| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
Since applications are allowed to specify some set of bindings which need
not be dense they also need not be in order. For most things, this doesn't
matter, but it could result getting the wrong dynamic offsets. This adds a
quick-and-dirty sort to ensure that everything is always in increasing
order of binding index.
Signed-off-by: Jason Ekstrand <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
| |
This allows for some extra validation and makes it easier to see what's
going on when poking around in gdb.
Signed-off-by: Jason Ekstrand <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Jason Ekstrand <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
With glx of gstreamer-vaapi, the temporary pixmap for front buffer gets
renewed in each frame, so when we receive a new pixmap, should get a new
front buffer for it.
This also fixes Totem player playback corruption.
Signed-off-by: Leo Liu <[email protected]>
Reviewed-by: Michel Dänzer <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From original commit, the macro "if HAVE_DRI3" was in Makefile.sources,
this file is shared with SCons, SCons is not able to parse this marco,
the SCons build failed. Jose quickly gave two approaches and quick fix
with his second approach, thanks Jose for the solutions and fixes.
This patch is Jose's first approach, and it's more proper, because the
dri3 c file should not be included to build when DRI3 is not enabled.
Signed-off-by: Leo Liu <[email protected]>
Acked-by: Emil Velikov <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
| |
Besides the old JIT bug, it seems the X86 backend on LLVM 3.3 doesn't
handle llvm.fmuladd and instead it fall backs to a C function. Which in
turn causes a segfault on Windows.
Reviewed-by: Roland Scheidegger <[email protected]>
|
|
|
|
| |
Reviewed-by: Roland Scheidegger <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
As suggested by Roland Scheidegger.
Use the same logic as f16c, since fma requires VEX encoding.
But disable FMA on LLVM 3.3 without MCJIT.
Reviewed-by: Roland Scheidegger <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes a problem with the CE preamble and restoring only stuff in the
preamble when needed.
To illustrate suppose we have two graphics IB's 1 and 2, which are submitted in
that order. Furthermore suppose IB 1 does not use CE ram, but IB 2 does, and we
have a context switch at the start of IB 1, but not between IB 1 and IB 2.
The old code put the CE RAM loads in the preamble of IB 2. As the preamble of
IB 1 does not have the loads and the preamble of IB 2 does not get executed, the
old values are not load into CE RAM.
Fix this by always restoring the entire CE RAM.
v2: - Just load all descriptor set buffers instead of load and store the entire
CE RAM.
- Leave the ce_ram_dirty tracking in place for the non-preamble case.
v3: - Fixed parameter alignment.
- Rebased to master (Nicolai's descriptor series).
Signed-off-by: Bas Nieuwenhuizen <[email protected]>
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
| |
This should fix the MSVC linker failures that arose with commit
5e2d25894b962aae9158261897e13843377e3b95.
Trivial.
|
|
|
|
|
|
|
|
|
|
|
| |
The reality is that this doesn't matter, because we manually emit the
ARL to the sampler reladdr, and those arguments don't get an extra load
later, so it's effectively just a boolean. However having the types be
wrong is confusing and could trigger very odd bugs should usage change
down the line.
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
| |
Adding 64-bit integers support was going to make this file worse,
just remove the tabs from it now.
Acked-by: Timothy Arceri <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
| |
An update in graphics specs has deleted the halign and valign fields
from XY_FAST_COPY_BLT command. See mesa commit 97f0f91.
Cc: Ben Widawsky <[email protected]>
Signed-off-by: Anuj Phogat <[email protected]>
Reviewed-by: Ben Widawsky <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Anuj Phogat <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
The extension is already advertised in compatibility profile, but
the _mesa_has_compute_shaders only returns true in core profile.
If we advertise it, we should allow it to work.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ian Romanick <[email protected]>
Reviewed-by: Nanley Chery <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
v2: only load the clip vertex once
v3: fix clip enable logic, add cullDistance
v4: remove duplicate fields in vs jit key, fix test of clip fixup needed
v5: fix clipdistance linkage for slot!=0,4
v6: support clip+cull; passes most piglit clip (failures understood)
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GLX documentation states:
glXCreateNewContext can generate the following errors: (...)
GLXBadFBConfig if config is not a valid GLXFBConfig
Function checks if the given config is a valid config and sets proper
error code.
Fixes currently crashing glx-fbconfig-bad Piglit test.
v2: coding style cleanups (Emil, Topi)
use DefaultScreen macro (Emil)
Signed-off-by: Matt Roper <[email protected]>
Signed-off-by: Tapani Pälli <[email protected]>
Reviewed-by: Emil Velikov <[email protected]>
Cc: "11.2" <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Nayan Deshmukh <[email protected]>
Reviewed-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Apply the luma key filter to the YCbCr values during the CSC conversion
in video buffer shader. The initial values of max and min luma are set
to opposite values to disable the filter initially and will be set when
enabling it.
Add extra parmeters min and max luma for the luma key filter in
vl_compositor_set_csc_matrix in va, xvmc. Setting them
to opposite value 1.f and 0.f respectively won't effect the CSC
conversion
v2: -Squash 1,2 and 3 into one patch to avoid breaking build of
other components. (Christian)
-use ureg_swizzle. (Christian)
-change name of the variables. (Christian)
v3: -Squash all patches in one to avoid breaking of build. (Emil)
-wrap functions properly. (Emil)
-use 0.0f and 1.0f instead of 0.f and 1.f respectively. (Emil)
v4: -Divide it in two patches one which introduces the functionality
and assigs dummy values to the changed functions and second which
implements the lumakey filter. (Christian)
-use ureg_scalar instead ureg_swizzle. (Christian)
Signed-off-by: Nayan Deshmukh <[email protected]>
Reviewed-by: Christian König <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
When Kristian implemented GL_TEXTURE_EXTERNAL_OES, he hooked it up for gen8
but not for gen7 or earlier. It all works, we just need to emit the states
for the extra planes.
Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Kristian Høgsberg <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When calling virgl_fence_wait() with timeout=0,
virgl_{drm,vtest}_resource_is_busy() is called. However, it returns TRUE
for a busy resource, whereace virgl_fence_wait() should return TRUE for
a completed (non-busy) resource.
This fixes running supertuxkart in a VM (I could not reproduce locally
with vtest though there is a similar fix)
Signed-off-by: Marc-André Lureau <[email protected]>
Cc: "11.1 11.2 12.0" <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
| |
In the future int64 support will have the same requirements.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
| |
This uses the new types interfaces to check for 64-bit types,
as futureproofing against int64 support.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
| |
This is just some better type safety that I noticed while working
on 64-bit integer support.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
| |
This just moves to the new interfaces in advance of int64.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
| |
This is prep work for int64 support.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
| |
This is just prep work for int64 support, changing
places where 64-bit matters no doubles.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
| |
This just moves code to the new check in advance of int64 support.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
This adds an inline and type query for if a type is 64-bit.
Fow now this is equivalent to double, but int64 will change
this.
Reviewed-by: Ilia Mirkin <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Coverity complains that the computed sizes can lead to negative lengths
passed to memcpy. If that happens we've been handed invalid arguments
anyway, so just bomb out.
The funky "0%s" is because the size string for the variable-length part
of the request is of the form "+ safe_pad() ...", and a unary + would
coerce the result to always be positive, defeating the overflow check.
Signed-off-by: Adam Jackson <[email protected]>
Reviewed-by: Matt Turner <[email protected]>
|
|
|
|
|
|
| |
2% isn't much. If you think the number should be decreased, please speak up.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
| |
This was for some old unsupported LLVM version.
Only si_create_context creates the target machine now.
r600g doesn't use this function.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Diff from shader-db:
Scratch: 3221504 -> 17408 (-99.46 %) bytes per wave
v2: add "break;"
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
| |
This reverts commit ffd54d1936fcd07424265b780e1d049222a01e94.
No, it doesn't work. The test case is "glxgears -samples 2".
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The width0/height0/depth0 on stObj may not have been set at this point.
Observed in a trace that set up levels 2..9 of a 2d texture, and set the base
level to 2, with height 1. This made the guess logic always bail.
Originally investigated by Ilia Mirkin, this patch gets rid of the somewhat
redundant storage of width0/height0/depth0 and makes sure we always compute
pipe texture sizes that are compatible with the base level image of the
GL texture.
Fixes the gl-1.2-texture-base-level piglit test provided by Brian Paul.
v2:
- try to re-use an existing pipe texture when possible
- handle a corner case where the base level is not level 0 and it is of
size 1x1x1
v3:
- ptHeight = ptWidth in cube map 1x1 case (suggested by Brian)
Cc: "12.0" <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
|
| |
This just stops counting and assigning a storage location for
these uniforms, the count is only used to create the uniform storage.
These uniform types don't use this storage.
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
We were previously unconditionally doing this for arrays and ubo's, and
ignoring texture/storage/atomic buffers. Instead use the usage history
to determine which atoms need to be revalidated.
Signed-off-by: Ilia Mirkin <[email protected]>
Reviewed-by: Nicolai Hähnle <[email protected]>
Cc: "12.0" <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With this change, to enable precise SIN and COS instructions
on Intel hardware, one can put
<option name="precise_trig" value="true"/>
in the proper drirc file.
V2: Make option name more generic
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Stephane Marchesin <[email protected]>
|
|
|
|
|
|
| |
disabled by 4f1cccf570112f93265a4cace504eb763fa8f73e
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
| |
Since DCC is enabled almost everywhere now, it's important not to disable
this fast path.
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
| |
No idea why this was disabled, but it works fine.
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
| |
Allocating it has no effect, but it adds overhead (useless DCC clear).
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
| |
If first_level > 0 and DCC is disabled for that level, let's skip DCC
reads entirely.
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
| |
Also add dcc_fast_clear_size for clearing only the necessary subset
of DCC. For no AA, it's equal to the size of the whole DCC level.
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
| |
We want to keep DCC enabled to save bandwidth. It was a bad idea to disable
it here.
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|