On 07.06.2016 19:04, Marek Olšák wrote:
>
> @@ -286,6 +283,16 @@ static int r600_init_surface(struct r600_common_screen
> *rscreen,
> ptex->format == PIPE_FORMAT_R9G9B9E5_FLOAT)
> surface->flags |= RADEON_SURF_DISABLE_DCC;
>
> + if (ptex->bind & PIPE_BIND_SCANOUT) {
Reviewed-by: Dave Airlie
On 8 June 2016 at 10:04, Timothy Arceri wrote:
> Ping. This just stops counting and assigning a storage location for
> these uniforms, the count is only used to create the Uniform storage
> [1].
>
> These uniform types don't use this storage.
>
> [1] https://cgit.freedes
This replaces the current bash generator with a python based generator
using mako. It's quite fast and works with both python 2.7 and python
3.5, and should work with 3.3+ and maybe even 3.2.
It produces an almost identical file except for a minor layout changes,
and the addition of a "generated f
Ping. This just stops counting and assigning a storage location for
these uniforms, the count is only used to create the Uniform storage
[1].
These uniform types don't use this storage.
[1] https://cgit.freedesktop.org/mesa/mesa/tree/src/compiler/glsl/link_
uniforms.cpp#n1080
On Thu, 2016-06-02
On 06/07/2016 04:28 PM, Nanley Chery wrote:
> On Wed, May 25, 2016 at 12:52:38PM -0700, Matt Turner wrote:
>> On Wed, May 25, 2016 at 12:17 PM, Samuel Pitoiset
>> wrote:
>>> This fixes use of compute shaders with some NVIDIA GL 4.3 samples which
>>> require a compat profile.
>>
>> I'm confused.
>>
On Tue, Jun 07, 2016 at 04:28:48PM -0700, Nanley Chery wrote:
> On Wed, May 25, 2016 at 12:52:38PM -0700, Matt Turner wrote:
> > On Wed, May 25, 2016 at 12:17 PM, Samuel Pitoiset
> > wrote:
> > > This fixes use of compute shaders with some NVIDIA GL 4.3 samples which
> > > require a compat profile
On Wed, May 25, 2016 at 12:52:38PM -0700, Matt Turner wrote:
> On Wed, May 25, 2016 at 12:17 PM, Samuel Pitoiset
> wrote:
> > This fixes use of compute shaders with some NVIDIA GL 4.3 samples which
> > require a compat profile.
>
> I'm confused.
>
> The samples you're talking about... use 4.3 co
Turn off warnings for -Wpointer-arith, -Wno-missing-field-initializers,
-Wno-initializer-overrides, and -Wno-mismatched-tags. These are all deemed
pointless, on purpose or no plans to fix.
Signed-off-by: Rob Herring
---
Android.common.mk | 4
1 file changed, 4 insertions(+)
diff --git a/An
Fix warnings like these due to HAVE_LIBDRM being inconsistently defined:
external/libdrm/include/drm/drm.h:839:30: warning: redefinition of typedef
'drm_clip_rect_t' is a C11 feature [-Wtypedef-redefinition]
typedef struct drm_clip_rect drm_clip_rect_t;
HAVE_LIBDRM needs to be set project wide t
On 08.06.2016 01:05, Marek Olšák wrote:
From: Marek Olšák
Diff from shader-db:
Scratch: 3221504 -> 17408 (-99.46 %) bytes per wave
Wow.
This really needs to be fixed in LLVM. In the meantime, this series is
Reviewed-by: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 1
On 08.06.2016 01:03, Ilia Mirkin wrote:
On Tue, Jun 7, 2016 at 7:00 PM, Nicolai Hähnle wrote:
@@ -2554,16 +2564,23 @@ st_finalize_texture(struct gl_context *ctx,
/* Need to import images in main memory or held in other textures.
*/
if (stImage && stObj->pt != s
On Tue, Jun 7, 2016 at 7:05 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> Diff from shader-db:
> Scratch: 3221504 -> 17408 (-99.46 %) bytes per wave
> ---
> src/gallium/drivers/radeonsi/si_shader.c | 17 +++--
> 1 file changed, 15 insertions(+), 2 deletions(-)
>
> diff --git a/src
On 08.06.2016 01:00, Nicolai Hähnle wrote:
From: Nicolai Hähnle
The width0/height0/depth0 on stObj may not have been set at this point.
Observed in a trace that set up levels 2..9 of a 2d texture, and set the base
level to 2, with height 1. This made the guess logic always bail.
Originally inv
From: Marek Olšák
This was for some old unsupported LLVM version.
Only si_create_context creates the target machine now.
r600g doesn't use this function.
---
src/gallium/drivers/radeon/radeon_llvm_emit.c | 23 ---
src/gallium/drivers/radeon/radeon_llvm_emit.h | 1 -
src/gall
From: Marek Olšák
Diff from shader-db:
Scratch: 3221504 -> 17408 (-99.46 %) bytes per wave
---
src/gallium/drivers/radeonsi/si_shader.c | 17 +++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeons
On Tue, Jun 7, 2016 at 7:00 PM, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> The width0/height0/depth0 on stObj may not have been set at this point.
> Observed in a trace that set up levels 2..9 of a 2d texture, and set the base
> level to 2, with height 1. This made the guess logic always ba
From: Nicolai Hähnle
The width0/height0/depth0 on stObj may not have been set at this point.
Observed in a trace that set up levels 2..9 of a 2d texture, and set the base
level to 2, with height 1. This made the guess logic always bail.
Originally investigated by Ilia Mirkin, this patch gets rid
On 08.06.2016 00:20, Brian Paul wrote:
I took a stab at fixing this too. My raw patch is attached below. Very
little testing.
Something like that would probably cover most bases, except that a
program that just uploads a base level=2 image of size 1x1 will still be
screwed.
I ended up not
Reviewed-by: Ilia Mirkin
On Tue, Jun 7, 2016 at 6:25 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/mesa/state_tracker/st_context.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/src/mesa/state_tracker/st_context.c
> b/src/mesa/state_tracker/st_context.c
> index 1ee0aec..4
On Wed, May 11, 2016 at 1:32 PM, Gurchetan Singh
wrote:
> With this change, to enable precise SIN and COS instructions
> on Intel hardware, one can put
>
>
>
> in the proper drirc file.
>
> V2: Make option name more generic
Reviewed-by: Stéphane Marchesin
> ---
> src/mesa/drivers/dri/common/x
From: Marek Olšák
---
src/mesa/state_tracker/st_context.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/mesa/state_tracker/st_context.c
b/src/mesa/state_tracker/st_context.c
index 1ee0aec..4b32399 100644
--- a/src/mesa/state_tracker/st_context.c
+++ b/src/mesa/state_tracker/st_context
I took a stab at fixing this too. My raw patch is attached below. Very
little testing.
-Brian
On 06/07/2016 04:06 PM, Nicolai Hähnle wrote:
Scratch that, there's a good reason to try to preserve the previous
texture size... working on yet another alternative.
On 07.06.2016 23:19, Nicolai Hä
Scratch that, there's a good reason to try to preserve the previous
texture size... working on yet another alternative.
On 07.06.2016 23:19, Nicolai Hähnle wrote:
From: Nicolai Hähnle
The width0/height0/depth0 on stObj may not have been set at this point.
Observed in a trace that set up level
For the series: Reviewed-by: Charmaine Lee
From: Brian Paul
Sent: Tuesday, June 7, 2016 12:03 PM
To: mesa-dev@lists.freedesktop.org
Cc: Neha Bhende; Charmaine Lee
Subject: [PATCH 6/6] llvmpipe: turn on pipe cap for GL_ARB_copy_image support
---
src/galli
From: Nicolai Hähnle
These are no longer needed since the last commit.
---
src/mesa/state_tracker/st_cb_eglimage.c | 3 --
src/mesa/state_tracker/st_cb_texture.c | 60 +++--
src/mesa/state_tracker/st_manager.c | 3 --
src/mesa/state_tracker/st_texture.h |
From: Nicolai Hähnle
The width0/height0/depth0 on stObj may not have been set at this point.
Observed in a trace that set up levels 2..9 of a 2d texture, and set the base
level to 2, with height 1. This made the guess logic always bail.
Originally investigated by Ilia Mirkin, this patch makes su
On 7 June 2016 at 19:26, Ian Romanick wrote:
> Patches 2, 3, 4 (with the trivial comment message nit addressed), 5, 6,
> 7, 8 (with the trivial comment message nit addressed), 10, 14, and 15 are
>
> Reviewed-by: Ian Romanick
>
> Patch 13 is
>
> Acked-by: Ian Romanick
>
> I'd like to do some arch
Reviewed-by: Nicolai Hähnle
On 07.06.2016 21:39, Marek Olšák wrote:
From: Marek Olšák
disabled by 4f1cccf570112f93265a4cace504eb763fa8f73e
---
src/gallium/drivers/radeonsi/si_state.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/s
On 07.06.2016 17:19, Rob Clark wrote:
On Tue, Jun 7, 2016 at 10:13 AM, Emil Velikov wrote:
On 7 June 2016 at 14:03, Nicolai Hähnle wrote:
On 04.06.2016 05:26, Timothy Arceri wrote:
On Fri, 2016-06-03 at 13:12 +0200, wrote:
Hello
Situation: Looking at the content displayed by the web br
On 07.06.2016 16:08, Ilia Mirkin wrote:
On Tue, Jun 7, 2016 at 8:50 AM, Nicolai Hähnle wrote:
On 04.06.2016 19:28, Ilia Mirkin wrote:
We were previously unconditionally doing this for arrays and ubo's, and
ignoring texture/storage/atomic buffers. Instead use the usage history
to determine whi
For the series:
Reviewed-by: Nicolai Hähnle
On 07.06.2016 12:04, Marek Olšák wrote:
From: Marek Olšák
Since DCC is enabled almost everywhere now, it's important not to disable
this fast path.
---
src/gallium/drivers/radeonsi/si_blit.c | 15 +--
src/gallium/drivers/radeonsi/si
From: Marek Olšák
disabled by 4f1cccf570112f93265a4cace504eb763fa8f73e
---
src/gallium/drivers/radeonsi/si_state.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_state.c
b/src/gallium/drivers/radeonsi/si_state.c
index 6c879f3..e99d30
On 07.06.2016 20:26, Ian Romanick wrote:
On 06/07/2016 07:26 AM, Juha-Pekka Heikkila wrote:
ralloc originally had had idea of using malloc but somehow
had slipped in calloc. Without these changes rzalloc did double
job of zeroing the memory, first calloc and then memset.
Now change ralloc to use
This primarily means added support for copying between compressed
and uncompressed formats.
---
src/gallium/auxiliary/util/u_surface.c | 115 +++--
1 file changed, 95 insertions(+), 20 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_surface.c
b/src/gallium/auxi
---
src/gallium/drivers/softpipe/sp_screen.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/softpipe/sp_screen.c
b/src/gallium/drivers/softpipe/sp_screen.c
index e74f2d2..514009c 100644
--- a/src/gallium/drivers/softpipe/sp_screen.c
+++ b/src/gallium/dri
See earlier for comment for softpipe (same story).
Note that we were not supporting 3-component formats for render
targets before, now we don't for textures either.
---
src/gallium/drivers/llvmpipe/lp_screen.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/drivers/llvmpipe/lp_s
I never finished checking in my changes for GL_ARB_copy_image support
for softpipe/llvmpipe last fall. This series just rebases the old
patches on current master with minor updates.
Some of these were previously reviewed by Marek.
-Brian
___
mesa-dev
1. Try to choose R8G8B8A8 unorm/srgb formats before others in an
effort to try to match component ordering for UINT/SINT/etc.
2. If we can't get a format such as PIPE_FORMAT_A16_UNORM, try
PIPE_FORMAT_R16G16B16A16_UNORM before shallower formats.
Reviewed-by: Marek Olšák
---
src/mesa/state_track
---
src/gallium/drivers/llvmpipe/lp_screen.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/llvmpipe/lp_screen.c
b/src/gallium/drivers/llvmpipe/lp_screen.c
index 50a6513..d252c7a 100644
--- a/src/gallium/drivers/llvmpipe/lp_screen.c
+++ b/src/gallium/dri
Mesa and gallium don't have a complete set of matching 3-component
texture formats. For example, 8-bit sRGB unorm. To fully support
the GL_ARB_copy_image extension we need to have support for all of
these formats: RGB8_UNORM, RGB8_SNORM, RGB8_SRGB, RGB8_UINT, and
RGB8_SINT using the same componen
On 07.06.2016 18:07, Ilia Mirkin wrote:
On Tue, Jun 7, 2016 at 10:26 AM, Juha-Pekka Heikkila
wrote:
Signed-off-by: Juha-Pekka Heikkila
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 2 +-
src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp| 2 +-
src/mesa/drivers/dri/
Patches 2, 3, 4 (with the trivial comment message nit addressed), 5, 6,
7, 8 (with the trivial comment message nit addressed), 10, 14, and 15 are
Reviewed-by: Ian Romanick
Patch 13 is
Acked-by: Ian Romanick
I'd like to do some archaeology for patch 9 to be more sure about it. I
think it's co
LGTM. Nuke it!
Reviewed-by: Jason Ekstrand
On Tue, Jun 7, 2016 at 9:33 AM, Emil Velikov
wrote:
> From: Emil Velikov
>
> ... and associated file(s).
>
> No longer needed since commit 057259655e7 ("i965: Don't link libmesa or
> libdri_test_stubs into tests")
>
> Cc: Jason Ekstrand
> Signed-of
On 06/07/2016 09:33 AM, Emil Velikov wrote:
> From: Emil Velikov
>
> Signed-off-by: Emil Velikov
> ---
> src/mesa/main/context.c | 45 +
> 1 file changed, 21 insertions(+), 24 deletions(-)
>
> diff --git a/src/mesa/main/context.c b/src/mesa/main/cont
On 06/07/2016 09:33 AM, Emil Velikov wrote:
> From: Emil Velikov
>
> An earlier commit 16ee7a55ae2 ("mesa: Allow contexts of different APIs
> to coexist.") reworked the remapping so that the table and approach used
> is identical for all APIs.
>
> Yet again with function itself not API specific,
On 06/07/2016 09:33 AM, Emil Velikov wrote:
> From: Emil Velikov
>
> Unused as of commit 5a175127f38 ("dri: Remove all extension enabling
> utility functions")
...and the the patch before the previous patch.
>
> Signed-off-by: Emil Velikov
> ---
> src/mesa/main/remap.c | 44 -
2016-06-07 18:33 GMT+02:00 Emil Velikov :
> From: Emil Velikov
>
> ... and inline its contents in _mesa_do_init_remap_table().
>
in _mesa_init_remap_table().
Signed-off-by: Emil Velikov
> ---
> src/mesa/main/remap.c | 25 +++--
> 1 file changed, 7 insertions(+), 18 dele
Apply the luma key filter to the YCbCr values during the CSC conversion
in video buffer shader. The initial values of max and min luma are set
to opposite values to disable the filter initially and will be set when
enabling it.
Add extra parmeters min and max luma for the luma key filter in
vl_com
On 06/07/2016 07:26 AM, Juha-Pekka Heikkila wrote:
> Switch to use memory allocations which zero memory for places
> where needed.
This series is in backwards order. Assuming that rzalloc and friends
are necessary in these places, if a git-bisect hits the previous patch,
a person will (potentiall
On 06/07/2016 07:26 AM, Juha-Pekka Heikkila wrote:
> ralloc originally had had idea of using malloc but somehow
> had slipped in calloc. Without these changes rzalloc did double
> job of zeroing the memory, first calloc and then memset.
> Now change ralloc to use malloc, leave rzalloc to use calloc
Hi Nayan,
Since the core area is VL lease use "gallium/vl:" or just "vl:".
On 7 June 2016 at 14:32, Nayan Deshmukh wrote:
> Apply the luma key filter to the YCbCr values during the CSC conversion
> in video buffer shader. The initial values of max and min luma are set
> to opposite values to dis
On 06/07/2016 09:38 AM, Brian Paul wrote:
> On 06/07/2016 10:09 AM, Mathias Fröhlich wrote:
>> Hi,
>>
>> On Tuesday, June 07, 2016 16:15:58 Nicolai Hähnle wrote:
>>
>> > Thanks for staying with this!
>>
>> >
>>
>> > I know it's bike-shedding, and I'm not sure what other people's
>> opinions
>>
>
On 06/06/2016 10:20 PM, Dave Airlie wrote:
> From: Dave Airlie
>
> This fixes:
> GL45-CTS.shader_subroutine.subroutines_cannot_be_assigned_float_int_values_or_be_compared
>
> though I'm not 100% sure why this is illegal from the spec,
> but it makes us pass the test, and I really can't see a use
Hi Marc-André, Dave
On 7 June 2016 at 13:54, Marc-André Lureau wrote:
> When calling virgl_fence_wait() with timeout=0,
> virgl_{drm,vtest}_resource_is_busy() is called. However, it returns TRUE
> for a busy resource, whereace virgl_fence_wait() should return TRUE for
> a completed (non-busy) res
Reviewed-by: Bruce Cherniak
> On Jun 3, 2016, at 12:39 PM, Rowley, Timothy O
> wrote:
>
> Use rasterizer provoking vertex API.
>
> Fix rasterizer provoking vertex for tristrips and quad list/strips.
>
> v2: make provoking vertex tables static const
> ---
> .../drivers/swr/rasterizer/core/fr
On 06/07/2016 10:09 AM, Mathias Fröhlich wrote:
Hi,
On Tuesday, June 07, 2016 16:15:58 Nicolai Hähnle wrote:
> Thanks for staying with this!
>
> I know it's bike-shedding, and I'm not sure what other people's opinions
> are on this matter, but having the duplication of u_bit_scan and
>
From: Emil Velikov
Final user was killed with last commit.
Signed-off-by: Emil Velikov
---
src/mesa/main/remap.c | 18 --
src/mesa/main/remap.h | 9 -
2 files changed, 27 deletions(-)
diff --git a/src/mesa/main/remap.c b/src/mesa/main/remap.c
index c64b819..e76293c 10
From: Emil Velikov
Signed-off-by: Emil Velikov
---
src/mesa/main/context.c | 45 +
1 file changed, 21 insertions(+), 24 deletions(-)
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index 894c38a..f8961fc 100644
--- a/src/mesa/main/cont
From: Emil Velikov
The actual code of the function print_table_stats() is guarded
by a ifdef GET_DEBUG, which was not been defined in years.
The last fix in 2013 (7db6b5aa91a) indicates that it's rarely
used/tested. Since the issue has gone unnoticed for a whole year
(broken with 2ad4a475474).
From: Emil Velikov
It has been unused for a long time, plus makes the gallium dri modules
require an extra glapi symbol relative to their classic counterparts.
Signed-off-by: Emil Velikov
---
We can keep the hook, if other (non-public) state-trackers need/use it.
I've went ahead with the best
From: Emil Velikov
Signed-off-by: Emil Velikov
---
src/mesa/state_tracker/st_manager.c | 26 +-
1 file changed, 13 insertions(+), 13 deletions(-)
diff --git a/src/mesa/state_tracker/st_manager.c
b/src/mesa/state_tracker/st_manager.c
index 5e82f4c..b204658 100644
--- a/
From: Emil Velikov
We have return on the previous line, thus the break will never be
reached.
Signed-off-by: Emil Velikov
---
src/mesa/state_tracker/st_manager.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/mesa/state_tracker/st_manager.c
b/src/mesa/state_tracker/st_manager.c
index
From: Emil Velikov
Inline the function into it's only caller. This way it's more obvious
how the classic and gallium drivers (st/mesa) use _mesa_initialize_context.
Signed-off-by: Emil Velikov
---
src/mesa/main/context.c | 38 -
src/mesa/main/con
From: Emil Velikov
All of the functions and related data is internal, so there's no point
if using the GL types.
Signed-off-by: Emil Velikov
---
src/mesa/main/remap.c | 11 ++-
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/src/mesa/main/remap.c b/src/mesa/main/remap.c
i
From: Emil Velikov
An earlier commit 16ee7a55ae2 ("mesa: Allow contexts of different APIs
to coexist.") reworked the remapping so that the table and approach used
is identical for all APIs.
Yet again with function itself not API specific, thus one should not
need to invoke it multiple times.
Si
Hi all,
A selection of mostly dead code removal and "inline this one/two line function
into the only caller" patches.
Note: PATCH 09/15 removes the initialisation of _mesa_init_remap_table() from
once per API to once per context. Based on my humble understanding current
approach was a side effect
From: Emil Velikov
Unused as of commit 5a175127f38 ("dri: Remove all extension enabling
utility functions")
Signed-off-by: Emil Velikov
---
src/mesa/main/remap.c | 44
src/mesa/main/remap.h | 3 ---
2 files changed, 47 deletions(-)
diff --git a/sr
From: Emil Velikov
... and inline its contents in _mesa_do_init_remap_table().
Signed-off-by: Emil Velikov
---
src/mesa/main/remap.c | 25 +++--
1 file changed, 7 insertions(+), 18 deletions(-)
diff --git a/src/mesa/main/remap.c b/src/mesa/main/remap.c
index a756057..6dc42
From: Emil Velikov
Used only locally.
Signed-off-by: Emil Velikov
---
src/mesa/main/remap.c | 6 +++---
src/mesa/main/remap.h | 3 ---
2 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/src/mesa/main/remap.c b/src/mesa/main/remap.c
index e76293c..595d0a1 100644
--- a/src/mesa/main/
From: Emil Velikov
... and associated file(s).
No longer needed since commit 057259655e7 ("i965: Don't link libmesa or
libdri_test_stubs into tests")
Cc: Jason Ekstrand
Signed-off-by: Emil Velikov
---
src/mesa/drivers/dri/common/Makefile.am | 6 +-
src/mesa/drivers/dri/common/Makefile.
From: Emil Velikov
Unused as of commit 5a175127f38 ("dri: Remove all extension enabling
utility functions")
Cc: Ian Romanick
Signed-off-by: Emil Velikov
---
src/mesa/main/remap.c | 20
src/mesa/main/remap.h | 3 ---
2 files changed, 23 deletions(-)
diff --git a/src/mesa
From: Emil Velikov
The final user was nuked with last commit.
Signed-off-by: Emil Velikov
---
src/mapi/glapi/gen/remap_helper.py | 42 --
1 file changed, 42 deletions(-)
diff --git a/src/mapi/glapi/gen/remap_helper.py
b/src/mapi/glapi/gen/remap_helper.py
i
Hi,
On Tuesday, June 07, 2016 16:15:58 Nicolai Hähnle wrote:
> Thanks for staying with this!
>
> I know it's bike-shedding, and I'm not sure what other people's opinions
> are on this matter, but having the duplication of u_bit_scan and
> _mesa_bit_scan feels a bit annoying to me...
Well, yes,
On Tue, Jun 7, 2016 at 10:13 AM, Emil Velikov wrote:
>
> On 7 June 2016 at 14:03, Nicolai Hähnle wrote:
>>
>> On 04.06.2016 05:26, Timothy Arceri wrote:
>>>
>>> On Fri, 2016-06-03 at 13:12 +0200, wrote:
Hello
Situation: Looking at the content displayed by the web browser for
>
On Tue, Jun 7, 2016 at 10:26 AM, Juha-Pekka Heikkila
wrote:
> Signed-off-by: Juha-Pekka Heikkila
> ---
> src/mesa/drivers/dri/i965/brw_fs.cpp | 2 +-
> src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp| 2 +-
> src/mesa/drivers/dri/i965/brw_fs_dead_code_eliminate.cpp
Hmmm looks like you're right. I thought I was also seeing this crash
in dri swrast, but that's not the case after a second look.
Will investigate more, thanks!
Regards,
Lars Hamre
On Tue, Jun 7, 2016 at 10:10 AM, Nicolai Hähnle wrote:
> On 07.06.2016 03:21, Lars Hamre wrote:
>>
>> Unrolls the lo
Signed-off-by: Juha-Pekka Heikkila
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 2 +-
src/mesa/drivers/dri/i965/brw_fs_copy_propagation.cpp| 2 +-
src/mesa/drivers/dri/i965/brw_fs_dead_code_eliminate.cpp | 4 ++--
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git
Signed-off-by: Juha-Pekka Heikkila
---
src/mesa/drivers/dri/i965/brw_vec4_dead_code_eliminate.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_dead_code_eliminate.cpp
b/src/mesa/drivers/dri/i965/brw_vec4_dead_code_eliminate.cpp
index
rzalloc guarantee we get zeroed memory which is needed here.
Signed-off-by: Juha-Pekka Heikkila
---
src/util/set.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/util/set.c b/src/util/set.c
index 99abefd..d64291d 100644
--- a/src/util/set.c
+++ b/src/util/set.c
@@ -110,7
Signed-off-by: Juha-Pekka Heikkila
---
src/compiler/nir/nir.c | 6 +++---
src/compiler/nir/nir_opt_dce.c | 2 +-
src/compiler/nir/nir_phi_builder.c | 2 +-
src/compiler/nir/nir_search.c | 2 +-
src/compiler/nir/nir_to_ssa.c | 2 +-
src/compiler/nir/nir_worklist.c| 2
rzalloc guarantee we get zeroed memory which is needed here.
Signed-off-by: Juha-Pekka Heikkila
---
src/util/hash_table.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/util/hash_table.c b/src/util/hash_table.c
index 4cfe3d9..9097e85 100644
--- a/src/util/hash_table.c
++
ralloc originally had had idea of using malloc but somehow
had slipped in calloc. Without these changes rzalloc did double
job of zeroing the memory, first calloc and then memset.
Now change ralloc to use malloc, leave rzalloc to use calloc and
make needed changes in ralloc functions to get things
Switch to use memory allocations which zero memory for places
where needed.
Signed-off-by: Juha-Pekka Heikkila
---
src/compiler/glsl/ast_to_hir.cpp | 2 +-
src/compiler/glsl/glcpp/glcpp-parse.y | 4 ++--
src/compiler/glsl/link_uniform_blocks.cpp | 2 +-
src/compiler/glsl_types.cpp
I had somehow missed the discussion there had happened about my
earlier ralloc findings. I rebased and fixed my earlier patch set
so here it is for others to see.
I'm not at all confident this is somehow good set, it may explode on
different setup than mine. I've been running this on my IVB where
On Tue, 2016-06-07 at 15:25 +1000, Dave Airlie wrote:
> From: Dave Airlie
>
> One piece of ARB_shader_subroutine I ignored was the fact that it
> needs to store the subroutine index data per context and not per
> shader program.
>
> There is one CTS test that tests this:
> GL45-CTS.shader_subrou
Hi Mathias,
Thanks for staying with this!
I know it's bike-shedding, and I'm not sure what other people's opinions
are on this matter, but having the duplication of u_bit_scan and
_mesa_bit_scan feels a bit annoying to me...
Cheers,
Nicolai
On 07.06.2016 07:29, mathias.froehl...@gmx.net wro
On 7 June 2016 at 14:03, Nicolai Hähnle wrote:
> On 04.06.2016 05:26, Timothy Arceri wrote:
>
>> On Fri, 2016-06-03 at 13:12 +0200, ⚛ wrote:
>>
>>> Hello
>>>
>>> Situation: Looking at the content displayed by the web browser for
>>> URL
>>> http://patchwork.freedesktop.org/project/mesa/series and
On 07.06.2016 03:21, Lars Hamre wrote:
Unrolls the loop with a count of 1 if it contains an unconditional
return statement.
Fixes the following piglit test:
/spec/glsl-1.10/execution/fs-loop-return
from crashing at this assert:
tgsi/tgsi_exec.c:5952:tgsi_exec_machine_run: Assertion `mach->ContS
On Tue, Jun 7, 2016 at 8:50 AM, Nicolai Hähnle wrote:
> On 04.06.2016 19:28, Ilia Mirkin wrote:
>>
>> We were previously unconditionally doing this for arrays and ubo's, and
>> ignoring texture/storage/atomic buffers. Instead use the usage history
>> to determine which atoms need to be revalidated
On 06.06.2016 11:10, Michel Dänzer wrote:
On 04.06.2016 00:10, Marek Olšák wrote:
On Fri, Jun 3, 2016 at 4:33 PM, Dieter Nützel wrote:
Am 03.06.2016 11:47, schrieb Michel Dänzer:
On 03.06.2016 18:34, Marek =?UNKNOWN?B?T2zFocOhaw==?= wrote:
Module: Mesa
Branch: master
Commit: 8c361e84ad0105
Apply the luma key filter to the YCbCr values during the CSC conversion
in video buffer shader. The initial values of max and min luma are set
to opposite values to disable the filter initially and will be set when
enabling it.
Add extra parmeters min and max luma for the luma key filter in
vl_com
On Tue, Jun 7, 2016 at 12:04 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/gallium/drivers/radeon/r600_texture.c | 6 +++---
> src/gallium/drivers/radeon/radeon_winsys.h | 1 +
> src/gallium/drivers/radeonsi/si_blit.c | 7 +++
> src/gallium/drivers/radeonsi/si_
On 04.06.2016 14:38, Bas Nieuwenhuizen wrote:
On Fri, Jun 3, 2016 at 7:01 PM, Nicolai Hähnle wrote:
Hi,
it is quite common that few or even no descriptors (except for vertex
buffers) change between draw calls, but the function call overhead and
branchiness of the code that is involved in figur
On 04.06.2016 05:26, Timothy Arceri wrote:
On Fri, 2016-06-03 at 13:12 +0200, ⚛ wrote:
Hello
Situation: Looking at the content displayed by the web browser for
URL
http://patchwork.freedesktop.org/project/mesa/series and sub-pages
accessible via the links.
The following questions are troubling
When calling virgl_fence_wait() with timeout=0,
virgl_{drm,vtest}_resource_is_busy() is called. However, it returns TRUE
for a busy resource, whereace virgl_fence_wait() should return TRUE for
a completed (non-busy) resource.
This fixes running supertuxkart in a VM (I could not reproduce locally
w
Your version is better, thanks!
Reviewed-by: Samuel Pitoiset
On 06/07/2016 03:26 AM, Ilia Mirkin wrote:
Sometimes a register source can actually be double- or even quad-wide.
We must make sure that the inserted texbars take that width into
account.
Based on an earlier patch by Samuel Pitoiset
Reviewed-by: Nicolai Hähnle
On 04.06.2016 19:28, Ilia Mirkin wrote:
A texture may be redefined with _NEW_TEXTURE, which might have been
bound to a shader image slot. We have to revalidate the image atoms to
pick up on the new resource.
Signed-off-by: Ilia Mirkin
Cc: "12.0"
---
src/mesa/sta
On 04.06.2016 19:28, Ilia Mirkin wrote:
We were previously unconditionally doing this for arrays and ubo's, and
ignoring texture/storage/atomic buffers. Instead use the usage history
to determine which atoms need to be revalidated.
Does this need to handle shader images as well?
Nicolai
Sig
Am 07.06.2016 um 14:19 schrieb Mathieu Malaterre:
On Tue, Jun 7, 2016 at 8:05 AM, Mathieu Malaterre wrote:
Hi Alex,
On Mon, Jun 6, 2016 at 7:20 PM, Alex Deucher wrote:
On Mon, Jun 6, 2016 at 1:16 PM, Marek Olšák wrote:
[+ dri-devel]
On Mon, Jun 6, 2016 at 8:42 AM, Mathieu Malaterre wrote
On Tue, Jun 7, 2016 at 8:05 AM, Mathieu Malaterre wrote:
> Hi Alex,
>
> On Mon, Jun 6, 2016 at 7:20 PM, Alex Deucher wrote:
>> On Mon, Jun 6, 2016 at 1:16 PM, Marek Olšák wrote:
>>> [+ dri-devel]
>>>
>>> On Mon, Jun 6, 2016 at 8:42 AM, Mathieu Malaterre wrote:
Hi,
Before reportin
1 - 100 of 137 matches
Mail list logo
