commit: e7f0f2bac9d6d07e6056700534da627f854014c0 Author: Patrick Lauer <patrick <AT> gentoo <DOT> org> AuthorDate: Thu Sep 19 08:48:35 2024 +0000 Commit: Patrick Lauer <patrick <AT> gentoo <DOT> org> CommitDate: Thu Sep 19 08:48:58 2024 +0000 URL: https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=e7f0f2ba
dev-libs/rocm-opencl-runtime: Add segfault fix Signed-off-by: Patrick Lauer <patrick <AT> gentoo.org> .../rocm-opencl-runtime-6.1.2-fix-segfault.patch | 31 +++++++ .../rocm-opencl-runtime-6.1.2-r1.ebuild | 94 ++++++++++++++++++++++ 2 files changed, 125 insertions(+) diff --git a/dev-libs/rocm-opencl-runtime/files/rocm-opencl-runtime-6.1.2-fix-segfault.patch b/dev-libs/rocm-opencl-runtime/files/rocm-opencl-runtime-6.1.2-fix-segfault.patch new file mode 100644 index 000000000000..78f76303468f --- /dev/null +++ b/dev-libs/rocm-opencl-runtime/files/rocm-opencl-runtime-6.1.2-fix-segfault.patch @@ -0,0 +1,31 @@ +From d1fff7cea2231e7980d85790486edb70d093a1dc Mon Sep 17 00:00:00 2001 +From: Rakesh Roy <[email protected]> +Date: Wed, 20 Mar 2024 22:50:22 +0530 +Subject: [PATCH] SWDEV-445096 - Fix -O0 crash in OpenCL tests + +- With https://gerrit-git.amd.com/c/lightning/ec/llvm-project/+/1002628 applied, at -O0 Kernel::dynamicParallelism() returns true but virtual queue isn't created +- This causes segfault inside VirtualGPU::submitKernelInternal() when getVQVirtualAddress() is called + +Change-Id: Ia7af042adad2329e870c142caaac3e8fa886f8b8 +--- + rocclr/device/rocm/rocvirtual.cpp | 7 +++++-- + 1 file changed, 5 insertions(+), 2 deletions(-) + +diff --git a/rocclr/device/rocm/rocvirtual.cpp b/rocclr/device/rocm/rocvirtual.cpp +index 51080013b..1610aecfe 100644 +--- a/rocclr/device/rocm/rocvirtual.cpp ++++ b/rocclr/device/rocm/rocvirtual.cpp +@@ -3326,8 +3326,11 @@ bool VirtualGPU::submitKernelInternal(const amd::NDRangeContainer& sizes, + + if (gpuKernel.dynamicParallelism()) { + dispatchBarrierPacket(kBarrierPacketHeader, true); +- static_cast<KernelBlitManager&>(blitMgr()).runScheduler( +- getVQVirtualAddress(), schedulerParam_, schedulerQueue_, schedulerSignal_, schedulerThreads_); ++ if (virtualQueue_ != nullptr) { ++ static_cast<KernelBlitManager&>(blitMgr()).runScheduler( ++ getVQVirtualAddress(), schedulerParam_, schedulerQueue_, ++ schedulerSignal_, schedulerThreads_); ++ } + } + + // Check if image buffer write back is required diff --git a/dev-libs/rocm-opencl-runtime/rocm-opencl-runtime-6.1.2-r1.ebuild b/dev-libs/rocm-opencl-runtime/rocm-opencl-runtime-6.1.2-r1.ebuild new file mode 100644 index 000000000000..69ddac919e07 --- /dev/null +++ b/dev-libs/rocm-opencl-runtime/rocm-opencl-runtime-6.1.2-r1.ebuild @@ -0,0 +1,94 @@ +# Copyright 1999-2024 Gentoo Authors +# Distributed under the terms of the GNU General Public License v2 + +EAPI=8 + +ROCM_SKIP_GLOBALS=1 +inherit cmake edo flag-o-matic rocm + +DESCRIPTION="Radeon Open Compute OpenCL Compatible Runtime" +HOMEPAGE="https://github.com/ROCm-Developer-Tools/clr"; + +SRC_URI="https://github.com/ROCm-Developer-Tools/clr/archive/refs/tags/rocm-${PV}.tar.gz -> rocm-clr-${PV}.tar.gz" +S="${WORKDIR}/clr-rocm-${PV}/" + +LICENSE="Apache-2.0 MIT" +SLOT="0/$(ver_cut 1-2)" +KEYWORDS="~amd64" +IUSE="debug test" +RESTRICT="!test? ( test )" + +RDEPEND=">=dev-libs/rocr-runtime-6.0 + >=dev-libs/rocm-comgr-6.0 + >=dev-libs/rocm-device-libs-6.0 + >=virtual/opencl-3 + media-libs/mesa[-opencl]" +DEPEND="${RDEPEND}" +BDEPEND=">=dev-build/rocm-cmake-5.3 + media-libs/glew + test? ( >=x11-apps/mesa-progs-8.5.0[X] ) +" +PATCHES=( ${FILESDIR}/rocm-opencl-runtime-6.1.2-fix-segfault.patch ) + +src_configure() { + # -Werror=strict-aliasing + # https://bugs.gentoo.org/856088 + # https://github.com/ROCm/clr/issues/64 + # + # Do not trust it for LTO either + append-flags -fno-strict-aliasing + filter-lto + + # Fix ld.lld linker error: https://github.com/RadeonOpenCompute/ROCm-OpenCL-Runtime/issues/155 + append-ldflags $(test-flags-CCLD -Wl,--undefined-version) + + # Reported upstream: https://github.com/RadeonOpenCompute/ROCm-OpenCL-Runtime/issues/120 + append-cflags -fcommon + + local mycmakeargs=( + -Wno-dev + -DROCM_PATH="${EPREFIX}/usr" + -DBUILD_TESTS=$(usex test ON OFF) + -DEMU_ENV=ON + -DBUILD_ICD=ON + -DFILE_REORG_BACKWARD_COMPATIBILITY=OFF + -DCLR_BUILD_OCL=on + ) + cmake_src_configure +} + +src_install() { + insinto /etc/OpenCL/vendors + doins opencl/config/amdocl64.icd + + cd "${BUILD_DIR}"/opencl || die + insinto /usr/lib64 + doins amdocl/libamdocl64.so + doins tools/cltrace/libcltrace.so +} + +src_test() { + check_amdgpu + cd "${BUILD_DIR}"/tests/ocltst || die + export OCL_ICD_FILENAMES="${BUILD_DIR}"/amdocl/libamdocl64.so + local instruction1="Please start an X server using amdgpu driver (not Xvfb!)," + local instruction2="and export OCLGL_DISPLAY=\${DISPLAY} OCLGL_XAUTHORITY=\${XAUTHORITY} before reruning the test." + if [[ -n ${OCLGL_DISPLAY+x} ]]; then + export DISPLAY=${OCLGL_DISPLAY} + export XAUTHORITY=${OCLGL_XAUTHORITY} + ebegin "Running oclgl test under DISPLAY ${OCLGL_DISPLAY}" + if ! glxinfo | grep "OpenGL vendor string: AMD"; then + ewarn "${instruction1}" + ewarn "${instruction2}" + die "This display does not have AMD OpenGL vendor!" + fi + ./ocltst -m $(realpath liboclgl.so) -A ogl.exclude + eend $? || die "oclgl test failed" + else + ewarn "${instruction1}" + ewarn "${instruction2}" + die "\${OCLGL_DISPLAY} not set." + fi + edob ./ocltst -m $(realpath liboclruntime.so) -A oclruntime.exclude + edob ./ocltst -m $(realpath liboclperf.so) -A oclperf.exclude +}
