From: Petr Tesařík <p...@tesarici.cz> Sent: Wednesday, July 10, 2024 4:32 AM > > On Wed, 10 Jul 2024 07:55:20 +0200 > Christoph Hellwig <h...@lst.de> wrote: > > > On Tue, Jul 09, 2024 at 09:08:18PM +0200, Petr Tesařík wrote: > > > I'm confused. If you're not a big fan, why are we effectively adding > > > them to more places now than before the patch? > > > > Because I didn't want to second guess the patch author too much. > > Fair enough. I don't have any relevant test cases either, so when/if > somebody encounters an issue, let them change it then. >
Works for me. FWIW, I don't have a preference either way. I was trying to carry over the existing occurrences of unlikely() and not to introduce gratuitous and unrelated changes. But because of the inconsistencies in the use of unlikely() in the 9 occurrences that were collapsed into 3 wrappers, some change was unavoidable. I did not do any analysis of the generated code with and without unlikely() to make a decision. Michael
