On Monday, November 7, 2016 12:28:34 PM PST Eric Anholt wrote:
> NIR bools are ~0 for true, so ((unsigned)a >> 31) != 0 -> ((int)a >> 31).
> ---
> src/compiler/nir/nir_lower_idiv.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/src/compiler/nir/nir_lower_idiv.c
> b/sr
Check for Clang before GCC.
Clang defines __GNUC__ == 4 and __GNUC_MINOR__ == 2 and matches the GCC
check but not the GCC version for trivial destructor.
Fixes: 98ab905af0e0 ("mesa: Define introspection macro to determine
whether a type is trivially destructible.")
Bugzilla: https://bugs.freedesk
On Friday, November 11, 2016 9:41:11 AM PST Juan A. Suarez Romero wrote:
> Do not evaluate spill costs for registers that were already marked as
> no_spill.
> ---
> src/mesa/drivers/dri/i965/brw_vec4_reg_allocate.cpp | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/src
On Sunday, November 13, 2016 9:29:44 AM PST Timothy Arceri wrote:
> On Sat, 2016-11-12 at 13:34 -0800, Jason Ekstrand wrote:
> > ---
> > src/intel/vulkan/gen7_pipeline.c | 29 +---
> > src/intel/vulkan/gen8_pipeline.c | 57 ++-
> > src/intel/vulkan/genX_pipeli
On Sunday, November 13, 2016 10:21:11 AM PST Eduardo Lima Mitev wrote:
> This option is currently ignored when packing compressed textures.
>
> Fixes CTS test (on gen8+):
> * GL45-CTS.gtf32.GL3Tests.packed_pixels.packed_pixels_pixelstore
> ---
> src/mesa/drivers/common/meta.c | 13 ++---
>
Hello Francisco,
next little change needed.
Thanks,
Dieter
make[4]: Entering directory
'/opt/mesa/src/gallium/state_trackers/clover'
CXX llvm/codegen/libclllvm_la-bitcode.lo
llvm/codegen/bitcode.cpp: In function 'std::unique_ptr
clover::llvm::parse_module_library(const clover::module&
https://bugs.freedesktop.org/show_bug.cgi?id=98714
Bug ID: 98714
Summary: Weird whitish spots and green stuff mostly on black.
Product: Mesa
Version: git
Hardware: x86-64 (AMD64)
OS: Linux (All)
Status: NEW
Hello,
These two mails are two proposals to fix
https://bugs.freedesktop.org/show_bug.cgi?id=98632
Either properly fix the code, or just define PATH_MAX and be "done" with
it.
Samuel
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lis
Does using stdint.h help?
On 11/11/2016 07:48 AM, Brian Paul wrote:
> inttypes.h doesn't always define the PRIx64 macro when compiling C++
> code with MinGW. For that case, simply define PRIx64 to be "lx".
> This generates a format/argument type mismatch warning. And using
> "llx" generates othe
GNU/Hurd does not define PATH_MAX since it doesn't have such arbitrary
limitation. This drops the use of PATH_MAX by just allocating
the proper buffer size, and disables the path length check in
_mesa_get_shader_capture_path.
Signed-off-by: Samuel Thibault
Bugzilla: https://bugs.freedesktop.org/s
Could you also add commit 9b121512ac0f78d0996613664b456005d88370d2? It's a
correctness fix for radv (ported from anv) which fixes segfaults which I'd
observed in Talos Principle.
--
| _ | Darren Salt, using Debian GNU/Linux (and Android)
| ( ) |
| X | ASCII Ribbon campaign against HTML e-mail
GNU/Hurd does not define PATH_MAX since it doesn't have such arbitrary
limitation. This works around it in arbprogram.c and shaderapi.c by
defining PATH_MAX to an arbitrary value (4096, as on Linux).
Signed-off-by: Samuel Thibault
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98632
---
On Sat, Nov 12, 2016 at 5:29 PM, Thomas Helland
wrote:
> After looking into how optimization passes are run, it became apparent
> that most glsl optimization passes only show benefit of one run.
> My idea was that some optimization passes will snowball on each other,
> while others that provide no
On Nov 13, 2016 14:32, "Marek Olšák" wrote:
>
> Interesting idea. Also added this:
>
Oh, nice! Gotta catch them all :-)
> diff --git a/src/compiler/glsl/linker.cpp b/src/compiler/glsl/linker.cpp
> index cc28b26..295d10b 100644
> --- a/src/compiler/glsl/linker.cpp
> +++ b/src/compiler/glsl/linker
Forwarding to the mailing list because I'm an idiot and am apparently not
able to email.
> 2016-11-13 10:39 GMT+01:00 Timothy Arceri :
> > On Sun, 2016-11-13 at 02:29 +0100, Thomas Helland wrote:
> >> It turns out that only 8% of shaders benefit from more than one run
> >> of the do_common_optimiz
From: Roland Scheidegger
The code for elts and linear paths was nearly 100% identical by now - with
the elts path simply having some additional gather for the elements in the
main loop (with some additional small differences before the main loop).
Hence nuke the separate functions and decide thi
From: Roland Scheidegger
Don't keep the ofbit. This is just a minor simplification, just adjust
the buffer size so that there will always be an overflow if buffers aren't
valid to fetch from.
Also, get rid of control flow from the instanced path too. Not worried about
performance, but it's simple
From: Roland Scheidegger
It turns out that noone actually cares if the address computations overflow,
be it the stride mul or the offset adds.
Wrap around seems to be explicitly permitted even by some other API (which
is a _very_ surprising result, as these overflow computations were added just
f
Overflow handling is simplified quite a bit both in jit code and vsplit
paths (basically just let things wrap around everywhere). This seems to
be good enough for all apis.
Also, elts and linear jit code is unified since the differences are minimal
(even more so at the end of the series). The cost
From: Roland Scheidegger
This was kind of strange, since it replaced indices which were only
overflowing due to bias with MAX_UINT. This would cause an overflow later
in the shader, except if stride was 0, however the vertex id would be
essentially random then (-1 + eltBias). No test cared about
From: Roland Scheidegger
This is a bit simpler. Mostly to make it easier to unify the paths later...
---
src/gallium/auxiliary/draw/draw_llvm.c | 48 ++
src/gallium/auxiliary/draw/draw_llvm.h | 8 ++--
.../draw/draw_pt_fetch_shade_pipeline_llvm.c
From: Marek Olšák
---
src/gallium/auxiliary/gallivm/lp_bld_misc.cpp | 14 ++
src/gallium/auxiliary/gallivm/lp_bld_misc.h | 3 +++
2 files changed, 17 insertions(+)
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp
b/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp
index bd
From: Marek Olšák
---
src/gallium/drivers/radeonsi/si_shader.c | 25 +-
src/gallium/drivers/radeonsi/si_shader_internal.h | 4 ++-
.../drivers/radeonsi/si_shader_tgsi_setup.c| 30 +++---
3 files changed, 30 insertions(+), 29 deletions(-)
diff
From: Marek Olšák
---
src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c
b/src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c
index 624a167..2f38949 100644
--- a/src/ga
From: Marek Olšák
---
src/compiler/Makefile.glsl.am| 4
src/compiler/Makefile.sources| 4
src/compiler/glsl/tests/cache_test.c | 2 +-
src/util/Makefile.sources| 2 ++
src/{compiler/glsl/cache.c => util/dis
From: Marek Olšák
---
src/compiler/glsl/tests/cache_test.c | 140 +--
src/util/disk_cache.c| 30
src/util/disk_cache.h| 56 +++---
3 files changed, 114 insertions(+), 112 deletions(-)
diff --git a/src/compiler/gl
> I've tested my private shader-db (26011 shaders, many games). I see
> only 2 regressions on radeonsi, both are changes in the code size:
> WORST REGRESSIONS - Code Size
> Before After Delta Percentage
> shaders/private/alien_isolation/788.shader_test [0]
> 8784 9080 2963
Interesting idea. Also added this:
diff --git a/src/compiler/glsl/linker.cpp b/src/compiler/glsl/linker.cpp
index cc28b26..295d10b 100644
--- a/src/compiler/glsl/linker.cpp
+++ b/src/compiler/glsl/linker.cpp
@@ -4985,10 +4985,9 @@ link_shaders(struct gl_context *ctx, struct
gl_shader_program *prog
https://bugs.freedesktop.org/show_bug.cgi?id=98699
Karol Herbst changed:
What|Removed |Added
Attachment #127941|0 |1
is obsolete|
https://bugs.freedesktop.org/show_bug.cgi?id=98699
Karol Herbst changed:
What|Removed |Added
Summary|"float[a+4 ? 1:1] f;" |"float[a+++4 ? 1:1] f;"
On Sun, 2016-11-13 at 02:29 +0100, Thomas Helland wrote:
> It turns out that only 8% of shaders benefit from more than one run
> of the do_common_optimization function, now that it has the loop.
> Doing only one run makes no difference to instruction count with NIR.
>
> 8324 shaders were run
This
This option is currently ignored when packing compressed textures.
Fixes CTS test (on gen8+):
* GL45-CTS.gtf32.GL3Tests.packed_pixels.packed_pixels_pixelstore
---
src/mesa/drivers/common/meta.c | 13 ++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/src/mesa/drivers/com
32 matches
Mail list logo
