On Thursday, June 22, 2017 3:17:21 PM PDT Chad Versace wrote:
> Fix all lines in src/mesa/main/marshal_generated.c that declare
> double-const variables. Below is all such lines, with duplicates
> removed:
>
>$ grep 'const const' marshal_generated.c | sort -u
>const const GLboolean * point
---
src/gallium/drivers/radeonsi/si_shader.c | 4 ++--
src/gallium/drivers/radeonsi/si_shader_tgsi_alu.c | 10 ++
src/mesa/state_tracker/st_context.c | 2 ++
3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_shader.c
b
---
src/mesa/state_tracker/st_atom_constbuf.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/src/mesa/state_tracker/st_atom_constbuf.c
b/src/mesa/state_tracker/st_atom_constbuf.c
index 7f17546..eb8f6b3 100644
--- a/src/mesa/state_tracker/st_atom_constbuf.c
+++ b/src/me
These seems to be handled differently by the backend and are not
packed with this series anyway.
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
b/src/mesa/state_tracker/st_glsl_to_tgsi
This is wrong fix me, just set to 4 to reduce piglit failures for now.
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index e0f5718..111e5d2
Currently asm programs don't support packing and are always padded to 4.
---
src/mesa/state_tracker/st_mesa_to_tgsi.c | 22 +-
1 file changed, 13 insertions(+), 9 deletions(-)
diff --git a/src/mesa/state_tracker/st_mesa_to_tgsi.c
b/src/mesa/state_tracker/st_mesa_to_tgsi.c
ind
---
src/mesa/program/ir_to_mesa.cpp| 45 +++---
src/mesa/program/ir_to_mesa.h | 3 +-
src/mesa/state_tracker/st_glsl_to_nir.cpp | 2 +-
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 2 +-
4 files changed, 39 insertions(+), 13 deletions(-)
diff -
The extra mul should end up const folded away in the backend.
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 33 ++
1 file changed, 33 insertions(+)
diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index dbb01b5..e
This is more inline with what the functions name suggests it should
do, and makes the code much easier to follow.
This will also make adding uniform packing support much simpler.
---
src/compiler/glsl/shader_cache.cpp | 4 +--
src/mesa/program/ir_to_mesa.cpp| 19 ++---
src/mesa/progr
---
src/mesa/main/uniform_query.cpp | 16 +---
1 file changed, 13 insertions(+), 3 deletions(-)
diff --git a/src/mesa/main/uniform_query.cpp b/src/mesa/main/uniform_query.cpp
index 0b22683..dadad87 100644
--- a/src/mesa/main/uniform_query.cpp
+++ b/src/mesa/main/uniform_query.cpp
@@ -
For uniform packing we multiply the const buffer index by its componets
in the st rather than in the backend.
---
src/gallium/auxiliary/gallivm/lp_bld_tgsi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.c
b/src/gallium/auxiliary/gal
---
src/mesa/state_tracker/st_atifs_to_tgsi.c | 7 +--
src/mesa/state_tracker/st_atifs_to_tgsi.h | 1 +
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 13 -
src/mesa/state_tracker/st_mesa_to_tgsi.c | 5 +++--
src/mesa/state_tracker/st_program.c| 2 +-
5 files changed,
There are still a handful of piglit tests failing and I'm yet to test
that there are no regressions in the non-packed path, but I'd really
like some feedback on the approach as Dave has flagged it as a possible
controversial tgsi change.
In order to avoid complicated swizzling and array element ad
---
src/mesa/program/prog_parameter.c | 53 +-
src/mesa/program/prog_parameter.h | 14 +---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 5 +--
3 files changed, 43 insertions(+), 29 deletions(-)
diff --git a/src/mesa/program/prog_parameter.c
b/s
These will be used in the following patch to allow copying directly
to the param list when packing is enabled.
---
src/mesa/main/uniform_query.cpp | 154
1 file changed, 91 insertions(+), 63 deletions(-)
diff --git a/src/mesa/main/uniform_query.cpp b/src/m
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 2 +-
src/mesa/state_tracker/st_mesa_to_tgsi.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index cf58212..0535ae1 100644
--- a/src/
---
src/mesa/main/uniform_query.cpp | 73 ++---
1 file changed, 54 insertions(+), 19 deletions(-)
diff --git a/src/mesa/main/uniform_query.cpp b/src/mesa/main/uniform_query.cpp
index b13f506..0b22683 100644
--- a/src/mesa/main/uniform_query.cpp
+++ b/src/mesa/m
Currently everything is padded to 4 elements. Making the list
more flexible will allow us to do uniform packing.
This change requires us to remove the existing packing in
_mesa_add_typed_unnamed_constant() in order to avoid regressions.
This packing will no longer be required once we add full pack
Will be used to determine whether to take packing code paths or not.
---
src/mesa/main/mtypes.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 0cb0024..49c8f63 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -3980,20
On 24/06/17 11:45, Ian Romanick wrote:
On 06/22/2017 03:56 PM, Timothy Arceri wrote:
---
src/mesa/main/uniform_query.cpp | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/mesa/main/uniform_query.cpp b/src/mesa/main/uniform_query.cpp
index 1570770..9683fa8 100644
Please try the series from Marek which reduces the batch size [1], the
reduced size helps reduce the impact of syncs. MARSHAL_MAX_CMD_SIZE is
also greatly reduced to help reduce thrashing the cache so its possible
this patch won't be as effective anymore. However you might not even
need it.
[
This is effectively a revert of 388f02729bbf88ba104f4f8ee1fdf005a240969c
though much code has been added since. Kristian initially moved it to
try and avoid locking problems with meta-based resolves. Now that meta
is gone from the resolve path (for good this time, we hope), we can move
it back.
If the fileIndex is different, that means they are in logically
different spaces. However if there's also a relative offset, then they
could end up pointing at the same spot again.
Also add a note about potential for multiple buffers to overlap even if
they're at different file indexes. However th
Previously the logic would decide that the record is kept, which
translates into keep = false in the caller, which meant that these
passes did not run.
While it's right that keep = false which means that a new record does
not need to be added, we do still have to perform the usual list
maintenance
This has no effect since in practice this will only play for
memory-backed files, for which VFETCH will never happen.
Signed-off-by: Ilia Mirkin
---
src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/drivers/no
On Sat, Jun 24, 2017 at 9:59 AM, Grigori Goronzy wrote:
> These entry points are used by Alien Isolation and caused
> synchronization with glthread. The async marshalling implementation
> is similar to glBuffer(Sub)Data.
>
> Results in an approximately 6x drop in glthread synchronizations and a
>
Reviewed-by: Bas Nieuwenhuizen
for the series.
On Tue, Jun 20, 2017 at 6:38 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> Fixes:
> dEQP-VK.pipeline.framebuffer_attachment.no_attachments
> dEQP-VK.pipeline.framebuffer_attachment.no_attachments_ms
>
> Signed-off-by: Dave Airlie
> ---
> src/am
On Tue, Jun 20, 2017 at 6:38 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> Noticed in passing.
>
> Signed-off-by: Dave Airlie
> ---
> src/amd/common/ac_nir_to_llvm.c | 5 +++--
> src/amd/common/ac_nir_to_llvm.h | 1 +
> src/amd/vulkan/si_cmd_buffer.c | 6 ++
> 3 files changed, 10 insertio
Signed-off-by: Bas Nieuwenhuizen
Fixes: 0628580eff6 "radv: Specify semantics of HTILE layout helpers."
---
src/amd/vulkan/radv_meta_blit2d.c | 29 +++--
1 file changed, 11 insertions(+), 18 deletions(-)
diff --git a/src/amd/vulkan/radv_meta_blit2d.c
b/src/amd/vulkan/radv
The idxbuf could linger, and when a clear happened, which also uses the
3d bufctx, we could get an error trying to access it.
This fixes spurious crashes/errors in CTS tests.
Fixes: 61d8f3387d ("nv50,nvc0: clear index buffer bufctx bin unconditionally")
Signed-off-by: Ilia Mirkin
Cc: mesa-sta...
These entry points are used by Alien Isolation and caused
synchronization with glthread. The async marshalling implementation
is similar to glBuffer(Sub)Data.
Results in an approximately 6x drop in glthread synchronizations and a
~30% FPS jump in Alien Isolation (Medium preset, Athlon 860K, RX 480
All the BuildUtil helpers just insert the operation into the current BB.
So we have to take care that any fetchSrc() operations happen before the
operation whose setIndirect() it goes into.
Signed-off-by: Ilia Mirkin
Cc: mesa-sta...@lists.freedesktop.org
---
I don't know how any of this really w
In preparation for bindless images, we have to retrieve the
target/format info from the instruction directly, as there will be no
declaration. Furthermore, for bound images, this information is still
available in the instruction, so we can drop the declaration-based
mechanism entirely.
Signed-off-
This keeps a list of resident textures (per context), and dumps that
list into the active buffer list when submitting. We also treat bindless
texture fetches slightly differently, wrt the meaning of indirect, and
not requiring the SAMPLER file to be used.
Signed-off-by: Ilia Mirkin
---
src/galli
A part of the driver constbuf area is allocated for bindless images. Any
update requires uploading to all driver constbufs. This also extends the
driver constbuf to 64KB, up from 2KB.
As a side-effect, this also fixes indirect image access, as fetchSrc()
was done *after* the op was already created
This is RFC because it doesn't work. All the piglits pass, but all the games
hang with bindless enabled.
I can't immediately think of what I've missed though.
I also totally hate the way I implemented images. It should be possible to do
it much better on Maxwell, since images are represented by T
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=100785
v2: I was too much twiddling whether to initialize nsys_inputs at the beginning
of shader initialization or for allocation of system values, and by the time I
decided to go with the first one, I forgot to change it back.
Signed-off-by:
…and print error in such case. Which probably is not a rare event btw
because fopen doesn't expand ~ to $HOME.
Also get rid of unused "bool ret" variable.
Signed-off-by: Constantine Kharlamov
---
src/gallium/drivers/r600/r600_hw_context.c | 9 +
1 file changed, 5 insertions(+), 4 deleti
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=100785
Signed-off-by: Constantine Kharlamov
---
src/gallium/drivers/r600/r600_shader.c | 8
src/gallium/drivers/r600/r600_shader.h | 5 +++--
2 files changed, 7 insertions(+), 6 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_
Note, I didn't manage to piglit-test it because of a sporadic lockup. Funnily
though, by sheer luck the test with fixes passed, but not the master.
I tried fixing it, but stuck, and had to retreat to reporting a bug
https://bugs.freedesktop.org/show_bug.cgi?id=101575
Anyway, the series is trivia,
Signed-off-by: Constantine Kharlamov
---
src/gallium/drivers/r600/r600_shader.c | 44 +-
1 file changed, 22 insertions(+), 22 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_shader.c
b/src/gallium/drivers/r600/r600_shader.c
index bdaf28ced2..156dba085d 10
On Jun 24, 2017 12:01 PM, "Nicolai Hähnle" wrote:
On 22.06.2017 17:10, Marek Olšák wrote:
> From: Marek Olšák
>
> +2.3% better score on Fiji. It might be better without HBM.
>
Can this be a drirc option?
The answer is the same as for: Can sisched be a flag defined in
p_defines.h, drirc, and
https://bugs.freedesktop.org/show_bug.cgi?id=99467
Grazvydas Ignotas changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
Reviewed-by: Pierre Moreau
On 09:28 pm - Jun 22 2017, Karol Herbst wrote:
> It's the same for all supported chipsets.
>
> Signed-off-by: Karol Herbst
> ---
> .../drivers/nouveau/codegen/nv50_ir_build_util.cpp| 16
> .../drivers/nouveau/codegen/nv50_ir_build_util.h |
On 22.06.2017 17:10, Marek Olšák wrote:
From: Marek Olšák
+2.3% better score on Fiji. It might be better without HBM.
Can this be a drirc option?
---
src/gallium/drivers/radeonsi/si_pipe.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_pipe.c
Patches 3 & 4:
Reviewed-by: Nicolai Hähnle
On 22.06.2017 16:35, Samuel Pitoiset wrote:
This prevents glViewport() and friends to always flush and
trigger _NEW_VIEWPORT.
Signed-off-by: Samuel Pitoiset
---
src/mesa/main/viewport.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
On 22.06.2017 14:34, Marek Olšák wrote:
From: Marek Olšák
It's not needed. The hw doesn't fetch ahead over page boundaries.
Reviewed-by: Nicolai Hähnle
---
src/gallium/drivers/radeonsi/si_shader.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_
Hi
2017-06-22 14:39 GMT+02:00 Wladimir J. van der Laan :
> On Wed, Jun 21, 2017 at 10:36:46PM +0200, Christian Gmeiner wrote:
>> Passes all ext_texture_swizzle piglits.
>
> You seem to have dropped the check in etnaviv_screen.c for HALTI0 when texture
> formats have implicit non-identity swizzle.
48 matches
Mail list logo
