Commit 53c89c67f33639afef951e178f93f4e29acc5d53 ("i965: Avoid generating
MOVs for assignments of expressions.") added the line "this->result =
reg_undef" all over the code. Unfortunately, since Eric developed his
patch before I landed Ivybridge support, he missed adding it to
fs_visitor::emit_text
On Thu, 21 Jul 2011 12:16:56 -0700, "Ian Romanick" wrote:
> From: Ian Romanick
>
> Perviously the code would just look at deref->array->type to see if it
> was a constant. This isn't good enough because deref->array might be
> another ir_dereference_array... of a constant. As a result,
> deref
On Thu, 21 Jul 2011 12:16:58 -0700, "Ian Romanick" wrote:
> From: Ian Romanick
>
> Just like the non-constant array index lowering pass, compare all N
> indices at once. For accesses to a vec4, this saves 3 comparison
> instructions on a vector architecture.
> ---
> src/glsl/lower_vec_index_to
I have already installed LLVM but still i get the same error. But when i
pass --with-gallium-drivers="" it started working :)
Thanks
Manish
On Thu, Jul 21, 2011 at 10:38 PM, Kenneth Graunke wrote:
> On 07/21/2011 08:00 AM, manish sharma wrote:
> > Hi,
> >
> > I am trying to compile wayland
> > (
The blend_quad function clobbers the actual render target color/alpha
values while applying the destination blend factor, which results in
restoring the wrong value during the masking stage for write-disabled
channels.
---
src/gallium/drivers/softpipe/sp_quad_blend.c | 185 +--
TGSI's set of signed and unsigned integer opcodes is mostly complete and
usable, but there is one important operation missing: loading an integer
into the address register. Currently, the only way to do this is to use
I2F followed by ARL. This is very counterproductive, since every driver
has to
https://bugs.freedesktop.org/show_bug.cgi?id=39286
Sven Arvidsson changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|
On 07/21/2011 12:16 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> Leaving the unused registers with other values caused assertion
> failures and other problems in places that blindly iterate over all
> sources.
>
> brw_vs_emit.c:1381: get_src_reg: Assertion `c->regs[file][index].nr !=
> 0' fa
On 07/18/2011 02:15 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> There's no reason for it to be there, and another class that may not
> have access to the visitor will need it soon.
> ---
> src/glsl/lower_variable_index_to_cond_assign.cpp | 11 ++-
> 1 files changed, 6 insertions(+
From: Ian Romanick
Just like the non-constant array index lowering pass, compare all N
indices at once. For accesses to a vec4, this saves 3 comparison
instructions on a vector architecture.
---
src/glsl/lower_vec_index_to_cond_assign.cpp | 57 --
1 files changed, 35 i
From: Ian Romanick
---
src/glsl/ir_optimization.h |4 +
src/glsl/lower_variable_index_to_cond_assign.cpp | 111 +
2 files changed, 73 insertions(+), 42 deletions(-)
diff --git a/src/glsl/ir_optimization.h b/src/glsl/ir_optimization.h
index 59a0407.
From: Ian Romanick
Perviously the code would just look at deref->array->type to see if it
was a constant. This isn't good enough because deref->array might be
another ir_dereference_array... of a constant. As a result,
deref->array->type wouldn't be a constant, but
deref->variable_referenced()
From: Ian Romanick
Other code will soon need to know if an array needs lowering based
exclusively on the storage mode.
---
src/glsl/lower_variable_index_to_cond_assign.cpp | 15 ++-
1 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/src/glsl/lower_variable_index_to_con
From: Ian Romanick
Fixes i965 piglit vs-varying-array-mat[234]-row-rd.
---
src/mesa/drivers/dri/i965/brw_vs_emit.c | 17 +++--
1 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_vs_emit.c
b/src/mesa/drivers/dri/i965/brw_vs_emit.c
index 5e
From: Ian Romanick
Fixes i965 piglit:
vs-varying-array-mat[234]-col-row-wr
vs-varying-array-mat[234]-index-col-row-wr
vs-varying-array-mat[234]-index-row-wr
vs-varying-array-mat[234]-row-wr
vs-varying-mat[234]-col-row-wr
vs-varying-mat[234]-row-wr
Reviewed-by: Eric Anhol
From: Ian Romanick
Leaving the unused registers with other values caused assertion
failures and other problems in places that blindly iterate over all
sources.
brw_vs_emit.c:1381: get_src_reg: Assertion `c->regs[file][index].nr !=
0' failed.
Fixes i965 piglit:
vs-uniform-array-mat[234]-col
From: Ian Romanick
Fixes i965 piglit:
vs-temp-array-mat[234]-col-row-wr
vs-temp-array-mat[234]-index-col-row-wr
vs-temp-array-mat[234]-index-row-wr
vs-temp-mat[234]-col-row-wr
Fixes swrast piglit:
fs-temp-array-mat[234]-col-row-wr
fs-temp-array-mat[234]-index-col-row-wr
From: Ian Romanick
If the non-constant index was in the LHS of an assignment, any
existing condititon on that assignment would be lost.
---
src/glsl/lower_vec_index_to_cond_assign.cpp | 29 ++
1 files changed, 24 insertions(+), 5 deletions(-)
diff --git a/src/glsl/lowe
From: Ian Romanick
This fixes many cases of accessing arrays of matrices using
non-constant indices at each level.
Fixes i965 piglit:
vs-temp-array-mat[234]-index-col-rd
vs-temp-array-mat[234]-index-col-row-rd
vs-temp-array-mat[234]-index-col-wr
vs-uniform-array-mat[234]-index-c
From: Ian Romanick
If the non-constant index was in the LHS of an assignment, any
existing condititon on that assignment would be lost.
Fixes i965 piglit:
fs-temp-array-mat[234]-col-row-wr
fs-temp-array-mat[234]-index-col-row-wr
fs-temp-array-mat[234]-index-col-wr
fs-temp-array-
From: Ian Romanick
The previous implementation could easily get tricked if the LHS of an
assignment included a non-constant index that was "inside" another
dereference. For example:
mat4 m[2];
m[0][i] = vec4(0.0);
Due to the way it tracked whether the array was being assigned, it
would
From: Ian Romanick
There's no reason for it to be there, and another class that may not
have access to the visitor will need it soon.
---
src/glsl/lower_variable_index_to_cond_assign.cpp | 11 ++-
1 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/src/glsl/lower_variable_in
As can be seen from the giant pile of tests recently sent to the
piglit mailing list for review, the
lower_variable_index_to_cond_assign pass had some issues. With this
updated patch series, *all* of the variable indexing tests posted to the
piglit list pass with (classic) swrast and i965.
Patche
On 07/21/2011 08:48 AM, Brian Paul wrote:
> On 07/20/2011 04:53 PM, Bryan Cain wrote:
>> ---
>> src/mesa/state_tracker/st_extensions.c | 12 +++-
>> 1 files changed, 11 insertions(+), 1 deletions(-)
>>
>> diff --git a/src/mesa/state_tracker/st_extensions.c
>> b/src/mesa/state_tracker/s
https://bugs.freedesktop.org/show_bug.cgi?id=23525
--- Comment #6 from Brian Paul 2011-07-21 10:55:28 PDT
---
The patch was committed to master as commit
cfec000e7514342fd51859906e173ba2d474a55c
How confident are you in the patch? Should it go into the 7.11 release?
--
Configure bugmail: htt
https://bugs.freedesktop.org/show_bug.cgi?id=39420
--- Comment #1 from Younes Manton 2011-07-21 10:18:06 PDT
---
The original bug is actually in softpipe of all places. HW drivers seem to
agree with what was going on originally.
--
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?
On 07/21/2011 08:00 AM, manish sharma wrote:
> Hi,
>
> I am trying to compile wayland
> (http://wayland.freedesktop.org/building.html) and as part of that i
> need to compile mesa with below options.
>
> ./autogen.sh --prefix=$WLD --enable-gles2 --disable-gallium-egl
> --with-egl-platforms=x11,w
You need a devel package of LLVM. On Debian-based distros, it's
usually called llvm-dev.
Marek
On Thu, Jul 21, 2011 at 5:00 PM, manish sharma <83.man...@gmail.com> wrote:
> Hi,
>
> I am trying to compile wayland
> (http://wayland.freedesktop.org/building.html) and as part of that i need to
> comp
> I've been hacking around trying to make this test work and discovered
> writing some bits up at bit 16 of the exported word seems to act like
> some sort of mask. However I've no idea what this is or why this fixes
> this test. However it breaks the stencilreaddraw test in mesa-demos, which
> mea
From: Dave Airlie
I've been hacking around trying to make this test work and discovered
writing some bits up at bit 16 of the exported word seems to act like
some sort of mask. However I've no idea what this is or why this fixes
this test. However it breaks the stencilreaddraw test in mesa-demos,
https://bugs.freedesktop.org/show_bug.cgi?id=23525
--- Comment #5 from Marc Pignat 2011-07-21 08:12:08 PDT ---
Created an attachment (id=49386)
View: https://bugs.freedesktop.org/attachment.cgi?id=49386
Review: https://bugs.freedesktop.org/review?bug=23525&attachment=49386
fix 24 bits colors /
Hi,
I am trying to compile wayland (http://wayland.freedesktop.org/building.html)
and as part of that i need to compile mesa with below options.
./autogen.sh --prefix=$WLD --enable-gles2 --disable-gallium-egl
--with-egl-platforms=x11,wayland,drm --enable-gbm
--enable-shared-glapi
Which takes me
On 07/20/2011 04:53 PM, Bryan Cain wrote:
The purpose of the following two patches is to make st/mesa expose the S3TC
extensions when the force_s3tc_enable environment variable is used. This is
to match the behavior of the DRI drivers, in which force_s3tc_enable is an
option in driconf. Althoug
On 07/20/2011 04:53 PM, Bryan Cain wrote:
---
src/mesa/state_tracker/st_extensions.c | 12 +++-
1 files changed, 11 insertions(+), 1 deletions(-)
diff --git a/src/mesa/state_tracker/st_extensions.c
b/src/mesa/state_tracker/st_extensions.c
index 99b231d..073e72c 100644
--- a/src/mes
On Thu, Jul 21, 2011 at 9:48 AM, Chia-I Wu wrote:
> On Wed, Jul 20, 2011 at 11:32 PM, Marek Olšák wrote:
>> Well spotted, thanks. Some comments below.
>>
>> On Wed, Jul 20, 2011 at 3:25 PM, Chia-I Wu wrote:
>>> From: Chia-I Wu
>>>
>>> u_vbuf_upload_buffers modifies the buffer offsets. If they
https://bugs.freedesktop.org/show_bug.cgi?id=38869
--- Comment #4 from Jonathan Kirkham 2011-07-21
05:00:18 PDT ---
(In reply to comment #3)
> (In reply to comment #2)
> > (In reply to comment #0)
> > > Both the EGL and GLX specifications allow pbuffers to be created with a
> > > height
> > > a
https://bugs.freedesktop.org/show_bug.cgi?id=39420
Summary: g3dvl: Preserve previously rendered components for MC
output. Regresses R600 XVMC
Product: Mesa
Version: git
Platform: x86 (IA32)
OS/Version: Linux (All)
S
On Wed, Jul 20, 2011 at 11:32 PM, Marek Olšák wrote:
> Well spotted, thanks. Some comments below.
>
> On Wed, Jul 20, 2011 at 3:25 PM, Chia-I Wu wrote:
>> From: Chia-I Wu
>>
>> u_vbuf_upload_buffers modifies the buffer offsets. If they are not
>> restored, and any of the vertex formats is not s
38 matches
Mail list logo