On Sat, 2017-02-18 at 18:58 -0800, Francisco Jerez wrote:
> Samuel Iglesias Gonsálvez writes:
>
> > The lowered BSW/BXT indirect move instructions had incorrect
> > source types, which luckily wasn't causing incorrect assembly to be
> > generated due to the bug fixed in the next patch, but would
Hi;
On 02/17/2017 05:12 PM, Eric Engestrom wrote:
From: Brendan King
This is a DRI3 version of a change made for DRI2
(4d6d4f939e0af4252e0b, "egl/dri2: implement query surface hook"),
that fixed failures in dEQP-EGL.functional.resize.surface_size.grow
and dEQP-EGL.functional.resize.surface_siz
On 20/02/17 15:38, Edward O'Callaghan wrote:
On 02/20/2017 11:15 AM, Timothy Arceri wrote:
---
src/gallium/drivers/r600/r600_pipe.c | 19 +++
src/gallium/drivers/radeon/r600_pipe_common.c | 8 +++-
src/gallium/drivers/radeon/r600_pipe_common.h | 5 +
src/gal
On Fri, Feb 17, 2017 at 04:03:48PM -0800, Jason Ekstrand wrote:
> The isl_surf_init call that each of these helpers make can, in theory,
> fail. We should propagate that up to the caller rather than just
> silently ignoring it.
Reviewed-by: Topi Pohjolainen
> ---
> src/intel/isl/isl.c
Reviewed-by: Edward O'Callaghan
On 02/20/2017 04:17 PM, Dave Airlie wrote:
> From: Dave Airlie
>
> This splits out the cache flush bit setting code
> dependent on the src/dest access flags.
>
> It then calls it from the subpass barrier code.
>
> It also marks a TODO to remove the aggressive C
From: Dave Airlie
This splits out the cache flush bit setting code
dependent on the src/dest access flags.
It then calls it from the subpass barrier code.
It also marks a TODO to remove the aggressive CS/PS
flushes at some point.
This fixes a bunch of the
dEQP-VK.renderpass.attachment_allocati
On 02/20/2017 11:15 AM, Timothy Arceri wrote:
> ---
> src/gallium/drivers/r600/r600_pipe.c | 19 +++
> src/gallium/drivers/radeon/r600_pipe_common.c | 8 +++-
> src/gallium/drivers/radeon/r600_pipe_common.h | 5 +
> src/gallium/drivers/radeonsi/si_pipe.c
On Feb 19, 2017 2:33 PM, "Dave Airlie" wrote:
From: Dave Airlie
This adds support to radv_GetPhysicalDeviceXlibPresentationSupportKHR
and radv_GetPhysicalDeviceXcbPresentationSupportKHR to check if the
local device file descriptor is compatible with the descriptor
retrieved from the X server vi
tic bool
> +wsi_x11_check_dri3_compatible(xcb_connection_t *conn, int local_fd)
> +{
> + xcb_screen_iterator_t screen_iter =
> + xcb_setup_roots_iterator(xcb_get_setup(conn));
> + xcb_screen_t *screen = screen_iter.data;
> +
> + int dri3_fd = wsi_dri3_open(conn, screen->root, None);
> +
On Feb 19, 2017 2:33 PM, "Dave Airlie" wrote:
From: Dave Airlie
This adds support to radv_GetPhysicalDeviceXlibPresentationSupportKHR
and radv_GetPhysicalDeviceXcbPresentationSupportKHR to check if the
local device file descriptor is compatible with the descriptor
retrieved from the X server vi
On Sun, Feb 19, 2017 at 6:11 AM, Max Qian wrote:
> This fixes some situations where glBlitFrameBuffer would error in Mesa,
glBlitFramebuffer
> but wouldn't error in other drivers (e.g Nvidia)
Please be descriptive here. What situations does this fix?
Also, there's not a whole lot of caring in
V2: The functions we are now sharing used different return values
update the callers to reflect the new common api.
---
src/amd/vulkan/radv_device.c | 22 +++---
src/util/disk_cache.h| 17 +
2 files changed, 20 insertions(+), 19 deletions(-)
diff --git a/sr
---
src/gallium/drivers/r600/r600_pipe.c | 19 +++
src/gallium/drivers/radeon/r600_pipe_common.c | 8 +++-
src/gallium/drivers/radeon/r600_pipe_common.h | 5 +
src/gallium/drivers/radeonsi/si_pipe.c| 19 +++
src/mesa/state_tracker/st_conte
Implements a tgsi cache for the OpenGL state tracker.
V2: add support for compute shaders
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 222 +
src/mesa/state_tracker/st_program.c| 133 -
2 files changed, 350 insertions(+), 5 deletions(-)
dif
---
src/gallium/include/pipe/p_screen.h | 6 ++
1 file changed, 6 insertions(+)
diff --git a/src/gallium/include/pipe/p_screen.h
b/src/gallium/include/pipe/p_screen.h
index b6203f1..43dcdf0 100644
--- a/src/gallium/include/pipe/p_screen.h
+++ b/src/gallium/include/pipe/p_screen.h
@@ -58,6 +5
Changes in V3:
- split out change to pipe_screen into its own patch, and rework
enablement patch.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
On Sunday, 2017-02-19 18:54:51 -0500, Ilia Mirkin wrote:
> On Sun, Feb 19, 2017 at 6:51 PM, Eric Engestrom wrote:
> > On Sunday, 2017-02-19 18:33:16 -0500, Ilia Mirkin wrote:
> >> Why are patches 1-3 necessary?
> >
> > They allow patch #4, which makes use of a macro to simplify the list,
> > (requ
On Sun, Feb 19, 2017 at 6:51 PM, Eric Engestrom wrote:
> On Sunday, 2017-02-19 18:33:16 -0500, Ilia Mirkin wrote:
>> Why are patches 1-3 necessary?
>
> They allow patch #4, which makes use of a macro to simplify the list,
> (requires a 1:1 mapping of entrypoint to function) which also allows for
>
On Sunday, 2017-02-19 15:38:05 -0800, Matt Turner wrote:
> On Sun, Feb 19, 2017 at 3:23 PM, Eric Engestrom wrote:
> > Tested with dEQP-EGL.functional.get_proc_address.*
> >
> > Signed-off-by: Eric Engestrom
> > ---
> > src/egl/main/eglapi.c | 37 -
> > 1 file
On Sunday, 2017-02-19 18:33:16 -0500, Ilia Mirkin wrote:
> Why are patches 1-3 necessary?
They allow patch #4, which makes use of a macro to simplify the list,
(requires a 1:1 mapping of entrypoint to function) which also allows for
the simple sorting test script in patch #7.
I could do without, b
On Sun, Feb 19, 2017 at 3:23 PM, Eric Engestrom wrote:
> Tested with dEQP-EGL.functional.get_proc_address.*
>
> Signed-off-by: Eric Engestrom
> ---
> src/egl/main/eglapi.c | 37 -
> 1 file changed, 24 insertions(+), 13 deletions(-)
>
> diff --git a/src/egl/mai
Why are patches 1-3 necessary?
On Sun, Feb 19, 2017 at 6:23 PM, Eric Engestrom wrote:
> Signed-off-by: Eric Engestrom
> ---
> src/egl/main/eglapi.c | 31 ---
> 1 file changed, 24 insertions(+), 7 deletions(-)
>
> diff --git a/src/egl/main/eglapi.c b/src/egl/main/egla
As of the last 3 commits, there's a function for each entrypoint.
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 149 +-
1 file changed, 75 insertions(+), 74 deletions(-)
diff --git a/src/egl/main/eglapi.c b/src/egl/main/eglapi.c
index
Let's make that comment true.
If will also be necessary in a couple commits (using bsearch).
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 74 +--
1 file changed, 37 insertions(+), 37 deletions(-)
diff --git a/src/egl/main/eglapi.c b/s
Tested with dEQP-EGL.functional.get_proc_address.*
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 37 -
1 file changed, 24 insertions(+), 13 deletions(-)
diff --git a/src/egl/main/eglapi.c b/src/egl/main/eglapi.c
index 1038916cd2..b6a5b1c978 100644
Starting with the next commit, badly sorting this list will break the
eglGetProcAddress().
Signed-off-by: Eric Engestrom
---
src/egl/Makefile.am | 3 ++-
src/egl/egl-entrypoint-check | 4
2 files changed, 6 insertions(+), 1 deletion(-)
create mode 100755 src/egl/egl-entrypoint-che
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 31 ---
1 file changed, 24 insertions(+), 7 deletions(-)
diff --git a/src/egl/main/eglapi.c b/src/egl/main/eglapi.c
index e149c0f8d1..e44375a106 100644
--- a/src/egl/main/eglapi.c
+++ b/src/egl/main/eglapi.c
@@
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 28 +---
1 file changed, 21 insertions(+), 7 deletions(-)
diff --git a/src/egl/main/eglapi.c b/src/egl/main/eglapi.c
index cab05c2301..251855cc3b 100644
--- a/src/egl/main/eglapi.c
+++ b/src/egl/main/eglapi.c
@@ -1
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 28 +---
1 file changed, 21 insertions(+), 7 deletions(-)
diff --git a/src/egl/main/eglapi.c b/src/egl/main/eglapi.c
index 251855cc3b..e149c0f8d1 100644
--- a/src/egl/main/eglapi.c
+++ b/src/egl/main/eglapi.c
@@ -1
This will allow us to make sure the list is always sorted in the next
commit.
Signed-off-by: Eric Engestrom
---
src/egl/main/eglapi.c | 78 +-
src/egl/main/eglentrypoint.def | 77 +
2 files changed, 78 inser
This assert might have made sense before but we no longer use
gl_linked_shader here. Unless the caller has really done something
crazy this assert is fairly useless.
We also do some small tidy ups in this change.
---
src/mesa/main/uniforms.c | 9 ++---
1 file changed, 2 insertions(+), 7 delet
Reviewed-by: Bas Nieuwenhuizen
On Sun, Feb 19, 2017 at 11:48 PM, Dave Airlie wrote:
> From: Dave Airlie
>
> This just keeps popping up minor problems and regressions we should
> revisit in a more sustainable manner later.
>
> This also reverts:
> Revert "radv: query cmds should mark a cmd buffe
From: Dave Airlie
This just keeps popping up minor problems and regressions we should
revisit in a more sustainable manner later.
This also reverts:
Revert "radv: query cmds should mark a cmd buffer as having draws."
Revert "radv: also fixup event emission to not get culled."
This reverts commi
From: Dave Airlie
This adds support to radv_GetPhysicalDeviceXlibPresentationSupportKHR
and radv_GetPhysicalDeviceXcbPresentationSupportKHR to check if the
local device file descriptor is compatible with the descriptor
retrieved from the X server via DRI3.
This will stop radv binding to an X ser
From: Dave Airlie
This adds support to radv_GetPhysicalDeviceXlibPresentationSupportKHR
and radv_GetPhysicalDeviceXcbPresentationSupportKHR to check if the
local device file descriptor is compatible with the descriptor
retrieved from the X server via DRI3.
This will stop radv binding to an X ser
On 20 February 2017 at 07:06, Bas Nieuwenhuizen
wrote:
> On Sun, Feb 19, 2017 at 9:56 PM, Dave Airlie wrote:
>> From: Dave Airlie
>>
>> physical devices can be enumerated multiple time per instance,
>> this was causing the realloc to be hit for no reason, and
>> the CTS allocation verification
> @@ -270,6 +318,25 @@ VkBool32
> wsi_get_physical_device_xcb_presentation_support(
>return false;
> }
>
> + xcb_screen_iterator_t screen_iter =
> + xcb_setup_roots_iterator(xcb_get_setup(connection));
> + xcb_screen_t *screen = screen_iter.data;
> +
> + char *local_dev = drm
On Feb 18, 2017 9:48 PM, "Dave Airlie" wrote:
From: Dave Airlie
This adds support to radv_GetPhysicalDeviceXlibPresentationSupportKHR
and radv_GetPhysicalDeviceXcbPresentationSupportKHR to check if the
local device file descriptor is compatible with the descriptor
retrieved from the X server vi
On Sun, Feb 19, 2017 at 9:56 PM, Dave Airlie wrote:
> From: Dave Airlie
>
> physical devices can be enumerated multiple time per instance,
> this was causing the realloc to be hit for no reason, and
> the CTS allocation verification would get angry.
We only do the radv_physical_device_init on th
From: Dave Airlie
physical devices can be enumerated multiple time per instance,
this was causing the realloc to be hit for no reason, and
the CTS allocation verification would get angry.
We'd also be allocating common extensions onto the list
each time for no reason.
This hopefully fixes a bun
From: Nicolai Hähnle
Allocating huge buffers in VRAM is not a problem, but when those buffers
start being migrated, the kernel runs into errors because it cannot split
those buffer up for moving through GTT.
This should fix intermittent failures of
GL45-CTS.texture_buffer.texture_buffer_max_size
https://bugs.freedesktop.org/show_bug.cgi?id=99863
Bug ID: 99863
Summary: [llvmpipe] gtkopengl rendering glitches
Product: Mesa
Version: unspecified
Hardware: Other
OS: All
Status: NEW
Severity: normal
From: Marek Olšák
This removes a lot of useless LDS stores.
A few games read TESSINNER/OUTER, but not any other outputs. Most games
don't read any outputs.
The only app doing LDS output reads is UE4 Lightsroom Interior.
---
src/gallium/drivers/radeonsi/si_shader.c | 17 -
1 fil
From: Marek Olšák
v2: define lp_float_mode
---
src/gallium/auxiliary/gallivm/lp_bld_misc.cpp | 15 ---
src/gallium/auxiliary/gallivm/lp_bld_misc.h | 8 +++-
src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c | 6 +-
3 files changed, 24 insertions(+), 5 deleti
From: Marek Olšák
We were unconditionally storing these outputs, sometimes even one component
at a time, but apps never read them in TES.
Move the TESSINNER/OUTER buffer stores into the TCS epilog where we can
easily disable them on demand.
---
src/gallium/drivers/radeonsi/si_shader.c|
From: Marek Olšák
not all of them will be used immediately
---
src/gallium/auxiliary/tgsi/tgsi_scan.c | 26 ++
src/gallium/auxiliary/tgsi/tgsi_scan.h | 8
2 files changed, 34 insertions(+)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_scan.c
b/src/gallium/auxil
From: Marek Olšák
R600_DEBUG=mono has had no effect since:
commit 1fabb297177069e95ec1bb7053acb32f8ec3e092
Author: Marek Olšák
Date: Tue Feb 14 22:08:32 2017 +0100
radeonsi: have separate LS and ES main shader parts in the shader selector
Also, this assertion was failing:
From: Marek Olšák
Recommended by Matt Arsenault.
46757 shaders in 28742 tests
Totals:
SGPRS: 2068851 -> 2066907 (-0.09 %)
VGPRS: 1604056 -> 1602676 (-0.09 %)
Spilled SGPRs: 1402 -> 1382 (-1.43 %)
Spilled VGPRs: 113 -> 113 (0.00 %)
Private memory VGPRs: 1332 -> 1332 (0.00 %)
Scratch size: 3224 ->
Patches 1-3 are:
Reviewed-by: Eduardo Lima Mitev
I would let somebody else look at patch 4 because I'm not very familiar
with entrypoints generation scripts.
cheers,
Eduardo
On 02/19/2017 12:49 PM, Emil Velikov wrote:
> Fixes: 4c9dec80ede ("anv: Get rid of the ANV_CALL macro")
> Signed-off-by
Signed-off-by: Bas Nieuwenhuizen
---
src/amd/vulkan/radv_meta_fast_clear.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/amd/vulkan/radv_meta_fast_clear.c
b/src/amd/vulkan/radv_meta_fast_clear.c
index 0902c455037..ffaa9a3d5e8 100644
--- a/src/amd/vulkan/radv_meta_fa
Used only within the generated source file.
Fixes: 12301c54186 ("radv: drop the RADV_CALL macro.")
Signed-off-by: Emil Velikov
---
src/amd/vulkan/radv_entrypoints_gen.py | 2 +-
src/amd/vulkan/radv_private.h | 1 -
2 files changed, 1 insertion(+), 2 deletions(-)
diff --git a/src/amd/vu
Fixes: 12301c54186 ("radv: drop the RADV_CALL macro.")
Signed-off-by: Emil Velikov
---
src/amd/vulkan/radv_device.c | 1 -
src/amd/vulkan/radv_private.h | 2 --
2 files changed, 3 deletions(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index 6f2fac2a9f..7039052a2d
Fixes: 4c9dec80ede ("anv: Get rid of the ANV_CALL macro")
Signed-off-by: Emil Velikov
---
src/intel/vulkan/anv_device.c | 2 --
src/intel/vulkan/anv_private.h | 2 --
2 files changed, 4 deletions(-)
diff --git a/src/intel/vulkan/anv_device.c b/src/intel/vulkan/anv_device.c
index cae5feff94..91a
This changes the way radv_entrypoints_gen.py works from generating a
table containing every single entrypoint in the XML to just the ones
that we actually need. There's no reason for us to burn entrypoint
table space on a bunch of NV extensions we never plan to implement.
RADV implements VK_AMD_d
This fixes some situations where glBlitFrameBuffer would error in Mesa,
but wouldn't error in other drivers (e.g Nvidia)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97921
v2 : fix unintentional line removal
v3 : author correction, add commit message, and change the error message
---
sr
Yes, that's correct. Mesa was throwing errors for glBlitFramebuffer in
the same situations other (e.g nvidia) drivers weren't, causing
artifacting for some applications (e.g wine). The bugzilla link has a
test apitrace for it.
I'll work on the commit and post the revision when I'm done.
__
Jason Ekstrand wrote on 19.02.2017 06:01:
> On Feb 18, 2017 12:37 PM, "Kai Wasserbäch"
> wrote:
>
> Hey Jacob,
> sorry for not spotting this the first time, but I have an additional
> comment.
> Please see below.
>
> Jacob Lifshay wrote on 18.02.2017 18:48:> This commit improves the message
> by
Hi Max,
can you explain a bit more about the error you're seeing? It looks like
the actual functional change is that previously an INVALID_OPERATION
error was generated when /mask/ includes DEPTH_BUFFER_BIT but either
source or destination had no depth buffer, and with your change no error
is
Series is
Reviewed-by: Bas Nieuwenhuizen
On Sun, Feb 19, 2017 at 7:33 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> This iterates the fast clear flush across the layers in the
> specified range.
>
> It also moves the compute resolve flush into the function
> and builds the range in there.
>
>
Reviewed-by: Bas Nieuwenhuizen
On Sun, Feb 19, 2017 at 7:38 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> This fixes:
> dEQP-VK.renderpass.formats.a2b10g10r10_unorm_pack32*
> regressions.
>
> Signed-off-by: Dave Airlie
> ---
> src/amd/vulkan/radv_formats.c | 2 +-
> 1 file changed, 1 inserti
On 17.02.2017 11:27, Marek Olšák wrote:
From: Marek Olšák
all drivers support it
Patches 4 & 5:
Reviewed-by: Nicolai Hähnle
1-3 look good to me as well, but I'd wait for feedback from the
corresponding driver authors.
---
src/gallium/docs/source/screen.rst | 4
src/
On 17.02.2017 14:26, Marek Olšák wrote:
From: Marek Olšák
sffbh needs to be suffixed with ".i32"
Reviewed-by: Nicolai Hähnle
---
src/amd/common/ac_llvm_build.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/amd/common/ac_llvm_build.c b/src/amd/common/ac_llvm_
On 17.02.2017 13:10, Marek Olšák wrote:
From: Marek Olšák
It's OK for r300g (because r300g can't write to buffers via the GPU), but
not later hardware. This issue was spotted randomly.
Cc: mesa-sta...@lists.freedesktop.org
Both patches:
Reviewed-by: Nicolai Hähnle
---
src/gallium/auxili
63 matches
Mail list logo
