From: Dave Airlie
This passes the CTS and piglit tests.
This also disable sb for helper invocations until it doesn't
mess up the VPM flags.
Thanks to Ilia and Glenn for advice, and Roland for working
out the working evergreen path.
---
src/gallium/drivers/r600/r600_asm.c| 7 +-
src/galli
https://bugs.freedesktop.org/show_bug.cgi?id=104732
--- Comment #1 from Samuel Pitoiset ---
This should fix the issue
https://cgit.freedesktop.org/~hakzsam/mesa/commit/?id=4922e7f25c1a42779bab11aff6489d8a39a17061
Can you confirm? Thanks!
--
You are receiving this mail because:
You are the QA C
Am Mittwoch, 17. Januar 2018, 18:41:00 CET schrieb Dylan Baker:
> I dug into what autotools is doing. It's using -Wl,--whole-archive to link
> everything, in this target, not just the state tracker. I think I'm okay
> with this patch given that, so I'm going to apply the same change to the
> other
On Wed, Jan 31, 2018 at 10:39 PM, Brian Paul wrote:
> And rename indirect_params -> indirect_draw_count_buffer and
> indirect_params_offset -> indirect_draw_count_offset to be more
> specific.
> ---
> src/mesa/vbo/vbo.h | 57
> --
> 1 file chan
You might be interested with https://hastebin.com/vekuhojapo.
I have patches that do the same optimization for RADV, and my pipeline
stats are quite similar.
No regressions detected with CTS.
Series is:
Reviewed-by: Samuel Pitoiset
On 01/30/2018 10:46 PM, Marek Olšák wrote:
From: Marek Ol
Totals from affected shaders:
SGPRS: 142032 -> 137651 (-3.08 %)
VGPRS: 93280 -> 92992 (-0.31 %)
Spilled SGPRs: 104 -> 129 (24.04 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Private memory VGPRs: 0 -> 0 (0.00 %)
Scratch size: 0 -> 0 (0.00 %) dwords per thread
Code Size: 3309852 -> 3301444 (-0.25 %) bytes
LDS:
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_shader_info.c | 13 +
src/amd/common/ac_shader_info.h | 3 +++
2 files changed, 16 insertions(+)
diff --git a/src/amd/common/ac_shader_info.c b/src/amd/common/ac_shader_info.c
index d771cd250d..e713263817 100644
--- a/src/amd/comm
There's a bug about quiting on unsupported tags:
https://bugzilla.redhat.com/show_bug.cgi?id=1348200
Should be fixed in v1.10
http://pkgs.fedoraproject.org/cgit/rpms/dwarves.git/commit/?h=f25
On Wed, Jan 31, 2018 at 5:57 PM, Brian Paul wrote:
> On 01/31/2018 01:48 AM, Tapani Pälli wrote:
>
>> R
On 31 January 2018 at 19:33, Kenneth Graunke wrote:
> In commit 3f353342a6b6744773c26ed66b12afed42bd57af (present in 17.3.0)
> we started unconditionally using I915_EXEC_NO_RELOC, which was
> introduced in Linux v3.9. ChromeOS kernel 3.8 has backported this,
> so it should work too.
>
> Running o
On 25 January 2018 at 15:11, Stephan Gerhold wrote:
> On Thu, Jan 25, 2018 at 11:22:10AM +, Emil Velikov wrote:
>> On 24 January 2018 at 14:13, Stephan Gerhold wrote:
>> > build_id_find_nhdr_for_addr() fails to find the build-id if the first LOAD
>> > segment has a virtual address other than
This fixes the issue;
Tested-by: Tapani Pälli
On 02/01/2018 03:41 AM, Jason Ekstrand wrote:
This fixes some rendering corruption in a couple of Android apps that
use window-system MSAA.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104741
Cc: mesa-sta...@lists.freedesktop.org
Cc: Chad
On Mon, 2018-01-29 at 16:25 -0800, Eric Anholt wrote:
> "Juan A. Suarez Romero" writes:
>
> > Enable these kind of 3D texture when proper extensions are available.
> >
> > Fixes KHR-GLES2.texture_3d.* with these textures.
> >
> > v2: add better clarification (Eric)
> > ---
> > src/mesa/main/te
On 28 January 2018 at 14:24, Jon Turney wrote:
> From: Dylan Baker
>
> ---
> meson.build | 8 +++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/meson.build b/meson.build
> index 824e3c46bc5..178743eddb1 100644
> --- a/meson.build
> +++ b/meson.build
> @@ -642,7 +642,13 @
Only dive into the windows subdir if windows platform is selected.
Signed-off-by: Marc Dietrich
---
src/glx/meson.build | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/glx/meson.build b/src/glx/meson.build
index 04cd647ee4..e645b22b5f 100644
--- a/src/glx/meson.build
+
Hats off for sorting these out Jon!
On 28 January 2018 at 14:18, Jon Turney wrote:
> Jon Turney (6):
> configure: Default to gbm=no on osx
> osx: ld doesn't support --build-id
> glx/apple: include util/debug.h for env_var_as_boolean prototype
> glx/apple: locate dispatch table functions t
On 28 January 2018 at 14:24, Jon Turney wrote:
> ---
> .travis.yml | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/.travis.yml b/.travis.yml
> index 9cecf2f615f..7e6bbfe306b 100644
> --- a/.travis.yml
> +++ b/.travis.yml
> @@ -403,6 +403,11 @@ matrix:
> - MAKE_CHECK_COMMAND
On 31 January 2018 at 19:03, Andres Rodriguez wrote:
> Don't want an overly large numBufferBarriers/numTextureBarriers to blow
> up the stack.
>
> v2: handle malloc errors
> v3: fix patch
>
> Suggested-by: Emil Velikov
> Signed-off-by: Andres Rodriguez
> ---
> src/mesa/main/externalobjects.c |
https://bugs.freedesktop.org/show_bug.cgi?id=104732
Józef Kucia changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
On 1 February 2018 at 11:13, Adrian M Negreanu wrote:
> There's a bug about quiting on unsupported tags:
> https://bugzilla.redhat.com/show_bug.cgi?id=1348200
>
> Should be fixed in v1.10
> http://pkgs.fedoraproject.org/cgit/rpms/dwarves.git/commit/?h=f25
>
Seems like distros should update their p
On Thursday, 2018-02-01 13:27:28 +0100, Marc Dietrich wrote:
> Only dive into the windows subdir if windows platform is selected.
>
> Signed-off-by: Marc Dietrich
Good point, although if I'm reading it right, it only installs a couple
unnecessary headers right now; we got lucky :P
Fixes: 5ef75c
For the series,
Reviewed-by: Neha Bhende
From: Brian Paul
Sent: Wednesday, January 31, 2018 7:37:23 PM
To: mesa-dev@lists.freedesktop.org
Cc: Charmaine Lee; Neha Bhende
Subject: [PATCH 4/4] svga: remove unneeded #includes in svga_pipe_draw.c
---
src/gallium/driv
Am Donnerstag, 1. Februar 2018, 15:07:55 CET schrieb Eric Engestrom:
> On Thursday, 2018-02-01 13:27:28 +0100, Marc Dietrich wrote:
> > Only dive into the windows subdir if windows platform is selected.
> >
> > Signed-off-by: Marc Dietrich
>
> Good point, although if I'm reading it right, it onl
On 2018-02-01 08:24 AM, Emil Velikov wrote:
On 31 January 2018 at 19:03, Andres Rodriguez wrote:
Don't want an overly large numBufferBarriers/numTextureBarriers to blow
up the stack.
v2: handle malloc errors
v3: fix patch
Suggested-by: Emil Velikov
Signed-off-by: Andres Rodriguez
---
sr
https://bugs.freedesktop.org/show_bug.cgi?id=104905
Bug ID: 104905
Summary: SpvOpFOrdEqual doesn't return correct results for NaNs
Product: Mesa
Version: git
Hardware: All
OS: All
Status: NEW
Severity: nor
downsize_format_if_needed takes an integer as number of uploads
parameter. Hence, let's do an integer comparation instead of a boolean
check, since that is confusing.
Since we are at it, fix a couple of wrongly tabbed indents.
Cc: Alejandro Piñeiro
Cc: Kenneth Graunke
Signed-off-by: Andres Gome
downsize_format_if_needed takes an integer as number of uploads
parameter. Hence, let's do an integer comparation instead of a boolean
check, since that is confusing.
Since we are at it, fix a couple of wrongly tabbed indents.
Cc: Alejandro Piñeiro
Cc: Kenneth Graunke
Signed-off-by: Andres Gome
The old one generates useless instructions in there, found while
comparing geometry shaders between RadeonSI and RADV.
This improves all Vulkan demos that use geometry shaders, +4%
for deferredshadows, +9% for viewportarray, +7% for
geometryshader on Polaris10.
This seems to also improve DOW3 a l
On Thursday, 2018-02-01 15:29:25 +0100, Marc Dietrich wrote:
> Am Donnerstag, 1. Februar 2018, 15:07:55 CET schrieb Eric Engestrom:
> > On Thursday, 2018-02-01 13:27:28 +0100, Marc Dietrich wrote:
> > > Only dive into the windows subdir if windows platform is selected.
> > >
> > > Signed-off-by: M
On 02/01/2018 03:49 AM, Ilia Mirkin wrote:
On Wed, Jan 31, 2018 at 10:39 PM, Brian Paul wrote:
And rename indirect_params -> indirect_draw_count_buffer and
indirect_params_offset -> indirect_draw_count_offset to be more
specific.
---
src/mesa/vbo/vbo.h | 57 +++
On Feb 1st all xorg members were expired as part of the regular
process to remove inactive members. If you would still like to be a
member of the X.Org Foundation, please renew your membership. To
renew or to become a first time member, go to https://members.x.org/ .
For renewals, log in and clic
On Thursday, February 1, 2018 7:15:14 AM PST Andres Gomez wrote:
> downsize_format_if_needed takes an integer as number of uploads
> parameter. Hence, let's do an integer comparation instead of a boolean
> check, since that is confusing.
>
> Since we are at it, fix a couple of wrongly tabbed inden
Am 01.02.2018 um 09:21 schrieb Dave Airlie:
> From: Dave Airlie
>
> This passes the CTS and piglit tests.
>
> This also disable sb for helper invocations until it doesn't
> mess up the VPM flags.
>
> Thanks to Ilia and Glenn for advice, and Roland for working
> out the working evergreen path.
>
I'm surprised it does not do a idxen and swizzling, but with the current
code, this is
Reviewed-by: Bas Nieuwenhuizen
On Thu, Feb 1, 2018 at 4:37 PM, Samuel Pitoiset
wrote:
> The old one generates useless instructions in there, found while
> comparing geometry shaders between RadeonSI and RAD
On 02/01/2018 12:32 AM, mathias.froehl...@gmx.net wrote:
From: Mathias Fröhlich
Hi,
Thanks for the review!
This is the starting point to a patch series that I intent to feed.
The aim is to get rid of some VERT_ATTRIB_MAX long loops that currently
happen at about any draw call.
The series tri
Reviewed-by: Marek Olšák
Marek
On Thu, Feb 1, 2018 at 4:57 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> If the lowering flags for int64 were set we'd try and lower
> the wrong things sometimes.
>
> Signed-off-by: Dave Airlie
> ---
> src/compiler/glsl/lower_64bit.cpp | 12
> 1
https://bugs.freedesktop.org/show_bug.cgi?id=104908
Bug ID: 104908
Summary: Texture Compression Hint not converted to enum16
Product: Mesa
Version: git
Hardware: All
OS: All
Status: NEW
Severity: minor
Move generated files from codegen/meson.build to other directories, in order
to satisfy generated include file dependencies
Add correct file lists for architecture-specific libraries.
cc: mesa-sta...@lists.freedesktop.org
cc: dy...@pnwbakers.com
---
src/gallium/drivers/swr/meson.build
cc: mesa-sta...@lists.freedesktop.org
cc: dy...@pnwbakers.com
---
src/gallium/drivers/swr/meson.build | 2 +-
src/gallium/drivers/swr/rasterizer/codegen/meson.build | 2 +-
src/gallium/drivers/swr/rasterizer/core/backends/meson.build | 2 +-
src/gallium/drivers/swr/r
With the comments on patch 3 addressed, the series is:
Reviewed-by: Marek Olšák
Marek
On Thu, Feb 1, 2018 at 4:39 AM, Brian Paul wrote:
> In vbo_draw_indirect_prims() pass the 'indirect_data' argument to
> vbo->draw_prims(). All the callers are passing ctx->DrawIndirectBuffer
> so this should
Quoting Eric Engestrom (2018-01-31 03:47:25)
> The big comment is taken from the equivalent block in configure.ac
>
> Signed-off-by: Eric Engestrom
> ---
> meson.build | 30
> +
> src/gallium/targets/d3dadapter9/meson.build | 2 +-
>
Thanks!
Quoting Eric Engestrom (2018-01-31 03:14:50)
> On Thursday, 2018-01-25 16:14:45 -0800, Dylan Baker wrote:
> > Signed-off-by: Dylan Baker
>
> Reviewed-by: Eric Engestrom
>
> > ---
> >
> > I have tested building every mesa driver against this (with and without
> > udev!)
> > so I'm pre
Quoting Emil Velikov (2018-01-31 03:42:00)
> On 30 January 2018 at 21:31, Dylan Baker wrote:
> > Quoting Emil Velikov (2018-01-30 10:43:06)
> >> On 29 January 2018 at 18:57, Dylan Baker wrote:
> >> > Quoting Eric Engestrom (2018-01-29 10:15:50)
> >> >> The big comment is taken from the equivalent
I suppose this is a bit inefficient if the hw actually supports some
double ops?
I mean for uint->double you could just convert the low 8 bits and high
24 bits (or different split) separately to float, convert both to double
and add them. (I think should work for int->double exactly the same
despit
On 31 January 2018 at 17:11, Dylan Baker wrote:
> Quoting Emil Velikov (2018-01-31 03:42:00)
>> On 30 January 2018 at 21:31, Dylan Baker wrote:
>> > Quoting Emil Velikov (2018-01-30 10:43:06)
>> >> On 29 January 2018 at 18:57, Dylan Baker wrote:
>> >> > Quoting Eric Engestrom (2018-01-29 10:15:5
On Thu, 2018-02-01 at 12:45 +0100, Juan A. Suarez Romero wrote:
> On Mon, 2018-01-29 at 16:25 -0800, Eric Anholt wrote:
> > "Juan A. Suarez Romero" writes:
> >
> > > Enable these kind of 3D texture when proper extensions are available.
> > >
> > > Fixes KHR-GLES2.texture_3d.* with these textures
An additional stub for applegl_create_context() is needed
Cannot test indirect API as it's not built on osx, currently
Signed-off-by: Jon Turney
---
src/glx/tests/fake_glx_screen.cpp | 11 +++
src/glx/tests/indirect_api.cpp| 4
2 files changed, 15 insertions(+)
diff --git a/sr
Jon Turney (4):
glx/test: fix building for osx
travis: conditionalize building of prerequisites on if OS=linux
travis: pip -> pip2
travis: add osx autotools build
.travis.yml | 123 --
src/glx/tests/fake_glx_screen.cpp | 11
Use a '|' YAML literal block to avoid the convoluted syntax needed to put
the entire conditional on a single line.
Signed-off-by: Jon Turney
---
.travis.yml | 91 +++--
1 file changed, 47 insertions(+), 44 deletions(-)
diff --git a/.travis
On travis, for OSX, python2 from homebrew is pre-installed. per [1]:
python points to the macOS system Python (with no manual PATH modification)
python2 points to Homebrew’s Python 2.7.x (if installed)
python3 points to Homebrew’s Python 3.x (if installed)
pip doesn't exist
pip2 points to Hom
Signed-off-by: Jon Turney
---
.travis.yml | 30 ++
1 file changed, 30 insertions(+)
diff --git a/.travis.yml b/.travis.yml
index e18327bec9..fc137c9667 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -396,6 +396,36 @@ matrix:
- libexpat1-dev
-
On 01/02/2018 11:49, Emil Velikov wrote:
On 28 January 2018 at 14:24, Jon Turney wrote:
From: Dylan Baker
---
meson.build | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/meson.build b/meson.build
index 824e3c46bc5..178743eddb1 100644
--- a/meson.build
+++ b/meson
https://bugs.freedesktop.org/show_bug.cgi?id=104908
--- Comment #1 from Brian Paul ---
The patch looks good to me. I'll push it in a bit.
I don't see any other issues related to Hint.TextureCompression.
Unfortunately, there's no Piglit test that hits this.
--
You are receiving this mail becau
Hi Brian,
with the updated comment that Ilia found, also
Reviewed-by: Mathias Fröhlich
best
Mathias
> ---
> src/mesa/vbo/vbo_save_draw.c | 14 +++---
> 1 file changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/src/mesa/vbo/vbo_save_draw.c b/src/mesa/vbo/vbo_save_draw.c
> inde
Hi Brian,
On Thursday, 1 February 2018 17:17:57 CET Brian Paul wrote:
> Looks good.
>
> Reviewed-by: Brian Paul
Thanks!!
> I don't remember, do you need me to push these for you?
I used to have an account for mesa. I have not used that for some time, but I
assume that it is still functional.
Am Mittwoch, den 31.01.2018, 11:37 + schrieb Daniel Stone:
> On 30 January 2018 at 14:22, Lucas Stach
> wrote:
> > Take into account the resource format, instead of applying a
> > hardcoded
> > 32bpp. This not only over-allocates 16bpp formats, but also results
> > in
> > a wrong stride being
https://bugs.freedesktop.org/show_bug.cgi?id=104908
Brian Paul changed:
What|Removed |Added
Resolution|--- |FIXED
Status|NEW
With the below nits fixed, for the series:
Reviewed-by: Dylan Baker
Quoting George Kyriazis (2018-02-01 08:54:14)
> Move generated files from codegen/meson.build to other directories, in order
> to satisfy generated include file dependencies
>
> Add correct file lists for architecture-specific l
Quoting Emil Velikov (2018-02-01 03:49:12)
> On 28 January 2018 at 14:24, Jon Turney wrote:
> > From: Dylan Baker
> >
> > ---
> > meson.build | 8 +++-
> > 1 file changed, 7 insertions(+), 1 deletion(-)
> >
> > diff --git a/meson.build b/meson.build
> > index 824e3c46bc5..178743eddb1 100644
Emil had requested some changes, and it's on my list to get these respun
today and back on the list. I'll CC you on them.
Thanks for reminding me :)
Dylan
Quoting Marc Dietrich (2018-02-01 02:44:25)
> Am Mittwoch, 17. Januar 2018, 18:41:00 CET schrieb Dylan Baker:
> > I dug into what autotools i
There are changes from Makefile.sources that need be ported to corresponding
meson.build files as well.
It looks like builtin_float64.h is generated from float64.glsl, is that right?
I'm not very happy about checking in a generated file for all of the reasons
that we don't check generated files in
On Thu, Feb 01, 2018 at 10:00:12AM -0800, Dylan Baker wrote:
> There are changes from Makefile.sources that need be ported to corresponding
> meson.build files as well.
Oh yes, nice spot. I will do it on my local fp64 branch too.
>
> It looks like builtin_float64.h is generated from float64.glsl,
Once a lp_build_sampler_soa or lp_build_sampler_aos object is created,
it should never be modified. Found by inspection.
---
src/gallium/auxiliary/draw/draw_llvm.c | 2 +-
src/gallium/auxiliary/gallivm/lp_bld_tgsi.h | 8
src/gallium/auxiliary/gallivm/lp_bld_tgsi_aos.c | 2 +-
Makes sense.
Reviewed-by: Roland Scheidegger
Am 01.02.2018 um 21:18 schrieb Brian Paul:
> Once a lp_build_sampler_soa or lp_build_sampler_aos object is created,
> it should never be modified. Found by inspection.
> ---
> src/gallium/auxiliary/draw/draw_llvm.c | 2 +-
> src/gallium/auxi
https://bugs.freedesktop.org/show_bug.cgi?id=104374
--- Comment #2 from João Henrique ---
This is the same issue as https://bugs.freedesktop.org/show_bug.cgi?id=104902
and it is reproducible on non mesa_mild branches (such as Padoka PPA)
Included in that bug report is a piece of code to reproduc
Fixes fbo-generatemipmap-formats, fbo-alphatest-formats, etc. tests for
GL_RGBA4, GL_RGB4, GL_RGBA2, etc.
---
src/gallium/drivers/vc5/v3dx_format_table.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/src/gallium/drivers/vc5/v3dx_format_table.c
b/src/gallium/drivers/vc5/v3d
swapBytes operates on bytes, not 4-bit channels, so you can't just take
non-swapBytes cases and flip the REV flag.
Avoids piglit texture-packed-formats regressions when enabling the
ABGR format.
Fixes: c5a5c9a7db89 ("mesa/formats: add new mesa formats and their pack/unpack
functions.")
---
The VC5 HW puts A in the low bits and R in the high bits. We can't just
swizzle in the shaders because the blending HW can't pick what channel A
is in, so make a new format to match it.
---
src/gallium/auxiliary/util/u_format.csv | 1 +
src/gallium/drivers/svga/svga_format.c | 1 +
src/gallium/i
On Tue, Jan 30, 2018 at 12:07 AM, wrote:
> From: Roland Scheidegger
>
> By the looks of it it seems hemlock is treated separately to cypress, but
> certainly it won't need the stack workarounds cedar/redwood (and
> seemingly every other eg chip except cypress/juniper) need.
> (Discovered by acci
Quoting Elie Tournier (2018-02-01 12:13:53)
> On Thu, Feb 01, 2018 at 10:00:12AM -0800, Dylan Baker wrote:
> > There are changes from Makefile.sources that need be ported to corresponding
> > meson.build files as well.
> Oh yes, nice spot. I will do it on my local fp64 branch too.
> >
> > It looks
This was added in 425dc4c4b366 but never used. Also since
100796c15c3a native has superseded llvm.
---
src/gallium/include/pipe/p_state.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/gallium/include/pipe/p_state.h
b/src/gallium/include/pipe/p_state.h
index 2b56d60b5e..640e6ed26d 100644
Patch 1 & 2:
Reviewed-by: Marek Olšák
Marek
On Thu, Feb 1, 2018 at 8:20 PM, Eric Anholt wrote:
> The VC5 HW puts A in the low bits and R in the high bits. We can't just
> swizzle in the shaders because the blending HW can't pick what channel A
> is in, so make a new format to match it.
> ---
On 2 February 2018 at 02:05, Roland Scheidegger wrote:
> Am 01.02.2018 um 09:21 schrieb Dave Airlie:
>> From: Dave Airlie
>>
>> This passes the CTS and piglit tests.
>>
>> This also disable sb for helper invocations until it doesn't
>> mess up the VPM flags.
>>
>> Thanks to Ilia and Glenn for adv
On 2 February 2018 at 09:23, Dave Airlie wrote:
> On 2 February 2018 at 02:05, Roland Scheidegger wrote:
>> Am 01.02.2018 um 09:21 schrieb Dave Airlie:
>>> From: Dave Airlie
>>>
>>> This passes the CTS and piglit tests.
>>>
>>> This also disable sb for helper invocations until it doesn't
>>> mes
Am 02.02.2018 um 00:23 schrieb Dave Airlie:
> On 2 February 2018 at 09:23, Dave Airlie wrote:
>> On 2 February 2018 at 02:05, Roland Scheidegger wrote:
>>> Am 01.02.2018 um 09:21 schrieb Dave Airlie:
From: Dave Airlie
This passes the CTS and piglit tests.
This also disab
>>
>> It might be for this test, but I don't think it is in general. We want
>> boolean
>> which I think is 0 or 0x, not 0.0 or 1.0.
>
> The initial alu reg write was 0x, and the resinfo gives you back
> a fixed 0, so...
Doh I just noticed it now, my brain had a copy of the older
Am 30.01.2018 um 09:21 schrieb Gert Wollny:
> Although gl_SampleMaskIn is declared as an array in GLSL, it is
> effectively a 32 bit mask on all hardware supported by mesa, so the
> array indexing is ignored (Thanks Glenn Kennard for the explanation).
>
> Add a comment that the assert is not made
That's not much "indirect" the way it's done, but looks alright to me ;-).
Reviewed-by: Roland Scheidegger
Am 30.01.2018 um 03:22 schrieb Dave Airlie:
> From: Dave Airlie
>
> I think the cp packets can be made work, but I think it might
> need a kernel change, so for now just do the worst thi
---
src/intel/vulkan/anv_image.c | 14 +-
src/intel/vulkan/anv_private.h | 9 -
src/intel/vulkan/genX_cmd_buffer.c | 34 --
3 files changed, 33 insertions(+), 24 deletions(-)
diff --git a/src/intel/vulkan/anv_image.c b/src/intel/vulka
Reviewed-by: Ian Romanick
I also ran it through our CI.
On 01/31/2018 11:54 AM, Marek Olšák wrote:
> From: Marek Olšák
>
> This only affects drivers that set DriverFlags.NewBlend.
>
> v2: - fix typo advanded -> advanced
> - return "enum gl_advanced_blend_mode" from
> _mesa_get_advan
---
src/amd/common/ac_nir_to_llvm.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 570c798d4e..b2e72e109e 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -6758,8
---
src/gallium/drivers/radeonsi/si_shader_nir.c | 30 +++-
1 file changed, 29 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c
b/src/gallium/drivers/radeonsi/si_shader_nir.c
index 128be585cd..28b031d935 100644
--- a/src/gallium/drive
---
src/amd/common/ac_nir_to_llvm.c | 3 +++
src/amd/common/ac_shader_abi.h | 2 ++
src/gallium/drivers/radeonsi/si_shader.c | 1 +
3 files changed, 6 insertions(+)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index b2e72e109e..4c346119c3 10064
---
src/amd/common/ac_nir_to_llvm.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 4c346119c3..bf36c7ef84 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -4448,6 +4448,9 @@ static
Confiliting values returned by PIPE_SHADER_CAP_PREFERRED_IR meant
that compute for NIR was always disable before this series (see
patch 11).
This series should contain everything except shared
variable/atomic support, I intend to tackle those in a follow up
series.
Please review.
_
---
src/amd/common/ac_nir_to_llvm.c | 5 ++---
src/amd/common/ac_shader_abi.h| 2 ++
src/gallium/drivers/radeonsi/si_shader.c | 4 ++--
src/gallium/drivers/radeonsi/si_shader_internal.h | 1 -
4 files changed, 6 insertions(+), 6 deletions(-)
diff --g
This will be reused by the nir backend in a later patch.
---
src/gallium/drivers/radeonsi/si_shader.c | 47 ++--
1 file changed, 27 insertions(+), 20 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index 74
---
src/gallium/drivers/radeonsi/si_compute.c | 44 ++-
src/gallium/drivers/radeonsi/si_compute.h | 5 +++-
src/gallium/drivers/radeonsi/si_get.c | 8 +++---
3 files changed, 39 insertions(+), 18 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_compute.c
---
src/amd/common/ac_nir_to_llvm.c | 5 ++---
src/amd/common/ac_shader_abi.h| 1 +
src/gallium/drivers/radeonsi/si_shader.c | 4 ++--
src/gallium/drivers/radeonsi/si_shader_internal.h | 1 -
4 files changed, 5 insertions(+), 6 deletions(-)
diff --gi
---
src/amd/common/ac_nir_to_llvm.c | 9 -
src/amd/common/ac_shader_abi.h| 1 +
src/gallium/drivers/radeonsi/si_shader.c | 9 -
src/gallium/drivers/radeonsi/si_shader_internal.h | 1 -
4 files changed, 9 insertions(+), 11 deletions(-)
When PIPE_SHADER_IR_LLVM existed this query made sense but now it
always returns PIPE_SHADER_IR_NATIVE. Also it is now conlicting
with PIPE_SHADER_IR_NIR for compute shaders, so just assume this
is always PIPE_SHADER_IR_NATIVE for clover.
This change indirectly enables NIR support for compute shad
---
src/amd/common/ac_llvm_build.c | 8
src/amd/common/ac_llvm_build.h | 2 ++
src/gallium/drivers/radeonsi/si_shader.c | 6 +-
3 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/src/amd/common/ac_llvm_build.c b/src/amd/common/ac_llvm_build.c
index 6af
Seems like the proper solution to this problem is to use something
like ST_DEBUG=nir (which would in turn check the SUPPORTED_IRS bitmap)
rather than trying to hack drivers...
On Thu, Feb 1, 2018 at 11:03 PM, Timothy Arceri wrote:
> When PIPE_SHADER_IR_LLVM existed this query made sense but now i
From: Dave Airlie
Doing a straight uint/int->fp32->fp64 conversion causes
some precision issues, Roland suggested splitting the
integer into two portions and doing two separate
int->fp32->fp64 conversions then adding the results.
This passes the tests in CTS and piglit.
Signed-off-by: Dave Airl
On 02/02/18 15:12, Ilia Mirkin wrote:
Seems like the proper solution to this problem is to use something
like ST_DEBUG=nir (which would in turn check the SUPPORTED_IRS bitmap)
rather than trying to hack drivers...
That sounds like more of a hack to me. What happend when nir becomes the
default
On 2 February 2018 at 03:40, Roland Scheidegger wrote:
> I suppose this is a bit inefficient if the hw actually supports some
> double ops?
> I mean for uint->double you could just convert the low 8 bits and high
> 24 bits (or different split) separately to float, convert both to double
> and add
From: Dave Airlie
This fixes a few CTS cases in :
KHR-GL45.texture_view.view_sampling
some multisample cases are still broken, but not sure this is
the same problem.
v2: fix more cases
Signed-off-by: Dave Airlie
---
src/gallium/drivers/r600/evergreen_state.c | 17 ++---
1 file ch
On Fri, 2018-02-02 at 15:03 +1100, Timothy Arceri wrote:
> When PIPE_SHADER_IR_LLVM existed this query made sense but now it
> always returns PIPE_SHADER_IR_NATIVE. Also it is now conlicting
> with PIPE_SHADER_IR_NIR for compute shaders, so just assume this
> is always PIPE_SHADER_IR_NATIVE for clo
We move the nir check before the shader cache call so that we can
call a nir based caching function in a following patch.
Also with this change we simply check if vertex shaders support
NIR rather than looping over the stages as mixing of shader types
is not supported anyway.
---
src/mesa/state_t
Reviewed-by: Marek Olšák
---
src/gallium/drivers/radeonsi/si_pipe.c | 4
src/mesa/state_tracker/st_program.c| 15 +++
2 files changed, 11 insertions(+), 8 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_pipe.c
b/src/gallium/drivers/radeonsi/si_pipe.c
index b44c2f
v2: include compute shader support
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 2 +-
src/mesa/state_tracker/st_program.c| 6 +-
src/mesa/state_tracker/st_shader_cache.c | 163 ++---
src/mesa/state_tracker/st_shader_cache.h | 8 +-
4 files changed, 135
1 - 100 of 107 matches
Mail list logo
