On Fri, 2011-12-23 at 20:57 -0800, Kenneth Graunke wrote:
> In this code, 'i' loops over the number of virtual GRFs, while 'j' loops
> over the number of vector components (0 <= j <= 3).
>
> It can't possibly be correct to see if bit 'i' is set in the destination
> writemask, as it will have value
On Thu, 22 Dec 2011 17:43:57 -0800, "Ian Romanick" wrote:
> From: Ian Romanick
>
> There are a couple scenarios where the source could be zero and the
> operand could be either SRC_ALPHA or ONE_MINUS_SRC_ALPHA. For
> example, if the source was ZERO. This would result in something like
> (0).w,
On Fri, 23 Dec 2011 13:11:58 +0800, Yuanhan Liu
wrote:
> On Thu, Dec 22, 2011 at 07:51:46PM -0800, Kenneth Graunke wrote:
> > On 12/22/2011 07:04 PM, Yuanhan Liu wrote:
> > > On Thu, Dec 22, 2011 at 02:33:03PM -0800, Kenneth Graunke wrote:
> > >> On 12/21/2011 01:33 AM, Yuanhan Liu wrote:
> > [sn
On Fri, 23 Dec 2011 11:13:09 +0800, Yuanhan Liu
wrote:
> Here is the fixed patch:
>
> From 7c8b8bc87846df9513a0c32cc8a388fb62f5476a Mon Sep 17 00:00:00 2001
> From: Yuanhan Liu
> Date: Wed, 21 Dec 2011 15:32:02 +0800
> Subject: [PATCH] i965: call next_insn() before referencing a instruction by
On Thu, 22 Dec 2011 19:22:16 -0800, Paul Berry wrote:
> On 22 December 2011 16:54, Eric Anholt wrote:
>
> > Fixes almost all of the transform feedback piglit tests. Remaining
> > are a few tests related to tesselation for
> > quads/trifans/tristrips/polygons with flat shading.
> > ---
> > src/
On Fri, Dec 23, 2011 at 11:57 PM, Kenneth Graunke wrote:
> In this code, 'i' loops over the number of virtual GRFs, while 'j' loops
> over the number of vector components (0 <= j <= 3).
>
> It can't possibly be correct to see if bit 'i' is set in the destination
> writemask, as it will have values
https://bugs.freedesktop.org/show_bug.cgi?id=44101
Dan Callaghan changed:
What|Removed |Added
CC||d...@djc.id.au
--
Configure bugmail: ht
Having the direct copy case short-circuit the loop doesn't make sense.
In both cases, we need to update the destination's current value, as
well as the current value of any register that pointed to our
destination. The only difference is whether or not we know our
destination's current value.
By
Consider the following code:
MOV A.x, B.x
MOV B.x, C.x
After the first line, cur_value[A][0] == B, indicating that A.x's
current value came from register B.
When processing the second line, we update cur_value[B][0] to C.
However, for drect copies, we fail to reset cur_value[A][0] to NULL.
This
In this code, 'i' loops over the number of virtual GRFs, while 'j' loops
over the number of vector components (0 <= j <= 3).
It can't possibly be correct to see if bit 'i' is set in the destination
writemask, as it will have values much larger than 3. Clearly this is
supposed to be 'j'.
Found by
https://bugs.freedesktop.org/show_bug.cgi?id=44101
Alexandre Demers changed:
What|Removed |Added
Summary|[bisected] |[PATCH]
|egl_g3d_ap
https://bugs.freedesktop.org/show_bug.cgi?id=44101
--- Comment #2 from Alexandre Demers 2011-12-23
20:55:25 PST ---
Created attachment 54776
--> https://bugs.freedesktop.org/attachment.cgi?id=54776
Fixes stapi->createContext usage
Fixed the bug, fixed a warning where attributs and error argum
* Several platforms try to build static libGLU.a
however, the libGLU makefile only will build a shared .so
* Haiku links libGLU into libGL, so this makes the process easier
---
configs/haiku|3 +++
src/glu/sgi/Makefile | 10 +-
2 files changed, 12 insertions(+), 1 deletio
---
Makefile |1 +
acinclude.m4 |2 +-
configs/haiku | 66
+
3 files changed, 68 insertions(+), 1 deletions(-)
create mode 100644 configs/haiku
diff --git a/Makefile b/Makefile
index cf6555c..4caa8ce 100644
--- a/Makefil
---
bin/mklib | 37 +
1 files changed, 37 insertions(+), 0 deletions(-)
diff --git a/bin/mklib b/bin/mklib
index 70bd1a2..93dd989 100755
--- a/bin/mklib
+++ b/bin/mklib
@@ -959,6 +959,43 @@ case $ARCH in
fi
;;
+ 'Haiku')
+
* Use generic pthread barrier on Haiku.
* Re-add debug_printf define for Haiku.
* Add mapi headers to r300 build (to not rely on makedepend?)
* Remove NDEBUG define of union, as the union is used
regardless of NDEFINE setting.
* Remove BeOS reference to not building GLU as BeOS is no
longer a
https://bugs.freedesktop.org/show_bug.cgi?id=44113
Bug #: 44113
Summary: git mesa gallium/state_trackers/egl fails to compile
Classification: Unclassified
Product: Mesa
Version: git
Platform: x86-64 (AMD64)
OS/Version: Linux (All)
On Fri, Dec 23, 2011 at 8:37 PM, Kenneth Graunke wrote:
> On 12/23/2011 04:21 PM, Matt Turner wrote:
>> # Flex and Bison for GLSL compiler
>> -FLEX = @FLEX@
>> -BISON = @BISON@
>> +FLEX = @LEX@
>> +BISON = @YACC@
>
> Question about this: will this still require Flex and Bison, or accept
> any Lex
On 12/23/2011 05:38 PM, Gaetan Nadon wrote:
>
> On 11-12-23 08:09 PM, Matt Turner wrote:
>> On Fri, Dec 23, 2011 at 8:06 PM, Kenneth Graunke
> wrote:
>>> On 12/23/2011 04:21 PM, Matt Turner wrote:
---
configure.ac | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
di
On 12/23/2011 05:31 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> Other parts of the compiler assume that expressions will have
> well-formed types or the error type. Just using the type of the thing
> being operated on can cause expressions like ~3.14 or ~false to not
> have a well-formed ty
On 12/23/2011 04:21 PM, Matt Turner wrote:
> The glsl_{lexer.lpp,parser.ypp} symlinks allow automake to figure out
> dependency information, since .ll generates a .cc file, and .lpp
> generates a .cpp. They were originally named .lpp/.ypp, but were changed
> to work around scons.
> ---
>
> This is
From: Ian Romanick
Other parts of the compiler assume that expressions will have
well-formed types or the error type. Just using the type of the thing
being operated on can cause expressions like ~3.14 or ~false to not
have a well-formed type. This could then result in an assertion
failure in t
On 12/23/2011 04:21 PM, Matt Turner wrote:
> See 9e7a4147.
> ---
> configure.ac | 77
> --
> 1 files changed, 0 insertions(+), 77 deletions(-)
Patches 1-4 are
Reviewed-by: Kenneth Graunke
I went ahead and pushed 4 (remove unused autogen
On Fri, Dec 23, 2011 at 8:06 PM, Kenneth Graunke wrote:
> On 12/23/2011 04:21 PM, Matt Turner wrote:
>> ---
>> configure.ac | 2 +-
>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/configure.ac b/configure.ac
>> index c0d6882..0d75353 100644
>> --- a/configure.ac
>> +++ b
On 12/23/2011 04:21 PM, Matt Turner wrote:
> ---
> configure.ac |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/configure.ac b/configure.ac
> index c0d6882..0d75353 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -1,6 +1,6 @@
> dnl Process this file with auto
The glsl_{lexer.lpp,parser.ypp} symlinks allow automake to figure out
dependency information, since .ll generates a .cc file, and .lpp
generates a .cpp. They were originally named .lpp/.ypp, but were changed
to work around scons.
---
This isn't free from breakage, but it is good enough to build
---
src/glsl/autogen.sh | 12
1 files changed, 0 insertions(+), 12 deletions(-)
delete mode 100755 src/glsl/autogen.sh
diff --git a/src/glsl/autogen.sh b/src/glsl/autogen.sh
deleted file mode 100755
index 904cd67..000
--- a/src/glsl/autogen.sh
+++ /dev/null
@@ -1,12 +0,0 @@
-#
Signed-off-by: Matt Turner
---
src/glsl/glsl_lexer.ll |2 +-
src/glsl/glsl_parser.yy |4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/glsl/glsl_lexer.ll b/src/glsl/glsl_lexer.ll
index c7cfedd..936a907 100644
--- a/src/glsl/glsl_lexer.ll
+++ b/src/glsl/glsl_lexe
---
configure.ac |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/configure.ac b/configure.ac
index c0d6882..0d75353 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1,6 +1,6 @@
dnl Process this file with autoconf to create configure.
-AC_PREREQ([2.59])
+AC_PREREQ([2.60
See 9e7a4147.
---
configure.ac | 77 --
1 files changed, 0 insertions(+), 77 deletions(-)
diff --git a/configure.ac b/configure.ac
index a4943e1..c0d6882 100644
--- a/configure.ac
+++ b/configure.ac
@@ -709,83 +709,6 @@ if test "x$enable_s
https://bugs.freedesktop.org/show_bug.cgi?id=44101
Ian Romanick changed:
What|Removed |Added
CC||v...@vmware.com
--- Comment #1 from Ian R
https://bugs.freedesktop.org/show_bug.cgi?id=44105
Ian Romanick changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|
https://bugs.freedesktop.org/show_bug.cgi?id=44105
Bug #: 44105
Summary: egl_g3d_api.c:176: error: too few arguments to
function ‘gctx->stapi->create_context’
Classification: Unclassified
Product: Mesa
Version: git
Pla
When updating SOL indices, we were accidentally putting the starting
index in dword 1 and the SVBI number to increment in dword 2--these
should be reversed. Usually both of these values are zero, so we
didn't see any problem. However, if a transform feedback operation
spans multiple batch buffers
From: Ian Romanick
The various l-value errors this was designed to catch are now caught
by other means. Marking the temporaries as read-only now just
prevents sensible error messages from being generated. It's
0:0(0): error: function parameter 'out p' references the read-only variable
'_post_
From: Ian Romanick
Somethings, like pre-increment operations, were not previously caught.
After the 8.0 release, this code needs some major refactoring and
clean-up. It's a mess. :(
Signed-off-by: Ian Romanick
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42755
---
src/glsl/ast_funct
From: Ian Romanick
Signed-off-by: Ian Romanick
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42755
---
src/glsl/ast_to_hir.cpp | 28 +---
1 files changed, 21 insertions(+), 7 deletions(-)
diff --git a/src/glsl/ast_to_hir.cpp b/src/glsl/ast_to_hir.cpp
index 93
From: Ian Romanick
Signed-off-by: Ian Romanick
---
src/glsl/ast.h | 13 +
src/glsl/ast_to_hir.cpp |5 +
src/glsl/glsl_parser_extras.cpp |1 +
3 files changed, 19 insertions(+), 0 deletions(-)
diff --git a/src/glsl/ast.h b/src/glsl/ast.h
index
On 22 December 2011 19:37, Paul Berry wrote:
> On 22 December 2011 16:54, Eric Anholt wrote:
>
> Also, it looks like none of the current piglit tests test transform
>> feedback across batchbuffers. I don't expect major problems there,
>> given that we have a userland count of verts emitted.
>>
On 12/23/2011 12:19 PM, Matt Turner wrote:
The (swrast, i965, gallium, r600g) tuples are inconsistent and
confusing. If swrast, i965, and gallium support something, let's simply
say DONE without qualifying it.
---
I've checked as many of these as I could on my Sandybridge and r600g
systems. I al
On Thu, 22 Dec 2011 09:45:01 +0100, Maarten Lankhorst wrote:
Hey Alexander,
On 12/21/2011 07:16 PM, Alexander von Gluck wrote:
* Doesn't reintroduce legacy drivers
* Adds Haiku mklib code
* Removes some broken PIPE_OS_HAIKU defines
* Removes an NDEBUG ifdef in link_uniforms.cpp,
there is an
When updating SOL indices, we were accidentally putting the starting
index in dword 1 and the SVBI number to increment in dword 2--these
should be reversed. Usually both of these values are zero, so we
didn't see any problem. However, if a transform feedback operation
spans multiple batch buffers
On 23 December 2011 13:02, Kenneth Graunke wrote:
> On 12/23/2011 09:37 AM, Paul Berry wrote:
> > When rendering triangle strips, vertices come down the pipeline in the
> > order specified, even though this causes alternate triangles to have
> > reversed winding order. For example, if the vertic
On 12/23/2011 09:37 AM, Paul Berry wrote:
> When rendering triangle strips, vertices come down the pipeline in the
> order specified, even though this causes alternate triangles to have
> reversed winding order. For example, if the vertices are ABCDE, then
> the GS is invoked on triangles ABC, BCD
The (swrast, i965, gallium, r600g) tuples are inconsistent and
confusing. If swrast, i965, and gallium support something, let's simply
say DONE without qualifying it.
---
I've checked as many of these as I could on my Sandybridge and r600g
systems. I also checked against swrast and llvmpipe.
doc
https://bugs.freedesktop.org/show_bug.cgi?id=44039
Nikita changed:
What|Removed |Added
Component|Other |Mesa core
--
Configure bugmail: https://bugs.f
On Fri, Dec 23, 2011 at 7:51 PM, Chia-I Wu wrote:
> On Sat, Dec 24, 2011 at 12:37 AM, Ian Romanick wrote:
>> On 12/23/2011 07:20 AM, Jose Fonseca wrote:
>>>
>>> - Original Message -
Hi list,
Multiple driver support in EGL is hard to get right, if not
impossible.
>
On 12/23/2011 10:51 AM, Chia-I Wu wrote:
> On Sat, Dec 24, 2011 at 12:37 AM, Ian Romanick wrote:
>> On 12/23/2011 07:20 AM, Jose Fonseca wrote:
>>> Couldn't we restructure this such that OpenVG (and all other state
>>> trackers on Linux) still use a dri2 driver, but one which exposes an
>>> additi
https://bugs.freedesktop.org/show_bug.cgi?id=44101
Bug #: 44101
Summary: [bisected] egl_g3d_api.c:176:10: error: too few
arguments to function ‘gctx->stapi->create_context’
Classification: Unclassified
Product: Mesa
Version: git
Maarten Lankhorst wrote:
Feel free to commit the result with those 3 changes without further review. :)
Mesa master is now failing with everything again for me (with -vc
ffmpeg12vdpau )
vl/vl_vlc.h:228:vl_vlc_eatbits: Assertion `vl_vlc_valid_bits(vlc) <=
num_bits' failed.
__
On Sat, Dec 24, 2011 at 12:37 AM, Ian Romanick wrote:
> On 12/23/2011 07:20 AM, Jose Fonseca wrote:
>>
>> - Original Message -
>>>
>>> Hi list,
>>>
>>> Multiple driver support in EGL is hard to get right, if not
>>> impossible.
>>> On Linux desktop, we almost always want to use egl_dri2.
When rendering triangle strips, vertices come down the pipeline in the
order specified, even though this causes alternate triangles to have
reversed winding order. For example, if the vertices are ABCDE, then
the GS is invoked on triangles ABC, BCD, and CDE, even though this
means that triangle BC
On 12/23/2011 07:20 AM, Jose Fonseca wrote:
- Original Message -
Hi list,
Multiple driver support in EGL is hard to get right, if not
impossible.
On Linux desktop, we almost always want to use egl_dri2. It allows
EGL
to loads DRI2 drivers, thus allowing it to share DRI2 drivers with
li
https://bugs.freedesktop.org/show_bug.cgi?id=30493
--- Comment #2 from Stefano Teso 2011-12-23 07:20:30
PST ---
(In reply to comment #1)
> The read-front piglit test fails with llvmpipe / softpipe on mesa 7.9-devel /
> 7.10-devel.
>
> The test usually fails on swrast and may fail with other dri
- Original Message -
> Hi list,
>
> Multiple driver support in EGL is hard to get right, if not
> impossible.
> On Linux desktop, we almost always want to use egl_dri2. It allows
> EGL
> to loads DRI2 drivers, thus allowing it to share DRI2 drivers with
> libGL.
>
> In one case where t
From: Jian Zhao
Move the format and type check before select_tex_image, or it will fail to
report
the mismatch error if the teximage is null.
Reported-by: Anuj Phogat
Signed-off-by: Jian Zhao
Reviewed-by: Yuanhan Liu
---
src/mesa/main/texgetimage.c | 12 ++--
1 files changed, 6 in
On 12/22/2011 02:25 PM, Christian König wrote:
> Hi Maarten,
>
> On 21.12.2011 21:52, Maarten Lankhorst wrote:
>> It would be nice if you inlined patches for easier reviewing. :)
> Well I can try, but I can't promise that Thunderbird isn't badly fucking up
> all whitespaces, newest version of the
57 matches
Mail list logo