On Thu, 19 Dec 2019, Chris Wilson <[email protected]> wrote: > Quoting Jani Nikula (2019-12-19 15:56:52) >> From: Masahiro Yamada <[email protected]> >> >> I implemented a small build rule in drivers/gpu/drm/i915/Makefile >> without relying on the special header-test-y syntax that was removed in >> commit fcbb8461fd23 ("kbuild: remove header compile test"). >> >> I excluded some headers from the test coverage. I hope somebody >> intrested can take a closer look at them. >> >> Dummy subdir Makefiles can be removed altogether as single target build >> use case is now covered by commit 394053f4a4b3 ("kbuild: make single >> targets work more correctly"). >> >> v2 by Jani: >> - add selftests/i915_perf_selftests.h to no-header-test >> - add .gitignore for *.hdrtest >> >> Cc: Chris Wilson <[email protected]> >> Cc: Masahiro Yamada <[email protected]> >> Signed-off-by: Masahiro Yamada <[email protected]> >> Signed-off-by: Jani Nikula <[email protected]> > > Whelp, I guess that explains why CI didn't tell me off for breaking the > self-contained headers. The pattern matching works, I didn't see any > residue after a make clean, and git status is not swamped with the > artifacts. Single object build still works. O=_build still works > (O=_build -j10 did not, hopefully that is not significant???)
I couldn't reproduce any issues with parallel out-of-tree builds. Fingers crossed. > Everything looks (mostly) in order, > Reviewed-by: Chris Wilson <[email protected]> Thanks, pushed. BR, Jani. > -Chris > _______________________________________________ > Intel-gfx mailing list > [email protected] > https://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Jani Nikula, Intel Open Source Graphics Center _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
