NetCube-Systems-Austria/kumquat-buildroot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
e2c0f234ed
kumquat-buildroot/package/libclc/libclc.mk

47 lines
1.4 KiB
Makefile
Raw Normal View History

package/libclc: new package This patch provides libclc, an open source implementation of the library requirements of the OpenCL C programming language, as specified by the OpenCL 1.1 Specification. It is intended to be used with Clover (Mesa3D's OpenCL implementation for AMD GPUs. It needs to be compiled with host-clang, as it generates LLVM IR bitcode files containing device builtin functions for each target. Currently, libclc supports AMDGCN, R600 and NVPTX targets. As OpenCL kernels can be built dynamically on the target using libclang and libLLVM, it is necessary to have clc headers installed on the target. Buildroot removes /usr/include in its target-finalize step, so clc headers are installed to /usr/share/clc. Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr> Signed-off-by: Valentin Korenblit <valentinkorenblit@gmail.com> Reviewed-by: Romain Naour <romain.naour@smile.fr> [Thomas: - fix license information as noticed by Romain - add comment to explain why --includedir=/usr/share is used] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
2018-10-20 16:34:49 +02:00
################################################################################
#
# libclc
#
################################################################################
package/libclc: bump version to latest commit in release_90 branch Since v9.0.0, it was relicensed to the Apache License 2.0 with LLVM Exceptions. Update the license file hash. Signed-off-by: Romain Naour <romain.naour@gmail.com> Cc: Valentin Korenblit <valentinkorenblit@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
2019-10-05 19:04:12 +02:00
# Use the latest commit from release_90 branch.
LIBCLC_VERSION = d1cbc92e2ceee59963f5c3a576382e5bba31f060
package/libclc: new package This patch provides libclc, an open source implementation of the library requirements of the OpenCL C programming language, as specified by the OpenCL 1.1 Specification. It is intended to be used with Clover (Mesa3D's OpenCL implementation for AMD GPUs. It needs to be compiled with host-clang, as it generates LLVM IR bitcode files containing device builtin functions for each target. Currently, libclc supports AMDGCN, R600 and NVPTX targets. As OpenCL kernels can be built dynamically on the target using libclang and libLLVM, it is necessary to have clc headers installed on the target. Buildroot removes /usr/include in its target-finalize step, so clc headers are installed to /usr/share/clc. Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr> Signed-off-by: Valentin Korenblit <valentinkorenblit@gmail.com> Reviewed-by: Romain Naour <romain.naour@smile.fr> [Thomas: - fix license information as noticed by Romain - add comment to explain why --includedir=/usr/share is used] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
2018-10-20 16:34:49 +02:00
LIBCLC_SITE = https://git.llvm.org/git/libclc
LIBCLC_SITE_METHOD = git
package/libclc: bump version to latest commit in release_90 branch Since v9.0.0, it was relicensed to the Apache License 2.0 with LLVM Exceptions. Update the license file hash. Signed-off-by: Romain Naour <romain.naour@gmail.com> Cc: Valentin Korenblit <valentinkorenblit@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
2019-10-05 19:04:12 +02:00
LIBCLC_LICENSE = Apache-2.0 with exceptions or MIT
package/libclc: new package This patch provides libclc, an open source implementation of the library requirements of the OpenCL C programming language, as specified by the OpenCL 1.1 Specification. It is intended to be used with Clover (Mesa3D's OpenCL implementation for AMD GPUs. It needs to be compiled with host-clang, as it generates LLVM IR bitcode files containing device builtin functions for each target. Currently, libclc supports AMDGCN, R600 and NVPTX targets. As OpenCL kernels can be built dynamically on the target using libclang and libLLVM, it is necessary to have clc headers installed on the target. Buildroot removes /usr/include in its target-finalize step, so clc headers are installed to /usr/share/clc. Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr> Signed-off-by: Valentin Korenblit <valentinkorenblit@gmail.com> Reviewed-by: Romain Naour <romain.naour@smile.fr> [Thomas: - fix license information as noticed by Romain - add comment to explain why --includedir=/usr/share is used] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
2018-10-20 16:34:49 +02:00
LIBCLC_LICENSE_FILES = LICENSE.TXT
LIBCLC_DEPENDENCIES = host-clang host-llvm
LIBCLC_INSTALL_STAGING = YES
# C++ compiler is used to build a small tool (prepare-builtins) for the host.
# It must be built with the C++ compiler from the host.
#
# The headers are installed in /usr/share and not /usr/include,
# because they are needed at runtime on the target to build the OpenCL
# kernels.
LIBCLC_CONF_OPTS = \
--with-llvm-config=$(HOST_DIR)/usr/bin/llvm-config \
--prefix=/usr \
--includedir=/usr/share \
--pkgconfigdir=/usr/lib/pkgconfig \
package/libclc: disable ccache Configure.py extracts arch, vendor, and os from the passed toolchain string on line 180. When using ccache, the passed path string is "/usr/lib64/ccache/g++" which breaks the logic, causing the following error: Traceback (most recent call last): File "./configure.py", line 180, in <module> (t_arch, t_vendor, t_os) = target.split('-') ValueError: not enough values to unpack (expected 3, got 1) Use --with-cxx-compiler=$(HOSTCXX_NOCCACHE) instead of $(HOSTCXX) to fix the above error. Signed-off-by: Adam Duskett <Aduskett@gmail.com> Reviewed-by: Romain Naour <romain.naour@gmail.com> Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
2020-04-06 18:50:44 +02:00
--with-cxx-compiler=$(HOSTCXX_NOCCACHE)
package/libclc: new package This patch provides libclc, an open source implementation of the library requirements of the OpenCL C programming language, as specified by the OpenCL 1.1 Specification. It is intended to be used with Clover (Mesa3D's OpenCL implementation for AMD GPUs. It needs to be compiled with host-clang, as it generates LLVM IR bitcode files containing device builtin functions for each target. Currently, libclc supports AMDGCN, R600 and NVPTX targets. As OpenCL kernels can be built dynamically on the target using libclang and libLLVM, it is necessary to have clc headers installed on the target. Buildroot removes /usr/include in its target-finalize step, so clc headers are installed to /usr/share/clc. Signed-off-by: Valentin Korenblit <valentin.korenblit@smile.fr> Signed-off-by: Valentin Korenblit <valentinkorenblit@gmail.com> Reviewed-by: Romain Naour <romain.naour@smile.fr> [Thomas: - fix license information as noticed by Romain - add comment to explain why --includedir=/usr/share is used] Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
2018-10-20 16:34:49 +02:00
define LIBCLC_CONFIGURE_CMDS
(cd $(@D); $(TARGET_CONFIGURE_OPTS) ./configure.py $(LIBCLC_CONF_OPTS))
endef
define LIBCLC_BUILD_CMDS
$(TARGET_MAKE_ENV) $(MAKE) -C $(@D)
endef
define LIBCLC_INSTALL_TARGET_CMDS
$(TARGET_MAKE_ENV) $(MAKE) -C $(@D) DESTDIR=$(TARGET_DIR) install
endef
define LIBCLC_INSTALL_STAGING_CMDS
$(TARGET_MAKE_ENV) $(MAKE) -C $(@D) DESTDIR=$(STAGING_DIR) install
endef
$(eval $(generic-package))
Reference in New Issue Copy Permalink