Please, no. You could make a case for changing the default in
debugoptimized builds (which I would still be against) but we should
definitely still compile it. Why is it so hard to set NIR_VALIDATE=0 when
you really care about compiler times?
--Jason
On May 10, 2019 20:15:56 Marek Olšák
From: Marek Olšák
This reverts commit 7b85b9b8773b119360a31b66b321ae560a77cb6d.
---
src/compiler/nir/nir.h | 8
src/compiler/nir/nir_metadata.c | 2 +-
src/compiler/nir/nir_validate.c | 2 +-
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/src/compiler/nir/nir.h
We have unit tests for that pass. Maybe you could write one which
exercises the issue? It'd help in debugging.
On Thu, May 9, 2019 at 8:12 PM Dave Airlie wrote:
> I've got a bunch of cases where copy prop vars is getting things wrong
> around casts, it finds a store to an vec2 but ends up
Reviewed-by: Jason Ekstrand
On Fri, May 10, 2019 at 12:59 PM Alyssa Rosenzweig
wrote:
> This allows algebraic optimizations to check if the argument accesses
> multiple distinct components of a vector. So a swizzle like "xyz" will
> return true, but "yyy" will return false, as will a scalar.
https://bugs.freedesktop.org/show_bug.cgi?id=110625
Clayton Craft changed:
What|Removed |Added
Depends on||110357
Referenced Bugs:
https://bugs.freedesktop.org/show_bug.cgi?id=110625
Clayton Craft changed:
What|Removed |Added
Depends on||110302
Referenced Bugs:
glibc < 2.27 defines OVERFLOW in /usr/include/math.h.
This patch fixes this build error.
In file included from ../include/c99_math.h:37:0,
from ../src/util/u_math.h:44,
from ../src/mesa/main/macros.h:35,
from ../src/intel/compiler/brw_reg.h:47,
https://bugs.freedesktop.org/show_bug.cgi?id=110662
--- Comment #6 from tempel.jul...@gmail.com ---
Created attachment 144229
--> https://bugs.freedesktop.org/attachment.cgi?id=144229=edit
renderdoc capture part 6/6
--
You are receiving this mail because:
You are the QA Contact for the bug.
https://bugs.freedesktop.org/show_bug.cgi?id=110663
Bug ID: 110663
Summary: threads_posix.h:96: undefined reference to
`pthread_once'
Product: Mesa
Version: git
Hardware: x86-64 (AMD64)
OS: Linux (All)
https://bugs.freedesktop.org/show_bug.cgi?id=110662
--- Comment #5 from tempel.jul...@gmail.com ---
Created attachment 144228
--> https://bugs.freedesktop.org/attachment.cgi?id=144228=edit
renderdoc capture part 5
--
You are receiving this mail because:
You are the QA Contact for the bug.
You
https://bugs.freedesktop.org/show_bug.cgi?id=110662
--- Comment #4 from tempel.jul...@gmail.com ---
Created attachment 144227
--> https://bugs.freedesktop.org/attachment.cgi?id=144227=edit
renderdoc capture part 4
--
You are receiving this mail because:
You are the QA Contact for the bug.
You
https://bugs.freedesktop.org/show_bug.cgi?id=110662
--- Comment #3 from tempel.jul...@gmail.com ---
Created attachment 144226
--> https://bugs.freedesktop.org/attachment.cgi?id=144226=edit
renderdoc capture part 3
--
You are receiving this mail because:
You are the assignee for the bug.
You
https://bugs.freedesktop.org/show_bug.cgi?id=110662
--- Comment #2 from tempel.jul...@gmail.com ---
Created attachment 144225
--> https://bugs.freedesktop.org/attachment.cgi?id=144225=edit
renderdoc capture part 2
--
You are receiving this mail because:
You are the assignee for the bug.
You
On 2019-05-10 at 19:28, Vinson Lee wrote:
> glibc < 2.27 defines OVERFLOW in /usr/include/math.h.
>
> This patch fixes this build error.
>
> In file included from ../include/c99_math.h:37:0,
> from ../src/util/u_math.h:44,
> from ../src/mesa/main/macros.h:35,
>
https://bugs.freedesktop.org/show_bug.cgi?id=110662
--- Comment #1 from tempel.jul...@gmail.com ---
Created attachment 144224
--> https://bugs.freedesktop.org/attachment.cgi?id=144224=edit
renderdoc capture part 1
--
You are receiving this mail because:
You are the assignee for the bug.
You
https://bugs.freedesktop.org/show_bug.cgi?id=110662
Bug ID: 110662
Summary: shadow artifacts in Assassin's Creed: Unity
Product: Mesa
Version: git
Hardware: x86-64 (AMD64)
OS: Linux (All)
Status: NEW
On Fri, May 10, 2019 at 1:29 PM Marek Olšák wrote:
>
> From: Marek Olšák
>
> src/mesa/state_tracker/st_tgsi_lower_yuv.c:68: void reg_dst(struct
> tgsi_full_dst_register *, const struct tgsi_full_dst_register *, unsigned
> int): assertion "dst->Register.WriteMask" failed
>
> The second crash
From: Marek Olšák
src/mesa/state_tracker/st_tgsi_lower_yuv.c:68: void reg_dst(struct
tgsi_full_dst_register *, const struct tgsi_full_dst_register *, unsigned
int): assertion "dst->Register.WriteMask" failed
The second crash was due to insufficient allocated size for TGSI
instructions.
Cc:
Reviewed-by: Bas Nieuwenhuizen
Fixes: 5010436e09f "radv: bail out when binding the same vertex buffers"
I'll push after running it through the testsuite.
Thanks!
On Fri, May 10, 2019 at 9:38 PM Józef Kucia wrote:
>
> Only vertex inputs accessed by vertex shader must have valid buffers
>
Only vertex inputs accessed by vertex shader must have valid buffers
bound.
Signed-off-by: Józef Kucia
---
src/amd/vulkan/radv_cmd_buffer.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/amd/vulkan/radv_cmd_buffer.c b/src/amd/vulkan/radv_cmd_buffer.c
index
On Fri, May 10, 2019 at 12:28 PM Rob Clark wrote:
>
> On Fri, May 10, 2019 at 7:43 AM Connor Abbott wrote:
> >
> > On Fri, May 10, 2019 at 11:47 AM Connor Abbott wrote:
> > >
> > >
> > > This way of representing liveness, and then using a coloring register
> > > allocator, is a common
On Fri, May 10, 2019 at 7:43 AM Connor Abbott wrote:
>
> On Fri, May 10, 2019 at 11:47 AM Connor Abbott wrote:
> >
> >
> > This way of representing liveness, and then using a coloring register
> > allocator, is a common anti-pattern in Mesa, that was initially copied
> > from i965 which dates
glibc < 2.27 defines OVERFLOW in /usr/include/math.h.
This patch fixes this build error.
In file included from ../include/c99_math.h:37:0,
from ../src/util/u_math.h:44,
from ../src/mesa/main/macros.h:35,
from ../src/intel/compiler/brw_reg.h:47,
This allows algebraic optimizations to check if the argument accesses
multiple distinct components of a vector. So a swizzle like "xyz" will
return true, but "yyy" will return false, as will a scalar. This can be
useful for optimizations on vector processors, where a scalar
swizzle can be handled
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 151 +++---
.../panfrost/midgard/midgard_nir_algebraic.py | 12 +-
2 files changed, 37 insertions(+), 126 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
Hi Sumit,
I haven't looked too closely at this, but there are a couple of things right
away that jumped out to me. I've noted them inline below
Quoting Sumit Semwal (2019-05-09 23:02:57)
> Android version O and later would like to use altxmlconfig
> instead of libexpat based parsing.
>
> We
Oh, forgot to amend the message itself, thank you :)
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Thank you. I'm so sorry for the inconvenience... I've learned my CI
lesson ;)
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
On Fri, May 10, 2019 at 09:25:06AM -0700, Caio Marcelo de Oliveira Filho wrote:
> On Fri, May 10, 2019 at 04:22:05PM +, Alyssa Rosenzweig wrote:
> > This avoids a conflict with the new (driver-agnostic) blend_func enum in
> > shader_enum.h, which broke the build of swrast (and i965 by
On Fri, May 10, 2019 at 04:22:05PM +, Alyssa Rosenzweig wrote:
> This avoids a conflict with the new (driver-agnostic) blend_func enum in
> shader_enum.h, which broke the build of swrast (and i965 by extension).
>
> My apologies :(
>
> Signed-off-by: Alyssa Rosenzweig
> Fixes: f41be53a
Perhaps change the subject?
On Fri, May 10, 2019 at 12:08 PM Alyssa Rosenzweig wrote:
>
> This allows algebraic optimizations to check if the argument accesses
> multiple distinct components of a vector. So a swizzle like "xyz" will
> return true, but "yyy" will return false, as will a scalar.
This avoids a conflict with the new (driver-agnostic) blend_func enum in
shader_enum.h, which broke the build of swrast (and i965 by extension).
My apologies :(
Signed-off-by: Alyssa Rosenzweig
Fixes: f41be53a ("compiler: Add enums for blend state")
Cc: Caio Marcelo de Oliveira Filho
---
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 151 +++---
.../panfrost/midgard/midgard_nir_algebraic.py | 12 +-
2 files changed, 37 insertions(+), 126 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
This allows algebraic optimizations to check if the argument accesses
multiple distinct components of a vector. So a swizzle like "xyz" will
return true, but "yyy" will return false, as will a scalar. This can be
useful for optimizations on vector processors, where a convergent
swizzle can be done
On Fri, May 10, 2019 at 5:47 AM Connor Abbott wrote:
>
> On Fri, May 10, 2019 at 11:09 AM Christian Gmeiner
> wrote:
> >
> > Signed-off-by: Christian Gmeiner
> > ---
> > src/etnaviv/compiler/eir.h| 3 +
> > src/etnaviv/compiler/eir_live_variables.c | 258
On Fri, May 10, 2019 at 11:47 AM Connor Abbott wrote:
>
>
> This way of representing liveness, and then using a coloring register
> allocator, is a common anti-pattern in Mesa, that was initially copied
> from i965 which dates back to before we knew any better. I really
> don't want to see it
https://bugs.freedesktop.org/show_bug.cgi?id=110345
--- Comment #22 from Samuel Pitoiset ---
Created attachment 144219
--> https://bugs.freedesktop.org/attachment.cgi?id=144219=edit
force wd switch on eop
Does the attached patch help?
--
You are receiving this mail because:
You are the
https://bugs.freedesktop.org/show_bug.cgi?id=110345
--- Comment #21 from Samuel Pitoiset ---
Can you attach the output of "RADV_DEBUG=info vulkaninfo" please?
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the
https://bugs.freedesktop.org/show_bug.cgi?id=109822
Samuel Pitoiset changed:
What|Removed |Added
Status|NEEDINFO|RESOLVED
Resolution|---
https://bugs.freedesktop.org/show_bug.cgi?id=109599
Samuel Pitoiset changed:
What|Removed |Added
Resolution|--- |NOTABUG
Status|NEW
On Fri, May 10, 2019 at 11:09 AM Christian Gmeiner
wrote:
>
> Signed-off-by: Christian Gmeiner
> ---
> src/etnaviv/compiler/eir.h| 3 +
> src/etnaviv/compiler/eir_live_variables.c | 258 ++
> src/etnaviv/compiler/meson.build | 1 +
>
There has been interest in NIR support for etnaviv for a while, for the
obvious reasons: gaining access to common optimizations, better support
for non-trivial code transformations, better register allocation, and
the promise of OpenCL and SPIR-V support in the future.
This time I give it a try
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir_compiler.c | 61 ++
src/etnaviv/compiler/eir_compiler.h | 29 +
src/etnaviv/compiler/eir_compiler_nir.c | 1035 +++
src/etnaviv/compiler/eir_shader.c | 312 +++
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir_nir.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/etnaviv/compiler/eir_nir.c b/src/etnaviv/compiler/eir_nir.c
index 6e75f2c5fd8..a579037a667 100644
--- a/src/etnaviv/compiler/eir_nir.c
+++
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir.h | 3 +
src/etnaviv/compiler/eir_print.c | 222 +++
src/etnaviv/compiler/meson.build | 1 +
3 files changed, 226 insertions(+)
create mode 100644 src/etnaviv/compiler/eir_print.c
diff --git
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir.c | 239 ++
src/etnaviv/compiler/eir.h | 499
src/etnaviv/compiler/eir_uniform.c | 108 +
src/etnaviv/compiler/meson.build| 38 ++
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir.h| 3 +
src/etnaviv/compiler/eir_live_variables.c | 258 ++
src/etnaviv/compiler/meson.build | 1 +
.../compiler/tests/eir_live_variables.cpp | 162 +++
This change adds the eir_optimize_nir(..) function which gets used
to optimize and transform a nir shader for our needs.
Besides this there are some other nir helpers for compiler options
and tgsi handling.
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir_nir.c | 131
Signed-off-by: Christian Gmeiner
---
src/gallium/drivers/etnaviv/eir_cmdline.c | 189 ++
src/gallium/drivers/etnaviv/meson.build | 20 +++
2 files changed, 209 insertions(+)
create mode 100644 src/gallium/drivers/etnaviv/eir_cmdline.c
diff --git
Some created shaders can contain instructions like:
mov t0 t0
This simple pass removes that kind of mov instructions.
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir_opt_peephole.c | 63 +
src/etnaviv/compiler/eir_optimize.c | 4 +-
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir.h | 3 ++
src/etnaviv/compiler/eir_compiler_nir.c | 6
src/etnaviv/compiler/eir_optimize.c | 48 +
src/etnaviv/compiler/eir_optimize.h | 31
Add a debug option ETNA_MESA_DEBUG="nir" that will cause the etnaviv
gallium pipe driver to advertise support and preference for NIR shaders.
Signed-off-by: Philipp Zabel
Signed-off-by: Michael Tretter
---
src/gallium/drivers/etnaviv/etnaviv_debug.h | 3 +++
Some operations are scalar and need special treatment. Those
are exp and log2.
dst := exp2(src2.x)
dst := log2(src2.x)
The result is broadcasted over all active destination components.
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir_nir.h| 3 +
Signed-off-by: Christian Gmeiner
---
src/gallium/drivers/etnaviv/etnaviv_context.c | 23 ++--
src/gallium/drivers/etnaviv/etnaviv_shader.c | 55 +--
2 files changed, 68 insertions(+), 10 deletions(-)
diff --git a/src/gallium/drivers/etnaviv/etnaviv_context.c
This change preps etnaviv to be able to swap the used compiler
backend easily. It takes care of uniform stuff and the actual
creation of shader variants. In the long run - if one backend
won - we can remove this again.
Signed-off-by: Christian Gmeiner
---
- if shader is empty add a NOP instruction
- avoid multiple uniform src for alu ops
- resolve jump target
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir.h | 3 +
src/etnaviv/compiler/eir_legalize.c | 177
Some unit tests would be great.
Signed-off-by: Christian Gmeiner
---
src/etnaviv/compiler/eir.h | 10 +-
src/etnaviv/compiler/eir_register_allocate.c | 176 ++-
2 files changed, 184 insertions(+), 2 deletions(-)
diff --git a/src/etnaviv/compiler/eir.h
Since all threads share a global temporary vec4 register file, it is
important to reduce temporary register use of shaders.
Using source swizzles and destination write mask of ALU operations we
can layer smaller virtual registers on top of the physical base
registers that overlap with their base
Signed-off-by: Christian Gmeiner
---
src/gallium/drivers/etnaviv/Makefile.sources | 2 +
src/gallium/drivers/etnaviv/etnaviv_eir.c| 454 +++
src/gallium/drivers/etnaviv/etnaviv_eir.h| 71 +++
src/gallium/drivers/etnaviv/etnaviv_screen.c | 15 +
Hello Dear Sir,
My name is Shubham Verma. I am a B tech 2nd-year I.T. student I would like
to look into working with X.Org for the Endless Vacation of Code program
2019 this summer.
I have just recently started learning and looking into the contributions
that I can make to GitHub issues. As this
Core mesa uses libexpat to parse default configuration information
stored as xml in header files. This potentially allows userspaces to
supply their own configuration information, without a need to rebuild.
However, for userspaces that don't want or need to allow external
configuration, libexpat
Mesa uses libexpat for many configuration parsing needs; however some
userspaces like Android may not want to use libexpat for various reasons -
eg some might restrict reading of any config xml files from filesystems.
This patchset proposes a simple lookup mechanism for the default values
as per
Android version O and later would like to use altxmlconfig
instead of libexpat based parsing.
We enable altxmlconfig instead of xmlconfig.c for Android O+.
Change-Id: Iaa51b2a1e9c16a1fdf6f120104429b19dc68f9eb
Signed-off-by: Sumit Semwal
---
src/gallium/targets/dri/Android.mk | 8 +++-
63 matches
Mail list logo
