This has been ported to NIR now so we don'tneed to keep the GLSL IR
lowering any more.
---
src/mesa/drivers/dri/i965/Makefile.sources | 1 -
src/mesa/drivers/dri/i965/brw_context.h| 2 -
src/mesa/drivers/dri/i965/brw_fs_generator.cpp | 2 +-
This gets the lowering on the Vulkan driver too, which is required for
hardware that does not have the sample_l_d message (up to IvyBridge).
---
src/mesa/drivers/dri/i965/brw_nir.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/mesa/drivers/dri/i965/brw_nir.c
This is ported from the Intel lowering pass that we use with GLSL IR.
This takes care of lowering texture gradients on shadow samplers other
than cube maps. Intel hardware requires this for gen < 8.
---
src/compiler/nir/nir.h | 7 +++
src/compiler/nir/nir_lower_tex.c | 40
We need this lowering in Vulkan too do texture gradients with cube map textures
(and also with shadow samplers in IvyBridge).
Currently, we have the brw_lower_texture_gradients() pass that works on GLSL IR
and is called from the OpenGL driver. This pass is actually two passes in one:
1. A
This gets the lowering on the Vulkan driver too.
Fixes Vulkan CTS cube map texture gradient tests in:
dEQP-VK.glsl.texture_functions.texturegrad.*
---
src/mesa/drivers/dri/i965/brw_nir.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/mesa/drivers/dri/i965/brw_nir.c
This was written specifically for RECT samplers. Make it more generic so
we can call this from the gradient lowerings too.
---
src/compiler/nir/nir_lower_tex.c | 15 ++-
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/src/compiler/nir/nir_lower_tex.c
Hi,
On Fri, Nov 25, 2016 at 12:58 PM, Liu Zhiquan wrote:
> Some dri drivers will pass multiple bits in buffer_mask parameter
> to droid_image_get_buffer(), more than the actual supported buffer
> type combination. For such case, will go through all the bits, and
> will not
This will be useful for splitting blits into smaller sizes.
We also make the coordinates of type double rather than float. Since
we will be splitting and scaling the coordinates, we might require
extra precision in the calculations.
v2:
* Use double instead of float. (Jason)
Signed-off-by:
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98297
Signed-off-by: Jordan Justen
---
src/mesa/drivers/dri/i965/brw_context.c | 13 ++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_context.c
We rename do_blorp_blit() to try_blorp_blit(), and add a return error
if the surface size for the blit is too large. Now, do_blorp_blit() is
rewritten to try to split the blit into smaller operations if
try_blorp_blit() fails.
Note: In this commit, try_blorp_blit() will always attempt to blit and
If try_blorp_blit() previously returned that a blit was too large,
shrink_surface_params() will be used to update the surface parameters
for the smaller blit so the blit operation can proceed.
v2:
* Use double instead of float. (Jason)
Signed-off-by: Jordan Justen
Enabling this debug switch causes surface shrinking to happen by
default, and lowers the surface size limit which causes blorp blits to
be split.
Signed-off-by: Jordan Justen
---
src/intel/blorp/blorp_blit.c | 11 ---
1 file changed, 8 insertions(+), 3
Detect when the surface sizes are too large for a blorp blit. When it
is too large, the blorp blit will be split into a smaller operation
and attempted again.
For gen7, this fixes the cts test:
In blorp_copy, when RGB surfaces are copied, we convert the
destination surface to a Red only surface, but 3 times as wide. This
introduces an implicit restriction of "mod 3" for the destination
width.
It is easier to handle the blorp split buffer offsetting with the
original RGB surface, and do
git://people.freedesktop.org/~jljusten/mesa split-blorp-blit-v2
(These patches have been applied on top of my ISL surface dumping
patches which are included in the branch above.)
v1:
https://lists.freedesktop.org/archives/mesa-dev/2016-November/136235.html
v2:
* Simplify the logic to shrink
A couple of notes on existing weirdness here:
- Naming of GEN9_PSX_SHADER_NORMAL_COVERAGE_MASK_SHIFT is bizarre (not your
fault)
- Is BRW_PSICMS_INNER really the right thing for the normal mode? Why not
BRW_PSICMS_NORMAL? Perhaps whoever added this stuff can shed some light
here?
Actual change
Enabling this option causes the source and destination surfaces to be
dumped out to debug files. The destination is dumped both before and
after the blit operation.
Signed-off-by: Jordan Justen
---
src/intel/blorp/blorp_blit.c | 18 ++
1 file changed,
v2:
* Use new blorp_surf_dump function. (Jason)
Signed-off-by: Jordan Justen
Acked-by: Matt Turner
Cc: Jason Ekstrand
---
src/mesa/drivers/dri/i965/brw_blorp.c | 54 +++
1 file changed, 36
Signed-off-by: Jordan Justen
---
src/intel/Makefile.sources | 1 +
src/intel/isl/isl.h| 14 +++
src/intel/isl/isl_dump.c | 217 +
3 files changed, 232 insertions(+)
create mode 100644 src/intel/isl/isl_dump.c
This function takes care of mapping and unmapping the buffers, and
then calls isl_surf_dump.
Signed-off-by: Jordan Justen
---
src/intel/blorp/blorp.c | 35 +++
src/intel/blorp/blorp.h | 6 ++
2 files changed, 41 insertions(+)
diff
Signed-off-by: Jordan Justen
---
src/intel/blorp/blorp.h | 8 +++
src/mesa/drivers/dri/i965/brw_blorp.c | 39 +++
2 files changed, 47 insertions(+)
diff --git a/src/intel/blorp/blorp.h b/src/intel/blorp/blorp.h
index
On Tue, 2016-11-22 at 23:53 +0200, Plamena Manolova wrote:
> This extension allows the fragment shader to control whether values
> in
> gl_SampleMaskIn[] reflect the coverage after application of the early
> depth and stencil tests.
>
> Signed-off-by: Plamena Manolova
https://bugs.freedesktop.org/show_bug.cgi?id=97524
--- Comment #16 from Marek Olšák ---
Workaround for radeonsi:
CI-VI:
If the resource type and instruction mismatch (e.g. a buffer constant with an
image instruction, or an image resource with a buffer instruction), the
On Wed, Nov 30, 2016 at 3:00 PM, Plamena Manolova
wrote:
> This extension allows the fragment shader to control whether values in
> gl_SampleMaskIn[] reflect the coverage after application of the early
> depth and stencil tests.
>
> Signed-off-by: Plamena Manolova
It's possible that nir_shader was cloned and it no longer contains
a pointer to the shader_info in gl_program. So we need to copy
shader_info back to gl_program if that is the case.
Fixes a regression with NIR_TEST_CLONE=true
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98840
---
On Wed, Nov 30, 2016 at 5:04 PM, Ilia Mirkin wrote:
> We were previously also verifying that no backing buffers were available
> when an array wasn't enabled. This is has no basis in the spec, and it
> causes GLupeN64 to fail as a result.
>
> Fixes: c2e146f487 ("mesa: error
On Wed, Nov 30, 2016 at 2:53 PM, Plamena Manolova
wrote:
> This extension allows the fragment shader to control whether values in
> gl_SampleMaskIn[] reflect the coverage after application of the early
> depth and stencil tests.
>
> Signed-off-by: Plamena Manolova
Excellent, disregard that. Patch looks good.
On Thu, Dec 1, 2016 at 3:10 PM, Ilia Mirkin wrote:
> On Wed, Nov 30, 2016 at 9:10 PM, Chris Forbes wrote:
> > This patch misses adding the #define to the GLSL preprocessor. Other than
>
> The future is today.
On Wed, Nov 30, 2016 at 9:10 PM, Chris Forbes wrote:
> This patch misses adding the #define to the GLSL preprocessor. Other than
The future is today. That's no longer necessary :)
> that it looks good though, so with that fixed:
>
> Reviewed-by: Chris Forbes
This patch misses adding the #define to the GLSL preprocessor. Other than
that it looks good though, so with that fixed:
Reviewed-by: Chris Forbes
On Thu, Dec 1, 2016 at 8:53 AM, Plamena Manolova wrote:
> This extension allows the fragment
Build mesa 2856 completed
Commit fda1d0187d by Ilia Mirkin on 11/30/2016 5:56 AM:
anv: expose support for VK_KHR_sampler_mirror_clamp_to_edge\n\nThis is already supported in genX_state.c, expose the extension string.\n\nSigned-off-by: Ilia Mirkin
On Wed, Nov 30, 2016 at 10:20 AM, Nanley Chery
wrote:
> On Tue, Nov 29, 2016 at 05:41:58PM -0800, Jason Ekstrand wrote:
> > In an attempt to fix 3DSTATE_DEPTH_BUFFER for stencil-only cases, I
> > accidentally kept setting the SurfaceType to 2D in the stencil-only case
> >
Build mesa 2855 failed
Commit 27433b26b1 by Jason Ekstrand on 11/29/2016 10:00 PM:
anv/cmd_buffer: Actually use the stencil dimension\n\nIn an attempt to fix 3DSTATE_DEPTH_BUFFER for stencil-only cases, I\naccidentally kept setting the SurfaceType to 2D in the
https://bugs.freedesktop.org/show_bug.cgi?id=98911
Kenneth Graunke changed:
What|Removed |Added
Component|Mesa core
Weren't we going to call it 17.0?...
On Nov 30, 2016 12:23 PM, "Emil Velikov" wrote:
> Hi all,
>
> With holidays not far off, it might be a nice idea to consider the
> branchpoint/release schedule for the next release.
>
> I will be having limited internet access
Build mesa 2852 failed
Commit 966567aa12 by Timothy Arceri on 11/30/2016 9:52 PM:
mesa: reset linked_stages bitmask when re-linking\n\n34953f8907fdd added this bitmask but it wasn't being reset when\na program was relinked. If a stage was removed from the
Signed-off-by: Bas Nieuwenhuizen
---
src/amd/common/ac_nir_to_llvm.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index df121d8..87bb19a 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:23 PM, Ilia Mirkin
> wrote:
The buffer_size does not take the offset into account. Just add the
offset into the pointer which lines up
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:23 PM, Ilia Mirkin
> wrote:
Signed-off-by: Ilia Mirkin >
---
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:23 PM, Ilia Mirkin
> wrote:
The number has to be less than or equal to the max, not just less than.
Signed-off-by: Ilia Mirkin
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:23 PM, Ilia Mirkin
> wrote:
The components count the number of individual values, not the number of
slots.
Signed-off-by: Ilia Mirkin
We were previously also verifying that no backing buffers were available
when an array wasn't enabled. This is has no basis in the spec, and it
causes GLupeN64 to fail as a result.
Fixes: c2e146f487 ("mesa: error out in indirect draw when vertex bindings
mismatch")
Cc:
34953f8907fdd added this bitmask but it wasn't being reset when
a program was relinked. If a stage was removed from the new
program then it could case a crash as we expect the linked shader
for that stage to not be null.
Fixes crashes in:
On 11/28, Ian Romanick wrote:
From: Ian Romanick
I believe that I have addressed all of the review feedback from the
previous iteration. Many of the patches have been reviewed, and they
should be ready to go.
Patches marked with "vN" in the subject have changed in a
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:23 PM, Ilia Mirkin
> wrote:
There is no support for resuming streamout. Furthermore, this also
controls glDrawTransformFeedback
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:05 PM, Ilia Mirkin
> wrote:
We now support clearing these, and actually rendering to multiple layers
would require GS support, which
Reviewed-by: Tim Rowley
>
On Nov 16, 2016, at 9:04 PM, Ilia Mirkin
> wrote:
Since we don't pass a renderTargetArrayIndex in, and the current hot
tile may be for a different index, we
Am 30.11.2016 um 20:19 schrieb Nicolai Hähnle:
> On 30.11.2016 19:06, Roland Scheidegger wrote:
>> Am 30.11.2016 um 14:35 schrieb Nicolai Hähnle:
>>> From: Nicolai Hähnle
>>>
>>> This is for geometry shader outputs. Without it, drivers have no way of
>>> knowing which
Not seeing this assert fire on our tests either.
Reviewed-by: Tim Rowley
>
On Nov 29, 2016, at 8:04 PM, Ilia Mirkin
> wrote:
Signed-off-by: Ilia Mirkin
On 11/28, Ian Romanick wrote:
From: Ian Romanick
These functions are directly available in shaders. A #define is added
to detect the presence. This allows these functions to be tested using
piglit regardless of whether the driver uses them for lowering. The
GLSL
On 11/28, Ian Romanick wrote:
From: Ian Romanick
Signed-off-by: Ian Romanick
---
src/compiler/Makefile.glsl.am| 1 +
src/compiler/Makefile.sources| 1 +
src/compiler/glsl/ir_optimization.h | 6 +
Hi all,
With holidays not far off, it might be a nice idea to consider the
branchpoint/release schedule for the next release.
I will be having limited internet access during 20 Dec - 7 Jan, thus
the I'm leaning towards following:
Jan 13 2017 - Feature freeze/Release candidate 1
Jan 20 2017 -
On 11/28, Ian Romanick wrote:
From: Ian Romanick
These functions are directly available in shaders. A #define is added
to detect the presence. This allows these functions to be tested using
piglit regardless of whether the driver uses them for lowering. The
GLSL
This extension allows the fragment shader to control whether values in
gl_SampleMaskIn[] reflect the coverage after application of the early
depth and stencil tests.
Signed-off-by: Plamena Manolova
---
docs/relnotes/13.1.0.html| 1 +
This extension allows the fragment shader to control whether values in
gl_SampleMaskIn[] reflect the coverage after application of the early
depth and stencil tests.
Signed-off-by: Plamena Manolova
---
src/compiler/glsl/ast.h | 5 +
On 11/28, Ian Romanick wrote:
From: Ian Romanick
Signed-off-by: Ian Romanick
---
src/compiler/glsl/ir_builder_print_visitor.cpp | 16 +---
src/compiler/glsl/standalone_scaffolding.cpp | 1 +
2 files changed, 14 insertions(+), 3
https://bugs.freedesktop.org/show_bug.cgi?id=98428
Jonathan Dieter changed:
What|Removed |Added
CC||jdie...@lesbg.com
On 30.11.2016 19:06, Roland Scheidegger wrote:
Am 30.11.2016 um 14:35 schrieb Nicolai Hähnle:
From: Nicolai Hähnle
This is for geometry shader outputs. Without it, drivers have no way of
knowing which stream each output is intended for, and have to
conservatively
Am 30.11.2016 um 14:35 schrieb Nicolai Hähnle:
> From: Nicolai Hähnle
>
> This is for geometry shader outputs. Without it, drivers have no way of
> knowing which stream each output is intended for, and have to
> conservatively write all outputs to all streams.
>
>
On Tue, Nov 29, 2016 at 05:41:58PM -0800, Jason Ekstrand wrote:
> In an attempt to fix 3DSTATE_DEPTH_BUFFER for stencil-only cases, I
> accidentally kept setting the SurfaceType to 2D in the stencil-only case
> thanks to a copy+paste error.
>
> Cc: Nanley Chery
> ---
>
On Tue, Nov 29, 2016 at 08:54:26AM -0800, Jason Ekstrand wrote:
> On Tue, Nov 29, 2016 at 8:47 AM, Ville Syrjälä <
> ville.syrj...@linux.intel.com> wrote:
>
> > On Tue, Nov 29, 2016 at 08:28:55AM -0800, Jason Ekstrand wrote:
> > > On Tue, Nov 29, 2016 at 1:23 AM,
Reviewed-by: Bruce Cherniak
> On Nov 29, 2016, at 8:18 PM, Ilia Mirkin wrote:
>
> We need to take the instance divisor and number of instances into
> account for instanced client-side arrays, rather than the vertex
> parameters.
>
> Loosely
https://bugs.freedesktop.org/show_bug.cgi?id=98911
Mark Janes changed:
What|Removed |Added
CC|
On Tue 29 Nov 2016, ville.syrj...@linux.intel.com wrote:
> From: Ville Syrjälä
>
> Set the include paths to consider in-tree headers before out-of-tree
> headers.
>
> Avoids the build failing due to stale headers being present in
> $prefix. Previosuly 'make -ki
On 30.11.2016 14:35, Nicolai Hähnle wrote:
From: Nicolai Hähnle
Vertex emits beyond the specified maximum number of vertices are supposed to
have no effect, which is why we used to always kill GS that reached the limit.
However, if the GS also writes to memory (SSBO,
On Wed, Nov 30, 2016 at 4:36 PM, Ilia Mirkin wrote:
> On Wed, Nov 30, 2016 at 9:19 AM, Lionel Landwerlin
> wrote:
> > Also forgot that (like Ilia suggested for NV_image_formats) you can
> update
> > docs/relnotes/ to list the new feature.
>
>
Thank you for reviewing guys! I'll go ahead and make that change.
Pam
On Wed, Nov 30, 2016 at 3:33 PM, Ilia Mirkin wrote:
> On Nov 30, 2016 7:17 AM, "Lionel Landwerlin" <
> lionel.g.landwer...@intel.com> wrote:
> >
> > On 22/11/16 21:53, Plamena Manolova wrote:
> >>
> >>
On 30.11.2016 15:01, Nicolai Hähnle wrote:
On 30.11.2016 02:36, Marek Olšák wrote:
From: Marek Olšák
Internal docs don't mention it, but they also don't mention that the bug
has been fixed (like other CI bugs fixed in VI).
Vulkan does this too.
Cc: 13.0
On Wed, Nov 30, 2016 at 9:19 AM, Lionel Landwerlin
wrote:
> Also forgot that (like Ilia suggested for NV_image_formats) you can update
> docs/relnotes/ to list the new feature.
That would be appropriate for the patch that actually exposes the
feature. This one just
Also forgot that (like Ilia suggested for NV_image_formats) you can
update docs/relnotes/ to list the new feature.
Cheers,
-
Lionel
On 30/11/16 13:33, Ilia Mirkin wrote:
On Nov 30, 2016 7:17 AM, "Lionel Landwerlin"
>
I didn't spot any relevant discussion on
http://help.appveyor.com/discussions/problems or
https://appveyor.statuspage.io/
It must have been something transient.
Jose
On 30/11/16 03:52, Brian Paul wrote:
We've had a few instances in the past where complex macro code caused
super long MSVC
https://bugs.freedesktop.org/show_bug.cgi?id=98911
--- Comment #1 from paed808 ---
Here is a video I recorded of the issue.
https://www.youtube.com/watch?v=Iatb2PQPSFI
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for
Apart from the comment on patch 7, the series is
Reviewed-by: Nicolai Hähnle
On 30.11.2016 02:36, Marek Olšák wrote:
From: Marek Olšák
This one is easy to miss, because it's not documented in any internal doc.
---
src/amd/common/sid.h | 6
On 30.11.2016 02:36, Marek Olšák wrote:
From: Marek Olšák
Internal docs don't mention it, but they also don't mention that the bug
has been fixed (like other CI bugs fixed in VI).
Vulkan does this too.
Cc: 13.0
---
Hi all,
this time I send the patchset to the correct ml.
Had no coffee when sending it out for the first time :)
greets
--
Christian Gmeiner, MSc
https://soundcloud.com/christian-gmeiner
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
The imx (stub) driver is needed to get hardware acceleration from
etnaviv on a platform using imx-drm kms driver. This adds support
for wayland and native kms egl apps.
Signed-off-by: Christian Gmeiner
---
configure.ac | 12
This a very lightweight library to add basic support for
renderonly GPUs. It does all the magic regarding in/exporting
buffers etc. This library will likely break android support and
hopefully will get replaced with a better solution based on gbm2.
Signed-off-by: Christian Gmeiner
As the original patchstack is now about 300 patches, I have choosen to
squash the patches together into three different parts.
- renderonly library
A lightweight library to add basic infrastructure for renderonly
GPUs. With this library it is possible to run wayland or and other
kms egl
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c| 26 +++--
src/gallium/drivers/radeonsi/si_state_shaders.c | 37 +++--
2 files changed, 39 insertions(+), 24 deletions(-)
diff --git
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 41 +++-
1 file changed, 25 insertions(+), 16 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index
From: Nicolai Hähnle
---
src/compiler/glsl/ir.h | 3 +++
src/compiler/glsl/ir_print_visitor.cpp | 14 --
src/compiler/glsl/lower_packed_varyings.cpp | 9 +
3 files changed, 24 insertions(+), 2 deletions(-)
diff --git
From: Nicolai Hähnle
We can hardcode all of the fields for swizzling in the geometry shader.
The advantage is that we use fewer descriptor slots and we no longer have to
update any of the (ring) descriptors when the geometry shader changes.
---
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_state_shaders.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_state_shaders.c
b/src/gallium/drivers/radeonsi/si_state_shaders.c
index 151ed17..4a7f638 100644
From: Nicolai Hähnle
The iteration is not needed for normal vertex shaders.
---
src/gallium/drivers/radeonsi/si_shader.c | 62 +++-
1 file changed, 37 insertions(+), 25 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
From: Nicolai Hähnle
Allow drivers to emit GS outputs in a smarter way.
---
src/gallium/auxiliary/tgsi/tgsi_ureg.c | 16 +++-
src/gallium/auxiliary/tgsi/tgsi_ureg.h | 1 +
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 10 ++
3 files changed,
From: Nicolai Hähnle
When running the copy shader for vertex streams != 0, the SX does not need
any data from us (there is no rasterization for the higher vertex streams,
only streamout).
---
src/gallium/drivers/radeonsi/si_shader.c | 31 +++
From: Nicolai Hähnle
So that we can pass pointers to individual array entries around.
---
src/gallium/include/pipe/p_state.h | 21 +
1 file changed, 13 insertions(+), 8 deletions(-)
diff --git a/src/gallium/include/pipe/p_state.h
From: Nicolai Hähnle
Note that the memory layout of one vertex stream inside one "item" (= memory
written by one GS wave) on the GSVS ring is:
t0v0c0 ... t15v0c0 t0v1c0 ... t15v1c0 ... t0vLc0 ... t15vLc0
t0v0c1 ... t15v0c1 t0v1c1 ... t15v1c1 ... t0vLc1 ... t15vLc1
From: Nicolai Hähnle
This affects for GS copy shaders. When an output is meant for vertex
stream != 0, then we don't have to make it available to the pixel
shader.
There is a minor inefficiency here because the GLSL varying packing pass
does not group varyings of the
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 32 ++--
1 file changed, 22 insertions(+), 10 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 91 ++--
1 file changed, 52 insertions(+), 39 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 10 +-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index c488f91..b420a7b 100644
---
From: Nicolai Hähnle
Vertex emits beyond the specified maximum number of vertices are supposed to
have no effect, which is why we used to always kill GS that reached the limit.
However, if the GS also writes to memory (SSBO, atomics, shader images), then
we must keep
From: Nicolai Hähnle
SimplifyCFG generates a switch instruction anyway when all four streams
are present, but is simultaneously not smart enough to eliminate some
redundant jumps that it generates.
The generated assembly is still a bit silly, probably because the
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index 0eeff19..b861831 100644
---
From: Nicolai Hähnle
This is for geometry shader outputs. Without it, drivers have no way of
knowing which stream each output is intended for, and have to
conservatively write all outputs to all streams.
Separate stream numbers for each component are required due to
From: Nicolai Hähnle
Be a bit more verbose and avoid confusion in future patches.
---
src/gallium/drivers/radeonsi/si_shader.c | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
From: Nicolai Hähnle
LLVM can still decide to hoist the loads since they're marked invariant.
---
src/gallium/drivers/radeonsi/si_shader.c | 25 +++--
1 file changed, 11 insertions(+), 14 deletions(-)
diff --git
From: Nicolai Hähnle
Fixes
GL45-CTS.gtf40.GL3Tests.transform_feedback3.transform_feedback3_geometry_instanced.
Cc: mesa-sta...@lists.freedesktop.org
---
src/gallium/drivers/radeonsi/si_state_shaders.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff
From: Nicolai Hähnle
---
src/gallium/auxiliary/tgsi/tgsi_scan.c | 9 +
src/gallium/auxiliary/tgsi/tgsi_scan.h | 2 ++
2 files changed, 11 insertions(+)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_scan.c
b/src/gallium/auxiliary/tgsi/tgsi_scan.c
index
1 - 100 of 110 matches
Mail list logo