---
src/gallium/drivers/swr/rasterizer/common/os.h | 18 +-
1 file changed, 17 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/swr/rasterizer/common/os.h
b/src/gallium/drivers/swr/rasterizer/common/os.h
index 45517f6..32eca68 100644
---
Enabling KNOB_SIMD_WIDTH = 16 for AVX512 pre-work and low level simd utils
---
.../drivers/swr/rasterizer/common/simdintrin.h | 644 +
.../drivers/swr/rasterizer/core/format_types.h | 194 ++-
src/gallium/drivers/swr/rasterizer/core/knobs.h| 25 +-
Makes these names semantically correct.
---
src/gallium/drivers/swr/rasterizer/core/api.cpp| 6 +++---
src/gallium/drivers/swr/rasterizer/core/backend.cpp| 18 +-
src/gallium/drivers/swr/rasterizer/core/multisample.h | 4 ++--
Moved the setting into the existing component control code. Fixes bad
interaction between attribute/component setting for vertex/instance ID
and component packing.
---
.../drivers/swr/rasterizer/jitter/fetch_jit.cpp| 78 ++
.../drivers/swr/rasterizer/jitter/fetch_jit.h
Adjust viewport rounding when scissor rect is disabled during macro
tile scissor setup.
---
src/gallium/drivers/swr/rasterizer/core/api.cpp | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/swr/rasterizer/core/api.cpp
On 2016-07-18 09:14:48, Jason Ekstrand wrote:
> Thanks to rebase fail, recent surface state changes effectively reverted
This happened in 09b5a71517fadd6c20b72e7ad9ea1f7539c93a42, right?
Should we mention that commit?
Reviewed-by: Jordan Justen
> 727a9b24933 and
Hi Emil,
On Mon, Jul 18, 2016 at 4:44 PM, Emil Velikov
wrote:
> Hi Nayan,
>
> A couple of suggestions, which I hope will be less controversial than
> the "s/.f/.0f/" from earlier ;-)
>
> On 14 July 2016 at 21:58, Nayan Deshmukh
> wrote:
> >
Thanks to rebase fail, recent surface state changes effectively reverted
727a9b24933 and 367cf3a2e3e which was unintentional. This should bring it
back.
Signed-off-by: Jason Ekstrand
Cc: Jordan Justen
---
Will do. Thanks for pinging me. Early June was very busy and I didn't see
this come through. I'll get back to you within a couple days.
--Jeremy
> On Jul 18, 2016, at 01:12, Tapani Pälli wrote:
>
> Jeremy, would you have time to test these changes? Patch 1 was
On Mon, Jul 18, 2016 at 11:58 PM, Emil Velikov wrote:
> On 18 July 2016 at 13:02, Tomasz Figa wrote:
>> On Mon, Jul 18, 2016 at 7:28 PM, Emil Velikov
>> wrote:
>>> Hi Tomasz,
>>>
>>> On 15 July 2016 at 08:53, Tomasz Figa
July 18 2016 9:20 AM, "Emil Velikov" wrote:
> On 18 July 2016 at 14:39, Alexander von Gluck IV
> wrote:
>
>> July 18 2016 3:29 AM, "Nicolai Hähnle" wrote:
>>> A comment further up in the same file says
>>>
>>> /* Haiku
On Mon, Jul 18, 2016 at 3:14 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/gallium/auxiliary/os/os_time.c | 15 +--
> src/gallium/auxiliary/os/os_time.h | 4
> 2 files changed, 13 insertions(+), 6 deletions(-)
>
> diff --git
On 18 July 2016 at 13:02, Tomasz Figa wrote:
> On Mon, Jul 18, 2016 at 7:28 PM, Emil Velikov
> wrote:
>> Hi Tomasz,
>>
>> On 15 July 2016 at 08:53, Tomasz Figa wrote:
>>> We can support render nodes alone without any private
Structurally, this is very similar to the existing Apple-DRI code, except I
have chosen to implement this using the __GLXDRIdisplay, etc. vtables (as
suggested originally in [1]), rather than a maze of ifdefs. This also means
that LIBGL_ALWAYS_SOFTWARE and LIBGL_ALWAYS_INDIRECT work as expected.
On Mon, Jul 18, 2016 at 3:31 PM, Ilia Mirkin wrote:
> On Mon, Jul 18, 2016 at 9:11 AM, Marek Olšák wrote:
>> From: Marek Olšák
>>
>> ---
>> src/mesa/state_tracker/st_atom.c | 15 ++-
>> src/mesa/state_tracker/st_atom.h |
Hi,
On 18.07.2016 16:11, Marek Olšák wrote:
torcs is 2% faster between the previous patch and the end of this series.
"Driver" & "Driver2" tests included with GFXBench v4 could be good to
check with this. They're fully CPU bound, synthetic 3D driver tests
doing subset of stuff that the
Am 16.07.2016 um 00:41 schrieb Boyuan Zhang:
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
On Mon, Jul 18, 2016 at 3:55 PM, Rob Clark wrote:
> On Mon, Jul 18, 2016 at 9:24 AM, Marek Olšák wrote:
>> On Mon, Jul 18, 2016 at 2:25 PM, Rob Clark wrote:
>>> On Mon, Jul 18, 2016 at 8:16 AM, Marek Olšák wrote:
On 18 July 2016 at 14:11, Marek Olšák wrote:
> From: Marek Olšák
>
> The goal is to do this in st_validate_state:
>while (dirty)
> atoms[u_bit_scan()]->update(st);
>
> That implies that atoms can't specify which flags they consume.
> There is
On 18 July 2016 at 14:39, Alexander von Gluck IV wrote:
> July 18 2016 3:29 AM, "Nicolai Hähnle" wrote:
>> A comment further up in the same file says
>>
>> /* Haiku provides debug_printf in libroot with OS.h */
>>
>> Is that no longer true?
>>
>>
On Mon, Jul 18, 2016 at 9:24 AM, Marek Olšák wrote:
> On Mon, Jul 18, 2016 at 2:25 PM, Rob Clark wrote:
>> On Mon, Jul 18, 2016 at 8:16 AM, Marek Olšák wrote:
>>> From: Marek Olšák
>>>
>>> There are 2 uses:
>>> -
Reviewed-by: Iago Toral Quiroga
On Mon, 2016-07-18 at 16:39 +0300, Andres Gomez wrote:
> subroutine variables are to be used just in the way functions are
> called. Although the spec doesn't say it explicitely, this means that
> these variables are not to be used in any other
Reviewed-by: Iago Toral Quiroga
On Mon, 2016-07-11 at 12:11 +1000, Timothy Arceri wrote:
> At this point there is no reason not to be using the linked shaders,
> using the linked shaders should be faster and will make things
> simpler
> for upcoming shader cache work.
>
>
subroutine variables are to be used just in the way functions are
called. Although the spec doesn't say it explicitely, this means that
these variables are not to be used in any other way than those left
for function calls. Therefore, a comparison between 2 subroutine
variables should also cause a
July 18 2016 3:29 AM, "Nicolai Hähnle" wrote:
> A comment further up in the same file says
>
> /* Haiku provides debug_printf in libroot with OS.h */
>
> Is that no longer true?
>
> Nicolai
>
> On 16.07.2016 16:27, Alexander von Gluck IV wrote:
>
>> ---
>>
Am 16.07.2016 um 00:41 schrieb Boyuan Zhang:
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
Am 16.07.2016 um 00:41 schrieb Boyuan Zhang:
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
On Mon, Jul 18, 2016 at 9:11 AM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/mesa/state_tracker/st_atom.c | 15 ++-
> src/mesa/state_tracker/st_atom.h | 22 ++
> 2 files changed, 28 insertions(+), 9 deletions(-)
>
>
On Mon, Jul 18, 2016 at 2:25 PM, Rob Clark wrote:
> On Mon, Jul 18, 2016 at 8:16 AM, Marek Olšák wrote:
>> From: Marek Olšák
>>
>> There are 2 uses:
>> - Asynchronous flushing for multithreaded drivers.
>> - Return a fence without
On Mon, Jul 18, 2016 at 2:38 PM, Eric Engestrom
wrote:
> On Mon, Jul 18, 2016 at 02:14:49PM +0200, Marek Olšák wrote:
>> From: Marek Olšák
>>
>> ---
>> src/gallium/auxiliary/os/os_time.c | 15 +--
>> src/gallium/auxiliary/os/os_time.h
From: Marek Olšák
finally unused
---
src/mesa/main/dd.h | 1 -
src/mesa/main/pipelineobj.c | 3 ---
src/mesa/main/shaderapi.c | 6 --
3 files changed, 10 deletions(-)
diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h
index 114cbd2..a3dc191 100644
---
From: Marek Olšák
---
src/mesa/state_tracker/st_atom.c | 15 ++-
src/mesa/state_tracker/st_atom.h | 22 ++
2 files changed, 28 insertions(+), 9 deletions(-)
diff --git a/src/mesa/state_tracker/st_atom.c b/src/mesa/state_tracker/st_atom.c
From: Marek Olšák
This just needs to be done by st_validate_state.
---
src/mesa/state_tracker/st_atom.c | 9 ++
src/mesa/state_tracker/st_cb_program.c | 53 --
2 files changed, 9 insertions(+), 53 deletions(-)
diff --git
From: Marek Olšák
This bug seems to have always been there. Applications changing shaders
but not textures between draw calls would have gotten undefined behavior.
---
src/mesa/state_tracker/st_atom.h | 18 --
1 file changed, 12 insertions(+), 6 deletions(-)
From: Marek Olšák
This won't be needed after the rewrite.
---
src/mesa/state_tracker/st_atom.c | 70 ++--
1 file changed, 3 insertions(+), 67 deletions(-)
diff --git a/src/mesa/state_tracker/st_atom.c b/src/mesa/state_tracker/st_atom.c
From: Marek Olšák
Already done as part of ST_NEW_FRAGMENT_PROGRAM in st_validate_state.
---
src/mesa/state_tracker/st_context.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/src/mesa/state_tracker/st_context.c
b/src/mesa/state_tracker/st_context.c
From: Marek Olšák
---
src/mesa/state_tracker/st_atom_array.c | 1 -
src/mesa/state_tracker/st_atom_atomicbuf.c | 6 --
src/mesa/state_tracker/st_atom_blend.c | 1 -
src/mesa/state_tracker/st_atom_clip.c | 1 -
From: Marek Olšák
---
src/mesa/state_tracker/st_atom_shader.c | 4
1 file changed, 4 insertions(+)
diff --git a/src/mesa/state_tracker/st_atom_shader.c
b/src/mesa/state_tracker/st_atom_shader.c
index b3401c4..7a23469 100644
---
From: Marek Olšák
The goal is to do this in st_validate_state:
while (dirty)
atoms[u_bit_scan()]->update(st);
That implies that atoms can't specify which flags they consume.
There is exactly one ST_NEW_* flag for each atom. (58 flags in total)
There are macros
Hi,
This series rewrites state tracking in st/mesa. The advantages of the new
approach are described in the main patch. Here I'll just say it that it was
motivated by a little inefficient code in st_validate_state and the mapping
of _NEW_* flags to atoms. The new approach works with one ST_NEW_*
Hi,
On 15/07/16 22:46, Francisco Jerez wrote:
> Alejandro Piñeiro writes:
>
>> On 14/07/16 21:24, Francisco Jerez wrote:
>>> Alejandro Piñeiro writes:
>>>
Without this commit, a image is considered valid if the level of the
texture bound to
On Mon, Jul 18, 2016 at 1:35 PM, Marek Olšák wrote:
> FWIW, r600 and radeonsi should return 8.
>
> Marek
Thanks, I'll send patches.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
On Mon, Jul 18, 2016 at 02:14:49PM +0200, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/gallium/auxiliary/os/os_time.c | 15 +--
> src/gallium/auxiliary/os/os_time.h | 4
> 2 files changed, 13 insertions(+), 6 deletions(-)
>
> diff --git
From: Marek Olšák
This makes Bioshock Infinite with deferred flushing 2% faster.
---
src/gallium/auxiliary/pipebuffer/pb_cache.c | 15 +++
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/src/gallium/auxiliary/pipebuffer/pb_cache.c
From: Marek Olšák
---
src/gallium/auxiliary/pipebuffer/pb_cache.c | 16 +---
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/gallium/auxiliary/pipebuffer/pb_cache.c
b/src/gallium/auxiliary/pipebuffer/pb_cache.c
index 664867b..a1ca678 100644
From: Marek Olšák
This makes Bioshock Infinite with deferred flushing 2.2% faster.
---
src/gallium/winsys/radeon/drm/radeon_drm_bo.c | 29 ---
1 file changed, 22 insertions(+), 7 deletions(-)
diff --git
On Sat, Jul 16, 2016 at 11:51 PM, Ilia Mirkin wrote:
> Of course none of this actually changes backends to return > 0 bits...
> did you mean to follow up with further patches?
Yes, I wanted to try to change GL_VIEWPORT_SUBPIXEL_BITS at least for
some drivers (r600,
Hi,
These are small optimizations for reducing pb_cache overhead with Bioshock
Infinite.
Please review.
Marek
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
From: Marek Olšák
---
src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c | 4 +-
src/gallium/auxiliary/pipebuffer/pb_cache.c| 49 ++
src/gallium/auxiliary/pipebuffer/pb_cache.h| 12 --
src/gallium/winsys/amdgpu/drm/amdgpu_bo.c
From: Marek Olšák
---
src/gallium/winsys/amdgpu/drm/amdgpu_bo.c | 27 +--
1 file changed, 21 insertions(+), 6 deletions(-)
diff --git a/src/gallium/winsys/amdgpu/drm/amdgpu_bo.c
b/src/gallium/winsys/amdgpu/drm/amdgpu_bo.c
index 3545b78..1805ce6
From: Marek Olšák
All drivers were already doing it except virgl.
---
src/gallium/auxiliary/util/u_inlines.h | 11 ++-
src/gallium/drivers/virgl/virgl_resource.c | 5 +
2 files changed, 7 insertions(+), 9 deletions(-)
diff --git
From: Marek Olšák
to reduce the call indirections with u_resource_vtbl.
The worst call tree you could get was:
- u_transfer_inline_write_vtbl
- u_default_transfer_inline_write
- u_transfer_map_vtbl
- driver_transfer_map
- u_transfer_unmap_vtbl
From: Marek Olšák
There is less noise in CPU profile data now.
---
src/gallium/drivers/r600/r600_pipe.c| 2 +-
src/gallium/drivers/radeon/r600_buffer_common.c | 22 ++
src/gallium/drivers/radeon/r600_pipe_common.c | 13 +++--
On Mon, Jul 18, 2016 at 8:16 AM, Marek Olšák wrote:
> From: Marek Olšák
>
> There are 2 uses:
> - Asynchronous flushing for multithreaded drivers.
> - Return a fence without flushing (mid-command-buffer fence). The driver
> can defer flushing until
This series is,
Reviewed-by: Edward O'Callaghan
On 07/18/2016 10:16 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> There are 2 uses:
> - Asynchronous flushing for multithreaded drivers.
> - Return a fence without flushing (mid-command-buffer
From: Marek Olšák
---
src/gallium/drivers/radeon/r600_pipe_common.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/gallium/drivers/radeon/r600_pipe_common.c
b/src/gallium/drivers/radeon/r600_pipe_common.c
index cd4908f..647832b 100644
---
From: Marek Olšák
There are 2 uses:
- Asynchronous flushing for multithreaded drivers.
- Return a fence without flushing (mid-command-buffer fence). The driver
can defer flushing until fence_finish is called.
This is required to make Bioshock Infinite faster, which
From: Marek Olšák
---
src/gallium/auxiliary/os/os_time.c | 15 +--
src/gallium/auxiliary/os/os_time.h | 4
2 files changed, 13 insertions(+), 6 deletions(-)
diff --git a/src/gallium/auxiliary/os/os_time.c
b/src/gallium/auxiliary/os/os_time.c
index
From: Marek Olšák
---
src/gallium/drivers/radeonsi/si_perfcounter.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_perfcounter.c
b/src/gallium/drivers/radeonsi/si_perfcounter.c
index 96007a5..0ced617 100644
---
From: Marek Olšák
ported from Vulkan
---
src/gallium/drivers/radeonsi/si_compute.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_compute.c
b/src/gallium/drivers/radeonsi/si_compute.c
index 5a40286..e75f6bd
From: Marek Olšák
This effectively removes s_waitcnt instructions after FP16 exports.
Before:
v_cvt_pkrtz_f16_f32_e32 v0, v0, v1 ; 5E000300
v_cvt_pkrtz_f16_f32_e32 v1, v2, v3 ; 5E020702
exp 15, 0, 1, 0, 0, v0, v1, v0, v0 ; F800040F 0100
s_waitcnt
From: Marek Olšák
always set
---
src/gallium/drivers/r300/r300_flush.c | 2 --
src/gallium/drivers/r600/r600_hw_context.c| 3 ---
src/gallium/drivers/radeon/radeon_winsys.h| 3 +--
src/gallium/drivers/radeonsi/si_hw_context.c | 3 ---
On Mon, Jul 18, 2016 at 7:28 PM, Emil Velikov wrote:
> Hi Tomasz,
>
> On 15 July 2016 at 08:53, Tomasz Figa wrote:
>> We can support render nodes alone without any private headers, so let's
>> make support for control nodes depend on presence of
On Sat, Jul 16, 2016 at 11:51 PM, Ilia Mirkin wrote:
> On Sat, Jul 16, 2016 at 7:47 AM, Józef Kucia wrote:
>> This allows Gallium drivers to advertise the subpixel precision
>> for floating point viewports bounds.
>>
>> Signed-off-by: Józef Kucia
For the series:
Reviewed-by: Marek Olšák
Marek
On Sat, Jul 16, 2016 at 8:44 PM, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> Since commit d938b8c, the sample locations are no longer set unconditionally,
> so we need to set
Hi Nayan,
A couple of suggestions, which I hope will be less controversial than
the "s/.f/.0f/" from earlier ;-)
On 14 July 2016 at 21:58, Nayan Deshmukh wrote:
> HIGH_QUALITY_SCALING_L2 to HIGH_QUALTIY_SCALING_L9
> uses lanczos filter with number representing the
On 15 July 2016 at 08:53, Tomasz Figa wrote:
> This patch adds support for YV12 pixel format to the Android platform
> backend. Only creating EGL images is supported, it is not added to the
> list of available visuals.
>
> Signed-off-by: Tomasz Figa
>
On 15 July 2016 at 08:53, Tomasz Figa wrote:
> There are DRI_IMAGE_FOURCC macros, for which there are no corresponding
> DRI_IMAGE_FORMAT macros. To support such formats we need to make the
> lookup function take the native format directly. As a side effect, it
> simplifies
Hi Alan,
we could adjust the HQ options on either side i.e we could remove L2 as the
quality of the output
was really bad. I haven't tested the L8 and L9 personally as my hardware
won't allow
for it. But we can adjust it in case we want to implement other scaling
algorithms.
Regards,
Nayan.
On
Hi Tomasz,
On 15 July 2016 at 08:53, Tomasz Figa wrote:
> We can support render nodes alone without any private headers, so let's
> make support for control nodes depend on presence of private drm_gralloc
> headers.
>
> Signed-off-by: Tomasz Figa
> ---
>
On 18 July 2016 at 04:19, Nicolas Boichat wrote:
> On Fri, Jul 15, 2016 at 9:03 PM, Emil Velikov
> wrote:
>> On 15 July 2016 at 09:28, Nicolas Boichat wrote:
>>> android.opengl.cts.WrapperTest#testGetIntegerv1 CTS test
Reviewed-by: Samuel Pitoiset
On 07/17/2016 11:18 PM, Ilia Mirkin wrote:
Mark both L8_SRGB and L8A8_SRGB as non-renderable (the latter already
didn't have the bind flags). This makes the state tracker pick a
different format when rendering is required, or mark the fb
Jeremy, would you have time to test these changes? Patch 1 was already
reviewed by Emil.
// Tapani
On 06/07/2016 01:33 PM, Tapani Pälli wrote:
Signed-off-by: Tapani Pälli
Cc: Jeremy Huddleston Sequoia
---
src/glx/apple/apple_xgl_api_stereo.c |
A comment further up in the same file says
/* Haiku provides debug_printf in libroot with OS.h */
Is that no longer true?
Nicolai
On 16.07.2016 16:27, Alexander von Gluck IV wrote:
---
src/gallium/auxiliary/util/u_debug.h | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff
On 18.07.2016 01:46, Eric Engestrom wrote:
A few weeks ago, Jose Fonseca suggested [0] we use .editorconfig files
to try and enforce the formatting of the code, to which Michel Dänzer
suggested [1] we start by importing the existing .dir-locals.el
settings.
This is a first draft, with settings
101 - 175 of 175 matches
Mail list logo