On Friday, December 19, 2014 04:36:55 PM Jason Ekstrand wrote:
> All we have right now are stub implementations of these that do nothing.
> Let's avoid confusing apps by advertising an extension we don't actually
> support.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=87516
> ---
>
We don't actually have the code for the shader cache just yet, but
this configure machinery puts everything in place so that the shader
cache can be optionally compiled in.
Specifically, if the user passes no option (neither
--disable-shader-cache, nor --enable-shader-cache), then this feature
wil
The upcoming shader cache uses the SHA-1 algorithm for cryptographic
naming. These new mesa_sha1 functions are implemented with any one of
several differeny cryptographics libraries.
This code was copied from the xserver repository, (where it has
apparently been functioning well on a variety of op
On Fri, Dec 12 2014, Jose Fonseca wrote:
> Yes, ideally we'd have something small that we could bundle into mesa
> source tree, for sake of non Linux OSes.
Ken has pointed to an implementation that might be suitable for this.
I haven't reviewed that code myself, nor am I signing up to maintain a
On Fri, Dec 19, 2014 at 2:20 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> With the switch-statement, GCC 4.8.3 produces a small pile of code with
> a branch.
>
> :
> 00: 8b 54 24 04 mov0x4(%esp),%edx
> 04: b8 01 00 00 00 mov$0x1,%e
I wrote basically this same patch when I was trying to speed up basically
the same test. Unfortunately, I wasn't able to get much of an improvement
out of it. Other than the one nit below,
Reviewed-by: Jason Ekstrand
On Fri, Dec 19, 2014 at 2:20 PM, Ian Romanick wrote:
>
> From: Ian Romanick
On Fri, Dec 19, 2014 at 2:20 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> ...so that it can be inlined in the two places that call it.
if (brw->meta_in_progress != _mesa_meta_in_progress(ctx)) {
brw->meta_in_progress = _mesa_meta_in_progress(ctx);
brw->state.dirty.brw |= BRW_N
On Fri, Dec 19, 2014 at 6:02 PM, Matt Turner wrote:
>
> On Fri, Dec 19, 2014 at 5:04 PM, Jason Ekstrand
> wrote:
> > v3: fmin and fmax technically aren't commutative or associative. Things
> > get funny when one of the arguments is a NaN.
>
> I have a difficult time believing that the GLSL d
On Fri, Dec 19, 2014 at 5:04 PM, Jason Ekstrand wrote:
> v3: fmin and fmax technically aren't commutative or associative. Things
> get funny when one of the arguments is a NaN.
I have a difficult time believing that the GLSL definitions of min()
and max() intentionally have that effect.
To
On Sat, 20 Dec 2014 02:02:27 +0100, Jason Ekstrand
wrote:
This should make debugging a lot easier as GDB handles static inlines
much
better than macros. Also, static inlines are typesafe.
---
src/glsl/nir/nir.h | 109
+++--
1 file changed
https://bugs.freedesktop.org/show_bug.cgi?id=87516
--- Comment #2 from Leith Bade ---
Yeah the spec is a bit ambiguous. It is almost as if they can't see why anyone
would implement ARB_get_program_binary without supporting at least one
binaryFormat. Of course they then made it core, so IDK.
OK I
https://bugs.freedesktop.org/show_bug.cgi?id=87516
--- Comment #1 from Ian Romanick ---
(In reply to Leith Bade from comment #0)
> The current implementation of ARB_get_program_binary is incomplete and
> currently violates the spec.
>
> Specifically these are the major issues:
> * glGet with GL_
Glenn,
Thanks for taking the time to look over things.
On Fri, Dec 19, 2014 at 1:24 PM, Glenn Kennard
wrote:
>
> On Tue, 16 Dec 2014 07:04:14 +0100, Jason Ekstrand
> wrote:
>
> From: Connor Abbott
>>
>> This includes all the instructions, ifs, loops, functions, etc. This is
>> similar to the i
This commit adds some algebraic properties to the metadata of each opcode
in NIR. In particular, you now know, just from the metadata, if a given
opcode is commutative or associative. This will be useful for algebraic
transformation passes that want to be able to match a + b as well as b + a
in o
This was a left-over relic of GLSL IR that we aren't using for anything.
If we ever want that value again, we can add it back, but NIR constant
folding should be just as good as GLSL IR's if not better pretty soon, so
I'm not worried about it.
---
src/glsl/nir/glsl_to_nir.cpp | 6 ++
src/glsl
This should be much better for debugging as GDB will pick up on the fact
that it's an enum and actually tell you what you're looking at instead of
giving you some arbitrary hex value you have to go look up.
---
src/glsl/nir/nir.h | 28 ++--
1 file changed, 14 insertions(+),
On Fri, Dec 19, 2014 at 5:35 PM, Roland Scheidegger wrote:
> I'm not a fan of using semantics for consts (fwiw I think it would
> probably be a good idea to use different namespaces for system values
> and inputs too but that's another topic).
> I don't quite share the concerns about possibly indi
This should make debugging a lot easier as GDB handles static inlines much
better than macros. Also, static inlines are typesafe.
---
src/glsl/nir/nir.h | 109 +++--
1 file changed, 81 insertions(+), 28 deletions(-)
diff --git a/src/glsl/nir/nir.h
https://bugs.freedesktop.org/show_bug.cgi?id=87516
Jason Ekstrand changed:
What|Removed |Added
Priority|high|medium
Assignee|mesa-dev@list
scratch that... We need it for ES 3.0. I'll look at the bug more.
On Fri, Dec 19, 2014 at 4:36 PM, Jason Ekstrand
wrote:
>
> All we have right now are stub implementations of these that do nothing.
> Let's avoid confusing apps by advertising an extension we don't actually
> support.
>
> Bugzilla
All we have right now are stub implementations of these that do nothing.
Let's avoid confusing apps by advertising an extension we don't actually
support.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=87516
---
src/mesa/main/extensions.c | 4 ++--
1 file changed, 2 insertions(+), 2 delet
https://bugs.freedesktop.org/show_bug.cgi?id=87516
Leith Bade changed:
What|Removed |Added
Priority|medium |high
--
You are receiving this mail becaus
https://bugs.freedesktop.org/show_bug.cgi?id=87516
Bug ID: 87516
Summary: glProgramBinary violates spec
Product: Mesa
Version: git
Hardware: All
OS: All
Status: NEW
Severity: major
Priority: medi
This commit adds some algebraic properties to the metadata of each opcode
in NIR. In particular, you now know, just from the metadata, if a given
opcode is commutative or associative. This will be useful for algebraic
transformation passes that want to be able to match a + b as well as b + a
in o
I'm not a fan of using semantics for consts (fwiw I think it would
probably be a good idea to use different namespaces for system values
and inputs too but that's another topic).
I don't quite share the concerns about possibly indirectly addressing
such a const though (because if const buffers are
On Fri, Dec 19, 2014 at 5:18 PM, Chris Forbes wrote:
> On Sat, Dec 20, 2014 at 8:11 AM, Rob Clark wrote:
>> @@ -1452,6 +1593,20 @@ tgsi_transform_lowering(const struct
>> tgsi_lowering_config *config,
>>}
>> }
>>
>> + if ((info->processor == TGSI_PROCESSOR_FRAGMENT) &&
>> + c
I should have also mentioned that this series is on the
byt-cpu-optimizations branch of my fd.o tree.
On 12/19/2014 02:20 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> ...so that it can be inlined in the two places that call it.
>
> On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtu
From: Ian Romanick
The previous check was insufficient (as it did not take 'indices' into
consideration), and DX10 hardware does not need this check anyway.
Since index_bytes is no longer used, remove it.
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32
From: Ian Romanick
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Multithread:
32-bit: Difference at 95.0% confidence 0.416027% +/- 0.163529% (n=40)
64-bit: Difference at 95.0% confidence 0.494771% +/- 0.
From: Ian Romanick
You would not believe the mess GCC 4.8.3 generated for the old
switch-statement.
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Batch7:
32-bit: Difference at 95.0% confidence -0.37374%
From: Ian Romanick
With the switch-statement, GCC 4.8.3 produces a small pile of code with
a branch.
:
00: 8b 54 24 04 mov0x4(%esp),%edx
04: b8 01 00 00 00 mov$0x1,%eax
09: 81 fa 03 14 00 00 cmp$0x1403,%edx
000
From: Ian Romanick
This doesn't affect performance, but it feels more correct.
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Batch7:
32-bit: No difference proven at 95.0% confidence (n=120)
64-bit: No d
From: Ian Romanick
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Batch7:
32-bit: Difference at 95.0% confidence 0.495267% +/- 0.202063% (n=40)
64-bit: Difference at 95.0% confidence 3.57576% +/- 0.288175
From: Ian Romanick
Instead of having an extra pointer indirection in one of the hottest
loops in the driver.
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Batch7:
32-bit: Difference at 95.0% confidence
From: Ian Romanick
...so that it can be inlined in the two places that call it.
On Bay Trail-D using Fedora 20 compile flags (-m64 -O2 -mtune=generic
for 64-bit and -m32 -march=i686 -mtune=atom for 32-bit), affects
Gl32Batch7:
32-bit: No difference proven at 95.0% confidence (n=120)
64-bit: Dif
On Sat, Dec 20, 2014 at 8:11 AM, Rob Clark wrote:
> @@ -1452,6 +1593,20 @@ tgsi_transform_lowering(const struct
> tgsi_lowering_config *config,
>}
> }
>
> + if ((info->processor == TGSI_PROCESSOR_FRAGMENT) &&
> + config->lower_alpha_test &&
> + (config->alpha_func != PIP
On Fri, Dec 19, 2014 at 5:06 PM, Brian Paul wrote:
> On 12/19/2014 02:47 PM, Rob Clark wrote:
>>
>> On Fri, Dec 19, 2014 at 4:26 PM, Ilia Mirkin wrote:
>>>
>>> On Fri, Dec 19, 2014 at 4:04 PM, Rob Clark wrote:
On Fri, Dec 19, 2014 at 3:50 PM, Ilia Mirkin
wrote:
>
> On Fri
On Fri, Dec 19, 2014 at 4:47 PM, Rob Clark wrote:
> On Fri, Dec 19, 2014 at 4:26 PM, Ilia Mirkin wrote:
>> On Fri, Dec 19, 2014 at 4:04 PM, Rob Clark wrote:
>>> On Fri, Dec 19, 2014 at 3:50 PM, Ilia Mirkin wrote:
On Fri, Dec 19, 2014 at 2:11 PM, Rob Clark wrote:
> From: Rob Clark
>>>
On 12/19/2014 02:47 PM, Rob Clark wrote:
On Fri, Dec 19, 2014 at 4:26 PM, Ilia Mirkin wrote:
On Fri, Dec 19, 2014 at 4:04 PM, Rob Clark wrote:
On Fri, Dec 19, 2014 at 3:50 PM, Ilia Mirkin wrote:
On Fri, Dec 19, 2014 at 2:11 PM, Rob Clark wrote:
From: Rob Clark
This emulates alpha-test w
On Fri, Dec 19, 2014 at 4:26 PM, Ilia Mirkin wrote:
> On Fri, Dec 19, 2014 at 4:04 PM, Rob Clark wrote:
>> On Fri, Dec 19, 2014 at 3:50 PM, Ilia Mirkin wrote:
>>> On Fri, Dec 19, 2014 at 2:11 PM, Rob Clark wrote:
From: Rob Clark
This emulates alpha-test with a compare + KILL_IF.
On Fri, Dec 19, 2014 at 4:04 PM, Rob Clark wrote:
> On Fri, Dec 19, 2014 at 3:50 PM, Ilia Mirkin wrote:
>> On Fri, Dec 19, 2014 at 2:11 PM, Rob Clark wrote:
>>> From: Rob Clark
>>>
>>> This emulates alpha-test with a compare + KILL_IF. The alpha-ref value
>>> is passed to the shader via consta
On Tue, 16 Dec 2014 07:04:14 +0100, Jason Ekstrand
wrote:
From: Connor Abbott
This includes all the instructions, ifs, loops, functions, etc. This is
similar to the information in ir.h.
v2: Jason Ekstrand :
Include ralloc and hash_table from the util directory
---
src/glsl/Makefile.sou
On Fri, Dec 19, 2014 at 10:21:59PM +0100, EdB wrote:
> On Friday, December 19, 2014 01:08:31 PM Tom Stellard wrote:
> > On Sun, Dec 14, 2014 at 11:31:22AM +0100, EdB wrote:
> > > ---
> > >
> > > src/gallium/state_trackers/clover/llvm/invocation.cpp | 6 +++---
> > > 1 file changed, 3 insertions(+
On Friday, December 19, 2014 01:14:47 PM Tom Stellard wrote:
> On Sun, Dec 14, 2014 at 11:31:23AM +0100, EdB wrote:
> > OpenCL 1.2 make distinction between 3 binaries kind:
> > executable as produce by clBuildProgram and clLinkProgam
> > compiled as produce by clCompileProgram
> > library as pro
On Friday, December 19, 2014 01:08:31 PM Tom Stellard wrote:
> On Sun, Dec 14, 2014 at 11:31:22AM +0100, EdB wrote:
> > ---
> >
> > src/gallium/state_trackers/clover/llvm/invocation.cpp | 6 +++---
> > 1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/src/gallium/state_tracker
On Sun, Dec 14, 2014 at 11:31:23AM +0100, EdB wrote:
> OpenCL 1.2 make distinction between 3 binaries kind:
> executable as produce by clBuildProgram and clLinkProgam
> compiled as produce by clCompileProgram
> library as produce by clLinkProgam with -create-library option
s/produce/produced/
On Sun, Dec 14, 2014 at 11:31:22AM +0100, EdB wrote:
> ---
> src/gallium/state_trackers/clover/llvm/invocation.cpp | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/src/gallium/state_trackers/clover/llvm/invocation.cpp
> b/src/gallium/state_trackers/clover/llvm/invoc
On Fri, Dec 19, 2014 at 3:50 PM, Ilia Mirkin wrote:
> On Fri, Dec 19, 2014 at 2:11 PM, Rob Clark wrote:
>> From: Rob Clark
>>
>> This emulates alpha-test with a compare + KILL_IF. The alpha-ref value
>> is passed to the shader via constant tagged with new ALPHAREF semantic.
>> For example:
>>
>
On Fri, Dec 19, 2014 at 2:11 PM, Rob Clark wrote:
> From: Rob Clark
>
> This emulates alpha-test with a compare + KILL_IF. The alpha-ref value
> is passed to the shader via constant tagged with new ALPHAREF semantic.
> For example:
>
> FRAG
> PROPERTY FS_COLOR0_WRITES_ALL_CBUFS 1
> DCL IN[
Jason Ekstrand writes:
> Cc: Eric Anholt
Nice. With this squashed in, and maybe a s/var_defs/var_decls/, that
patch is:
Reviewed-by: Eric Anholt
signature.asc
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
htt
On Sunday, December 14, 2014 03:06:41 PM Emil Velikov wrote:
> On 11/12/14 21:51, Carl Worth wrote:
> > From: Kristian Høgsberg
> >
> > The upcoming shader cache uses the SHA-1 algorithm for cryptographic
> > naming. These new mesa_sha1 functions are implemented with the nettle
> > library.
> > -
Connor Abbott writes:
> On Thu, Dec 18, 2014 at 1:49 PM, Eric Anholt wrote:
>> Connor Abbott writes:
>>
>>> On Thu, Dec 18, 2014 at 2:01 AM, Eric Anholt wrote:
Jason Ekstrand writes:
> From: Connor Abbott
>
> This is similar to ir_validate.cpp.
>
> v2: Jason Eks
From: Rob Clark
This emulates alpha-test with a compare + KILL_IF. The alpha-ref value
is passed to the shader via constant tagged with new ALPHAREF semantic.
For example:
FRAG
PROPERTY FS_COLOR0_WRITES_ALL_CBUFS 1
DCL IN[0], COLOR, COLOR
DCL OUT[0], COLOR
0: MOV OUT[0], IN[0]
1
On 12/18/2014 10:14 PM, Kenneth Graunke wrote:
> On Tuesday, November 18, 2014 09:11:26 PM Chad Versace wrote:
>> This patch reduces the likelihood of pointer arithmetic overflow bugs in
>> gather_oa_results(), like the one fixed by b69c7c5dac.
>>
>> I haven't yet encountered any overflow bugs in t
Am 19.12.2014 um 18:37 schrieb Brian Paul:
> v2: move initialization of llvm_gs to declaration.
> ---
> src/gallium/auxiliary/draw/draw_gs.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/gallium/auxiliary/draw/draw_gs.c
> b/src/gallium/auxiliary/draw/draw_gs.c
> in
On Tue, Dec 16, 2014 at 3:47 PM, Eric Anholt wrote:
>
> Jason Ekstrand writes:
>
> > From: Connor Abbott
> >
> > These include functions for adding and removing various bits of IR and
> > helpers for iterating over all the sources and destinations of an
> > instruction. This is similar to ir.cpp
v2: move initialization of llvm_gs to declaration.
---
src/gallium/auxiliary/draw/draw_gs.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/draw/draw_gs.c
b/src/gallium/auxiliary/draw/draw_gs.c
index 40f7ed3..6375d41 100644
--- a/src/gallium/auxiliary/dra
On 19/12/14 16:37, Brian Paul wrote:
Silence warnings about possibly uninitialized variables when making a
release build.
---
src/gallium/auxiliary/gallivm/lp_bld_sample.c | 3 +++
src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c | 2 +-
2 files changed, 4 insertions(+), 1 deletion(-)
diff
---
src/gallium/auxiliary/draw/draw_gs.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/gallium/auxiliary/draw/draw_gs.c
b/src/gallium/auxiliary/draw/draw_gs.c
index 40f7ed3..7b1cc0b 100644
--- a/src/gallium/auxiliary/draw/draw_gs.c
+++ b/src/gallium/auxiliary/draw/draw_gs.c
@@ -733,6 +7
Silence warnings about possibly uninitialized variables when making a
release build.
---
src/gallium/auxiliary/gallivm/lp_bld_sample.c | 3 +++
src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c | 2 +-
2 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/gallivm/lp_b
Am 19.12.2014 um 15:04 schrieb Emil Velikov:
Due to the current implementation of the gl-vdpau interop, we depend on
the dri module having a working set of util VL functions. Otherwise we
will end up calling the stubs.
This commit partially reverts the following commits:
f093c1c8ec1(auxiliary/vl
---
src/gallium/state_trackers/clover/api/device.cpp | 21 +
src/gallium/state_trackers/clover/core/device.cpp | 11 +++
src/gallium/state_trackers/clover/core/device.hpp | 2 ++
3 files changed, 34 insertions(+)
diff --git a/src/gallium/state_trackers/clover/api/dev
patch 1 depends on clLinkProgram serie
patch 2 should be fine on master
EdB (2):
clover: support CL_PROGRAM_BINARY_TYPE (CL1.2)
clover: add missing clGetDeviceInfo CL1.2 queries
src/gallium/state_trackers/clover/api/device.cpp | 21 +++
src/gallium/state_trackers/clover/api/pro
CL_PROGRAM_BINARY_TYPE have been added to clGetProgramBuildInfo in CL1.2
---
src/gallium/state_trackers/clover/api/program.cpp | 4 +++
src/gallium/state_trackers/clover/core/program.cpp | 31 +-
src/gallium/state_trackers/clover/core/program.hpp | 1 +
3 files changed, 29 i
On Fri, Dec 19, 2014 at 8:11 AM, David Heidelberg wrote:
> Same as ARL, just has extra rounding.
> Useful for st/nine.
>
> Tested-by: Pavel Ondračka
> Reviewed-by: Marek Olšák
> Signed-off-by: David Heidelberg
Reviewed-by: Alex Deucher
> ---
> src/gallium/drivers/r300/compiler/r3xx_vertprog
On Fri, Dec 19, 2014 at 8:13 AM, David Heidelberg wrote:
> Signed-off-by: David Heidelberg
Reviewed-by: Alex Deucher
> ---
> src/gallium/state_trackers/nine/device9.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/gallium/state_trackers/nine/device9.c
> b/src/ga
https://bugs.freedesktop.org/show_bug.cgi?id=86837
Christian König changed:
What|Removed |Added
URL|http://patchwork.freedeskto |
|p.org/patch/39400/
https://bugs.freedesktop.org/show_bug.cgi?id=86837
Emil Velikov changed:
What|Removed |Added
URL||http://patchwork.freedeskto
Due to the current implementation of the gl-vdpau interop, we depend on
the dri module having a working set of util VL functions. Otherwise we
will end up calling the stubs.
This commit partially reverts the following commits:
f093c1c8ec1(auxiliary/vl: add galliumvl_stub.la)
c642e87d9f4(auxiliary/
Signed-off-by: David Heidelberg
---
src/gallium/state_trackers/nine/device9.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/state_trackers/nine/device9.c
b/src/gallium/state_trackers/nine/device9.c
index b0e42eb..1d97688 100644
--- a/src/gallium/state_trackers/n
Same as ARL, just has extra rounding.
Useful for st/nine.
Tested-by: Pavel Ondračka
Reviewed-by: Marek Olšák
Signed-off-by: David Heidelberg
---
src/gallium/drivers/r300/compiler/r3xx_vertprog.c | 7 ---
src/gallium/drivers/r300/compiler/radeon_opcodes.c | 7 +++
src/gallium/drivers/r
https://bugs.freedesktop.org/show_bug.cgi?id=86837
Emil Velikov changed:
What|Removed |Added
CC||deathsim...@vodafone.de
--- Comment #6 fr
https://bugs.freedesktop.org/show_bug.cgi?id=86837
--- Comment #5 from Emil Velikov ---
(In reply to Christian König from comment #3)
> Most likely we end up with the implementation with the dummy symbols loaded
> first (e.g. kodi loads OGL first the VDPAU). This results in a couple of
> problems
73 matches
Mail list logo
