On Wed, 15 Apr 2015 19:02:50 +0300
Marko Moberg wrote:
> Thanks for the replies.
>
> I tried to upgrade LLVM from 3.3 to 3.4 after I sent the e-mail but ended
> up having "an internal compiler error" from gcc linaro toolchain compiler.
> Nice. I could try a more up-to-date version of Mesa to see
Rob Clark writes:
> From: Rob Clark
>
> v2: move ishl into ttn (instead of driver backend) to keep the units
> consistent between immediate and indirect offsets
This looks good to me.
Reviewed-by: Eric Anholt
signature.asc
Description: PGP signature
_
Jason Ekstrand writes:
> On Sat, Apr 11, 2015 at 4:25 PM, Thomas Helland
> wrote:
>> The performance numbers (shader-db runtime) are:
>>
>> Difference at 95.0% confidence
>> -14.7608 +/- 3.36786
>> -9.05064% +/- 2.06501%
>> (Original runtime was 160 seconds)
>
> Good Work!
>
> I had o
Reviewed-by: Jason Ekstrand
On Wed, Apr 15, 2015 at 6:18 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> Suggested by Jason on a different patch after some comments /
> questions by Ilia.
>
> Signed-off-by: Ian Romanick
> Cc: Jason Ekstrand
> ---
> src/glsl/nir/nir_search.c | 4 ++--
> 1 fi
On Wed, Apr 15, 2015 at 5:13 PM, Jason Ekstrand wrote:
> On Wed, Apr 15, 2015 at 1:31 PM, Matt Turner wrote:
>> On Wed, Apr 1, 2015 at 6:19 PM, Jason Ekstrand wrote:
>>> Instead of the complicated and broken-by-design pile of heuristics we had
>>> before, we now have a straightforward lowering:
Reviewed-by: Connor Abbott
On Wed, Apr 15, 2015 at 9:18 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> Suggested by Jason on a different patch after some comments /
> questions by Ilia.
>
> Signed-off-by: Ian Romanick
> Cc: Jason Ekstrand
> ---
> src/glsl/nir/nir_search.c | 4 ++--
> 1 fil
From: Ian Romanick
Suggested by Jason on a different patch after some comments /
questions by Ilia.
Signed-off-by: Ian Romanick
Cc: Jason Ekstrand
---
src/glsl/nir/nir_search.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/glsl/nir/nir_search.c b/src/glsl/nir/nir
On 04/15/2015 04:28 PM, Kenneth Graunke wrote:
> On Wednesday, April 15, 2015 02:47:01 PM Ian Romanick wrote:
>> On 04/15/2015 12:41 PM, Kristian Høgsberg wrote:
>>> The ir_tex opcode turns into a sample or sample_c message, which will try to
>>> compute derivatives to determine the lod. This produ
On Wed, Apr 15, 2015 at 1:31 PM, Matt Turner wrote:
> On Wed, Apr 1, 2015 at 6:19 PM, Jason Ekstrand wrote:
>> Instead of the complicated and broken-by-design pile of heuristics we had
>> before, we now have a straightforward lowering:
>>
>> 1) All header sources are copied directly using force_
On Wed, Apr 15, 2015 at 4:28 PM, Kenneth Graunke wrote:
> On Wednesday, April 15, 2015 02:47:01 PM Ian Romanick wrote:
>> On 04/15/2015 12:41 PM, Kristian Høgsberg wrote:
>> > The ir_tex opcode turns into a sample or sample_c message, which will try
>> > to
>> > compute derivatives to determine t
On Wednesday, April 15, 2015 11:38:51 AM Anuj Phogat wrote:
> This enables using _mesa_meta_pbo_TexSubImage() to upload data
> to R16G16B16X16 texture. Earlier it fell back to slower paths.
>
> Jenkins run shows no piglit regressions.
>
> Signed-off-by: Anuj Phogat
> ---
> src/mesa/drivers/dri/
On Wednesday, April 15, 2015 02:47:01 PM Ian Romanick wrote:
> On 04/15/2015 12:41 PM, Kristian Høgsberg wrote:
> > The ir_tex opcode turns into a sample or sample_c message, which will try to
> > compute derivatives to determine the lod. This produces garbage for
> > non-fragment shaders where the
On 04/14/2015 02:20 PM, Jose Fonseca wrote:
On 14/04/15 17:56, Brian Paul wrote:
Subroutines need labels so they can be identied by CAL instructions.
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 19 +--
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/src/mes
On Wed, Apr 15, 2015 at 2:14 PM, Thomas Helland
wrote:
> This follows the same pattern as in the hash_table.
>
> Reviewed-by: Jason Ekstrand
Fix the email address before committing, whoever commits.
> Signed-off-by: Thomas Helland
> ---
___
mesa-dev
On Wed, Apr 15, 2015 at 6:15 PM, Ian Romanick wrote:
> On 04/15/2015 10:56 AM, Emil Velikov wrote:
>> On 15 April 2015 at 18:33, Matt Turner wrote:
>>> On Wed, Apr 15, 2015 at 7:08 AM, Emil Velikov
>>> wrote:
Rather than forcing everyone to provide their own definition of the symbol
p
On 04/15/2015 10:56 AM, Emil Velikov wrote:
> On 15 April 2015 at 18:33, Matt Turner wrote:
>> On Wed, Apr 15, 2015 at 7:08 AM, Emil Velikov
>> wrote:
>>> Rather than forcing everyone to provide their own definition of the symbol
>>> provide a common weak one, which anyone can override if needed
https://bugs.freedesktop.org/show_bug.cgi?id=90032
--- Comment #11 from Felix von Leitner ---
Good news! It works now that I recompiled llvm with RTTI support!
I thought I had RTTI in my LLVM but I was doing it wrong.
For the record: the right way to build LLVM with RTTI support is:
cmake [path
On 04/15/2015 12:41 PM, Kristian Høgsberg wrote:
> The ir_tex opcode turns into a sample or sample_c message, which will try to
> compute derivatives to determine the lod. This produces garbage for
> non-fragment shaders where the sample coordinates don't correspond to
> subspans.
>
> We fix this
Reviewed-by: Marek Olšák
Marek
On Wed, Apr 15, 2015 at 4:08 PM, Emil Velikov wrote:
> Accidentally added since the introduction of the file.
>
> Cc: "10.4 10.5"
> Signed-off-by: Emil Velikov
> ---
> src/gallium/drivers/r300/Automake.inc | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
From: Rob Clark
v2: move ishl into ttn (instead of driver backend) to keep the units
consistent between immediate and indirect offsets
Signed-off-by: Rob Clark
---
src/gallium/auxiliary/nir/tgsi_to_nir.c | 69 -
1 file changed, 60 insertions(+), 9 deletions(
From: Rob Clark
For lowering if/else, I need a way to insert at the end of the previous
block.
Signed-off-by: Rob Clark
---
src/glsl/nir/nir_builder.h | 20 ++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/src/glsl/nir/nir_builder.h b/src/glsl/nir/nir_builder
That's good to know. I've always thought it's a bit of a pity that it
isn't optimized at all for arm (doesn't use any arm intrinsics, though
with newer llvm versions it is probably not all that important any more,
and I'm not even sure which neon intrinsics would be useful) but someone
would need t
The same rationale applies here as for the hash table.
Power of two size should give better performance,
and using the algorithm hash = sh + i/2 + i*i/2
should result in only distinct hash values when hitting collisions.
Difference at 95.0% confidence
-7.9505 +/- 2.44011
-5.04357% +/- 1.54
This follows the same pattern as in the hash_table.
Reviewed-by: Jason Ekstrand
Signed-off-by: Thomas Helland
---
src/util/set.c | 8 ++--
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/src/util/set.c b/src/util/set.c
index 2bcc743..7ff9520 100644
--- a/src/util/set.c
+++ b/s
On Wed 15 Apr 2015, Anuj Phogat wrote:
Signed-off-by: Anuj Phogat
---
src/mesa/drivers/dri/i965/brw_blorp_blit.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_blorp_blit.cpp
b/src/mesa/drivers/dri/i965/brw_blorp_blit.cpp
index 644cb41..7f
On Tue 14 Apr 2015, Ian Romanick wrote:
From: Ian Romanick
Previously GLX_EXT_create_context_es2_profile was marked as "direct
only" so that it would not depend on server support. Since the
extension required functions that are part of
GLX_ARB_create_context_profile, support for the EXT was di
On Wed 15 Apr 2015, Emil Velikov wrote:
On 14 April 2015 at 17:22, Ian Romanick wrote:
On 04/14/2015 08:36 AM, Emil Velikov wrote:
On 14 April 2015 at 13:52, Jose Fonseca wrote:
On 13/04/15 22:59, Ian Romanick wrote:
On 04/10/2015 03:36 PM, Jose Fonseca wrote:
From: José Fonseca
diff
On Wed, Apr 1, 2015 at 6:19 PM, Jason Ekstrand wrote:
> Instead of the complicated and broken-by-design pile of heuristics we had
> before, we now have a straightforward lowering:
>
> 1) All header sources are copied directly using force_writemask_all and,
> since they are guaranteed to be a
On Fri, Apr 3, 2015 at 7:28 AM, Francisco Jerez wrote:
> Jason Ekstrand writes:
>
>> On Thu, Apr 2, 2015 at 3:01 AM, Francisco Jerez
>> wrote:
>>> Jason Ekstrand writes:
>>>
Instead of the complicated and broken-by-design pile of heuristics we had
before, we now have a straightforwar
On 14/04/15 17:22, Ian Romanick wrote:
On 04/14/2015 08:36 AM, Emil Velikov wrote:
On 14 April 2015 at 13:52, Jose Fonseca wrote:
On 13/04/15 22:59, Ian Romanick wrote:
On 04/10/2015 03:36 PM, Jose Fonseca wrote:
From: José Fonseca
The latest version of GLX_EXT_create_context_es2_profile
On 14/04/15 17:35, Ian Romanick wrote:
From: Ian Romanick
Previously GLX_EXT_create_context_es2_profile was marked as "direct
only" so that it would not depend on server support. Since the
extension required functions that are part of
GLX_ARB_create_context_profile, support for the EXT was dis
Rob Clark writes:
> On Tue, Apr 14, 2015 at 1:04 PM, Eric Anholt wrote:
>> Rob Clark writes:
>>
>>> From: Rob Clark
>>>
>>> Signed-off-by: Rob Clark
>>
>> 1-3 (with the fix to 1 that you posted in irc) are:
>>
>> Reviewed-by: Eric Anholt
>>
>> I don't like the mismatch on bytes vs vec4s in t
On Wed, Apr 15, 2015 at 2:24 AM, Kenneth Graunke wrote:
> Jason noticed that shader_time was bumping the reference count on the
> gl_shader_program and gl_program structures, in code called during
> compilation.
>
> Not only were these never unreferenced, but it meant fragment shaders
> might be r
On 15/04/15 20:38, Brian Paul wrote:
On 04/15/2015 11:56 AM, Emil Velikov wrote:
On 15 April 2015 at 18:33, Matt Turner wrote:
On Wed, Apr 15, 2015 at 7:08 AM, Emil Velikov
wrote:
Rather than forcing everyone to provide their own definition of the
symbol
provide a common weak one, which anyo
The ir_tex opcode turns into a sample or sample_c message, which will try to
compute derivatives to determine the lod. This produces garbage for
non-fragment shaders where the sample coordinates don't correspond to
subspans.
We fix this by rewriting the opcode from ir_tex to ir_txl and setting the
On 04/15/2015 11:56 AM, Emil Velikov wrote:
On 15 April 2015 at 18:33, Matt Turner wrote:
On Wed, Apr 15, 2015 at 7:08 AM, Emil Velikov wrote:
Rather than forcing everyone to provide their own definition of the symbol
provide a common weak one, which anyone can override if needed.
This resol
Reduces the number of conditions tested in if to one in case of
non-integer formats. Makes no functional changes.
Signed-off-by: Anuj Phogat
---
src/mesa/drivers/dri/i965/brw_surface_formats.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_
Signed-off-by: Anuj Phogat
---
src/mesa/drivers/dri/i965/brw_blorp_blit.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_blorp_blit.cpp
b/src/mesa/drivers/dri/i965/brw_blorp_blit.cpp
index 644cb41..7fd812e 100644
--- a/src/mesa/drivers/dri
This enables using _mesa_meta_pbo_TexSubImage() to upload data
to R16G16B16X16 texture. Earlier it fell back to slower paths.
Jenkins run shows no piglit regressions.
Signed-off-by: Anuj Phogat
---
src/mesa/drivers/dri/i965/brw_surface_formats.c | 6 ++
1 file changed, 6 insertions(+)
diff
From: Ian Romanick
In response to another patch, Emil asked for some clarification how this
stuff works. Rather than just reply to the e-mail, I decided to update
the exlanation in the code.
Signed-off-by: Ian Romanick
Cc: Emil Velikov
---
src/glx/glxextensions.c | 69 +++
On Wed, Apr 15, 2015 at 9:50 AM, Ilia Mirkin wrote:
> On Wed, Apr 15, 2015 at 12:37 PM, Ian Romanick wrote:
>> From: Ian Romanick
>>
>> Shader-db results:
>>
>> GM45 NIR:
>> total instructions in shared programs: 4082044 -> 4081919 (-0.00%)
>> instructions in affected programs: 27609 -> 2748
On Wed, Apr 15, 2015 at 06:58:02PM +0100, Neil Roberts wrote:
> Commit 5a06ee738 added a step to the generator to set up the message
> header when generating the VS_OPCODE_PULL_CONSTANT_LOAD_GEN7
> instruction. That pseudo opcode is implemented in terms of multiple
> actual opcodes, one of which wr
On Wed, Apr 15, 2015 at 06:58:01PM +0100, Neil Roberts wrote:
> There were three places in the visitor that had a similar chunk of
> code to emit the VS_OPCODE_PULL_CONSTANT_LOAD opcode using a register
> for the offset. This patch combines the chunks into a helper function
> to reduce the code dup
There were three places in the visitor that had a similar chunk of
code to emit the VS_OPCODE_PULL_CONSTANT_LOAD opcode using a register
for the offset. This patch combines the chunks into a helper function
to reduce the code duplication. It will also be useful in the next
patch to expand what happ
Commit 5a06ee738 added a step to the generator to set up the message
header when generating the VS_OPCODE_PULL_CONSTANT_LOAD_GEN7
instruction. That pseudo opcode is implemented in terms of multiple
actual opcodes, one of which writes to one of the source registers in
order to set up the message hea
On 15 April 2015 at 18:33, Matt Turner wrote:
> On Wed, Apr 15, 2015 at 7:08 AM, Emil Velikov
> wrote:
>> Rather than forcing everyone to provide their own definition of the symbol
>> provide a common weak one, which anyone can override if needed.
>>
>> This resolved the build of the standalone
On Wed, Apr 1, 2015 at 6:19 PM, Jason Ekstrand wrote:
> ---
> src/mesa/drivers/dri/i965/brw_fs_cse.cpp | 75
>
> 1 file changed, 38 insertions(+), 37 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_cse.cpp
> b/src/mesa/drivers/dri/i965/brw_fs_cse.
On 15 April 2015 at 10:24, Michel Dänzer wrote:
> On 09.04.2015 03:27, Emil Velikov wrote:
>> As mentioned by Michel Dänzer for LLVM >= 3.6 we create the
>> LLVMTargetMachine (with triple amdgcn--), as we setup the radeonsi
>> context. For older LLVM or hardware (r600) the triple is always r600--
https://bugs.freedesktop.org/show_bug.cgi?id=90032
--- Comment #10 from Jan Vesely ---
(In reply to Felix von Leitner from comment #9)
> That fixed that problem, but there is another one:
> ^
> Makefile:893: recipe for target 'api/libclover_la-event.lo' failed
>
> That appears t
On Wed, Apr 15, 2015 at 7:08 AM, Emil Velikov wrote:
> Rather than forcing everyone to provide their own definition of the symbol
> provide a common weak one, which anyone can override if needed.
>
> This resolved the build of the standalone pipe-drivers, as it provides a
> default symbol which wa
On Wed, Apr 15, 2015 at 12:37 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> Shader-db results:
>
> GM45 NIR:
> total instructions in shared programs: 4082044 -> 4081919 (-0.00%)
> instructions in affected programs: 27609 -> 27484 (-0.45%)
> helped:44
>
> Iro
From: Ian Romanick
Shader-db results:
GM45 NIR:
total instructions in shared programs: 4082044 -> 4081919 (-0.00%)
instructions in affected programs: 27609 -> 27484 (-0.45%)
helped:44
Iron Lake NIR:
total instructions in shared programs: 5678776 -> 5678646 (-
Thanks for the replies.
I tried to upgrade LLVM from 3.3 to 3.4 after I sent the e-mail but ended
up having "an internal compiler error" from gcc linaro toolchain compiler.
Nice. I could try a more up-to-date version of Mesa to see if it helps.
Although I have learned that I can not take the most
Hi,
I can run arm llvmpipe glxgears on IFC6410 (Snapdragon 600 APQ8064,
Krait 300) without problems.
llvm 3.5
mesa 10.4.3
both are fedora 21 standard packages without modifications.
I haven't tried anything more complex than glxgears
jan
On Wed, 2015-04-15 at 16:30 +0200, Roland Scheidegger wrote
On Wed, Apr 15, 2015 at 10:32 AM, Connor Abbott wrote:
>>
>> But more immediately, I hit a sort of snag: I cannot seem to narrow
>> from 32b to 16b at the same time I move to address register. Which
>> ends up meaning I need a mov from 32b to 16b followed by a 2nd mov to
>> get it into address r
On Tue, Apr 14, 2015 at 7:08 PM, Rob Clark wrote:
> On Tue, Apr 14, 2015 at 6:24 PM, Connor Abbott wrote:
>> On Tue, Apr 14, 2015 at 5:16 PM, Rob Clark wrote:
>>> On Tue, Apr 14, 2015 at 4:59 PM, Jason Ekstrand
>>> wrote:
> + /**
> +* Addressing mode for corresponding _in
There were reports with varying success of running llvmpipe with arm in
the past but I haven't heard anything for a while. In theory it should
work (and crashing in the pixel shader is an indication the vertex
shader worked or at least didn't crash).
But it's not really tested and hence might easil
On 14 April 2015 at 17:35, Ian Romanick wrote:
> From: Ian Romanick
>
> Previously GLX_EXT_create_context_es2_profile was marked as "direct
> only" so that it would not depend on server support. Since the
> extension required functions that are part of
> GLX_ARB_create_context_profile, support f
On 14 April 2015 at 17:22, Ian Romanick wrote:
> On 04/14/2015 08:36 AM, Emil Velikov wrote:
>> On 14 April 2015 at 13:52, Jose Fonseca wrote:
>>> On 13/04/15 22:59, Ian Romanick wrote:
On 04/10/2015 03:36 PM, Jose Fonseca wrote:
>
> From: José Fonseca
>
> The latest ve
Ben Widawsky writes:
>> + void generate_set_simd4x2_header_gen9(vec4_instruction *inst,
>> + struct brw_reg dst);
>
> super bikesheddy, but this function name sounds pretty bad, though I
> understand it's a side-effect of the name of the opcode (which see
Similar to commit 127f8767e0a and 101142c4010.
Signed-off-by: Emil Velikov
---
src/gallium/targets/d3dadapter9/Makefile.am | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/gallium/targets/d3dadapter9/Makefile.am
b/src/gallium/targets/d3dadapter9/Makefile.am
index 1dc55f5..591978f 10064
Rather than forcing everyone to provide their own definition of the symbol
provide a common weak one, which anyone can override if needed.
This resolved the build of the standalone pipe-drivers, as it provides a
default symbol which was missing previously.
Cc: Rob Clark
Signed-off-by: Emil Velik
Accidentally added since the introduction of the file.
Cc: "10.4 10.5"
Signed-off-by: Emil Velikov
---
src/gallium/drivers/r300/Automake.inc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/drivers/r300/Automake.inc
b/src/gallium/drivers/r300/Automake.inc
index 9
We no longer need to (although we can) provide this symbol ourselves.
Cc: Rob Clark
Signed-off-by: Emil Velikov
---
src/gallium/drivers/freedreno/ir3/ir3_cmdline.c | 8
1 file changed, 8 deletions(-)
diff --git a/src/gallium/drivers/freedreno/ir3/ir3_cmdline.c
b/src/gallium/drivers/f
Otherwise we'll error out due to unresolved nir symbols. Note that we
still fail to link due to unresolved _mesa_error_no_memory().
Based on commit 101142c4010(xa: support for drivers which use NIR)
Cc: Rob Clark
Signed-off-by: Emil Velikov
---
src/gallium/targets/pipe-loader/Makefile.am | 2 +
Signed-off-by: Emil Velikov
---
src/glsl/Makefile.am| 13 -
src/glsl/SConscript | 2 +-
src/glsl/main.cpp | 6 --
src/glsl/tests/common.c | 30 --
4 files changed, 5 insertions(+), 46 deletions(-)
delete mode 100644 src/glsl/tests/commo
Thus we can remove the workaround that we previously had.
Cc: Rob Clark
Signed-off-by: Emil Velikov
---
src/gallium/state_trackers/xa/xa_tracker.c | 12
1 file changed, 12 deletions(-)
diff --git a/src/gallium/state_trackers/xa/xa_tracker.c
b/src/gallium/state_trackers/xa/xa_trac
Cc: Rob Clark
Signed-off-by: Emil Velikov
---
src/gallium/drivers/freedreno/Makefile.am | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/freedreno/Makefile.am
b/src/gallium/drivers/freedreno/Makefile.am
index 4b2629f..e798e44 100644
--- a/src/gallium/dri
Just like every other place in gallium.
Signed-off-by: Emil Velikov
---
src/gallium/drivers/freedreno/Makefile.am | 2 +-
src/gallium/drivers/nouveau/Makefile.am | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/freedreno/Makefile.am
b/src/gallium/driv
On Tue, Apr 14, 2015 at 1:04 PM, Eric Anholt wrote:
> Rob Clark writes:
>
>> From: Rob Clark
>>
>> Signed-off-by: Rob Clark
>
> 1-3 (with the fix to 1 that you posted in irc) are:
>
> Reviewed-by: Eric Anholt
>
> I don't like the mismatch on bytes vs vec4s in the load_ubo_indirect
> arguments
Hi,
I have been trying to get OpenGL sw rastering to work on Linux based ARM
CortexA15 device but for some reason gallium llvmpipe gives me a
segmentation fault. Has anybody managed to run sw rastering using gallium
llvmpipe on ARM? Softpipe seems to be working ok.
I am using the following versio
On 09.04.2015 03:27, Emil Velikov wrote:
> As mentioned by Michel Dänzer for LLVM >= 3.6 we create the
> LLVMTargetMachine (with triple amdgcn--), as we setup the radeonsi
> context. For older LLVM or hardware (r600) the triple is always r600--
> and is created at a later stage - radeon_llvm_compil
It is true that a gl_shader_program with ID 0 will be a fixed-function
fragment program; a gl_program with ID 0 but NULL gl_shader_program
means that it's a fixed-function vertex shader.
But that's not terribly interesting or relevant to what we're doing.
We just need to know that ID 0 means "fixe
0 is not a valid GLSL shader or ARB program ID. For some reason,
shader_time used -1 instead...so we had code to detect 0, then override
it to -1.
We can just delete that.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_program.c | 14 ++
1 file changed, 6 insertio
Jason noticed that shader_time was bumping the reference count on the
gl_shader_program and gl_program structures, in code called during
compilation.
Not only were these never unreferenced, but it meant fragment shaders
might be referenced twice (SIMD8 and SIMD16)...or only once.
We don't actuall
On 09.04.2015 20:27, eveli...@kemper.freedesktop.org (Emil Velikov) wrote:
> @@ -2740,7 +2741,45 @@ static int si_generate_gs_copy_shader(struct si_screen
> *sscreen,
> return r;
> }
>
> -int si_shader_create(struct si_screen *sscreen, struct si_shader *shader)
> +static void si_dump_key(
On 12.04.2015 04:11, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/gallium/drivers/radeon/radeon_setup_tgsi_llvm.c | 6 ++
> 1 file changed, 6 insertions(+)
>
> diff --git a/src/gallium/drivers/radeon/radeon_setup_tgsi_llvm.c
> b/src/gallium/drivers/radeon/radeon_setup_tgsi_llvm.c
>
77 matches
Mail list logo
