On Mon, 2026-04-27 at 16:39 -0700, Sean Christopherson wrote: > However, for mman.h, we actually need both. linux/mman.h provides the > bleeding > header flags and definitions, but sys/mman.h proves the syscall wrappers that > selftests expect/use. The below builds because sys/mman.h is getting included > elsewhere, more than likely by > tools/testing/selftests/kvm/include/test_util.h. > > Rather than updating individual tests on an ad hoc basis, what if we try our > best to handle this once?
Sounds reasonable to me. But I couldn't find the base commit that let me apply this in order to test it. Do you want me to turn it into a patch?
