Mesa 11.0.4 is now available.
With this release we have a wide selection of core mesa, glsl and driver
fixes, a few build related patches and an omx bugfix.
Alejandro Piñeiro (2):
i965/vec4: check writemask when bailing out at register coalesce
i965/vec4: fill src_reg type using the
On Sat, Oct 24, 2015 at 7:08 PM, Rob Clark wrote:
> From: Rob Clark
>
> +#define NIR_PASS_PROGRESS(pass, nir, ...) ({ \
> + assert(nir_shader_is_mutable(nir)); \
> + bool __ret = pass(nir, ##__VA_ARGS__);
Reviewed-by: Marek Olšák
Marek
On Sun, Oct 25, 2015 at 4:27 AM, Kenneth Graunke wrote:
> When a TCS is present, the TES input gl_PatchVerticesIn is actually a
> constant - it's simply the # of output vertices specified by the TCS
> layout qualifiers.
On 10/25/15, Ivan Kalvachev wrote:
> Some constants (like 1.0 and 0.5) could be inlined as immediate inputs
> without using their literal value. The r600_bytecode_special_constants()
> function emulates the negative of these constants by using NEG modifier.
>
> However some
On Sun, Oct 25, 2015 at 8:05 AM, Erik Faye-Lund wrote:
> On Sat, Oct 24, 2015 at 7:08 PM, Rob Clark wrote:
>> From: Rob Clark
>>
>> +#define NIR_PASS_PROGRESS(pass, nir, ...) ({ \
>> +
On 10/22/2015 01:16 AM, Julien Isorce wrote:
The real fix is in nouveau_drm_winsys.c by setting dev to 0.
Which means dev's ownership has been passed to previous call.
Other changes are there to be consistent with what the
screen_create functions already do on errors.
This actually happens
On 10/22/2015 01:02 AM, Julien Isorce wrote:
Sorry this patch should not have gone in the v2 since it has been
already reviewed by Emil. But thx for your review.
I experienced the crash when testing patch 5/7 of this patch series,
around "resource = pscreen->resource_from_handle" in the new
From: Marek Olšák
I wonder if the craziness was worth it.
---
src/mesa/Makefile.sources| 2 +
src/mesa/state_tracker/st_cb_copyimage.c | 609 +++
src/mesa/state_tracker/st_cb_copyimage.h | 33 ++
From: Marek Olšák
For ARB_copy_image.
---
src/gallium/docs/source/screen.rst | 4 +++-
src/gallium/drivers/freedreno/freedreno_screen.c | 1 +
src/gallium/drivers/i915/i915_screen.c | 1 +
src/gallium/drivers/ilo/ilo_screen.c | 1 +
From: Marek Olšák
which is where a block in src maps to a pixel in dst and vice versa.
e.g. DXT1 <-> R32G32_UINT
DXT5 <-> R32G32B32A32_UINT
---
src/gallium/drivers/radeonsi/si_blit.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
On Fri, Oct 23, 2015 at 7:10 PM, Kenneth Graunke wrote:
> diff --git a/src/mesa/drivers/dri/i965/brw_vec4_generator.cpp
> b/src/mesa/drivers/dri/i965/brw_vec4_generator.cpp
> index 96a52c6..6569d81 100644
> --- a/src/mesa/drivers/dri/i965/brw_vec4_generator.cpp
> +++
2015-10-25 13:58 GMT+01:00 Rob Clark :
> On Sun, Oct 25, 2015 at 8:05 AM, Erik Faye-Lund wrote:
>> On Sat, Oct 24, 2015 at 7:08 PM, Rob Clark wrote:
>>> From: Rob Clark
>>>
>>> +#define
On Sat, Oct 24, 2015 at 01:20:22PM -0700, Matt Turner wrote:
> ---
> src/mesa/drivers/dri/i965/brw_vec4_generator.cpp | 3 ---
> 1 file changed, 3 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_vec4_generator.cpp
> b/src/mesa/drivers/dri/i965/brw_vec4_generator.cpp
> index
v4:
Rebase against lastest master and fix bad variable name in assert.
A new test was developed to properly check commit 4 behavior.
I ran most of the piglit test without regression.
v3:
Squash old commit 1&2
* Use a better name for the new attribute: always_active_io
* Use ir_variable directly
GL_ARB_separate_shader_objects allow to match by name variable or block
interface. Input varying can't be removed because it is will impact the
location assignment.
It fixes the bug 79783 and likely any application that uses
GL_ARB_separate_shader_objects extension.
piglit test:
The value will be set in separate-shader program when an input/output
must remains active (i.e. deadcode removal isn't allowed because it will create
interface location/name-matching mismatch)
v3:
* Rename the attribute
* Use ir_variable directly instead of ir_variable_refcount_visitor
* Move the
Current behavior on the interface matching:
layout (location = 0) out0; // Assigned to VARYING_SLOT_VAR0 by user
out1; // Assigned to VARYING_SLOT_VAR0 by the linker
New behavior on the interface matching:
layout (location = 0) out0; // Assigned to VARYING_SLOT_VAR0 by user
out1; // Assigned to
Current issue is the addition of FLAT qualifier on varying_matches::record()
which break the varying expected order
Future issue is the removal of the interpolation qualifier matching constrain
In my humble opinion, it is the responsability of the GL developer to optimize
their slots assignment
On Sun, Oct 25, 2015 at 7:27 AM, Ivan Kalvachev
wrote:
> On 10/25/15, Ivan Kalvachev wrote:
> > Some constants (like 1.0 and 0.5) could be inlined as immediate inputs
> > without using their literal value. The r600_bytecode_special_constants()
> >
On Sun, Oct 25, 2015 at 9:09 AM, Ben Widawsky wrote:
> On Sat, Oct 24, 2015 at 01:20:22PM -0700, Matt Turner wrote:
>> ---
>> src/mesa/drivers/dri/i965/brw_vec4_generator.cpp | 3 ---
>> 1 file changed, 3 deletions(-)
>>
>> diff --git
20 matches
Mail list logo