| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
See my LLVM patch which fixes the root cause.
Users have to apply this patch and then they have 2 choices:
- Downgrade to LLVM 5.0
- Update to LLVM git after my LLVM patch is pushed.
It won't be possible to use current and earlier development version
of LLVM 6.0.
Reviewed-by: Nicolai Hähnle <[email protected]>
Cc: 17.3 <[email protected]>
|
|
|
|
|
|
|
| |
Obtained from: FreeBSD ports
Signed-off-by: Emil Velikov <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
| |
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
| |
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
| |
This is different from pipe->invalidate_resource()..
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
This should be OUT_RELOC() since the operation isn't writing to the
buffer. Technically it doesn't matter much currently, since we'd
anyways to a gmem2mem later. But that will change.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
| |
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
For CPU access.
Signed-off-by: Leo Liu <[email protected]>
Cc: [email protected]
Reviewed-by: Christian König <[email protected]>
|
|
|
|
|
|
|
| |
This was the same between si and ac.
Reviewed-by: Timothy Arceri <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
| |
This was duplicated between both drivers, share here.
Reviewed-by: Timothy Arceri <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
v2: use a more generic compat function
v3: rename and formatting cleanup
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103388
Signed-off-by: Jan Vesely <[email protected]>
Reviewed-by: Francisco Jerez <[email protected]>
CC: <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These assertions were revisited a couple of times in the past, and they
still weren't quite right.
The problem I was seeing (with some other state tracker) was a copy between
two 512x512 s3tc textures, but from mip level 0 to mip level 8. Therefore,
the destination has only size 2x2 (not a full block), so the box width/height
was only 2, causing the assertion to trigger for src alignment.
As far as I can tell, such a copy is completely legal, and because a correct
assertion would get ridiculously complicated just get rid of it for good.
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
| |
Mostly copy/pasta from Dylan Baker's conversion of nouveau and i965.
Signed-off-by: Rob Clark <[email protected]>
Reviewed-by: Eric Engestrom <[email protected]>
Reviewed-by: Dylan Baker <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
Similar to 848da662224326ccfbe6647bc82f4f89ca22c762, pass an arg to
ir3_nir_trig.py to add to python path, rather than using $PYTHONPATH,
to prep for meson build support.
Signed-off-by: Rob Clark <[email protected]>
Reviewed-by: Eric Engestrom <[email protected]>
Reviewed-by: Dylan Baker <[email protected]>
|
|
|
|
|
|
|
|
| |
To enable per-context priorities, we need to have per-context pipe's.
Unfortunately we still need to keep the global screen pipe, mostly just
for screen->get_timestamp().
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
To add context priority support we need to have an fd_pipe per context,
rather than per-screen. Which conflicts with existing ctx->pipe (which
is actually a visibility stream pipe (hw resource). So just rename it.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
| |
Prep work for later patch.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Instead of plain snprintf(). To fix the MSVC build.
snprintf() is used in various places in Mesa/gallium, but apparently,
not in code built with MSVC.
Reviewed-by: Eric Engestrom <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This restores performance for the drirc workaround, i.e.
KILL_IF does:
visible = src0 >= 0;
kill_flag &= visible; // accumulate kills
amdgcn_kill(wqm_vote(visible)); // kill fully dead quads only
And all helper pixels are killed at the end of the shader:
amdgcn_kill(kill_flag);
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
| |
This will be a new LLVM intrinsic and will also work nicely with
llvm.amdgcn.wqm.vote.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
Compiling with MSVC options /we4995 /we4996 (a subset of /sdl) generates
a warning that the gethostbyname() function is deprecated in favor of
getaddrinfo() or GetAddrInfoW(). Replace the call with getaddrinfo().
Untested. There are no callers to u_socket_connect() in Gallium.
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Dylan Baker <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Dylan Baker <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Vulkan CTS does not expect the value to be clamped (at least for D32),
and it makes a differences even though depth is in [0,1], due
to strict inequalities.
I couldn't find anything in the Vulkan spec about this, but the test
seemed to be copied from GL tests and the GL spec only specifies
clamping for fixed point formats. Hence I expect radeonsi to run into
this at some point as well, but given that they still have a usecase
with the Z16->Z32 promotion, I'll leave that for someone else to clean
up.
This at least fixes radv dEQP-VK.texture.shadow.* on VI.
Fixes: 0f9e32519bb 'ac/nir: clamp shadow texture comparison value on VI'
Reviewed-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
| |
This was causing Android clang version 3.8.256229 to miscompile,
presumably due to strict aliasing.
Fixes: 14dc281c1332 ("vc4: Enforce one-uniform-per-instruction after optimization.")
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is part of a cooperative scheduling approach used by radv. All
drivers in the stack must opt-in to resource arbitration, otherwise GL
based apps will be able to ignore system priorities.
We always hardcode the field to its maximum value, instead of attempting
to calculate an approximate usage. In testing, there were no benefits to
using anything other than the maximum.
Signed-off-by: Andres Rodriguez <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
| |
We weren't ever filling in the texture state record, so we'd dereference
NULL from the shader.
|
|
|
|
|
|
|
|
|
| |
I was generating some stub values to bring the driver up, but fill them in
properly now. We now set 1.0 or 1u as appropriate, and thanks to being in
their own BO it fixes piglit failures on the 7268 (where our 4-byte
alignment was insufficient).
Fixes const-packHalf2x16.shader_test
|
|
|
|
|
| |
This should reduce our draw-time overhead, and puts the code where it
should go long term.
|
|
|
|
| |
Fixes: 59fb59ad54d3 ("nir: Get rid of nir_shader::stage")
|
|
|
|
|
|
|
|
| |
It's redundant with nir_shader::info::stage.
Acked-by: Timothy Arceri <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
Reviewed-by: Jordan Justen <[email protected]>
|
|
|
|
|
|
|
| |
Galliums query_type used in APIs is unsigned.
Signed-off-by: Christian Gmeiner <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Christian Gmeiner <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Passes most occlusion query piglits. The following piglits are broken:
- spec@arb_occlusion_query@occlusion_query_meta_fragments
- spec@arb_occlusion_query@occlusion_query_meta_save
- spec@arb_occlusion_query2@render
v1 -> v2:
- use one sample provider for all occlusion queries tyes
- add comment about 'magic' value 0x1DF5E76
Signed-off-by: Christian Gmeiner <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
No hardware query is supported yet.
v1 -> v2
- removed query_type from strcut etna_hw_sample_provider
Signed-off-by: Christian Gmeiner <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
|
|
| |
Update to etna_viv commit 6c9c706.
Signed-off-by: Christian Gmeiner <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Jump through the layers of abstraction between egl and dri in order to
feed the context priority attribute through to the backend. This
requires us to read the value from the base _egl_context, convert it to
a DRI attribute, parse it again in the generic context creator before
passing it to the driver as a function parameter.
In order to not require us to pass back the actual value of the context
priority after creation, we impose that drivers should report the
available set of priorities during screen setup (and then they may chose
to fail if given an invalid value as that should have been checked at
the user boundary.)
Signed-off-by: Chris Wilson <[email protected]>
Acked-by: Ben Widawsky <[email protected]> # i915/i965
Reviewed-by: Emil Velikov <[email protected]>
Reviewed-by: Eric Engestrom <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove allocation of > 2kbyte buffers into context memory in
swr_copy_to_scatch_space() (which is used to copy small vertex/index buffers
and shader constants to a scratch space to be used by the upcoming draw.)
Large shader constant allocations need to be done in the circular scratch
buffer instead of context memory, because their values persist across
render calls.
Also lower SCRATCH_SINGLE_ALLOCATION_LIMIT to 8k, since allocations of larger
buffers will get too large for the circular scratch space.
Fixes render issues with CEI Ensight.
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Otherwise -Dgallium-drivers= will cause libmesa_gallium to be built and
the megadriver install script to attempt to install drivers without any
actual drivers being built.
fixes: 66f97f6640f5316b36177fd1053f0027eb6ec6cc ("meson: build radeonsi")
Reported-by: Rafael Antognolli <[email protected]>
Signed-off-by: Dylan Baker <[email protected]>
Tested-by: Rafael Antognolli <[email protected]>
Reviewed-by: Lyude Paul <[email protected]>
|
|
|
|
|
|
|
|
| |
Architecture benefits from having more threads/work outstanding.
Patch by Jan Zielinski.
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
|
|
| |
Allow draws in flight to be overridden via SWR_CREATECONTEXT_INFO.
Patch by Jan Zielinski.
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
|
| |
Refactored the gather operation to process 16 elements at a time via
paired SIMD8 operations.
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
| |
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
| |
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
| |
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
| |
Reviewed-by: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A number of double/int64 operations don't have matching
read and write usage masks, which the fallthrough case of
tgsi_util_get_inst_usage_mask assumes for componentwise
tagged instructions.
No regressions in llvmpipe piglit; fixes a large number of
swr regressions.
Reviewed-by: Roland Scheidegger <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|