Zou Nan hai [Wed, 4 Jul 2007 02:51:57 +0000 (10:51 +0800)]
fix issue with output as src register.
Zou Nan hai [Tue, 3 Jul 2007 07:27:43 +0000 (15:27 +0800)]
Fix a nasty bug...
Zou Nan hai [Thu, 21 Jun 2007 02:22:28 +0000 (10:22 +0800)]
support branch and loop in pixel shader
most of the sample working with some small modification
Zou Nan hai [Wed, 9 May 2007 02:08:10 +0000 (10:08 +0800)]
Support loop, conditional update fix
Zou Nan hai [Mon, 30 Apr 2007 07:24:38 +0000 (15:24 +0800)]
support nested function call
else instruction fix.
Zou Nan hai [Thu, 12 Apr 2007 01:43:00 +0000 (09:43 +0800)]
Initial 965 GLSL support
Brian [Wed, 11 Apr 2007 18:30:31 +0000 (12:30 -0600)]
regenerated
Brian [Wed, 11 Apr 2007 18:30:03 +0000 (12:30 -0600)]
checkpoint: updating non-square matrix constructors and operators
Brian [Wed, 11 Apr 2007 18:29:38 +0000 (12:29 -0600)]
move/fix texture sampling funcs
Brian [Wed, 11 Apr 2007 17:13:37 +0000 (11:13 -0600)]
s/Shadow/SHADOW/
Brian [Wed, 11 Apr 2007 15:04:18 +0000 (09:04 -0600)]
as with prev commit, also fix color index mode zoomed depth copies (bug 10608)
Brian [Wed, 11 Apr 2007 15:00:56 +0000 (09:00 -0600)]
fix invalid error detection problem in _mesa_bind_attrib_location(), bug 10602
Brian [Wed, 11 Apr 2007 14:47:05 +0000 (08:47 -0600)]
fix zoomed depth copies (bug 10608)
Brian [Wed, 11 Apr 2007 14:11:52 +0000 (08:11 -0600)]
use _mesa_reference_renderbuffer(), fix typo
Brian [Wed, 11 Apr 2007 03:51:27 +0000 (21:51 -0600)]
fix/work-around allocation bugs for non-square matrices
Brian [Wed, 11 Apr 2007 03:50:53 +0000 (21:50 -0600)]
update SLANG version test
Dave Airlie [Tue, 10 Apr 2007 10:46:58 +0000 (11:46 +0100)]
rs480: set vap cntl to what fglrx uses for non-TCL cards
Michel Dänzer [Tue, 10 Apr 2007 09:05:17 +0000 (11:05 +0200)]
i915: Bring test for vsync to pipe B in line with i915tex.
Michel Dänzer [Tue, 10 Apr 2007 09:04:19 +0000 (11:04 +0200)]
Add missing generated file.
Brian [Tue, 10 Apr 2007 01:07:13 +0000 (19:07 -0600)]
NULL ptr check
Brian [Tue, 10 Apr 2007 01:04:45 +0000 (19:04 -0600)]
plug in GLSL 1.20 unit
Brian [Tue, 10 Apr 2007 00:36:55 +0000 (18:36 -0600)]
undo a debug/test change
Dave Airlie [Mon, 9 Apr 2007 23:32:13 +0000 (09:32 +1000)]
r300: don't enable VAP/TCL on cards that don't support it
Brian [Mon, 9 Apr 2007 17:21:51 +0000 (11:21 -0600)]
remove unneeded initialization code (see bug 10569)
Brian [Mon, 9 Apr 2007 14:49:26 +0000 (08:49 -0600)]
allocate __GLXdisplayPrivate w/ Xcalloc instead of Xmalloc (bug 10569)
Michel Dänzer [Mon, 9 Apr 2007 12:57:08 +0000 (14:57 +0200)]
i915tex: Fix some mismatches between texels or bytes for pitch/stride.
Michel Dänzer [Mon, 9 Apr 2007 12:52:18 +0000 (14:52 +0200)]
i915tex: Make sure texture format fetch hooks are initialized.
Dave Airlie [Mon, 9 Apr 2007 12:03:31 +0000 (22:03 +1000)]
the RS400 definitely doesn't work at this point so don't let it init
Brian [Mon, 9 Apr 2007 03:48:19 +0000 (21:48 -0600)]
enable GLSL 1.20
Brian [Sun, 8 Apr 2007 22:47:32 +0000 (16:47 -0600)]
support for GLSL 1.20 non-square matrices
Brian [Sun, 8 Apr 2007 21:50:31 +0000 (15:50 -0600)]
add glsl to PROGRAM_DIRS
Brian [Sun, 8 Apr 2007 21:42:26 +0000 (15:42 -0600)]
fix some dependencies, remove some CFLAGS
Brian [Sun, 8 Apr 2007 21:41:02 +0000 (15:41 -0600)]
add 2.1 functions
Adam Jackson [Sun, 8 Apr 2007 18:40:03 +0000 (14:40 -0400)]
Make sure GLX entrypoints are marked PUBLIC.
Fedora bug #229808.
Wei Wang [Sun, 8 Apr 2007 17:24:08 +0000 (11:24 -0600)]
fix bug 9823: GL_CLIENT_ATTRIB_STACK_DEPTH query fails for indirect rendering
Michel Dänzer [Sun, 8 Apr 2007 12:04:15 +0000 (14:04 +0200)]
i915tex: Clean up resizing of renderbuffers.
Michel Dänzer [Sun, 8 Apr 2007 12:02:14 +0000 (14:02 +0200)]
driUpdateFramebufferSize: Use ctx->Driver.ResizeBuffers.
Brian [Sun, 8 Apr 2007 02:38:25 +0000 (20:38 -0600)]
MAX_TEXTURE_COORDS_ARB and MAX_TEXTURE_IMAGE_UNITS_ARB had wrong glGet info.
Fixes bug 10371.
Brian [Sun, 8 Apr 2007 01:36:58 +0000 (19:36 -0600)]
fix etags command to find headers
George Sapountzis [Sat, 7 Apr 2007 15:41:14 +0000 (18:41 +0300)]
Drop XFree86Server from execmem.c
This is a remnant from the libcwrapper days that can be removed now. It is the
last XFree86 glitch in Mesa core.
Reverts the following commit (thanks to Brian Paul for pointing to it):
commit
8b1dc68662f2a18ea4e02c6d3247c5d1da741aa8
Author: Alan Hourihane <alanh@tungstengraphics.com>
Date: Tue Jan 24 21:40:53 2006 +0000
pull in fix from 6.4 branch for XFree86Server definition
commit
cc7b6810e2fe6ab485fd96d5f17bf5d9329bd153
Author: Ian Romanick <idr@us.ibm.com>
Date: Fri Oct 21 18:09:24 2005 +0000
Make execmem.c compile in the server (libGLcore). Since xf86mmap
doesn't know about MAP_ANONYMOUS, this is the only viable fix. This
issue will likely have to be revisited at some point.
Richard Hughes [Sat, 7 Apr 2007 14:48:07 +0000 (08:48 -0600)]
set version for tarballs to 6.5.3pre for now
Richard Hughes [Sat, 7 Apr 2007 14:46:44 +0000 (08:46 -0600)]
fix references to non-existant sz4 field
Richard Hughes [Sat, 7 Apr 2007 14:46:10 +0000 (08:46 -0600)]
fix some CHAN_TYPE==GL_FLOAT breakage
Brian [Sat, 7 Apr 2007 14:43:40 +0000 (08:43 -0600)]
fix comment typo
Richard Hughes [Sat, 7 Apr 2007 14:39:24 +0000 (08:39 -0600)]
remove array_cache/, add vbo/
Brian [Sat, 7 Apr 2007 14:38:36 +0000 (08:38 -0600)]
added glActiveStencilFaceEXT
Brian [Fri, 6 Apr 2007 21:45:11 +0000 (15:45 -0600)]
for evaluators, loop to VBO_ATTRIB_TEX7, not VBO_ATTRIB_INDEX. See bug 10543
George Sapountzis [Fri, 6 Apr 2007 10:28:39 +0000 (13:28 +0300)]
xmesa: use newly added xm_image.[hc]
Keep external includes to glxheader.h and xmesa includes to xmesaP.h.
Drop the following from xm_image.h:
- dix-config.h (comes from glheader.h)
- xfree86 includes (come from GL/xmesa_xf86.h)
- ifdef __CYGWIN__ (leftover ? xm_api.c has something similar)
George Sapountzis [Fri, 6 Apr 2007 10:27:43 +0000 (13:27 +0300)]
xmesa: add xf86glx_util.[hc] from xserver as xm_image.[hc]
George Sapountzis [Fri, 6 Apr 2007 10:22:41 +0000 (13:22 +0300)]
Drop the funky SSE exception test on linux.
Replace the check for IN_DRI_DRIVER with the appropriate kernel version check
and just disable SSE on older (pre 2.4) kernels.
Brian [Fri, 6 Apr 2007 03:12:44 +0000 (21:12 -0600)]
fix color storage bug in convolution path
Brian [Thu, 5 Apr 2007 17:43:07 +0000 (11:43 -0600)]
include points.h to fix warnings
George Sapountzis [Thu, 5 Apr 2007 16:01:51 +0000 (19:01 +0300)]
Remove SI imports/exports remnants.
George Sapountzis [Thu, 5 Apr 2007 16:00:47 +0000 (19:00 +0300)]
glx: minor cosmetic in glcontextmodes.c
move memory macros to separate block and unwrap malloc/free for miniglx towards
cleaning Mesa core glitches in glx...
George Sapountzis [Thu, 5 Apr 2007 15:59:42 +0000 (18:59 +0300)]
Move glcontextmodes.c to glx.
It is no longer linked with DRI drivers, libGL passes function pointers through
the DRI interface.
Brian [Thu, 5 Apr 2007 15:28:09 +0000 (09:28 -0600)]
temporary add some extra renderbuffer debug code
Brian [Thu, 5 Apr 2007 04:18:53 +0000 (22:18 -0600)]
Remove the never-used SI-style imports/exports code.
Brian [Wed, 4 Apr 2007 15:33:12 +0000 (09:33 -0600)]
assorted updates
Brian [Wed, 4 Apr 2007 15:31:41 +0000 (09:31 -0600)]
assorted documentation updates
Brian [Wed, 4 Apr 2007 15:30:28 +0000 (09:30 -0600)]
init span.y=0 to silence warnings
Brian [Wed, 4 Apr 2007 14:48:06 +0000 (08:48 -0600)]
Overlapping copies with zoomZ=-1 were broken. See bug 10521.
Need to check for overlapping src/dest regions before computing bottom-to-top
vs. top-to-bottom order.
Brian [Wed, 4 Apr 2007 14:26:41 +0000 (08:26 -0600)]
remove debug printf
Xiang, Haihao [Wed, 4 Apr 2007 08:50:47 +0000 (16:50 +0800)]
i810/i915/i915tex: reinitialize the context point state
Nicolai Haehnle [Mon, 2 Apr 2007 18:59:51 +0000 (20:59 +0200)]
Fix in t_vp_build: Missed necessary updates sometimes
_mesa_update_state doesn't always reset VertexProgram._Current to NULL.
This caused us to skip a necessary update of the fixed function vertex
program sometimes.
Change the logic such that we check for updates whenever the _Current
program is either NULL or the program generated by t_vp_build.
Brian [Mon, 2 Apr 2007 16:03:34 +0000 (10:03 -0600)]
Call _glapi_set_context(NULL) in _glapi_check_multithread() to make sure the _glapi_Context global gets cleared properly.
Brian [Mon, 2 Apr 2007 15:56:28 +0000 (09:56 -0600)]
use _mesa_reference_renderbuffer() in a few more places
Brian [Mon, 2 Apr 2007 15:55:55 +0000 (09:55 -0600)]
updated debug printf
Brian [Mon, 2 Apr 2007 00:30:28 +0000 (18:30 -0600)]
Fix some bugs/issues related to alpha channel support. See bug 10483.
Brian [Mon, 2 Apr 2007 00:28:28 +0000 (18:28 -0600)]
If using PF_8A8B8G8R, be sure alpha is correctly handled.
Brian [Mon, 2 Apr 2007 00:27:23 +0000 (18:27 -0600)]
re-order tokens, fix comments
Brian [Sat, 31 Mar 2007 15:09:48 +0000 (09:09 -0600)]
always emit BGN/ENDSUB, for now anyway
Brian [Sat, 31 Mar 2007 15:09:21 +0000 (09:09 -0600)]
fix scoping mistake in previous commit that checked for writable LHSs
Brian [Fri, 30 Mar 2007 20:52:23 +0000 (14:52 -0600)]
check that LHS of assignment is writable
Eric Anholt [Fri, 30 Mar 2007 20:18:27 +0000 (13:18 -0700)]
Merge branch 'origin'
Eric Anholt [Fri, 30 Mar 2007 20:08:14 +0000 (13:08 -0700)]
Merge branch 'crestline-qa', adding support for the 965GM chipset.
Gustavo Pichorim Boiko [Fri, 30 Mar 2007 20:10:10 +0000 (14:10 -0600)]
call DRI_VALIDATE_DRAWABLE_INFO(), bug 10477
George Sapountzis [Wed, 6 Dec 2006 04:49:28 +0000 (06:49 +0200)]
Clean and update XMesa/XFree86 interface.
Drop XMesaSetVisualDisplay(), XMesaReset(), no longer used.
Add XMesaCopyContext() and move the GlxSetRenderTables() call for XGL within
XMesaForceCurrent(). This is to make xserver/GL/mesa/X/xf86glx.c unaware of
Mesa internals.
Also, clean some ifdef's to make it clear that USE_XSHM and XFree86Server are
mutually exclusive.
Lastly,
- move gcstruct.h from glxheader.h up to xmesa_xf86.h since it calls *gc->ops
- drop GL/glxtokens.h from xm_api|dd.c, GLX tokens come from glcore.h and are
used irrelevant of XFree86.
Brian [Wed, 28 Mar 2007 23:14:35 +0000 (17:14 -0600)]
more tips, validation info
Brian [Wed, 28 Mar 2007 23:08:17 +0000 (17:08 -0600)]
cond code fix
Brian [Wed, 28 Mar 2007 21:45:24 +0000 (15:45 -0600)]
fix cond code swizzle bug
Brian [Wed, 28 Mar 2007 21:16:33 +0000 (15:16 -0600)]
Handle logical NOT and XOR without library functions. Results in much tighter code.
Brian [Wed, 28 Mar 2007 20:49:33 +0000 (14:49 -0600)]
optimization for the emit_not() function
Brian [Wed, 28 Mar 2007 20:38:36 +0000 (14:38 -0600)]
remove IR_BREAK_IF_FALSE
Brian [Wed, 28 Mar 2007 20:33:25 +0000 (14:33 -0600)]
don't generate IR_BREAK_IF_FALSE
Brian [Wed, 28 Mar 2007 20:25:30 +0000 (14:25 -0600)]
Get rid of IR_CONT_IF_FALSE
Brian [Wed, 28 Mar 2007 20:23:33 +0000 (14:23 -0600)]
Get rid of BRK0, BRK1, CONT0, CONT1 instructions.
Brian [Wed, 28 Mar 2007 20:21:26 +0000 (14:21 -0600)]
Don't emit OPCODE_CONT0/1, BRK0/1 instructions, clean-ups elsewhere.
Brian [Wed, 28 Mar 2007 20:14:00 +0000 (14:14 -0600)]
simplify, clean-up break/cont code
Brian [Wed, 28 Mar 2007 19:29:57 +0000 (13:29 -0600)]
added some null ptr checks to handle error recovery
Brian [Wed, 28 Mar 2007 19:23:44 +0000 (13:23 -0600)]
Use constant_to_src_reg() to simplify some code
Brian [Wed, 28 Mar 2007 18:48:27 +0000 (12:48 -0600)]
added missing returns after slang_info_log_error() calls
Brian [Wed, 28 Mar 2007 17:06:34 +0000 (11:06 -0600)]
Fix, clean-up code related to comparisons, condition codes, etc.
Brian [Wed, 28 Mar 2007 17:01:28 +0000 (11:01 -0600)]
print condcodes if DEBUG_PROG
Brian [Wed, 28 Mar 2007 17:01:09 +0000 (11:01 -0600)]
expose _mesa_condcode_string(), fix some printing
Brian [Wed, 28 Mar 2007 16:44:38 +0000 (10:44 -0600)]
check that if/while/do-while condition is boolean or scalar
Brian [Wed, 28 Mar 2007 15:07:41 +0000 (09:07 -0600)]
bump version to 6.5.3 to match version.h
Alan Hourihane [Wed, 28 Mar 2007 14:58:08 +0000 (15:58 +0100)]
Add missing OSMesaColorClamp function
Brian [Tue, 27 Mar 2007 22:06:20 +0000 (16:06 -0600)]
disable MakeTexture()
Brian [Tue, 27 Mar 2007 22:05:25 +0000 (16:05 -0600)]
added ftransform() comment
Brian [Tue, 27 Mar 2007 22:04:36 +0000 (16:04 -0600)]
Implement true CAL/RET subroutines. Some optimizations, clean-ups coming...