---
src/intel/vulkan/anv_allocator.c | 31 +--
src/intel/vulkan/anv_private.h | 4
2 files changed, 33 insertions(+), 2 deletions(-)
diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c
index b576101..835bcfa 100644
---
This commit just fixes up the English a bit and re-flows the comment.
---
src/intel/vulkan/anv_allocator.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c
index 05666ec..fa7101b 100644
---
Now that everything is going through the state pools, the block pool no
longer needs to be able to handle re-use.
---
src/intel/vulkan/anv_allocator.c | 33 ++---
src/intel/vulkan/anv_private.h | 3 ---
2 files changed, 2 insertions(+), 34 deletions(-)
diff --git
Since the state_stream is now pulling from a state_pool, the only thing
pulling directly off the block pool is the state pool so we can just
move the block_size there. The one exception is when we allocate
binding tables but we can just reference the state pool there as well.
The only functional
---
src/intel/vulkan/anv_allocator.c | 195 +--
1 file changed, 104 insertions(+), 91 deletions(-)
diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c
index a168af5..183d2cb 100644
--- a/src/intel/vulkan/anv_allocator.c
+++
---
src/intel/vulkan/anv_batch_chain.c | 47 +-
src/intel/vulkan/anv_private.h | 2 +-
2 files changed, 22 insertions(+), 27 deletions(-)
diff --git a/src/intel/vulkan/anv_batch_chain.c
b/src/intel/vulkan/anv_batch_chain.c
index 28193c5..3714d05 100644
---
src/intel/vulkan/anv_allocator.c | 9 +
src/intel/vulkan/anv_private.h | 2 ++
2 files changed, 11 insertions(+)
diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c
index 594cf49..d93d4c9 100644
--- a/src/intel/vulkan/anv_allocator.c
+++
This absurdly long series does something fairly simple: It pulls the
block_pool into the state_pool and makes the state pool capable of
allocating states larger than the block size. Unfortunately, there was no
good way to do it in less than 18 patches. :-(
Cc: Juan A. Suarez Romero
The helper functions aren't really gaining us as much as they claim and
are actually about to be in the way.
---
src/intel/vulkan/anv_allocator.c | 85 +---
1 file changed, 35 insertions(+), 50 deletions(-)
diff --git a/src/intel/vulkan/anv_allocator.c
---
src/intel/vulkan/anv_allocator.c | 21 ++---
src/intel/vulkan/anv_private.h | 1 -
2 files changed, 10 insertions(+), 12 deletions(-)
diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c
index 2dad400..7091fd4 100644
---
We should only use size_t when referring to sizes of bits of CPU memory.
Anything on the GPU or just a regular array length should be a type that
has the same size on both 32 and 64-bit architectures. For state
objects, we use a uint32_t because we'll never allocate a piece of
driver-internal GPU
---
src/intel/vulkan/anv_allocator.c | 71 +--
src/intel/vulkan/anv_cmd_buffer.c | 8 ++--
src/intel/vulkan/anv_descriptor_set.c | 4 +-
src/intel/vulkan/anv_private.h| 21 ++-
4 files changed, 52 insertions(+), 52 deletions(-)
diff
---
src/intel/vulkan/anv_allocator.c | 24 +++-
1 file changed, 19 insertions(+), 5 deletions(-)
diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c
index 784191e..594cf49 100644
--- a/src/intel/vulkan/anv_allocator.c
+++
Am Mittwoch, 26. April 2017, 16:16:35 CEST schrieb Constantine Kharlamov:
> Thank you for testing! What is "UH4" btw?
old GL3.3 benchmark Unigine Heaven 4.0
https://benchmark.unigine.com/heaven
Marc
>
> On 26.04.2017 17:03, Marc Dietrich wrote:
> > Am Dienstag, 25. April 2017, 13:59:44 CEST
Thank you for testing! What is "UH4" btw?
On 26.04.2017 17:03, Marc Dietrich wrote:
> Am Dienstag, 25. April 2017, 13:59:44 CEST schrieb Constantine Kharlamov:
>> Split out of "avoid redundant CB register updates" for easier review.
>>
>> Signed-off-by: Constantine Kharlamov
On 04/26/2017 12:19 PM, Samuel Pitoiset wrote:
On 04/26/2017 04:44 AM, Timothy Arceri wrote:
Can we just update is_scalar instead?
That should work but only if vector_elements is 1 for samplers.
With
https://lists.freedesktop.org/archives/mesa-dev/2017-April/153166.html
and with
Hi Tapani,
Sure, I can share them, but they are QNX based. And as far as I remember it
was an issue in the past :)
I think the simplest way for you is to add following calls to function
eglutCreateWindow():
https://cgit.freedesktop.org/mesa/demos/tree/src/egl/eglut/eglut.c#n321
if
Am Dienstag, 25. April 2017, 13:59:44 CEST schrieb Constantine Kharlamov:
> Split out of "avoid redundant CB register updates" for easier review.
>
> Signed-off-by: Constantine Kharlamov
> Tested-by: Dieter Nützel
> ---
This one generates GPU crashes
Reviewed-by: Ilia Mirkin
On Wed, Apr 26, 2017 at 9:21 AM, Boyan Ding wrote:
> v2:
> Also handle the the new format in indirect dispatch
> Use compute class check instead of chipset check
>
> Signed-off-by: Boyan Ding
> ---
>
Hi Bruce,
On 26 April 2017 at 14:24, Bruce Cherniak wrote:
> This patch moves msaa resolve down into core/StoreTiles where it can
> take advantage of all the surface formats - previous resolve
> was limited to 8-bit unsigned. This fixes a number of piglit msaa
> tests
On 26 April 2017 at 12:56, Jean Hertel wrote:
>>On 2017-04-26 at 08:13, Emil Velikov wrote:
>
>>Thanks again for doing this Jean. A handful of, mostly trivial, questions.
>>
>> - Is there a website where one can check the output of specific
This patch moves msaa resolve down into core/StoreTiles where it can
take advantage of all the surface formats - previous resolve
was limited to 8-bit unsigned. This fixes a number of piglit msaa
tests that were crashing. MSAA performance is also greatly improved
because resolve is done in
v2:
Also handle the the new format in indirect dispatch
Use compute class check instead of chipset check
Signed-off-by: Boyan Ding
---
src/gallium/drivers/nouveau/nvc0/nve4_compute.c | 164 +++-
src/gallium/drivers/nouveau/nvc0/nve4_compute.h | 67
On 26 April 2017 at 12:18, Eric Engestrom wrote:
> On Wednesday, 2017-04-26 12:13:17 +0100, Emil Velikov wrote:
>> Eric,
>> Not sure that squashing 200+ changes (resulting in a total of +37K
>> -41K change stat) in one big blob is a wise move.
>> Since those are going
Hello,
I'm not an r600g person but I will be pleased to have a look to this work.
I'm currently working on the fp64 for r600g. Hope to be able to send a
series in fews days.
The fp64 code is not HW specific, I can test it on my laptop with an
Intel card. I don't even own a r600g card.
Have a
query and return supported dmabuf format modifiers for
EGL_EXT_image_dma_buf_import_modifiers.
v2: rebase to master.
v3: move format check to the driver instead of making format queries
from here and then checking.
Signed-off-by: Varad Gautam
---
From: Varad Gautam
ask the driver for supported modifiers for a given format.
bump __DRIimageExtension to 16.
v2: move to __DRIimageExtension v16.
v3: fail if the supplied format is not supported by driver.
Signed-off-by: Varad Gautam
From: Varad Gautam
allow egl clients to query the dmabuf formats supported on this platform.
v2: rebase to master.
v3: return EGL_FALSE upon failure.
Signed-off-by: Louis-Francis Ratté-Boulianne
Signed-off-by: Varad Gautam
allows drivers to be queried for supported formats and format
modifiers for EGL_EXT_image_dma_buf_import_modifiers. drivers that
implement format/modifier queries must advertise these under
PIPE_CAP_QUERY_DMABUF_ATTRIBS.
v2: rebase to master.
v3: drivers must return false on query failure.
From: Varad Gautam
these allow querying the driver for supported dmabuf formats and
modifiers.
v2: move to __DRIimageExtension version 16.
v3: return GLBoolean for error reporting, document params better.
Signed-off-by: Varad Gautam
---
From: Varad Gautam
ask the driver for supported dmabuf formats
v2: rebase to master.
v3: return false on failure.
Signed-off-by: Varad Gautam
---
src/gallium/state_trackers/dri/dri2.c | 15 +++
1 file changed, 15
>On 2017-04-26 06:23, Eric Engestrom wrote:
>
>
>Perfect, thanks!
>I left a couple comments on 0658cf3db95c77dbe590.
>
Ok, I will focus on fixing the issues pointed out.
>>
>> Foot Note: Some changes are still necessary because some specification files
>> (patents.txt, libGL.txt, features.txt,
The regioning parameters are now properly set by convert_to_hw_regs()
and we don't need to fix them in the generator.
Signed-off-by: Samuel Iglesias Gonsálvez
Cc: "17.1"
---
src/intel/compiler/brw_vec4_generator.cpp | 9 +
1 file
On gen7, the swizzles used in DF align16 instructions works for element
size of 32 bits, so we can address only 2 consecutive DFs. As we assumed that
in the rest of the code and prepare the instructions for this (scalarize_df()),
we need to set it to two again.
However, for DF align1
From IVB PRM, vol4, part3, "General Restrictions on Regioning
Parameters":
"If ExecSize = Width and HorzStride ≠ 0, VertStride must
be set to Width * HorzStride."
In next patch, we are going to modify the region parameter for
uniforms and vgrf. For uniforms that are the source of
DF align1
>On 2017-04-26 at 08:13, Emil Velikov wrote:
>Thanks again for doing this Jean. A handful of, mostly trivial, questions.
>
> - Is there a website where one can check the output of specific
>Sphinx markdown.
I have searched for such a thing too, but have not found :(
>
I don't see any reasons why vector_elements is 1 for images and
0 for samplers. This increases consistency and allows to clean
up some code a bit.
This will also help for ARB_bindless_texture.
No piglit regressions with RadeonSI.
Signed-off-by: Samuel Pitoiset
---
On Wednesday, 2017-04-26 12:13:17 +0100, Emil Velikov wrote:
> Eric,
> Not sure that squashing 200+ changes (resulting in a total of +37K
> -41K change stat) in one big blob is a wise move.
> Since those are going to be pushed at once, there should be
> limited-to-no side effects to mesa3d.org.
On 25 April 2017 at 19:56, Jean Hertel wrote:
>>On 2017-04-24 12:40, Eric Engestrom wrote:
>>
>>Sorry I haven't had time to look at this until now.
>>
>>I just had a look, and besides style that can be tweaked later, I think
>>the main issue is that the homepage a visitor
Reorder the uniforms to load first the dvec4-aligned variables
in the push constant buffer and then push the vec4-aligned ones.
This fixes a bug were the dvec3/4 might be loaded one part on a GRF and
the rest in next GRF, so the region parameters to read that could break
the HW rules.
It was setting XYWZ swizzle and writemask to all uniforms, no matter if they
were a vector or scalar, so this can lead to problems when loading them
to the push constant buffer.
Moreover, 'shift' calculation was designed to calculate the offset in
DWORDS, but it doesn't take into account DFs, so
On 04/26/2017 04:44 AM, Timothy Arceri wrote:
Can we just update is_scalar instead?
That should work but only if vector_elements is 1 for samplers.
On 24/04/17 20:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Samplers are represented using 64-bit integer handles".
On 26/04/17 20:14, Timothy Arceri wrote:
On 26/04/17 19:50, Timothy Arceri wrote:
On 26/04/17 17:28, Nicolai Hähnle wrote:
On 26.04.2017 04:45, Timothy Arceri wrote:
Reviewed-by: Timothy Arceri
On 24/04/17 20:35, Samuel Pitoiset wrote:
The spec doesn't clearly
On 04/26/2017 12:14 PM, Timothy Arceri wrote:
On 26/04/17 19:50, Timothy Arceri wrote:
On 26/04/17 17:28, Nicolai Hähnle wrote:
On 26.04.2017 04:45, Timothy Arceri wrote:
Reviewed-by: Timothy Arceri
On 24/04/17 20:35, Samuel Pitoiset wrote:
The spec doesn't
On 26/04/17 19:50, Timothy Arceri wrote:
On 26/04/17 17:28, Nicolai Hähnle wrote:
On 26.04.2017 04:45, Timothy Arceri wrote:
Reviewed-by: Timothy Arceri
On 24/04/17 20:35, Samuel Pitoiset wrote:
The spec doesn't clearly state this, but I have got clarifiation
from
On 4/26/17, Michel Dänzer wrote:
> On 26/04/17 05:07 PM, Gregory Hainaut wrote:
>>
>> Note: those dri2* functions are typically called by gallium/mesa state
>> tracker to handle new backbuffer allocation. When the old backbuffer was
>> previously invalidated due to vsync.
>
>
On 04/26/2017 11:44 AM, Timothy Arceri wrote:
On 26/04/17 17:47, Samuel Pitoiset wrote:
On 04/26/2017 03:46 AM, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
Bindless samplers/images are 64-bit unsigned integers, which
means they consume two components as specified by
On 4/26/17, Michel Dänzer wrote:
> On 26/04/17 05:07 PM, Gregory Hainaut wrote:
>> Following the discussion in "[PATCH v4 0/3] asynchronous pbo transfer with
>> glthread"
>>
>> It will help apps that are ported to XCB.
>
> How so?
I didn't test it (yet). But I think it is
On 26/04/17 17:28, Nicolai Hähnle wrote:
On 26.04.2017 04:45, Timothy Arceri wrote:
Reviewed-by: Timothy Arceri
On 24/04/17 20:35, Samuel Pitoiset wrote:
The spec doesn't clearly state this, but I have got clarifiation
from the spec authors.
Okay, that answers my
On 26/04/17 17:47, Samuel Pitoiset wrote:
On 04/26/2017 03:46 AM, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
Bindless samplers/images are 64-bit unsigned integers, which
means they consume two components as specified by
ARB_bindless_texture.
It looks like we are not
From: Marek Olšák
We could also remove index_bounds_valid and use max_index != ~0 instead.
Opinions on that are welcome.
---
src/mesa/vbo/vbo_context.c| 2 +-
src/mesa/vbo/vbo_exec_array.c | 6 +++---
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git
From: Marek Olšák
also remove the incorrect comment about primitive restart.
---
src/mesa/state_tracker/st_draw.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/src/mesa/state_tracker/st_draw.c b/src/mesa/state_tracker/st_draw.c
index
On Tuesday, 2017-04-25 18:56:53 +, Jean Hertel wrote:
> >On 2017-04-24 12:40, Eric Engestrom wrote:
> >
> >Sorry I haven't had time to look at this until now.
> >
> >I just had a look, and besides style that can be tweaked later, I think
> >the main issue is that the homepage a visitor lands
On 26/04/17 05:07 PM, Gregory Hainaut wrote:
>
> Note: those dri2* functions are typically called by gallium/mesa state
> tracker to handle new backbuffer allocation. When the old backbuffer was
> previously invalidated due to vsync.
FWIW, DRI2 buffer invalidation isn't directly related to
On 26/04/17 05:07 PM, Gregory Hainaut wrote:
> Following the discussion in "[PATCH v4 0/3] asynchronous pbo transfer with
> glthread"
>
> It will help apps that are ported to XCB.
How so?
> But Xlib (without XInitThread) apps will still crash when glthread is
> enabled on DRI2.
Do the
Hi Emil,
Thanks for review my patch,
2017-04-26 0:58 GMT+09:00 Daniel Stone :
> Hi,
>
> On 25 April 2017 at 17:03, Emil Velikov wrote:
>> From a quick look most of this code is copy/pasted from platform_wayland.c.
>>
>> At the same time, large
On 04/26/2017 10:05 AM, Nicolai Hähnle wrote:
On 24.04.2017 12:36, Samuel Pitoiset wrote:
Because the variable declaration holds more information than
the dereference. Note that an image is considered bindless either
if it has been declared in the default uniform block with the
On 04/26/2017 10:02 AM, Nicolai Hähnle wrote:
On 26.04.2017 06:45, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 5.4.1, Conversion and Scalar Constructors, p. 60"
(add the following constructors:)
//
On 04/26/2017 09:54 AM, Nicolai Hähnle wrote:
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture allows sampler and image types to be
declared as temporary variables, to be converted from/to uvec2
using explicit constructors, so they can be assigned.
Signed-off-by: Samuel
On 26.04.2017 01:12, Dave Airlie wrote:
From: Dave Airlie
This code can be shared by radv, we bump the max to
VARYING_SLOT_MAX here, but that shouldn't have too
much fallout.
Signed-off-by: Dave Airlie
Reviewed-by: Nicolai Hähnle
On Wed, 26 Apr 2017 11:03:11 +0900
Michel Dänzer wrote:
> On 25/04/17 06:14 PM, Gregory Hainaut wrote:
> > Hello,
> >
> > I did more tests on my side. DRI3 + recent stack is fine. Older
> > (Debian Jessie, ~2y old) XCB hangs/deadlock. So all DRI3 drivers
> > should be fine
On 04/26/2017 10:07 AM, Timothy Arceri wrote:
On 26/04/17 17:40, Samuel Pitoiset wrote:
Nope, I'm pretty sure this allows to declare "out bool foo" in a
fragment shader because is_varying_var() only returns TRUE for inputs.
The code below what this patch removes validates frag outputs.
On 26.04.2017 00:36, Samuel Pitoiset wrote:
Because the buffer is new, it can't be referenced by any CS.
This can save few CPU cycles by skipping the whole
PIPE_TRANSFER_UNSYNCHRONIZED if in amdgpu_bo_map().
Signed-off-by: Samuel Pitoiset
Reviewed-by: Nicolai
On 25.04.2017 22:15, Eric Engestrom wrote:
Signed-off-by: Eric Engestrom
Makes sense.
Reviewed-by: Nicolai Hähnle
---
configure.ac | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/configure.ac b/configure.ac
index
By default Xlib isn't thread safe so we better avoid it when gl thread
is enabled.
It will help applications that use XCB. But it will still crash if
applications are still relying on Xlib (without XInitThread).
Note: those dri2* functions are typically called by gallium/mesa state
tracker to
Following the discussion in "[PATCH v4 0/3] asynchronous pbo transfer with
glthread"
It will help apps that are ported to XCB. But Xlib (without XInitThread) apps
will still crash when glthread is enabled on DRI2.
I only tested the patch on Nouveau
There are 3 remaining possibilities
* Won't
On 26/04/17 17:40, Samuel Pitoiset wrote:
Nope, I'm pretty sure this allows to declare "out bool foo" in a
fragment shader because is_varying_var() only returns TRUE for inputs.
The code below what this patch removes validates frag outputs. That
should catch your example.
On 04/26/2017
On 04/26/2017 09:37 AM, Nicolai Hähnle wrote:
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 4.3.4, Inputs, p. 34"
"(modify last paragraph, p. 35, allowing samplers and images as
fragment shader inputs) ... Fragment inputs can only
On 24.04.2017 12:36, Samuel Pitoiset wrote:
Because the variable declaration holds more information than
the dereference. Note that an image is considered bindless either
if it has been declared in the default uniform block with the
bindless_sampler layout qualifier, or when its storage is not
On 24.04.2017 12:36, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 2.14.8, Shader Execution, p. 122"
"(modify second paragraph, p. 126) ... against the
MAX_COMBINED_TEXTURE_IMAGE_UNITS limit. Samplers accessed
using texture handles (section
On 04/26/2017 09:30 AM, Nicolai Hähnle wrote:
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The GL_ARB_bindless_texture allows to declare images inside
structures which means that qualifiers like writeonly should
be allowed.
I have a got a confirmation from Jeff Bolz (one author of the spec),
On 24.04.2017 12:35, Samuel Pitoiset wrote:
Signed-off-by: Samuel Pitoiset
---
src/compiler/glsl/lower_packed_varyings.cpp | 24
1 file changed, 20 insertions(+), 4 deletions(-)
diff --git a/src/compiler/glsl/lower_packed_varyings.cpp
On 26.04.2017 06:45, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 5.4.1, Conversion and Scalar Constructors, p. 60"
(add the following constructors:)
// In the following four constructors, the low 32
On 04/26/2017 09:28 AM, Nicolai Hähnle wrote:
On 26.04.2017 04:45, Timothy Arceri wrote:
Reviewed-by: Timothy Arceri
On 24/04/17 20:35, Samuel Pitoiset wrote:
The spec doesn't clearly state this, but I have got clarifiation
from the spec authors.
Okay, that answers
With Timothy's suggestion, patches 24-26:
Reviewed-by: Nicolai Hähnle
On 24.04.2017 12:35, Samuel Pitoiset wrote:
For the explicit conversions.
Signed-off-by: Samuel Pitoiset
---
src/compiler/glsl/ast_function.cpp | 3 ++-
1 file changed,
With Timothy's suggestion:
Reviewed-by: Nicolai Hähnle
On 24.04.2017 12:35, Samuel Pitoiset wrote:
Signed-off-by: Samuel Pitoiset
---
src/compiler/glsl/ast_to_hir.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture allows sampler and image types to be
declared as temporary variables, to be converted from/to uvec2
using explicit constructors, so they can be assigned.
Signed-off-by: Samuel Pitoiset
---
On 04/26/2017 06:00 AM, Timothy Arceri wrote:
On 26/04/17 12:21, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
Signed-off-by: Samuel Pitoiset
---
src/compiler/glsl/ast_to_hir.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff
On 04/26/2017 09:10 AM, Nicolai Hähnle wrote:
On 24.04.2017 12:35, Samuel Pitoiset wrote:
Bindless samplers/images are 64-bit unsigned integers, which
means they consume two components as specified by
ARB_bindless_texture.
It looks like we are not wasting uniform storage by changing
this
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Replace Section 4.1.7 (Samplers), p. 25"
"Samplers can be used as l-values, so can be assigned into and
used as "out" and "inout" function parameters."
"Replace Section 4.1.X, (Images)"
"Images
On 04/26/2017 03:46 AM, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
Bindless samplers/images are 64-bit unsigned integers, which
means they consume two components as specified by
ARB_bindless_texture.
It looks like we are not wasting uniform storage by changing
this
On 26.04.2017 06:25, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
Yes, this is a bit hacky but we don't really have the choice.
Plain GLSL doesn't accept bindless samplers/images as l-values
while it's allowed when ARB_bindless_texture is enabled.
Are you sure we need this?
On 04/26/2017 03:36 AM, Timothy Arceri wrote:
On 24/04/17 20:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Samplers are represented using 64-bit integer handles."
and,
"Images are represented using 64-bit integer handles."
It seems simpler to always consider
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Replace Section 4.1.7 (Samplers), p. 25"
"Samplers aggregated into arrays within a shader (using square
brackets []) can be indexed with arbitrary integer expressions."
Signed-off-by: Samuel Pitoiset
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Interactions with GLSL 4.20"
"Without GLSL 4.20 support, sampler and image uniforms may only
be initialized through the OpenGL API. With GLSL 4.20, sampler
and image uniforms may be initialized in
Nope, I'm pretty sure this allows to declare "out bool foo" in a
fragment shader because is_varying_var() only returns TRUE for inputs.
On 04/26/2017 05:56 AM, Timothy Arceri wrote:
Varying types have already been validated in
apply_type_qualifier_to_variable() by this point.
---
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 4.3.4, Inputs, p. 34"
"(modify last paragraph, p. 35, allowing samplers and images as
fragment shader inputs) ... Fragment inputs can only be signed
and unsigned integers and integer
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 4.3.4, Inputs, p. 34"
"(modify third paragraph of the section to allow sampler and
image types) ... Vertex shader inputs can only be float,
single-precision floating-point scalars,
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The GL_ARB_bindless_texture allows to declare images inside
structures which means that qualifiers like writeonly should
be allowed.
I have a got a confirmation from Jeff Bolz (one author of the spec),
because the spec doesn't clearly explain this.
On 04/26/2017 05:56 AM, Timothy Arceri wrote:
Here get_scalar_type() was just geing use to remove the array
after that we converted it back to bast_type anyway so just
use the without_array() helper.
Type: geing -> being
I think we can just get rid of get_scalar_type() everywhere because
On 26.04.2017 04:45, Timothy Arceri wrote:
Reviewed-by: Timothy Arceri
On 24/04/17 20:35, Samuel Pitoiset wrote:
The spec doesn't clearly state this, but I have got clarifiation
from the spec authors.
Okay, that answers my question from the other email.
BTW, what
With Timothy's comments addressed, patch 9 & 10:
Reviewed-by: Nicolai Hähnle
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Modify Section 4.3.7, Interface Blocks, p. 38"
"(remove the following bullet from the last list on p.
On 26.04.2017 04:44, Timothy Arceri wrote:
Can we just update is_scalar instead?
It's probably worth a try.
Cheers,
Nicolai
On 24/04/17 20:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Samplers are represented using 64-bit integer handles".
Signed-off-by: Samuel
Reviewed-by: Nicolai Hähnle
On 24.04.2017 12:35, Samuel Pitoiset wrote:
Signed-off-by: Samuel Pitoiset
---
src/compiler/glsl/ast_to_hir.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
On 24.04.2017 12:35, Samuel Pitoiset wrote:
The ARB_bindless_texture spec says:
"Replace Section 4.1.7 (Samplers), p. 25"
"Samplers may be declared as shader inputs and outputs, as uniform
variables, as temporary variables, and as function parameters."
"Replace Section 4.1.X,
On 24.04.2017 12:35, Samuel Pitoiset wrote:
This adds bindless_sampler and bound_sampler (and respectively
bindless_image and bound_image) to the parser.
Signed-off-by: Samuel Pitoiset
---
src/compiler/glsl/ast.h | 8
Ken?
On Sat, Apr 22, 2017 at 6:09 PM, Grazvydas Ignotas
wrote:
> Fixes: 651ec926fc1 "anv: Add support for 48-bit addresses"
> Tested-by: Grazvydas Ignotas
>
> On Sun, Apr 23, 2017 at 1:55 AM, Jason Ekstrand
> wrote:
>
>> This fixes
On 24.04.2017 12:35, Samuel Pitoiset wrote:
Bindless samplers/images are 64-bit unsigned integers, which
means they consume two components as specified by
ARB_bindless_texture.
It looks like we are not wasting uniform storage by changing
this because default-block uniforms are not packed. So,
---
src/intel/vulkan/genX_query.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/src/intel/vulkan/genX_query.c b/src/intel/vulkan/genX_query.c
index 126431b..22de3c3 100644
--- a/src/intel/vulkan/genX_query.c
+++ b/src/intel/vulkan/genX_query.c
@@ -566,6 +566,11 @@
On 26/04/17 05:56, Timothy Arceri wrote:
> Here get_scalar_type() was just geing use to remove the array
> after that we converted it back to bast_type anyway so just
Typo: base_type (instead of bast_type)
> use the without_array() helper.
Makes sense, assuming no test regression:
Reviewed-by:
101 - 200 of 202 matches
Mail list logo