This commit makes mir_pack_swizzle set rep_low, rep_high and half in a
more explicit way, helping in the transition to a unified expand_mode
enum in the following commit.
Signed-off-by: Italo Nicola <italonicola@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9461>
This code is taken from src/freedreno/isa/decode.c.
Since we need a similar function in panfrost, it's probably good to move
it to utils.
Signed-off-by: Italo Nicola <italonicola@collabora.com>
Acked-by: Rob Clark <robclark@freedesktop.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9461>
This fixes a bunch of MSAA related CTS regressions. This restores
previous behaviour on GFX9 but it should be fixed properly.
Cc: 21.1 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10374>
This will prevent the driver to take the prime blit path for presentation
in scenarios where it can avoid it, which can substantially improve
performance, particularly at high resolutions.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5917>
Until now, the WSI code would rely on VK_EXT_pci_bus_info to check if the
WSI device matched the DRM device used for display. If they matched, then
it would allow direct presentation of the swapchain images, otherwise
it would fallback to a blit path through a linear buffer.
Unfortunately, this only works for PCI devices, so embedded GPUs would
always fail the test and fallback to the prime blit path, incurring in a
performance penalty due to the extra blit.
With this change driver backends have the possibility to attach a
callback to the WSI device to take control over that decision and have
freedom to make that choice according to their own platform particularities.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5917>
Patch workarounds the problem like commit eec904d29c did, this fixes
compilation on older distributions that still use old header.
Fixes: 01ba8a8d02 ("glx: Implement GLX_EXT_no_config_context")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10447>
This commit contains a hack to retrieve extended buffer info, which will
be fixed by moving the logic into gralloc HAL implementation.
With this hack, the rendering result is correct.
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10259>
$(MESA_TOP)/src/compiler/nir include path is added
for both clarity and build errors preventive reasons
Cc: 21.0 21.1 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10443>
Since generated nir headers are included, it makes sense to include nir path
Fixes the following building error:
FAILED: out/target/product/x86_64/obj/STATIC_LIBRARIES/libmesa_pipe_radeonsi_intermediates/si_shader_nir.o
...
In file included from external/mesa/src/gallium/drivers/radeonsi/si_shader_nir.c:26:
external/mesa/src/amd/common/ac_nir.h:29:10: fatal error: 'nir.h' file not found
^~~~~~~
1 error generated.
Cc: 21.0 21.1 <mesa-stable@lists.freedesktop.org>
Fixes: 1c702a8239 ("ac: move ac_lower_indirect_derefs() outside of the LLVM dir")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10443>
this only needs to be set if the mode is LAST, otherwise the normal
pipeline state can be used and this one can be omitted
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10458>
ideally we would want to require transformFeedbackPreservesTriangleFanProvokingVertex,
but as there is no fallback path, any amount of provoking vertex support
is still better than none, so tests related to this will continue to fail
for that hardware (intel)
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10458>
Present operations do not necessarily complete after the requested
time; if the CRTC has been blanked, they will complete
immediately. This means we cannot use the MSC value to tell when a
present has finished.
Instead of using MSC, track whether each present is complete by serial
number, which is the low 32-bits of the SBC value. This value is
provided in the present_pixmap request and returned in the present
complete notify event.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
This avoids the use of uninitialized memory:
Conditional jump or move depends on uninitialised value(s)
at 0x5EDBF87: r600_draw_vbo (r600_state_common.c:2232)
by 0x5DEFA00: u_vbuf_draw_vbo (u_vbuf.c:1329)
by 0x5D990EF: cso_multi_draw (cso_context.c:1445)
by 0x59B7EF4: _mesa_draw_arrays (draw.c:1346)
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10442>
Do all the necessary lowering in one place, during finalization, and
stop uselessly calling nir_lower_indirect_derefs in turnip. Splitting
i/o to elements should no longer be necessary since we use the i/o
semantics instead of variables now.
This has the side effect that we no longer generate enormous if-ladders
for tess/GS shaders with turnip.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7274>
Don't setup inputs and outputs if we aren't using
load_input/store_output intrinsics. While it's mostly harmless, there
may be more outputs than expected which would lead to an oob write of
the outputs array when setting the register id to INVALID_REG.
Also be more paranoid with asserts to catch this.
Fixes: a6291b1 ("freedreno/ir3: rework setup_{input,output} to make struct varyings work")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7274>
VK_KHR_swapchain requires VK_KHR_surface in the instance to be enabled.
This fixes a validation-error.
Fixes: bbeee415ee ("zink: Learn about VK_KHR_swapchain")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10461>
This makes the vertex order of TRISTRIP and TRISTRIP_ADJ primitves
consistent between XFB output and GS input. Technically, the Vulkan
spec allows us to XFB out in whatever order we want but being consistent
with GS inputs is probably nicer to apps.
Fixes: 36ee2fd61c "anv: Implement the basic form of VK_EXT_transform_feedback"
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10460>
PIPE_CAP_IMAGE_LOAD_FORMATTED doesn't depend on
shaderStorageImageExtendedFormats or
shaderStorageImageWriteWithoutFormat.
PIPE_SHADER_CAP_MAX_SHADER_IMAGES enables
GL_EXT_shader_image_load_store, which *does* require
shaderStorageImageExtendedFormats. Having
shaderStorageImageWriteWithoutFormat and
shaderStorageImageReadWithoutFormat isn't enough to support this.
It *might* be possible to lower extended formats to format-less
reads or writes, but we don't currently do that, so we should
just correct the test for now.
Fixes: 3f9a6d333b ("zink: export shader image caps using features")
Fixes: 5282210c0b ("zink: check correct caps for PIPE_CAP_IMAGE_LOAD_FORMATTED")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10456>
