This reverts (partial) commit
e39ad96136, and
26642e4cc0, but also fixed additional
sites where the !BR2_RISCV_32 dependency is no longer needed, thanks
to the recently added
0001-src-xshmfence_futex.h-fix-build-on-32-bit-architectu.patch in
xlib_libxshmfence.
Signed-off-by: Thomas Devoogdt <thomas@devoogdt.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Since wayland 1.15 (upstream commit
549a5ea710f4da1a5749587176d39fef1ded4077), libwayland-egl.so is
provided by the wayland package, so there is no longer a question of
whether libwayland-egl.so is provided by the particular EGL
implementation. See the Wayland commit log:
wayland-egl: import libwayland-egl.so frontend library from Mesa
Currently the client-facing libwayland-egl API is defined by a header
file shipped by Wayland, but the implementation is left to each vendor.
This can cause collisions when multiple implementations are installed on
the same system. Importing the implementation into Wayland with a stable
and versioned driver-facing ABI allows multiple drivers to coexist on
the same system.
Pull the sample implementation from Mesa commit 677edff5cfd
("wayland-egl: rework and simplify wl_egl_window initialization")
It has been used by the Mesa open source drivers, NVIDIA and others[1].
v2: Reword commit message, rebase on top of newer Mesa.
[1] https://github.com/thayama/wayland-egl
Consequently, we remove the BR2_PACKAGE_HAS_LIBEGL_WAYLAND
option. Packages that rely on BR2_PACKAGE_HAS_LIBGLES and
BR2_PACKAGE_WAYLAND are guaranteed to have libwayland-egl.so.
Note that this doesn't solve the problem that libwayland-egl.so will be
provided both by wayland itself and by by the implementation
(rockchip-mali, sunxi-mali-utgard, ...). Still, there is a dependency
from the implementation on wayland so at least it is predictable which
one will end up on the target.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Arnout: remove remaining references in sway and sunxi-mali-utgard]
Signed-off-by: Arnout Vandecappelle <arnout@mind.be>
While fixing a build error
../src/egl/drivers/dri2/platform_x11.c: In function 'dri2_x11_get_msc_rate':
../src/egl/drivers/dri2/platform_x11.c:1229:44:
error: 'struct dri2_egl_display' has no member named 'screen_resources'
with this defconfig:
BR2_x86_64=y
BR2_TOOLCHAIN_BUILDROOT_CXX=y
BR2_PACKAGE_MESA3D=y
BR2_PACKAGE_MESA3D_GALLIUM_DRIVER_IRIS=y
BR2_PACKAGE_MESA3D_OPENGL_GLX=y
BR2_PACKAGE_MESA3D_OPENGL_EGL=y
BR2_PACKAGE_MESA3D_OPENGL_ES=y
BR2_PACKAGE_XORG7=y
(crocus and i915 drivers are also affected) it turns out that we can
assume the need for dri3 support when X.org is enabled as a hard depen-
dency even if mesa3d's configure does not throw errors when missing,
like for the Intel drivers.
Before this patch these Config.in options were used:
config BR2_PACKAGE_MESA3D_DRI3
select BR2_PACKAGE_XLIB_LIBXSHMFENCE
select BR2_PACKAGE_MESA3D_DRI3 if BR2_PACKAGE_XORG7
which can be translated into:
select BR2_PACKAGE_XLIB_LIBXSHMFENCE if BR2_PACKAGE_XORG7
and used at option BR2_PACKAGE_MESA3D_DRIVER.
Configure option -Ddri3=enabled is passed to mesa3d when at least one
driver is enabled along with X.org:
ifeq ($(BR2_PACKAGE_MESA3D_DRIVER)$(BR2_PACKAGE_XORG7),yy)
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
Signed-off-by: Arnout Vandecappelle <arnout@mind.be>
codesourcery arm/aarch64 toolchains are old (2014) and use glibc
2.18/2.20 which are not compatible with 64-bit time_t raising the
following build failure with libcgroup since commit
1c2dbcdcf0:
In file included from ./libcgroup-internal.h:25:0,
from parse.y:21:
/home/buildroot/autobuild/run/instance-3/output-1/host/arm-buildroot-linux-gnueabi/sysroot/usr/include/fts.h:41:3: error: #error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
# error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
^
Fixes: 1c2dbcdcf0
- http://autobuild.buildroot.org/results/e28f955f2b360f6e7bb231a5a3800cfbd17a23d7
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Peter: add Config.in.legacy entries]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
mesa3d 23.3 added dri3_priv.h header and dri3_check_multibuffer() function in
drisw (part of swrast gallium driver) that can be build without dri3.
i686-buildroot-linux-gnu/bin/ld: src/glx/libglx.a.p/drisw_glx.c.o: in function `driswCreateScreenDriver':
drisw_glx.c:(.text.driswCreateScreenDriver+0x3a0): undefined reference to `dri3_check_multibuffer'
collect2: error: ld returned 1 exit status
Add HAVE_DRI3 guard around dri3_priv.h header and the zink code using
dri3_check_multibuffer().
While testing again the TestGlxinfo runtime test, a new unexpected error line
appeared in the glxinfo log:
MESA-LOADER: failed to open zink: /usr/lib/dri/zink_dri.so
mesa3d try to load zink gallium driver even if it was not enabled at
build time, indeed there is no such option in Buildroot.
Apply patches sent upstream:
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27478
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/6093854455
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Buildroot commit eb60820c0a disabled
elfutils for musl toolchains in 2015. Current code builds fine with musl
so remove the exceptions.
Signed-off-by: Bernd Kuhls <bernd@kuhls.net>
[yann.morin.1998@free.fr:
- move all libc-related conditional blocks together
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
- Add new build dependencies to lld and llvm
- Add host-gcc and gcc >= 7.x dependency since clang doesn't build with gcc 6.3
[ 28%] Building CXX object utils/TableGen/CMakeFiles/obj.clang-tblgen.dir/TableGen.cpp.o
output/build/host-clang-15.0.3/utils/TableGen/RISCVVEmitter.cpp: In instantiation of ‘{anonymous}::SemaSignatureTable::init(llvm::ArrayRef<{anonymous}::SemaRecord>)::<lambda(auto:2&)> [with auto:2 = const llvm::SmallVector<clang::RISCV::PrototypeDescriptor>]’:
/usr/include/c++/6/bits/stl_algo.h:3769:5: required from ‘_Funct std::for_each(_IIter, _IIter, _Funct) [with _IIter = std::_Rb_tree_const_iterator<llvm::SmallVector<clang::RISCV::PrototypeDescriptor> >; _Funct = {anonymous}::SemaSignatureTable::init(llvm::ArrayRef<{anonymous}::SemaRecord>)::<lambda(auto:2&)>]’
output/host/include/llvm/ADT/STLExtras.h:1611:23: required from ‘UnaryFunction llvm::for_each(R&&, UnaryFunction) [with R = std::set<llvm::SmallVector<clang::RISCV::PrototypeDescriptor>, {anonymous}::SemaSignatureTable::init(llvm::ArrayRef<{anonymous}::SemaRecord>)::Compare>&; UnaryFunction = {anonymous}::SemaSignatureTable::init(llvm::ArrayRef<{anonymous}::SemaRecord>)::<lambda(auto:2&)>]’
output/build/host-clang-15.0.3/utils/TableGen/RISCVVEmitter.cpp:249:64: required from here
output/build/host-clang-15.0.3/utils/TableGen/RISCVVEmitter.cpp:249:50: error: cannot call member function ‘void {anonymous}::SemaSignatureTable::insert(llvm::ArrayRef<clang::RISCV::PrototypeDescriptor>)’ without object
llvm::for_each(Signatures, [this](auto &Sig) { insert(Sig); });
^~~~~~
utils/TableGen/CMakeFiles/obj.clang-tblgen.dir/build.make:313: recipe for target 'utils/TableGen/CMakeFiles/obj.clang-tblgen.dir/RISCVVEmitter.cpp.o' failed
make[4]: *** [utils/TableGen/CMakeFiles/obj.clang-tblgen.dir/RISCVVEmitter.cpp.o] Error 1
Indeed LLVM_TEMPORARILY_ALLOW_OLD_TOOLCHAIN is OFF, the CheckCompilerVersion.cmake
trigger an error if gcc < 7.x. Indded since llvm 15 this option check for C++17 [1]
The last gcc 7.x version (7.5) has been released in 2019.
Fixing llvm/clang issues for old compiler maybe it's not worth it.
- Remove LLVM_TEMPORARILY_ALLOW_OLD_TOOLCHAIN since
gcc >= 7 is mendatory.
When updating LLVM, we also need to update the version of
SPIRV_LLVM_TRANSLATOR because its CMakeLists.txt verifies the LLVM
base version [2].
The patch
package/llvm/0002-Add-missing-cstdint-header-to-Signals.h.patch has
been included in tag llvmorg-15.0.0 of llvm-project.
[1] https://github.com/llvm/llvm-project/commit/4c72deb613d9d8838785b431facb3eb480fb2f51
[2] https://github.com/KhronosGroup/SPIRV-LLVM-Translator/blob/5be08b8bb91fc6ca62857792c177bc7479e5e3db/CMakeLists.txt#LL82C6-L82C6
Signed-off-by: Daniel Lang <d.lang@abatec.at>
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Sebastian Weyer <sebastian.weyer@smile.fr>
Tested-by: Adam Duskett <adam.duskett@amarulasolutions.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Update 002-vc4-add-meson-option-to-disable-optional-neon-suppor.patch to
fix an syntax error introduced by commit 'package/{mesa3d, mesa3d-headers}:
bump version to 23.1.0' ([1])
Fixes:
.../build/mesa3d-23.1.2/src/gallium/drivers/vc4/meson.build:87:76: ERROR:
Trying to compare values of different types (UserFeatureOption, str) using !=.
This was deprecated and undefined behavior previously and is as of 0.60.0 a hard error.
[1] https://git.buildroot.net/buildroot/commit/?id=e6203db11ed40122a2d9f29012db5683a7eb10d0
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
No release announcement available from upstream this time so we had to
computed the tarball hash ourselves.
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Thomas Devoogdt