On Mon, Oct 21, 2019 at 09:31:19PM -0700, Yonghong Song wrote: > llvm alu32 was introduced in llvm7: > https://reviews.llvm.org/rL325987 > https://reviews.llvm.org/rL325989 > Experiments showed that in general performance > is better with alu32 enabled: > https://lwn.net/Articles/775316/ > > This patch turned on alu32 with no-flavor test_progs > which is tested most often. The flavor test at > no_alu32/test_progs can be used to test without > alu32 enabled. The Makefile check for whether > llvm supports '-mattr=+alu32 -mcpu=v3' is > removed as llvm7 should be available for recent > distributions and also latest llvm is preferred > to run bpf selftests. > > Note that jmp32 is checked by -mcpu=probe and > will be enabled if the host kernel supports it. > > Cc: Jiong Wang <jiong.wang@xxxxxxxxxxxxx> > Acked-by: Andrii Nakryiko <andriin@xxxxxx> > Signed-off-by: Yonghong Song <yhs@xxxxxx> Applied, thanks! Would it make sense to include -mattr=+alu32 also into -mcpu=probe on LLVM side or is the rationale to not do it that this causes a penalty for various other, non-x86 archs when done by default (although they could opt-out at the same time via -mattr=-alu32)?
