From: Nicolai Hähnle
This is in preparation for the new image intrinsics.
---
src/amd/common/ac_llvm_build.c | 101 +-
src/amd/common/ac_llvm_build.h | 14 +-
src/amd/common/ac_nir_to_llvm.c| 365 +++--
src/gallium
From: Nicolai Hähnle
In preparation of dimension-aware LLVM image intrinsics.
---
src/amd/common/ac_llvm_build.c | 105 +++---
src/amd/common/ac_llvm_build.h | 37 +++-
src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.h | 2 +-
src/gallium/drivers/
Hi all,
one of the things I've been recently working on is a new set of image
intrinsics in LLVM to have explicit parameters for all the different
components of the address argument.
This is motivated by the A16 feature in gfx9, which allows address
components (coordinates, explicit derivatives)
From: Nicolai Hähnle
This is in preparation for the new, dimension-aware LLVM image
intrinsics.
---
src/amd/common/ac_llvm_build.c| 10 -
src/amd/common/ac_llvm_build.h| 13 +-
src/amd/common/ac_nir_to_llvm.c | 54
From: Nicolai Hähnle
It looks as if the structure fields array is fully initialized below,
but in fact at least gcc in debug builds will not actually overwrite
the unused bits of bit fields.
---
src/compiler/glsl_types.cpp | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git
From: Nicolai Hähnle
The LLVM instruction returns { i32, i1 }, where the i1 indicates success.
We're only interested in the first part, which is the loaded value.
Fixes dEQP-GLES31.functional.compute.shared_var.atomic.compswap.*
---
src/amd/common/ac_nir_to_llvm.c | 1 +
1 file chang
From: Nicolai Hähnle
trans is zero-initialized, but trans->resource is setup immediately so
needs to be dereferenced.
---
src/gallium/drivers/radeonsi/si_texture.c | 25 -
1 file changed, 12 insertions(+), 13 deletions(-)
diff --git a/src/gallium/drivers/radeo
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader_nir.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c
b/src/gallium/drivers/radeonsi/si_shader_nir.c
index c0e08c79a56..b4fba8b8812 100644
--- a/src/gallium
From: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index 8c62d53e2ad..3e224b083e6 100644
--- a/src/gallium/drivers/radeonsi
Hi all,
just a collection of random fixes I've accumulated over time, most
of them stuff that I ran into while testing out NIR stuff.
Please review!
Thanks,
Nicolai
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org
Nice, did you actually get it to work entirely on a big endian machine?
Bit fields aren't super portable, but this looks good enough. However, I
think we should use the PIPE_ARCH_LITTLE_ENDIAN define from u_endian.h
Cheers,
Nicolai
On 20.03.2018 15:21, Bas Vermeulen wrote:
Using mesa OpenCL
Nothing wrong with them, really. All the LLVM stuff has caused me to
fall behind on going through Mesa patches, sorry for that. 9-13 are finally
Reviewed-by: Nicolai Hähnle
On 25.02.2018 02:04, Marek Olšák wrote:
So what is wrong with patches 9-13?
We can do cleanups after those.
Marek
On
With a small comment on patch 6, patches 1-8:
Reviewed-by: Nicolai Hähnle
for now.
However, I'm unhappy about how complex this is all getting. 32- vs.
64-bit, merged vs. non-merged, monolithic vs. non-monolithic, and then
special user SGPR uses like for blits and soon VBO descriptor
On 17.02.2018 20:43, Marek Olšák wrote:
From: Marek Olšák
TCS_OUT_LAYOUT has 13 unused bits. That's enough for a 32-bit address
aligned to 512KB. Hey, it's a 13-bit pointer!
Heh :)
---
src/gallium/drivers/radeonsi/si_shader.c | 104 +++---
src/gallium/drivers/ra
Why? 3D images are not arrays.
On 20.02.2018 11:11, Samuel Pitoiset wrote:
This doesn't fix anything known but it should definitely be set.
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_nir_to_llvm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/amd/common
Patches 1 & 2:
Reviewed-by: Nicolai Hähnle
On 31.01.2018 00:41, Dave Airlie wrote:
From: Dave Airlie
This reduces the size from 28->24 bytes.
Signed-off-by: Dave Airlie
---
src/amd/common/ac_shader_info.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/
Series is:
Reviewed-by: Nicolai Hähnle
On 27.01.2018 19:01, Marek Olšák wrote:
From: Marek Olšák
it's useless and shader-db stats only report the main shader part.
---
src/gallium/drivers/radeon/r600_pipe_common.h | 1 -
src/gallium/drivers/radeonsi/si_pipe.c | 1 -
From: Nicolai Hähnle
LLVM now emits labels as part of the disassembly string, which is very
useful but breaks the old parsing approach.
Use the semicolon to detect the boundary of instructions instead of going
by line breaks.
---
src/gallium/drivers/radeonsi/si_debug.c | 63
From: Nicolai Hähnle
This will give us meaningful wave information in the case of a hang where
shaders are still running in an infinite loop.
Note that we call umr multiple times for different sections of the ddebug
hang dump, and so the wave information will not necessarily match up
between
-by: Nicolai Hähnle
---
src/amd/common/ac_nir_to_llvm.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 4f240db4ac..e88f64fb3b 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common
On 19.01.2018 06:50, Timothy Arceri wrote:
This fixes a segfault for varyings not starting at component 0.
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_llvm_build.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/common/ac_llvm_build.c b/src/amd/common
On 20.01.2018 02:38, Timothy Arceri wrote:
Fixes the following piglit test on radeonsi:
./bin/arb_enhanced_layouts-gs-stream-location-aliasing
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_nir_to_llvm.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src
arb_shader_image_load_store/max-size/image3d max size test/8x8x2048x1
arb_shader_image_load_store/semantics/imageload/vertex shader/rgba32f/image3d
test
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_nir_to_llvm.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/amd/common
On 19.01.2018 07:05, Tapani Pälli wrote:
On 01/18/2018 05:49 PM, Nicolai Hähnle wrote:
On 17.01.2018 17:10, Tapani Pälli wrote:
On 17.01.2018 13:34, Nicolai Hähnle wrote:
On 15.01.2018 13:31, Tapani Pälli wrote:
Hello;
Here's a refactored series of EGL_ANDROID_blob_cache. Now
-no-gs-strip-first-vertex
Reviewed-by: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader_nir.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c
b/src/gallium/drivers/radeonsi/si_shader_nir.c
index 4455322392..2f19bae813 100644
--- a
On 19.01.2018 12:03, Timothy Arceri wrote:
Fixes a bunch of ARB_sample_shading piglit tests.
Reviewed-by: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader_nir.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c
b/src/gallium
On 18.01.2018 08:17, Timothy Arceri wrote:
This is what radeonsi does. Fixes remaing piglit subtest in:
./bin/arb_shader_image_size-builtin --quick -auto -fbo
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_nir_to_llvm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff
Both patches:
Reviewed-by: Nicolai Hähnle
On 19.01.2018 05:12, Timothy Arceri wrote:
This is needed for ARB_bindless_texture support.
---
src/compiler/glsl/glsl_to_nir.cpp | 4
1 file changed, 4 insertions(+)
diff --git a/src/compiler/glsl/glsl_to_nir.cpp
b/src/compiler/glsl
On 10.01.2018 20:49, Marek Olšák wrote:
From: Marek Olšák
Please add a bug report or reference to what fails without it.
It does make sense though, so
Reviewed-by: Nicolai Hähnle
Cc: 17.2 17.3
---
src/gallium/drivers/radeon/r600_texture.c | 14 --
1 file changed, 12
. Can you reduce it to
something like 64MB? Unless there's concrete evidence that having it
higher is beneficial, of course.
With that, patches 9 & 10:
Reviewed-by: Nicolai Hähnle
+
typedef void (*tc_replace_buffer_storage_func)(struct pipe_context *ctx,
For the series:
Reviewed-by: Nicolai Hähnle
On 10.01.2018 20:12, Samuel Pitoiset wrote:
This allows to reduce the number of dwords that are loaded
with buffer_load_format_xyzw. For example, when the only used
channel is 1, the driver will emit buffer_load_format_x instead.
Shader stats for
d you please add a comment /* Adjust for dual source blending */ or
similar?
With that:
Reviewed-by: Nicolai Hähnle
+ }
} else {
tgsi_get_gl_varying_semantic(variable->data.location,
true,
rinsic == nir_intrinsic_read_invocation) {
+ num_args = 2;
+ intr_name = intr_names[0];
Please just inline the intrinsic names here.
With that fixed:
Reviewed-by: Nicolai Hähnle
+
+ /* Invocation */
+ args[1] = get_src(ctx,
Reviewed-by: Nicolai Hähnle
On 22.01.2018 00:52, Timothy Arceri wrote:
---
src/amd/common/ac_nir_to_llvm.c | 15 +++
1 file changed, 15 insertions(+)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index ac3a949bd3..eabdc83f16 100644
--- a/src
Reviewed-by: Nicolai Hähnle
On 19.01.2018 00:05, Timothy Arceri wrote:
---
src/amd/common/ac_nir_to_llvm.c | 14 ++
1 file changed, 14 insertions(+)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index cf0b3d998c..4f240db4ac 100644
--- a/src/amd
Why hasn't this been noticed before? Does Vulkan not have multiple streams?
Anyway, series is:
Reviewed-by: Nicolai Hähnle
On 18.01.2018 23:26, Timothy Arceri wrote:
---
src/amd/common/ac_nir_to_llvm.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/sr
shot again after
updating my systems.
Kind Regards,
Jean Hertel
De: Jean Hertel
Enviado: domingo, 7 de janeiro de 2018 20:10
Para: Michel Dänzer; Nicolai Hähnle
Cc: mesa-dev@lists.freedesktop.org
Assunto: Re: [Mesa-dev] [GSOC] DriConf Replacement
Hello Michael,
Finally I found some free
Reviewed-by: Nicolai Hähnle
On 18.01.2018 02:03, Timothy Arceri wrote:
Fixes: f4e499ec7914 "radv: add initial non-conformant radv vulkan driver"
---
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_ll
;< multiplier) - 1u) << (i *
multiplier);
I think using util_bitscan for the loop would be cleaner.
Apart from that, both patches:
Reviewed-by: Nicolai Hähnle
+ return new_mask;
+}
+
static void visit_store_ssbo(struct ac_nir_context *ctx,
On 11.01.2018 12:22, Ilia Mirkin wrote:
On Thu, Jan 11, 2018 at 3:51 AM, Gert Wollny wrote:
Am Donnerstag, den 11.01.2018, 08:09 +0100 schrieb Gert Wollny:
Am Mittwoch, den 10.01.2018, 15:27 -0500 schrieb Ilia Mirkin:
[...]
If your hardware executes all the vertices in parallel, then a
barri
On 17.01.2018 17:10, Tapani Pälli wrote:
On 17.01.2018 13:34, Nicolai Hähnle wrote:
On 15.01.2018 13:31, Tapani Pälli wrote:
Hello;
Here's a refactored series of EGL_ANDROID_blob_cache. Now cache
functions are stored in disk_cache struct and the functionality
called within exi
On 18.01.2018 03:52, Roland Scheidegger wrote:
Am 17.01.2018 um 13:54 schrieb Nicolai Hähnle:
On 11.01.2018 23:54, srol...@vmware.com wrote:
From: Roland Scheidegger
The command parser is very sad if we don't emit the relocs per hw
query...
However, don't enable it. It mostly
On 04.01.2018 18:37, Andres Rodriguez wrote:
On 2018-01-04 12:33 PM, Marek Olšák wrote:
Is the renaming necessary? It looks like everything would be fine if
we used the "fence" name for semaphores.
The rename was requested by nha. We could keep going with the fences
name. Or we could do t
On 11.01.2018 23:54, srol...@vmware.com wrote:
From: Roland Scheidegger
The command parser is very sad if we don't emit the relocs per hw query...
However, don't enable it. It mostly works, but piglit
arb_transform_feedback_overflow_query-basic shows 2 failures (it's really the
same case for t
Cool stuff :)
Except for a remark on patch #2, the series is:
Reviewed-by: Nicolai Hähnle
On 12.01.2018 04:55, Timothy Arceri wrote:
We still have more work to do but piglit results are looking
pretty good.
At GLSL 1.50 we have 30647/31118 piglit tests passing.
Al GLSL 4.50 we have 37927
On 12.01.2018 03:55, Timothy Arceri wrote:
This shares mode code and calls the new shared load_tess_varyings()
s/mode/more/?
abi so that the radeonsi nir path now supports tcs output loads.
---
src/amd/common/ac_nir_to_llvm.c | 126 +---
1 file changed,
On 16.01.2018 22:18, Timothy Arceri wrote:
This resolves a game bug in Deal Island. The game doesn't properly
s/Deal/Dead/
Series is:
Reviewed-by: Nicolai Hähnle
handle ARB_get_program_binary with 0 supported formats, and ends up
crashing.
This will enable ARB_get_program_binary b
On 17.01.2018 01:23, Jordan Justen wrote:
On 2018-01-16 13:57:37, Kenneth Graunke wrote:
On Tuesday, January 16, 2018 11:18:13 AM PST Emil Velikov wrote:
Hi all,
As you've know the Mesa 18.0.0 release plan has been available for a while
on the mesa3d.org website [1].
In case you've missed it
On 17.01.2018 02:12, Ian Romanick wrote:
On 01/16/2018 10:11 AM, Ian Romanick wrote:
On 01/13/2018 01:47 PM, Jason Ekstrand wrote:
On January 12, 2018 14:56:26 "Ian Romanick" wrote:
From: Ian Romanick
With the exception of NVIDIA hardware, these are is the values that all
hardware and Gall
s really quite odd. I'd love to hear the
history behind it.
(I think it might be useful to add this information in the comment, but
I'll leave that decision up to you.)
Patches 1, 3, & the core + radeon parts of 4:
Reviewed-by: Nicolai Hähnle
+ */
+enum PACKED
For the series:
Reviewed-by: Nicolai Hähnle
On 17.01.2018 06:17, Brian Paul wrote:
The two headers already have the right extern "C" annotations.
---
src/compiler/glsl/serialize.cpp | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/src/compiler/glsl/seria
On 17.01.2018 06:18, Brian Paul wrote:
Both state->prog->info.inputs_read and state->InputsBound are GLbitfield64
so it seems that the OR of those values should be of the same type.
I'm not sure this fixes any actual issues though.
It's certainly less surprising this way.
R
On 15.01.2018 13:31, Tapani Pälli wrote:
Hello;
Here's a refactored series of EGL_ANDROID_blob_cache. Now cache
functions are stored in disk_cache struct and the functionality
called within existing disk_cache put/get code. Problems/errors
that existed with earlier series are gone.
On Android c
On 16.01.2018 18:45, Emil Velikov wrote:
Hi Tapani,
On 15 January 2018 at 12:31, Tapani Pälli wrote:
+static void
+update_blob_cache_functions(struct dri2_egl_display *dri2_dpy,
+struct dri2_egl_context *dri2_ctx)
+{
+ if (!dri2_dpy || !dri2_ctx)
+ return;
that, both patches:
Reviewed-by: Nicolai Hähnle
dvec3.
---
src/amd/common/ac_nir_to_llvm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c
b/src/amd/common/ac_nir_to_llvm.c
index c24e563695..bb4f90b3a0 100644
--- a/src/amd/common/ac_nir
), false);
}
static void visit_store_ssbo(struct ac_nir_context *ctx,
nir_intrinsic_instr *instr)
Agreed that "index" makes more sense here, as it is an index in OpenGL.
Reviewed-by: Nicolai Hähnle
--
Lerne, wie die Welt wirklich ist,
Abe
sually add the version difference as part of the commit message.
Reviewed-by: Nicolai Hähnle
You may want to wait a while though to see if somebody else speaks up.
Cheers,
Nicolai
---
src/glx/g_glxglvnddispatchfuncs.c | 4
1 file changed, 4 insertions(+)
diff --git
On 15.01.2018 15:15, Samuel Thibault wrote:
Nicolai Hähnle, on lun. 15 janv. 2018 15:07:03 +0100, wrote:
On 13.01.2018 12:36, Samuel Thibault wrote:
glXGetDriverConfig parameters do not provide a context to dynamically
check for the presence of the function, so the dispatcher directly calls
On 13.01.2018 12:36, Samuel Thibault wrote:
glXGetDriverConfig parameters do not provide a context to dynamically
check for the presence of the function, so the dispatcher directly calls
glXGetDriverConfig, but in non-dri builds dri_glx.c didn't provide
glXGetDriverConfig.
This change makes it p
On 15.12.2017 12:37, Rob Clark wrote:
On Fri, Dec 15, 2017 at 4:41 AM, Nicolai Hähnle wrote:
On 15.12.2017 00:56, Rob Clark wrote:
On Wed, Dec 6, 2017 at 3:31 PM, Ian Romanick wrote:
On 12/05/2017 08:25 AM, Ilia Mirkin wrote:
On Tue, Dec 5, 2017 at 8:18 AM, Emil Velikov
wrote:
Hi Rob
On 15.12.2017 00:56, Rob Clark wrote:
On Wed, Dec 6, 2017 at 3:31 PM, Ian Romanick wrote:
On 12/05/2017 08:25 AM, Ilia Mirkin wrote:
On Tue, Dec 5, 2017 at 8:18 AM, Emil Velikov wrote:
Hi Rob,
On 5 December 2017 at 12:54, Rob Clark wrote:
This is a bit sad/annoying. But with current GPU
On 12.12.2017 11:33, Timothy Arceri wrote:
On 12/12/17 21:21, Timothy Arceri wrote:
On 12/12/17 20:42, Nicolai Hähnle wrote:
On 11.12.2017 03:43, Timothy Arceri wrote:
---
src/amd/common/ac_nir_to_llvm.c | 66
+---
src/amd/common/ac_shader_abi.h
case of the comments on refactored/moved code) should
probably be in separate commits.
Patches 1, 2, 4-12, 14 (with a trivial change), 15-19:
Reviewed-by: Nicolai Hähnle
... with the caveat that some of the patches will probably be
(trivially) affected by the following:
Considering patch
What about reads_pervertex_outputs?
Cheers,
Nicolai
On 11.12.2017 03:43, Timothy Arceri wrote:
---
src/gallium/drivers/radeonsi/si_shader_nir.c | 29
1 file changed, 29 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c
b/src/gallium/driver
Typo in the subject: radeonsi
Cheers,
Nicolai
On 11.12.2017 03:43, Timothy Arceri wrote:
---
src/gallium/drivers/radeonsi/si_shader.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_sha
On 11.12.2017 03:43, Timothy Arceri wrote:
We will call these from the radeonsi NIR backend.
---
src/amd/common/ac_llvm_build.c | 24 +
src/amd/common/ac_llvm_build.h | 8 ++
src/amd/common/ac_nir_to_llvm.c | 58 +
3 files change
On 11.12.2017 03:43, Timothy Arceri wrote:
---
src/amd/common/ac_nir_to_llvm.c | 63 +
src/amd/common/ac_shader_abi.h | 12
2 files changed, 51 insertions(+), 24 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_ni
On 11.12.2017 03:43, Timothy Arceri wrote:
---
src/gallium/drivers/radeonsi/si_shader.c | 47
1 file changed, 47 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index ba8301b3deb..0561d2ee9b5 1006
On 11.12.2017 03:43, Timothy Arceri wrote:
---
src/amd/common/ac_nir_to_llvm.c | 66 +---
src/amd/common/ac_shader_abi.h | 12 ++
src/gallium/drivers/radeonsi/si_shader.c | 1 +
3 files changed, 57 insertions(+), 22 deletions(-)
diff --git
I have a slight preference for Jason's suggestion, but either way, both
patches:
Reviewed-by: Nicolai Hähnle
On 12.12.2017 04:47, Timothy Arceri wrote:
---
src/compiler/nir/nir_lower_io_arrays_to_elements.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
With this patch we now get:
%44 = and i32 %43, 1065353216
%45 = bitcast i32 %44 to float
%46 = fpext float %45 to double
Yay for informative commit messages :)
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_nir_to_llvm.c | 1 +
1 file changed, 1 insertion(+)
diff --git
On 11.12.2017 16:31, Marek Olšák wrote:
From: Marek Olšák
Cc: 17.3
Reviewed-by: Nicolai Hähnle
---
src/gallium/winsys/amdgpu/drm/amdgpu_bo.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/gallium/winsys/amdgpu/drm/amdgpu_bo.c
b/src/gallium/winsys/amdgpu
ruct cso_rasterizer));
if (!cso)
Sure, why not.
Reviewed-by: Nicolai Hähnle
--
Lerne, wie die Welt wirklich ist,
Aber vergiss niemals, wie sie sein sollte.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.o
On 06.12.2017 13:43, Marek Olšák wrote:
On Dec 6, 2017 12:34 PM, "Nicolai Hähnle" <mailto:nhaeh...@gmail.com>> wrote:
On 05.12.2017 20:05, Marek Olšák wrote:
From: Marek Olšák mailto:marek.ol...@amd.com>>
Cc: 17.3 mailto:mesa-sta
On 06.12.2017 14:25, Rob Clark wrote:
On Wed, Dec 6, 2017 at 2:07 AM, James Jones wrote:
Note I also posed the question of whether things like cached (and similarly
compression, since I view compression as roughly an equivalent mechanism to
a cache) in one of the open issues on my XDC 2017 slid
On 05.12.2017 20:05, Marek Olšák wrote:
From: Marek Olšák
Cc: 17.3
---
src/gallium/drivers/radeon/r600_texture.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeon/r600_texture.c
b/src/gallium/drivers/radeon/r600_texture.c
index 2aa47b5..07f
Except for the comment on patch 1, this series is:
Reviewed-by: Nicolai Hähnle
On 05.12.2017 20:05, Marek Olšák wrote:
From: Marek Olšák
and anything that clones these uploaders, like u_threaded_context.
---
src/gallium/drivers/radeon/r600_pipe_common.c | 7 +--
1 file changed, 5
On 06.12.2017 08:07, James Jones wrote:
[snip]
So lets say you have a setup where both display and GPU supported
FOO/tiled, but only GPU supported compressed (FOO/CC) and cached
(FOO/cached). But the GPU supported the following transitions:
trans_a: FOO/CC -> null
trans_b: FOO/cached ->
On 06.12.2017 08:01, James Jones wrote:
On 12/01/2017 10:34 AM, Nicolai Hähnle wrote:
On 01.12.2017 18:09, Nicolai Hähnle wrote:
[snip]
As for the actual transition API, I accept that some metadata may be
required, and the metadata probably needs to depend on the memory
layout,
which is
Signed-off-by: Adam Jackson
Makes sense.
Reviewed-by: Nicolai Hähnle
---
src/glx/glxextensions.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/glx/glxextensions.c b/src/glx/glxextensions.c
index af6ffbf660..638d8bcbbe 100644
--- a/src/glx/glxextensions.c
+++
S on cayman, it appears to work.
Signed-off-by: Dave Airlie
I can't say much about the assembly, but the CP packets look reasonable
to me.
Acked-by: Nicolai Hähnle
---
src/gallium/drivers/r600/evergreen_state.c | 57 +-
src/gallium/drivers/r600/r6
Patches 1 and 3-7:
Reviewed-by: Nicolai Hähnle
On 01.12.2017 21:19, Marek Olšák wrote:
From: Marek Olšák
---
src/gallium/auxiliary/util/u_upload_mgr.c| 9 ++---
src/gallium/auxiliary/util/u_upload_mgr.h| 2 +-
src/gallium/drivers/freedreno/a3xx/fd3_context.c | 2
On 01.12.2017 21:19, Marek Olšák wrote:
From: Marek Olšák
and anything that clones these uploaders, like u_threaded_context.
Bottom/top-of-pipe fences will have to be moved away from using the
stream_uploader.
---
src/gallium/drivers/radeon/r600_pipe_common.c | 6 --
1 file change
On 01.12.2017 21:19, Marek Olšák wrote:
From: Marek Olšák
Cc: 17.3
What's the use-case for this?
What if somebody exports as DMABUF, then re-imports in a different API
and exports as FD for inter-process sharing from there?
---
src/gallium/drivers/radeon/r600_texture.c | 4 +++-
1 f
Reviewed-by: Nicolai Hähnle
On 04.12.2017 05:17, Timothy Arceri wrote:
---
src/gallium/drivers/radeonsi/si_shader.c | 15 +++
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
On 01.12.2017 18:09, Nicolai Hähnle wrote:
[snip]
As for the actual transition API, I accept that some metadata may be
required, and the metadata probably needs to depend on the memory
layout,
which is often vendor-specific. But even linear layouts need some
transitions for caches. We probably
On 01.12.2017 16:06, Rob Clark wrote:
On Thu, Nov 30, 2017 at 5:43 PM, Nicolai Hähnle wrote:
Hi,
I've had a chance to look a bit more closely at the allocator prototype
repository now. There's a whole bunch of low-level API design feedback, but
for now let's focus on the h
On 01.12.2017 08:20, Samuel Iglesias Gonsálvez wrote:
On Thu, 2017-11-30 at 15:47 +0100, Nicolai Hähnle wrote:
Can you add an explanation / spec quote for this?
Yes.
"We validate that the interface block array type's definition matches.
However, the function could be previously ca
On 01.12.2017 06:06, Dave Airlie wrote:
From: Dave Airlie
On Cayman we don't use the append/consume counters (fglrx doesn't)
and they don't seem to work well with compute shaders.
This just uses GDS instead to do the atomic operations.
Interesting. This is kind of what I'd have expected to b
Hi,
I've had a chance to look a bit more closely at the allocator prototype
repository now. There's a whole bunch of low-level API design feedback,
but for now let's focus on the high-level stuff first.
Going by the 4.5 major object types (as also seen on slide 5 of your
presentation [0]), a
On 30.11.2017 19:52, Rob Clark wrote:
On Thu, Nov 30, 2017 at 4:21 AM, Nicolai Hähnle wrote:
On 30.11.2017 01:09, Miguel Angel Vico wrote:
It seems to me that $new_thing should grow as a separate thing whether
it ends up replacing GBM or GBM internals are somewhat rewritten on top
of it. If
On 30.11.2017 02:20, Marek Olšák wrote:
From: Marek Olšák
The same code for VI doesn't check for scanout either.
Both patches:
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_surface.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/amd/common/ac_surface.c b/src/amd/c
On 30.11.2017 16:14, Rob Clark wrote:
On Thu, Nov 30, 2017 at 9:53 AM, Nicolai Hähnle wrote:
On 28.11.2017 15:01, Rob Clark wrote:
On Tue, Nov 21, 2017 at 4:13 PM, Eric Anholt wrote:
v2: Remove the callback, leave avoiding the recursion up to the caller
(probably by rewriting the
On 29.11.2017 05:36, Dave Airlie wrote:
From: Dave Airlie
Just reuse the cs atomics bit and emit the hw atomic state.
Reviewed-by: Nicolai Hähnle
---
src/mesa/state_tracker/st_atom_atomicbuf.c | 4
src/mesa/state_tracker/st_context.c| 2 +-
2 files changed, 5 insertions
arr_bar[n];
Anyway, that's clearly way beyond the scope of this change.
I still think the assertion makes sense, but it's not too important, so
either way:
Reviewed-by: Nicolai Hähnle
---
src/compiler/Makefile.sources | 1 +
src/compiler/nir/meson.build
On 29.11.2017 19:20, Emil Velikov wrote:
From: Emil Velikov
Cc: Juan A. Suárez
Cc: Andres Gomez
Signed-off-by: Emil Velikov
Acked-by: Nicolai Hähnle
---
Gents, I have no objections if you want to do any of the releases listed
;-)
*crickets* ;)
---
docs/release-calendar.html
Reviewed-by: Nicolai Hähnle
On 29.11.2017 16:05, Eric Engestrom wrote:
Signed-off-by: Eric Engestrom
---
src/gallium/auxiliary/hud/hud_context.c | 10 +-
src/gallium/auxiliary/hud/hud_cpufreq.c | 2 +-
src/gallium/auxiliary/hud/hud_diskstat.c | 2 +-
src/gallium
On 29.11.2017 14:48, Rob Clark wrote:
Add a new helper that drivers can use to emulate various things that
need special handling in particular in transfer_map:
1) z32_s8x24.. gl/gallium treats this as a single buffer with depth
and stencil interleaved but hardware frequently treats this a
On 28.11.2017 15:01, Rob Clark wrote:
On Tue, Nov 21, 2017 at 4:13 PM, Eric Anholt wrote:
v2: Remove the callback, leave avoiding the recursion up to the caller
(probably by rewriting the vtbl either in pctx or u_resource_vtbl)
hmm, that is still a bit ugly.. and looking at the equiv th
Can you add an explanation / spec quote for this?
On 09.11.2017 12:48, Samuel Iglesias Gonsálvez wrote:
Signed-off-by: Samuel Iglesias Gonsálvez
---
src/compiler/glsl/link_interface_blocks.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/compiler/glsl/link_interfac
201 - 300 of 4145 matches
Mail list logo