for z buffer
this is wrong since even if ddx has not set up a surface reg to cover the z
buffer we should pretend it has on those rv100 chips since they presumably do
not do z buffer tiling if not using hyperz, so we can use linear addressing
just the same. Doesn't seem to fix #13080, but it's wrong anyway and the bug
almost certainly broke newer non-tcl chips.
commit 28c9930888ffeb66274c27e9a5f980e115df6ee7
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 13 09:51:14 2007 -0700
add pointer to Gallium3D info
commit e3a35a123efc48e0a69898fd7bda396c9fbe3fff
Author: Brian <brian.paul@tungstengraphics.com>
Date: Tue Nov 13 09:50:42 2007 -0700
DRI memory manager info (fixes dangling link)
commit 412168f2e884d3dd8318c28405683c15c6912c93
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Nov 12 08:21:21 2007 -0700
remove dependency on libGLU
commit fc0fa0d636a7ca27735d580777d4d4844a515bb4
Author: Brian <brian.paul@tungstengraphics.com>
Date: Mon Nov 12 07:52:20 2007 -0700
add glw.pc.in to tarball list, remove from DEPEND_FILES
diff --git a/Makefile b/Makefile
index fe0c354..dcc60f6 100644
--- a/Makefile
+++ b/Makefile
@@ -166,10 +166,10 @@ ultrix-gcc:
# Rules for making release tarballs
-DIRECTORY = Mesa-7.0.2
-LIB_NAME = MesaLib-7.0.2
-DEMO_NAME = MesaDemos-7.0.2
-GLUT_NAME = MesaGLUT-7.0.2
+DIRECTORY = Mesa-7.0.3
+LIB_NAME = MesaLib-7.0.3
+DEMO_NAME = MesaDemos-7.0.3
+GLUT_NAME = MesaGLUT-7.0.3
MAIN_FILES = \
$(DIRECTORY)/Makefile* \
@@ -355,6 +355,7 @@ GLW_FILES = \
$(DIRECTORY)/src/glw/*.[ch] \
$(DIRECTORY)/src/glw/Makefile* \
$(DIRECTORY)/src/glw/README \
+ $(DIRECTORY)/src/glw/glw.pc.in \
$(DIRECTORY)/src/glw/depend
DEMO_FILES = \
@@ -424,7 +425,6 @@ DEPEND_FILES = \
$(TOP)/src/mesa/depend \
$(TOP)/src/glx/x11/depend \
$(TOP)/src/glw/depend \
- $(TOP)/src/glw/glw.pc.in \
$(TOP)/src/glut/glx/depend \
$(TOP)/src/glu/sgi/depend
diff --git a/bin/mklib b/bin/mklib
index 499e789..0fb9930 100755
--- a/bin/mklib
+++ b/bin/mklib
@@ -209,8 +209,13 @@ case $ARCH in
if [ $NOPREFIX = 1 ] ; then
# No "lib" or ".so" part
echo "mklib: Making" $ARCH "shared library: " ${LIBNAME}
- #OPTS="-shared -Wl,-soname,${LIBNAME}" # soname???
- OPTS="-shared"
+ case $ARCH in 'Linux' | 'GNU' | GNU/*)
+ OPTS="-Xlinker -Bsymbolic -shared"
+ ;;
+ *)
+ OPTS="-shared"
+ ;;
+ esac
# Check if objects are 32-bit and we're running in 64-bit
# environment. If so, pass -m32 flag to linker.
diff --git a/configs/linux-static b/configs/linux-static
index 2a77d4c..1ee16be 100644
--- a/configs/linux-static
+++ b/configs/linux-static
@@ -22,4 +22,5 @@ GLUT_LIB_DEPS =
GLW_LIB_DEPS =
# Need to specify all libraries we may need
-APP_LIB_DEPS = $(EXTRA_LIB_PATH) -lX11 -lXmu -lXt -lXi -lpthread -lstdc++ -lm
+APP_LIB_DEPS = $(EXTRA_LIB_PATH) -lX11 -lXext -lXmu -lXt -lXi -lpthread \
+ -lstdc++ -lm
diff --git a/configs/linux-x86-64-static b/configs/linux-x86-64-static
index 06e6fc3..611abf0 100644
--- a/configs/linux-x86-64-static
+++ b/configs/linux-x86-64-static
@@ -20,4 +20,5 @@ OSMESA_LIB_DEPS =
GLU_LIB_DEPS =
GLUT_LIB_DEPS =
GLW_LIB_DEPS =
-APP_LIB_DEPS = $(EXTRA_LIB_PATH) -lX11 -lXmu -lXt -lXi -lpthread -lstdc++ -lm
+APP_LIB_DEPS = $(EXTRA_LIB_PATH) -lX11 -lXext -lXmu -lXt -lXi -lpthread \
+ -lstdc++ -lm
diff --git a/configs/linux-x86-static b/configs/linux-x86-static
index 2b6478b..645196c 100644
--- a/configs/linux-x86-static
+++ b/configs/linux-x86-static
@@ -20,4 +20,5 @@ OSMESA_LIB_DEPS =
GLU_LIB_DEPS =
GLUT_LIB_DEPS =
GLW_LIB_DEPS =
-APP_LIB_DEPS = $(EXTRA_LIB_PATH) -lX11 -lXmu -lXt -lXi -lpthread -lstdc++ -lm
+APP_LIB_DEPS = $(EXTRA_LIB_PATH) -lX11 -lXext -lXmu -lXt -lXi -lpthread \
+ -lstdc++ -lm
diff --git a/docs/fbdev-dri.html b/docs/fbdev-dri.html
index 0d9e52c..c3724b3 100644
--- a/docs/fbdev-dri.html
+++ b/docs/fbdev-dri.html
@@ -34,6 +34,11 @@ in the normal Mesa releases so you'll need to get the latest sources
sources from the <a href="repository.html">git repository</a>.
</p>
+<p>
+This fbdev/DRI environment isn't well supported.
+Code and documentation updates/patches are welcomed.
+</p>
+
<h1>2. Compilation</h1>
@@ -66,6 +71,7 @@ You'll need fbdev header files. Check with:
</pre>
<p>
+You'll need to get Mesa from git (see above).
Compile Mesa with the 'linux-solo' configuration:
</p>
<pre>
diff --git a/docs/memory.html b/docs/memory.html
new file mode 100644
index 0000000..cd59bd7
--- /dev/null
+++ b/docs/memory.html
@@ -0,0 +1,21 @@
+<HTML>
+
+<TITLE>Mesa News</TITLE>
+
+<head><link rel="stylesheet" type="text/css" href="mesa.css"></head>
+
+<BODY>
+
+<body bgcolor="#eeeeee">
+
+<H1>DRI Memory Management</H1>
+
+<p>
+Thomas Hellström's
+<a href="http://www.tungstengraphics.com/mm.pdf">Memory Management
+whitepaper</a> describes the goals, design and implementation of the
+new DRI memory management system.
+</p>
+
+</body>
+</html>
diff --git a/docs/news.html b/docs/news.html
index 79fc599..13e43f1 100644
--- a/docs/news.html
+++ b/docs/news.html
@@ -11,6 +11,23 @@
<H1>News</H1>
+<h2>November 13, 2007</h2>
+
+<p>
+Gallium3D is the codename for the new Mesa device driver architecture
+which is currently under development.
+A <a href="http://www.tungstengraphics.com/wiki/index.php/Gallium3D"
+target="_parent"> summary</a> of the architecture can be found on the
+Tungsten Graphics website.
+</p>
+<p>
+Gallium3D development is taking place on the <em>gallium-0.1</em> branch
+of the git repository.
+Currently, there's only a software-only driver and an Intel i915/945 driver
+but other drivers will be coming...
+</p>
+
+
<h2>November 10, 2007</h2>
<p>
<a href="relnotes-7.0.2.html">Mesa 7.0.2</a> is released.
diff --git a/docs/relnotes-7.0.3.html b/docs/relnotes-7.0.3.html
new file mode 100644
index 0000000..59090da
--- /dev/null
+++ b/docs/relnotes-7.0.3.html
@@ -0,0 +1,60 @@
+<HTML>
+
+<TITLE>Mesa Release Notes</TITLE>
+
+<head><link rel="stylesheet" type="text/css" href="mesa.css"></head>
+
+<BODY>
+
+<body bgcolor="#eeeeee">
+
+<H1>Mesa 7.0.3 Release Notes / (TBD) 2007</H1>
+
+<p>
+Mesa 7.0.3 is a stable release with bug fixes since version 7.0.2.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+</pre>
+
+
+<h2>New features</h2>
+<ul>
+</ul>
+
+<h2>Bug fixes</h2>
+<ul>
+<li>Added missing glw.pc.in file to release tarball
+<li>Fix GLUT/Fortran issues
+<li>GLSL gl_FrontLightModelProduct.sceneColor variable wasn't defined
+<li>Fix crash upon GLSL variable array indexes (not yet supported)
+<li>Two-sided stencil test didn't work in software rendering
+<li>Fix two-sided lighting bugs/crashes (bug 13368)
+<li>GLSL gl_FrontFacing didn't work properly
+<li>glGetActiveUniform returned incorrect sizes (bug 13751)
+</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>
diff --git a/docs/relnotes.html b/docs/relnotes.html
index 7464f5c..8469c7f 100644
--- 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-7.0.3.html">7.0.3 release notes</A>
<LI><A HREF="relnotes-7.0.2.html">7.0.2 release notes</A>
<LI><A HREF="relnotes-7.0.1.html">7.0.1 release notes</A>
<LI><A HREF="relnotes-7.0.html">7.0 release notes</A>
diff --git a/docs/shading.html b/docs/shading.html
index 0e1a5e1..0a9f5f3 100644
--- a/docs/shading.html
+++ b/docs/shading.html
@@ -48,6 +48,7 @@ in Mesa:
<li>The inverse trig functions asin(), acos(), and atan() are not implemented
<li>The gl_Color and gl_SecondaryColor varying vars are interpolated
without perspective correction
+<li>Floating point literal suffixes 'f' and 'F' aren't allowed.
</ul>
<p>
{
diff --git a/progs/xdemos/offset.c b/progs/xdemos/offset.c
index 3e92e68..97170f5 100644
--- a/progs/xdemos/offset.c
+++ b/progs/xdemos/offset.c
@@ -47,7 +47,6 @@ PERFORMANCE OF THIS SOFTWARE.
#include <GL/glx.h>
-#include <GL/glu.h>
#include <X11/keysym.h>
#include <stdlib.h>
#include <stdio.h>
@@ -134,7 +133,7 @@ int main(int argc, char** argv) {
/* set up viewing parameters */
glMatrixMode(GL_PROJECTION);
- gluPerspective(20, 1, 0.1, 20);
+ glFrustum(-1, 1, -1, 1, 6, 20);
glMatrixMode(GL_MODELVIEW);
glTranslatef(0, 0, -15);
diff --git a/src/mesa/Makefile b/src/mesa/Makefile
index eea8eb9..d0c19f5 100644
--- a/src/mesa/Makefile
+++ b/src/mesa/Makefile
@@ -65,7 +65,7 @@ linux-solo: depend subdirs libmesa.a
# Stand-alone Mesa libGL, no built-in drivers (DirectFB)
libgl-core: $(CORE_OBJECTS)
- @ $(TOP)/bin/mklib -o $(GL_LIB) -linker '$(CC)' -ldflags '$(LDFLAGS) \
+ @ $(TOP)/bin/mklib -o $(GL_LIB) -linker '$(CC)' -ldflags '$(LDFLAGS)' \
-major $(MESA_MAJOR) -minor $(MESA_MINOR) -patch $(MESA_TINY) \
-install $(TOP)/$(LIB_DIR) $(MKLIB_OPTIONS) $(CORE_OBJECTS) \
$(GL_LIB_DEPS)
diff --git a/src/mesa/drivers/common/driverfuncs.c b/src/mesa/drivers/common/driverfuncs.c
index adf9aaf..e9b364e 100644
--- a/src/mesa/drivers/common/driverfuncs.c
+++ b/src/mesa/drivers/common/driverfuncs.c
@@ -112,6 +112,8 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
driver->DeleteTexture = _mesa_delete_texture_object;
driver->NewTextureImage = _mesa_new_texture_image;
driver->FreeTexImageData = _mesa_free_texture_image_data;
+ driver->MapTexture = NULL;
+ driver->UnmapTexture = NULL;
driver->TextureMemCpy = _mesa_memcpy;
driver->IsTextureResident = NULL;
driver->PrioritizeTexture = NULL;
diff --git a/src/mesa/drivers/dri/Makefile.template b/src/mesa/drivers/dri/Makefile.template
index 9e84dbc..aeeb267 100644
--- a/src/mesa/drivers/dri/Makefile.template
+++ b/src/mesa/drivers/dri/Makefile.template
@@ -50,7 +50,6 @@ SHARED_INCLUDES = \
-I$(TOP)/src/mesa/main \
-I$(TOP)/src/mesa/glapi \
-I$(TOP)/src/mesa/math \
- -I$(TOP)/src/mesa/transform \
-I$(TOP)/src/mesa/shader \
-I$(TOP)/src/mesa/swrast \
-I$(TOP)/src/mesa/swrast_setup \
diff --git a/src/mesa/drivers/dri/gamma/gamma_render.c b/src/mesa/drivers/dri/gamma/gamma_render.c
index 4b462f2..a8fba49 100644
--- a/src/mesa/drivers/dri/gamma/gamma_render.c
+++ b/src/mesa/drivers/dri/gamma/gamma_render.c
@@ -193,7 +193,7 @@ static GLboolean gamma_run_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/i810/i810render.c b/src/mesa/drivers/dri/i810/i810render.c
index a31d542..d022596 100644
--- a/src/mesa/drivers/dri/i810/i810render.c
+++ b/src/mesa/drivers/dri/i810/i810render.c
@@ -144,7 +144,7 @@ static GLboolean i810_run_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/i915/intel_render.c b/src/mesa/drivers/dri/i915/intel_render.c
index d9438ba..773779a 100644
--- a/src/mesa/drivers/dri/i915/intel_render.c
+++ b/src/mesa/drivers/dri/i915/intel_render.c
@@ -213,7 +213,7 @@ static GLboolean intel_run_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/i915tex/i830_vtbl.c b/src/mesa/drivers/dri/i915tex/i830_vtbl.c
index 5555b74..e76e2e7 100644
--- a/src/mesa/drivers/dri/i915tex/i830_vtbl.c
+++ b/src/mesa/drivers/dri/i915tex/i830_vtbl.c
@@ -378,9 +378,15 @@ do { \
} while (0)
static GLuint
+get_dirty(struct i830_hw_state *state)
+{
+ return state->active & ~state->emitted;
+}
+
+static GLuint
get_state_size(struct i830_hw_state *state)
{
- GLuint dirty = state->active & ~state->emitted;
+ GLuint dirty = get_dirty(state);
GLuint sz = 0;
GLuint i;
@@ -411,7 +417,7 @@ get_state_size(struct i830_hw_state *state)
/* Push the state into the sarea and/or texture memory.
*/
static void
-i830_emit_state(struct intel_context *intel)
+i830_do_emit_state(struct intel_context *intel)
{
struct i830_context *i830 = i830_context(&intel->ctx);
struct i830_hw_state *state = i830->current;
@@ -428,10 +434,32 @@ i830_emit_state(struct intel_context *intel)
*/
intel_batchbuffer_require_space(intel->batch, get_state_size(state), 0);
+ /* Workaround. There are cases I haven't been able to track down
+ * where we aren't emitting a full state at the start of a new
+ * batchbuffer. This code spots that we are on a new batchbuffer
+ * and forces a full state emit no matter what.
+ *
+ * In the normal case state->emitted is already zero, this code is
+ * another set of checks to make sure it really is.
+ */
+ if (intel->batch->id != intel->last_state_batch_id ||
+ intel->batch->map == intel->batch->ptr)
+ {
+ state->emitted = 0;
+ intel_batchbuffer_require_space(intel->batch, get_state_size(state), 0);
+ }
+
/* Do this here as we may have flushed the batchbuffer above,
* causing more state to be dirty!
*/
- dirty = state->active & ~state->emitted;
+ dirty = get_dirty(state);
+ state->emitted |= dirty;
+ assert(get_dirty(state) == 0);
+
+ if (intel->batch->id != intel->last_state_batch_id) {
+ assert(dirty & I830_UPLOAD_CTX);
+ intel->last_state_batch_id = intel->batch->id;
+ }
if (dirty & I830_UPLOAD_INVARIENT) {
DBG("I830_UPLOAD_INVARIENT:\n");
@@ -513,7 +541,30 @@ i830_emit_state(struct intel_context *intel)
}
}
- state->emitted |= dirty;
+ intel->batch->dirty_state &= ~dirty;
+ assert(get_dirty(state) == 0);
+}
+
+static void
+i830_emit_state(struct intel_context *intel)
+{
+ struct i830_context *i830 = i830_context(&intel->ctx);
+
+ i830_do_emit_state( intel );
+
+ /* Second chance - catch batchbuffer wrap in the middle of state
+ * emit. This shouldn't happen but it has been observed in
+ * testing.
+ */
+ if (get_dirty( i830->current )) {
+ /* Force a full re-emit if this happens.
+ */
+ i830->current->emitted = 0;
+ i830_do_emit_state( intel );
+ }
+
+ assert(get_dirty(i830->current) == 0);
+ assert((intel->batch->dirty_state & (1<<1)) == 0);
}
static void
@@ -650,8 +701,7 @@ i830_assert_not_dirty( struct intel_context *intel )
{
struct i830_context *i830 = i830_context(&intel->ctx);
struct i830_hw_state *state = i830->current;
- GLuint dirty = state->active & ~state->emitted;
- assert(!dirty);
+ assert(!get_dirty(state));
}
diff --git a/src/mesa/drivers/dri/i915tex/i915_vtbl.c b/src/mesa/drivers/dri/i915tex/i915_vtbl.c
index e911fc4..1b055f6 100644
--- a/src/mesa/drivers/dri/i915tex/i915_vtbl.c
+++ b/src/mesa/drivers/dri/i915tex/i915_vtbl.c
@@ -198,7 +198,7 @@ i915_emit_invarient_state(struct intel_context *intel)
/* Need to initialize this to zero.
*/
- OUT_BATCH(_3DSTATE_LOAD_STATE_IMMEDIATE_1 | I1_LOAD_S(3) | (1));
+ OUT_BATCH(_3DSTATE_LOAD_STATE_IMMEDIATE_1 | I1_LOAD_S(3) | (0));
OUT_BATCH(0);
/* XXX: Use this */
@@ -216,6 +216,7 @@ i915_emit_invarient_state(struct intel_context *intel)
/* Don't support twosided stencil yet */
OUT_BATCH(_3DSTATE_BACKFACE_STENCIL_OPS | BFO_ENABLE_STENCIL_TWO_SIDE | 0);
+ OUT_BATCH(0);
ADVANCE_BATCH();
}
@@ -247,6 +248,9 @@ get_state_size(struct i915_hw_state *state)
GLuint i;
GLuint sz = 0;
+ if (dirty & I915_UPLOAD_INVARIENT)
+ sz += 30 * 4;
+
if (dirty & I915_UPLOAD_CTX)
sz += sizeof(state->Ctx);
@@ -281,7 +285,7 @@ get_state_size(struct i915_hw_state *state)
/* Push the state into the sarea and/or texture memory.
*/
static void
-i915_emit_state(struct intel_context *intel)
+i915_do_emit_state(struct intel_context *intel)
{
struct i915_context *i915 = i915_context(&intel->ctx);
struct i915_hw_state *state = i915->current;
@@ -298,10 +302,33 @@ i915_emit_state(struct intel_context *intel)
*/
intel_batchbuffer_require_space(intel->batch, get_state_size(state), 0);
+
+ /* Workaround. There are cases I haven't been able to track down
+ * where we aren't emitting a full state at the start of a new
+ * batchbuffer. This code spots that we are on a new batchbuffer
+ * and forces a full state emit no matter what.
+ *
+ * In the normal case state->emitted is already zero, this code is
+ * another set of checks to make sure it really is.
+ */
+ if (intel->batch->id != intel->last_state_batch_id ||
+ intel->batch->map == intel->batch->ptr)
+ {
+ state->emitted = 0;
+ intel_batchbuffer_require_space(intel->batch, get_state_size(state), 0);
+ }
+
/* Do this here as we may have flushed the batchbuffer above,
* causing more state to be dirty!
*/
dirty = get_dirty(state);
+ state->emitted |= dirty;
+ assert(get_dirty(state) == 0);
+
+ if (intel->batch->id != intel->last_state_batch_id) {
+ assert(dirty & I915_UPLOAD_CTX);
+ intel->last_state_batch_id = intel->batch->id;
+ }
if (INTEL_DEBUG & DEBUG_STATE)
fprintf(stderr, "%s dirty: %x\n", __FUNCTION__, dirty);
@@ -424,7 +451,29 @@ i915_emit_state(struct intel_context *intel)
i915_disassemble_program(state->Program, state->ProgramSize);
}
- state->emitted |= dirty;
+ intel->batch->dirty_state &= ~dirty;
+}
+
+static void
+i915_emit_state(struct intel_context *intel)
+{
+ struct i915_context *i915 = i915_context(&intel->ctx);
+
+ i915_do_emit_state( intel );
+
+ /* Second chance - catch batchbuffer wrap in the middle of state
+ * emit. This shouldn't happen but it has been observed in
+ * testing.
+ */
+ if (get_dirty( i915->current )) {
+ /* Force a full re-emit if this happens.
+ */
+ i915->current->emitted = 0;
+ i915_do_emit_state( intel );
+ }
+
+ assert(get_dirty(i915->current) == 0);
+ assert((intel->batch->dirty_state & (1<<1)) == 0);
}
static void
diff --git a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.c b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.c
index c92b83b..d7079a2 100644
--- a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.c
@@ -118,6 +118,8 @@ intel_batchbuffer_reset(struct intel_batchbuffer *batch)
batch->map = driBOMap(batch->buffer, DRM_BO_FLAG_WRITE, 0);
batch->ptr = batch->map;
+ batch->dirty_state = ~0;
+ batch->id = batch->intel->intelScreen->batch_id++;
}
/*======================================================================
diff --git a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h
index 59261f7..76e5b30 100644
--- a/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h
+++ b/src/mesa/drivers/dri/i915tex/intel_batchbuffer.h
@@ -38,6 +38,9 @@ struct intel_batchbuffer
struct buffer_reloc reloc[MAX_RELOCS];
GLuint nr_relocs;
GLuint size;
+
+ GLuint dirty_state;
+ GLuint id;
};
struct intel_batchbuffer *intel_batchbuffer_alloc(struct intel_context
diff --git a/src/mesa/drivers/dri/i915tex/intel_blit.c b/src/mesa/drivers/dri/i915tex/intel_blit.c
index a060151..eef94c7 100644
--- a/src/mesa/drivers/dri/i915tex/intel_blit.c
+++ b/src/mesa/drivers/dri/i915tex/intel_blit.c
@@ -112,8 +112,8 @@ intelCopyBuffer(const __DRIdrawablePrivate * dPriv,
for (i = 0; i < nbox; i++, pbox++) {
drm_clip_rect_t box;
- if (pbox->x1 > pbox->x2 ||
- pbox->y1 > pbox->y2 ||
+ if (pbox->x1 >= pbox->x2 ||
+ pbox->y1 >= pbox->y2 ||
pbox->x2 > intelScreen->width || pbox->y2 > intelScreen->height)
continue;
@@ -129,19 +129,22 @@ intelCopyBuffer(const __DRIdrawablePrivate * dPriv,
if (rect->y2 < box.y2)
box.y2 = rect->y2;
- if (box.x1 > box.x2 || box.y1 > box.y2)
+ if (box.x1 >= box.x2 || box.y1 >= box.y2)
continue;
}
+ assert(box.x1 < box.x2);
+ assert(box.y1 < box.y2);
+
BEGIN_BATCH(8, INTEL_BATCH_NO_CLIPRECTS);
OUT_BATCH(CMD);
OUT_BATCH(BR13);
- OUT_BATCH((pbox->y1 << 16) | pbox->x1);
- OUT_BATCH((pbox->y2 << 16) | pbox->x2);
+ OUT_BATCH((box.y1 << 16) | box.x1);
+ OUT_BATCH((box.y2 << 16) | box.x2);
OUT_RELOC(frontRegion->buffer, DRM_BO_FLAG_MEM_TT | DRM_BO_FLAG_WRITE,
DRM_BO_MASK_MEM | DRM_BO_FLAG_WRITE, 0);
- OUT_BATCH((pbox->y1 << 16) | pbox->x1);
+ OUT_BATCH((box.y1 << 16) | box.x1);
OUT_BATCH(BR13 & 0xffff);
OUT_RELOC(backRegion->buffer, DRM_BO_FLAG_MEM_TT | DRM_BO_FLAG_READ,
DRM_BO_MASK_MEM | DRM_BO_FLAG_READ, 0);
@@ -152,7 +155,8 @@ intelCopyBuffer(const __DRIdrawablePrivate * dPriv,
if (intel->first_swap_fence)
driFenceUnReference(intel->first_swap_fence);
intel->first_swap_fence = intel_batchbuffer_flush(intel->batch);
- driFenceReference(intel->first_swap_fence);
+ if (intel->first_swap_fence)
+ driFenceReference(intel->first_swap_fence);
}
UNLOCK_HARDWARE(intel);
@@ -193,6 +197,8 @@ intelEmitFillBlit(struct intel_context *intel,
DBG("%s dst:buf(%p)/%d+%d %d,%d sz:%dx%d\n",
__FUNCTION__, dst_buffer, dst_pitch, dst_offset, x, y, w, h);
+ assert(w > 0);
+ assert(h > 0);
BEGIN_BATCH(6, INTEL_BATCH_NO_CLIPRECTS);
OUT_BATCH(CMD);
@@ -280,10 +286,11 @@ intelEmitCopyBlit(struct intel_context *intel,
return;
}
- if (dst_y2 < dst_y || dst_x2 < dst_x) {
+ if (dst_y2 <= dst_y || dst_x2 <= dst_x) {
return;
}
+
/* Initial y values don't seem to work with negative pitches. If
* we adjust the offsets manually (below), it seems to work fine.
*
@@ -292,6 +299,9 @@ intelEmitCopyBlit(struct intel_context *intel,
* the wrong result.
*/
if (dst_pitch > 0 && src_pitch > 0) {
+ assert(dst_x < dst_x2);
+ assert(dst_y < dst_y2);
+
BEGIN_BATCH(8, INTEL_BATCH_NO_CLIPRECTS);
OUT_BATCH(CMD);
OUT_BATCH(BR13);
@@ -306,6 +316,9 @@ intelEmitCopyBlit(struct intel_context *intel,
ADVANCE_BATCH();
}
else {
+ assert(dst_x < dst_x2);
+ assert(h > 0);
+
BEGIN_BATCH(8, INTEL_BATCH_NO_CLIPRECTS);
OUT_BATCH(CMD);
OUT_BATCH(BR13);
@@ -480,6 +493,9 @@ intelClearWithBlit(GLcontext * ctx, GLbitfield mask)
*/
intel_wait_flips(intel, INTEL_BATCH_NO_CLIPRECTS);
+ assert(b.x1 < b.x2);
+ assert(b.y1 < b.y2);
+
BEGIN_BATCH(6, INTEL_BATCH_NO_CLIPRECTS);
OUT_BATCH(CMD);
OUT_BATCH(BR13);
diff --git a/src/mesa/drivers/dri/i915tex/intel_buffers.c b/src/mesa/drivers/dri/i915tex/intel_buffers.c
index 78fb720..6778fdf 100644
--- a/src/mesa/drivers/dri/i915tex/intel_buffers.c
+++ b/src/mesa/drivers/dri/i915tex/intel_buffers.c
@@ -714,7 +714,8 @@ intel_wait_flips(struct intel_context *intel, GLuint batch_flags)
BUFFER_BIT_FRONT_LEFT ? BUFFER_FRONT_LEFT :
BUFFER_BACK_LEFT);
- if (intel_fb->Base.Name == 0 && intel_rb->pf_pending == intel_fb->pf_seq) {
+ if (intel_fb->Base.Name == 0 && intel_rb &&
+ intel_rb->pf_pending == intel_fb->pf_seq) {
GLint pf_pipes = intel_fb->pf_pipes;
BATCH_LOCALS;
@@ -1019,16 +1020,11 @@ intel_draw_buffer(GLcontext * ctx, struct gl_framebuffer *fb)
/*
* How many color buffers are we drawing into?
*/
- if (fb->_NumColorDrawBuffers[0] != 1
-#if 0
- /* XXX FBO temporary - always use software rendering */
- || 1
-#endif
- ) {
+ if (fb->_NumColorDrawBuffers[0] != 1) {
/* writing to 0 or 2 or 4 color buffers */
/*_mesa_debug(ctx, "Software rendering\n");*/
FALLBACK(intel, INTEL_FALLBACK_DRAW_BUFFER, GL_TRUE);
- front = 1; /* might not have back color buffer */
+ colorRegion = NULL;
}
else {
/* draw to exactly one color buffer */
@@ -1037,30 +1033,30 @@ intel_draw_buffer(GLcontext * ctx, struct gl_framebuffer *fb)
if (fb->_ColorDrawBufferMask[0] == BUFFER_BIT_FRONT_LEFT) {
front = 1;
}
- }
- /*
- * Get the intel_renderbuffer for the colorbuffer we're drawing into.
- * And set up cliprects.
- */
- if (fb->Name == 0) {
- /* drawing to window system buffer */
- if (front) {
- intelSetFrontClipRects(intel);
- colorRegion = intel_get_rb_region(fb, BUFFER_FRONT_LEFT);
+ /*
+ * Get the intel_renderbuffer for the colorbuffer we're drawing into.
+ * And set up cliprects.
+ */
+ if (fb->Name == 0) {
+ /* drawing to window system buffer */
+ if (front) {
+ intelSetFrontClipRects(intel);
+ colorRegion = intel_get_rb_region(fb, BUFFER_FRONT_LEFT);
+ }
+ else {
+ intelSetBackClipRects(intel);
+ colorRegion = intel_get_rb_region(fb, BUFFER_BACK_LEFT);
+ }
}
else {
- intelSetBackClipRects(intel);
- colorRegion = intel_get_rb_region(fb, BUFFER_BACK_LEFT);
+ /* drawing to user-created FBO */
+ struct intel_renderbuffer *irb;
+ intelSetRenderbufferClipRects(intel);
+ irb = intel_renderbuffer(fb->_ColorDrawBuffers[0][0]);
+ colorRegion = (irb && irb->region) ? irb->region : NULL;
}
}
- else {
- /* drawing to user-created FBO */
- struct intel_renderbuffer *irb;
- intelSetRenderbufferClipRects(intel);
- irb = intel_renderbuffer(fb->_ColorDrawBuffers[0][0]);
- colorRegion = (irb && irb->region) ? irb->region : NULL;
- }
/* Update culling direction which changes depending on the
* orientation of the buffer:
diff --git a/src/mesa/drivers/dri/i915tex/intel_context.c b/src/mesa/drivers/dri/i915tex/intel_context.c
index 40ea756..722ef28 100644
--- a/src/mesa/drivers/dri/i915tex/intel_context.c
+++ b/src/mesa/drivers/dri/i915tex/intel_context.c
@@ -315,7 +315,7 @@ intelFinish(GLcontext * ctx)
intelFlush(ctx);
if (intel->batch->last_fence) {
driFenceFinish(intel->batch->last_fence,
- 0, GL_FALSE);
+ DRM_FENCE_TYPE_EXE | DRM_I915_FENCE_TYPE_RW, GL_FALSE);
driFenceUnReference(intel->batch->last_fence);
intel->batch->last_fence = NULL;
}
diff --git a/src/mesa/drivers/dri/i915tex/intel_context.h b/src/mesa/drivers/dri/i915tex/intel_context.h
index 5fc8eb3..6f893bd 100644
--- a/src/mesa/drivers/dri/i915tex/intel_context.h
+++ b/src/mesa/drivers/dri/i915tex/intel_context.h
@@ -189,6 +189,7 @@ struct intel_context
struct _DriFenceObject *first_swap_fence;
struct intel_batchbuffer *batch;
+ GLuint last_state_batch_id;
struct
{
diff --git a/src/mesa/drivers/dri/i915tex/intel_render.c b/src/mesa/drivers/dri/i915tex/intel_render.c
index f9fa550..3162462 100644
--- a/src/mesa/drivers/dri/i915tex/intel_render.c
+++ b/src/mesa/drivers/dri/i915tex/intel_render.c
@@ -214,7 +214,7 @@ intel_run_render(GLcontext * ctx, struct tnl_pipeline_stage *stage)
tnl->Driver.Render.Start(ctx);
for (i = 0; i < VB->PrimitiveCount; i++) {
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/i915tex/intel_screen.h b/src/mesa/drivers/dri/i915tex/intel_screen.h
index bac43aa..760a1eb 100644
--- a/src/mesa/drivers/dri/i915tex/intel_screen.h
+++ b/src/mesa/drivers/dri/i915tex/intel_screen.h
@@ -95,6 +95,7 @@ typedef struct
struct _DriBufferPool *regionPool;
struct _DriBufferPool *staticPool;
unsigned int maxBatchSize;
+ unsigned batch_id;
GLboolean havePools;
} intelScreenPrivate;
diff --git a/src/mesa/drivers/dri/i915tex/intel_tris.c b/src/mesa/drivers/dri/i915tex/intel_tris.c
index 5fe3d45..06dd505 100644
--- a/src/mesa/drivers/dri/i915tex/intel_tris.c
+++ b/src/mesa/drivers/dri/i915tex/intel_tris.c
@@ -111,6 +111,9 @@ intelStartInlinePrimitive(struct intel_context *intel,
BEGIN_BATCH(2, batch_flags);
OUT_BATCH(0);
+ assert(intel->batch->id == intel->last_state_batch_id);
+ assert((intel->batch->dirty_state & (1<<1)) == 0);
+
intel->prim.start_ptr = intel->batch->ptr;
intel->prim.primitive = prim;
intel->prim.flush = intel_flush_inline_primitive;
diff --git a/src/mesa/drivers/dri/i965/brw_sf.c b/src/mesa/drivers/dri/i965/brw_sf.c
index d517539..df18a48 100644
--- a/src/mesa/drivers/dri/i965/brw_sf.c
+++ b/src/mesa/drivers/dri/i965/brw_sf.c
@@ -82,15 +82,15 @@ static void compile_sf_prog( struct brw_context *brw,
switch (key->primitive) {
case SF_TRIANGLES:
c.nr_verts = 3;
- brw_emit_tri_setup( &c );
+ brw_emit_tri_setup( &c, GL_TRUE );
break;
case SF_LINES:
c.nr_verts = 2;
- brw_emit_line_setup( &c );
+ brw_emit_line_setup( &c, GL_TRUE );
break;
case SF_POINTS:
c.nr_verts = 1;
- brw_emit_point_setup( &c );
+ brw_emit_point_setup( &c, GL_TRUE );
break;
case SF_UNFILLED_TRIS:
c.nr_verts = 3;
diff --git a/src/mesa/drivers/dri/i965/brw_sf.h b/src/mesa/drivers/dri/i965/brw_sf.h
index fb72b84..b321cda 100644
--- a/src/mesa/drivers/dri/i965/brw_sf.h
+++ b/src/mesa/drivers/dri/i965/brw_sf.h
@@ -97,9 +97,9 @@ struct brw_sf_compile {
};
-void brw_emit_tri_setup( struct brw_sf_compile *c );
-void brw_emit_line_setup( struct brw_sf_compile *c );
-void brw_emit_point_setup( struct brw_sf_compile *c );
+void brw_emit_tri_setup( struct brw_sf_compile *c, GLboolean allocate );
+void brw_emit_line_setup( struct brw_sf_compile *c, GLboolean allocate );
+void brw_emit_point_setup( struct brw_sf_compile *c, GLboolean allocate );
void brw_emit_anyprim_setup( struct brw_sf_compile *c );
#endif
diff --git a/src/mesa/drivers/dri/i965/brw_sf_emit.c b/src/mesa/drivers/dri/i965/brw_sf_emit.c
index 2ea9816..94be815 100644
--- a/src/mesa/drivers/dri/i965/brw_sf_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_sf_emit.c
@@ -343,13 +343,16 @@ static GLboolean calculate_masks( struct brw_sf_compile *c,
-void brw_emit_tri_setup( struct brw_sf_compile *c )
+void brw_emit_tri_setup( struct brw_sf_compile *c, GLboolean allocate)
{
struct brw_compile *p = &c->func;
GLuint i;
c->nr_verts = 3;
- alloc_regs(c);
+
+ if (allocate)
+ alloc_regs(c);
+
invert_det(c);
copy_z_inv_w(c);
@@ -428,14 +431,17 @@ void brw_emit_tri_setup( struct brw_sf_compile *c )
-void brw_emit_line_setup( struct brw_sf_compile *c )
+void brw_emit_line_setup( struct brw_sf_compile *c, GLboolean allocate)
{
struct brw_compile *p = &c->func;
GLuint i;
c->nr_verts = 2;
- alloc_regs(c);
+
+ if (allocate)
+ alloc_regs(c);
+
invert_det(c);
copy_z_inv_w(c);
@@ -497,17 +503,19 @@ void brw_emit_line_setup( struct brw_sf_compile *c )
}
}
-
/* Points setup - several simplifications as all attributes are
* constant across the face of the point (point sprites excluded!)
*/
-void brw_emit_point_setup( struct brw_sf_compile *c )
+void brw_emit_point_setup( struct brw_sf_compile *c, GLboolean allocate)
{
struct brw_compile *p = &c->func;
GLuint i;
c->nr_verts = 1;
- alloc_regs(c);
+
+ if (allocate)
+ alloc_regs(c);
+
copy_z_inv_w(c);
brw_MOV(p, c->m1Cx, brw_imm_ud(0)); /* zero - move out of loop */
@@ -564,7 +572,10 @@ void brw_emit_anyprim_setup( struct brw_sf_compile *c )
struct brw_reg primmask;
struct brw_instruction *jmp;
struct brw_reg v1_null_ud = vec1(retype(brw_null_reg(), BRW_REGISTER_TYPE_UD));
+
+ GLuint saveflag;
+ c->nr_verts = 3;
alloc_regs(c);
primmask = retype(get_element(c->tmp, 0), BRW_REGISTER_TYPE_UD);
@@ -582,10 +593,15 @@ void brw_emit_anyprim_setup( struct brw_sf_compile *c )
(1<<_3DPRIM_TRIFAN_NOSTIPPLE)));
jmp = brw_JMPI(p, ip, ip, brw_imm_w(0));
{
- brw_push_insn_state(p);
- brw_emit_tri_setup( c );
+ saveflag = p->flag_value;
+ brw_push_insn_state(p);
+ brw_emit_tri_setup( c, GL_FALSE );
brw_pop_insn_state(p);
- /* note - thread killed in subroutine */
+ p->flag_value = saveflag;
+ /* note - thread killed in subroutine, so must
+ * restore the flag which is changed when building
+ * the subroutine. fix #13240
+ */
}
brw_land_fwd_jump(p, jmp);
@@ -598,14 +614,16 @@ void brw_emit_anyprim_setup( struct brw_sf_compile *c )
(1<<_3DPRIM_LINESTRIP_CONT_BF)));
jmp = brw_JMPI(p, ip, ip, brw_imm_w(0));
{
+ saveflag = p->flag_value;
brw_push_insn_state(p);
- brw_emit_line_setup( c );
+ brw_emit_line_setup( c, GL_FALSE );
brw_pop_insn_state(p);
+ p->flag_value = saveflag;
/* note - thread killed in subroutine */
}
brw_land_fwd_jump(p, jmp);
- brw_emit_point_setup( c );
+ brw_emit_point_setup( c, GL_FALSE );
}
diff --git a/src/mesa/drivers/dri/i965/brw_wm_emit.c b/src/mesa/drivers/dri/i965/brw_wm_emit.c
index fd60515..80bd576 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_emit.c
@@ -777,6 +777,7 @@ static void emit_kil( struct brw_wm_compile *c,
brw_push_insn_state(p);
brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_GE, arg0[i], brw_imm_f(0));
brw_set_predicate_control_flag_value(p, 0xff);
+ brw_set_compression_control(p, BRW_COMPRESSION_NONE);
brw_AND(p, r0uw, brw_flag_reg(), r0uw);
brw_pop_insn_state(p);
}
diff --git a/src/mesa/drivers/dri/mga/mgarender.c b/src/mesa/drivers/dri/mga/mgarender.c
index 3080cea..c151f79 100644
--- a/src/mesa/drivers/dri/mga/mgarender.c
+++ b/src/mesa/drivers/dri/mga/mgarender.c
@@ -144,7 +144,7 @@ static GLboolean mga_run_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/r200/r200_tcl.c b/src/mesa/drivers/dri/r200/r200_tcl.c
index e0c32b2..1ff0cf9 100644
--- a/src/mesa/drivers/dri/r200/r200_tcl.c
+++ b/src/mesa/drivers/dri/r200/r200_tcl.c
@@ -489,7 +489,7 @@ static GLboolean r200_run_tcl_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/r200/r200_texstate.c b/src/mesa/drivers/dri/r200/r200_texstate.c
index af5092b..d12c3bc 100644
--- a/src/mesa/drivers/dri/r200/r200_texstate.c
+++ b/src/mesa/drivers/dri/r200/r200_texstate.c
@@ -72,7 +72,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
#define _INVALID(f) \
[ MESA_FORMAT_ ## f ] = { 0xffffffff, 0 }
#define VALID_FORMAT(f) ( ((f) <= MESA_FORMAT_RGBA_DXT5) \
- && (tx_table_le[f].format != 0xffffffff) )
+ && (tx_table_be[f].format != 0xffffffff) )
struct tx_table {
GLuint format, filter;
@@ -165,12 +165,15 @@ static void r200SetTexImages( r200ContextPtr rmesa,
*/
if ( !t->image_override ) {
if ( VALID_FORMAT( baseImage->TexFormat->MesaFormat ) ) {
+ const struct tx_table *table = _mesa_little_endian() ? tx_table_le :
+ tx_table_be;
+
t->pp_txformat &= ~(R200_TXFORMAT_FORMAT_MASK |
R200_TXFORMAT_ALPHA_IN_MAP);
t->pp_txfilter &= ~R200_YUV_TO_RGB;
- t->pp_txformat |= tx_table_le[ baseImage->TexFormat->MesaFormat ].format;
- t->pp_txfilter |= tx_table_le[ baseImage->TexFormat->MesaFormat ].filter;
+ t->pp_txformat |= table[ baseImage->TexFormat->MesaFormat ].format;
+ t->pp_txfilter |= table[ baseImage->TexFormat->MesaFormat ].filter;
}
else {
_mesa_problem(NULL, "unexpected texture format in %s", __FUNCTION__);
diff --git a/src/mesa/drivers/dri/radeon/radeon_screen.c b/src/mesa/drivers/dri/radeon/radeon_screen.c
index 682cf3a..e5d48d4 100644
--- a/src/mesa/drivers/dri/radeon/radeon_screen.c
+++ b/src/mesa/drivers/dri/radeon/radeon_screen.c
@@ -702,8 +702,11 @@ radeonCreateScreen( __DRIscreenPrivate *sPriv )
screen->depthPitch = dri_priv->depthPitch;
/* Check if ddx has set up a surface reg to cover depth buffer */
- screen->depthHasSurface = ((sPriv->ddxMajor > 4) &&
- (screen->chip_flags & RADEON_CHIPSET_TCL));
+ screen->depthHasSurface = (sPriv->ddxMajor > 4) ||
+ /* these chips don't use tiled z without hyperz. So always pretend
+ we have set up a surface which will cause linear reads/writes */
+ ((screen->chip_family & RADEON_CLASS_R100) &&
+ !(screen->chip_flags & RADEON_CHIPSET_TCL));
if ( dri_priv->textureSize == 0 ) {
screen->texOffset[RADEON_LOCAL_TEX_HEAP] = screen->gart_texture_offset;
diff --git a/src/mesa/drivers/dri/radeon/radeon_tcl.c b/src/mesa/drivers/dri/radeon/radeon_tcl.c
index 5ad044c..0f4baf2 100644
--- a/src/mesa/drivers/dri/radeon/radeon_tcl.c
+++ b/src/mesa/drivers/dri/radeon/radeon_tcl.c
@@ -418,7 +418,7 @@ static GLboolean radeon_run_tcl_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/s3v/s3v_render.c b/src/mesa/drivers/dri/s3v/s3v_render.c
index 6aaa949..6b86e0e 100644
--- a/src/mesa/drivers/dri/s3v/s3v_render.c
+++ b/src/mesa/drivers/dri/s3v/s3v_render.c
@@ -170,7 +170,7 @@ static GLboolean s3v_run_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++ )
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/savage/savagerender.c b/src/mesa/drivers/dri/savage/savagerender.c
index 514434c..f2f1998 100644
--- a/src/mesa/drivers/dri/savage/savagerender.c
+++ b/src/mesa/drivers/dri/savage/savagerender.c
@@ -198,7 +198,7 @@ static GLboolean savage_run_render( GLcontext *ctx,
for (i = 0 ; i < VB->PrimitiveCount ; i++)
{
- GLuint prim = VB->Primitive[i].mode;
+ GLuint prim = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
diff --git a/src/mesa/drivers/dri/unichrome/via_render.c b/src/mesa/drivers/dri/unichrome/via_render.c
index 387473e..d7e373d 100644
--- a/src/mesa/drivers/dri/unichrome/via_render.c
+++ b/src/mesa/drivers/dri/unichrome/via_render.c
@@ -106,7 +106,7 @@ static GLboolean via_run_fastrender(GLcontext *ctx,
tnl->clipspace.new_inputs |= VERT_BIT_POS;
for (i = 0; i < VB->PrimitiveCount; ++i) {
- GLuint mode = VB->Primitive[i].mode;
+ GLuint mode = _tnl_translate_prim(&VB->Primitive[i]);
GLuint start = VB->Primitive[i].start;
GLuint length = VB->Primitive[i].count;
if (length)
diff --git a/src/mesa/drivers/windows/gdi/wgl.c b/src/mesa/drivers/windows/gdi/wgl.c
index dad3dc1..04e3497 100644
--- a/src/mesa/drivers/windows/gdi/wgl.c
+++ b/src/mesa/drivers/windows/gdi/wgl.c
@@ -55,7 +55,7 @@
#include <windows.h>
#endif
-
+#include "config.h"
#include "glapi.h"
#include "GL/wmesa.h" /* protos for wmesa* functions */
@@ -70,10 +70,12 @@ struct __pixelformat__
GLboolean doubleBuffered;
};
+
+
/* These are the PFD's supported by this driver. */
struct __pixelformat__ pfd[] =
{
-#if 0
+#if 0
/* Double Buffer, alpha */
{
{
@@ -87,7 +89,7 @@ struct __pixelformat__ pfd[] =
8, 16,
8, 24,
0, 0, 0, 0, 0,
- 16, 8,
+ DEFAULT_SOFTWARE_DEPTH_BITS, 8,
0, 0, 0,
0, 0, 0
},
@@ -106,13 +108,13 @@ struct __pixelformat__ pfd[] =
8, 16,
8, 24,
0, 0, 0, 0, 0,
- 16, 8,
+ DEFAULT_SOFTWARE_DEPTH_BITS, 8,
0, 0, 0,
0, 0, 0
},
GL_FALSE
},
-#endif
+#endif
/* Double Buffer, no alpha */
{
{
@@ -126,7 +128,7 @@ struct __pixelformat__ pfd[] =
8, 16,
0, 0,
0, 0, 0, 0, 0,
- 16, 8,
+ DEFAULT_SOFTWARE_DEPTH_BITS, 8,
0, 0, 0,
0, 0, 0
},
@@ -145,7 +147,7 @@ struct __pixelformat__ pfd[] =
8, 16,
0, 0,
0, 0, 0, 0, 0,
- 16, 8,
+ DEFAULT_SOFTWARE_DEPTH_BITS, 8,
0, 0, 0,
0, 0, 0
},
diff --git a/src/mesa/drivers/windows/gdi/wmesa.c b/src/mesa/drivers/windows/gdi/wmesa.c
index c1d26a3..86c2663 100644
--- a/src/mesa/drivers/windows/gdi/wmesa.c
+++ b/src/mesa/drivers/windows/gdi/wmesa.c
@@ -1533,5 +1533,6 @@ void gl_dispatch_stub_749(void){}
void gl_dispatch_stub_769(void){}
void gl_dispatch_stub_770(void){}
void gl_dispatch_stub_771(void){}
+void gl_dispatch_stub_772(void){}
#endif
diff --git a/src/mesa/main/config.h b/src/mesa/main/config.h
index 24c2567..0d9c981 100644
--- a/src/mesa/main/config.h
+++ b/src/mesa/main/config.h
@@ -209,7 +209,7 @@
/** For GL_ARB_vertex_shader */
/*@{*/
#define MAX_VERTEX_ATTRIBS 16
-#define MAX_VERTEX_TEXTURE_IMAGE_UNITS 0
+#define MAX_VERTEX_TEXTURE_IMAGE_UNITS MAX_TEXTURE_UNITS
#define MAX_COMBINED_TEXTURE_IMAGE_UNITS (MAX_TEXTURE_IMAGE_UNITS + MAX_VERTEX_TEXTURE_IMAGE_UNITS)
/*@}*/
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index 754b1a7..b599638 100644
--- a/src/mesa/main/context.c
+++ b/src/mesa/main/context.c
@@ -977,7 +977,6 @@ init_attrib_groups(GLcontext *ctx)
/* Miscellaneous */
ctx->NewState = _NEW_ALL;
ctx->ErrorValue = (GLenum) GL_NO_ERROR;
- ctx->_Facing = 0;
return GL_TRUE;
}
diff --git a/src/mesa/main/context.h b/src/mesa/main/context.h
index b6a9d13..7a9d97c 100644
--- a/src/mesa/main/context.h
+++ b/src/mesa/main/context.h
@@ -272,10 +272,10 @@ do { \
(((CTX)->Light.Enabled && \
(CTX)->Light.Model.ColorControl == GL_SEPARATE_SPECULAR_COLOR) \
|| (CTX)->Fog.ColorSumEnabled \
- || ((CTX)->VertexProgram._Enabled && \
- ((CTX)->VertexProgram.Current->Base.InputsRead & VERT_BIT_COLOR1)) \
- || ((CTX)->FragmentProgram._Enabled && \
- ((CTX)->FragmentProgram.Current->Base.InputsRead & FRAG_BIT_COL1)) \
+ || ((CTX)->VertexProgram._Current && \
+ ((CTX)->VertexProgram._Current->Base.InputsRead & VERT_BIT_COLOR1)) \
+ || ((CTX)->FragmentProgram._Current && \
+ ((CTX)->FragmentProgram._Current->Base.InputsRead & FRAG_BIT_COL1)) \
)
diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h
index 88f3394..bb47cd4 100644
--- a/src/mesa/main/dd.h
+++ b/src/mesa/main/dd.h
@@ -493,6 +493,11 @@ struct dd_function_table {
*/
void (*FreeTexImageData)( GLcontext *ctx, struct gl_texture_image *tImage );
+ /** Map texture image data into user space */
+ void (*MapTexture)( GLcontext *ctx, struct gl_texture_object *tObj );
+ /** Unmap texture images from user space */
+ void (*UnmapTexture)( GLcontext *ctx, struct gl_texture_object *tObj );
+
/**
* Note: no context argument. This function doesn't initially look
* like it belongs here, except that the driver is the only entity
diff --git a/src/mesa/main/enable.c b/src/mesa/main/enable.c
index 1c6167b..0c2bf8c 100644
--- a/src/mesa/main/enable.c
+++ b/src/mesa/main/enable.c
@@ -5,9 +5,9 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.1
+ * Version: 7.0.3
*
- * Copyright (C) 1999-2006 Brian Paul All Rights Reserved.
+ * Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -364,12 +364,12 @@ _mesa_set_enable(GLcontext *ctx, GLenum cap, GLboolean state)
case GL_LIGHTING:
if (ctx->Light.Enabled == state)
return;
+ FLUSH_VERTICES(ctx, _NEW_LIGHT);
+ ctx->Light.Enabled = state;
if (ctx->Light.Enabled && ctx->Light.Model.TwoSide)
ctx->_TriangleCaps |= DD_TRI_LIGHT_TWOSIDE;
else
ctx->_TriangleCaps &= ~DD_TRI_LIGHT_TWOSIDE;
- FLUSH_VERTICES(ctx, _NEW_LIGHT);
- ctx->Light.Enabled = state;
break;
case GL_LINE_SMOOTH:
if (ctx->Line.SmoothFlag == state)
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index d00f2ef..96c3515 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -7,7 +7,7 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -1063,7 +1063,7 @@ struct gl_point_attrib
{
GLboolean SmoothFlag; /**< True if GL_POINT_SMOOTH is enabled */
GLfloat Size; /**< User-specified point size */
- GLfloat _Size; /**< Size clamped to Const.Min/MaxPointSize */
+ GLfloat _Size; /**< Size clamped to user limits */
GLfloat Params[3]; /**< GL_EXT_point_parameters */
GLfloat MinSize, MaxSize; /**< GL_EXT_point_parameters */
GLfloat Threshold; /**< GL_EXT_point_parameters */
@@ -3032,12 +3032,6 @@ struct __GLcontextRec
struct gl_list_extensions ListExt; /**< driver dlist extensions */
-
- GLuint _Facing; /**< This is a hack for 2-sided stencil test.
- *
- * We don't have a better way to communicate this value from
- * swrast_setup to swrast. */
-
/** \name For debugging/development only */
/*@{*/
GLboolean FirstTimeCurrent;
diff --git a/src/mesa/main/points.c b/src/mesa/main/points.c
index 8825bb1..5fa905f 100644
--- a/src/mesa/main/points.c
+++ b/src/mesa/main/points.c
@@ -5,7 +5,7 @@
/*
* Mesa 3-D graphics library
- * Version: 7.0.1
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -57,6 +57,7 @@ _mesa_PointSize( GLfloat size )
FLUSH_VERTICES(ctx, _NEW_POINT);
ctx->Point.Size = size;
+ /* _Size is only used for non-attenuated path */
ctx->Point._Size = CLAMP(ctx->Point.Size,
ctx->Point.MinSize,
ctx->Point.MaxSize);
@@ -150,6 +151,10 @@ _mesa_PointParameterfvEXT( GLenum pname, const GLfloat *params)
return;
FLUSH_VERTICES(ctx, _NEW_POINT);
ctx->Point.MinSize = params[0];
+ /* re-clamp _Size */
+ ctx->Point._Size = CLAMP(ctx->Point.Size,
+ ctx->Point.MinSize,
+ ctx->Point.MaxSize);
}
else {
_mesa_error(ctx, GL_INVALID_ENUM,
@@ -168,6 +173,10 @@ _mesa_PointParameterfvEXT( GLenum pname, const GLfloat *params)
return;
FLUSH_VERTICES(ctx, _NEW_POINT);
ctx->Point.MaxSize = params[0];
+ /* re-clamp _Size */
+ ctx->Point._Size = CLAMP(ctx->Point.Size,
+ ctx->Point.MinSize,
+ ctx->Point.MaxSize);
}
else {
_mesa_error(ctx, GL_INVALID_ENUM,
diff --git a/src/mesa/main/version.h b/src/mesa/main/version.h
index 2d3c68b..5e2747a 100644
--- a/src/mesa/main/version.h
+++ b/src/mesa/main/version.h
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 7.0.2
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -30,8 +30,8 @@
/* Mesa version */
#define MESA_MAJOR 7
#define MESA_MINOR 0
-#define MESA_PATCH 2
-#define MESA_VERSION_STRING "7.0.2"
+#define MESA_PATCH 3
+#define MESA_VERSION_STRING "7.0.3"
/* To make version comparison easy */
#define MESA_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c))
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index 6a87a17..e787524 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -1759,7 +1759,7 @@ parse_param_elements (GLcontext * ctx, const GLubyte ** inst,
{
GLint idx;
GLuint err = 0;
- gl_state_index state_tokens[STATE_LENGTH];
+ gl_state_index state_tokens[STATE_LENGTH] = {0, 0, 0, 0, 0};
GLfloat const_values[4];
switch (*(*inst)++) {
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index 9faf9d8..f4a12af 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -46,9 +46,6 @@
#include "slang_library_noise.h"
-/* See comments below for info about this */
-#define LAMBDA_ZERO 1
-
/* debug predicate */
#define DEBUG_PROG 0
@@ -303,6 +300,36 @@ fetch_vector1(const struct prog_src_register *source,
/**
+ * Fetch texel from texture. Use partial derivatives when possible.
+ */
+static INLINE void
+fetch_texel(GLcontext *ctx,
+ const struct gl_program_machine *machine,
+ const struct prog_instruction *inst,
+ const GLfloat texcoord[4], GLfloat lodBias,
+ GLfloat color[4])
+{
+ /* Note: we only have the right derivatives for fragment input attribs.
+ */
+ if (machine->NumDeriv > 0 &&
+ inst->SrcReg[0].File == PROGRAM_INPUT &&
+ inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit) {
+ /* simple texture fetch for which we should have derivatives */
+ GLuint attr = inst->SrcReg[0].Index;
+ machine->FetchTexelDeriv(ctx, texcoord,
+ machine->DerivX[attr],
+ machine->DerivY[attr],
+ lodBias,
+ inst->TexSrcUnit, color);
+ }
+ else {
+ machine->FetchTexelLod(ctx, texcoord, lodBias,
+ inst->TexSrcUnit, color);
+ }
+}
+
+
+/**
* Test value against zero and return GT, LT, EQ or UN if NaN.
*/
static INLINE GLuint
@@ -1306,33 +1333,18 @@ _mesa_execute_program(GLcontext * ctx,
}
break;
case OPCODE_TEX: /* Both ARB and NV frag prog */
- /* Texel lookup */
+ /* Simple texel lookup */
{
- /* Note: only use the precomputed lambda value when we're
- * sampling texture unit [K] with texcoord[K].
- * Otherwise, the lambda value may have no relation to the
- * instruction's texcoord or texture image. Using the wrong
- * lambda is usually bad news.
- * The rest of the time, just use zero (until we get a more
- * sophisticated way of computing lambda).
- */
- GLfloat coord[4], color[4], lambda;
-#if 0
- if (inst->SrcReg[0].File == PROGRAM_INPUT &&
- inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
- lambda = span->array->lambda[inst->TexSrcUnit][column];
- else
-#endif
- lambda = 0.0;
- fetch_vector4(&inst->SrcReg[0], machine, coord);
- machine->FetchTexelLod(ctx, coord, lambda, inst->TexSrcUnit,
- color);
+ GLfloat texcoord[4], color[4];
+ fetch_vector4(&inst->SrcReg[0], machine, texcoord);
+
+ fetch_texel(ctx, machine, inst, texcoord, 0.0, color);
+
if (DEBUG_PROG) {
- printf("TEX (%g, %g, %g, %g) = texture[%d][%g, %g, %g, %g], "
- "lod %f\n",
+ printf("TEX (%g, %g, %g, %g) = texture[%d][%g, %g, %g, %g]\n",
color[0], color[1], color[2], color[3],
inst->TexSrcUnit,
- coord[0], coord[1], coord[2], coord[3], lambda);
+ texcoord[0], texcoord[1], texcoord[2], texcoord[3]);
}
store_vector4(inst, machine, color);
}
@@ -1342,21 +1354,18 @@ _mesa_execute_program(GLcontext * ctx,
{
const struct gl_texture_unit *texUnit
= &ctx->Texture.Unit[inst->TexSrcUnit];
- GLfloat coord[4], color[4], lambda, bias;
-#if 0
- if (inst->SrcReg[0].File == PROGRAM_INPUT &&
- inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
- lambda = span->array->lambda[inst->TexSrcUnit][column];
- else
-#endif
- lambda = 0.0;
- fetch_vector4(&inst->SrcReg[0], machine, coord);
- /* coord[3] is the bias to add to lambda */
- bias = texUnit->LodBias + coord[3];
- if (texUnit->_Current)
- bias += texUnit->_Current->LodBias;
- machine->FetchTexelLod(ctx, coord, lambda + bias,
- inst->TexSrcUnit, color);
+ GLfloat texcoord[4], color[4], lodBias;
+
+ fetch_vector4(&inst->SrcReg[0], machine, texcoord);
+
+ /* texcoord[3] is the bias to add to lambda */
+ lodBias = texUnit->LodBias + texcoord[3];
+ if (texUnit->_Current) {
+ lodBias += texUnit->_Current->LodBias;
+ }
+
+ fetch_texel(ctx, machine, inst, texcoord, lodBias, color);
+
store_vector4(inst, machine, color);
}
break;
@@ -1368,6 +1377,7 @@ _mesa_execute_program(GLcontext * ctx,
fetch_vector4(&inst->SrcReg[1], machine, dtdx);
fetch_vector4(&inst->SrcReg[2], machine, dtdy);
machine->FetchTexelDeriv(ctx, texcoord, dtdx, dtdy,
+ 0.0, /* lodBias */
inst->TexSrcUnit, color);
store_vector4(inst, machine, color);
}
@@ -1375,14 +1385,8 @@ _mesa_execute_program(GLcontext * ctx,
case OPCODE_TXP: /* GL_ARB_fragment_program only */
/* Texture lookup w/ projective divide */
{
- GLfloat texcoord[4], color[4], lambda;
-#if 0
- if (inst->SrcReg[0].File == PROGRAM_INPUT &&
- inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
- lambda = span->array->lambda[inst->TexSrcUnit][column];
- else
-#endif
- lambda = 0.0;
+ GLfloat texcoord[4], color[4];
+
fetch_vector4(&inst->SrcReg[0], machine, texcoord);
/* Not so sure about this test - if texcoord[3] is
* zero, we'd probably be fine except for an ASSERT in
@@ -1393,22 +1397,19 @@ _mesa_execute_program(GLcontext * ctx,
texcoord[1] /= texcoord[3];
texcoord[2] /= texcoord[3];
}
- machine->FetchTexelLod(ctx, texcoord, lambda,
- inst->TexSrcUnit, color);
+
+ fetch_texel(ctx, machine, inst, texcoord, 0.0, color);
+
store_vector4(inst, machine, color);
}
break;
case OPCODE_TXP_NV: /* GL_NV_fragment_program only */
- /* Texture lookup w/ projective divide */
+ /* Texture lookup w/ projective divide, as above, but do not
+ * do the divide by w if sampling from a cube map.
+ */
{
- GLfloat texcoord[4], color[4], lambda;
-#if 0
- if (inst->SrcReg[0].File == PROGRAM_INPUT &&
- inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
- lambda = span->array->lambda[inst->TexSrcUnit][column];
- else
-#endif
- lambda = 0.0;
+ GLfloat texcoord[4], color[4];
+
fetch_vector4(&inst->SrcReg[0], machine, texcoord);
if (inst->TexSrcTarget != TEXTURE_CUBE_INDEX &&
texcoord[3] != 0.0) {
@@ -1416,8 +1417,9 @@ _mesa_execute_program(GLcontext * ctx,
texcoord[1] /= texcoord[3];
texcoord[2] /= texcoord[3];
}
- machine->FetchTexelLod(ctx, texcoord, lambda,
- inst->TexSrcUnit, color);
+
+ fetch_texel(ctx, machine, inst, texcoord, 0.0, color);
+
store_vector4(inst, machine, color);
}
break;
diff --git a/src/mesa/shader/prog_execute.h b/src/mesa/shader/prog_execute.h
index be29ece..3ea0ba1 100644
--- a/src/mesa/shader/prog_execute.h
+++ b/src/mesa/shader/prog_execute.h
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -32,6 +32,7 @@ typedef void (*FetchTexelLodFunc)(GLcontext *ctx, const GLfloat texcoord[4],
typedef void (*FetchTexelDerivFunc)(GLcontext *ctx, const GLfloat texcoord[4],
const GLfloat texdx[4],
const GLfloat texdy[4],
+ GLfloat lodBias,
GLuint unit, GLfloat color[4]);
diff --git a/src/mesa/shader/prog_parameter.c b/src/mesa/shader/prog_parameter.c
index 9e3d3fe..46d3087 100644
--- a/src/mesa/shader/prog_parameter.c
+++ b/src/mesa/shader/prog_parameter.c
@@ -384,7 +384,7 @@ sizeof_state_reference(const GLint *stateTokens)
* PARAM ambient = state.material.front.ambient;
*
* \param paramList the parameter list
- * \param state an array of 6 (STATE_LENGTH) state tokens
+ * \param stateTokens an array of 5 (STATE_LENGTH) state tokens
* \return index of the new parameter.
*/
GLint
diff --git a/src/mesa/shader/prog_statevars.c b/src/mesa/shader/prog_statevars.c
index d37d7fb..aa1e6e1 100644
--- a/src/mesa/shader/prog_statevars.c
+++ b/src/mesa/shader/prog_statevars.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 7.0
+ * Version: 7.1
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -824,3 +824,94 @@ _mesa_load_state_parameters(GLcontext *ctx,
}
}
+
+/**
+ * Copy the 16 elements of a matrix into four consecutive program
+ * registers starting at 'pos'.
+ */
+static void
+load_matrix(GLfloat registers[][4], GLuint pos, const GLfloat mat[16])
+{
+ GLuint i;
+ for (i = 0; i < 4; i++) {
+ registers[pos + i][0] = mat[0 + i];
+ registers[pos + i][1] = mat[4 + i];
+ registers[pos + i][2] = mat[8 + i];
+ registers[pos + i][3] = mat[12 + i];
+ }
+}
+
+
+/**
+ * As above, but transpose the matrix.
+ */
+static void
+load_transpose_matrix(GLfloat registers[][4], GLuint pos,
+ const GLfloat mat[16])
+{
+ MEMCPY(registers[pos], mat, 16 * sizeof(GLfloat));
+}
+
+
+/**
+ * Load current vertex program's parameter registers with tracked
+ * matrices (if NV program). This only needs to be done per
+ * glBegin/glEnd, not per-vertex.
+ */
+void
+_mesa_load_tracked_matrices(GLcontext *ctx)
+{
+ GLuint i;
+
+ for (i = 0; i < MAX_NV_VERTEX_PROGRAM_PARAMS / 4; i++) {
+ /* point 'mat' at source matrix */
+ GLmatrix *mat;
+ if (ctx->VertexProgram.TrackMatrix[i] == GL_MODELVIEW) {
+ mat = ctx->ModelviewMatrixStack.Top;
+ }
+ else if (ctx->VertexProgram.TrackMatrix[i] == GL_PROJECTION) {
+ mat = ctx->ProjectionMatrixStack.Top;
+ }
+ else if (ctx->VertexProgram.TrackMatrix[i] == GL_TEXTURE) {
+ mat = ctx->TextureMatrixStack[ctx->Texture.CurrentUnit].Top;
+ }
+ else if (ctx->VertexProgram.TrackMatrix[i] == GL_COLOR) {
+ mat = ctx->ColorMatrixStack.Top;
+ }
+ else if (ctx->VertexProgram.TrackMatrix[i]==GL_MODELVIEW_PROJECTION_NV) {
+ /* XXX verify the combined matrix is up to date */
+ mat = &ctx->_ModelProjectMatrix;
+ }
+ else if (ctx->VertexProgram.TrackMatrix[i] >= GL_MATRIX0_NV &&
+ ctx->VertexProgram.TrackMatrix[i] <= GL_MATRIX7_NV) {
+ GLuint n = ctx->VertexProgram.TrackMatrix[i] - GL_MATRIX0_NV;
+ ASSERT(n < MAX_PROGRAM_MATRICES);
+ mat = ctx->ProgramMatrixStack[n].Top;
+ }
+ else {
+ /* no matrix is tracked, but we leave the register values as-is */
+ assert(ctx->VertexProgram.TrackMatrix[i] == GL_NONE);
+ continue;
+ }
+
+ /* load the matrix values into sequential registers */
+ if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_IDENTITY_NV) {
+ load_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
+ }
+ else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_INVERSE_NV) {
+ _math_matrix_analyse(mat); /* update the inverse */
+ ASSERT(!_math_matrix_is_dirty(mat));
+ load_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
+ }
+ else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_TRANSPOSE_NV) {
+ load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
+ }
+ else {
+ assert(ctx->VertexProgram.TrackMatrixTransform[i]
+ == GL_INVERSE_TRANSPOSE_NV);
+ _math_matrix_analyse(mat); /* update the inverse */
+ ASSERT(!_math_matrix_is_dirty(mat));
+ load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
+ }
+ }
+}
diff --git a/src/mesa/shader/prog_statevars.h b/src/mesa/shader/prog_statevars.h
index 3281a4a..22bb8e0 100644
--- a/src/mesa/shader/prog_statevars.h
+++ b/src/mesa/shader/prog_statevars.h
@@ -1,8 +1,8 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.2
+ * Version: 7.1
*
- * Copyright (C) 1999-2006 Brian Paul All Rights Reserved.
+ * Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -126,4 +126,8 @@ extern const char *
_mesa_program_state_string(const gl_state_index state[STATE_LENGTH]);
+extern void
+_mesa_load_tracked_matrices(GLcontext *ctx);
+
+
#endif /* PROG_STATEVARS_H */
diff --git a/src/mesa/shader/shader_api.c b/src/mesa/shader/shader_api.c
index 06d24b4..aba35dc 100644
--- a/src/mesa/shader/shader_api.c
+++ b/src/mesa/shader/shader_api.c
@@ -370,6 +370,54 @@ copy_string(GLchar *dst, GLsizei maxLength, GLsizei *length, const GLchar *src)
/**
+ * Return size (in floats) of the given GLSL type.
+ * See also _slang_sizeof_type_specifier().
+ */
+static GLint
+sizeof_glsl_type(GLenum type)
+{
+ switch (type) {
+ case GL_BOOL:
+ case GL_FLOAT:
+ case GL_INT:
+ return 1;
+ case GL_BOOL_VEC2:
+ case GL_FLOAT_VEC2:
+ case GL_INT_VEC2:
+ return 2;
+ case GL_BOOL_VEC3:
+ case GL_FLOAT_VEC3:
+ case GL_INT_VEC3:
+ return 3;
+ case GL_BOOL_VEC4:
+ case GL_FLOAT_VEC4:
+ case GL_INT_VEC4:
+ return 4;
+ case GL_FLOAT_MAT2:
+ return 8; /* 2 rows of 4, actually */
+ case GL_FLOAT_MAT3:
+ return 12; /* 3 rows of 4, actually */
+ case GL_FLOAT_MAT4:
+ return 16;
+ case GL_FLOAT_MAT2x3:
+ return 6;
+ case GL_FLOAT_MAT2x4:
+ return 8;
+ case GL_FLOAT_MAT3x2:
+ return 12; /* 3 rows of 4, actually */
+ case GL_FLOAT_MAT3x4:
+ return 12;
+ case GL_FLOAT_MAT4x2:
+ return 16; /* 4 rows of 4, actually */
+ case GL_FLOAT_MAT4x3:
+ return 12;
+ default:
+ return 0; /* error */
+ }
+}
+
+
+/**
* Called via ctx->Driver.AttachShader()
*/
void
@@ -666,13 +714,17 @@ _mesa_get_active_uniform(GLcontext *ctx, GLuint program, GLuint index,
if (shProg->Uniforms->Parameters[j].Type == PROGRAM_UNIFORM ||
shProg->Uniforms->Parameters[j].Type == PROGRAM_SAMPLER) {
if (ind == index) {
+ GLuint uSize = shProg->Uniforms->Parameters[j].Size;
+ GLenum uType = shProg->Uniforms->Parameters[j].DataType;
/* found it */
copy_string(nameOut, maxLength, length,
shProg->Uniforms->Parameters[j].Name);
- if (size)
- *size = shProg->Uniforms->Parameters[j].Size;
+ if (size) {
+ /* convert from floats to 'type' (eg: sizeof(mat4x4)=1) */
+ *size = uSize / sizeof_glsl_type(uType);
+ }
if (type)
- *type = shProg->Uniforms->Parameters[j].DataType;
+ *type = uType;
return;
}
ind++;
diff --git a/src/mesa/shader/slang/slang_builtin.c b/src/mesa/shader/slang/slang_builtin.c
index 6ee0fd3..7351f26 100644
--- a/src/mesa/shader/slang/slang_builtin.c
+++ b/src/mesa/shader/slang/slang_builtin.c
@@ -250,7 +250,7 @@ lookup_statevar(const char *var, GLint index1, GLint index2, const char *field,
}
}
else if (strcmp(var, "gl_FrontLightModelProduct") == 0) {
- if (strcmp(field, "ambient") == 0) {
+ if (strcmp(field, "sceneColor") == 0) {
tokens[0] = STATE_LIGHTMODEL_SCENECOLOR;
tokens[1] = 0;
}
@@ -259,7 +259,7 @@ lookup_statevar(const char *var, GLint index1, GLint index2, const char *field,
}
}
else if (strcmp(var, "gl_BackLightModelProduct") == 0) {
- if (strcmp(field, "ambient") == 0) {
+ if (strcmp(field, "sceneColor") == 0) {
tokens[0] = STATE_LIGHTMODEL_SCENECOLOR;
tokens[1] = 1;
}
@@ -397,6 +397,8 @@ lookup_statevar(const char *var, GLint index1, GLint index2, const char *field,
* var.field
* var[i].field
* var[i][j]
+ *
+ * \return -1 upon error, else position in paramList of the state var/data
*/
GLint
_slang_alloc_statevar(slang_ir_node *n,
@@ -414,9 +416,13 @@ _slang_alloc_statevar(slang_ir_node *n,
if (n->Opcode == IR_ELEMENT) {
/* XXX can only handle constant indexes for now */
- assert(n->Children[1]->Opcode == IR_FLOAT);
- index1 = (GLint) n->Children[1]->Value[0];
- n = n->Children[0];
+ if (n->Children[1]->Opcode == IR_FLOAT) {
+ index1 = (GLint) n->Children[1]->Value[0];
+ n = n->Children[0];
+ }
+ else {
+ return -1;
+ }
}
if (n->Opcode == IR_ELEMENT) {
diff --git a/src/mesa/shader/slang/slang_codegen.c b/src/mesa/shader/slang/slang_codegen.c
index 675dd83..ef9c0ab 100644
--- a/src/mesa/shader/slang/slang_codegen.c
+++ b/src/mesa/shader/slang/slang_codegen.c
@@ -2344,7 +2344,8 @@ _slang_gen_field(slang_assemble_ctx * A, slang_operation *oper)
return n;
}
else if ( ti.spec.type == SLANG_SPEC_FLOAT
- || ti.spec.type == SLANG_SPEC_INT) {
+ || ti.spec.type == SLANG_SPEC_INT
+ || ti.spec.type == SLANG_SPEC_BOOL) {
const GLuint rows = 1;
slang_swizzle swz;
slang_ir_node *n;
diff --git a/src/mesa/shader/slang/slang_emit.c b/src/mesa/shader/slang/slang_emit.c
index fe13f28..9947544 100644
--- a/src/mesa/shader/slang/slang_emit.c
+++ b/src/mesa/shader/slang/slang_emit.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 2005-2007 Brian Paul All Rights Reserved.
*
@@ -859,12 +859,18 @@ emit_return(slang_emit_info *emitInfo, slang_ir_node *n)
static struct prog_instruction *
emit_kill(slang_emit_info *emitInfo)
{
+ struct gl_fragment_program *fp;
struct prog_instruction *inst;
/* NV-KILL - discard fragment depending on condition code.
* Note that ARB-KILL depends on sign of vector operand.
*/
inst = new_instruction(emitInfo, OPCODE_KIL_NV);
inst->DstReg.CondMask = COND_TR; /* always branch */
+
+ assert(emitInfo->prog->Target == GL_FRAGMENT_PROGRAM_ARB);
+ fp = (struct gl_fragment_program *) emitInfo->prog;
+ fp->UsesKill = GL_TRUE;
+
return inst;
}
@@ -1486,6 +1492,10 @@ emit_struct_field(slang_emit_info *emitInfo, slang_ir_node *n)
{
if (n->Store->File == PROGRAM_STATE_VAR) {
n->Store->Index = _slang_alloc_statevar(n, emitInfo->prog->Parameters);
+ if (n->Store->Index < 0) {
+ slang_info_log_error(emitInfo->log, "Error parsing state variable");
+ return NULL;
+ }
}
else {
GLint offset = n->FieldOffset / 4;
diff --git a/src/mesa/swrast/s_aalinetemp.h b/src/mesa/swrast/s_aalinetemp.h
index 80cec0b..074ffe8 100644
--- a/src/mesa/swrast/s_aalinetemp.h
+++ b/src/mesa/swrast/s_aalinetemp.h
@@ -141,6 +141,7 @@ NAME(line)(GLcontext *ctx, const SWvertex *v0, const SWvertex *v1)
return;
INIT_SPAN(line.span, GL_LINE, 0, 0, SPAN_XY | SPAN_COVERAGE);
+ line.span.facing = swrast->PointLineFacing;
line.xAdj = line.dx / line.len * line.halfWidth;
line.yAdj = line.dy / line.len * line.halfWidth;
diff --git a/src/mesa/swrast/s_aatritemp.h b/src/mesa/swrast/s_aatritemp.h
index 4162ed6..b26f21f 100644
--- a/src/mesa/swrast/s_aatritemp.h
+++ b/src/mesa/swrast/s_aatritemp.h
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -78,7 +78,7 @@
GLfloat texWidth[FRAG_ATTRIB_MAX];
GLfloat texHeight[FRAG_ATTRIB_MAX];
#endif
- GLfloat bf = SWRAST_CONTEXT(ctx)->_BackfaceSign;
+ GLfloat bf = SWRAST_CONTEXT(ctx)->_BackfaceCullSign;
(void) swrast;
@@ -116,6 +116,7 @@
majDx = vMax->win[0] - vMin->win[0];
majDy = vMax->win[1] - vMin->win[1];
+ /* front/back-face determination and cullling */
{
const GLfloat botDx = vMid->win[0] - vMin->win[0];
const GLfloat botDy = vMid->win[1] - vMin->win[1];
@@ -124,6 +125,8 @@
if (area * bf < 0 || area == 0 || IS_INF_OR_NAN(area))
return;
ltor = (GLboolean) (area < 0.0F);
+
+ span.facing = area * swrast->_BackfaceSign > 0.0F;
}
/* Plane equation setup:
@@ -336,7 +339,7 @@
}
/* skip fragments with zero coverage */
- while (startX >= 0) {
+ while (startX > 0) {
coverage = compute_coveragef(pMin, pMax, pMid, startX, iy);
if (coverage > 0.0F)
break;
@@ -350,6 +353,7 @@
/* (cx,cy) = center of fragment */
const GLfloat cx = ix + 0.5F, cy = iy + 0.5F;
SWspanarrays *array = span.array;
+ ASSERT(ix >= 0);
#ifdef DO_INDEX
array->coverage[ix] = (GLfloat) compute_coveragei(pMin, pMax, pMid, ix, iy);
#else
diff --git a/src/mesa/swrast/s_context.c b/src/mesa/swrast/s_context.c
index 2f25edb..9b9b73f 100644
--- a/src/mesa/swrast/s_context.c
+++ b/src/mesa/swrast/s_context.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -117,8 +117,8 @@ _swrast_update_rasterflags( GLcontext *ctx )
/**
- * Examine polycon culls tate to compute the _BackfaceSign field.
- * _BackfaceSign will be 0 if no culling, -1 if culling back-faces,
+ * Examine polycon culls tate to compute the _BackfaceCullSign field.
+ * _BackfaceCullSign will be 0 if no culling, -1 if culling back-faces,
* and 1 if culling front-faces. The Polygon FrontFace state also
* factors in.
*/
@@ -128,31 +128,32 @@ _swrast_update_polygon( GLcontext *ctx )
GLfloat backface_sign;
if (ctx->Polygon.CullFlag) {
- backface_sign = 1.0;
switch (ctx->Polygon.CullFaceMode) {
case GL_BACK:
- if (ctx->Polygon.FrontFace == GL_CCW)
- backface_sign = -1.0;
+ backface_sign = -1.0;
break;
case GL_FRONT:
- if (ctx->Polygon.FrontFace != GL_CCW)
- backface_sign = -1.0;
+ backface_sign = 1.0;
break;
case GL_FRONT_AND_BACK:
/* fallthrough */
default:
backface_sign = 0.0;
- break;
}
}
else {
backface_sign = 0.0;
}
- SWRAST_CONTEXT(ctx)->_BackfaceSign = backface_sign;
+ SWRAST_CONTEXT(ctx)->_BackfaceCullSign = backface_sign;
+
+ /* This is for front/back-face determination, but not for culling */
+ SWRAST_CONTEXT(ctx)->_BackfaceSign
+ = (ctx->Polygon.FrontFace == GL_CW) ? -1.0 : 1.0;
}
+
/**
* Update the _PreferPixelFog field to indicate if we need to compute
* fog blend factors (from the fog coords) per-fragment.
@@ -721,6 +722,12 @@ _swrast_ResetLineStipple( GLcontext *ctx )
}
void
+_swrast_SetFacing(GLcontext *ctx, GLuint facing)
+{
+ SWRAST_CONTEXT(ctx)->PointLineFacing = facing;
+}
+
+void
_swrast_allow_vertex_fog( GLcontext *ctx, GLboolean value )
{
if (SWRAST_DEBUG) {
diff --git a/src/mesa/swrast/s_context.h b/src/mesa/swrast/s_context.h
index c8333b8..58841ad 100644
--- a/src/mesa/swrast/s_context.h
+++ b/src/mesa/swrast/s_context.h
@@ -128,7 +128,8 @@ typedef struct
* _swrast_validate_derived():
*/
GLbitfield _RasterMask;
- GLfloat _BackfaceSign;
+ GLfloat _BackfaceSign; /** +1 or -1 */
+ GLfloat _BackfaceCullSign; /** +1, 0, or -1 */
GLboolean _PreferPixelFog; /* Compute fog blend factor per fragment? */
GLboolean _AnyTextureCombine;
GLboolean _FogEnabled;
@@ -151,6 +152,7 @@ typedef struct
/* Working values:
*/
GLuint StippleCounter; /**< Line stipple counter */
+ GLuint PointLineFacing;
GLbitfield NewState;
GLuint StateChanges;
GLenum Primitive; /* current primitive being drawn (ala glBegin) */
diff --git a/src/mesa/swrast/s_fragprog.c b/src/mesa/swrast/s_fragprog.c
index e47dbbd..89114eb 100644
--- a/src/mesa/swrast/s_fragprog.c
+++ b/src/mesa/swrast/s_fragprog.c
@@ -1,8 +1,8 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.2
+ * Version: 7.0.3
*
- * Copyright (C) 1999-2006 Brian Paul All Rights Reserved.
+ * Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -40,10 +40,12 @@ fetch_texel( GLcontext *ctx, const GLfloat texcoord[4], GLfloat lambda,
{
GLchan rgba[4];
SWcontext *swrast = SWRAST_CONTEXT(ctx);
+ const struct gl_texture_object *texObj = ctx->Texture.Unit[unit]._Current;
+
+ lambda = CLAMP(lambda, texObj->MinLod, texObj->MaxLod);
/* XXX use a float-valued TextureSample routine here!!! */
- swrast->TextureSample[unit](ctx, ctx->Texture.Unit[unit]._Current,
- 1, (const GLfloat (*)[4]) texcoord,
+ swrast->TextureSample[unit](ctx, texObj, 1, (const GLfloat (*)[4]) texcoord,
&lambda, &rgba);
color[0] = CHAN_TO_FLOAT(rgba[0]);
color[1] = CHAN_TO_FLOAT(rgba[1]);
@@ -59,7 +61,7 @@ fetch_texel( GLcontext *ctx, const GLfloat texcoord[4], GLfloat lambda,
static void
fetch_texel_deriv( GLcontext *ctx, const GLfloat texcoord[4],
const GLfloat texdx[4], const GLfloat texdy[4],
- GLuint unit, GLfloat color[4] )
+ GLfloat lodBias, GLuint unit, GLfloat color[4] )
{
SWcontext *swrast = SWRAST_CONTEXT(ctx);
const struct gl_texture_object *texObj = ctx->Texture.Unit[unit]._Current;
@@ -68,15 +70,17 @@ fetch_texel_deriv( GLcontext *ctx, const GLfloat texcoord[4],
const GLfloat texH = (GLfloat) texImg->HeightScale;
GLchan rgba[4];
- GLfloat lambda = _swrast_compute_lambda(texdx[0], texdy[0], /* ds/dx, ds/dy */
- texdx[1], texdy[1], /* dt/dx, dt/dy */
- texdx[3], texdy[2], /* dq/dx, dq/dy */
- texW, texH,
- texcoord[0], texcoord[1], texcoord[3],
- 1.0F / texcoord[3]);
+ GLfloat lambda
+ = _swrast_compute_lambda(texdx[0], texdy[0], /* ds/dx, ds/dy */
+ texdx[1], texdy[1], /* dt/dx, dt/dy */
+ texdx[3], texdy[2], /* dq/dx, dq/dy */
+ texW, texH,
+ texcoord[0], texcoord[1], texcoord[3],
+ 1.0F / texcoord[3]) + lodBias;
+
+ lambda = CLAMP(lambda, texObj->MinLod, texObj->MaxLod);
- swrast->TextureSample[unit](ctx, ctx->Texture.Unit[unit]._Current,
- 1, (const GLfloat (*)[4]) texcoord,
+ swrast->TextureSample[unit](ctx, texObj, 1, (const GLfloat (*)[4]) texcoord,
&lambda, &rgba);
color[0] = CHAN_TO_FLOAT(rgba[0]);
color[1] = CHAN_TO_FLOAT(rgba[1]);
@@ -114,7 +118,7 @@ init_machine(GLcontext *ctx, struct gl_program_machine *machine,
if (ctx->Shader.CurrentProgram) {
/* Store front/back facing value in register FOGC.Y */
- machine->Attribs[FRAG_ATTRIB_FOGC][col][1] = (GLfloat) ctx->_Facing;
+ machine->Attribs[FRAG_ATTRIB_FOGC][col][1] = 1.0 - span->facing;
}
machine->CurElement = col;
diff --git a/src/mesa/swrast/s_linetemp.h b/src/mesa/swrast/s_linetemp.h
index b6e8f28..9e240a7 100644
--- a/src/mesa/swrast/s_linetemp.h
+++ b/src/mesa/swrast/s_linetemp.h
@@ -317,6 +317,9 @@ NAME( GLcontext *ctx, const SWvertex *vert0, const SWvertex *vert1 )
span.attrStepX[FRAG_ATTRIB_WPOS][3] = 0.0F;
span.attrStepY[FRAG_ATTRIB_WPOS][3] = 0.0F;
+ span.facing = swrast->PointLineFacing;
+
+
/*
* Draw
*/
diff --git a/src/mesa/swrast/s_pointtemp.h b/src/mesa/swrast/s_pointtemp.h
index 9436464..7dcc1fc 100644
--- a/src/mesa/swrast/s_pointtemp.h
+++ b/src/mesa/swrast/s_pointtemp.h
@@ -105,6 +105,7 @@ NAME ( GLcontext *ctx, const SWvertex *vert )
*/
span->interpMask = SPAN_FOG;
span->arrayMask = SPAN_XY | SPAN_Z;
+ span->facing = swrast->PointLineFacing;
span->attrStart[FRAG_ATTRIB_FOGC][0] = vert->attrib[FRAG_ATTRIB_FOGC][0];
span->attrStepX[FRAG_ATTRIB_FOGC][0] = 0.0;
span->attrStepY[FRAG_ATTRIB_FOGC][0] = 0.0;
diff --git a/src/mesa/swrast/s_readpix.c b/src/mesa/swrast/s_readpix.c
index fe9a70f..916ddc1 100644
--- a/src/mesa/swrast/s_readpix.c
+++ b/src/mesa/swrast/s_readpix.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -129,7 +129,8 @@ read_depth_pixels( GLcontext *ctx,
rb->GetRow(ctx, rb, width, x, y, dest);
/* convert range from 24-bit to 32-bit */
for (k = 0; k < width; k++) {
- dest[k] = (dest[k] << 8) | (dest[k] >> 24);
+ /* Note: put MSByte of 24-bit value into LSByte */
+ dest[k] = (dest[k] << 8) | ((dest[k] >> 16) & 0xff);
}
}
}
diff --git a/src/mesa/swrast/s_texfilter.c b/src/mesa/swrast/s_texfilter.c
index 5413fc0..084d65f 100644
--- a/src/mesa/swrast/s_texfilter.c
+++ b/src/mesa/swrast/s_texfilter.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.0.3
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -1043,7 +1043,7 @@ sample_2d_linear_repeat(GLcontext *ctx,
ASSERT(tObj->WrapS == GL_REPEAT);
ASSERT(tObj->WrapT == GL_REPEAT);
ASSERT(img->Border == 0);
- ASSERT(img->_BaseFormat != GL_COLOR_INDEX);
+ ASSERT(img->TexFormat->BaseFormat != GL_COLOR_INDEX);
ASSERT(img->_IsPowerOfTwo);
COMPUTE_LINEAR_REPEAT_TEXEL_LOCATION(texcoord[0], u, width, i0, i1);
@@ -1198,6 +1198,7 @@ sample_linear_2d( GLcontext *ctx,
(void) lambda;
if (tObj->WrapS == GL_REPEAT &&
tObj->WrapT == GL_REPEAT &&
+ image->Border == 0 &&
image->_IsPowerOfTwo) {
for (i=0;i<n;i++) {
sample_2d_linear_repeat(ctx, tObj, image, texcoords[i], rgba[i]);
@@ -1237,7 +1238,7 @@ opt_sample_rgb_2d( GLcontext *ctx,
ASSERT(tObj->WrapS==GL_REPEAT);
ASSERT(tObj->WrapT==GL_REPEAT);
ASSERT(img->Border==0);
- ASSERT(img->_BaseFormat==GL_RGB);
+ ASSERT(img->TexFormat->MesaFormat==MESA_FORMAT_RGB);
ASSERT(img->_IsPowerOfTwo);
for (k=0; k<n; k++) {
@@ -1278,7 +1279,7 @@ opt_sample_rgba_2d( GLcontext *ctx,
ASSERT(tObj->WrapS==GL_REPEAT);
ASSERT(tObj->WrapT==GL_REPEAT);
ASSERT(img->Border==0);
- ASSERT(img->_BaseFormat==GL_RGBA);
+ ASSERT(img->TexFormat->MesaFormat==MESA_FORMAT_RGBA);
ASSERT(img->_IsPowerOfTwo);
for (i = 0; i < n; i++) {
@@ -1308,7 +1309,7 @@ sample_lambda_2d( GLcontext *ctx,
const GLboolean repeatNoBorderPOT = (tObj->WrapS == GL_REPEAT)
&& (tObj->WrapT == GL_REPEAT)
&& (tImg->Border == 0 && (tImg->Width == tImg->RowStride))
- && (tImg->_BaseFormat != GL_COLOR_INDEX)
+ && (tImg->TexFormat->BaseFormat != GL_COLOR_INDEX)
&& tImg->_IsPowerOfTwo;
ASSERT(lambda != NULL);
@@ -1323,16 +1324,10 @@ sample_lambda_2d( GLcontext *ctx,
if (repeatNoBorderPOT) {
switch (tImg->TexFormat->MesaFormat) {
case MESA_FORMAT_RGB:
- case MESA_FORMAT_RGB888:
- /*case MESA_FORMAT_BGR888:*/
opt_sample_rgb_2d(ctx, tObj, m, texcoords + minStart,
NULL, rgba + minStart);
break;
case MESA_FORMAT_RGBA:
- case MESA_FORMAT_RGBA8888:
- case MESA_FORMAT_ARGB8888:
- /*case MESA_FORMAT_ABGR8888:*/
- /*case MESA_FORMAT_BGRA8888:*/
opt_sample_rgba_2d(ctx, tObj, m, texcoords + minStart,
NULL, rgba + minStart);
break;
@@ -1386,16 +1381,10 @@ sample_lambda_2d( GLcontext *ctx,
if (repeatNoBorderPOT) {
switch (tImg->TexFormat->MesaFormat) {
case MESA_FORMAT_RGB:
- case MESA_FORMAT_RGB888:
- /*case MESA_FORMAT_BGR888:*/
opt_sample_rgb_2d(ctx, tObj, m, texcoords + magStart,
NULL, rgba + magStart);
break;
case MESA_FORMAT_RGBA:
- case MESA_FORMAT_RGBA8888:
- case MESA_FORMAT_ARGB8888:
- /*case MESA_FORMAT_ABGR8888:*/
- /*case MESA_FORMAT_BGRA8888:*/
opt_sample_rgba_2d(ctx, tObj, m, texcoords + magStart,
NULL, rgba + magStart);
break;
@@ -2102,7 +2091,7 @@ sample_nearest_rect(GLcontext *ctx,
ASSERT(tObj->WrapT == GL_CLAMP ||
tObj->WrapT == GL_CLAMP_TO_EDGE ||
tObj->WrapT == GL_CLAMP_TO_BORDER);
- ASSERT(img->_BaseFormat != GL_COLOR_INDEX);
+ ASSERT(img->TexFormat->BaseFormat != GL_COLOR_INDEX);
for (i = 0; i < n; i++) {
GLint row, col;
@@ -2138,7 +2127,7 @@ sample_linear_rect(GLcontext *ctx,
ASSERT(tObj->WrapT == GL_CLAMP ||
tObj->WrapT == GL_CLAMP_TO_EDGE ||
tObj->WrapT == GL_CLAMP_TO_BORDER);
- ASSERT(img->_BaseFormat != GL_COLOR_INDEX);
+ ASSERT(img->TexFormat->BaseFormat != GL_COLOR_INDEX);
/* XXX lots of opportunity for optimization in this loop */
for (i = 0; i < n; i++) {
@@ -2286,8 +2275,8 @@ sample_depth_texture( GLcontext *ctx,
(void) lambda;
- ASSERT(tObj->Image[0][tObj->BaseLevel]->_BaseFormat == GL_DEPTH_COMPONENT ||
- tObj->Image[0][tObj->BaseLevel]->_BaseFormat == GL_DEPTH_STENCIL_EXT);
+ ASSERT(img->TexFormat->BaseFormat == GL_DEPTH_COMPONENT ||
+ img->TexFormat->BaseFormat == GL_DEPTH_STENCIL_EXT);
ASSERT(tObj->Target == GL_TEXTURE_1D ||
tObj->Target == GL_TEXTURE_2D ||
@@ -2604,7 +2593,7 @@ sample_depth_texture2(const GLcontext *ctx,
* GL_TEXTURE_COMPARE_SGIX == GL_TRUE but the current texture object
* isn't a depth texture.
*/
- if (texImage->_BaseFormat != GL_DEPTH_COMPONENT) {
+ if (texImage->TexFormat->BaseFormat != GL_DEPTH_COMPONENT) {
_mesa_problem(ctx,"GL_TEXTURE_COMPARE_SGIX enabled with non-depth texture");
return;
}
@@ -2707,7 +2696,7 @@ _swrast_choose_texture_sample_func( GLcontext *ctx,
}
else {
const GLboolean needLambda = (GLboolean) (t->MinFilter != t->MagFilter);
- const GLenum format = t->Image[0][t->BaseLevel]->_BaseFormat;
+ const GLenum format = t->Image[0][t->BaseLevel]->TexFormat->BaseFormat;
switch (t->Target) {
case GL_TEXTURE_1D:
diff --git a/src/mesa/swrast/s_triangle.c b/src/mesa/swrast/s_triangle.c
index fc9d29b..fc40084 100644
--- a/src/mesa/swrast/s_triangle.c
+++ b/src/mesa/swrast/s_triangle.c
@@ -58,7 +58,7 @@ _swrast_culltriangle( GLcontext *ctx,
GLfloat fy = v2->win[1] - v0->win[1];
GLfloat c = ex*fy-ey*fx;
- if (c * SWRAST_CONTEXT(ctx)->_BackfaceSign > 0)
+ if (c * SWRAST_CONTEXT(ctx)->_BackfaceCullSign > 0)
return 0;
return 1;
diff --git a/src/mesa/swrast/s_tritemp.h b/src/mesa/swrast/s_tritemp.h
index dcc3e95..891eaf6 100644
--- a/src/mesa/swrast/s_tritemp.h
+++ b/src/mesa/swrast/s_tritemp.h
@@ -292,18 +292,17 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
#else
const GLfloat area = eMaj.dx * eBot.dy - eBot.dx * eMaj.dy;
#endif
- /* Do backface culling */
- if (area * bf < 0.0)
+ if (IS_INF_OR_NAN(area) || area == 0.0F)
return;
- if (IS_INF_OR_NAN(area) || area == 0.0F)
+ if (area * bf * swrast->_BackfaceCullSign < 0.0)
return;
oneOverArea = 1.0F / area;
- }
-
- span.facing = ctx->_Facing; /* for 2-sided stencil test */
+ /* 0 = front, 1 = back */
+ span.facing = oneOverArea * bf > 0.0F;
+ }
/* Edge setup. For a triangle strip these could be reused... */
{
diff --git a/src/mesa/swrast/swrast.h b/src/mesa/swrast/swrast.h
index 12264a1..8640550 100644
--- a/src/mesa/swrast/swrast.h
+++ b/src/mesa/swrast/swrast.h
@@ -140,6 +140,13 @@ _swrast_Accum(GLcontext *ctx, GLenum op, GLfloat value);
extern void
_swrast_ResetLineStipple( GLcontext *ctx );
+/**
+ * Indicates front/back facing for subsequent points/lines when drawing
+ * unfilled polygons. Needed for two-side stencil.
+ */
+extern void
+_swrast_SetFacing(GLcontext *ctx, GLuint facing);
+
/* These will always render the correct point/line/triangle for the
* current state.
*
diff --git a/src/mesa/swrast_setup/ss_context.c b/src/mesa/swrast_setup/ss_context.c
index 3f6d294..43f2696 100644
--- a/src/mesa/swrast_setup/ss_context.c
+++ b/src/mesa/swrast_setup/ss_context.c
@@ -186,6 +186,9 @@ _swsetup_RenderStart( GLcontext *ctx )
swsetup->NewState = 0;
+ /* This will change if drawing unfilled tris */
+ _swrast_SetFacing(ctx, 0);
+
_swrast_render_start(ctx);
/* Important */
diff --git a/src/mesa/swrast_setup/ss_triangle.c b/src/mesa/swrast_setup/ss_triangle.c
index 628e928..0585673 100644
--- a/src/mesa/swrast_setup/ss_triangle.c
+++ b/src/mesa/swrast_setup/ss_triangle.c
@@ -1,8 +1,8 @@
/*
* Mesa 3-D graphics library
- * Version: 6.1
+ * Version: 7.1
*
- * Copyright (C) 1999-2004 Brian Paul All Rights Reserved.
+ * Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -68,6 +68,8 @@ static void _swsetup_render_line_tri( GLcontext *ctx,
return;
}
+ _swrast_SetFacing(ctx, facing);
+
if (ctx->Light.ShadeModel == GL_FLAT) {
COPY_CHAN4(c[0], v0->color);
COPY_CHAN4(c[1], v1->color);
@@ -127,6 +129,8 @@ static void _swsetup_render_point_tri( GLcontext *ctx,
return;
}
+ _swrast_SetFacing(ctx, facing);
+
if (ctx->Light.ShadeModel == GL_FLAT) {
/* save colors/indexes for v0, v1 vertices */
COPY_CHAN4(c[0], v0->color);
@@ -290,10 +294,8 @@ void _swsetup_choose_trifuncs( GLcontext *ctx )
ctx->Polygon.OffsetFill)
ind |= SS_OFFSET_BIT;
- /* Note: gl_FrontFacing lives in fragment input FOGC.Y at this time */
if ((ctx->Light.Enabled && ctx->Light.Model.TwoSide) ||
- (ctx->VertexProgram._Enabled && ctx->VertexProgram.TwoSideEnabled) ||
- (ctx->FragmentProgram._Current && ctx->FragmentProgram._Current->Base.InputsRead & (1 << FRAG_ATTRIB_FOGC)))
+ (ctx->VertexProgram._Current && ctx->VertexProgram.TwoSideEnabled))
ind |= SS_TWOSIDE_BIT;
/* We piggyback the two-sided stencil front/back determination on the
@@ -311,6 +313,4 @@ void _swsetup_choose_trifuncs( GLcontext *ctx )
tnl->Driver.Render.Quad = quad_tab[ind];
tnl->Driver.Render.Line = swsetup_line;
tnl->Driver.Render.Points = swsetup_points;
-
- ctx->_Facing = 0;
}
diff --git a/src/mesa/swrast_setup/ss_tritmp.h b/src/mesa/swrast_setup/ss_tritmp.h
index 59c534e..0e0cbce 100644
--- a/src/mesa/swrast_setup/ss_tritmp.h
+++ b/src/mesa/swrast_setup/ss_tritmp.h
@@ -43,7 +43,6 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
v[1] = &verts[e1];
v[2] = &verts[e2];
-
if (IND & (SS_TWOSIDE_BIT | SS_OFFSET_BIT | SS_UNFILLED_BIT))
{
GLfloat ex = v[0]->win[0] - v[2]->win[0];
@@ -55,7 +54,6 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
if (IND & (SS_TWOSIDE_BIT | SS_UNFILLED_BIT))
{
facing = (cc < 0.0) ^ ctx->Polygon._FrontBit;
- ctx->_Facing = facing;
if (IND & SS_UNFILLED_BIT)
mode = facing ? ctx->Polygon.BackMode : ctx->Polygon.FrontMode;
diff --git a/src/mesa/tnl/t_vb_program.c b/src/mesa/tnl/t_vb_program.c
index 9961af7..03e1ad2 100644
--- a/src/mesa/tnl/t_vb_program.c
+++ b/src/mesa/tnl/t_vb_program.c
@@ -1,6 +1,6 @@
/*
* Mesa 3-D graphics library
- * Version: 6.5.3
+ * Version: 7.1
*
* Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
@@ -30,29 +30,150 @@
*/
-#include "glheader.h"
-#include "colormac.h"
-#include "context.h"
-#include "macros.h"
-#include "imports.h"
-#include "prog_instruction.h"
-#include "prog_statevars.h"
-#include "prog_execute.h"
+#include "main/glheader.h"
+#include "main/colormac.h"
+#include "main/context.h"
+#include "main/macros.h"
+#include "main/imports.h"
+#include "shader/prog_instruction.h"
+#include "shader/prog_statevars.h"
+#include "shader/prog_execute.h"
+#include "swrast/s_context.h"
+#include "swrast/s_texfilter.h"
-#include "tnl.h"
-#include "t_context.h"
-#include "t_pipeline.h"
+#include "tnl/tnl.h"
+#include "tnl/t_context.h"
+#include "tnl/t_pipeline.h"
#include "swrast/s_context.h"
#include "swrast/s_texfilter.h"
+
+/*!
+ * Private storage for the vertex program pipeline stage.
+ */
+struct vp_stage_data {
+ /** The results of running the vertex program go into these arrays. */
+ GLvector4f results[VERT_RESULT_MAX];
+
+ GLvector4f ndcCoords; /**< normalized device coords */
+ GLubyte *clipmask; /**< clip flags */
+ GLubyte ormask, andmask; /**< for clipping */
+};
+
+
+#define VP_STAGE_DATA(stage) ((struct vp_stage_data *)(stage->privatePtr))
+
+
+static void
+userclip( GLcontext *ctx,
+ GLvector4f *clip,
+ GLubyte *clipmask,
+ GLubyte *clipormask,
+ GLubyte *clipandmask )
+{
+ GLuint p;
+
+ for (p = 0; p < ctx->Const.MaxClipPlanes; p++) {
+ if (ctx->Transform.ClipPlanesEnabled & (1 << p)) {
+ GLuint nr, i;
+ const GLfloat a = ctx->Transform._ClipUserPlane[p][0];
+ const GLfloat b = ctx->Transform._ClipUserPlane[p][1];
+ const GLfloat c = ctx->Transform._ClipUserPlane[p][2];
+ const GLfloat d = ctx->Transform._ClipUserPlane[p][3];
+ GLfloat *coord = (GLfloat *)clip->data;
+ GLuint stride = clip->stride;
+ GLuint count = clip->count;
+
+ for (nr = 0, i = 0 ; i < count ; i++) {
+ GLfloat dp = (coord[0] * a +
+ coord[1] * b +
+ coord[2] * c +
+ coord[3] * d);
+
+ if (dp < 0) {
+ nr++;
+ clipmask[i] |= CLIP_USER_BIT;
+ }
+
+ STRIDE_F(coord, stride);
+ }
+
+ if (nr > 0) {
+ *clipormask |= CLIP_USER_BIT;
+ if (nr == count) {
+ *clipandmask |= CLIP_USER_BIT;
+ return;
+ }
+ }
+ }
+ }
+}
+
+
+static GLboolean
+do_ndc_cliptest(GLcontext *ctx, struct vp_stage_data *store)
+{
+ TNLcontext *tnl = TNL_CONTEXT(ctx);
+ struct vertex_buffer *VB = &tnl->vb;
+ /* Cliptest and perspective divide. Clip functions must clear
+ * the clipmask.
+ */
+ store->ormask = 0;
+ store->andmask = CLIP_FRUSTUM_BITS;
+
+ if (tnl->NeedNdcCoords) {
+ VB->NdcPtr =
+ _mesa_clip_tab[VB->ClipPtr->size]( VB->ClipPtr,
+ &store->ndcCoords,
+ store->clipmask,
+ &store->ormask,
+ &store->andmask );
+ }
+ else {
+ VB->NdcPtr = NULL;
+ _mesa_clip_np_tab[VB->ClipPtr->size]( VB->ClipPtr,
+ NULL,
+ store->clipmask,
+ &store->ormask,
+ &store->andmask );
+ }
+
+ if (store->andmask) {
+ /* All vertices are outside the frustum */
+ return GL_FALSE;
+ }
+
+ /* Test userclip planes. This contributes to VB->ClipMask.
+ */
+ /** XXX NEW_SLANG _Enabled ??? */
+ if (ctx->Transform.ClipPlanesEnabled && (!ctx->VertexProgram._Enabled ||
+ ctx->VertexProgram.Current->IsPositionInvariant)) {
+ userclip( ctx,
+ VB->ClipPtr,
+ store->clipmask,
+ &store->ormask,
+ &store->andmask );
+
+ if (store->andmask) {
+ return GL_FALSE;
+ }
+ }
+
+ VB->ClipAndMask = store->andmask;
+ VB->ClipOrMask = store->ormask;
+ VB->ClipMask = store->clipmask;
+
+ return GL_TRUE;
+}
+
+
/**
* XXX the texture sampling code in this module is a bit of a hack.
* The texture sampling code is in swrast, though it doesn't have any
* real dependencies on the rest of swrast. It should probably be
* moved into main/ someday.
*/
-
static void
vp_fetch_texel(GLcontext *ctx, const GLfloat texcoord[4], GLfloat lambda,
GLuint unit, GLfloat color[4])
@@ -85,22 +206,6 @@ _tnl_program_string(GLcontext *ctx, GLenum target, struct gl_program *program)
}
-/*!
- * Private storage for the vertex program pipeline stage.
- */
-struct vp_stage_data {
- /** The results of running the vertex program go into these arrays. */
- GLvector4f results[VERT_RESULT_MAX];
-
- GLvector4f ndcCoords; /**< normalized device coords */
- GLubyte *clipmask; /**< clip flags */
- GLubyte ormask, andmask; /**< for clipping */
-};
-
-
-#define VP_STAGE_DATA(stage) ((struct vp_stage_data *)(stage->privatePtr))
-
-
/**
* Initialize virtual machine state prior to executing vertex program.
*/
@@ -143,92 +248,44 @@ init_machine(GLcontext *ctx, struct gl_program_machine *machine)
/**
- * Copy the 16 elements of a matrix into four consecutive program
- * registers starting at 'pos'.
+ * Map the texture images which the vertex program will access (if any).
*/
static void
-load_matrix(GLfloat registers[][4], GLuint pos, const GLfloat mat[16])
+map_textures(GLcontext *ctx, const struct gl_vertex_program *vp)
{
- GLuint i;
- for (i = 0; i < 4; i++) {
- registers[pos + i][0] = mat[0 + i];
- registers[pos + i][1] = mat[4 + i];
- registers[pos + i][2] = mat[8 + i];
- registers[pos + i][3] = mat[12 + i];
- }
-}
+ GLuint u;
+ if (!ctx->Driver.MapTexture)
+ return;
-/**
- * As above, but transpose the matrix.
- */
-static void
-load_transpose_matrix(GLfloat registers[][4], GLuint pos,
- const GLfloat mat[16])
-{
- MEMCPY(registers[pos], mat, 16 * sizeof(GLfloat));
+ for (u = 0; u < ctx->Const.MaxVertexTextureImageUnits; u++) {
+ if (vp->Base.TexturesUsed[u]) {
+ /* Note: _Current *should* correspond to the target indicated
+ * in TexturesUsed[u].
+ */
+ ctx->Driver.MapTexture(ctx, ctx->Texture.Unit[u]._Current);
+ }
+ }
}
/**
- * Load current vertex program's parameter registers with tracked
- * matrices (if NV program). This only needs to be done per
- * glBegin/glEnd, not per-vertex.
+ * Unmap the texture images which were used by the vertex program (if any).
*/
-void
-_mesa_load_tracked_matrices(GLcontext *ctx)
+static void
+unmap_textures(GLcontext *ctx, const struct gl_vertex_program *vp)
{
- GLuint i;
+ GLuint u;
- for (i = 0; i < MAX_NV_VERTEX_PROGRAM_PARAMS / 4; i++) {
- /* point 'mat' at source matrix */
- GLmatrix *mat;
- if (ctx->VertexProgram.TrackMatrix[i] == GL_MODELVIEW) {
- mat = ctx->ModelviewMatrixStack.Top;
- }
- else if (ctx->VertexProgram.TrackMatrix[i] == GL_PROJECTION) {
- mat = ctx->ProjectionMatrixStack.Top;
- }
- else if (ctx->VertexProgram.TrackMatrix[i] == GL_TEXTURE) {
- mat = ctx->TextureMatrixStack[ctx->Texture.CurrentUnit].Top;
- }
- else if (ctx->VertexProgram.TrackMatrix[i] == GL_COLOR) {
- mat = ctx->ColorMatrixStack.Top;
- }
- else if (ctx->VertexProgram.TrackMatrix[i]==GL_MODELVIEW_PROJECTION_NV) {
- /* XXX verify the combined matrix is up to date */
- mat = &ctx->_ModelProjectMatrix;
- }
- else if (ctx->VertexProgram.TrackMatrix[i] >= GL_MATRIX0_NV &&
- ctx->VertexProgram.TrackMatrix[i] <= GL_MATRIX7_NV) {
- GLuint n = ctx->VertexProgram.TrackMatrix[i] - GL_MATRIX0_NV;
- ASSERT(n < MAX_PROGRAM_MATRICES);
- mat = ctx->ProgramMatrixStack[n].Top;
- }
- else {
- /* no matrix is tracked, but we leave the register values as-is */
- assert(ctx->VertexProgram.TrackMatrix[i] == GL_NONE);
- continue;
- }
+ if (!ctx->Driver.MapTexture)
+ return;
- /* load the matrix values into sequential registers */
- if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_IDENTITY_NV) {
- load_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
- }
- else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_INVERSE_NV) {
- _math_matrix_analyse(mat); /* update the inverse */
- ASSERT(!_math_matrix_is_dirty(mat));
- load_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
- }
- else if (ctx->VertexProgram.TrackMatrixTransform[i] == GL_TRANSPOSE_NV) {
- load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->m);
- }
- else {
- assert(ctx->VertexProgram.TrackMatrixTransform[i]
- == GL_INVERSE_TRANSPOSE_NV);
- _math_matrix_analyse(mat); /* update the inverse */
- ASSERT(!_math_matrix_is_dirty(mat));
- load_transpose_matrix(ctx->VertexProgram.Parameters, i*4, mat->inv);
+ for (u = 0; u < ctx->Const.MaxVertexTextureImageUnits; u++) {
+ if (vp->Base.TexturesUsed[u]) {
+ /* Note: _Current *should* correspond to the target indicated
+ * in TexturesUsed[u].
+ */
+ ctx->Driver.UnmapTexture(ctx, ctx->Texture.Unit[u]._Current);
}
}
}
@@ -259,6 +316,7 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
_mesa_load_state_parameters(ctx, program->Base.Parameters);
}
+ /* make list of outputs to save some time below */
numOutputs = 0;
for (i = 0; i < VERT_RESULT_MAX; i++) {
if (program->Base.OutputsWritten & (1 << i)) {
@@ -266,6 +324,8 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
}
}
+ map_textures(ctx, program);
+
for (i = 0; i < VB->Count; i++) {
GLuint attr;
@@ -317,6 +377,8 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
#endif
}
+ unmap_textures(ctx, program);
+
/* Fixup fog and point size results if needed */
if (program->IsNVProgram) {
if (ctx->Fog.Enabled &&
@@ -334,12 +396,39 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
}
}
- /* Setup the VB pointers so that the next pipeline stages get
- * their data from the right place (the program output arrays).
- */
- VB->ClipPtr = &store->results[VERT_RESULT_HPOS];
- VB->ClipPtr->size = 4;
- VB->ClipPtr->count = VB->Count;
+ if (program->IsPositionInvariant) {
+ /* We need the exact same transform as in the fixed function path here
+ * to guarantee invariance, depending on compiler optimization flags
+ * results could be different otherwise.
+ */
+ VB->ClipPtr = TransformRaw( &store->results[0],
+ &ctx->_ModelProjectMatrix,
+ VB->AttribPtr[0] );
+
+ /* Drivers expect this to be clean to element 4...
+ */
+ switch (VB->ClipPtr->size) {
+ case 1:
+ /* impossible */
+ case 2:
+ _mesa_vector4f_clean_elem( VB->ClipPtr, VB->Count, 2 );
+ /* fall-through */
+ case 3:
+ _mesa_vector4f_clean_elem( VB->ClipPtr, VB->Count, 3 );
+ /* fall-through */
+ case 4:
+ break;
+ }
+ }
+ else {
+ /* Setup the VB pointers so that the next pipeline stages get
+ * their data from the right place (the program output arrays).
+ */
+ VB->ClipPtr = &store->results[VERT_RESULT_HPOS];
+ VB->ClipPtr->size = 4;
+ VB->ClipPtr->count = VB->Count;
+ }
+
VB->ColorPtr[0] = &store->results[VERT_RESULT_COL0];
VB->ColorPtr[1] = &store->results[VERT_RESULT_BFC0];
VB->SecondaryColorPtr[0] = &store->results[VERT_RESULT_COL1];
@@ -365,41 +454,10 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
}
}
- /* Cliptest and perspective divide. Clip functions must clear
- * the clipmask.
- */
- store->ormask = 0;
- store->andmask = CLIP_FRUSTUM_BITS;
-
- if (tnl->NeedNdcCoords) {
- VB->NdcPtr =
- _mesa_clip_tab[VB->ClipPtr->size]( VB->ClipPtr,
- &store->ndcCoords,
- store->clipmask,
- &store->ormask,
- &store->andmask );
- }
- else {
- VB->NdcPtr = NULL;
- _mesa_clip_np_tab[VB->ClipPtr->size]( VB->ClipPtr,
- NULL,
- store->clipmask,
- &store->ormask,
- &store->andmask );
- }
-
- if (store->andmask) /* All vertices are outside the frustum */
- return GL_FALSE;
-
- /* This is where we'd do clip testing against the user-defined
- * clipping planes, but they're not supported by vertex programs.
+ /* Perform NDC and cliptest operations:
*/
-
- VB->ClipOrMask = store->ormask;
- VB->ClipMask = store->clipmask;
-
- return GL_TRUE;
+ return do_ndc_cliptest(ctx, store);
}
diff --git a/src/mesa/tnl/tnl.h b/src/mesa/tnl/tnl.h
index 20bed55..c8c0093 100644
--- a/src/mesa/tnl/tnl.h
+++ b/src/mesa/tnl/tnl.h
@@ -1,9 +1,8 @@
-
/*
* Mesa 3-D graphics library
- * Version: 3.5
+ * Version: 7.1
*
- * Copyright (C) 1999-2001 Brian Paul All Rights Reserved.
+ * Copyright (C) 1999-2007 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -82,7 +81,4 @@ _tnl_draw_prims( GLcontext *ctx,
GLuint min_index,
GLuint max_index);
-extern void
-_mesa_load_tracked_matrices(GLcontext *ctx);
-
#endif
diff --git a/src/mesa/tnl_dd/t_dd_tritmp.h b/src/mesa/tnl_dd/t_dd_tritmp.h
index 0bf32ff..6c2378f 100644
--- a/src/mesa/tnl_dd/t_dd_tritmp.h
+++ b/src/mesa/tnl_dd/t_dd_tritmp.h
@@ -155,10 +155,6 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
{
facing = AREA_IS_CCW( cc ) ^ ctx->Polygon._FrontBit;
- if (DO_TWOSTENCIL && ctx->Stencil.TestTwoSide) {
- ctx->_Facing = facing; /* mixed mode rendering: for 2-sided stencil test */
- }
-
if (DO_UNFILLED) {
if (facing) {
mode = ctx->Polygon.BackMode;
@@ -421,10 +417,6 @@ static void TAG(quad)( GLcontext *ctx,
{
facing = AREA_IS_CCW( cc ) ^ ctx->Polygon._FrontBit;
- if (DO_TWOSTENCIL && ctx->Stencil.TestTwoSide) {
- ctx->_Facing = facing; /* mixed mode rendering: for 2-sided stencil test */
- }
-
if (DO_UNFILLED) {
if (facing) {
mode = ctx->Polygon.BackMode;