Trivial. There's some regressions internally, related to overflow
behavior. I'll have to look at it at another time, some interactions
with vsplit/vcache are actually mind-blowing.
This reverts commit 3fa10ffb49.
The first IF statement disabled the second one.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98599
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
This patch does two things:
1. It separates the host-CPU code generation from the generic code
generation. This guards against accidently breaking things for
radeonsi in the future.
2. It makes sure we actually use both arguments and don't just compute
a square :-p
Fixes a regression introduced by commit 29279f44b3
Cc: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Instead of doing all the math with scalars, use vectors. This means the
overflow math needs to be done manually, albeit that's only really
problematic for the stride/index mul, the rest has been pretty much
moved outside the shader loop (albeit the mul could actually be optimized
away too), where things are still scalar. Because llvm is complete fail
with the zero-extend widening mul, roll our own even...
To eliminate control flow in the main shader loop fetch, provide fake
buffers (so index 0 is always valid to fetch).
Still uses aos fetch though in the end - mostly because some more code
would be needed to handle unaligned fetches in that path, and because for
most formats it won't make a difference anyway (we generate some truly
horrendous code for things like R16G16_something for instance).
Instanced fetch however stays roughly the same as before, except that
no longer the same element is fetched multiple times (I've seen a reduction
of ~3 times in main shader loop size due to apparently llvm not being able
to deduce it's really all the same with a couple instanced elements).
Also, for elts gathering, use vectorized code as well - provide a fake
elt buffer if there's no valid one bound.
The generated shaders are smaller and faster to compile (not entirely sure
about execution speed, but generally unless there's just single vertices
to handle I would expect it to be faster - there's more opportunities
for future improvements by using soa fetch).
No piglit change.
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
This is used by shader umul_hi/imul_hi functions (and soon by draw).
It's actually useful separating this out on its own, however the real
reason for doing it is because we're using an optimized sse2 version,
since the code llvm generates is atrocious (since there's no widening
mul in llvm, and it does not recognize the widening mul pattern, so
it generates code for real 64x64->64bit mul, which the cpu can't do
natively, in contrast to 32x32->64bit mul which it could do).
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
In the event that multiple threads attempt to install a graph
concurrently, protect the shared list.
Signed-off-by: Steven Toth <stoth@kernellabs.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
We're missing the closedir() to the matching opendir().
Signed-off-by: Steven Toth <stoth@kernellabs.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Instead of trying to maintain a reference counted list of valid HUD
objects, and freeing them accordingly, creating race conditions
between unanticipated multiple threads, simply accept they're
allocated once and never released until the process terminates.
They're a shared resource between multiple threads, so accept
they're always available for use.
Signed-off-by: Steven Toth <stoth@kernellabs.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
v3 [Francisco Jerez]: Loosely based on Serge's v1 of this patch in
order to avoid CL-specific enums in the clover module binary
format. In addition to other changes made in v2: Represent the CL
program binary type as the section type instead of adding a CL
API-specific enum, check that the binary types of the input objects
are valid during clLinkProgram(), pass section type as argument to
build_module_library() instead of using separate function.
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Instead, hardcode group sigsel because there are a bunch of unknown
groups, especially on SM50/SM52.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
This is similar to NVC0 and GK110 emitters where we emit
reduction operations instead of atomic operations when the
destination is not used.
Found after writing some tests which check if performance counters
return the expected value. In that case, gred_count returned 0
on gm107 while at least gk106 returned the correct value.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Fixes the following building error introduced with commit 7115e56
and related amd/common dependencies:
external/mesa/src/gallium/drivers/radeonsi/si_shader.c:6861: error: undefined reference to 'ac_is_sgpr_param'
external/mesa/src/gallium/drivers/radeonsi/si_shader.c:6951: error: undefined reference to 'ac_is_sgpr_param'
clang++: error: linker command failed with exit code 1 (use -v to see invocation)
ninja: build stopped: subcommand failed.
build/core/ninja.mk:148: recipe for target 'ninja_wrapper' failed
make: *** [ninja_wrapper] Error 1
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
The 1/W was apparently not accurate enough, and we were getting sparklies
in the distance. The closed driver also did a N-R step here.
Cc: <mesa-stable@lists.freedesktop.org>
The hardware always treats the alpha channel as unsigned, so add a shader
workaround. This is rare enough that we'll just build a monolithic vertex
shader.
The SINT case cannot actually happen in OpenGL, but I've included it for
completeness since it's just a mix of the other cases.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
This reverts commit d180de3532.
This is a radeon specific hack that causes problems on nouveau
when combined with the SHARED flag later. If radeonsi needs a fix
for this, please fix it in the driver.
[chk]
Using linear surfaces for this makes sense because tilling isn't
beneficial and the surfaces can potentially be shared with other GPUs
using the VDPAU OpenGL interop.
[airlied]
I think we need a flag that isn't SHARED/LINEAR that is more
SHARED_OTHER_GPU.
[mareko]
Does radeonsi need PIPE_BIND_VIDEO_DECODE_OUTPUT that it would translate
into linear ?
[mareko]
My only concern is decoding performance. If the decoder works in 64x1
blocks, tiling will hurt. That's the theory. I don't know how the
decoder works.
Cc: 12.0 13.0 <mesa-stable@lists.freedesktop.org>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Tested-by: Ilia Mirkin <imirkin@alum.mit.edu>
Tested-by: Nayan Deshmukh <nayan26deshmukh@gmail.com> (I+A)
This fixes a crash in Deus Ex: Mankind Divided. Release builds were
unaffected, so it's not too serious.
Cc: 11.2 12.0 13.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Fixes valgrind warnings about using uninitialized memory when starting X.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Fixes spurious assertion failure in surf_level_drm_to_winsys when
starting X, due to processing a miplevel which was never initialized.
Fixes: e9c76eeeaa ("gallium/radeon: remove radeon_surf_level::pitch_bytes")
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Previous fixes were incomplete - some code still iterated through the number
of elements provided by velem layout instead of the number stored in the key
(which is the same as the number defined by the vs). And also actually
accessed the elements from the layout directly instead of those in the key.
This mismatch could still cause crashes.
(Besides, it is a very good idea to only use data stored in the key anyway.)
v2: move null format check, remove now unnecessary function parameter,
some minor prettify
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Invalidate a "new" surface before it is bound to a render target view or
depth stencil view in order to avoid the unnecessary host side copy
of the surface data before it is rendered to.
Note that, recycled surface is already invalidated before it is reused.
Reviewed-by: Brian Paul <brianp@vmware.com>
Using untyped surface formats causes huge performance degradation on Fusion.
This reverts commit eb0ced74f6 until
the backend has a better solution to address typeless surface formats.
Currently blitter will fail if the blit format is different and
view-incompatible to the resource format. Instead of punting
to software blit which will stall the pipeline, we will
create temporary resource to allow blitter to work.
Fixes piglit test arb_copy_image-formats.
Also tested with MTT piglit, glretrace.
Reviewed-by: Brian Paul <brianp@vmware.com>
Currently we adjust the view format when we are asked to create a
BGRA render target view for BGRX surface. But we only look for
SVGA3D_B8G8R8X8_TYPELESS surface format.
With this patch, we will also check for SVGA3D_B8G8R8X8_UNORM surface format,
and use SVGA3D_B8G8R8X8_UNORM as the view format for that case.
Reviewed-by: Brian Paul <brianp@vmware.com>
This patch adds a helper function svga_format_is_typeless() which
returns TRUE if the specified format is typeless.
Reviewed-by: Brian Paul <brianp@vmware.com>
We may need to re-emit texture bindings when the framebuffer state
changes. In particular, emitting the texture binding can also involve
updating a texture from its backing copy during sampler view validation.
The backing copy is made during framebuffer validation.
This helps to fix an issue with Photoshop on VGPU9 (VMware bug 1723971).
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
With this patch, we will allow blit with copy_region if the
source and destination textures have the same sample counts.
Fixes failures with piglit tests
spec@arb_texture_float@multisample-formats 2 gl_arb_texture_float
spec@arb_texture_rg@multisample-formats 2 gl_arb_texture_rg-float
Reviewed-by: Brian Paul <brianp@vmware.com>
This patch sets the rendered-to flag for the subresource after it is
updated using the PredCopyRegion command. This is to ensure that the GB surface
will be sync up properly before it will be directly mapped to.
Tested with MTT piglit, glretrace.
Reviewed-by: Brian Paul <brianp@vmware.com>
This patch adds a flag "can_use_upload" to svga_texture structure
to avoid some checking of the upload availability at each transfer map time.
Tested with Lightsmark2008, Tropics, MTT glretrace, piglit.
Reviewed-by: Brian Paul <brianp@vmware.com>
As Thomas suggested, we'll first try to map directly to a GB surface.
If it is blocked, then we'll use texture upload buffer.
Also if a texture is already "rendered to", that is, the GB surface
is already out of sync, then we'll use the texture upload buffer
to avoid syncing the GB surface.
Tested with Lightsmark2008, Tropics, MTT piglit, glretrace.
Reviewed-by: Brian Paul <brianp@vmware.com>
This patch sets the rendered_to flag for the texture subresource that
is uploaded using the TransferFromBuffer command. This is to ensure that
the subresource will be read back or invalidated before it will be
directly mapped to. This makes sure that the content of the GB surface
will not be accidentally overwritten by the device at suspend/resume time.
Reviewed-by: Brian Paul <brianp@vmware.com>
set render_condition flag when driver performs conditional rendering.
Blit using DXPredCopyRegion command gets affected by conditional rendering so
We should check this flag while performing blit operation
Tested with piglit tests.
v2: As per Charmaine's comment, setting render_condition flag if svga_query is valid.
Tested with pigit tests.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
