Perf data?
On Fri, Jul 10, 2015 at 6:41 PM, Timothy Arceri wrote:
> This implementation should be faster and there was no
> need to store a data field.
> ---
> src/glsl/ir_validate.cpp | 24
> 1 file changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/src/glsl/ir
On Fri, 2015-07-10 at 19:07 +1200, Chris Forbes wrote:
> Perf data?
I can create some if you like, but wasn't program/hash_table.c meant to die
along time ago [1] anyway.
[1] http://lists.freedesktop.org/archives/mesa-dev/2013-December/050524.html
>
> On Fri, Jul 10, 2015 at 6:41 PM, Timothy Ar
On Fri, 2015-07-10 at 19:07 +1200, Chris Forbes wrote:
> Perf data?
When the new hash table implementation was added to Mesa by Eric it
claimed to be much faster, see commits 35fd61bd99c1 and 72e55bb6888ff.
The set implementation seems to follow the same implementation strategy,
so I suppose it m
Hello Renaud,
I am going to comment a few things to improve your patch submission.
In the email subject we usually put [PATCH vN] to indicate it is the
Nth version of the patch, so the reviewers don't get confused about
which version is the more recent. You can do that by editing the .patch
file
Sent a new patch with subject [Mesa-dev] [PATCH] clover: Pass image
attributes to the kernel.
2015.07.06. 17:58 ezt írta ("Zoltán Gilián" ):
> > Hint: you'll need new
> > module::argument::semantic enums
>
> I see. Reworked it a bit.
>
> On Mon, Jul 6, 2015 at 1:13 PM, Francisco Jerez
> wrote:
>
From: Samuel Iglesias Gonsalvez
Section 4.3.7 "Buffer Variables" of the GLSL 4.30 spec:
"Buffer variables cannot have initializers."
v2:
- Rewrite error message (Jordan)
Signed-off-by: Samuel Iglesias Gonsalvez
Reviewed-by: Jordan Justen
---
src/glsl/ast_to_hir.cpp | 9 +
1 file
From: Kristian Høgsberg
This will be used to identify buffer variables inside shader storage
buffer objects, which are very similar to uniforms except for a few
differences, most important of which is that they are writable.
Since buffer variables are so similar to uniforms, we will almost alway
Reviewed-by: Jordan Justen
---
src/glsl/nir/glsl_to_nir.cpp | 4
src/glsl/nir/nir.h | 1 +
src/glsl/nir/nir_lower_atomics.c | 3 ++-
src/glsl/nir/nir_lower_io.c | 9 ++---
src/glsl/nir/nir_print.c | 5 +++--
src/glsl/nir/nir_validate.c | 6 --
6 fi
Since this now checks if a variable is inside a uniform or a shader
storage block.
Reviewed-by: Jordan Justen
---
src/glsl/ast_to_hir.cpp| 2 +-
src/glsl/ir.h | 5 +++--
src/glsl/link_uniform_block_active_visitor.cpp | 6 +++---
src/glsl/l
As discussed with Jordan, this v3-part1 series contains a good part of the
frontend stuff (most of which has already been reviewed). The idea is to get
this landed ahead, since some of the remaining patches (specifically the i965
backend stuff) depends on ther patches from Curro that have not lande
v2:
- Fix error message (Jordan)
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c | 110 ++
1 file changed, 110 insertions(+)
diff --git a/src/mesa/main/bufferobj.c b/src/mesa/main/bufferobj.c
index 27638ca..0a9ffe4 100644
--- a/src/mesa/main/
Otherwise we can lose writes into the buffers backing the variables.
Reviewed-by: Jordan Justen
---
src/glsl/opt_tree_grafting.cpp | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/src/glsl/opt_tree_grafting.cpp b/src/glsl/opt_tree_grafting.cpp
index d47613c..7f2ee6c 1
From: Samuel Iglesias Gonsalvez
Signed-off-by: Samuel Iglesias Gonsalvez
Reviewed-by: Jordan Justen
---
src/glsl/glcpp/glcpp-parse.y| 3 ++
src/glsl/glsl_parser_extras.cpp | 63 +
src/glsl/glsl_parser_extras.h | 7 +
src/mesa/main/extensions.
v2:
- Remove the extra spaces (Jordan)
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c | 11 +++
1 file changed, 11 insertions(+)
diff --git a/src/mesa/main/bufferobj.c b/src/mesa/main/bufferobj.c
index 2d70f7b..ee920a9 100644
--- a/src/mesa/main/bufferobj.c
+++ b/src/mesa/main
From: Samuel Iglesias Gonsalvez
Due to GL_ARB_shader_storage_buffer_object extension, shader storage blocks
have the same limitations as uniform blocks.
This patch fixes the corresponding error messages.
Signed-off-by: Samuel Iglesias Gonsalvez
Reviewed-by: Jordan Justen
---
src/glsl/ast_to_
This includes the array of bindings, the current buffer bound to the
GL_SHADER_STORAGE_BUFFER target and a set of general limits and default
values for shader storage buffers.
v2:
- Use spec values for the new defined constants (Jordan)
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c |
v2:
- Fix indention, used tabs instead of whitespaces. (Jordan)
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c | 19 +++
1 file changed, 19 insertions(+)
diff --git a/src/mesa/main/bufferobj.c b/src/mesa/main/bufferobj.c
index c5d4ada..2d70f7b 100644
--- a/src/mesa/mai
From: Samuel Iglesias Gonsalvez
See GLSL 4.30 spec, section 4.4.5 "Uniform and Shader Storage Block
Layout Qualifiers".
v2:
- Add whitespace in an error message. Delete period '.' at the end of that
error message (Jordan).
Signed-off-by: Samuel Iglesias Gonsalvez
Reviewed-by: Jordan Justen
--
From: Kristian Høgsberg
This is used to identify shader storage buffer interface blocks where
buffer variables are declared.
Reviewed-by: Jordan Justen
---
src/glsl/ast.h | 1 +
src/glsl/ast_to_hir.cpp | 14 ++
src/glsl/ast_type.cpp | 3 ++-
src
From: Samuel Iglesias Gonsalvez
Section 4.3.7 "Buffer Variables", GLSL 4.30 spec:
"Buffer variables may only be declared inside interface blocks
(section 4.3.9 “Interface Blocks”), which are then referred to as
shader storage blocks. It is a compile-time error to declare buffer
variables at glob
From: Samuel Iglesias Gonsalvez
v2:
- Set MaxShaderStorageBlocks to 8.
Signed-off-by: Samuel Iglesias Gonsalvez
Reviewed-by: Jordan Justen
---
src/mesa/main/context.c | 2 ++
src/mesa/main/mtypes.h | 3 +++
2 files changed, 5 insertions(+)
diff --git a/src/mesa/main/context.c b/src/mesa/mai
Since the backing storage for these is shared we cannot ensure that
the value won't change by writes from other threads. Normally SSBO
accesses are not guaranteed to be syncronized with other threads,
except when memoryBarrier is used. So, we might be able to optimize
some SSBO accesses, but for no
Since the backing storage for these is shared we cannot ensure that
the value won't change by writes from other threads. Normally SSBO
accesses are not guaranteed to be syncronized with other threads,
except when memoryBarrier is used. So, we might be able to optimize
some SSBO accesses, but for no
Reviewed-by: Jordan Justen
---
src/glsl/link_uniform_block_active_visitor.cpp | 1 +
src/glsl/link_uniform_block_active_visitor.h | 1 +
src/glsl/link_uniform_blocks.cpp | 4
src/mesa/main/mtypes.h | 5 +
4 files changed, 11 insertions(+)
diff --g
If we kill dead assignments we lose the buffer writes.
Also, we never kill UBO declarations even if they are never referenced
by the shader, they are always considered active. Although the spec
does not seem say this specifically for SSBOs, it is probably implied
since SSBOs are pretty much the sa
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c | 56 +++
1 file changed, 56 insertions(+)
diff --git a/src/mesa/main/bufferobj.c b/src/mesa/main/bufferobj.c
index 0a9ffe4..c3548b5 100644
--- a/src/mesa/main/bufferobj.c
+++ b/src/mesa/main/buf
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c | 37 +
1 file changed, 37 insertions(+)
diff --git a/src/mesa/main/bufferobj.c b/src/mesa/main/bufferobj.c
index c3548b5..4e25a72 100644
--- a/src/mesa/main/bufferobj.c
+++ b/src/mesa/main/bufferobj.c
@
v2:
- Add space before const (Jordan)
Reviewed-by: Jordan Justen
---
src/mesa/main/bufferobj.c | 142 ++
src/mesa/main/mtypes.h| 7 +++
2 files changed, 149 insertions(+)
diff --git a/src/mesa/main/bufferobj.c b/src/mesa/main/bufferobj.c
index e
From: Kristian Høgsberg
Reviewed-by: Jordan Justen
---
src/glsl/link_interface_blocks.cpp | 15 ---
src/glsl/link_uniform_initializers.cpp | 3 ++-
src/glsl/link_uniforms.cpp | 8 +---
src/glsl/linker.cpp| 4 ++--
4 files changed, 21 insert
From: Samuel Iglesias Gonsalvez
Extend the existing lower_ubo_reference pass to also detect SSBO writes
and lower them to __intrinsic_store_ssbo intrinsics.
Signed-off-by: Samuel Iglesias Gonsalvez
---
src/glsl/lower_ubo_reference.cpp | 441 +++
1 file chang
Since the backing storage for these is shared we cannot ensure that
the value won't change by writes from other threads. Normally SSBO
accesses are not guaranteed to be syncronized with other threads,
except when memoryBarrier is used. So, we might be able to optimize
some SSBO accesses, but for no
From: Samuel Iglesias Gonsalvez
Extend the existing lower_ubo_reference pass to also detect SSBO loads
and lower them to __intrinsic_load_ssbo intrinsics.
Signed-off-by: Samuel Iglesias Gonsalvez
---
src/glsl/lower_ubo_reference.cpp | 73 +++-
1 file changed
On Fri, 2015-07-10 at 12:13 +0200, Iago Toral Quiroga wrote:
> As discussed with Jordan, this v3-part1 series contains a good part of the
> frontend stuff (most of which has already been reviewed). The idea is to get
> this landed ahead, since some of the remaining patches (specifically the i965
>
Reviewed-by: Marek Olšák
Marek
On Wed, Jul 8, 2015 at 10:34 PM, Rob Clark wrote:
> From: Rob Clark
>
> No where was it spelled out that the state tracker may expect the pipe
> driver to unref the old fence.
>
> Signed-off-by: Rob Clark
> ---
> src/gallium/include/pipe/p_context.h | 8 +++
I wonder... do we still need XA considering that everybody can just
use glamor instead?
Marek
On Wed, Jul 8, 2015 at 7:39 PM, Rob Clark wrote:
> From: Rob Clark
>
> XA was never unref'ing last_fence in the various call paths to
> pipe->flush(). Add this to xa_context_flush() and update the ot
Shouldn't this stuff be in src/util?
Marek
On Wed, Jul 8, 2015 at 11:07 PM, Ben Widawsky
wrote:
> WARNING: No perf data, please keep reading though)
>
> This implements the suggestion provided by the paper, "Fast USWC to WB Memory
> Copy"
> (https://software.intel.com/en-us/articles/copying-acce
Am 10.07.2015 um 05:44 schrieb Michel Dänzer:
> On 10.07.2015 05:13, Emil Velikov wrote:
>> The original code only half considered hyperz as an option. As per
>> previous commit "major != 2 cannot occur" we can simply things, and
>> allow users to set the option if they choose to do so.
>>
>> Signe
Jason Ekstrand writes:
> On Jul 9, 2015 7:57 AM, "Francisco Jerez" wrote:
>>
>> We were passing src0 alpha and oMask in reverse order. There seems to
>> be no good way to pass them in the correct order to the new-style
>> LOAD_PAYLOAD (how surprising) because src0 alpha is per-channel while
>>
well, freedreno and vmware ddx still use XA.. they both could probably
be ported to use glamor instead, but that hasn't been done yet..
At least for freedreno with upstream drm/kms driver, one can just use
-modesetting ddx instead. But that doesn't work w/ android fbdev
driver. I need to check a
This looks all pretty reasonable, though I can't really figure out if we
could translate from d3d11 to that. In particular, it's UAVs for all of
buffers and images, and I'm not entirely sure if it's actually possible
to figure out which UAV entries are a shader buffer and which ones are a
shader im
https://bugs.freedesktop.org/show_bug.cgi?id=91292
Bug ID: 91292
Summary: glVertexAttribDivisor not working Linux 10.5.8 Intel
HD 5500?
Product: Mesa
Version: 10.5
Hardware: x86-64 (AMD64)
OS: Linux (All)
https://bugs.freedesktop.org/show_bug.cgi?id=91292
--- Comment #1 from Hugh Fisher ---
Created attachment 117030
--> https://bugs.freedesktop.org/attachment.cgi?id=117030&action=edit
Expected output from running program
--
You are receiving this mail because:
You are the QA Contact for the bu
On Thu, Jul 9, 2015 at 8:43 AM, Dave Airlie wrote:
> From: Glenn Kennard
>
> This adds the main chunk of the geometry shader multiple stream
> support to the r600 driver.
>
> Glenn wrote the original pass, and I took his code and hacked
> it into a working state.
>
> Signed-off-by: Dave Airlie
>
On Thu, Jul 9, 2015 at 4:11 PM, Francisco Jerez wrote:
> Ilia Mirkin writes:
>
>> FYI there's already a lowering pass that does this in the GLSL IR
>> (CARRY_TO_ARITH in lower_instructions). Perhaps the right place to do
>> this is NIR though, just wanted to let you know.
>>
> Ah, I wasn't aware
For Evergreen (the first DX11 radeon), which only supports 12 writable
resources, we'll only expose 8 images, 8 shader storage buffers and 8
atomic counter buffers. There will be a CAP saying that only a total
of 12 is allowed to be used by a shader and all bindings slots will
thus be considered vi
https://bugs.freedesktop.org/show_bug.cgi?id=91292
Neil Roberts changed:
What|Removed |Added
CC||n...@linux.intel.com
Summary|g
https://bugs.freedesktop.org/show_bug.cgi?id=91292
Neil Roberts changed:
What|Removed |Added
Status|NEW |ASSIGNED
Component|Mesa core
On 06/30/2015 06:58 PM, Jason Ekstrand wrote:
> On Fri, Jun 26, 2015 at 1:06 AM, Eduardo Lima Mitev wrote:
>> This patch makes public the is_scalar_shader_stage() method in brw_shader,
>> and
>> renames it to brw_compiler_is_scalar_shader_stage(). The plan is to later
>> reuse it
>> in brw_nir,
On Fri, Jul 10, 2015 at 5:25 AM, Francisco Jerez wrote:
> Jason Ekstrand writes:
>
>> On Jul 9, 2015 7:57 AM, "Francisco Jerez" wrote:
>>>
>>> We were passing src0 alpha and oMask in reverse order. There seems to
>>> be no good way to pass them in the correct order to the new-style
>>> LOAD_PAY
Jason Ekstrand writes:
> On Fri, Jul 10, 2015 at 5:25 AM, Francisco Jerez
> wrote:
>> Jason Ekstrand writes:
>>
>>> On Jul 9, 2015 7:57 AM, "Francisco Jerez" wrote:
We were passing src0 alpha and oMask in reverse order. There seems to
be no good way to pass them in the correct
When the edge flag element is enabled then the elements are slightly
reordered so that the edge flag is always the last one. This was
confusing the code to upload the 3DSTATE_VF_INSTANCING state because
that is uploaded with a separate loop which has an instruction for
each element. The indices use
Booleans are represented as 0/-1 on modern hardware which means we can
just negate them to convert them into a numeric type. Negation has
the benefit that it can be implemented using a source modifier which
can easily be propagated into some other instruction. shader-db
results on HSW:
total in
This gets rid of two no16() fall-backs and should allow better
scheduling of the generated IR. There are no uses of usubBorrow() or
uaddCarry() in shader-db so no changes are expected. However the
"arb_gpu_shader5/execution/built-in-functions/fs-usubBorrow" and
"arb_gpu_shader5/execution/built-in
On 10 July 2015 at 13:18, Roland Scheidegger wrote:
> Am 10.07.2015 um 05:44 schrieb Michel Dänzer:
>> On 10.07.2015 05:13, Emil Velikov wrote:
>>> The original code only half considered hyperz as an option. As per
>>> previous commit "major != 2 cannot occur" we can simply things, and
>>> allow u
Inspired (copied) from Marek's commit for egl/x11
commit 0b56e23e7f3(egl/dri2: use the correct screen index)
Cc: 10.6
Cc: Marek Olšák
Signed-off-by: Emil Velikov
---
src/gallium/auxiliary/vl/vl_winsys_dri.c | 18 +++---
1 file changed, 15 insertions(+), 3 deletions(-)
diff --git a
... and update the documentation to reflect reality.
null and gdi are gone, and surfaceless is a recent addition.
Signed-off-by: Emil Velikov
---
configure.ac | 3 ---
docs/egl.html | 6 +++---
2 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/configure.ac b/configure.ac
index 9ffd
All of these are already defined in the headers provided.
Signed-off-by: Emil Velikov
---
src/egl/main/eglapi.c | 36 +---
src/egl/main/eglapi.h | 24
src/egl/main/eglfallbacks.c | 6 --
3 files changed, 1 insertion(+), 65
Cc: Axel Davy
Signed-off-by: Emil Velikov
---
configure.ac| 2 +-
src/egl/drivers/dri2/platform_wayland.c | 20 ++--
2 files changed, 3 insertions(+), 19 deletions(-)
diff --git a/configure.ac b/configure.ac
index e648a8d..e6efa01 100644
--- a/config
The macro HAVE_MKOSTEMP was never defined.
Cc: Axel Davy
Signed-off-by: Emil Velikov
---
src/egl/drivers/dri2/platform_wayland.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/src/egl/drivers/dri2/platform_wayland.c
b/src/egl/drivers/dri2/platform_wayland.c
index 9005eb7..bdccee3 100
Prompt at configure time if it's missing otherwise we'll fail later on
in the build. Remove ambiguous HAVE_LIBDRM guard.
Cc: 10.6
Signed-off-by: Emil Velikov
---
configure.ac| 3 +++
src/egl/drivers/dri2/egl_dri2.c | 2 --
2 files changed, 3 insertions(+), 2 deletions(-)
di
Move the code around rather than having it scattered. No functional
change.
Signed-off-by: Emil Velikov
---
src/egl/drivers/dri2/egl_dri2.c | 210 +++-
1 file changed, 102 insertions(+), 108 deletions(-)
diff --git a/src/egl/drivers/dri2/egl_dri2.c b/src/egl/
... and update the documentation to reflect reality.
null and gdi are gone, and surfaceless is a recent addition.
Signed-off-by: Emil Velikov
---
configure.ac | 3 ---
docs/egl.html | 6 +++---
2 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/configure.ac b/configure.ac
index 9ffd
Signed-off-by: Emil Velikov
---
src/egl/drivers/dri2/platform_wayland.c | 26 +-
1 file changed, 13 insertions(+), 13 deletions(-)
diff --git a/src/egl/drivers/dri2/platform_wayland.c
b/src/egl/drivers/dri2/platform_wayland.c
index 1e12760..0b1adf8 100644
--- a/src/egl/d
https://bugs.freedesktop.org/show_bug.cgi?id=66346
--- Comment #21 from Emil Velikov ---
(In reply to Vinson Lee from comment #20)
> (In reply to José Fonseca from comment #19)
> > (In reply to Vinson Lee from comment #18)
> > > mesa: 0166b4c165271bd7525a91049e58e390cb596c60 (master 10.7.0-devel)
Hi folks,
I've pushed the latest version of my shader-cache work to a branch named
"shader-cache" at:
git://people.freedesktop.org/~cworth/mesa
I've rebased this against the latest master branch and verified that it
at least compiles and works with at least some trivial test programs.
I
Zoltan Gilian writes:
> Read-only and write-only image arguments are recognized and
> distinguished.
> Attributes of the image arguments are passed to the kernel as implicit
> arguments.
Thanks, this looks much better. One thing that still seems kind of
unfortunate is the fact that you've added
On Fri, Jul 10, 2015 at 10:06 AM, Francisco Jerez wrote:
> Booleans are represented as 0/-1 on modern hardware which means we can
> just negate them to convert them into a numeric type. Negation has
> the benefit that it can be implemented using a source modifier which
> can easily be propagated
On Fri, Jul 10, 2015 at 10:49 AM, Emil Velikov wrote:
> The macro HAVE_MKOSTEMP was never defined.
>
> Cc: Axel Davy
> Signed-off-by: Emil Velikov
> ---
> src/egl/drivers/dri2/platform_wayland.c | 6 --
> 1 file changed, 6 deletions(-)
>
> diff --git a/src/egl/drivers/dri2/platform_wayland.
https://bugs.freedesktop.org/show_bug.cgi?id=91263
--- Comment #2 from Marek Olšák ---
Proposed fix:
http://patchwork.freedesktop.org/patch/52969/
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
__
On Fri, Jul 10, 2015 at 4:20 AM, Marek Olšák wrote:
> Shouldn't this stuff be in src/util?
I guess? I'm not opposed to moving it, if other drivers can make use
of it, but I don't want this patch to do the move.
I wrote it in November 2013 and src/util only came into being in
August 2014 if you'r
I'm sure this doesn't build. "dri2_dpy" is not declared and "screen" is int.
Marek
On Fri, Jul 10, 2015 at 7:46 PM, Emil Velikov wrote:
> Inspired (copied) from Marek's commit for egl/x11
> commit 0b56e23e7f3(egl/dri2: use the correct screen index)
>
> Cc: 10.6
> Cc: Marek Olšák
> Signed-off-b
Matt Turner writes:
> On Fri, Jul 10, 2015 at 10:06 AM, Francisco Jerez
> wrote:
>> Booleans are represented as 0/-1 on modern hardware which means we can
>> just negate them to convert them into a numeric type. Negation has
>> the benefit that it can be implemented using a source modifier whi
Previously OUT_BATCH was just a macro around an inline function which
does
brw->batch.map[brw->batch.used++] = dword;
When making consecutive calls to intel_batchbuffer_emit_dword() the
compiler isn't able to recognize that we're writing consecutive memory
locations or that it doesn't need to
BEGIN_BATCH() and ADVANCE_BATCH() will contain "do {" and "} while (0)"
respectively to allow declaring local variables used by intervening
OUT_BATCH macros. As such, BEGIN_BATCH() and ADVANCE_BATCH() need to be
in the same control flow.
---
src/mesa/drivers/dri/i965/brw_draw.c | 4 ++--
1 file ch
Its uses of OUT_BATCH will need a local variable defined by BEGIN_BATCH.
Increases .text size by 528 bytes.
---
src/mesa/drivers/dri/i965/intel_blit.c | 55 --
1 file changed, 25 insertions(+), 30 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/intel_blit.c
b
Its uses of OUT_BATCH will need a local variable defined by BEGIN_BATCH.
Increases .text size by 8 bytes.
---
src/mesa/drivers/dri/i965/brw_draw_upload.c | 79 +
1 file changed, 36 insertions(+), 43 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_draw_upload.c
So that everything writing to the batch between BEGIN_BATCH() and
ADVANCE_BATCH() goes through OUT_BATCH.
---
src/mesa/drivers/dri/i965/intel_batchbuffer.c | 30 ++-
src/mesa/drivers/dri/i965/intel_batchbuffer.h | 34 ++-
2 files changed, 30 insertions(+
On Thu, Jul 9, 2015 at 11:00 AM, Ben Widawsky
wrote:
> While implementing the workaround in the previous patch I noticed things were
> starting to get a bit messy. Since gen8 works differently enough from gen7, I
> thought splitting it out with be good.
>
> While here, get rid of gen8 MOCS which d
https://bugs.freedesktop.org/show_bug.cgi?id=91290
Vinson Lee changed:
What|Removed |Added
Keywords||regression
CC|
On Wed, Jul 8, 2015 at 10:58 PM, Zhenyu Wang wrote:
> NOP actually has no compact version, but we use it for instruction
> alignment for compact kernel. Although it seems working on HW, it is
> illegal and might not be valid for any future one.
>
> This trys to get a temporary compact instruction
On Fri, Jul 10, 2015 at 9:53 AM, Francisco Jerez wrote:
> Jason Ekstrand writes:
>
>> On Fri, Jul 10, 2015 at 5:25 AM, Francisco Jerez
>> wrote:
>>> Jason Ekstrand writes:
>>>
On Jul 9, 2015 7:57 AM, "Francisco Jerez" wrote:
>
> We were passing src0 alpha and oMask in reverse ord
On 10/07/15 19:22, Matt Turner wrote:
> On Fri, Jul 10, 2015 at 10:49 AM, Emil Velikov
> wrote:
>> The macro HAVE_MKOSTEMP was never defined.
>>
>> Cc: Axel Davy
>> Signed-off-by: Emil Velikov
>> ---
>> src/egl/drivers/dri2/platform_wayland.c | 6 --
>> 1 file changed, 6 deletions(-)
>>
>>
Inspired (copied) from Marek's commit for egl/x11
commit 0b56e23e7f3(egl/dri2: use the correct screen index)
v2: Fix copy/pasta errors.
Cc: 10.6
Cc: Marek Olšák
Signed-off-by: Emil Velikov
---
src/gallium/auxiliary/vl/vl_winsys_dri.c | 17 +
1 file changed, 13 insertions(+), 4
On 10 July 2015 at 19:33, Marek Olšák wrote:
> I'm sure this doesn't build. "dri2_dpy" is not declared and "screen" is int.
>
In all honesty I "wrote" it at 3am, but that is no excuse for such a
goof-up. A correct version should be out.
Thanks,
Emil
___
Fixes the following piglits:
fbo-alphatest-nocolor
fbo-alphatest-nocolor-ff
Signed-off-by: Ilia Mirkin
Cc: mesa-sta...@lists.freedesktop.org
---
The nv50 bits need testing, only have a GK208 on-hand. Will be sure to test
before pushing.
src/gallium/drivers/nouveau/nv50/nv50_state_validate.
On 2015-07-10 03:13:44, Iago Toral Quiroga wrote:
> From: Samuel Iglesias Gonsalvez
>
> Extend the existing lower_ubo_reference pass to also detect SSBO writes
> and lower them to __intrinsic_store_ssbo intrinsics.
>
> Signed-off-by: Samuel Iglesias Gonsalvez
> ---
> src/glsl/lower_ubo_referen
Reviewed-by: Jordan Justen
On 2015-07-10 03:13:45, Iago Toral Quiroga wrote:
> From: Samuel Iglesias Gonsalvez
>
> Extend the existing lower_ubo_reference pass to also detect SSBO loads
> and lower them to __intrinsic_load_ssbo intrinsics.
>
> Signed-off-by: Samuel Iglesias Gonsalvez
> ---
>
Am 10.07.2015 um 19:41 schrieb Emil Velikov:
> On 10 July 2015 at 13:18, Roland Scheidegger wrote:
>> Am 10.07.2015 um 05:44 schrieb Michel Dänzer:
>>> On 10.07.2015 05:13, Emil Velikov wrote:
The original code only half considered hyperz as an option. As per
previous commit "major != 2
On Wed, Jul 1, 2015 at 11:51 AM, Connor Abbott wrote:
> sources with file == HW_REG get all their information from the
> fixed_hw_reg field, so we need to get the stride and type from there
> when computing the size.
>
> Signed-off-by: Connor Abbott
> ---
Patches 1 and 2 are
Reviewed-by: Matt T
On Fri, Jul 10, 2015 at 12:03:54PM -0700, Matt Turner wrote:
> On Thu, Jul 9, 2015 at 11:00 AM, Ben Widawsky
> wrote:
> > While implementing the workaround in the previous patch I noticed things
> > were
> > starting to get a bit messy. Since gen8 works differently enough from gen7,
> > I
> > th
On Wed, Jul 1, 2015 at 7:28 PM, Kenneth Graunke wrote:
> Paul's original code had emit_control_data_bits() skip the URB write if
> vertex_count was 0. This meant wrapping every control data write in a
> conditional write.
>
> We accumulate control data bits in a single UD (32-bit) register. For
Confirmed that git show -w shows only the removal of matching braces.
Reviewed-by: Matt Turner
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
On 2015-06-03 00:01:23, Iago Toral Quiroga wrote:
> Since these are a special kind of UBOs we emit them together reusing the
> same infrastructure, however, we use a RAW surface so we can reuse
> existing untyped read/write/atomic messages which include a pixel mask
> header that we need to set to
Reviewed-by: Jordan Justen
On 2015-06-03 00:01:24, Iago Toral Quiroga wrote:
> ---
> src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> b/src/mesa/drivers/dri/i965/brw_vec
Rob Clark writes:
> well, freedreno and vmware ddx still use XA.. they both could probably
> be ported to use glamor instead, but that hasn't been done yet..
>
> At least for freedreno with upstream drm/kms driver, one can just use
> -modesetting ddx instead. But that doesn't work w/ android fbd
Reviewed-by: Jordan Justen
On 2015-06-03 00:01:25, Iago Toral Quiroga wrote:
> Buffer variables are the same as uniforms, only that read/write, so we want
> the same treatment.
> ---
> src/mesa/drivers/dri/i965/brw_fs_vector_splitting.cpp | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a
On Fri, Jul 10, 2015 at 6:20 PM, Eric Anholt wrote:
> Rob Clark writes:
>
>> well, freedreno and vmware ddx still use XA.. they both could probably
>> be ported to use glamor instead, but that hasn't been done yet..
>>
>> At least for freedreno with upstream drm/kms driver, one can just use
>> -m
Surely the *right* thing would be to have the correct order expressed
in brw->vb.*, instead so you don't have this workaround in multiple
places.
As a minimal fix for stable though, this seems OK, so -
Reviewed-by: Chris Forbes
On Sat, Jul 11, 2015 at 5:04 AM, Neil Roberts wrote:
> When the
s/interpolater/interpolator/g
On Fri, Jul 10, 2015 at 1:31 AM, Neil Roberts wrote:
> For centroid interpolation we can just directly use the values set up
> in the shader payload instead of querying the pixel interpolator. To
> do this we need to modify brw_compute_barycentric_interp_modes to
> d
Nitpicks aside, I don't think this is a great idea now that you've got
the SKL PI working.
I also think it's broken -- you need to arrange to have the centroid
barycentric coords delivered to the FS thread, which won't be
happening if this is the *only* use of them. Masked in the tests,
because th
1 - 100 of 106 matches
Mail list logo
