2016-07-14 5:33 GMT+02:00 Mauro Rossi :
> Sending patches to implement DRI2_Fence extension for i915,
> to enable EGL_KHR_fence_sync and EGL_KHR_wait_sync.
>
> It is the step-by-step porting to i915 of i965/sync,
> plus patch to support {server,client}_wait_sync without bound context.
>
> [PATCH 1
Am 05.07.2016 um 13:17 schrieb Christian König:
Am 01.07.2016 um 18:18 schrieb Andy Furniss:
Christian König wrote:
Am 01.07.2016 um 18:02 schrieb Andy Furniss:
Christian König wrote:
Am 01.07.2016 um 16:42 schrieb Andy Furniss:
Boyuan Zhang wrote:
Signed-off-by: Boyuan Zhang
Is this sup
Hi,
On 12.07.2016 18:43, Brian Paul wrote:
Done.
Thanks!
I was going through the fixed bugs list from the release notes, and
noticed that at least this one wasn't marked as fixed although claimed to:
https://bugs.freedesktop.org/show_bug.cgi?id=93840
Is fixed bugs list generated fr
Am 14.07.2016 um 00:51 schrieb Boyuan Zhang:
Allow to specify more parameters in the encoding interface which previously
just hardcoded in the encoder
Signed-off-by: Boyuan Zhang
I'm still not convinced what exactly this is_idr flag should do, but for
now I've reviewed and push patch 1, 3 a
Am 14.07.2016 um 00:51 schrieb Boyuan Zhang:
Add function to copy from yv12 image to nv12 surface for VAAPI putimage call.
Existing function only work for copying from yv12 surface to nv12 image.
Signed-off-by: Boyuan Zhang
Two nitpicks here, the subject should read "vl/util: ..." and the co
This consolidates a bunch of hacks in a single place - by setting
the interpolation modes and locations on variables appropriately,
we can simply trust them in the rest of the code. This avoids
having to handle INTERP_QUALIFIER_NONE, gl_Color overrides,
sample-shading overrides, and Gen4-5 centroi
Rather than computing the barycentric mode each time we emit a LINTERP,
we can simply compute it once, as soon as we know we're doing non-flat
interpolation.
At that point, emit_linterp() doesn't do much, so fold it into the
call sites and drop it.
Signed-off-by: Kenneth Graunke
---
src/mesa/dr
brw_wm_barycentric_interp_mode is wordy, brw_barycentric_mode is less
typing and suffers from fewer line wrapping problems.
The enum values themselves don't really benefit from "WM" in the name,
either. Put "BARYCENTRIC" first instead of at the end and drop "WM".
Generated by:
for file in *.c *
This combines two copies of basically the same code.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 100 +--
1 file changed, 49 insertions(+), 51 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp
b/src/mesa/drivers/dri/i96
A bit tidier.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp
b/src/mesa/drivers/dri/i965/brw_fs.cpp
index c6f193a..1f6ff59 100644
--- a/src/mesa/drivers/dri/
Am 14.07.2016 um 00:51 schrieb Boyuan Zhang:
Add entry point for encoding which previously hardcoded for decoding purpose
only
I still can't figure out why we would want this? The variable doesn't
seem to be used in the whole patchset.
Christian.
Signed-off-by: Boyuan Zhang
---
src/ga
Signed-off-by: Vedran Miletić
---
docs/xlibdriver.html| 2 +-
src/gallium/docs/source/drivers/openswr/faq.rst | 4 ++--
src/gallium/include/pipe/p_defines.h| 2 +-
src/mesa/main/version.c | 2 +-
4 files changed, 5 insertions(+), 5 d
Den 13 juli 2016 9:48 em skrev "Kenneth Graunke" :
>
> On Wednesday, July 13, 2016 1:53:26 PM PDT Nils Wallménius wrote:
> > When possible, do the memcpy on larger blocks. This reduces cycles
> > spent in _mesa_propagate_uniforms_to_driver_storage from
> > 1.51 % to 0.62% according to perf during
I've played around with libglvnd support in Mesa12. I'm using libglvnd
from git master (commit id 1b30d15) and Mesa 12.0.1.
It tries to open the last fallback libGLX_indirect. I workarounded this via
setting __GLX_VENDOR_LIBRARY_NAME environment variable to "mesa". This
resulted in a crash when tr
This 'last' variable used in FindGLXFunction(...) may become negative,
but has been defined as unsigned int resulting in an overflow,
finally resulting in a segfault when accessing _glXDispatchTableStrings[...].
Fixed this by definining it as signed int. 'first' variable also needs to be
defined as
CovID: 1363008
Signed-off-by: Eric Engestrom
---
v2: avoid using malloc() altogether (Christian König)
---
src/gallium/auxiliary/vl/vl_bicubic_filter.c | 8 +---
1 file changed, 1 insertion(+), 7 deletions(-)
diff --git a/src/gallium/auxiliary/vl/vl_bicubic_filter.c
b/src/gallium/auxiliar
On Sat Jul 9 23:24:35 UTC 2016, Marek Olšák wrote:
> I'm open to having the on-disk shader cache for radeonsi just because
> of Borderlands 2. I have yet to discover other games that would
> benefit from it.
it is possible that wined3d does on-demand compilation? I read it in
talks about gallium
Any news? Or possible fix?
On Sunday 10 July 2016 18:26:28 Jason Ekstrand wrote:
> This smells like strange Python problems. Ccing the resident Python expert.
> Hello, compiling mesa from git is failing on this error:
>
> Making all in isl
> make[5]: Entering directory `/«PKGBUILDDIR»/build/dri/
On Wed, Jul 13, 2016 at 01:48:03PM +0200, Christian König wrote:
> Good catch, since we need only 16 offsets I think we could simply allocate
> it on the stack as well.
>
> Or am I missing something?
You're entirely right, and that's a much better fix too.
v2 coming in a few minutes.
>
> Regard
Thanks! In addition to Emil's suggestion, please be sure to definitely
reference the bugreport in the commit message:
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96835
With that fixed, the patch is
Reviewed-by: Nicolai Hähnle
On 13.07.2016 17:39, Tim Rowley wrote:
This reverts c
I've pushed 'em, thx
On Wed, Jul 13, 2016 at 1:03 PM, Francesco Ansanelli
wrote:
> Yes, please.
>
> Cheers,
> Francesco
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Both patches are:
Reviewed-by: Iago Toral Quiroga
On Wed, 2016-07-13 at 14:28 -0700, Jason Ekstrand wrote:
> It was returning true if the function types have different lengths
> rather
> than false. This was new with the SPIR-V to NIR pass and I thought
> I'd
> fixed it a while ago but it may ha
On Thu, Jul 14, 2016 at 03:21:20PM +0200, Stefan Dirsch wrote:
> This 'last' variable used in FindGLXFunction(...) may become negative,
> but has been defined as unsigned int resulting in an overflow,
> finally resulting in a segfault when accessing _glXDispatchTableStrings[...].
> Fixed this by de
From: Marek Olšák
not all places are fixed
---
si-report.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/si-report.py b/si-report.py
index c5855ff..523c452 100755
--- a/si-report.py
+++ b/si-report.py
@@ -39,7 +39,7 @@ def get_str(value, suffix = ' %'):
def calculate_pe
From: Marek Olšák
Scratch VGPRs = spilled VGPRs + private memory per thread +
an incorrectly-counted multiple of spilled SGPRs
---
si-report.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/si-report.py b/si-report.py
index 3e39350..c5855ff 100755
--- a/
From: Marek Olšák
We can remove the old reporting if people are OK with that.
The old reporting has a bug that it reports 0->N changes as 0 % in several
places. (should be inf %)
The new reporting shows:
- VGPR spilling shaders and apps (from the second file only)
- the same for SGPRs
- worst re
From: Marek Olšák
---
si-report.py | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/si-report.py b/si-report.py
index 69af89e..3e39350 100755
--- a/si-report.py
+++ b/si-report.py
@@ -58,6 +58,8 @@ class si_stats:
metrics = [
('sgprs', 'SGPRS', ''),
On Thu, Jul 14, 2016 at 12:24:32PM +0200, Pali Rohár wrote:
> Any news? Or possible fix?
Have you tried Emil's suggestion, ie. upgrading to at least 0.8.0?
Build system wizards:
Any way to check the version and abort the compilation before running
into this issue? If it helps, this prints the ve
From: Nicolai Hähnle
According to firmware guys, the new sequence that we added for Polaris should
work on all CIK parts, and should actually be faster on some parts.
---
src/gallium/drivers/radeonsi/si_debug.c | 5 -
src/gallium/drivers/radeonsi/si_state_draw.c | 7 +--
src/gallium
On Thu, Jul 14, 2016 at 10:41 AM, Eric Engestrom
wrote:
> On Thu, Jul 14, 2016 at 12:24:32PM +0200, Pali Rohár wrote:
>> Any news? Or possible fix?
>
> Have you tried Emil's suggestion, ie. upgrading to at least 0.8.0?
>
> Build system wizards:
> Any way to check the version and abort the compilat
On Thu, Jul 14, 2016 at 10:51:43AM -0400, Rob Clark wrote:
> On Thu, Jul 14, 2016 at 10:41 AM, Eric Engestrom
> wrote:
> > On Thu, Jul 14, 2016 at 12:24:32PM +0200, Pali Rohár wrote:
> >> Any news? Or possible fix?
> >
> > Have you tried Emil's suggestion, ie. upgrading to at least 0.8.0?
> >
> >
On Fri, Jul 01, 2016 at 02:29:25PM -0700, Jason Ekstrand wrote:
> seems fine to me
>
Should I use the Acked-by tag for this patch? Or push without a reviewer
tag?
Thanks,
Nanley
> On Fri, Jul 1, 2016 at 2:25 PM, Chad Versace wrote:
>
> > On Mon 27 Jun 2016, Nanley Chery wrote:
> > > In the ne
Inspired by fix for mem leak of vdpau interop, resource_from_handle
set texture reference count, that need to be decreased and released,
recall there is a similar case for DRI3, that is with VA-API glx
extension, there is temporary TFP(texture from pixmap), we target it
through dma-buf. leak happen
For example, in patch 5/11 when " VaCreateContext", we used to CALLOC_STRUCT
for "pps" and "sps" whenever we see video format is H.264. This is fine for
decode ONLY case. Now, since we added H.264 encoding, "pps" and "sps" shouldn't
be allocated. So we need to use the entry_point to determine th
On Thu, Jul 14, 2016 at 12:57 AM, Kenneth Graunke wrote:
> Rather than computing the barycentric mode each time we emit a LINTERP,
> we can simply compute it once, as soon as we know we're doing non-flat
> interpolation.
>
> At that point, emit_linterp() doesn't do much, so fold it into the
> call
If an internal user creates an image with Vulkan tiling VK_IMAGE_TILING_OPTIMAL
and an ISL tiling that isn't set, ISL will fail to create the image as
anv_image_create_info::isl_tiling_flags will be an invalid value.
Correct this by making anv_image_create_info::isl_tiling_flags an opt-in,
filteri
On 14 July 2016 at 15:41, Eric Engestrom wrote:
> On Thu, Jul 14, 2016 at 12:24:32PM +0200, Pali Rohár wrote:
>> Any news? Or possible fix?
>
> Have you tried Emil's suggestion, ie. upgrading to at least 0.8.0?
>
Other than this, I'm out of ideas. My python/mako isn't that great.
> Build system w
On 14 July 2016 at 04:11, Tomasz Figa wrote:
> Hi Emil,
>
> On Thu, Jul 14, 2016 at 1:28 AM, Emil Velikov
> wrote:
>> On 13 July 2016 at 04:29, Nicolas Boichat wrote:
>>> From: Tomasz Figa
>>>
>>> An earlier patch fixed the problem for classic drivers, however Gallium
>>> was still left broken
Hello,
It occurred to me that the llvmpipe might be a good basis for a GLSL
debugger. Has anyone thought about this and how it might be approached?
Rob Conde
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mai
On 14 July 2016 at 15:23, Eric Engestrom wrote:
> On Thu, Jul 14, 2016 at 03:21:20PM +0200, Stefan Dirsch wrote:
>> This 'last' variable used in FindGLXFunction(...) may become negative,
>> but has been defined as unsigned int resulting in an overflow,
>> finally resulting in a segfault when acces
Am 14.07.2016 um 15:25 schrieb Rob Conde:
> Hello,
>
>It occurred to me that the llvmpipe might be a good basis for a GLSL
> debugger. Has anyone thought about this and how it might be approached?
>
>
> Rob Conde
I'm not entirely sure what exactly you want to debug, but if you just
want to
I'll take a look at that. The idea would be to be able to step through a shader
and watch values etc.
Rob
From: Roland Scheidegger
Sent: Thursday, July 14, 2016 12:26:58 PM
To: Rob Conde; mesa-dev@lists.freedesktop.org
Subject: Re: [Mesa-dev] GLSL Debugging
Am
On Thu, Jul 14, 2016 at 04:01:13PM +0100, Eric Engestrom wrote:
> Oh right, there's already check for the Mako version, but the minimum is
> currently set to 0.3.4 (configure.ac:92).
>
> Emil, you were the one to mention 0.8.0; is that the actual minimum, or
> just a known working version?
OK, so
Note that various GPUs (definitely NVIDIA ones, but I think others
too) also include instruction stepping/breakpoints/etc functionality.
However it's never been clear to me how that sort of thing could be
effectively exposed to userspace.
-ilia
On Thu, Jul 14, 2016 at 12:29 PM, Rob Conde wrot
If you have an Nvidia card and are using a sufficiently new version of opengl
you can debug shaders in Visual Studio, but those are some significant
limitations. It would also be nice to have a reference driver to help
distinguish buggy driver behavior.
From: ibm
My point is ... any stepping you can do with softpipe/llvmpipe, you
can also do with an NVIDIA G80+ GPU. The trick is in coming up with an
interface to expose the debugging capability in a useful manner.
An application might do any number of draws, each of which might
generate any number of primit
On 14 July 2016 at 16:06, Zhang, Boyuan wrote:
> For example, in patch 5/11 when " VaCreateContext", we used to CALLOC_STRUCT
> for "pps" and "sps" whenever we see video format is H.264. This is fine for
> decode ONLY case. Now, since we added H.264 encoding, "pps" and "sps"
> shouldn't be allo
Likewise, rename the enum type to glsl_interp_mode.
Beyond the GLSL front-end, talking about "interpolation modes" seems
more natural than "interpolation qualifiers" - in the IR, we're removed
from how exactly the source language specifies how to interpolate an
input. Also, SPIR-V calls these "de
rb
On Thu, Jul 14, 2016 at 8:32 AM, Nanley Chery wrote:
> If an internal user creates an image with Vulkan tiling
> VK_IMAGE_TILING_OPTIMAL
> and an ISL tiling that isn't set, ISL will fail to create the image as
> anv_image_create_info::isl_tiling_flags will be an invalid value.
>
> Correct thi
On Thu, Jul 14, 2016 at 6:59 AM, Iago Toral wrote:
> Both patches are:
> Reviewed-by: Iago Toral Quiroga
>
Thanks!
>
> On Wed, 2016-07-13 at 14:28 -0700, Jason Ekstrand wrote:
> > It was returning true if the function types have different lengths
> > rather
> > than false. This was new with
R-B
On Thu, Jul 14, 2016 at 8:02 AM, Nanley Chery wrote:
> On Fri, Jul 01, 2016 at 02:29:25PM -0700, Jason Ekstrand wrote:
> > seems fine to me
> >
>
> Should I use the Acked-by tag for this patch? Or push without a reviewer
> tag?
>
> Thanks,
> Nanley
>
> > On Fri, Jul 1, 2016 at 2:25 PM, Chad
On Thursday, July 14, 2016 8:23:23 AM PDT Matt Turner wrote:
> On Thu, Jul 14, 2016 at 12:57 AM, Kenneth Graunke
> wrote:
> > Rather than computing the barycentric mode each time we emit a LINTERP,
> > we can simply compute it once, as soon as we know we're doing non-flat
> > interpolation.
> >
>
Less typing and word wrapping issues than intrin->variables[0]->var.
Signed-off-by: Kenneth Graunke
---
src/compiler/nir/nir_lower_io.c | 28
1 file changed, 12 insertions(+), 16 deletions(-)
diff --git a/src/compiler/nir/nir_lower_io.c b/src/compiler/nir/nir_lower_
Both loads and atomics had identical code to rewrite destinations,
and all cases had the same two lines to replace instructions.
Signed-off-by: Kenneth Graunke
---
src/compiler/nir/nir_lower_io.c | 44 ++---
1 file changed, 19 insertions(+), 25 deletions(-)
d
This can't happen, the caller asserts that mode is shader_out or shared.
Signed-off-by: Kenneth Graunke
---
src/compiler/nir/nir_lower_io.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/compiler/nir/nir_lower_io.c b/src/compiler/nir/nir_lower_io.c
index 3a8587a..914e0e1 100644
--- a/src
The original function was becoming a bit hard to read, with the details
of creating and filling out load/store/atomic atomics all in one
function.
This patch makes helpers for creating each type of intrinsic, and also
combines them with the *_op() helpers, as they're closely coupled and
not too la
Signed-off-by: Kenneth Graunke
---
src/compiler/nir/nir_intrinsics.h | 16 +---
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/compiler/nir/nir_intrinsics.h
b/src/compiler/nir/nir_intrinsics.h
index ccca1ff..2f74555 100644
--- a/src/compiler/nir/nir_intrinsics.h
++
The load/store/atomic cases all duplicated the get_io_offset code, with
a few tiny differences: stores didn't bother checking for per-vertex
inputs, because they can't be stored to, and atomics didn't check at
all, since shared variables aren't per-vertex.
However, it's harmless to check, and allo
Without this commit, a image is considered valid if the level of the
texture bound to the image is complete, something we can check as mesa
save independently if it is "base incomplete" of "mipmap incomplete".
But, from the OpenGL 4.3 Core Specification, section 8.25 ("Texture
Image Loads and Stor
On Thu, Jul 07, 2016 at 07:34:24PM -0700, Anuj Phogat wrote:
> Signed-off-by: Anuj Phogat
> ---
> src/mapi/glapi/registry/gl.xml | 1 +
> src/mesa/main/extensions_table.h | 1 +
> src/mesa/main/mtypes.h | 1 +
> src/mesa/main/teximage.c | 5 +++--
> 4 files changed, 6 insertio
On Thu, Jul 07, 2016 at 07:34:26PM -0700, Anuj Phogat wrote:
> Signed-off-by: Anuj Phogat
> ---
Patches 2 and 3 are:
Reviewed-by: Nanley Chery
> docs/GL3.txt | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/docs/GL3.txt b/docs/GL3.txt
> index ce34869..883604a 100644
> --- a/docs/GL3.t
On Thu 14 Jul 2016, Pohjolainen, Topi wrote:
>
> Subject says: "isl/state: Divide the aux qpitch by 2". Should be
> divide by 4 or shift by 2.
>
> Otherwise:
>
> Reviewed-by: Topi Pohjolainen
Ditto. Fix that and
Reviewed-by: Chad Versace
___
mesa-d
On Wed 13 Jul 2016, Jason Ekstrand wrote:
> ---
> src/intel/isl/isl.c | 7 +--
> 1 file changed, 5 insertions(+), 2 deletions(-)
Reviewed-by: Chad Versace
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailm
On Wed 13 Jul 2016, Jason Ekstrand wrote:
> ---
> src/intel/isl/isl.c | 7 +--
> 1 file changed, 5 insertions(+), 2 deletions(-)
Reviewed-by: Chad Versace
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailma
On Wed 13 Jul 2016, Jason Ekstrand wrote:
> ---
> src/mesa/drivers/dri/i965/brw_context.c | 2 ++
> src/mesa/drivers/dri/i965/brw_context.h | 4
> 2 files changed, 6 insertions(+)
Reviewed-by: Chad Versace
___
mesa-dev mailing list
mesa-dev@lists.
V2: Drop the changes to gl.xml.
Signed-off-by: Anuj Phogat
---
src/mesa/main/extensions_table.h | 1 +
src/mesa/main/mtypes.h | 1 +
src/mesa/main/teximage.c | 5 +++--
3 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/src/mesa/main/extensions_table.h b/src/mesa/ma
Few files were missing the copyright headers. This patch adds correct
copyright headers according to the commit author and date.
Signed-off-by: Vedran Miletić
---
src/gallium/state_trackers/nine/nine_dump.c | 21 +
src/gallium/state_trackers/nine/nine_dump.h | 21 ++
On Thu, Jul 14, 2016 at 11:12:45AM -0700, Anuj Phogat wrote:
> V2: Drop the changes to gl.xml.
>
> Signed-off-by: Anuj Phogat
> ---
> src/mesa/main/extensions_table.h | 1 +
> src/mesa/main/mtypes.h | 1 +
> src/mesa/main/teximage.c | 5 +++--
> 3 files changed, 5 insertions(+)
On Wed 13 Jul 2016, Jason Ekstrand wrote:
> Reviewed-by: Topi Pohjolainen
> ---
> src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 24
> src/mesa/drivers/dri/i965/intel_mipmap_tree.h | 4
> 2 files changed, 28 insertions(+)
>
> diff --git a/src/mesa/drivers/dri/i965
On Wed 13 Jul 2016, Jason Ekstrand wrote:
> Reviewed-by: Topi Pohjolainen
> ---
> src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 175
> +-
> src/mesa/drivers/dri/i965/intel_mipmap_tree.h | 6 +
> 2 files changed, 179 insertions(+), 2 deletions(-)
>
> diff --git a/src/
On Thu, Jul 14, 2016 at 11:42 AM, Chad Versace
wrote:
> On Wed 13 Jul 2016, Jason Ekstrand wrote:
> > Reviewed-by: Topi Pohjolainen
> > ---
> > src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 175
> +-
> > src/mesa/drivers/dri/i965/intel_mipmap_tree.h | 6 +
> > 2 file
On Thu, Jul 14, 2016 at 11:30 AM, Nanley Chery wrote:
> On Thu, Jul 14, 2016 at 11:12:45AM -0700, Anuj Phogat wrote:
>> V2: Drop the changes to gl.xml.
>>
>> Signed-off-by: Anuj Phogat
>> ---
>> src/mesa/main/extensions_table.h | 1 +
>> src/mesa/main/mtypes.h | 1 +
>> src/mesa/main/t
Adds a second optional cleanup callback, called after the fence is
signaled. This is needed if, for example, the queue has the last
reference to the object that embeds the util_queue_fence. In this
case we cannot drop the ref in the main callback, since that would
result in the fence being destro
Alejandro Piñeiro writes:
> Without this commit, a image is considered valid if the level of the
> texture bound to the image is complete, something we can check as mesa
> save independently if it is "base incomplete" of "mipmap incomplete".
>
> But, from the OpenGL 4.3 Core Specification, sectio
On Thu, Jul 14, 2016 at 10:49 AM, Kenneth Graunke
wrote:
> Less typing and word wrapping issues than intrin->variables[0]->var.
>
> Signed-off-by: Kenneth Graunke
> ---
> src/compiler/nir/nir_lower_io.c | 28
> 1 file changed, 12 insertions(+), 16 deletions(-)
>
> d
On Thu, Jul 14, 2016 at 10:49 AM, Kenneth Graunke
wrote:
> Both loads and atomics had identical code to rewrite destinations,
> and all cases had the same two lines to replace instructions.
>
> Signed-off-by: Kenneth Graunke
> ---
> src/compiler/nir/nir_lower_io.c | 44
> ++-
On Thu, Jul 14, 2016 at 10:49 AM, Kenneth Graunke
wrote:
> The original function was becoming a bit hard to read, with the details
> of creating and filling out load/store/atomic atomics all in one
> function.
>
> This patch makes helpers for creating each type of intrinsic, and also
> combines t
I gave a few trivial comments. Series is
Reviewed-by: Jason Ekstrand
On Thu, Jul 14, 2016 at 10:49 AM, Kenneth Graunke
wrote:
> Signed-off-by: Kenneth Graunke
> ---
> src/compiler/nir/nir_intrinsics.h | 16 +---
> 1 file changed, 9 insertions(+), 7 deletions(-)
>
> diff --git a/
In d035d50 this changed to 64b.. which I'm pretty sure was
unintentional. Revert it back to 32b so the entire state struct
is a nice round 64b (cache-line size).
(Note sure that it would actually be measurable, but I did notice
that check_state() was hot in some benchmarks.)
Signed-off-by: Rob C
2016-07-14 22:14 GMT+02:00 Rob Clark :
> In d035d50 this changed to 64b.. which I'm pretty sure was
> unintentional. Revert it back to 32b so the entire state struct
> is a nice round 64b (cache-line size).
Actually, cache line size, at least on most x86 and ARM processors, is
64B (bytes) not 64b
On 07/07/2016 07:11 PM, Miklós Máté wrote:
On 06/26/2016 09:48 PM, Miklós Máté wrote:
Only include the ones that can be used by the shader.
This fixes texture coordinates, which were completely wrong,
because WPOS was included in the list of attribs. It also
increases performance noticeably.
S
On Thu, Jul 14, 2016 at 4:41 PM, Gustaw Smolarczyk wrote:
> 2016-07-14 22:14 GMT+02:00 Rob Clark :
>> In d035d50 this changed to 64b.. which I'm pretty sure was
>> unintentional. Revert it back to 32b so the entire state struct
>> is a nice round 64b (cache-line size).
> Actually, cache line size
On Thu, Jul 14, 2016 at 11:56 AM, Jason Ekstrand
wrote:
>
>
> On Thu, Jul 14, 2016 at 11:42 AM, Chad Versace
> wrote:
>
>> On Wed 13 Jul 2016, Jason Ekstrand wrote:
>> > Reviewed-by: Topi Pohjolainen
>> > ---
>> > src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 175
>> +
On Thu 14 Jul 2016, Jason Ekstrand wrote:
> On Thu, Jul 14, 2016 at 11:42 AM, Chad Versace
> wrote:
>
> > On Wed 13 Jul 2016, Jason Ekstrand wrote:
> > > Reviewed-by: Topi Pohjolainen
> > > ---
> > > src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 175
> > +-
> > > src/m
On Thu 14 Jul 2016, Jason Ekstrand wrote:
> On Thu, Jul 14, 2016 at 11:56 AM, Jason Ekstrand
> wrote:
>
> >
> >
> > On Thu, Jul 14, 2016 at 11:42 AM, Chad Versace
> > wrote:
> >
> >> On Wed 13 Jul 2016, Jason Ekstrand wrote:
> >> > Reviewed-by: Topi Pohjolainen
> >> > ---
> >> > src/mesa/drive
On Wed 13 Jul 2016, Jason Ekstrand wrote:
> Reviewed-by: Topi Pohjolainen
> ---
> src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 120
> ++
> src/mesa/drivers/dri/i965/intel_mipmap_tree.h | 5 ++
> 2 files changed, 125 insertions(+)
>
> diff --git a/src/mesa/drivers/dr
HIGH_QUALITY_SCALING_L2 to HIGH_QUALTIY_SCALING_L9
uses lanczos filter with number representing the size
of the sinc window.
Signed-off-by: Nayan Deshmukh
---
src/gallium/state_trackers/vdpau/mixer.c | 150 ++-
src/gallium/state_trackers/vdpau/query.c | 8 ++
Hi guys,
I have tried to implement a simple lanczos filter
but I am getting line artifacts in the output.
I was not able to spot the error even after going over
the code many times. Please have a look at code and
suggest changes.
Thanks,
Nayan.
Nayan Deshmukh (2):
vl: add a lanczos interpolat
Signed-off-by: Nayan Deshmukh
---
src/gallium/auxiliary/Makefile.sources | 2 +
src/gallium/auxiliary/vl/vl_lanczos_filter.c | 436 +++
src/gallium/auxiliary/vl/vl_lanczos_filter.h | 63
3 files changed, 501 insertions(+)
create mode 100755 src/gallium/auxil
VAAPI passes PIPE_VIDEO_ENTRYPOINT_ENCODE as entry point for encoding case. We
will save this encode entry point in config. config_id was used as profile
previously. Now, config has both profile and entrypoint field, and config_id is
used to get the config object. Later on, we pass this entrypoi
Add necessary functions/changes for VAAPI encoding to buffer and picture. These
changes will allow driver to handle all Vaapi encode related operations. This
patch doesn't change the Vaapi decode behaviour.
Signed-off-by: Boyuan Zhang
---
src/gallium/state_trackers/va/buffer.c | 6 ++
sr
For putimage call, if image format is yv12 (or IYUV with U V field swap) and
surface format is nv12, then we need to convert yv12 to nv12 and then copy the
converted data from image to surface. We can't use the existing logic where
surface is destroyed and re-created with yv12 format.
Signed-of
Enable H.264 VAAPI encoding through config. Currently only H.264 baseline is
supported.
Signed-off-by: Boyuan Zhang
---
src/gallium/state_trackers/va/config.c | 32 ++--
1 file changed, 22 insertions(+), 10 deletions(-)
diff --git a/src/gallium/state_trackers/va/con
Add entrypoint to distinguish H.264 decode and encode. For example, in patch
5/11 when is calling "VaCreateContext", "pps" and "sps" shouldn't be allocated
for H.264 encoding. So we need to use the entry_point to determine this is
H.264 decode or H.264 encode. We can use config to determine the
Add function to copy from yv12 image to nv12 surface for VAAPI putimage call.
We need this function in VaPutImage call where copying from yv12 image to nv12
surface for encoding. Existing function can't be used because it only work for
copying from yv12 surface to nv12 image in Vaapi.
Signed-of
Add some hardcoded values hardware needs mainly for rate control purpose. With
previously hardcoded values for OMX, the rate control result is not correct.
This change fixed the rate control result by setting correct values for Vaapi.
Signed-off-by: Boyuan Zhang
---
src/gallium/state_trackers/
Thanks for the suggestion Emil. I modified each un-submitted patch based on the
article you provided. Please see the newly submitted patch set.
Regards,
Boyuan
-Original Message-
From: Emil Velikov [mailto:emil.l.veli...@gmail.com]
Sent: July-14-16 1:07 PM
To: Zhang, Boyuan
Cc: Christia
Fixed both subject and coding style in newly submitted patch set.
Regards,
Boyuan
-Original Message-
From: Christian König [mailto:deathsim...@vodafone.de]
Sent: July-14-16 3:56 AM
To: Zhang, Boyuan; mesa-dev@lists.freedesktop.org
Subject: Re: [PATCH 06/11] st/va: add copy function for y
On Thu, Jul 14, 2016 at 1:54 PM, Chad Versace
wrote:
> On Wed 13 Jul 2016, Jason Ekstrand wrote:
> > Reviewed-by: Topi Pohjolainen
> > ---
> > src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 120
> ++
> > src/mesa/drivers/dri/i965/intel_mipmap_tree.h | 5 ++
> > 2 file
Rate control method is passed from app to driver through config attrib list.
That is why we need to store this rate control method to config. And later on,
we will pass this value to context->desc.h264enc.rate_ctrl.rate_ctrl_method.
Signed-off-by: Boyuan Zhang
---
src/gallium/state_trackers/va
1 - 100 of 127 matches
Mail list logo
