[PATCH] D73699: [RISCV] Implement mayBeEmittedAsTailCall for tail call optimization

weiwei via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Feb 10 06:21:45 PST 2020


wwei marked 2 inline comments as done.
wwei added inline comments.


================
Comment at: llvm/test/CodeGen/RISCV/tail-calls.ll:177
+define i32 @duplicate_returns(i32 %a, i32 %b) nounwind {
+; CHECK-LABEL: duplicate_returns:
+; CHECK:       # %bb.3: # %if.then6
----------------
luismarques wrote:
> lewis-revill wrote:
> > Perhaps these check lines should be simplified to just the checks for the tail calls, rather than using 'CHECK-NEXT' to check every line. Just to potentially reduce the chance of this test failing for unrelated reasons. Ideally we'd be able to check the tail calls in any order too but I'm not sure how that would be done with filecheck.
> Since the called functions all have different names I guess a plain CHECK would do it?
done, using `CHECK`  should be better here.


================
Comment at: llvm/test/CodeGen/RISCV/tail-calls.ll:177
+define i32 @duplicate_returns(i32 %a, i32 %b) nounwind {
+; CHECK-LABEL: duplicate_returns:
+; CHECK:       # %bb.3: # %if.then6
----------------
wwei wrote:
> luismarques wrote:
> > lewis-revill wrote:
> > > Perhaps these check lines should be simplified to just the checks for the tail calls, rather than using 'CHECK-NEXT' to check every line. Just to potentially reduce the chance of this test failing for unrelated reasons. Ideally we'd be able to check the tail calls in any order too but I'm not sure how that would be done with filecheck.
> > Since the called functions all have different names I guess a plain CHECK would do it?
> done, using `CHECK`  should be better here.
done


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D73699/new/

https://reviews.llvm.org/D73699





More information about the llvm-commits mailing list