[PATCH] D38554: Fixed ppc32 function relocations in non-pic mode

vit9696 via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Sun Oct 22 10:43:00 PDT 2017


vit9696 added a comment.

@joerg, please read carefully the comment @hfinkel posted (https://reviews.llvm.org/D38554#889007) in case we cannot understand each other.

Firstly, I did test what I submitted, and I can tell that it works expectedly.
Secondly, just as stated above — PLT is an extra indirection layer which affects the performance.
Thirdly, this is not what gcc does, and this is what llvm did previously until @rafael broke it.

If you think that adding a single (and sane) if condition to fix a regression, conform to what another compiler does, and improve the performance is adding more complexity, I have to additionally remind you about the complexity of using the current executables on embedded hardware. While I definitely heard your opinion, I have to state that the decisions made should benefit the project in total and there is no place for personal preconceived actions.

So far I was unable to hear clear reasons that are based on something why this should cause issues (please point it out, if I did not notice something). Furthermore, we started to move to the area where the proposed change is simply "disliked", and it gets stuck because I have to convince you to read the arguments already the two people provided.


https://reviews.llvm.org/D38554





More information about the llvm-commits mailing list