This issue is always there, it just shows up with newer LLVM since
symbol mismatches are being found otherwise it was happily linking with
host libLLVM.so silently.
(From OE-Core rev: 941089fa63e2b38608291251d52a354f0f624926)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
llvm-17 support patch had redundant checks for llvm-17, Simplify them as
submitted in v3 upstream
(From OE-Core rev: 97f97ff02514303b71a07e53d2a486a12861814e)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Recently mesa added a patch that allows to cross-compile Intel Vulkan
ray tracing support.
- Backport this patch
- Build and install intel_clc for mesa-native
- Add a dependency on mesa-native to provide intel_clc for target build
- Add a dependency on python3-ply-native as needed to build intel-clc
- Automatically build Intel Vulkan ray tracing support if the opencl
packageconfig was added
(From OE-Core rev: 25fbe4d1dcfa329f0229eb3ed384b8b84583bd45)
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop patches:
0001-gbm-backend-fix-gbm-compile-without-dri.patch
0001-freedreno-pm4-Use-unsigned-instead-of-uint-to-fix-mu.patch
(backports)
0001-util-format-Check-for-NEON-before-using-it.patch
(upstream submission has stalled with unanswered concerns from upstream;
patch no longer applies; issue does not reproduce with 'raspberrypi0-wifi' MACHINE)
(From OE-Core rev: cdb5d1662bd9df5e9137a52187c94792d2d3247d)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Commit introduces a patch that allows for gbm to
be built with an empty backend. There are situation
where mesa-gl is the preferred provider for virtual/libgbm,
virtual/libgl, virtual/mesa, etc... But the x11 DISTRO_FEATURE
isn't included this leads to build errors such as
| /../../../ld: src/gbm/libgbm.so.1.0.0.p/main_backend.c.o: in function
`find_backend':
| backend.c:(.text.find_backend+0xa4): undefined reference to
`gbm_dri_backend'
| /../../../ld:
src/gbm/libgbm.so.1.0.0.p/main_backend.c.o:(.data.rel.ro.builtin_backends+0x4):
undefined reference to `gbm_dri_backend'
| collect2: error: ld returned 1 exit status
Add patch to bypass compilation issue by excluding gbm dri backend.
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20447842ca28465
Commit also includes the use of TARGET_CFLAGS variable in mesa-gl recipe
to fix bellow compilation bug.
/../recipe-sysroot/usr/include/xf86drm.h:40:10: fatal error: drm.h: No such file or directory
| 40 | #include <drm.h>
(From OE-Core rev: 00a662c5e28235fb64348aab4ce9aeb5e7f95d06)
Signed-off-by: Vincent Davis Jr <vince@underview.tech>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- add PACKAGECONFIG to build vulkan beta drivers
- add PACKAGECONFIG for zink and build it along with vulkan
- remove TLS-ELF patch and the associated PACKAGECONFIG. It looks like its unneeded
- remove swrast_kms backport patch
- remove the patch that reverts the deprecation of drm_handle as it is not applicable. Still needed ?
- fix patch fuzz warnings
License file has been changed, but no change of licenses
e6392fcf3d
(From OE-Core rev: a85d0566c0e09e7ccd394c6465305ab97ce5973a)
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It fails to compile mesa{,-native} when DEBUG_BUILD is enabled:
../mesa-22.1.3/src/compiler/nir/nir_inline_helpers.h: In function ‘nir_opt_move_block’:
../mesa-22.1.3/src/compiler/nir/nir_opt_move.c:55:1: error: inlining failed in call to
always_inline ‘src_is_ssa’: indirect function call with a yet undetermined callee
src_is_ssa(nir_src *src, void *state)
^~~~~~~~~~
Backport patch from mesa merge request to fix it.
(From OE-Core rev: 468715c42632a8467dc1f0abe0f3cd540de5f7c3)
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- remove 0001-v3dv-account-for-64bit-time_t-on-32bit-arches.patch that was applied upstream
- adjust 0002-meson.build-make-TLS-ELF-optional.patch
- fix patch fuzz
- DRI1 drivers have been removed! For all AMD/Nvidia Cards gallium-llvm is required
License-Update: removed references to deleted code.
RP: Fix dri PACKAGECONFIG reference to opengl instead
(From OE-Core rev: 325ea352a8326f3b35a8a591f8f7a7dfa890cbcb)
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Since 80923e8d ("util/format: Add some NEON intrinsics-based
u_format_unpack.") upstream the build fails on Arm platforms which use
the soft-float ABI, such as qemuarmv5:
arm_neon.h:31:2: error: #error "NEON intrinsics not available with the
soft-float ABI. Please use -mfloat-abi=softfp or -mfloat-abi=hard"
Take a patch from upstream to check the ABI being used before trying to
use NEON instructions.
(From OE-Core rev: f3d82abb781c2642a50a60033641194d1e40c212)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua's vc4/v3d/dma-patch was applied in 21.1.7
Release notes 21.1.7:
Bug fixes
Miscompilation of a switch case
[radeonsi][regression] CPU is being used ~10 times more than usual after c5478f9067f.
dEQP-VK.wsi.android.swapchain.create#image_swapchain_create_info crash on Android R
The image is distorted while use iGPU(Intel GPU) rendering and output via dGPU (AMD GPU)
lima: regression in plbu scissors cmd
Incorrect rendering
Changes
pan/bi: Restrict swizzles on same cycle temporaries
nir/lower_mediump_io: Don’t remap base unless needed
nir/lower_mediump: Fix metadata in all passes
util/fossilize_db: Update parsed_offset correctly.
util/fossilize_db: Reset file position to parsed_offset on cache_offset read failure.
util/fossilize_db: Flush files after header write.
util/fossilize_db: Be conservative about header length check for locking.
util/fossilize_db: Only allocate entries after full read.
util/fossilize_db: Use uint64_t for file size.
util/fossilize_db: Unlock the cache file if the entry already exists.
util/fossilize_db: Add extra flock mutex.
radv: Allocate space for inline push constants.
vulkan/wsi/x11: do not inherit last_present_mode
ir3: Preserve gl_ViewportIndex in the binning shader
teximage: return correct desktop GL error for compressedteximage
intel/fs: restrict max push length on older GPUs to a smaller amount
virgl: disable anisotropic filtering.
freedreno/ir3: Add build id to the disassembler test
.pick_status.json: Update to 7055282231cdad23e7b23fb49a2f953a99797f8a
.pick_status.json: Update to 2f06642b06c94f07f4bfe3df32b25659428c9144
.pick_status.json: Mark 6ccf11ac2b951963407601b5b1109b45a972e48e as denominated
.pick_status.json: Update to 2f06642b06c94f07f4bfe3df32b25659428c9144
.pick_status.json: Update to 353e6323936b812884fa677c25bdba91554881b7
.pick_status.json: Update to 4128acdee32eadfef1fb8c3c3ce2e512a267cdf4
pick-ui: drop assert that optional argument is passed
pick-ui: show nomination type in the UI
pick-ui: show commit date
lima: avoid crash with negative viewport values
zink: make codegen compatible with python 3.5
mesa/main: Check for fbo attachments when importing EGL images to textures
v3d, vc4: Fix dmabuf import for non-scanout buffers
gallium/hud: initialize query
nv30: fix emulated vertex index buffers
gallium: Reset {d,r}Priv in dri_unbind_context
loader/dri3: create linear buffer with scanout support
anv: fix submission batching with perf queries
anv: don’t try to access Android swapchains
renderonly: don’t complain when GPU import fails
etnaviv: always try to create KMS side handles for imported resources
glsl: evaluate switch expression once
zink: use array size in spirv bo length calculations
etnaviv: fix gbm_bo_get_handle_for_plane for multiplanar images
clover/spirv: Properly size 3-component vector args
clover/nir: Set constant buffer pointer size to host
gallium/va: don’t use key=NULL in hash tables
amd/registers: fix fields conflict detection
radeonsi: don’t create an infinite number of variants
nir/algebraic: fix imod by negative power-of-two
radv: fix selecting the first active CU when profiling with SQTT
etnaviv: fix renderonly check in etna_resource_alloc
etnaviv: fail in get_handle(TYPE_KMS) without a scanout resource
freedreno: fail in get_handle(TYPE_KMS) without a scanout resource
panfrost: fail in get_handle(TYPE_KMS) without a scanout resource
lima: fail in get_handle(TYPE_KMS) without a scanout resource
nine: Fix assert in tx_src_param
intel/compiler: make sure swizzle is applied to if condition
glsl: Add operator for .length() method on implicitly-sized arrays
glsl: Properly handle .length() of an unsized array
gallium/vl: don’t leak fd in vl_dri3_screen_create
Release notes 21.1.6:
Bug fixes
[RADV] FSR in Resident Evil: Village looks very pixelated on Polaris
Crash in update_buffers after closing KDE “splash screen” downloader
Crash in update_buffers after closing KDE “splash screen” downloader
radeonsi: persistent, read-only buffer maps are slow to read
substance painter flickering with jagged texture and masks shown black
radv: FP16 mode in FidelityFX FSR doesn’t look right
[RADV] Blocky corruption in Scarlet Nexus and vkd3d-proton 2.4
Radeon RX580 and 5700 XT: Reloading ARB assembly shaders causes very glitchy rendering
Changes
targets/graw-xlib: Add missing dep_x11
egl/surfaceless: try kms_swrast before swrast
meson: allow egl_native_platform to be specified
freedreno/computerator: Fix local_size typo
aco/optimizer: ensure to not erase high bits when propagating packed constants
egl/wayland: Error on invalid native window
egl/wayland: Allow EGLSurface to outlive wl_egl_window
softpipe/aniso: move DDQ calculation to after scaling.
llvmpipe: add support for time elapsed queries.
.pick_status.json: Mark a9c9a9938d6d2ead6f08ccf5256c95801fbc864f as backported
.pick_status.json: Update to 376fb4f55f080929df8ca5a2a975fc88a00eec06
.pick_status.json: Update to 1d6c790f9f6c373e667de955822b1048c7073ba3
.pick_status.json: Update to 73262d4bd4a4637a332264a165c62ce3209d03fb
zink: remove duplicate format-mapping on little-endian
vulkan: do not map zero-sized region of memory
vulkan: allocate host-visible memory for swapchain images
nir/gcm: Clear out pass_flags before starting
iris: Re-emit MEDIA_VFE_STATE for variable group size shaders
anv: Handle errors properly in anv_i915_query
v3d/driconf: Expose non-MSAA texture limits for mutter and gnome-shell
lavapipe: Use common Vulkan format helpers
broadcom/compiler: emit TMU flush before a jump
nv50/ir/nir: fix smem size for GL
radeonsi: don’t expose no-attachment MSAA 16x on all 1 RB chips due to issues
st/mesa: always use PIPE_USAGE_STAGING for GL_MAP_READ_BIT usage
egl/android: include “util/compiler.h” for FALLTHROUGH macro
util/foz: stop crashing on destroy if prepare hasn’t been called
radeonsi: Check aux_context on si_destroy_screen()
mesa: clear shader_info::is_lowered in prog_to_nir
mesa: fix bindless uniform samplers update
aco: don’t create v_madmk_f32/v_madak_f32 from v_fma_legacy_f16
freedreno/ir3: Don’t force RTNE if rounding mode is undefined
radv: only init the TC-compat ZRANGE metadata for the depth aspect
glsl: replace some C++ code with C
util: add workaround for Full Bore
glsl: relax rule on varying matching for shaders older than 4.20
lima: handle fp16 vertex formats
(From OE-Core rev: ce202e5f3f0eaceec2e23ebd1697f73c9fa00aaf)
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mesa would unnecessarily fail in the v3d and vc4 driver if a dmabuf
could not imported for scanout. In particular, this meant the
weston-simple-dmabuf-egl test program would fail on a Raspberry Pi.
(From OE-Core rev: 6396c42df4609b4515f197926388ff4acb11fe52)
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
0001-anv-fix-a-build-race-between-generating-a-header-and.patch
merged upstream.
osmesa is now a simple boolean option.
Swrast has been removed from dri drivers, so as we need at least one
driver to enable dri in native/nativesdk, replace it with nouveau
(which does not take long to build).
(From OE-Core rev: e4e4cc57a0a949400c557073af0cf78f977bd02d)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Two of the patches have been applied upstream:
0001-meson-Add-xcb-fixes-to-loader-when-using-x11-and-dri.patch
0005-vc4-use-intmax_t-for-formatted-output-of-timespec-me.patch
The 'surfaceless' platform is now enabled by default, therefore it is no
longer a valid choice. The meson build system is intelligent enough now to
decide if 'drm' needs to be enabled based on whether or not gbm is available.
Therefore 'drm' also is no longer a valid platform choice:
ERROR: Options "drm, surfaceless" are not in allowed choices: "auto, x11, wayland, haiku, android, windows"
RP: Add missing mesa-gl upgrade
(From OE-Core rev: de68fe9268092c0600a81bedfc805eaed9dcb87b)
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Backport fix for undefined symbol for xcb_xfixes_create_region in loader_dri3_helper.c.
(From OE-Core rev: fd55d41c9d2e79fd7f7665107d1640a2dd9a3dc3)
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build tweaks:
- drop separate dri option and the patch that adds it; just setting -Ddri-drivers='' is enough
- strip leading/trailing commas from drivers' lists, as upstream no longer allows blank entries if the list
is not empty (e.g. -Ddri-drivers=,driver1,driver2 is now producing an error).
- drop the long-obsolete llvm 3.3/3.2 logic
License-Update: reformatted into rst
(From OE-Core rev: 8682024ef029241c75e224174fd0b93835f88df3)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
meson is detecting it does not support 64bit attomics but then when
defining local suppliments confused clang
../mesa-19.3.1/src/util/u_atomic.c:38:1: error: cannot redeclare builtin function '__sync_add_and_fetch_8'
__sync_add_and_fetch_8(uint64_t *ptr, uint64_t val)
^
../mesa-19.3.1/src/util/u_atomic.c:38:1: note: '__sync_add_and_fetch_8' is a builtin with type 'long long (volatile long long *, long long, ...)'
../mesa-19.3.1/src/util/u_atomic.c:38:1: error: definition of builtin function '__sync_add_and_fetch_8'
__sync_add_and_fetch_8(uint64_t *ptr, uint64_t val)
^
(From OE-Core rev: a18ad964d000a15ac4039c893dd9d3f8e545c223)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Upgrade mesa to 19.3 and refresh the patches.
As MESA_EGL_NO_X11_HEADERS was renamed to EGL_NO_X11 in [1] we are going
to change the do_install and add a patch to libepoxy to use the new
define.
1: 6202a13b71
(From OE-Core rev: 7a3d575ba56311c46122dfe9fa88094859fe9713)
Signed-off-by: Alistair Francis <alistair@alistair23.me>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Upgrade mesa and mesa-gl to 19.2.4.
The license hash change was a trivial new line removal.
The glx-tls option was removed as it isn't included in the meson.build
file. It has been replaced with 'use-elf-tls' instead.
I have backported the asm removal as an attempt to fix the musl build
issue.
(From OE-Core rev: 82c2acc1f658f039becac04aa3dae696418bd1de)
Signed-off-by: Alistair Francis <alistair@alistair23.me>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is a typical scenario with arm platforms, with meson if DRIDRIVERS
is empty then dri is disabled automatically, but gbm needs dri2 enabled
for it to have proper backend, in autotools we could use --enable-gbm
--enable-dri and get correct behaviour but with meson enabling dri2 only
works if DRIDRIVERS is not empty
This fixes mesa-gl build for rpi machines when using userland graphics
(From OE-Core rev: 0c670a0fad379c3dc2a69ac5abfeaceda98a2f9c)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
- Patch 0005-egl-add-missing-include-stddef.h-in-egldevice.h.patch
was applied on commit e68777c87ceed02ab199b32f941778c3cf97c794.
- Refresh all patches
- mesa 19.0.0 deprecated the use of autotools and we need to add
--enable-autotools flag. For details see mesa commit:
e68777c87ceed02ab199b32f941778c3cf97c794
The complete change log can be found here:
https://www.mesa3d.org/relnotes/19.0.0.html
(From OE-Core rev: 50b7418869b7d2f40adbfbb0844f710d9aa5c396)
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When wayland and egl both are on then it enables the wayland
plugin which actually fails to build because its not finding wayland-protocol
xml templates in proper location which is recipe sysroot since
we are cross compiling
Fixes build errors e.g.
make[4]: *** No rule to make target `//usr/share/wayland-protocols/unstable/linux-dmabuf/linux-dmabuf-unstable-v1.xml', needed by `drivers/dri2/linux-dmabuf-unstable-v1-client-protocol.h'. Stop.
make[4]: *** Waiting for unfinished jobs....
(From OE-Core rev: ff4762e94faa095abd7175810d1053fdd361058f)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
