On 11/03/2016 12:18 AM, Timothy Arceri wrote:
This will allow us to move some fields from gl_linked_shader to
gl_program, which will allow us to do some clean-ups like storing
gl_program directly in the CurrentProgram array in gl_pipeline_object
enabling some small validation optimisations at
I am not really that much familiar with nir, so I apologize if what I write
below is wrong.
What if we had something like:
a = 0;
if (x) {
discard_if(y);
a = 1;
}
Then if (x && !y), we must not discard and both at the same time set a to
1. Your transformation would probably change it to:
a
On Wed, Nov 02, 2016 at 01:58:43PM -0700, Jason Ekstrand wrote:
> +/**
> + * This function applies the relocation for a command buffer and writes the
> + * actual addresses into the buffers as per what we were told by the kernel
> on
> + * the previous execbuf2 call. This should be safe to do
On 3 Nov. 2016 18:21, "Gustaw Smolarczyk" wrote:
>
> I am not really that much familiar with nir, so I apologize if what I
write below is wrong.
>
> What if we had something like:
>
> a = 0;
> if (x) {
> discard_if(y);
> a = 1;
> }
>
> Then if (x && !y), we must not
commit cc6aa1d161280f10ded7834d1ec2413bc97589fe changed to using rzalloc
for gl_program creation but one instance for program creation was still
using calloc.
Signed-off-by: Tapani Pälli
---
src/mesa/drivers/dri/i965/brw_program.c | 2 +-
1 file changed, 1 insertion(+),
On Wed, Nov 02, 2016 at 05:15:52PM -0700, Jason Ekstrand wrote:
> @@ -1562,22 +1580,98 @@ VkResult anv_WaitForFences(
> * best we can do is to clamp the timeout to INT64_MAX. This limits the
> * maximum timeout from 584 years to 292 years - likely not a big deal.
> */
> - if
On Thu, 2016-11-03 at 11:40 +0200, Tapani Pälli wrote:
> commit cc6aa1d161280f10ded7834d1ec2413bc97589fe changed to using
> rzalloc
> for gl_program creation but one instance for program creation was
> still
> using calloc.
>
> Signed-off-by: Tapani Pälli
> ---
>
The GL 4.5 spec says:
"If any enabled array’s buffer binding is zero when DrawArrays
or one of the other drawing commands defined in section 10.4 is
called, the result is undefined."
This commits avoids crashing the code, which is not a very good
"undefined result".
This fixes
On Thu, 2016-11-03 at 10:55 +0100, Juan A. Suarez Romero wrote:
> On Thu, 2016-11-03 at 11:40 +0200, Tapani Pälli wrote:
> >
> > commit cc6aa1d161280f10ded7834d1ec2413bc97589fe changed to using
> > rzalloc
> > for gl_program creation but one instance for program creation was
> > still
> > using
From: Nicolai Hähnle
From the manpage of asprintf:
"If memory allocation wasn't possible, or some other error occurs,
these functions will return -1, and the contents of strp are
undefined."
---
src/compiler/glsl/cache.c | 6 +++---
1 file changed, 3
From: Nicolai Hähnle
When splitting up loads, we have to add 16 bytes to the offset for
the high components, just like already happens for stores.
Fixes arb_gpu_shader_fp64@shader_storage@layout-std140-fp64-shader.
Cc: 13.0
---
From: Nicolai Hähnle
The hardware always treats the alpha channel as unsigned, so add a shader
workaround. This is rare enough that we'll just build a monolithic vertex
shader.
The SINT case cannot actually happen in OpenGL, but I've included it for
completeness since
For the series:
Reviewed-by: Nicolai Hähnle
On 02.11.2016 19:07, Marek Olšák wrote:
From: Marek Olšák
I need the definition of PUBLIC.
Cc: 12.0 13.0
---
src/egl/drivers/dri2/egl_dri2.h | 2 --
Reviewed-by: Iago Toral Quiroga
On Thu, 2016-11-03 at 11:03 +0100, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> From the manpage of asprintf:
>
> "If memory allocation wasn't possible, or some other error occurs,
> these functions will
If you incorporate Dieter's suggestion to fix FMASK, the series is:
Reviewed-by: Marek Olšák
Sorry for breaking it so badly. Testing SI on GBM apparently wasn't enough.
Marek
On Wed, Nov 2, 2016 at 5:09 PM, Nicolai Hähnle wrote:
> For the series:
>
>
From: Ben Widawsky
This patch will preserve the BO & offset, and not the miptree for the
aux_mcs buffer. Eventually it might make sense to pull put the sizing
function in miptree creation, but for now this should be sufficient
and not too hideous.
v2: Save BO's
Hi all,
Here is an updated version of a patchset to enable sampling from HiZ
in the i965 driver. The initial series was from Jordan and Ben made a
v2 [1].
Ben's version of the series enabled sampling from HiZ from gen8, but
unfortunately I don't think we can actually do this, mostly because
From: Ben Widawsky
libdrm may round up the allocation requested by mesa. As a result, accesses
through the gtt may end up accessing memory which does not belong to mesa. The
problem is described in the following commit:
commit 7ae870211ddc40ef6ed209a322c3a721214bb737
From: Ben Widawsky
This seems counter to the goal of consolidating hiz, mcs, and later ccs buffers.
Unfortunately, hiz on gen6 is a thing the code supports, and this wart will be
helpful to achieve that. Overall, I believe it does help unify AUX buffers on
gen7+.
I
From: Jordan Justen
This will allow us to treat HiZ and MCS the same when using as an
auxiliary surface buffer.
v2: (Ben) Minor rebase conflict resolution.
Rename mcs_buf to aux_buf to address upcoming change for hiz specific
buffers.
That second thing is
From: Ben Widawsky
The next patch will change the map type, and this will make sure there are no
regressions as a result of the other stuff. Since the miptree is newly created,
I believe it is always safe to just map.
It is possible to CPU map this buffer on LLC
From: Ben Widawsky
The original functionality this patch introduces was authored by a patch from
Ken (the commit subject was the same). Since I ended up changing so many patches
in the code before this one, I had some non-trivial decisions to make, and I
didn't feel
From: Jordan Justen
Currently it indicates that this is never supported, but soon it will
be supported for gen8+^w gen9+
v2 by Ben:
- Explicitly disable aux_hiz for gen < 9 (with comment)
- squashed in next patch to avoid unused and useless functions
i965: Support
From: Jordan Justen
For gen9+ this will indicate when we should allow hiz based sampling
during rendering.
Improves performance in :
- Synmark's OglDeferred by 2.2% (n=20)
- Synmark's OglShMapPcf by 0.44% (n=20)
v2 by Ben: Add spec reference, and make it fix with
Since we started releasing GLSL IR after linking the only time we can
print GLSL IR is during linking. When regenerating variants only NIR
will be available.
---
src/mesa/drivers/dri/i965/brw_cs.c | 6 --
src/mesa/drivers/dri/i965/brw_gs.c | 3 ---
On 03.11.2016 02:15, srol...@vmware.com wrote:
From: Roland Scheidegger
Previous fixes were incomplete - some code still iterated through the number
of elements provided by velem layout instead of the number stored in the key
(which is the same as the number defined by the
This was being initialized just once to 0, but we want to reset it
every time we call run() on the visitor, which can happen more than
once in scenarios where we failed the compilation using a method and
we are falling back to a different configuration. This is possible,
for example, with geometry
https://bugs.freedesktop.org/show_bug.cgi?id=98563
Lorenzo Bona changed:
What|Removed |Added
CC|
Mesa uses limits.h elsewhere, and this makes is possible to
compile anv_allocator.c on Android.
Signed-off-by: Tapani Pälli
---
src/intel/vulkan/anv_allocator.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/intel/vulkan/anv_allocator.c
Reviewed-by: Samuel Iglesias Gonsálvez
On 03/11/16 12:49, Tapani Pälli wrote:
> Mesa uses limits.h elsewhere, and this makes is possible to
> compile anv_allocator.c on Android.
>
> Signed-off-by: Tapani Pälli
> ---
>
On Wed, Nov 2, 2016 at 2:55 PM, Nicolai Hähnle wrote:
> On 29.10.2016 01:13, Marek Olšák wrote:
>>
>> From: Marek Olšák
>>
>> ---
>> src/compiler/glsl/ir_optimization.h | 3 ++-
>> src/compiler/glsl/lower_if_to_cond_assign.cpp | 23
>>
On 03.11.2016 13:02, Marek Olšák wrote:
On Wed, Nov 2, 2016 at 2:55 PM, Nicolai Hähnle wrote:
On 29.10.2016 01:13, Marek Olšák wrote:
From: Marek Olšák
---
src/compiler/glsl/ir_optimization.h | 3 ++-
https://bugs.freedesktop.org/show_bug.cgi?id=94232
Mike Lothian changed:
What|Removed |Added
Resolution|--- |FIXED
https://bugs.freedesktop.org/show_bug.cgi?id=61417
Mike Lothian changed:
What|Removed |Added
Resolution|--- |INVALID
Signed-off-by: Christian Gmeiner
---
src/egl/drivers/dri2/egl_dri2.c | 5 +++--
src/gbm/backends/dri/gbm_dri.c | 5 +++--
src/glx/dri_common.c| 5 ++---
3 files changed, 8 insertions(+), 7 deletions(-)
diff --git a/src/egl/drivers/dri2/egl_dri2.c
Signed-off-by: Christian Gmeiner
---
include/GL/internal/dri_interface.h | 14 ++
1 file changed, 14 insertions(+)
diff --git a/include/GL/internal/dri_interface.h
b/include/GL/internal/dri_interface.h
index d0b1bc6..36ba65e 100644
---
On Fri, Oct 28, 2016 at 02:17:00AM -0700, Jason Ekstrand wrote:
> There are some invariants such as number of samples on which we should
> assert. However, most other things should silently return false since
> they're much easier for isl_surf_get_ccs to check than the caller. We also
> update
This makes it possible to 'use' the imx-drm driver. Remeber that it
is not possible to have sysmbol names in C/C++ with a '-' in it.
Signed-off-by: Christian Gmeiner
---
include/GL/internal/dri_interface.h | 7 +++
1 file changed, 7 insertions(+)
diff --git
From: Nicolai Hähnle
This fixes a build regression of commit
7115e56c21ace07cf04f5073ba73a533e2182099.
Sorry for the breakage, this second location for link dependencies escaped
my build tests.
---
src/gallium/targets/pipe-loader/Makefile.am | 1 +
1 file changed, 1
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #5 from Nicolai Hähnle ---
This works just fine on radeonsi, so it's likely an r600-specific problem that
happened to be uncovered by that commit. Have you tried running with a debug
build of Mesa?
--
You are
On Fri, Oct 28, 2016 at 02:17:03AM -0700, Jason Ekstrand wrote:
> Eventually, we may want to just have a single blorp_ccs_op function that
> does both clears and resolves. For now we'll stick to just making the
> ccs_resolve function we have now a bit more configurable.
I had a question in patch
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #6 from Andy Furniss ---
(In reply to Nicolai Hähnle from comment #5)
> This works just fine on radeonsi, so it's likely an r600-specific problem
> that happened to be uncovered by that commit. Have you tried
On Fri, Oct 28, 2016 at 02:17:08AM -0700, Jason Ekstrand wrote:
> Signed-off-by: Jason Ekstrand
> ---
> src/intel/vulkan/anv_pass.c| 39 ---
> src/intel/vulkan/anv_private.h | 11 +++
> 2 files changed, 43 insertions(+), 7
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #7 from Christian König ---
Yeah, on SI we don't have a kernel command stream parser to double check those
things.
So instead of an error message you probably just get incorrect rendering for
maybe one frame
I'm probably not qualified enough to give a Reviewed-by line, but I can
confirm this fixes the build for me
On Thu, 3 Nov 2016 at 14:28 Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> This fixes a build regression of commit
>
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #9 from Christian König ---
(In reply to Nicolai Hähnle from comment #8)
> What flags am I missing?
Just an educated guess, but try adding "-vo opengl" to the command line.
--
You are receiving this mail
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #8 from Nicolai Hähnle ---
Hmm, so why don't I see that assertion? Following the linked mpv-player github
issue, I'm running
$ mpv --hwdec=auto Videos/Serenity.Trailer.H264.mp4
on a radeonsi debug build just
Reviewed-by: Marek Olšák
Marek
On Thu, Nov 3, 2016 at 11:04 AM, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> When splitting up loads, we have to add 16 bytes to the offset for
> the high components, just like already
Tested-by: Dieter Nützel
Argh,
should read, please go ahead and commit. ;-)
Dieter
Am 03.11.2016 15:28, schrieb Nicolai Hähnle:
From: Nicolai Hähnle
This fixes a build regression of commit
7115e56c21ace07cf04f5073ba73a533e2182099.
Sorry for
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #10 from Andy Furniss ---
(In reply to Nicolai Hähnle from comment #8)
> Hmm, so why don't I see that assertion? Following the linked mpv-player
> github issue, I'm running
>
> $ mpv --hwdec=auto
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #11 from Nicolai Hähnle ---
The -vo=opengl makes no difference. The one suspicious thing I'm seeing is:
Mesa: User error: GL_INVALID_OPERATION in VDPAUFiniNV
but mpv does call VDPAUInitNV later without error,
Tested-by: Dieter Nützel
Am 03.11.2016 15:28, schrieb Nicolai Hähnle:
From: Nicolai Hähnle
This fixes a build regression of commit
7115e56c21ace07cf04f5073ba73a533e2182099.
Sorry for the breakage, this second location for link dependencies
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #12 from Nicolai Hähnle ---
Andy, that's it. With an upgraded mpv, I get the assertion as well. I'm going
to look into it.
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA
https://bugs.freedesktop.org/show_bug.cgi?id=97967
--- Comment #3 from Emil Velikov ---
The "Failed to create foo" messages are expected and deliberate.
The actual issue is the next lines:
Error: Test 'cache_put eviction with MAX_SIZE=1K' failed: Expected=1, Actual=2
On 3 November 2016 at 14:28, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> This fixes a build regression of commit
> 7115e56c21ace07cf04f5073ba73a533e2182099.
> Sorry for the breakage, this second location for link dependencies escaped
> my build
On 3 November 2016 at 14:25, Christian Gmeiner
wrote:
> Signed-off-by: Christian Gmeiner
> ---
> include/GL/internal/dri_interface.h | 14 ++
> 1 file changed, 14 insertions(+)
>
> diff --git
On 2 November 2016 at 18:07, Marek Olšák wrote:
> From: Marek Olšák
>
> I need the definition of PUBLIC.
>
> Cc: 12.0 13.0
> ---
> src/egl/drivers/dri2/egl_dri2.h | 2 --
> src/egl/main/egldefines.h | 5 ++---
> 2
On 1 November 2016 at 19:06, Adam Jackson wrote:
> On Mon, 2016-10-31 at 20:28 +, Emil Velikov wrote:
>> >
>> > One user, xdriinfo(1), which is admittedly pretty useless. But the
>> > glvnd code in mesa definitely implements dispatch for it (libglvnd
>> > itself does not, but
On 27 October 2016 at 12:41, Emil Velikov wrote:
> Hi all,
>
> As pointed out by Mike, building Mesa with latest version of flex is
> not possible.
>
> Upon closer look it seems that the following functions are not expanded
> properly (the #define yyfoo ${prefix}foo is
On Fri, Oct 28, 2016 at 02:17:07AM -0700, Jason Ekstrand wrote:
> Signed-off-by: Jason Ekstrand
With the typo fixed in patch 10, patches 10 and 11 are:
Reviewed-by: Topi Pohjolainen
I'm working on patch 13. I'm stuck with it pretty much the
On Fri, Oct 28, 2016 at 02:17:11AM -0700, Jason Ekstrand wrote:
> Many of these UINT formats aren't available prior to Sky Lake so we used
> UNORM formats. Using UINT formats is a bit nicer because it guarantees we
> don't run into rounding issues. Also, we will need it in the next commit
> for
On Thu, 2016-11-03 at 16:21 +, Emil Velikov wrote:
> I'm not a fan of keeping legacy stuff with only one (pretty useless)
> user. Not to mention that the glXGetDriverConfig plumbing is missing
> [in here]. But as you feel so strongly in keeping it ...
Maybe I wasn't clear. I don't care at
https://bugs.freedesktop.org/show_bug.cgi?id=98578
Bug ID: 98578
Summary: AMDGPU white glitches in some games
Product: Mesa
Version: git
Hardware: Other
OS: All
Status: NEW
Severity: normal
https://bugs.freedesktop.org/show_bug.cgi?id=98578
Lukas Jirkovsky changed:
What|Removed |Added
OS|All |Linux (All)
https://bugs.freedesktop.org/show_bug.cgi?id=98578
Lukas Jirkovsky changed:
What|Removed |Added
Assignee|mesa-dev@lists.freedesktop.
On Thu, Nov 3, 2016 at 2:53 AM, Chris Wilson
wrote:
> On Wed, Nov 02, 2016 at 05:15:52PM -0700, Jason Ekstrand wrote:
> > @@ -1562,22 +1580,98 @@ VkResult anv_WaitForFences(
> > * best we can do is to clamp the timeout to INT64_MAX. This limits
> the
> > *
On 3 November 2016 at 17:12, Adam Jackson wrote:
> On Thu, 2016-11-03 at 16:21 +, Emil Velikov wrote:
>
>> I'm not a fan of keeping legacy stuff with only one (pretty useless)
>> user. Not to mention that the glXGetDriverConfig plumbing is missing
>> [in here]. But as you
Explain why it's 64-bit and its valid values.
---
src/intel/vulkan/anv_private.h | 4
1 file changed, 4 insertions(+)
diff --git a/src/intel/vulkan/anv_private.h b/src/intel/vulkan/anv_private.h
index 8f5a95b..0924ccf 100644
--- a/src/intel/vulkan/anv_private.h
+++
On Wed 02 Nov 2016, Jason Ekstrand wrote:
> ---
> src/intel/vulkan/gen8_cmd_buffer.c | 29 +
> 1 file changed, 29 insertions(+)
>
> diff --git a/src/intel/vulkan/gen8_cmd_buffer.c
> b/src/intel/vulkan/gen8_cmd_buffer.c
> index 0548a5e..5df2e50 100644
> ---
https://bugs.freedesktop.org/show_bug.cgi?id=98471
Emil Velikov changed:
What|Removed |Added
Depends on|96865 |
Referenced
https://bugs.freedesktop.org/show_bug.cgi?id=96865
Emil Velikov changed:
What|Removed |Added
Blocks|98471 |
--- Comment
Patch 1 looks good to me. Yes, not using compilation key is a recipe for
problems.
Regarding patch 2, I need a bit more time to review. I'd also like the new
lp_build_umul... function to be added to lp_bld_arit: there is a single caller
note, but that might change in the future, so might as
On Wed 02 Nov 2016, Jason Ekstrand wrote:
> ---
> src/intel/vulkan/anv_wsi.c | 12 +---
> 1 file changed, 9 insertions(+), 3 deletions(-)
>
> diff --git a/src/intel/vulkan/anv_wsi.c b/src/intel/vulkan/anv_wsi.c
> index 064581d..72f79f1 100644
> --- a/src/intel/vulkan/anv_wsi.c
> +++
Hi Emil,
2016-11-03 17:02 GMT+01:00 Emil Velikov :
> On 3 November 2016 at 14:25, Christian Gmeiner
> wrote:
>> Signed-off-by: Christian Gmeiner
>> ---
>> include/GL/internal/dri_interface.h | 14 ++
On Wed, Nov 02, 2016 at 08:00:04PM -0700, Jason Ekstrand wrote:
> Reviewed-by: Jason Ekstrand
>
Thanks!
> Wouldn't be a bad idea to do this for all compressed formats. It doesn't
> really matter though since blorp will nuke the format the moment it cones
> into
On Thu, 2016-11-03 at 17:28 +, Emil Velikov wrote:
> Your earlier reply did not have ack/nack on the topic of dropping it,
> thus the above reads as a justification why one would wants to keep
> it.
I thought "useless" expressed my opinion of the value of the feature
quite well. But I didn't
On Wed 02 Nov 2016, Jason Ekstrand wrote:
> Our previous fence implementation was very simple. Fences had two states:
> signaled and unsignaled. However, this didn't properly handle all of the
> edge-cases that we need to handle. In order to handle the case where the
> client calls
https://bugs.freedesktop.org/show_bug.cgi?id=98279
--- Comment #4 from Diego Vogogna ---
Getting the same errors in log, amd R9 390, Dota 2 and vulkan cube from sdk
tests freeze the system.
Running Arch x86_64
linux-git 4.9rc3.r206.g2a26d99-1 with no additional pathces
https://bugs.freedesktop.org/show_bug.cgi?id=98279
Diego Vogogna changed:
What|Removed |Added
CC||ars...@gmail.com
--
From: Marek Olšák
This fixes a crash in Deus Ex: Mankind Divided. Release builds were
unaffected, so it's not too serious.
Cc: 11.2 12.0 13.0
---
src/gallium/drivers/radeonsi/si_blit.c | 4 +++-
1 file changed, 3 insertions(+), 1
https://bugs.freedesktop.org/show_bug.cgi?id=96684
--- Comment #2 from Emil Velikov ---
I can reproduce this here so gave it a quick look:
The GLSL IR is identical with _and_ w/o the optimisations (MESA_GLSL=nopt), yet
the result changes to pass. Seems like the
On Thu, Nov 3, 2016 at 10:59 AM, Chad Versace
wrote:
> On Wed 02 Nov 2016, Jason Ekstrand wrote:
> > Our previous fence implementation was very simple. Fences had two
> states:
> > signaled and unsignaled. However, this didn't properly handle all of the
> > edge-cases
Reviewed-by: Bas Nieuwenhuizen
On Thu, Nov 3, 2016 at 7:22 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> This fixes a crash in Deus Ex: Mankind Divided. Release builds were
> unaffected, so it's not too serious.
>
> Cc: 11.2
https://bugs.freedesktop.org/show_bug.cgi?id=97863
Matt Turner changed:
What|Removed |Added
Assignee|mesa-dev@lists.freedesktop.
https://bugs.freedesktop.org/show_bug.cgi?id=77449
Bug 77449 depends on bug 97890, which changed state.
Bug 97890 Summary: Kerbal Space Program freezes randomly when DRI3 is enabled
https://bugs.freedesktop.org/show_bug.cgi?id=97890
What|Removed |Added
On 3 November 2016 at 17:53, Christian Gmeiner
wrote:
> Hi Emil,
>
> 2016-11-03 17:02 GMT+01:00 Emil Velikov :
>> On 3 November 2016 at 14:25, Christian Gmeiner
>> wrote:
>>> Signed-off-by: Christian Gmeiner
On 03.11.2016 19:22, Marek Olšák wrote:
From: Marek Olšák
This fixes a crash in Deus Ex: Mankind Divided. Release builds were
unaffected, so it's not too serious.
Cc: 11.2 12.0 13.0
Reviewed-by: Nicolai Hähnle
On 10/31/2016 07:06 PM, Timothy Arceri wrote:
> On Mon, 2016-10-31 at 16:51 +0100, Juan A. Suarez Romero wrote:
>> On Wed, 2016-10-26 at 13:42 +0200, Juan A. Suarez Romero wrote:
>>>
>>> Ignore source file, line number and column in glcpp_error() and
>>> glcpp_warning() if those are not available.
On 10/28/2016 04:13 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/compiler/glsl/ir_optimization.h | 3 ++-
> src/compiler/glsl/lower_if_to_cond_assign.cpp | 23 ---
> src/compiler/glsl/test_optpass.cpp| 2 +-
>
On 10/28/2016 04:13 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/compiler/glsl/ir_optimization.h | 2 +-
> src/compiler/glsl/lower_if_to_cond_assign.cpp | 55
> ---
> 2 files changed, 50 insertions(+), 7 deletions(-)
>
> diff
Ian Romanick writes:
> On 10/28/2016 04:13 PM, Marek Olšák wrote:
>> From: Marek Olšák
>>
>> ---
>> src/compiler/glsl/ir_optimization.h | 3 ++-
>> src/compiler/glsl/lower_if_to_cond_assign.cpp | 23 ---
>>
On 10/28/2016 04:13 PM, Marek Olšák wrote:
> Hi,
>
> this series enables lowering small branches for radeonsi, but other
> drivers can enable it too. The GLSL compiler computes a cost per
> branch and then either lowers the branch if the cost is low enough,
> or keeps it.
As an aside...
When
On Thu 03 Nov 2016, Jason Ekstrand wrote:
>
>
> On Thu, Nov 3, 2016 at 10:59 AM, Chad Versace
> wrote:
>
> On Wed 02 Nov 2016, Jason Ekstrand wrote:
> > @@ -1116,6 +1117,11 @@ VkResult anv_QueueSubmit(
> > result = anv_device_execbuf(device,
From: Nicolai Hähnle
A latent bug in VDPAU interop was exposed by commit
e5cc84dd43be066c1dd418e32f5ad258e31a150a.
Before that commit, the st_vdpau code created samplers with
first_layer == last_layer == 1 that the general texture handling code
would immediately delete
https://bugs.freedesktop.org/show_bug.cgi?id=98512
--- Comment #13 from Nicolai Hähnle ---
I think I understand the problem.
https://patchwork.freedesktop.org/patch/119855/ fixes the issue for me, could
you please check that it fixes r600 as well?
--
You are receiving this
On Thu, Nov 3, 2016 at 1:53 AM, Chris Wilson
wrote:
> On Wed, Nov 02, 2016 at 01:58:43PM -0700, Jason Ekstrand wrote:
> > +/**
> > + * This function applies the relocation for a command buffer and writes
> the
> > + * actual addresses into the buffers as per what we
Add code to detect avx512 cpu features, and allow llvm to generate
avx512 instructions with llvm-3.9+. gallivm will still treat avx512
capable processors as avx2 (256-bit vectors).
No piglit regressions with llvmpipe on llvm-3.9.0 and KNL.
Tim Rowley (2):
gallium: detect avx512 cpu features
v3: fix check for xmm/ymm test
v2: style code, add avx512 to cpu dump
---
src/gallium/auxiliary/util/u_cpu_detect.c | 26 ++
src/gallium/auxiliary/util/u_cpu_detect.h | 10 ++
2 files changed, 36 insertions(+)
diff --git a/src/gallium/auxiliary/util/u_cpu_detect.c
---
src/gallium/auxiliary/gallivm/lp_bld_misc.cpp | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp
b/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp
index bd4d4d3..bff2198 100644
--- a/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp
+++
On 3 November 2016 at 17:59, Adam Jackson wrote:
> On Thu, 2016-11-03 at 17:28 +, Emil Velikov wrote:
>
>> Your earlier reply did not have ack/nack on the topic of dropping it,
>> thus the above reads as a justification why one would wants to keep
>> it.
>
> I thought
1 - 100 of 133 matches
Mail list logo