| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| 0001-Clang-version-munging.patch | Clang version munging | Maxime Chambonnet <maxzor@maxzor.eu> | yes | 2022-02-11 | ||
| 0002-fix-cmake-library-notfound-check.patch | fix cmake library notfound check If find_library does not find the library, the given variable is set with a value that has a -NOTFOUND suffix. For example, the CLANGRT_BUILTINS variable will be set with the value CLANGRT_BUILTINS-NOTFOUND. |
Cordell Bloor <cgmb@slerp.xyz> | no | 2022-10-24 | ||
| 0003-hip-config.cmake.patch | hip-config.cmake | Maxime Chambonnet <maxzor@maxzor.eu> | no | 2022-01-27 | ||
| 0004-hip-cmake-install.patch | hip cmake install | Maxime Chambonnet <maxzor@maxzor.eu> | no | 2022-02-08 | ||
| 0005-clang-15.patch | cram clang-15 into build paths | tienne Mollier <emollier@debian.org> | yes | 2022-07-03 | ||
| 0007-isystem-removal.patch | -isystem breaks proper order of C and C++ include paths The addition of |-isystem "$HIP_CLANG_INCLUDE_PATH/.."| to HIPCXXFLAGS and HIPCFLAGS needs to be removed from hipcc [1]. Once HIP is installed in /usr, hipcc ends up passing -isystem /usr/include, which breakes the proper order of C and C++ include paths. If I understand correctly, it's unnecessary for hipcc to pass these flags anyway. A patch to remove them will eventually be made upstream. |
Cordell Bloor <cgmb-deb@slerp.xyz> | yes | 2022-06-01 | ||
| 0008-dotdots.patch | adjust dotdots directories around _IMPORT_PREFIX Note the patch has been partially applied around ROCm 5.2, so there could be a consistency issue if the below code were to not be necessary anymore. |
Cordell Bloor <cgmb-deb@slerp.xyz> | yes | 2022-06-02 | ||
| 0009-devicelib-searchpath.patch | remove check on amdgcn/bitcode The device libraries are not found correctly by default and the if (NOT EXISTS ${AMD_DEVICE_LIBS_PREFIX}/amdgcn/bitcode) check is triggered [6]. I think the goal upstream is to eventually remove this from the hip::device interface entirely, but for now just correcting the search paths would be a good step. |
Cordell Bloor <cgmb-deb@slerp.xyz> | yes | 2022-06-04 | ||
| 0010-gnuinstalldirs.patch | hipamd is not consistently using GNUInstallDirs Quoting the exchange: . the PACKAGE_PREFIX_DIR defined on line 27 of hip-config.cmake needs one more "../". Though, that seems to come from CMakePackageConfigHelpers [3], so I'm not sure why it's inaccurate. . I now believe the problem is that hipamd is not consistently using GNUInstallDirs. Jeremy Newton has proposed a more complete patch upstream, but a small change is sufficient to fix this particular issue: |
Cordell Bloor <cgmb-deb@slerp.xyz> | not-needed | upstream, https://lists.debian.org/debian-ai/2022/06/msg00002.html | 2022-06-14 | |
| 0011-enable-prevega.patch | enable support for pre-Vega GPU. This eases testing of unsupported cards upstream; may be kept around in case the debian project were to want to support those cards. |
tienne Mollier <emollier@debian.org> | not-needed | 2022-06-15 | ||
| 0012-typo.patch | fix typo caught by lintian. | tienne Mollier <emollier@debian.org> | yes | 2022-06-18 | ||
| 0013-hipcc-hip-version.patch | enforce --hip-version option within hipcc call. This change is needed since the normal mechanism in ROCm relies on reading a .hipVersion hidden configuration file which has no proper equivalent in a FHS compliant tree, as of ROCm 5.0. |
Cordell Bloor <cgmb-deb@slerp.xyz> | not-needed | upstream | 2022-07-04 | |
| 0014-gcc-12.patch | fix ftbfs with gcc-12 This is a, quite possibly ugly, hack to fix a failure to build from source due to the missing definition of `noinline`, which is supposed to be a compiler specific __attribute__ flag (for some reason it isn't recognized). |
tienne Mollier <emollier@debian.org> | yes | upstream | 2022-07-23 | |
| 0015-default-hip-clang-path.patch | HIP_CLANG_PATH: add check for clang in /usr/bin | Cordell Bloor <cgmb@slerp.xyz> | not-needed | 2022-09-18 | ||
| 0016-default-device-lib-path.patch | DEVICE_LIB_PATH: add check for multiarch dirs | Cordell Bloor <cgmb@slerp.xyz> | no | 2022-09-18 | ||
| 0017-default-rocm-path.patch | ROCM_PATH: use /usr as default | Cordell Bloor <cgmb@slerp.xyz> | not-needed | 2022-09-18 | ||
| 0018-rocm-path-flag.patch | pass rocm_path as flag | Cordell Bloor <cgmb@slerp.xyz> | not-needed | 2022-09-19 | ||
| 0019-fix-git-version-info.patch | [PATCH] SWDEV-335902 - Fix HIP when Git info unavailable Prior to this change, when Git_FOUND was false, HIP_VERSION_BUILD_ID would be undefined in the CMake code. The value of HIP_VERSION_BUILD_ID in <hip/hip_version.h> is taken from the CMake variable, so it was being defined as nothing in those cases. That would cause compilation failures, as src/hip_global.cpp contains the function: size_t amd_dbgapi_get_build_id() { return HIP_VERSION_BUILD_ID; } which would become size_t amd_dbgapi_get_build_id() { return ; } after preprocessing. To prevent this, we can define the version information to a default value when Git is not found. A related problem was reported by Harmen Stoppels in https://github.com/ROCm-Developer-Tools/HIP/pull/2218. When Git is not available (or if the library is being built from a tarball), the HIP_VERSION_GITHASH is not defined. This causes trouble because HIP_LIB_VERSION_STRING is defined as "X.Y.Z-${HIP_VERSION_GITHASH}" and therefore becomes "X.Y.Z-". The incomplete version string becomes a problem when it is appended to the shared library file name. File names that end with '-' confuse the linker. They cause strange errors when attempting to link to the HIP library. This problem can be prevented by dropping the trailing dash and using "X.Y.Z" as the version string when HIP_VERSION_GITHASH is not defined. |
Cordell Bloor <Cordell.Bloor@amd.com> | no | 2022-05-06 | ||
| 0020-hipcc-remove-rpath-flags.patch | hipcc remove rpath flags | Cordell Bloor <cgmb@slerp.xyz> | not-needed | 2022-10-21 | ||
| 0020-replace-x86_64-with-variables.patch | replace x86_64 with variables The changes to hipcc.pl and hip_embed_pch won't allow cross-compiling, but they will enable the build to proceed on other architectures. |
Cordell Bloor <cgmb@slerp.xyz> | yes | upstream | 2023-02-14 | |
| 0021-fix-default-cmake-build-on-unsupported-gpus.patch | fix default cmake build on unsupported gpus If AMDGPU_TARGETS is not set (or is set to nothing), then hipcc and clang will autodetect the installed hardware. This default, however, prevents that autodetection by setting it to a specific set of values. |
Cordell Bloor <cgmb@slerp.xyz> | not-needed | 2023-02-17 |