| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
v2:
- TG4 does not calculate derivatives (Ilia)
- also handle SAMPLE* instructions (Roland)
Cc: 12.0 <[email protected]>
Reviewed-by: Marek Olšák <[email protected]> (v1)
Reviewed-by: Brian Paul <[email protected]> (v1)
Reviewed-by: Ilia Mirkin <[email protected]>
Reviewed-by: Roland Scheidegger <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
Use rasterizer provoking vertex API.
Fix rasterizer provoking vertex for tristrips and quad list/strips.
v2: make provoking vertex tables static const
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
A texture may be redefined with _NEW_TEXTURE, which might have been
bound to a shader image slot. We have to revalidate the image atoms to
pick up on the new resource.
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Nicolai Hähnle <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Sometimes a register source can actually be double- or even quad-wide.
We must make sure that the inserted texbars take that width into
account.
Based on an earlier patch by Samuel Pitoiset.
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Samuel Pitoiset <[email protected]>
Cc: "12.0 11.2" <[email protected]>
|
|
|
|
|
|
|
| |
Reduces CPU load for draw calls that change none or few of the descriptors.
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
| |
This will simplify moving them to a per-context array.
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
| |
So that callers outside of si_descriptors.c need to worry less about the
details of descriptor handling.
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
| |
This mask is irrelevant for the generic descriptor set handling, and having it
outside simplifies subsequent changes slightly.
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
| |
Now that we emit guards for everything, we can just generate the files and
trust build flags to keep us safe. This should also fix the tarball
problems.
Reviewed-by: Emil Velikov <[email protected]>
|
|
|
|
| |
Reviewed-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To avoid blocking other EGL calls, release the display mutex before
we enqueue buffer to android frameworks and re-acquire the mutex
upon return.
v2: moved lock/unlock inside droid_window_enqueue_buffer().
TEST=verify pinch zoom in Photos app no longer causes hangs
Signed-off-by: Haixia Shi <[email protected]>
Reviewed-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the case of out-of-tree (OOT) builds, in particular when building
from tarball, we'll end up with the file in both srcdir and builddir.
We want the former to remain intact (since we need it on rebuild) while
the latter should be removed otherwise `make distclean' gets angry at
us.
Ideally there'll be a solution that feels a bit less of a hack. Until
then this does the job exactly as expected.
Cc: <[email protected]>
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
... when copied from git_sha1.h.
As the latter file can we lacking the write attribute, one should set it
explicitly. Otherwise we'll get a warning/failure at cleanup stage.
Cc: <[email protected]>
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
Otherwise the build will assume that we've talking about builddir, which
is not the case in the else statement.
Here the file is already generated and is part of the tarball.
Cc: <[email protected]>
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With earlier commit we introduced support for render_node devices, which
was couples with the use of the image loader extension.
As the work was inspired by egl/wayland we (erroneously) added the
extension for the !render_node path as well.
That works for wayland, as the implementations of the DRI2 and IMAGE
loader extensions converge behind the scenes. As that is not yet
the case for Android we shouldn't expose the extension.
Fixes: 34ddef39cef ("egl: android: add dma-buf fd support")
Cc: <[email protected]>
Reported-by: Mauro Rossi <[email protected]>
Tested-by: Mauro Rossi <[email protected]>
Acked-by: Rob Herring <[email protected]>
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
| |
v2: use a function for calculating WORD1 of bo metadata
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
|
| |
We don't import textures with DCC now, but soon we will.
v2: if we can't disable DCC for image writes, at least decompress DCC
at bind time
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
Could cause issues if you tried to read from an uninitialised pointer.
This just initalises the pointer to null to avoid that being a problem.
Discovered by Coverity.
CID: 1343616
Signed-off-by: Jakob Sinclair <[email protected]>
Reviewed-by: Iago Toral Quiroga <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
This fixes some cases of:
GL45-CTS.cull_distance.functional
on Skylake.
Reviewed-by: Chris Forbes <[email protected]>
Cc: "12.0" <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Samuel Pitoiset <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
| |
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Ian Romanick <[email protected]>
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We've had a FINISHME here since Eric originally wrote the code in 2011.
This patch implements his suggested approach, which makes us actually
able to copy propagate into the loops, at the unfortunate cost of making
this pass even more expensive.
The shader-db statistics are basically a wash:
No change in instruction counts.
total cycles in shared programs: 78685980 -> 78680730 (-0.01%)
cycles in affected programs: 2102646 -> 2097396 (-0.25%)
helped: 48
HURT: 83
I figured if we're going to do this for one copy propagation pass,
we may as well do it in both.
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Matt Turner <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We've had a FINISHME here since Eric originally wrote the code in 2010.
This patch implements his suggested approach, which makes us actually
able to copy propagate into the loops, at the unfortunate cost of making
this pass even more expensive.
The shader-db statistics are not terribly impressive:
total instructions in shared programs: 9008589 -> 9008613 (0.00%)
instructions in affected programs: 4293 -> 4317 (0.56%)
helped: 0
HURT: 10
total cycles in shared programs: 78550978 -> 78575760 (0.03%)
cycles in affected programs: 655426 -> 680208 (3.78%)
helped: 75
HURT: 88
GAINED: 2
Most of the "regressions" appear to be us successfully copy propagating
uniforms, which i965 is loading as pull constants instead of push, so we
occasionally have two pulls instead of one. That doesn't seem like this
pass's job - it's propagating correctly, and we should be smarter about
pull loads in the backend.
This patch is also useful for a couple of reasons:
1. It can clean up copies created by varying packing (previously, we
couldn't if the uses were inside a loop).
This fixes a bug when interpolateAt*() is used on a packed varying
inside a loop: glsl_to_nir struggles to see through the extra copy
and mistakenly believed the variable was not an input.
2. It will help propagate uniform array access created by
lower_const_array_to_uniforms().
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Matt Turner <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Like floats, we should use the round toward 0 mode instead of the
nearest one (which is the default) for doubles to integers.
This fixes all arb_gpu_shader_fp64 piglits which convert doubles to
integers (16 tests).
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
Cc: "11.2 12.0" <[email protected]>
|
|
|
|
|
|
| |
CMASK has no effect on metadata, because it's not sharable.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
| |
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94627
(against nouveau)
Acked-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
| |
Ported from Vulkan.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
| |
Ported from Vulkan.
v2: keep the comment
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
| |
This improves MSAA resolve performance.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
| |
This improves MSAA resolve performance.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The GL4.5 spec quote seems clear on this:
"The value -1 will be returned by either command if an error occurs,
if name does not identify an active variable on programInterface,
or if name identifies an active variable that does not have a valid
location assigned, as described above."
This fixes:
GL45-CTS.program_interface_query.output-built-in
[airlied: use _mesa_program_resource_location_index as
suggested by Eduardo]
Reviewed-by: Eduardo Lima Mitev <[email protected]>
Cc: "12.0" <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
| |
Found randomly while skimming the code. This might have caused VM faults in
robustness tests.
Cc: 12.0 <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
| |
Found by inspection after seeing
https://bugs.freedesktop.org/show_bug.cgi?id=96343
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Cast the unsigned semantic index to integer datatype before comparing
to max_generic, otherwise, max_generic which is initialized to -1
will be converted to unsigned int before the comparison, causing a wrong
semantic index to be assigned to a shader output.
Fixes the assert running TurboCAD_gl.trace. (VMware bug 1667265)
Also tested with glretrace, mesa demos pointblast, spriteblast and pointcoord.
v2: use the original max_generic variable but add the (int) cast
to the semantic index, as suggested by Brian.
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
| |
When TGSI debug flag is enabled, print the shader linkage info as well.
Tested with mesa demos with SVGA_DEBUG=tgsi
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
ARB_shader_image_load_store only requires a very fixed list of formats
to be supported, while textures may be in all kinds of formats, like
BGRA which are presently not supported on at least Kepler.
Signed-off-by: Ilia Mirkin <[email protected]>
Tested-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Switches to using truncf in micro_trunc.
Fixes the following piglit tests (for softpipe):
/spec/glsl-1.30/execution/built-in-functions/...
fs-trunc-float
fs-trunc-vec2
fs-trunc-vec3
fs-trunc-vec4
vs-trunc-float
vs-trunc-vec2
vs-trunc-vec3
vs-trunc-vec4
/spec/glsl-1.50/execution/built-in-functions/...
gs-trunc-float
gs-trunc-vec2
gs-trunc-vec3
gs-trunc-vec4
Signed-off-by: Lars Hamre <[email protected]>
Reviewed-by: Roland Scheidegger <[email protected]>
|
|
|
|
|
|
| |
Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]>
Reviewed-by: Timothy Arceri <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When we are not packing a double input varying, we might need to
read its data in a non-aligned to 64-bit offset, so we read
the wrong data. This is happening when using explicit locations
in varyings because Mesa disables packing varying for that case.
const_index is in 32-bit size units but offset() is multiplying
it by destination type size units. When operating with double
input varyings, const_index value could be not aligned to 64 bits.
To fix it, we load the double vector as if it was a float based vector
with twice the number of components.
Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]>
Reviewed-by: Timothy Arceri <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
Data starts at suboffet 3 in 32-bit units (12 bytes), so it is not
64-bit aligned and the current implementation fails to read the data
properly. Instead, when there is is a double input varying, read it as
vector of floats with twice the number of components.
Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]>
Reviewed-by: Timothy Arceri <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From GLSL 4.5 spec, "4.4.2.3 Geometry Outputs".
"all geometry shader output vertex count declarations in a
program must declare the same count."
Fixes:
GL45-CTS.geometry_shader.output.conflicted_output_vertices_max
Reviewed-by: Alejandro Piñeiro <[email protected]>
Cc: "11.2 12.0" <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|