When more than one instance of the shader compiler is running at
once, the output of ETNA_MESA_DEBUG=dump_shaders is almost useless
as output from multiple instances will get interleaved, making it
very hard to parse.
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/30604>
There is no reason to hide the shader machine code on release
builds when ETNA_MESA_DEBUG=dump_shaders is specified. This
debug option will dump lots of NIR already, so the machine
code is in good company.
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/30604>
While some bits for gl_VertexID were already present in the
compiler, it was still missing in lots of cases. Wire those
up to make gl_VertexID usable.
Fixes piglit spec@!opengl es 3.0@gl_vertexid.
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/30604>
The idea is to better abstract loading shader metadata parameters by
using the same layout for indirect/non-indirect compute pipelines.
Graphics support will added later.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30987>
Instead of storing the stride/rsrc_word3 as part of the VBO descriptors,
pass them as parameters. This is cleaner and this will allow us
to simplify VBO in DGC.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30918>
For the similar commit c738cfe8 a review comment was:
> The EGL 1.5 changelog says that it incorporates the 3D texture
> extension's methods, but does not seem to require that they succeed.
The same could be said for Reset Notification Strategy and Robust Buffer
Access as implementations seem to be explicitely allowed to not succeed.
The EGL 1.5 spec says:
> "An EGL_BAD_MATCH error is generated if an OpenGL or OpenGL ES
> context is requested with robust buffer access and with a
> specified reset notification behavior, and the implementation
> does not support that behavior."
As we do return these errors accordingly during context creation, stop
requiring drivers to implement successfull creation of robust contexts,
as indicated by the extension.
Signed-off-by: Robert Mader <robert.mader@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30764>
Placing mtl-fw.json in src/intel/ci/mtl-fw.json works for the
mesa build, but it fails to fetch in drm-ci. Move it to the
.gitlab-ci directory so it is included in the artifacts used
for building the kernel/rootfs in drm-ci.
Signed-off-by: Vignesh Raman <vignesh.raman@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30947>
One takes almost exactly 60 seconds these days, so it sometimes fails.
The other takes 43 seconds, which is too close for comfort; allocating
as much memory as possible is also not a great thing to do on shared
runners.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30993>
It looks like unvanquished is flaky on all freedreno, so just ignore it
for now. One of the SKQP tests equally seems to be pretty enthusiastic
about flipping its status.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30993>
Indeed, the object ring_bo was not freed.
For instance, this issue is triggered with:
"piglit/bin/arb_shader_image_load_store-host-mem-barrier -auto -fbo"
while setting GALLIUM_REFCNT_LOG=refcnt.log.
Fixes: 5438b19104 ("iris: enable generated indirect draws")
Signed-off-by: Patrick Lerda <patrick9876@free.fr>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30975>
The mutex needs to be locked before accessing the handle table.
After 64ca0fd2f2 ("frontends/va: Allocate surface buffers on demand")
the issue is now much more likely to happen and can be reproduced when
transcoding using ffmpeg.
Cc: mesa-stable
Reviewed-By: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30935>
