Also remove unused backend CEIL lowering.
Single regressed gnome-shell shader due to fceil followed by f2i32
where before nir_lower_int_to_float would recognize that we already
have integer and emit mov instead of trunc for the f2i32. We can
clean this up easily once we move ntt to the backend.
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Filip Gawin <filip.gawin@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23642>
According to the HW docs, TGL+ no longer requires that a CS stall be
added to a texture cache invalidate done in the compute pipeline.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18725>
The render target flush would have been needed if it was possible to:
1) pollute the render cache and write to the data port in one draw
call.
2) perform a subsequent operation that assumed the render cache was
up-to-date.
However, this is not possible for the two glMemoryBarrier barrier bits
that get translated to this pipe barrier:
* GL_TEXTURE_FETCH_BARRIER_BIT is only used for sampling operations.
It's possible to pollute the render cache and data cache with writes
to a texture in one draw call (1). However, the GL spec states that
apps cannot assume that any existing render caches are up-to-date for
sampling the written locations immediately afterwards. Apps are
required to use glTextureBarrier before the sampling operation, so
requirement #2 is not satisfied.
* GL_PIXEL_BUFFER_BARRIER_BIT could be used for a PBO upload (2), but
it's not possible to pollute the render cache and data cache with a
PBO access in one draw call. PBOs cannot be bound to framebuffers
for rendering, so requirement #1 is not satisfied.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> (v1)
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18725>
This reverts commit 0523607ebb.
The issue that commit worked around seems to have been fixed as of
commit 1c8b4940eb ("iris: Emit flushes for push constant source
buffers"). I could no longer reproduce it from that point onward with
this revert applied.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18725>
isl will assert out otherwise. Hit this with intel_stub_gpu on arm64, but it is
a legitimate bug since someone might plug a DG2 card into a workstation-grade
arm64 or ppc64 supporting PCIe (it exists).
This forward ports the logic from crocus, which checks for both SSE at a
compile-time level as well as in the CPU caps. This might be excessive since DG2
cards apparently wouldn't work properly on old non-SSE x86 boxes anyway? I just
crocus-and-pasted.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23608>
There's plenty of places we can use these new and shiny helpers, so
let's clean up the code a bit.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23460>
this should enable compression on more intermediate fb attachments
it also means that VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT can now be set
on images where ZINK_BIND_MUTABLE is not set, so non-resource APIs need
to check ZINK_BIND_MUTABLE
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23514>
There's no reason to differentiate between primitive types and structs here. `cl_prop_for_struct`
can handle primitive types just fine.
Drop `cl_prop_for_type` and rename the existing `cl_prop_for_struct` to `cl_prop_for_type`.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23652>
It's a layering violation and really the wrong tool for the job. Add a new fn to view a given slice
as a &[u8] instead of going though the clprop machinery which creates a new Vec.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23652>
The compiler will use the unsigned bit pattern of the check and combine this
with the 1 bit, which will always result in use_sb to be zero.
Fix this by making use_sb a bool
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23647>
This new approach handles things as follows:
1. Fences won't be attached to events anymore, applications only wait on
the cv attached to the event.
2. Only the queue is allowed to update event status for non user events.
This will eliminate all remaining status updating races between the
queue and applications waiting on events.
3. Queue minimized flushing by bundling events
4. Increase cv wait timeout as there is really no point in waking up too
often.
Reduces amount of emited fences on radeonsi in luxmark 3.1 luxball by 90%
Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed by Nora Allen <blackcatgames@protonmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23612>
Currently SAMPLER_CONFIG0 is always emitted to either update the active
configuration or disable the sampler. With NTE this always emits 32 state
dwords, while there are a lot of cases that only use a small number of
samplers and never change the other samplers from their disabled state.
Track the active samplers from the last emit, so we can skip the state
emission when the sampler is already disabled. Only emit the full state
after a context flush where we don't know the previous sampler state of
the GPU.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <cgmeiner@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23579>
