On 9 March 2018 at 16:53, Eric Engestrom wrote:
> On Friday, 2018-03-09 15:24:18 +, Emil Velikov wrote:
>> From: Emil Velikov
>>
>> Instead of having multiple guards littered through the code, simply
>> introduce static inline no-op
On 7 March 2018 at 23:13, Dylan Baker wrote:
> meson got the same fix.
>
> cc: Emil Veliov
> Signed-off-by: Dylan Baker
> ---
>
> I don't know hat the "right" fix is here Emil, I just picked one and applied
> it
> to both
https://bugs.freedesktop.org/show_bug.cgi?id=105436
Michel Dänzer changed:
What|Removed |Added
CC|
https://bugs.freedesktop.org/show_bug.cgi?id=105436
almos changed:
What|Removed |Added
Assignee|dri-devel@lists.freedesktop
The code to handle mat multipication by a scalar tries to pick either
imul or fmul depending on whether the matrix is float or integer.
However it was doing this by checking whether the base type is float.
This was making it choose the int path for doubles (and presumably
float16s).
---
This was
On Tue, Mar 13, 2018 at 8:33 AM, Emil Velikov
wrote:
> On 13 March 2018 at 15:02, Jason Ekstrand wrote:
> > On Tue, Mar 13, 2018 at 7:56 AM, Emil Velikov
> > wrote:
> >>
> >> On 12 March 2018 at 08:40, Iago Toral Quiroga
Daniel Vetter writes:
> On Mon, Mar 12, 2018 at 11:54:45PM -0700, Kenneth Graunke wrote:
>> On Friday, March 9, 2018 12:12:28 PM PDT Mark Janes wrote:
>> [snip]
>> > I've been doing this for Intel. Developers are on the hook to fix their
>> > bugs, but you can't make them do
Piglit+CTS+dEQP look good on Polaris12 with Mesa 17.3 and LLVM 6.0.
Marek
On Thu, Mar 8, 2018 at 6:22 AM, Emil Velikov wrote:
> [Adding Andres, Juan since they'll be doing the next 17.3]
>
> On 7 March 2018 at 16:35, Marek Olšák wrote:
>> I can test
On 13 March 2018 at 11:46, Tomasz Figa wrote:
> On Thu, Mar 8, 2018 at 7:39 AM, Lepton Wu wrote:
>> If user calls map twice for kms_sw_displaytarget, the first mapped
>> buffer could get leaked. Instead of calling mmap every time, just
>> reuse previous
Thanks Emil,
I'll include this in the next version :)
On 13/03/18 15:17, Emil Velikov wrote:
From: Emil Velikov
Cc: Lionel Landwerlin
---
src/mesa/drivers/dri/i965/Android.mk | 11 +---
On 13 March 2018 at 15:02, Jason Ekstrand wrote:
> On Tue, Mar 13, 2018 at 7:56 AM, Emil Velikov
> wrote:
>>
>> On 12 March 2018 at 08:40, Iago Toral Quiroga wrote:
>> > af5f2322d0c64 addressed this for extension commands, but
On 13 March 2018 at 14:44, Lionel Landwerlin
wrote:
> Performance metric numbers are calculated the following way :
>
>- out of the 256 bytes long OA reports, we accumulate the deltas
> into an array of uint64_t
>
>- the equations' generated code reads
From: Emil Velikov
Cc: Lionel Landwerlin
---
src/mesa/drivers/dri/i965/Android.mk | 11 +---
src/mesa/drivers/dri/i965/Makefile.am | 24 --
src/mesa/drivers/dri/i965/Makefile.sources | 40
On Tue, Mar 13, 2018 at 7:56 AM, Emil Velikov
wrote:
> On 12 March 2018 at 08:40, Iago Toral Quiroga wrote:
> > af5f2322d0c64 addressed this for extension commands, but the spec
> mandates
> > this behavior also for core API commands. From the Vulkan
On 12 March 2018 at 08:40, Iago Toral Quiroga wrote:
> af5f2322d0c64 addressed this for extension commands, but the spec mandates
> this behavior also for core API commands. From the Vulkan spec,
> Table 2. vkGetDeviceProcAddr behavior:
>
> device pname
On Monday, 2018-03-12 11:23:23 -0700, Dylan Baker wrote:
> Meson's compiler.has_header is completely useless, it only checks that a
> header exists, not whether it's usable. This creates problems if a
> header contains a conditional #error declaration, like so:
>
> > #if __x86_64__
> > # error
The equation code computes a float (percentage) yet the return type
was an uint64_t.
Signed-off-by: Lionel Landwerlin
---
src/mesa/drivers/dri/i965/brw_oa_hsw.xml | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
Performance metric numbers are calculated the following way :
- out of the 256 bytes long OA reports, we accumulate the deltas
into an array of uint64_t
- the equations' generated code reads the accumulated uint64_t
deltas and normalizes them for a particular platforms
Our
Hi all,
While adding new metrics for newer generations I wondered how much
code we could save by reusing the identical equation functions across
metrics & generations. Turns out quite a bit, the saving in patch 2 is
about 4Mb.
We might be able to improve further by sharing similar register
On 2018-03-12 08:05 PM, Dave Airlie wrote:
> On 13 March 2018 at 03:59, Marek Olšák wrote:
>> This is good, though some older distros only have libxcb 1.11.
>
> On those distros you likely just want to --disable-dri3 anyways.
Why would you? DRI3 has been working fine with XCB
both are
Reviewed-by: Jason Ekstrand
On Tue, Mar 13, 2018 at 4:12 AM, Iago Toral Quiroga
wrote:
> af5f2322d0c64 addressed this for extension commands, but the spec mandates
> this behavior also for core API commands. From the Vulkan spec,
> Table 2.
On Mon, Mar 12, 2018 at 4:20 PM, Dave Airlie wrote:
> On 13 March 2018 at 05:58, Marek Olšák wrote:
>> On Mon, Mar 12, 2018 at 3:05 PM, Dave Airlie wrote:
>>> On 13 March 2018 at 03:59, Marek Olšák wrote:
This is
Reviewed-by: Bas Nieuwenhuizen
for the series.
On Tue, Mar 13, 2018 at 3:06 PM, Samuel Pitoiset
wrote:
> Signed-off-by: Samuel Pitoiset
> ---
> src/amd/common/ac_nir_to_llvm.c | 6 +++---
>
Also replace ac_ by radv_.
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_nir_to_llvm.h | 72 --
src/amd/vulkan/radv_nir_to_llvm.c | 10 +++---
src/amd/vulkan/radv_pipeline.c| 6 ++--
src/amd/vulkan/radv_private.h |
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_nir_to_llvm.c | 4
src/amd/common/ac_nir_to_llvm.h | 6 +-
2 files changed, 1 insertion(+), 9 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_nir_to_llvm.c | 6 +++---
src/amd/common/ac_nir_to_llvm.h | 2 +-
src/amd/vulkan/radv_nir_to_llvm.c | 16
3 files changed, 12 insertions(+), 12 deletions(-)
diff --git
This is RADV specific code.
Signed-off-by: Samuel Pitoiset
---
src/amd/Makefile.sources | 2 -
src/amd/common/ac_nir_to_llvm.h| 1 -
src/amd/common/ac_shader_info.h| 76 --
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_nir_to_llvm.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 3a9cb7b920..dff3bca713 100644
---
Also replace ac_ by radv_.
Signed-off-by: Samuel Pitoiset
---
src/amd/common/ac_nir_to_llvm.h | 97 ---
src/amd/vulkan/radv_cmd_buffer.c | 14 ++---
src/amd/vulkan/radv_nir_to_llvm.c| 29 ++-
On Mon, Mar 12, 2018 at 11:54:45PM -0700, Kenneth Graunke wrote:
> On Friday, March 9, 2018 12:12:28 PM PDT Mark Janes wrote:
> [snip]
> > I've been doing this for Intel. Developers are on the hook to fix their
> > bugs, but you can't make them do it. They have many pressures on them,
> > and a
On 03/12/2018 05:21 PM, Marek Olšák wrote:
On Mon, Mar 12, 2018 at 6:18 AM, Samuel Pitoiset
wrote:
On 03/11/2018 04:41 PM, Marek Olšák wrote:
On Thu, Mar 8, 2018 at 5:48 PM, Ian Romanick wrote:
On 03/08/2018 06:50 AM, Samuel Pitoiset
Hi Jason, can you have a look at these two? We have some tests failing
because of this.
Iago
On Wed, 2018-03-07 at 10:33 +0100, Iago Toral wrote:
> This patches are still waiting for review.
>
> On Wed, 2018-02-28 at 09:57 +0100, Iago Toral Quiroga wrote:
> > When multiview is active a subpass
From: Iago Toral Quiroga
In this case we don't want to do sign-extension, since the value is
interpreted as unsigned. If we want sign-extension, OpSConvert should
be used.
v2:
- Use src/dst bitsize to get the proper conversion opcode. (Jason)
---
src/compiler/spirv/vtn_alu.c
OpSConvert interprets the MSB of the unsigned value as the sign bit and
extends it to the new type. If we want to preserve the value, we need
to use OpUConvert opcode.
v2:
- No need to check dst type.
- Fix typo in comment.
v3:
- Use src/dst bitsize to get the proper conversion opcode. (Jason)
Hi Emil,
On 13 March 2018 at 11:40, Emil Velikov wrote:
> During development the version was bumped, yet the comment did not get
> an update.
>
> Cc: Daniel Stone
> Fixes: c80c08e2260 ("vulkan/wsi/x11: Add support for DRI3 v1.2")
> Signed-off-by:
On Thu, Mar 8, 2018 at 7:39 AM, Lepton Wu wrote:
> If user calls map twice for kms_sw_displaytarget, the first mapped
> buffer could get leaked. Instead of calling mmap every time, just
> reuse previous mapping. Since user could map same displaytarget with
> different flags,
From: Emil Velikov
Use the handy macro instead of hard coded numbers.
Cc: Daniel Stone
Fixes: c80c08e2260 ("vulkan/wsi/x11: Add support for DRI3 v1.2")
Signed-off-by: Emil Velikov
---
From: Emil Velikov
During development the version was bumped, yet the comment did not get
an update.
Cc: Daniel Stone
Fixes: c80c08e2260 ("vulkan/wsi/x11: Add support for DRI3 v1.2")
Signed-off-by: Emil Velikov
---
On Tue, Mar 13, 2018 at 3:10 AM, Emil Velikov wrote:
> On 12 March 2018 at 17:45, Lepton Wu wrote:
>> Ping. Any more comments or missing stuff to get this commited into master?
>>
> As things have changed a bit (the original map/unmap behaviour is
I was surprised that missing that didn't cause any trouble. It seems
that the way the trampoline code was setup would still generate
trampoline functions, but interpreting the dispatch object as a command
buffer, and both anv_queue and anv_cmd_buffer have their device pointer
field at exactly the
af5f2322d0c64 addressed this for extension commands, but the spec mandates
this behavior also for core API commands. From the Vulkan spec,
Table 2. vkGetDeviceProcAddr behavior:
device pnamereturn
--
(..)
v2:
- Add trampoline functions (Jason)
- Add an assertion for unhandled trampoline cases
---
src/intel/vulkan/anv_entrypoints_gen.py | 9 +++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/src/intel/vulkan/anv_entrypoints_gen.py
b/src/intel/vulkan/anv_entrypoints_gen.py
https://bugs.freedesktop.org/show_bug.cgi?id=105425
Michel Dänzer changed:
What|Removed |Added
Component|Mesa core
On 12/03/18 20:42, Jason Ekstrand wrote:
> On Mon, Mar 5, 2018 at 10:21 PM, Samuel Iglesias Gonsálvez
> > wrote:
>
> OpSConvert interprets the MSB of the unsigned value as the sign
> bit and
> extends it to the new type. If we want
Hi Mario,
On 12 March 2018 at 20:45, Mario Kleiner wrote:
> This way the wayland server can signal support for these formats
> to wayland EGL clients. This is currently used by nouveau for 10
> bpc support.
>
> Tested with glmark2-wayland and glmark2-es2-wayland under
On Tue, Mar 13, 2018 at 02:28:04PM +1000, Dave Airlie wrote:
> On 13 March 2018 at 14:24, Dave Airlie wrote:
> > This is the main code for the soft fp64 work. It's mostly Elie's
> > code with a bunch of changes by me.
> >
>
> All the patches are in my tree here, along with
Kenneth Graunke writes:
> On Friday, March 9, 2018 12:12:28 PM PDT Mark Janes wrote:
> [snip]
>> I've been doing this for Intel. Developers are on the hook to fix their
>> bugs, but you can't make them do it. They have many pressures on them,
>> and a maintainer can't
On Friday, March 9, 2018 12:12:28 PM PDT Mark Janes wrote:
[snip]
> I've been doing this for Intel. Developers are on the hook to fix their
> bugs, but you can't make them do it. They have many pressures on them,
> and a maintainer can't make the call as to whether a rendering bug is
> more
On Mon, 2018-03-12 at 07:37 -0700, Jason Ekstrand wrote:
> On Mon, Mar 12, 2018 at 1:40 AM, Iago Toral Quiroga m> wrote:
> > af5f2322d0c64 addressed this for extension commands, but the spec
> > mandates
> >
> > this behavior also for core API commands. From the Vulkan spec,
>
101 - 149 of 149 matches
Mail list logo