v2:
Fix adding parameter attributes with LLVM < 4.0.
v3:
Fix typo.
Fix parameter index.
Add a gallivm enum for function attributes.
---
src/gallium/auxiliary/draw/draw_llvm.c| 6 +-
src/gallium/auxiliary/gallivm/lp_bld_intr.c | 70 ++-
src/gallium/au
On Wed, Nov 09, 2016 at 11:45:38PM +0100, Roland Scheidegger wrote:
> Am 09.11.2016 um 16:22 schrieb Tom Stellard:
> > v2:
> > Fix adding parameter attributes with LLVM < 4.0.
> >
> > v3:
> > Fix typo.
> > Fix parameter index.
> &
On Fri, Nov 11, 2016 at 02:00:26PM +0100, Laurent Carlier wrote:
Reviewed-by: Tom Stellard
> pretty trivial fix
> ---
> src/gallium/state_trackers/clover/llvm/codegen/bitcode.cpp | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/src/gallium/state_trackers/c
On Mon, Nov 14, 2016 at 01:44:18PM +0100, Dieter Nützel wrote:
> Tested-by: Dieter Nützel
>
> Thanks Vedran!
>
Pushed, thanks!
-Tom
> Dieter
>
> Am 14.11.2016 12:17, schrieb Vedran Miletić:
> > ---
> > src/gallium/state_trackers/clover/llvm/codegen/bitcode.cpp | 10
> > --
> > 1 fi
---
src/gallium/drivers/radeonsi/si_shader.c | 197 ++-
1 file changed, 143 insertions(+), 54 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index 0410a32..306e12f 100644
--- a/src/gallium/drivers/radeonsi
---
src/gallium/drivers/radeonsi/si_shader.c | 69 +++-
1 file changed, 50 insertions(+), 19 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b/src/gallium/drivers/radeonsi/si_shader.c
index 306e12f..ee4fe2f 100644
--- a/src/gallium/drivers/radeonsi/
mOnly, but I think
as long as it has non-pointer arguments this combination behaves the same as
ReadNone, which would be incorrect.
-Tom
> Cheers,
> Nicolai
>
> On 16.11.2016 03:14, Tom Stellard wrote:
> >---
> > src/gallium/drivers/radeonsi/si_shader.c | 69
> >
cd724208d3e1e3307f84a794f2c1fc83b69ccf8a added a call to
_mesa_lock_debug_state(ctx) but wasn't unlocking the debug state.
This fixes a hang in glsl-fs-loop piglit test with MESA_DEBUG=context.
v2:
- Remove unrelated changes.
---
src/mesa/main/debug_output.c | 4 +++-
1 file changed, 3 inserti
cd724208d3e1e3307f84a794f2c1fc83b69ccf8a added a call to
_mesa_lock_debug_state(ctx) but wasn't unlocking the debug state.
This fixes a hang in glsl-fs-loop piglit test with MESA_DEBUG=context.
---
src/gallium/drivers/radeonsi/si_pipe.c | 8 +---
src/mesa/main/debug_output.c | 5 +++
On Sat, Nov 19, 2016 at 01:09:00AM +0100, Marek Olšák wrote:
> On Wed, Nov 16, 2016 at 4:38 PM, Tom Stellard wrote:
> > On Wed, Nov 16, 2016 at 11:13:45AM +0100, Nicolai Hähnle wrote:
> >> Have you looked at the shader-db impact?
> >>
> >
> > shader-d
This prevents LLVM from using sext instructions for local memory offsets
and allows the backend to fold immediate offsets into the instruction.
This also prevents some incorrect code generation for ptrtoint and
inttoptr instructions.
---
src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c | 6 +++
Also add a need_range_checks parameter to this function, which can be
set to false to enable some additional optimizations. Currently, this
will cause the compiler to emit the llvm.SI.buffer.load.dword intrinsic
instead of llvm.amdgcn.buffer.load. Eventually, this information
will be passed to LL
On Mon, Dec 19, 2016 at 02:04:05PM -0500, Alex Deucher wrote:
> From: Junwei Zhang
>
> Signed-off-by: Junwei Zhang
> Reviewed-by: Nicolai Hähnle
> Acked-by: Christian König
> ---
> src/amd/addrlib/r800/ciaddrlib.cpp| 3 ++-
> src/amd/addrlib/r800/ciaddrlib.h | 1 +
>
On Sun, Jul 03, 2016 at 05:51:09PM -0700, Francisco Jerez wrote:
> This series is the result of a long back and forth of patches between
> Serge and me. Most of the changes are already reviewed by either of
> us, but I'm sending them again over the mailing list just in case
> somebody else wants t
On Tue, Jul 05, 2016 at 11:36:03AM +0200, Marek Olšák wrote:
> From: Marek Olšák
>
> RadeonSI stats: Mostly 0% difference, but Valley shows a small improvement:
>
Do you know which intrinsic this made a difference for? I'm guessing
this is required for all the intrinsics defined in the backend
On Wed, Jul 06, 2016 at 11:55:03PM +0200, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> Only setting R600_DEBUG doesn't set any debug callback. Conversely, the debug
> callback is only called when R600_DEBUG is set.
I don't get any output from shader-db with this patch.
-Tom
> ---
> src/ga
re using vtx#3.
With that change:
Reviewed-by: Tom Stellard
> ---
> Hi,
>
> I ran into problem when using VTX_READ from constant buffer would work only
> for 0 index. The LLVM code implied that it should work (or maybe they
> considered constant offsets only), but I could
On Tue, Jul 12, 2016 at 10:52:35PM +0200, Marek Olšák wrote:
> From: Marek Olšák
>
> This allows moving the loads arbitrarily in the Sinking pass.
>
> 26002 shaders in 14643 tests
> Totals:
> SGPRS: 2080160 -> 2080160 (0.00 %)
> VGPRS: 798875 -> 797826 (-0.13 %)
> Spilled SGPRs: 108485 -> 79165
On Tue, Jul 12, 2016 at 10:52:36PM +0200, Marek Olšák wrote:
> From: Marek Olšák
>
Reviewed-by: Tom Stellard
> no change in generated code thanks to dereferenceable(n)
> ---
> src/gallium/drivers/radeonsi/si_shader.c | 17 +
> 1 file changed, 5 insertion
On Wed, Jul 13, 2016 at 03:20:55PM -0400, Tom Stellard wrote:
> On Tue, Jul 12, 2016 at 10:52:35PM +0200, Marek Olšák wrote:
> > From: Marek Olšák
> >
> > This allows moving the loads arbitrarily in the Sinking pass.
> >
> > 26002 shaders in 14643 tests
> >
There was a patch committed to clang to remove unnecessary includes from
header files, so we now need to explicitly include
clang/Lex/PreprocessorOptions.h
---
src/gallium/state_trackers/clover/llvm/invocation.cpp | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/gallium/state_trackers/clo
The build was failing because the official CL headers have a few defines, like:
Which have the same name as some class members of clang's OpenCLOptions class.
If we include the cl headers first, this breaks the build because the member
names of this class are replaced by the literal 1.
---
.../st
There was a patch committed to clang to remove unnecessary includes from
header files, so we now need to explicitly include
clang/Lex/PreprocessorOptions.h
v2:
- Use <> instead of "" for the include path.
---
src/gallium/state_trackers/clover/llvm/invocation.cpp | 2 ++
1 file changed, 2 insert
The build was failing because the official CL headers have a few defines, like:
\# define cl_khr_gl_sharing 1
Which have the same name as some class members of clang's OpenCLOptions class.
If we include the cl headers first, this breaks the build because the member
names of this class are replace
This patche switches non-TGSI compute shaders over to using the HSA
ABI described here:
https://github.com/RadeonOpenCompute/ROCm-Docs/blob/master/AMDGPU-ABI.md
The HSA ABI provides a much cleaner interface for compute shaders and allows
us to share more code in the compiler with the HSA stack.
---
src/gallium/drivers/radeonsi/si_compute.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_compute.c
b/src/gallium/drivers/radeonsi/si_compute.c
index 5a40286..949ab1a 100644
--- a/src/gallium/drivers/radeonsi/si_compute.c
+++ b/src/gallium/drivers/radeon
On Wed, Sep 25, 2013 at 08:14:43PM +0200, Marek Olšák wrote:
> From: Marek Olšák
>
> This doesn't fix any known issue (I haven't run piglit with this yet),
> but the code was obviously completely wrong. It looks like copy-pasted from
> CMP.
Reviewed-by: Tom Stella
On Mon, Sep 23, 2013 at 09:09:14PM +0200, mar...@gmail.com wrote:
> From: Marek Olšák
>
> This fixes piglit:
> - shaders/glsl-fs-texture2d-masked
> - shaders/glsl-fs-texture2d-masked-4
>
> Signed-off-by: Marek Olšák
> ---
> lib/Target/R600/SIISelLowering.cpp | 27 +--
>
On Sun, Sep 22, 2013 at 09:29:27PM +0100, Emil Velikov wrote:
> libllvmradeon.la is available whenever NEED_RADEON_LLVM is set, using
> R600_NEED_RADEON_GALLIUM is rather ambiguous and unnecessary. Drop it
> in favour of NEED_RADEON_LLVM.
>
Reviewed-by: Tom Stellard
> Signed-off-b
On Sun, Sep 22, 2013 at 09:29:24PM +0100, Emil Velikov wrote:
> From: Johannes Obermayr
>
> libdricommon.la is available whenever a non swrast driver is built.
> All the classic dri drivers make use of the prebuild library but all
> of the gallium ones rebuild it explicitly.
>
> While we're here
On Sun, Sep 22, 2013 at 09:29:26PM +0100, Emil Velikov wrote:
Reviewed-by: Tom Stellard
> Signed-off-by: Emil Velikov
> ---
> src/gallium/drivers/radeon/Makefile.am | 3 ---
> 1 file changed, 3 deletions(-)
>
> diff --git a/src/gallium/drivers/radeon/Makefile.am
> b/src/g
On Sun, Sep 22, 2013 at 09:29:28PM +0100, Emil Velikov wrote:
> Signed-off-by: Emil Velikov
As long as you have build tested these with both automake and scons and
are prepared to deal with any fallout once they are committed.
Patches 5 through 29 are:
Reviewed-by: Tom Stellard
> ---
On Sat, Sep 28, 2013 at 04:42:34AM +0100, Emil Velikov wrote:
> On 28/09/13 04:36, Tom Stellard wrote:
> > On Sun, Sep 22, 2013 at 09:29:24PM +0100, Emil Velikov wrote:
> >> From: Johannes Obermayr
> >>
> >> libdricommon.la is available whenever a non swrast d
On Sat, Sep 28, 2013 at 03:01:15PM +0100, Emil Velikov wrote:
> On 28/09/13 04:48, Tom Stellard wrote:
> > On Sun, Sep 22, 2013 at 09:29:28PM +0100, Emil Velikov wrote:
> >> Signed-off-by: Emil Velikov
> >
> > As long as you have build tested these with both
SAMPLE_LB">;
> def TEX_SAMPLE_C_LB : R600_TEX <0x1A, "TEX_SAMPLE_C_LB">;
> def TEX_LD : R600_TEX <0x03, "TEX_LD">;
> +def TEX_LDPTR : R600_TEX <0x03, "TEX_LDPTR"> {
Can you use a named field here so it is easier to read:
field bi
On Mon, Sep 30, 2013 at 11:46:26PM +0200, Vincent Lejeune wrote:
Could you go through and wrap some on the long lines that you are adding
in the other patches.
With that fix, the entire series is:
Reviewed-by: Tom Stellard
> ---
> src/gallium/drivers/radeon/radeon_setup_tgsi_llvm
On Mon, Sep 30, 2013 at 07:55:34AM -0700, Tom Stellard wrote:
> On Sat, Sep 28, 2013 at 03:01:15PM +0100, Emil Velikov wrote:
> > On 28/09/13 04:48, Tom Stellard wrote:
> > > On Sun, Sep 22, 2013 at 09:29:28PM +0100, Emil Velikov wrote:
> > >> Signed-off-by: Emil Ve
From: Tom Stellard
---
src/gallium/drivers/radeon/r600_pipe_common.c | 4
src/gallium/drivers/radeonsi/radeonsi_shader.c | 4 ++--
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/radeon/r600_pipe_common.c
b/src/gallium/drivers/radeon/r600_pipe_common.c
On Fri, Oct 04, 2013 at 10:01:32AM -0700, Owen Anderson wrote:
>
> On Oct 3, 2013, at 5:36 PM, Tom Stellard wrote:
>
> > Unfortunately, we cannot enable the machine verifier on tests that have
> > branches, because the way IF and ELSE instructions are selected leav
ake.inc if you want to.
Either way, the series is
Reviewed-by: Tom Stellard
> ---
> src/gallium/drivers/r600/Makefile.am | 3 ++-
> src/gallium/drivers/radeon/Makefile.am | 8 +---
> src/gallium/drivers/radeonsi/Makefile.am | 4 +++-
> src/gallium/targets/r300
On Mon, Oct 07, 2013 at 01:05:15PM +0200, Marek Olšák wrote:
> From: Marek Olšák
>
Reviewed-by: Tom Stellard
> ---
> configure.ac | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/configure.ac b/configure.ac
> index 559c9a3..6158c6e 100
On Fri, Oct 04, 2013 at 09:22:30PM +0200, Laurent Carlier wrote:
> http://llvm.org/viewvc/llvm-project?view=revision&revision=191922
Pushed, thanks!
> ---
> src/gallium/state_trackers/clover/llvm/invocation.cpp | 6 +-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/src/g
On Sat, Sep 28, 2013 at 03:46:18PM +0100, Emil Velikov wrote:
> The function xf86GetEntityInfo() retrieves the entity rather than
> doing any changes. Remove this no-op code.
>
> Signed-off-by: Emil Velikov
Reviewed-by: Tom Stellard
> ---
> src/gallium/targets/r600/xorg/xor
On Sat, Sep 28, 2013 at 03:46:21PM +0100, Emil Velikov wrote:
> Use PKG_CHECK_MODULE over requesting the user to setup the
> option at configure time. Drop unused EXPAT_INCLUDE and add
> EXPAT_CFLAGS to all classic/dri and gallium/dri targets.
>
> Signed-off-by: Emil Velikov
> ---
> configure.ac
1b4cc4e
> --- /dev/null
> +++ b/test/CodeGen/R600/llvm.SI.sample-masked.ll
> @@ -0,0 +1,93 @@
> +;RUN: llc < %s -march=r600 -mcpu=verde | FileCheck %s
> +
> +; CHECK: @v1
I always forget to mention this, but we should use CHECK-LABEL: when
checking the function names. CHECK-LAB
On Thu, Oct 10, 2013 at 12:04:16AM +0200, Vincent Lejeune wrote:
What is the purpose of this change?
-Tom
> ---
> lib/Target/R600/AMDGPUCallingConv.td | 7 ++-
> lib/Target/R600/SIISelLowering.cpp | 3 ++-
> 2 files changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/lib/Target/R60
On Thu, Oct 10, 2013 at 11:29:17AM -0700, Matt Arsenault wrote:
> On 10/10/2013 10:55 AM, Tom Stellard wrote:
> > Hi,
> >
> > The attached patches simplify the handling of OpenCL private memory
> > space for VLIW4/VLIW5 GPUs and should fix a crash with pyrit on r600
Hi,
The attached patches fix a crash in the AMDILCFGStructurizer. The first
patch does some code cleanup and the second patch actually fixes the
crash.
-Tom
>From 8a95058178eba343cb6ee408677b42329e7069ed Mon Sep 17 00:00:00 2001
From: Tom Stellard
Date: Fri, 11 Oct 2013 09:18:22 -0700
Subj
From: Tom Stellard
This is required in order for clang to correctly handle the OpenCL C
barrier() builtin which has the following restrictions acording to
the OpenCL 1.1 Specification:
If barrier is inside a conditional statement, then all work-items must
enter the conditional if any work-item
From: Tom Stellard
---
src/gallium/drivers/r600/compute_memory_pool.c | 8 +---
src/gallium/drivers/r600/evergreen_compute.c | 4 ++--
2 files changed, 7 insertions(+), 5 deletions(-)
diff --git a/src/gallium/drivers/r600/compute_memory_pool.c
b/src/gallium/drivers/r600
From: Tom Stellard
Reading the special OQAP register pops the top value off the LDS
input queue and returns it to the instruction. This queue is
invalidated at the end of an ALU clause and leaving values in the queue
can lead to GPU hangs. This means that if we load a value into the queue,
we
On Fri, Oct 11, 2013 at 02:09:57PM +0100, Emil Velikov wrote:
> On 07/10/13 18:53, Emil Velikov wrote:
> > On 07/10/13 16:48, Tom Stellard wrote:
> >> On Sat, Sep 28, 2013 at 03:46:21PM +0100, Emil Velikov wrote:
> >>> Use PKG_CHECK_MODULE over requesting the use
On Thu, Oct 10, 2013 at 08:06:48PM -0500, Jay Cornwall wrote:
> With code dump enabled LLVM may generate disassembly during compilation.
> Show this disassembly when available and prefer it to SI bytecode dump.
>
I've pushed this and the LLVM patch. Thanks!
-Tom
> Signed-off-by: Jay Cornwall
Hi,
The attached patches fix various bugs in the handling of kernel arguments
in the R600 backend. The main problem with the old implementations was
that it disagreed with clover on type size and alignment, so vector
arguments were not being loaded correctly into the kernel.
This series depends
On Tue, Oct 15, 2013 at 01:59:55PM -0400, Alex Deucher wrote:
> Did you forget the attachments?
>
Yes I did. Here they are.
-Tom
> On Tue, Oct 15, 2013 at 12:40 PM, Tom Stellard wrote:
> > Hi,
> >
> > The attached patches fix various bugs in the handling of kerne
On Mon, Oct 21, 2013 at 10:02:12PM +0200, Vincent Lejeune wrote:
Can you add an explanation to the commit messages for both patches about
what was wrong with the old code?
Thanks,
Tom
> ---
> src/gallium/drivers/r600/r600_llvm.c | 29 +
> 1 file changed, 29 insertion
From: Tom Stellard
Without DataLayout, a lot of optimization passes aren't run and the ones
that are don't work as well.
---
src/gallium/drivers/radeon/radeon_llvm_util.c | 4
1 file changed, 4 insertions(+)
diff --git a/src/gallium/drivers/radeon/radeon_llvm_util.c
b/src/galli
ailed.
Any suggestions on where I should but it?
-Tom
On Mon, Oct 21, 2013 at 12:40:28PM -0700, Vincent Lejeune wrote:
>
>
>
>
> - Mail original -
> > De : Tom Stellard
> > À : llvm-comm...@cs.uiuc.edu
> > Cc : mesa-dev@lists.freedesktop.org; Tom Stellar
From: Tom Stellard
---
src/gallium/drivers/radeonsi/radeonsi_compute.c | 15 ---
1 file changed, 12 insertions(+), 3 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/radeonsi_compute.c
b/src/gallium/drivers/radeonsi/radeonsi_compute.c
index d3f5a23..f047d5d 100644
--- a/src
From: Tom Stellard
---
lib/Target/R600/AMDGPU.td| 2 ++
lib/Target/R600/AMDGPUAsmPrinter.cpp | 16 ++--
lib/Target/R600/AMDGPUCallingConv.td | 4 ++--
lib/Target/R600/AMDGPUSubtarget.h| 3 ++-
lib/Target/R600/Processors.td| 6 +++---
lib/Target/R600
Pushed, thanks.
-Tom
On Tue, Oct 22, 2013 at 02:15:01AM +0200, Marek Olk wrote:
> From: Marek Olk
>
> This fixes piglit:
> - shaders/glsl-fs-texture2d-masked
> - shaders/glsl-fs-texture2d-masked-4
>
> Signed-off-by: Marek Ol????k
> Reviewed-by: Tom Stellard
On Wed, Oct 23, 2013 at 09:47:30AM -0500, Jay Cornwall wrote:
> On 2013-10-22 21:05, Tom Stellard wrote:
>
> > ---
> > src/gallium/drivers/radeonsi/radeonsi_compute.c | 15 ---
> > 1 file changed, 12 insertions(+), 3 deletions(-)
> >
> [..]
&
On Wed, Oct 23, 2013 at 09:31:53AM +0200, Michel D??nzer wrote:
> On Die, 2013-10-22 at 22:07 -0400, Tom Stellard wrote:
> >
> > diff --git a/lib/Target/R600/AMDGPU.td b/lib/Target/R600/AMDGPU.td
> > index a722f55..f63617f 100644
> > --- a/lib/Target/R600/AMDGPU.t
From: Tom Stellard
This makes it possible to use clover with statically linked LLVM.
---
src/gallium/targets/pipe-loader/Makefile.am | 4
src/gallium/targets/pipe-loader/pipe.link | 3 +++
2 files changed, 7 insertions(+)
create mode 100644 src/gallium/targets/pipe-loader/pipe.link
From: Tom Stellard
This works now that pipe_*.so is no longer exporting LLVM symbols.
---
configure.ac | 6 --
1 file changed, 6 deletions(-)
diff --git a/configure.ac b/configure.ac
index dfa35b4..81403ee 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1530,12 +1530,6 @@ AC_ARG_WITH
Hi,
The attached patches introduce linker scripts to the pipe-loader and
egl-static targets. The linker scripts prevents these targets from
exporting LLVM (and other) symbols that they shouldn't be. This fixes
several crashes in the radeon drivers when statically linking LLVM.
With these patche
From: Tom Stellard
This fixes a crash in glamor when mesa links against static LLVM.
---
src/gallium/targets/egl-static/Makefile.am | 4
src/gallium/targets/egl-static/egl.link| 3 +++
2 files changed, 7 insertions(+)
create mode 100644 src/gallium/targets/egl-static/egl.link
diff
From: Tom Stellard
This makes it possible to use clover with statically linked LLVM.
v2:
- Inline LINKER_SCRIPT variable
---
src/gallium/targets/pipe-loader/Makefile.am | 2 ++
src/gallium/targets/pipe-loader/pipe.link | 3 +++
2 files changed, 5 insertions(+)
create mode 100644 src
From: Tom Stellard
This fixes a crash in glamor when mesa links against static LLVM.
v2:
- Inline LINKER_SCRIPT variable
---
src/gallium/targets/egl-static/Makefile.am | 2 ++
src/gallium/targets/egl-static/egl.link| 3 +++
2 files changed, 5 insertions(+)
create mode 100644 src/gallium
From: Tom Stellard
---
lib/Target/R600/AMDGPUISelLowering.cpp | 1 +
test/CodeGen/R600/llvm.sqrt.ll | 54 ++
2 files changed, 55 insertions(+)
create mode 100644 test/CodeGen/R600/llvm.sqrt.ll
diff --git a/lib/Target/R600/AMDGPUISelLowering.cpp
b/lib
From: Tom Stellard
v2:
- Fix LDS size calculation
---
lib/Target/R600/AMDGPU.td| 2 ++
lib/Target/R600/AMDGPUAsmPrinter.cpp | 16 ++--
lib/Target/R600/AMDGPUCallingConv.td | 4 ++--
lib/Target/R600/AMDGPUSubtarget.h| 3 ++-
lib/Target/R600/Processors.td
On Wed, Oct 23, 2013 at 06:41:56PM -0500, Aaron Watry wrote:
> The maximum workgroup size for a given kernel is based on the
> capabilities of the device that it's being run on. Previously,
> we were just returning the maximum value of a size_t which is
> obviously wrong.
>
> This patch uses the d
On Thu, Oct 24, 2013 at 02:55:36AM -0400, Yomi Ogunwumi wrote:
> Hello everyone.
>
> I was reading through the Help Wanted page and I was wondering if 6.
> Automatic Testing still needed to be done.
>
> The three continuous integration systems currently I'm looking at are :
>
> • Jenkins
> • Bui
On Thu, Oct 24, 2013 at 09:14:49AM -0600, Brian Paul wrote:
> On 10/23/2013 01:33 PM, David Heidelberger wrote:
> > link to similiar bug: https://bugs.freedesktop.org/show_bug.cgi?id=70804
> >
> > ---
> > src/gallium/state_trackers/clover/core/kernel.cpp | 7 ---
> > 1 file changed, 4 insert
On Wed, Oct 23, 2013 at 6:28 PM, Tom Stellard wrote:
> > From: Tom Stellard
> >
> > ---
> > lib/Target/R600/AMDGPUISelLowering.cpp | 1 +
> > test/CodeGen/R600/llvm.sqrt.ll | 54
> > ++
> > 2 files changed, 55 in
Kai Wasserbäch
This looks OK to me, which intrinsic names were too long?
Reviewed-by: Tom Stellard
> ---
> src/gallium/drivers/radeonsi/radeonsi_shader.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/gallium/drivers/radeonsi/radeonsi_shader
From: Tom Stellard
The ICD loader should be responsible for installing headers.
---
src/gallium/state_trackers/clover/Makefile.am | 21 +++--
1 file changed, 11 insertions(+), 10 deletions(-)
diff --git a/src/gallium/state_trackers/clover/Makefile.am
b/src/gallium
nly not compile on 4.7.0, if I recall corectly Tom Stellard
> said it compile on 4.7.1. This patch is tested.
I was mistaken. 4.7.3 is the oldest version that will work.
This patch doesn't apply for me. It looks like your mail client is
wrapping long lines, can you resend?
Thanks,
Tom
From: Tom Stellard
This results in huge performance improvements for applications like
pyrit, which depend on the implementation to determine the optimal
work group size.
---
src/gallium/state_trackers/clover/api/kernel.cpp | 9 -
src/gallium/state_trackers/clover/core/device.cpp | 19
From: Tom Stellard
This results in huge performance improvements for applications like
pyrit, which depend on the implementation to determine the optimal
work group size.
---
src/gallium/state_trackers/clover/api/kernel.cpp | 9 -
src/gallium/state_trackers/clover/core/device.cpp | 19
From: Tom Stellard
You can use the --enable-pipe-loader-render-nodes configure flag to
make the pipe-loader use render nodes for talking with the device.
---
configure.ac | 6 ++
src/gallium/auxiliary/pipe-loader/Makefile.am | 5 ++
.../auxiliary
From: Tom Stellard
DRM_IOCTL_VERSION is not currently a legal ioctl for render nodes, so we
need to hard code the drm version when drmGetVersion() returns NULL.
---
src/gallium/winsys/radeon/drm/radeon_drm_winsys.c | 36 ---
1 file changed, 25 insertions(+), 11 deletions
Both patches are:
Reviewed-by: Tom Stellard
On Wed, Oct 30, 2013 at 04:48:18PM +0100, Vincent Lejeune wrote:
> R600/R700 implementation of tex buffer fetch requires the result of the VFETCH
> instruction to be ANDed with R600_BUFFER_INFO_CONST_BUFFER, and the last
> channel
> to
ll LDS instructions
> that queues a value, I only use LDS_RET_READ in switch case.
>
> Vincent
>
>
>
> - Mail original -
> > De : Tom Stellard
> > À : Vincent Lejeune
> > Cc : "llvm-comm...@cs.uiuc.edu" ;
> > "mesa-dev@lists.fr
On Wed, Oct 30, 2013 at 08:15:43AM +0100, David Herrmann wrote:
> Hi Tom
>
> On Tue, Oct 29, 2013 at 9:00 PM, Tom Stellard wrote:
> > From: Tom Stellard
> >
> > You can use the --enable-pipe-loader-render-nodes configure flag to
> > make the pipe-loader use
On Tue, Oct 29, 2013 at 08:34:16PM +0100, Kai Wasserb??ch wrote:
> Tom Stellard schrieb am 29.10.2013 17:48:
> > On Sun, Oct 27, 2013 at 07:36:07PM +0100, Kai Wasserb??ch wrote:
> >> Fixes a boat load of Piglit tests for me, which crashed like fdo#70913
> >> before.
&
On Wed, Oct 30, 2013 at 08:38:37AM -0500, Aaron Watry wrote:
> Reviewed and Tested-by: Aaron Watry
I have pushed this patch, thanks for testing.
>
> On Tue, Oct 29, 2013 at 11:48 AM, Tom Stellard wrote:
> > From: Tom Stellard
> >
> > The ICD loader should be
, thanks for testing.
> --Aaron
>
> On Tue, Oct 22, 2013 at 11:28 AM, Tom Stellard wrote:
> > From: Tom Stellard
> >
> > Without DataLayout, a lot of optimization passes aren't run and the ones
> > that are don't work as well.
> > ---
> >
e in LLVM 3.4 and newer.
Otherwise, I think people may be confused about why this change was
made.
With that change, the patch is:
Reviewed-by: Tom Stellard
> Signed-off-by: Vinson Lee
> ---
> src/gallium/auxiliary/gallivm/lp_bld_misc.cpp | 2 ++
> 1 file changed, 2 insertions(+)
On Fri, Nov 01, 2013 at 10:25:43AM -0500, Aaron Watry wrote:
> dso_list was added as an argument for createInternalizePass in 3.4, and then
> it was removed again in the same llvm version.
I've pushed this patch, thanks!
-Tom
> ---
> src/gallium/state_trackers/clover/llvm/invocation.cpp | 5 ---
On Sun, Nov 03, 2013 at 06:58:18PM +0100, Vincent Lejeune wrote:
> ---
> src/gallium/drivers/r600/r600_llvm.c | 125
> -
> src/gallium/drivers/r600/r600_shader.c | 2 +
> src/gallium/drivers/radeon/radeon_llvm.h | 1 +
> .../drivers/rade
On Wed, Oct 23, 2013 at 04:26:18PM -0400, Tom Stellard wrote:
> Hi,
>
> The attached patches introduce linker scripts to the pipe-loader and
> egl-static targets. The linker scripts prevents these targets from
> exporting LLVM (and other) symbols that they shouldn't be.
On Wed, Nov 06, 2013 at 10:36:51AM -0600, Aaron Watry wrote:
> ---
> src/gallium/drivers/radeon/radeon_llvm_emit.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/src/gallium/drivers/radeon/radeon_llvm_emit.c
> b/src/gallium/drivers/radeon/radeon_llvm_emit.c
> index 8bf278b..20b3206
These look good, but the indentation seems wrong in patches 2 through 6.
-Tom
On Wed, Nov 06, 2013 at 10:36:49AM -0600, Aaron Watry wrote:
> I decided to have some fun and hooked valgrind up to my 7850 while running
> a few OpenCL tests in piglit. This is the first batch of fixes.
>
> Aaron Watr
On Wed, Nov 06, 2013 at 12:59:09PM -0600, Aaron Watry wrote:
> On Wed, Nov 6, 2013 at 12:15 PM, Tom Stellard wrote:
> > These look good, but the indentation seems wrong in patches 2 through 6.
>
> An artifact of expanding tabs to 4 spaces in the IDE... Although I'd
>
On Mon, Nov 11, 2013 at 03:45:53PM +0100, Vincent Lejeune wrote:
Reviewed-by: Tom Stellard
> ---
> src/gallium/drivers/r600/r600_llvm.c | 119
> +++
> src/gallium/drivers/r600/r600_shader.c | 1 +
> src/gallium/drivers/radeon/radeon_llvm
On Sat, Nov 02, 2013 at 07:00:45PM +, Emil Velikov wrote:
> Signed-off-by: Emil Velikov
Can you make sure make check still works with this patch, if it does,
then:
Reviewed-by: Tom Stellard
> ---
> src/gallium/drivers/r300/Makefile.am | 13 ++---
> src/gallium/
On Sat, Nov 02, 2013 at 07:00:44PM +, Emil Velikov wrote:
> Make automake's subdir-objects work.
> Update includes.
>
> Signed-off-by: Emil Velikov
> ---
> src/gallium/drivers/r300/Makefile.am | 6 +++---
> src/gallium/drivers/r300/ralloc.c| 1 +
> src/gallium/drivers/r30
On Thu, Nov 07, 2013 at 06:08:15PM -0600, Aaron Watry wrote:
> radeon_llvm_compile allocates memory for binary.code, binary.config, or
> neither depending on
> what's being done.
>
> We need to make sure to free that memory after it's no longer needed.
> ---
> src/gallium/drivers/r600/r600_llvm.
600/r600_llvm.c
> @@ -712,6 +712,7 @@ unsigned r600_llvm_compile(
> const char * gpu_family = r600_llvm_gpu_string(family);
> unsigned i;
>
> +memset(&binary, 0, sizeof(struct radeon_llvm_binary));
The indentation looks wrong here, otherwise this patch is:
Reviewe
101 - 200 of 1703 matches
Mail list logo
