On 26/06/17 09:49 AM, Marek Olšák wrote:
> From: Marek Olšák
>
> Broken by:
>
> commit 00173d91b70ae4dcea7c6324ee4858c498cae14b
> Author: Marek Olšák
> Date: Sat Jun 10 12:09:43 2017 +0200
>
> mesa: don't flag _NEW_TRANSFORM for st/mesa if possible
>
> It also optimizes the case slightl
Patches 2 & 3:
Reviewed-by: Nicolai Hähnle
On 23.06.2017 12:18, James Legg wrote:
---
src/amd/common/ac_nir_to_llvm.c | 17 -
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 3a26668..b32a9
On 26.06.2017 01:25, Timothy Arceri wrote:
---
src/mesa/main/uniforms.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/main/uniforms.c b/src/mesa/main/uniforms.c
index 91c3bf6..1c3d89b 100644
--- a/src/mesa/main/uniforms.c
+++ b/src/mesa/main/uniforms.c
@@ -104,7
For the series:
Reviewed-by: Nicolai Hähnle
On 26.06.2017 02:49, Timothy Arceri wrote:
---
src/mapi/glapi/gen/gl_API.xml | 14
src/mesa/main/teximage.c | 82 +++
src/mesa/main/teximage.h | 34 ++
3 files changed, 1
On 25.06.2017 03:31, Timothy Arceri wrote:
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/mt
On 25.06.2017 03:31, Timothy Arceri wrote:
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 wil
For the series:
Reviewed-by: Samuel Pitoiset
On 06/26/2017 02:49 AM, Timothy Arceri wrote:
This is used to inline KHR_no_error logic without inlining
the function into all its callers.
---
src/mesa/main/teximage.c | 40 ++--
1 file changed, 26 insertions(
With Nicolai's suggestion, patch is:
Reviewed-by: Samuel Pitoiset
On 06/26/2017 01:25 AM, Timothy Arceri wrote:
---
src/mesa/main/uniforms.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/main/uniforms.c b/src/mesa/main/uniforms.c
index 91c3bf6..1c3d89b 100644
Can you just keep the declarations of texObj and texFormat at the
beginning of the function?
Except this, patch is:
Reviewed-by: Samuel Pitoiset
On 06/26/2017 06:30 AM, Timothy Arceri wrote:
---
src/mesa/main/teximage.c | 17 -
1 file changed, 8 insertions(+), 9 deletions(
Except a little nitpick on patch 1, series is:
Reviewed-by: Samuel Pitoiset
On 06/26/2017 06:30 AM, Timothy Arceri wrote:
---
src/mesa/main/teximage.c | 17 -
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/src/mesa/main/teximage.c b/src/mesa/main/teximage.c
in
On 25.06.2017 03:31, Timothy Arceri wrote:
---
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/mes
I had a chance to look at the patches in detail now. Thanks for working
on this, it should be quite useful!
With the remark on patch 1 addressed, patches 1, 2, and 4-9 are
Reviewed-by: Nicolai Hähnle
I've sent some more extensive comments on patch 3, and I already
explained that I don't like
Looks good!
Reviewed-by: Nicolai Hähnle
On 26.06.2017 02:49, Marek Olšák wrote:
From: Marek Olšák
---
src/gallium/drivers/radeonsi/si_state.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeonsi/si_state.c
b/src/gallium/drivers/radeonsi/si
From: Nicolai Hähnle
Each field gets a distinct name, so we should never hit the case where
the name already exists in the parameter list.
---
src/mesa/program/ir_to_mesa.cpp | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/src/mesa/program/ir_to_mesa.cpp b/src/mes
From: Nicolai Hähnle
Otherwise, the padding bits remain undefined, which leads to valgrind
errors when storing the gl_shader_variable in the disk cache.
---
src/compiler/glsl/linker.cpp | 1 +
src/mesa/main/mtypes.h | 2 ++
2 files changed, 3 insertions(+)
diff --git a/src/compiler/glsl/l
From: Nicolai Hähnle
Save some passes over the IR.
---
src/compiler/glsl/linker.cpp | 88 +---
1 file changed, 50 insertions(+), 38 deletions(-)
diff --git a/src/compiler/glsl/linker.cpp b/src/compiler/glsl/linker.cpp
index 5366200..cfda263 100644
--- a/s
Hi all,
This is a bit of a random collection of small patches. What ties them
together is that they're all vaguely related to the glsl code, and that they
accumulated while I was working on ARB_gl_spirv. They do make sense
individually though. I see no regressions here with assertions enabled o
From: Nicolai Hähnle
This is helpful in debugging varying assignments.
---
src/compiler/glsl/ir_print_visitor.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/compiler/glsl/ir_print_visitor.cpp
b/src/compiler/glsl/ir_print_visitor.cpp
index ba3c1e2..86ddea6 100644
---
From: Nicolai Hähnle
---
src/mesa/program/ir_to_mesa.cpp | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/src/mesa/program/ir_to_mesa.cpp b/src/mesa/program/ir_to_mesa.cpp
index 3ed5bed..ac12b59 100644
--- a/src/mesa/program/ir_to_mesa.cpp
+++ b/src/mesa/program/ir_to
From: Nicolai Hähnle
We always have stage == first and stage == last when first == last, so
drop the special case. Also rephrase the comment to make the logic
clearer.
---
src/compiler/glsl/linker.cpp | 29 +++--
1 file changed, 11 insertions(+), 18 deletions(-)
diff --g
From: Nicolai Hähnle
---
src/mesa/state_tracker/st_program.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/mesa/state_tracker/st_program.h
b/src/mesa/state_tracker/st_program.h
index 57673e9..3f0cf49 100644
--- a/src/mesa/state_tracker/st_program.h
+++ b/src/mesa/state_tracker/st_progr
From: Nicolai Hähnle
---
src/util/disk_cache.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/util/disk_cache.h b/src/util/disk_cache.h
index 72f4463..9aade16 100644
--- a/src/util/disk_cache.h
+++ b/src/util/disk_cache.h
@@ -136,21 +136,21 @@ disk_cache_put(struct disk_
From: Nicolai Hähnle
Undefined data will eventually trigger a valgrind error while computing
its CRC32 while writing it into the disk cache, but at that point, it is
basically impossible to track down where the undefined data came from.
With this change, finding the origin of undefined data beco
On Fri, 2017-06-23 at 11:29 -0700, Jason Ekstrand wrote:
> OpAtomicCompareExchangeWeak is only available in OpenCL, not Vulkan.
> As such, we probably don't want to handle it yet as no one uses
> spirv_to_nir for OpenCL. The assert on AtomicCompareExchange is,
> however, a bug we should fix now.
We were hitting the
unreachable("Invalid image opcode")
near the end of vtn_handle_image when parsing the
SpvOpAtomicCompareExchange opcode.
v2: Add stable CC.
v3: Ignore SpvOpAtomicCompareExchangeWeak. It requires the Kernel
capability which is not exposed in Vulkan, and spirv_to_nir is n
Both patches:
Reviewed-by: Nicolai Hähnle
On 26.06.2017 02:49, Marek Olšák wrote:
From: Marek Olšák
And other client state changes don't have to call
update_derived_primitive_restart_state.
---
src/mesa/main/enable.c | 19 +--
1 file changed, 9 insertions(+), 10 deletions
For the series:
Reviewed-by: Samuel Pitoiset
On 06/26/2017 04:38 AM, Timothy Arceri wrote:
---
src/mesa/main/teximage.c | 175 +++
src/mesa/main/teximage.h | 9 ---
2 files changed, 85 insertions(+), 99 deletions(-)
diff --git a/src/mesa/main/
On 26/06/17 19:27, Nicolai Hähnle wrote:
On 25.06.2017 03:31, Timothy Arceri wrote:
---
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
https://bugs.freedesktop.org/show_bug.cgi?id=91724
--- Comment #9 from Frederic Devernay ---
> - Using/setting GL_GLEXT_PROTOTYPES is a _very_ bad idea.
Agreed. This patch *fixes* glext.h, but I do not use it anywhere. It is fixed
just in case someone crazy enough wants to use it.
More importan
On 26.06.2017 11:56, Timothy Arceri wrote:
On 26/06/17 19:27, Nicolai Hähnle wrote:
On 25.06.2017 03:31, Timothy Arceri wrote:
---
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
https://bugs.freedesktop.org/show_bug.cgi?id=91724
--- Comment #10 from Frederic Devernay ---
Created attachment 132249
--> https://bugs.freedesktop.org/attachment.cgi?id=132249&action=edit
patch file for glmangle.h
Here's the patch for mesa 17.1.3.
Note that in the #define section, some funct
On Fri, 2017-06-16 at 10:12 +0200, Juan A. Suarez Romero wrote:
> Commit 00620782c9 (i965: use nir_shader_gather_info() over
> do_set_program_inouts()) changed how we compute the outputs written.
>
> In the previous version it was using the IR declared outputs, while in
> the new one it uses NIR t
On 26/06/17 19:40, Nicolai Hähnle wrote:
From: Nicolai Hähnle
Otherwise, the padding bits remain undefined, which leads to valgrind
errors when storing the gl_shader_variable in the disk cache.
---
src/compiler/glsl/linker.cpp | 1 +
src/mesa/main/mtypes.h | 2 ++
2 files changed, 3
On Tue, 2017-06-13 at 11:14 +0200, Juan A. Suarez Romero wrote:
> v2: NIR fmax/fmin already handles NaN (Connor).
> ---
Implemented the functions using fmax/fmin.
Could you review it? Thank you!
J.A.
> src/compiler/spirv/vtn_glsl450.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> di
On 26/06/17 19:40, Nicolai Hähnle wrote:
From: Nicolai Hähnle
Save some passes over the IR.
---
src/compiler/glsl/linker.cpp | 88 +---
1 file changed, 50 insertions(+), 38 deletions(-)
diff --git a/src/compiler/glsl/linker.cpp b/src/compiler/glsl/li
Currently a resource flush may trigger a self resolve, even if a scanout buffer
exists, but is up to date. If a scanout buffer exists we only ever want to
flush the resource to the scanout buffer. This fixes a performance regression.
Fixes: dda956340ce9 (etnaviv: resolve tile status when flushing
On 26/06/17 20:06, Nicolai Hähnle wrote:
On 26.06.2017 11:56, Timothy Arceri wrote:
On 26/06/17 19:27, Nicolai Hähnle wrote:
On 25.06.2017 03:31, Timothy Arceri wrote:
---
src/mesa/state_tracker/st_atom_constbuf.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/
On 26.06.2017 12:18, Timothy Arceri wrote:
On 26/06/17 19:40, Nicolai Hähnle wrote:
From: Nicolai Hähnle
Save some passes over the IR.
---
src/compiler/glsl/linker.cpp | 88
+---
1 file changed, 50 insertions(+), 38 deletions(-)
diff --git a/src/c
On Fri, 2017-06-23 at 17:50 +0200, Lucas Stach wrote:
> The current way of importing the resource from renderonly after allocation
> is opaque and is taking away control from the driver, which it needs in
> order to implement more advanced scenarios, than the simple linear
> scanout with matching s
On Fri, 2017-06-23 at 17:50 +0200, Lucas Stach wrote:
> The layer stride information is used in various parts of the driver,
> so it needs to be present regardless if the driver allocated the
> buffer itself or merely imported it from an external source.
>
> Signed-off-by: Lucas Stach
> ---
> sr
On Fri, 2017-06-23 at 17:50 +0200, Lucas Stach wrote:
> The resource struct is already allocated at this point and should be
> free properly.
>
> Signed-off-by: Lucas Stach
> ---
> src/gallium/drivers/etnaviv/etnaviv_resource.c | 6 +-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> di
On Fri, 2017-06-23 at 17:50 +0200, Lucas Stach wrote:
> This fixes failures to import the scanout buffer with screen resolutions
> that don't satisfy teh RS alignment restrictions, like 1680x1050.
^^^ typo
> Signed-off-by: Lucas Stach
> ---
> src/gallium/drivers/etnaviv/etnav
On Fri, 2017-06-23 at 17:50 +0200, Lucas Stach wrote:
> If we import a dma-buf with a sampler/pixel pipe incompatible modifier,
> the imported buffer will end up in an external resource view. As
> resource_changed signals the change of the imported resource, we need
> to update the external view se
On Fri, Jun 23, 2017 at 06:43:37PM -0700, Ian Romanick wrote:
> On 06/23/2017 04:58 AM, ville.syrj...@linux.intel.com wrote:
> > From: Ville Syrjälä
> >
> > Pass the correct src coordinates to CopyTexSubImage()
> > when creating the temporary texture, and also take care to adjust
> > flipX/Y if t
On Mon, 2017-06-26 at 12:25 +0200, Lucas Stach wrote:
> Currently a resource flush may trigger a self resolve, even if a scanout
> buffer
> exists, but is up to date. If a scanout buffer exists we only ever want to
> flush the resource to the scanout buffer. This fixes a performance regression.
>
On 26/06/17 19:17, Nicolai Hähnle wrote:
On 25.06.2017 03:31, Timothy Arceri wrote:
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 o
2017-06-26 12:25 GMT+02:00 Lucas Stach :
> Currently a resource flush may trigger a self resolve, even if a scanout
> buffer
> exists, but is up to date. If a scanout buffer exists we only ever want to
> flush the resource to the scanout buffer. This fixes a performance regression.
>
> Fixes: dda9
2017-06-23 17:50 GMT+02:00 Lucas Stach :
> The resource struct is already allocated at this point and should be
> free properly.
>
> Signed-off-by: Lucas Stach
Reviewed-by: Christian Gmeiner
> ---
> src/gallium/drivers/etnaviv/etnaviv_resource.c | 6 +-
> 1 file changed, 5 insertions(+), 1
2017-06-23 17:50 GMT+02:00 Lucas Stach :
> The layer stride information is used in various parts of the driver,
> so it needs to be present regardless if the driver allocated the
> buffer itself or merely imported it from an external source.
>
> Signed-off-by: Lucas Stach
Reviewed-by: Christian G
This also fixes a conform mustpass failure we're seeing.
Reviewed-by: Roland Scheidegger
Am 26.06.2017 um 02:49 schrieb Marek Olšák:
> From: Marek Olšák
>
> Broken by:
>
> commit 00173d91b70ae4dcea7c6324ee4858c498cae14b
> Author: Marek Olšák
> Date: Sat Jun 10 12:09:43 2017 +0200
>
>
2017-06-23 17:50 GMT+02:00 Lucas Stach :
> This fixes failures to import the scanout buffer with screen resolutions
> that don't satisfy teh RS alignment restrictions, like 1680x1050.
>
> Signed-off-by: Lucas Stach
Reviewed-by: Christian Gmeiner
> ---
> src/gallium/drivers/etnaviv/etnaviv_reso
Am Montag, den 26.06.2017, 13:09 +0200 schrieb Christian Gmeiner:
> 2017-06-26 12:25 GMT+02:00 Lucas Stach :
> > Currently a resource flush may trigger a self resolve, even if a
> > scanout buffer
> > exists, but is up to date. If a scanout buffer exists we only ever
> > want to
> > flush the resou
The labels array may change its virtual address on a reallocation, so
it is invalid to cache pointers into the array. Rather than using the
pointer directly, remember the array index.
Fixes miscompilation of shaders in glmark2 ideas, leading to GPU hangs.
Signed-off-by: Lucas Stach
---
src/gall
Am Montag, den 26.06.2017, 13:40 +0200 schrieb Lucas Stach:
> The labels array may change its virtual address on a reallocation, so
> it is invalid to cache pointers into the array. Rather than using the
> pointer directly, remember the array index.
>
> Fixes miscompilation of shaders in glmark2 i
On 22 June 2017 at 20:14, Ben Crocker wrote:
> In swrastGetDrawableInfo, set *x and *y, not just *w and *h;
> this fixes a crash later in drisw_update_tex_buffer when the
> (formerly) uninitialized x and y values are used to construct
> an address in a call to llvmpipe_transfer_map.
>
> Fixes cras
On 20 June 2017 at 17:05, Eric Engestrom wrote:
> On Tuesday, 2017-06-20 16:04:47 +0100, Emil Velikov wrote:
>> From: Emil Velikov
>>
>> Function cannot fail and always returns true.
>>
>> Signed-off-by: Emil Velikov
>
> Both patches are:
> Reviewed-by: Eric Engestrom
>
>> ---
>> We could even
2017-06-26 13:30 GMT+02:00 Lucas Stach :
> Am Montag, den 26.06.2017, 13:09 +0200 schrieb Christian Gmeiner:
>> 2017-06-26 12:25 GMT+02:00 Lucas Stach :
>> > Currently a resource flush may trigger a self resolve, even if a
>> > scanout buffer
>> > exists, but is up to date. If a scanout buffer exis
On 25.06.2017 09:22, Gert Wollny wrote:
To prepare the implementation of a temp register lifetime tracker
some of the classes are moved into seperate header/implementation
files to make them accessible from other files.
Specifically these are:
class st_src_reg;
class st_dst_reg;
2017-06-23 17:50 GMT+02:00 Lucas Stach :
> There is no point in keeping this indirection. Makes the code easier to
> follow.
>
> Signed-off-by: Lucas Stach
Reviewed-by: Christian Gmeiner
> ---
> src/gallium/drivers/etnaviv/etnaviv_resource.c | 14 --
> src/gallium/drivers/etnaviv/e
https://bugs.freedesktop.org/show_bug.cgi?id=77449
Bug 77449 depends on bug 98238, which changed state.
Bug 98238 Summary: Witcher 2: objects are black when changing lod on Radeon
Pitcairn
https://bugs.freedesktop.org/show_bug.cgi?id=98238
What|Removed |Added
-
2017-06-26 13:40 GMT+02:00 Lucas Stach :
> The labels array may change its virtual address on a reallocation, so
> it is invalid to cache pointers into the array. Rather than using the
> pointer directly, remember the array index.
>
> Fixes miscompilation of shaders in glmark2 ideas, leading to GPU
2017-06-26 13:44 GMT+02:00 Lucas Stach :
> Am Montag, den 26.06.2017, 13:40 +0200 schrieb Lucas Stach:
>> The labels array may change its virtual address on a reallocation, so
>> it is invalid to cache pointers into the array. Rather than using the
>> pointer directly, remember the array index.
>>
Hi Tim,
On 22 June 2017 at 22:12, Tim Rowley wrote:
> Highlights include splitting the heavily templated files into multiple
> chunks to speed compile (2x for a large machine), and switching the
> simd intrinsic usage from a macro-based header to a more c++ feeling
> library.
>
Yay \o/. Out of cu
Thanks for the update. First off, you're still not tracking individual
components, but that's absolute necessary. Think:
BGNLOOP
MOV TEMP[1].x, ...
UIF ...
MOV TEMP[1].y, ...
ENDIF
use TEMP[1].y
ENDLOOP
Now for a scan through the patch:
On 25.06.2017 09:22, Gert Wollny wrote:
T
Hi Tim,
On 22 June 2017 at 22:13, Tim Rowley wrote:
> Hardcode split to four files currently. Decreases swr build
> time on a quad-core by ~10%.
> ---
> src/gallium/drivers/swr/Makefile.am| 26 +-
> src/gallium/drivers/swr/Makefile.sources |4 +
> src/gallium/dri
Signed-off-by: Eric Engestrom
---
src/amd/vulkan/radv_pipeline.c | 3 ++-
src/amd/vulkan/radv_pipeline_cache.c | 7 ---
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/amd/vulkan/radv_pipeline.c b/src/amd/vulkan/radv_pipeline.c
index 0c262799f0..efe641dc0b 100644
---
Signed-off-by: Eric Engestrom
---
src/intel/vulkan/anv_private.h | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/intel/vulkan/anv_private.h b/src/intel/vulkan/anv_private.h
index e5d88f2815..8079378ae6 100644
--- a/src/intel/vulkan/anv_private.h
+++ b/src/intel/vulkan
On 22 June 2017 at 22:13, Tim Rowley wrote:
> Hardcode split to four files currently. Decreases swr build
> time on KNL by over 50%.
Out of curiosity what is KNL?
Also, over 50% decrease - time to pop the champagne ;-)
> ---
> src/gallium/drivers/swr/Makefile.am| 36 +-
> src
Hi Tim,
On 22 June 2017 at 22:13, Tim Rowley wrote:
> Switch from a macro-based simd intrinsics layer to a more C++
> implementation, which also adds AVX512 optimizations to 128-bit
> and 256-bit SIMD.
> + rasterizer/common/simdlib_128_avx.inl \
> + rasterizer/common/simdlib_128_avx2
Hi,
Am Montag, 26. Juni 2017, 01:17:23 CEST schrieb Timothy Arceri:
> On 25/06/17 18:31, Grigori Goronzy wrote:
> > On 2017-06-25 02:37, Timothy Arceri wrote:
> >> Please try the series from Marek which reduces the batch size [1], the
> >> reduced size helps reduce the impact of syncs. MARSHAL_MAX
On 25.06.2017 09:22, Gert Wollny wrote:
This patch adds a set of unit tests for the new lifetime tracker.
---
configure.ac | 1 +
src/mesa/Makefile.am | 2 +-
src/mesa/state_tracker/tests/Makefile.am | 36 +
.
Thanks for the update. Do you have the series on an accessible git
repository somewhere? E.g. on GitHub or Gitlab or wherever? That would
be helpful.
Also, please don't continue to chain this thread for subsequent
versions. It's okay to do this for quick follow-up fixes to patches you
sent, b
On 22 June 2017 at 19:13, Eric Engestrom wrote:
> On Thursday, 2017-06-22 16:47:29 +0100, Emil Velikov wrote:
>> On 22 June 2017 at 00:29, Eric Engestrom wrote:
>> >> configure: error: Package requirements (x11 xext xdamage >= 1.1 xfixes
>> >> x11-xcb xcb xcb-glx >= 1.8.1 xcb-dri2 >= 1.8) were no
On 26/06/17 20:28, Nicolai Hähnle wrote:
On 26.06.2017 12:18, Timothy Arceri wrote:
On 26/06/17 19:40, Nicolai Hähnle wrote:
From: Nicolai Hähnle
Save some passes over the IR.
---
src/compiler/glsl/linker.cpp | 88
+---
1 file changed, 50 inserti
On 22 June 2017 at 19:14, Eric Engestrom wrote:
> From: Eric Engestrom
>
>> configure: error: Package requirements (x11 xext xdamage >= 1.1 xfixes
>> x11-xcb xcb xcb-glx >= 1.8.1 xcb-dri2 >= 1.8) were not met:
>> No package 'xdamage' found
>> No package 'xfixes' found
>
> Signed-off-by: Eric Enge
On 23 June 2017 at 00:07, Andres Gomez wrote:
> On Mon, 2017-06-19 at 12:36 +0100, Emil Velikov wrote:
>> Hi Nicolai,
>>
>> On 12 June 2017 at 20:33, Nicolai Hähnle wrote:
>> > From: Nicolai Hähnle
>> >
>> > The very last entry in the sid_strings_offsets table ended up missing,
>> > leading to o
On 25.06.2017 09:22, Gert Wollny wrote:
The remapping evaluator first sorts the temporary registers ascending
based on their first life time instruction, and then uses a binary search
to find merge canidates.
For the initial sorting it uses std::sort because qsort is quite slow in
comparison. By
On 25.06.2017 09:22, Gert Wollny wrote:
The patch adds tests for the register rename mapping evaluation.
---
.../tests/test_glsl_to_tgsi_lifetime.cpp | 94 ++
1 file changed, 94 insertions(+)
diff --git a/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cp
On 25.06.2017 09:22, Gert Wollny wrote:
This patch ties in the new temporary register lifetime estiamtion and
Tpyo: estimation
rename mapping evaluation. In order to enable it, the evironment
variable MESA_GLSL_TO_TGSI_NEW_MERGE must be set.
This make sense during development, but I'd say
Hi Lucas,
On 23 June 2017 at 16:50, Lucas Stach wrote:
> +const uint64_t priority_to_modifier[] = {
I think you can make this static.
[snip]
> res->texture =
> -etna_resource_alloc(pctx->screen, ETNA_LAYOUT_TILED, &templat);
> +etna_resource_alloc(pctx->screen,
On 26.06.2017 15:23, Emil Velikov wrote:
On 23 June 2017 at 00:07, Andres Gomez wrote:
On Mon, 2017-06-19 at 12:36 +0100, Emil Velikov wrote:
Hi Nicolai,
On 12 June 2017 at 20:33, Nicolai Hähnle wrote:
From: Nicolai Hähnle
The very last entry in the sid_strings_offsets table ended up miss
On 2017-06-26 15:11, Marc Dietrich wrote:
unfortunately, this change broke vmware/vmplayer here (bisected).
Windows
guest on linux host. Sig 11 in SVGA driver. All good if
mesa_glthread=false.
Can you provide instructions how to reproduce this problem? A backtrace
might help, too.
I don
Am Montag, 26. Juni 2017, 15:35:15 CEST schrieb Grigori Goronzy:
> On 2017-06-26 15:11, Marc Dietrich wrote:
> > unfortunately, this change broke vmware/vmplayer here (bisected).
> > Windows
> > guest on linux host. Sig 11 in SVGA driver. All good if
> > mesa_glthread=false.
>
> Can you provide in
Hi Constantine,
Thanks for giving r600 some much needed love.
While the patch has landed, just going to share some generic comments.
Most of which are documented here https://www.mesa3d.org/submittingpatches.html.
On 24 June 2017 at 15:06, Constantine Kharlamov wrote:
> Fixes: https://bugs.free
Some combinations of c++ compilers and standard libraries had problems
with the string::replace code we were using previously.
This should fix the travis-ci system.
---
.../drivers/swr/rasterizer/codegen/templates/gen_knobs.cpp | 12 +---
1 file changed, 9 insertions(+), 3 deletions(-)
It's enabled through message buffer for UVD
Signed-off-by: Leo Liu
---
src/gallium/drivers/radeon/radeon_vcn_dec.c | 1 +
src/gallium/drivers/radeon/radeon_vcn_dec.h | 2 ++
2 files changed, 3 insertions(+)
diff --git a/src/gallium/drivers/radeon/radeon_vcn_dec.c
b/src/gallium/drivers/radeon/r
On Fri, 2017-06-23 at 17:50 +0200, Lucas Stach wrote:
> This reworks the logic in etna_update_sampler_source to select the
> newest resource view for updating the texture view. This should make
> the logic easier to follow and fixes texture updates from imported
> dma-bufs.
>
> Signed-off-by: Luca
From: Nicolai Hähnle
Arguably this could convert to a load intrinsic, but other code paths
already expect this as a fragment shader input.
---
src/compiler/nir/nir_lower_system_values.c | 36 ++
1 file changed, 32 insertions(+), 4 deletions(-)
diff --git a/src/compil
From: Nicolai Hähnle
This is convenient for backends that support both Vulkan and OpenGL while
lowering samplers to derefs with nir_lower_samplers_as_deref.
---
src/compiler/glsl/glsl_to_nir.cpp | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/compiler/glsl/glsl_to_nir.cpp
b/src/compiler
From: Nicolai Hähnle
We will use this from radeonsi/nir, which we want to keep as pure C code.
---
src/compiler/glsl_types.h | 11 +++
1 file changed, 11 insertions(+)
diff --git a/src/compiler/glsl_types.h b/src/compiler/glsl_types.h
index 55faac2..2857dc9 100644
--- a/src/compiler/gls
From: Nicolai Hähnle
---
src/compiler/nir/nir_lower_wpos_ytransform.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/src/compiler/nir/nir_lower_wpos_ytransform.c
b/src/compiler/nir/nir_lower_wpos_ytransform.c
index 873d259..771c6ff 100644
--- a/src/compiler/nir/nir_lo
From: Nicolai Hähnle
This pass is a replacement for the nir_lower_samplers pass, which has the
advantage of keeping sampler references as derefs. This allows a unified
treatment of texture instructions and image intrinsics in the backend.
---
src/compiler/Makefile.sources | 1
Hi all,
A few weeks ago I mentioned that I was looking into support of
GL_ARB_gl_spirv. Based on that feedback, I continued an approach where
Mesa translates SPIR-V to NIR during shader specialization, and hands
off the lowered NIR to the driver. This makes it easier for a lot of
the program resou
From: Nicolai Hähnle
---
src/compiler/nir/nir.h | 19 +++
1 file changed, 19 insertions(+)
diff --git a/src/compiler/nir/nir.h b/src/compiler/nir/nir.h
index ab7ba14..d1b90cb 100644
--- a/src/compiler/nir/nir.h
+++ b/src/compiler/nir/nir.h
@@ -2301,47 +2301,66 @@ void nir_metada
From: Nicolai Hähnle
This is an alternative lowering of default-block uniforms that doesn't
use uniform load intrinsics. Instead, it translates the loads to UBO
loads, allowing a simpler backend.
---
src/compiler/Makefile.sources| 1 +
src/compiler/nir/nir.h
From: Nicolai Hähnle
---
src/gallium/drivers/ddebug/dd_screen.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/src/gallium/drivers/ddebug/dd_screen.c
b/src/gallium/drivers/ddebug/dd_screen.c
index a5d2be1..3f6368c 100644
--- a/src/gallium/drivers/ddebug/dd_screen.c
+++ b/src/ga
From: Nicolai Hähnle
In that case, prog->nir must be assigned at the end.
---
src/mesa/state_tracker/st_glsl_to_nir.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/mesa/state_tracker/st_glsl_to_nir.cpp
b/src/mesa/state_tracker/st_glsl_to_nir.cpp
index 524eefa..0a86
From: Nicolai Hähnle
Allows modifying a texture instruction's texture and sampler derefs.
---
src/compiler/nir/nir.c | 13 +
src/compiler/nir/nir.h | 2 ++
2 files changed, 15 insertions(+)
diff --git a/src/compiler/nir/nir.c b/src/compiler/nir/nir.c
index 491b908..f81f4ee 100644
-
From: Nicolai Hähnle
Allow the driver to configure the state tracker's nir_lower_tex pass when
using NIR.
---
src/gallium/docs/source/screen.rst | 20
src/gallium/drivers/freedreno/freedreno_screen.c | 14 ++
src/gallium/drivers/vc4/vc4_screen.c
From: Nicolai Hähnle
We have to mark the additional shader input as used, otherwise it will
be eliminated, and we have to setup its index correctly.
This is a bit of a hack, but so is everything surrounding edgeflag
passthrough.
---
src/mesa/state_tracker/st_glsl_to_nir.cpp | 3 +++
src/mesa/st
1 - 100 of 278 matches
Mail list logo
