mesa3d-dev

Re: [Mesa3d-dev] A small patch for mklib

[Dan N. <dbn...@gm...>]

On Sat, Feb 27, 2010 at 12:13 PM, Alex Weiss <al...@ca...> wrote:
> Hello everyone,
> I ran into a bit of trouble building Mesa on Mac OS X this afternoon. Turns
> out that when mklib was given one archive to merge into another on darwin,
> it was failing to unpack the first into it's component object files before
> calling ar, so we were winding up with libmesa.a containing the
> archive libglslpp.a itself rather than it's contents. Looks like it's
> related to http://bugs.freedesktop.org/show_bug.cgi?id=25735; the error I
> was getting is the same as the original reporter of that bug. In any case, I
> have a patch for mklib (attached) that fixes the problem, but I'm not
> certain where to send it.

Looks good, although I stripped out the commented commands/variables.
Probably at this point all the arches should be converted to doing the
expanded archives since they're all breaking, but that's another
patch.

Applied.

--
Dan

[Mesa3d-dev] [Bug 26788] Opengl Gallium driver for r500 crashes after "Loading Area" in NWN

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26788


Marek Olšák <ma...@gm...> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED




-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Mesa3d-dev] [Bug 26788] Opengl Gallium driver for r500 crashes after "Loading Area" in NWN

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26788


Marek Olšák <ma...@gm...> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Marek Olšák <ma...@gm...>  2010-02-27 15:50:52 PST ---
Thanks for testing, I pushed the patch to master. Closing...


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

Re: [Mesa3d-dev] [PATCH] st/mesa: do not advertise S3TC if the external lib is not available

[Marek O. <ma...@gm...>]

OK thanks, I pushed what you suggested.

-Marek

On Sat, Feb 27, 2010 at 9:44 PM, José Fonseca <jfo...@vm...> wrote:

> On Sat, 2010-02-27 at 09:15 -0800, Marek Olšák wrote:
> > Hi,
> >
> > S3TC shouldn't be advertised without the external lib and the attached
> > patch fixes that. Please review.
>
> Hi Marek,
>
> Actually it is a bit more subtle.
>
> S3TC can and should be advertised if the pipe driver supports ST3C
> decompression and *compression*.
>
> No real hardware has circuitry for S3TC compression, granted. But the
> vwmare svga gallium driver does it, by deferring everything to the host.
>
> It was the best solution we found to supporting S3TC without having to
> rely on closed source or questionable bits in the guest.
>
> So the right fix should be:
>
> screen->is_format_supported(screen, PIPE_FORMAT_DXT5_RGBA,
>                                    PIPE_TEXTURE_2D,
>                                    PIPE_TEXTURE_USAGE_SAMPLER, 0) &&
> (ctx->Mesa_DXTn || screen->is_format_supported(screen,
> PIPE_FORMAT_DXT5_RGBA, PIPE_TEXTURE_2D, PIPE_TEXTURE_USAGE_RENDERTARGET,
> 0)
>
> Jose
>
>

Re: [Mesa3d-dev] [PATCH] mesa: remove libmesagallium.a on make clean

[José F. <jfo...@vm...>]

On Sat, 2010-02-27 at 13:55 -0800, Marcin Slusarz wrote:
> ---
>  src/mesa/Makefile |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/src/mesa/Makefile b/src/mesa/Makefile
> index 0cb49e8..8c0ebf8 100644
> --- a/src/mesa/Makefile
> +++ b/src/mesa/Makefile
> @@ -154,7 +154,7 @@ tags:
>  clean:
>  	-rm -f */*.o
>  	-rm -f */*/*.o
> -	-rm -f depend depend.bak libmesa.a libglapi.a
> +	-rm -f depend depend.bak libmesa.a libglapi.a libmesagallium.a
>  	-rm -f drivers/*/*.o
>  	-rm -f *.pc
>  	-rm -f shader/slang/library/*_gc.h

Commited. Thanks.

Jose

[Mesa3d-dev] [Bug 26788] Opengl Gallium driver for r500 crashes after "Loading Area" in NWN

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26788





--- Comment #3 from Krzysztof A. Sobiecki <so...@gm...>  2010-02-27 14:35:10 PST ---
After applying your patch, I was unable to reproduce this bug any more. I
didn't saw any regressions so far.


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Mesa3d-dev] [PATCH] mesa: remove libmesagallium.a on make clean

[Marcin S. <mar...@gm...>]

---
 src/mesa/Makefile |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/src/mesa/Makefile b/src/mesa/Makefile
index 0cb49e8..8c0ebf8 100644
--- a/src/mesa/Makefile
+++ b/src/mesa/Makefile
@@ -154,7 +154,7 @@ tags:
 clean:
 	-rm -f */*.o
 	-rm -f */*/*.o
-	-rm -f depend depend.bak libmesa.a libglapi.a
+	-rm -f depend depend.bak libmesa.a libglapi.a libmesagallium.a
 	-rm -f drivers/*/*.o
 	-rm -f *.pc
 	-rm -f shader/slang/library/*_gc.h
-- 
1.6.6.1

Re: [Mesa3d-dev] [PATCH] st/mesa: do not advertise S3TC if the external lib is not available

[José F. <jfo...@vm...>]

On Sat, 2010-02-27 at 09:15 -0800, Marek Olšák wrote:
> Hi,
> 
> S3TC shouldn't be advertised without the external lib and the attached
> patch fixes that. Please review.

Hi Marek,

Actually it is a bit more subtle.

S3TC can and should be advertised if the pipe driver supports ST3C
decompression and *compression*.

No real hardware has circuitry for S3TC compression, granted. But the
vwmare svga gallium driver does it, by deferring everything to the host.

It was the best solution we found to supporting S3TC without having to
rely on closed source or questionable bits in the guest.

So the right fix should be:

screen->is_format_supported(screen, PIPE_FORMAT_DXT5_RGBA,
                                    PIPE_TEXTURE_2D, 
                                    PIPE_TEXTURE_USAGE_SAMPLER, 0) &&
(ctx->Mesa_DXTn || screen->is_format_supported(screen,
PIPE_FORMAT_DXT5_RGBA, PIPE_TEXTURE_2D, PIPE_TEXTURE_USAGE_RENDERTARGET,
0)

Jose

[Mesa3d-dev] [PATCH] st/mesa: do not advertise S3TC if the external lib is not available

[Marek O. <ma...@gm...>]

Hi,

S3TC shouldn't be advertised without the external lib and the attached patch
fixes that. Please review.

-Marek

Re: [Mesa3d-dev] Mesa (master): r300/compiler: Assert that array index is not negative.

[Marek O. <ma...@gm...>]

We actually hit this assertion when we get:

FRAG
DCL IN[0], COLOR, PERSPECTIVE
  0: END

See piglit/glsl-bug-22603.

-Marek

On Fri, Feb 26, 2010 at 10:18 AM, Corbin Simpson
<mos...@gm...>wrote:

> > Module: Mesa
> > Branch: master
> > Commit: e5c691f445e1c02e6e2f75b817b13d7024f7a3a6
> > URL:
> http://cgit.freedesktop.org/mesa/mesa/commit/?id=e5c691f445e1c02e6e2f75b817b13d7024f7a3a6
> >
> > Author: Vinson Lee <vlee at vmware.com>
> > Date:   Fri Feb 26 00:17:03 2010 -0800
> >
> > r300/compiler: Assert that array index is not negative.
> >
> > ---
> >
> > .../drivers/dri/r300/compiler/r500_fragprog_emit.c |    2 ++
> > 1 files changed, 2 insertions(+), 0 deletions(-)
> >
> > diff --git a/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
> b/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
> > index 829f028..710cae7 100644
> > --- a/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
> > +++ b/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
> > @@ -469,6 +469,8 @@ void r500BuildFragmentProgramHwCode(struct
> r300_fragment_program_compiler *compi
> > if (compiler->Base.Error)
> > return;
> >
> > + assert(code->inst_end >= 0);
> > +
> > if ((code->inst[code->inst_end].inst0 & R500_INST_TYPE_MASK) !=
> R500_INST_TYPE_OUT) {
> > /* This may happen when dead-code elimination is disabled or
> > * when most of the fragment program logic is leading to a KIL */
>
> Sorry, is this actually a problem? If this assertion is actually being
> hit, it sure would be nice to hear about it since it. Empty shaders
> shouldn't just be handled with debugging code.
>
> --
> Only fools are easily impressed by what is only
> barely beyond their reach. ~ Unknown
>
> Corbin Simpson
> <Mos...@gm...>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Mesa3d-dev mailing list
> Mes...@li...
> https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
>

[Mesa3d-dev] [Bug 26788] Opengl Gallium driver for r500 crashes after "Loading Area" in NWN

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26788





--- Comment #2 from Marek Olšák <ma...@gm...>  2010-02-27 08:12:58 PST ---
Created an attachment (id=33625)
 --> (http://bugs.freedesktop.org/attachment.cgi?id=33625)
Possible fix

Please let me know if the attached patch helps you.


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Mesa3d-dev] glapi cleanups

[George S. <gsa...@gm...>]

Hi,

I did some glapi cleanups at
http://cgit.freedesktop.org/~gsap7/mesa/log/?h=glapi-cleanup . Please
review.

regards,
George.

[Mesa3d-dev] [Bug 26788] Opengl Gallium driver for r500 crashes after "Loading Area" in NWN

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26788





--- Comment #1 from Krzysztof A. Sobiecki <so...@gm...>  2010-02-27 05:15:54 PST ---
Created an attachment (id=33621)
 --> (http://bugs.freedesktop.org/attachment.cgi?id=33621)
Small patch

It's only a workaround solution.


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Mesa3d-dev] [Bug 26788] New: Opengl Gallium driver for r500 crashes after "Loading Area" in NWN

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26788

           Summary: Opengl Gallium driver for r500 crashes after "Loading
                    Area" in NWN
           Product: Mesa
           Version: git
          Platform: x86-64 (AMD64)
        OS/Version: Linux (All)
            Status: NEW
          Severity: normal
          Priority: medium
         Component: Other
        AssignedTo: mes...@li...
        ReportedBy: so...@gm...


Created an attachment (id=33620)
 --> (http://bugs.freedesktop.org/attachment.cgi?id=33620)
Backtrace

Commit e6632b4bf7cba5fe8a77d54635a3b617fa67185f causes gallium driver to catch
assert in 
*__GI___assert_fail (assertion=0xf6c26f34 "boi->space_accounted",
file=0xf6c26f24 "radeon_cs_gem.c", line=129, function=0xf6c27039
"cs_gem_write_reloc#4  0xf6c258eb in cs_gem_write_reloc () from
/usr/lib32/libdrm_radeon.so.1

Removing assert from radeon_cs_gem.c:129 in drm "fixes" this problem.


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Mesa3d-dev] [PATCH] Nouveau_vieux "fixes"

[<ran...@gm...>]

Hello all!

After unsuccesfull battle with git-send-email I just  send these two patches  from Kmail. Botch as attachments and inlin, but inline 
version probably will be damaged in process.....

Patch 1: add XRGB8888 into nouveau_fbo.c, makes xmoto actually display its demo, not abort

diff --git a/src/mesa/drivers/dri/nouveau/nouveau_fbo.c b/src/mesa/drivers/dri/nouveau/nouveau_fbo.c
index 1db8c5d..8464786 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_fbo.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_fbo.c
@@ -215,6 +215,8 @@ get_tex_format(struct gl_texture_image *ti)
        switch (ti->TexFormat) {
        case MESA_FORMAT_ARGB8888:
                return GL_RGBA8;
+       case MESA_FORMAT_XRGB8888:
+               return GL_RGB8;
        case MESA_FORMAT_RGB565:
                return GL_RGB5;
        default:


Patch 2: add two stencil operation cases in nv04_state_raster.c, allow demos/reflect and demos/dinoshade actually work. Dinoshade 
still visible broken. Not sure about redbook/stencil, it looks very same on my modified driver with TNT2 and with swrast. But 
tests/stencil definitely wrong .... So, all cases are in, one just need to figure out  correct assignment.

diff --git a/src/mesa/drivers/dri/nouveau/nv04_state_raster.c b/src/mesa/drivers/dri/nouveau/nv04_state_raster.c
index 5e3788d..6d0b262 100644
--- a/src/mesa/drivers/dri/nouveau/nv04_state_raster.c
+++ b/src/mesa/drivers/dri/nouveau/nv04_state_raster.c
@@ -61,6 +61,10 @@ get_stencil_op(unsigned op)
        switch (op) {
        case GL_KEEP:
                return 0x1;
+       case GL_ZERO:
+               return 0x2;
+       case GL_REPLACE:
+               return 0x3;
        case GL_INCR:
                return 0x4;
        case GL_DECR:

-----

Tested-off-by: Andrew Randrianasulu <ran...@gm...>

Re: [Mesa3d-dev] Remove color index rendering?

[Ian R. <id...@fr...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ian Romanick wrote:
> While we're on the topic of removing dead weight, can we remove support
> for color index rendering?  None of the hardware drivers support color
> index rendering, and color index rendering is deprecated in OpenGL 3.0
> (and removed in 3.1).
> 
> Can it please die in a fire?

I have just pushed a branch that removes color-index rendering.  It is
available for review in the remove_ci_rendering branch of
git://anongit.freedesktop.org/~idr/mesa.git.  I have done on minimal
testing so far, but I intend to do more over the weekend.  It took most
of the week just to get all the changes made.

There are 38 individual patches.  The patches remove a total of 2,304
lines of code.  I tried to make it easier to review by keeping the
individual changes small.

There are probably some additional changes that could be made.  There is
still a bit of index tracking in TNL and other related places.  I think
some of this can be removed.  Such removals would be more invasive (grep
for MAT_*_INDEX), and may impact our adherence to the spec.  For
example, some queries would return incorrect values.

In any case, I'm satisfied with this set of removals for now.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkuIgUMACgkQX1gOwKyEAw8E+gCdGWJYc1sGhvkeT/Ikg57LcSQb
mNMAn3ks/Xa6Z1z2ziOvqlCP/742OCO+
=oXSE
-----END PGP SIGNATURE-----

Re: [Mesa3d-dev] [PATCH] Fix --enable-32-bit: honor -m32 for progs/glsl.

[Dan N. <dbn...@gm...>]

2010/2/26 Török Edwin <edw...@gm...>:
> On 02/26/2010 08:08 PM, Dan Nicholson wrote:
>>
>> Can you try the patch below? I think this fixes the issue more generally.
>
> Yes, this one fixes the linking issue too. You can commit this one
> instead of mine.

Thanks.

--
Dan

Re: [Mesa3d-dev] [PATCH] Fix --enable-32-bit: honor -m32 for progs/glsl.

[Török E. <edw...@gm...>]

On 02/26/2010 08:08 PM, Dan Nicholson wrote:
> 
> Can you try the patch below? I think this fixes the issue more generally.

Yes, this one fixes the linking issue too. You can commit this one
instead of mine.

> 
> --
> Dan
> 
> From bf51ca551c4e4b994d15f3d6b374c5e5b21ece4e Mon Sep 17 00:00:00 2001
> From: Dan Nicholson <dbn...@gm...>
> Date: Fri, 26 Feb 2010 10:01:37 -0800
> Subject: [PATCH] progs/glsl: Ensure CFLAGS are passed during linking
> 
> Commit ab6825b3e3fd04564b7840e945fa32b0a7b3c0c2 fixed an issue where the
> architecture option -m32 wasn't being passed to the compiler when
> linking .o files. This would only fix the issue with user's building
> with the hacky --enable-32-bit. The right way to fix this is to pass
> CFLAGS to the linker, which the GNU make builtin rules unfortunately
> don't do.
> 
> Signed-off-by: Dan Nicholson <dbn...@gm...>
> ---
>  progs/glsl/Makefile |    5 ++++-
>  1 files changed, 4 insertions(+), 1 deletions(-)
> 
> diff --git a/progs/glsl/Makefile b/progs/glsl/Makefile
> index aaa54cb..3b5a595 100644
> --- a/progs/glsl/Makefile
> +++ b/progs/glsl/Makefile
> @@ -16,7 +16,7 @@ LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLEW_LIB) -l$(GLU_LIB) \
>  # using : to avoid APP_CC pointing to CC loop
>  CC := $(APP_CC)
>  CFLAGS := -I$(INCDIR) $(CFLAGS)
> -LDLIBS = $(LIBS) $(ARCH_FLAGS)
> +LDLIBS = $(LIBS)
>  
>  PROG_SOURCES = \
>  	array.c \
> @@ -69,6 +69,9 @@ $(PROG_OBJS): $(UTIL_HEADERS)
>  
>  $(PROGS): $(UTIL_OBJS)
>  
> +.o:
> +	$(CC) $(CFLAGS) $(LDFLAGS) $^ $(LDLIBS) -o $@
> +
>  clean:
>  	-rm -f $(PROGS)
>  	-rm -f *.o *~

Re: [Mesa3d-dev] [PATCH] Fix --enable-32-bit: honor -m32 for progs/glsl.

[Dan N. <dbn...@gm...>]

On Fri, Feb 26, 2010 at 04:53:48PM +0200, Török Edwin wrote:
> On 2010-02-26 16:50, Dan Nicholson wrote:
> >> gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
> >> -lGL  -lm -m32 -o array
> > 
> > These programs aren't in progs/glsl. Are you sure this is solving your
> > problem? Can you just show me the make output for the failed run?
> 
> They are, I use git master, on this commit:
> commit e5c691f445e1c02e6e2f75b817b13d7024f7a3a6
> Author: Vinson Lee <vl...@vm...>
> Date:   Fri Feb 26 00:17:03 2010 -0800
> 
>     r300/compiler: Assert that array index is not negative.
> 
> ls -l progs/glsl/shaderutil.c progs/glsl/array.c
> -rw-r--r-- 1 edwin edwin 4838 Feb 23 22:26 progs/glsl/array.c
> 
> make[1]: Entering directory `/home/edwin/mesa/progs/glsl'
> cp ../../progs/util/extfuncs.h .
> cp ../../progs/util/shaderutil.h .
> cp ../../progs/util/readtex.h .
> gcc -I../../include -g -O2 -Wall -Wmissing-prototypes -std=c99
> -ffast-math -fvisibility=hidden -fno-strict-aliasing -m32  -fPIC -m32
> -DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -D_GNU_SOURCE
> -DPTHREADS -DHAVE_POSIX_MEMALIGN -DGLX_USE_TLS -DPTHREADS
> -DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING
> -DGLX_INDIRECT_RENDERING -DHAVE_ALIAS   -c -o array.o array.c
> cp ../../progs/util/shaderutil.c .
> gcc -I../../include -g -O2 -Wall -Wmissing-prototypes -std=c99
> -ffast-math -fvisibility=hidden -fno-strict-aliasing -m32  -fPIC -m32
> -DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -D_GNU_SOURCE
> -DPTHREADS -DHAVE_POSIX_MEMALIGN -DGLX_USE_TLS -DPTHREADS
> -DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING
> -DGLX_INDIRECT_RENDERING -DHAVE_ALIAS   -c -o shaderutil.o shaderutil.c
> cp ../../progs/util/readtex.c .
> gcc -I../../include -g -O2 -Wall -Wmissing-prototypes -std=c99
> -ffast-math -fvisibility=hidden -fno-strict-aliasing -m32  -fPIC -m32
> -DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -D_GNU_SOURCE
> -DPTHREADS -DHAVE_POSIX_MEMALIGN -DGLX_USE_TLS -DPTHREADS
> -DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING
> -DGLX_INDIRECT_RENDERING -DHAVE_ALIAS   -c -o readtex.o readtex.c
> gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
> -lGL  -lm -o array
> /usr/bin/ld: skipping incompatible ../../lib32/libglut.so when searching
> for -lglut
> /usr/bin/ld: cannot find -lglut
> collect2: ld returned 1 exit status
> make[1]: *** [array] Error 1
> make[1]: Leaving directory `/home/edwin/mesa/progs/glsl'

Can you try the patch below? I think this fixes the issue more generally.

--
Dan

>From bf51ca551c4e4b994d15f3d6b374c5e5b21ece4e Mon Sep 17 00:00:00 2001
From: Dan Nicholson <dbn...@gm...>
Date: Fri, 26 Feb 2010 10:01:37 -0800
Subject: [PATCH] progs/glsl: Ensure CFLAGS are passed during linking

Commit ab6825b3e3fd04564b7840e945fa32b0a7b3c0c2 fixed an issue where the
architecture option -m32 wasn't being passed to the compiler when
linking .o files. This would only fix the issue with user's building
with the hacky --enable-32-bit. The right way to fix this is to pass
CFLAGS to the linker, which the GNU make builtin rules unfortunately
don't do.

Signed-off-by: Dan Nicholson <dbn...@gm...>
---
 progs/glsl/Makefile |    5 ++++-
 1 files changed, 4 insertions(+), 1 deletions(-)

diff --git a/progs/glsl/Makefile b/progs/glsl/Makefile
index aaa54cb..3b5a595 100644
--- a/progs/glsl/Makefile
+++ b/progs/glsl/Makefile
@@ -16,7 +16,7 @@ LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLEW_LIB) -l$(GLU_LIB) \
 # using : to avoid APP_CC pointing to CC loop
 CC := $(APP_CC)
 CFLAGS := -I$(INCDIR) $(CFLAGS)
-LDLIBS = $(LIBS) $(ARCH_FLAGS)
+LDLIBS = $(LIBS)
 
 PROG_SOURCES = \
 	array.c \
@@ -69,6 +69,9 @@ $(PROG_OBJS): $(UTIL_HEADERS)
 
 $(PROGS): $(UTIL_OBJS)
 
+.o:
+	$(CC) $(CFLAGS) $(LDFLAGS) $^ $(LDLIBS) -o $@
+
 clean:
 	-rm -f $(PROGS)
 	-rm -f *.o *~
-- 
1.6.6.1

[Mesa3d-dev] [Bug 26768] Segmentation fault with indirect rendering

[<bug...@fr...>]

http://bugs.freedesktop.org/show_bug.cgi?id=26768


Brian Paul <bri...@gm...> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Brian Paul <bri...@gm...>  2010-02-26 09:36:21 PST ---
Thanks.  I've committed your fix.
83f1183e769777df5ada4bf7add3390881f88d02


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

Re: [Mesa3d-dev] Mesa (master): r300/compiler: Assert that array index is not negative.

[Corbin S. <mos...@gm...>]

On Fri, Feb 26, 2010 at 1:38 AM, José Fonseca <jfo...@vm...> wrote:
> On Fri, 2010-02-26 at 01:18 -0800, Corbin Simpson wrote:
>> > Module: Mesa
>> > Branch: master
>> > Commit: e5c691f445e1c02e6e2f75b817b13d7024f7a3a6
>> > URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=e5c691f445e1c02e6e2f75b817b13d7024f7a3a6
>> >
>> > Author: Vinson Lee <vlee at vmware.com>
>> > Date:   Fri Feb 26 00:17:03 2010 -0800
>> >
>> > r300/compiler: Assert that array index is not negative.
>> >
>> > ---
>> >
>> > .../drivers/dri/r300/compiler/r500_fragprog_emit.c |    2 ++
>> > 1 files changed, 2 insertions(+), 0 deletions(-)
>> >
>> > diff --git a/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c b/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
>> > index 829f028..710cae7 100644
>> > --- a/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
>> > +++ b/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
>> > @@ -469,6 +469,8 @@ void r500BuildFragmentProgramHwCode(struct r300_fragment_program_compiler *compi
>> > if (compiler->Base.Error)
>> > return;
>> >
>> > + assert(code->inst_end >= 0);
>> > +
>> > if ((code->inst[code->inst_end].inst0 & R500_INST_TYPE_MASK) != R500_INST_TYPE_OUT) {
>> > /* This may happen when dead-code elimination is disabled or
>> > * when most of the fragment program logic is leading to a KIL */
>>
>> Sorry, is this actually a problem? If this assertion is actually being
>> hit, it sure would be nice to hear about it since it. Empty shaders
>> shouldn't just be handled with debugging code.
>
> Vinson has been cleaning up code base based on the results of coverity
> static code analysis.
>
> Coverity generates thousands of errors/warnings for Mesa source code.
> Not all errors/warnings reported by coverity are actually hit in
> practice, but putting asserts and/or error handling code will at least
> inform coverity that that can't ever happen, and allow the real errors
> to stand out.
>
> I imagine this is the case here.
>
> Jose
>
>

Alright, thanks for the heads-up. I was not sure if this was static
analysis or not. I know that some of the r300 code is not so hot WRT
signed vs. unsigned comps and related things, so these kinds of
commits tend to get extra scrutiny from Nicolai and I.

Thanks for the explanation and hard work!

~ C.

-- 
Only fools are easily impressed by what is only
barely beyond their reach. ~ Unknown

Corbin Simpson
<Mos...@gm...>

[Mesa3d-dev] [RFC] Cygwin: Enable mesa to be built for Cygwin -with-driver=dri -with-dri-driver=swrast

[Jon T. <jon...@dr...>]

Why do I think I want to be able to do this?

Currently, libGL is built for cygwin using --with-driver=xlib

I've written a GLX provider for the Xwin Xserver which uses the native WGL 
interface, so now indirect GLX can be accelerated.

So now I need a libGL which can make use of this, but the realglx code in the x11 
driver which implemented the indirect path has been disabled from building for 
a while (see commit f440b0d and [1])

I've no idea when it last worked, and even if I could get it working again, I 
don't really want to be using code I am the only user of :-)

I've done various build tests on linux, which seem to work correctly, but, 
even if this is the correct approach, this could probably use a review to ensure 
I haven't broken anything.

[1] http://www.mail-archive.com/mes...@li.../msg05067.html

Re: [Mesa3d-dev] [PATCH] Fix --enable-32-bit: honor -m32 for progs/glsl.

[Török E. <edw...@gm...>]

On 2010-02-26 16:50, Dan Nicholson wrote:
>> gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
>> -lGL  -lm -m32 -o array
> 
> These programs aren't in progs/glsl. Are you sure this is solving your
> problem? Can you just show me the make output for the failed run?

They are, I use git master, on this commit:
commit e5c691f445e1c02e6e2f75b817b13d7024f7a3a6
Author: Vinson Lee <vl...@vm...>
Date:   Fri Feb 26 00:17:03 2010 -0800

    r300/compiler: Assert that array index is not negative.

ls -l progs/glsl/shaderutil.c progs/glsl/array.c
-rw-r--r-- 1 edwin edwin 4838 Feb 23 22:26 progs/glsl/array.c

make[1]: Entering directory `/home/edwin/mesa/progs/glsl'
cp ../../progs/util/extfuncs.h .
cp ../../progs/util/shaderutil.h .
cp ../../progs/util/readtex.h .
gcc -I../../include -g -O2 -Wall -Wmissing-prototypes -std=c99
-ffast-math -fvisibility=hidden -fno-strict-aliasing -m32  -fPIC -m32
-DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -D_GNU_SOURCE
-DPTHREADS -DHAVE_POSIX_MEMALIGN -DGLX_USE_TLS -DPTHREADS
-DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING
-DGLX_INDIRECT_RENDERING -DHAVE_ALIAS   -c -o array.o array.c
cp ../../progs/util/shaderutil.c .
gcc -I../../include -g -O2 -Wall -Wmissing-prototypes -std=c99
-ffast-math -fvisibility=hidden -fno-strict-aliasing -m32  -fPIC -m32
-DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -D_GNU_SOURCE
-DPTHREADS -DHAVE_POSIX_MEMALIGN -DGLX_USE_TLS -DPTHREADS
-DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING
-DGLX_INDIRECT_RENDERING -DHAVE_ALIAS   -c -o shaderutil.o shaderutil.c
cp ../../progs/util/readtex.c .
gcc -I../../include -g -O2 -Wall -Wmissing-prototypes -std=c99
-ffast-math -fvisibility=hidden -fno-strict-aliasing -m32  -fPIC -m32
-DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -D_GNU_SOURCE
-DPTHREADS -DHAVE_POSIX_MEMALIGN -DGLX_USE_TLS -DPTHREADS
-DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING
-DGLX_INDIRECT_RENDERING -DHAVE_ALIAS   -c -o readtex.o readtex.c
gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
-lGL  -lm -o array
/usr/bin/ld: skipping incompatible ../../lib32/libglut.so when searching
for -lglut
/usr/bin/ld: cannot find -lglut
collect2: ld returned 1 exit status
make[1]: *** [array] Error 1
make[1]: Leaving directory `/home/edwin/mesa/progs/glsl'

Re: [Mesa3d-dev] [PATCH] Fix --enable-32-bit: honor -m32 for progs/glsl.

[Dan N. <dbn...@gm...>]

2010/2/26 Török Edwin <edw...@gm...>:
> On 02/26/2010 04:19 PM, Dan Nicholson wrote:
>> 2010/2/25 Török Edvin <edw...@gm...>:
>>> ---
>>>  progs/glsl/Makefile |    2 +-
>>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>>
>>> diff --git a/progs/glsl/Makefile b/progs/glsl/Makefile
>>> index 8928c83..aaa54cb 100644
>>> --- a/progs/glsl/Makefile
>>> +++ b/progs/glsl/Makefile
>>> @@ -16,7 +16,7 @@ LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLEW_LIB) -l$(GLU_LIB) \
>>>  # using : to avoid APP_CC pointing to CC loop
>>>  CC := $(APP_CC)
>>>  CFLAGS := -I$(INCDIR) $(CFLAGS)
>>> -LDLIBS = $(LIBS)
>>> +LDLIBS = $(LIBS) $(ARCH_FLAGS)
>>
>> Can you show the error you're getting? This should already be handled
>> in CFLAGS, but I suspect using make's implicit rules is causing this
>> problem.
>
> CFLAGS is not used when linking .o files only, so the command-line
> doesn't have -m32 at all:

Yeah, the built-in make pattern rule for linking .o files doesn't pass
CFLAGS. We usually handle this by not using make's built-in rules for
linking.

> gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
> -lGL  -lm -o array
> /usr/bin/ld: skipping incompatible ../../lib32/libglut.so when searching
> for -lglut
> /usr/bin/ld: cannot find -lglut
> collect2: ld returned 1 exit status
>
> And with that patch applied it correctly adds -m32 to the commandline
> and links:
> gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
> -lGL  -lm -m32 -o array

These programs aren't in progs/glsl. Are you sure this is solving your
problem? Can you just show me the make output for the failed run?

--
Dan

Re: [Mesa3d-dev] [PATCH] Fix --enable-32-bit: honor -m32 for progs/glsl.

[Török E. <edw...@gm...>]

On 02/26/2010 04:19 PM, Dan Nicholson wrote:
> 2010/2/25 Török Edvin <edw...@gm...>:
>> ---
>>  progs/glsl/Makefile |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/progs/glsl/Makefile b/progs/glsl/Makefile
>> index 8928c83..aaa54cb 100644
>> --- a/progs/glsl/Makefile
>> +++ b/progs/glsl/Makefile
>> @@ -16,7 +16,7 @@ LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLEW_LIB) -l$(GLU_LIB) \
>>  # using : to avoid APP_CC pointing to CC loop
>>  CC := $(APP_CC)
>>  CFLAGS := -I$(INCDIR) $(CFLAGS)
>> -LDLIBS = $(LIBS)
>> +LDLIBS = $(LIBS) $(ARCH_FLAGS)
> 
> Can you show the error you're getting? This should already be handled
> in CFLAGS, but I suspect using make's implicit rules is causing this
> problem.

CFLAGS is not used when linking .o files only, so the command-line
doesn't have -m32 at all:

gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
-lGL  -lm -o array
/usr/bin/ld: skipping incompatible ../../lib32/libglut.so when searching
for -lglut
/usr/bin/ld: cannot find -lglut
collect2: ld returned 1 exit status

And with that patch applied it correctly adds -m32 to the commandline
and links:
gcc   array.o shaderutil.o readtex.o  -L../../lib32 -lglut -lGLEW -lGLU
-lGL  -lm -m32 -o array

Best regards,
--Edwin