On 09/06/2013 05:57 PM, Matt Turner wrote:
Because why doesn't GLSL allow you to do this already?
---
docs/specs/MESA_shader_integer_mix.spec | 125
src/glsl/builtin_functions.cpp | 15
src/glsl/glcpp/glcpp-parse.y| 3 +
src/glsl/
This series implements a small extension that adds int, uint, and bool
as allowable types to the mix() built-in function. We hope to make it
an EXT extension soon.
I found myself wanting a per-component conditional-select operation
when implementing the ldexp() and frexp() built-ins, and this
exte
Could be supported on Gen5 once it supports GLSL 1.30.
---
src/mesa/drivers/dri/i965/intel_extensions.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/mesa/drivers/dri/i965/intel_extensions.c
b/src/mesa/drivers/dri/i965/intel_extensions.c
index 655467c..6f9955d 100644
--- a/src/mesa/dr
Because why doesn't GLSL allow you to do this already?
---
docs/specs/MESA_shader_integer_mix.spec | 125
src/glsl/builtin_functions.cpp | 15
src/glsl/glcpp/glcpp-parse.y| 3 +
src/glsl/glsl_parser_extras.cpp | 1 +
src/glsl/
---
src/glsl/builtin_functions.cpp | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/src/glsl/builtin_functions.cpp b/src/glsl/builtin_functions.cpp
index 2a3cfcf..7b7bae6 100644
--- a/src/glsl/builtin_functions.cpp
+++ b/src/glsl/builtin_functions.cpp
@@ -2223,1
---
src/mesa/drivers/dri/i965/brw_fs_channel_expressions.cpp | 1 +
src/mesa/drivers/dri/i965/brw_fs_visitor.cpp | 6 ++
src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp | 6 ++
3 files changed, 13 insertions(+)
diff --git a/src/mesa/drivers/dri/i965/brw_fs_channel_exp
It's a ?: that operates per-component on vectors. Will be used in
upcoming lowering pass for ldexp and the implementation of frexp.
cond_sel(selector, a, b):
per-component result = selector ? a : b
---
src/glsl/ir.cpp| 2 ++
src/glsl/ir.h
278372b47e4db8a022d57f60302eec74819e9341 added the uninitialized pointer
field gl_query_object:Label. A free of this pointer resulted in a crash.
This patch fixes piglit regressions with swrast introduced by
6d8dd59cf53d2f47b817d79204a52bb3a46e8c77.
Bugzilla: https://bugs.freedesktop.org/show_bug
On 09/06/2013 04:40 PM, Kenneth Graunke wrote:
Paused transform feedback objects may refer to a program other than the
current program. If any active objects refer to a program, LinkProgram
must reject the request to relink.
The code to detect this is ugly since _mesa_HashWalk is awkward to use
Paused transform feedback objects may refer to a program other than the
current program. If any active objects refer to a program, LinkProgram
must reject the request to relink.
The code to detect this is ugly since _mesa_HashWalk is awkward to use,
but unfortunately we can't use hash_table_forea
This is actually a pretty important error condition: otherwise, you
could set up transform feedback with one program, and resume it with
a program that generates a completely different set of outputs.
Fixes a subcase of Piglit's new ARB_transform_feedback2 api-errors test.
Signed-off-by: Kenneth
The next few patches will use this for API error checking.
All of the drivers appear to CALLOC_STRUCT transform feedback objects,
so this should be properly NULL initialized on creation.
Signed-off-by: Kenneth Graunke
---
src/mesa/main/mtypes.h| 6 ++
src/mesa/main/transformfeed
Fixes a subcase of Piglit's new ARB_transform_feedback2 api-errors test.
Signed-off-by: Kenneth Graunke
---
src/mesa/main/transformfeedback.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/src/mesa/main/transformfeedback.c
b/src/mesa/main/transformfeedback.c
index 3f8a7f4..eca0
On Sat, Aug 31, 2013 at 12:10:52AM +0200, Marek Olšák wrote:
> The Gallium changes look good to me.
With Mareks OK on the gallium side changes, I'll go and commit this.
The intel changes look good and the DRI driver interface changes do
what we need.
thanks,
Kristian
> Marek
>
> On Thu, Aug 15,
https://bugs.freedesktop.org/show_bug.cgi?id=69047
Vinson Lee changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
This is now handled in the winsys.
Signed-off-by: Alex Deucher
---
src/gallium/drivers/r600/r600_pipe.c | 9 -
1 file changed, 9 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_pipe.c
b/src/gallium/drivers/r600/r600_pipe.c
index 2be5910..c684e6b 100644
--- a/src/gallium/drivers
This aligns the gfx, compute, and dma IBs to 8 DW boundries.
This aligns the the IB to the fetch size of the CP for optimal
performance. Additionally, r6xx hardware requires at least 4
DW alignment to avoid a hw bug. This also aligns the DMA
IBs to 8 DW which is required for the DMA engine. This
Kenneth Graunke writes:
> These are largely based on the similar fields in brw->wm.
>
> Signed-off-by: Kenneth Graunke
> ---
> src/mesa/drivers/dri/i965/brw_context.h | 7 +++
> 1 file changed, 7 insertions(+)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_context.h
> b/src/mesa/drivers/dr
On Fri, Sep 6, 2013 at 4:04 PM, Marek Olšák wrote:
> The CP_STRMOUT_CNTL register was moved again.
Reviewed-by: Alex Deucher
> ---
> src/gallium/drivers/radeon/r600_cs.h | 14 +
> src/gallium/drivers/radeon/r600_streamout.c | 46
>
> src/gallium/d
https://bugs.freedesktop.org/show_bug.cgi?id=69053
Priority: medium
Bug ID: 69053
Assignee: mesa-dev@lists.freedesktop.org
Summary: Account request
Severity: normal
Classification: Unclassified
OS: All
Reporter: joh
On 09/06/2013 01:39 PM, Vinson Lee wrote:
278372b47e4db8a022d57f60302eec74819e9341 added the uninitialized pointer
field gl_query_object:Label. A free of this pointer resulted in a crash.
This patch fixes piglit regressions with swrast introduced by
6d8dd59cf53d2f47b817d79204a52bb3a46e8c77.
Bug
For the series:
Reviewed-by: Marek Olšák
Marek
On Fri, Sep 6, 2013 at 11:00 PM, Alex Deucher wrote:
> This aligns the gfx, compute, and dma IBs to 8 DW boundries.
> This aligns the the IB to the fetch size of the CP for optimal
> performance. Additionally, r6xx hardware requires at least 4
> D
The CP_STRMOUT_CNTL register was moved again.
---
src/gallium/drivers/radeon/r600_cs.h | 14 +
src/gallium/drivers/radeon/r600_streamout.c | 46
src/gallium/drivers/radeon/r600d_common.h| 6
src/gallium/drivers/radeonsi/radeonsi_pipe.c | 2 +
https://bugs.freedesktop.org/show_bug.cgi?id=69047
--- Comment #1 from Vinson Lee ---
Looks like the field ql_query_object:Label was not initialized.
(gdb) frame 5
#5 0x7fd3839867e4 in _mesa_delete_query (ctx=0x7fd387f4a010, q=0x268e090)
at ../../../src/mesa/main/queryobj.c:129
129 fr
https://bugs.freedesktop.org/show_bug.cgi?id=69047
Priority: medium
Bug ID: 69047
Keywords: regression
CC: t_arc...@yahoo.com.au
Assignee: mesa-dev@lists.freedesktop.org
Summary: [swrast] piglit occulsion-query-discard regression
On 09/06/2013 05:58 AM, Erik Faye-Lund wrote:
diff --git a/src/glsl/ir_builder.cpp b/src/glsl/ir_builder.cpp
index e12ae3c..31a457d 100644
--- a/src/glsl/ir_builder.cpp
+++ b/src/glsl/ir_builder.cpp
@@ -264,6 +264,46 @@ abs(operand a)
return expr(ir_unop_abs, a);
}
+ir_expression *neg(ope
Events that are already signalled can be removed from the queue, too.
Signed-off-by: Niels Ole Salscheider
---
src/gallium/state_trackers/clover/core/queue.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/state_trackers/clover/core/queue.cpp
b/src/gallium/stat
While unreferencing fences as early as possible is not a bad idea, this patch
hides the underlying problem. That is, events are never deleted from the
queued_events list of the queue object if their fences are signalled before
the queue is flushed.
I will send a patch that fixes the problem shor
On Fre, 2013-09-06 at 10:36 -0400, Alex Deucher wrote:
> Align IBs to 8 DW to align with the fetch
> size of the CP.
[...]
> diff --git a/src/gallium/drivers/radeonsi/r600_hw_context.c
> b/src/gallium/drivers/radeonsi/r600_hw_context.c
> index 55daa82..fd86383 100644
> --- a/src/gallium/drivers/
> Any reason for this complicated logic, instead of simply:
>
> while (cs->cdw & 0x7)
> cs->buf[cs->cdw++] = 0x8000;
>
>
Ah, that is eloquently terse; I'm going to have to remember that.
Patrick
> Earthling Michel Dänzer | http://www.amd.co
The OpenCL spec says:
"Any blocking commands queued in a command-queue and clReleaseCommandQueue
perform an implicit flush of the command-queue. These blocking commands are
[...] or clWaitForEvents."
Flushing the queue unconditionally also helps to actually clear the
queued_events list of the queu
From: Tom Stellard
Also remove unused data fields from the DS_Load_Helper class.
v2:
- Merge fields for DS_WRITE
---
lib/Target/R600/SIInstrInfo.td| 35 +++
lib/Target/R600/SIInstructions.td | 8
2 files changed, 27 insertions(+), 16 deletions(-)
Align IBs to 8 DW to align with the fetch
size of the CP.
Signed-off-by: Alex Deucher
CC: "9.2"
CC: "9.1"
---
src/gallium/drivers/radeonsi/r600_hw_context.c | 15 +++
1 file changed, 15 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/r600_hw_context.c
b/src/gallium/driver
IBs need to be a multiple of 4 dwords on r6xx asics
to avoid a hw bug. Also align IBs to 8 DW to align
with the fetch size of the CP.
v2: apply to all asics, increase alignment to 8
Signed-off-by: Alex Deucher
CC: "9.2"
CC: "9.1"
---
src/gallium/drivers/r600/r600_hw_context.c | 12 +++
> diff --git a/src/glsl/ir_builder.cpp b/src/glsl/ir_builder.cpp
> index e12ae3c..31a457d 100644
> --- a/src/glsl/ir_builder.cpp
> +++ b/src/glsl/ir_builder.cpp
> @@ -264,6 +264,46 @@ abs(operand a)
> return expr(ir_unop_abs, a);
> }
>
> +ir_expression *neg(operand a)
> +{
> + return expr(ir
On Thu, Sep 5, 2013 at 9:57 PM, Chia-I Wu wrote:
> On Thu, Sep 5, 2013 at 5:12 PM, Chris Forbes wrote:
>> A possible explanation for the perf change is that Xonotic uses
>> anisotropic filtering at this quality level. Lowering to txl defeats
>> it.
> I had a look at that. gl_sampler->MaxAnisotro
On Thu, Sep 5, 2013 at 11:18 PM, Roland Scheidegger wrote:
> Hmm I don't think the math works out here actually, which may explain
> why it's faster.
> I believe the derivatives need to be transformed to cube coord system
> and I don't see that being done here (this is actually something I
> haven
Am 06.09.2013 04:52, schrieb cee1:
2013/9/6 Jerome Glisse :
On Thu, Sep 05, 2013 at 03:29:52PM -0400, Jerome Glisse wrote:
On Thu, Sep 05, 2013 at 10:14:32PM +0800, Chen Jie wrote:
Hi all,
This thread is about
http://lists.freedesktop.org/archives/dri-devel/2013-April/037598.html.
We recentl
Actually Type 2 packets are handled much faster on the R6xx compared to
most type 3 packets, cause they are handled by the PFP/fetch hw and
doesn't need to be forwarded to the ME.
Christian.
Am 06.09.2013 02:31, schrieb Dominik Behr:
0x8000 is Type 2 NOP.
You could make it a little better
On Don, 2013-09-05 at 14:36 -0400, Tom Stellard wrote:
> From: Tom Stellard
>
> Also remove unused data fields from the DS_Load_Helper class.
[...]
> diff --git a/lib/Target/R600/SIInstructions.td
> b/lib/Target/R600/SIInstructions.td
> index 31a5ad2..94dcf2c 100644
> --- a/lib/Target/R600/SII
On Don, 2013-09-05 at 14:36 -0400, Tom Stellard wrote:
>
> diff --git a/lib/Target/R600/SILowerControlFlow.cpp
> b/lib/Target/R600/SILowerControlFlow.cpp
> index a6c43bb..5a8836c 100644
> --- a/lib/Target/R600/SILowerControlFlow.cpp
> +++ b/lib/Target/R600/SILowerControlFlow.cpp
[...]
> @@ -487
On Thu, Sep 5, 2013 at 11:13 AM, Christophe Jarry
wrote:
> Then, what DRI driver should I use?
You shouldn't need one. The R600 gallium driver is the only thing
available for your hardware and is what you should use.
To disable all the classic dri drivers, use --with-dri-drivers=
___
https://bugs.freedesktop.org/show_bug.cgi?id=69013
Timothy Arceri changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://bugs.freedesktop.org/show_bug.cgi?id=69013
Priority: medium
Bug ID: 69013
Keywords: regression
CC: t_arc...@yahoo.com.au
Assignee: mesa-dev@lists.freedesktop.org
Summary: piglit arb_debug_output-api_error regression
> I unfortunately cannot make any sense out of your build error, but if
> you're building Mesa to run on an RS780E then you should just use
>
> --with-gallium-drivers=r600
Ok, I will try this.
> and maybe swrast, but the radeon driver specified in
> --with-dri-drivers is for old R100 cards, an
Hi all,
This thread is about
http://lists.freedesktop.org/archives/dri-devel/2013-April/037598.html.
We recently find some interesting thing about UVD based playback on
loongson 3a plaform, and also find a way to fix the problem.
First, we find memcpy in [mesa]src/gallium/drivers/radeon/radeon_u
46 matches
Mail list logo
