On 2/3/25 3:44 AM, Richard Biener wrote: > On Mon, Feb 3, 2025 at 10:32 AM H.J. Lu <hjl.to...@gmail.com> wrote: >> I believe the original patch should be reverted. Then my patch isn't needed. > > I'm OK with that, but it's not my call. I do wonder why the contributor did > not > address any of the fallout. Maybe he's gone? Peter?
Surya (she) is working on the fallout. In fact, one patch earlier this year was committed and reverted due to some aarch64 fallout. That said, Andrew mentioned on IRC that he was interested in getting that patch back in for aarch64 because it helps shrink-wrapping and he believes the patch itself wasn't bad, but exposed a latent issue that was causing the bootstrap issue on aarch64. Surya also just recently submitted another patch to help with the original fallout: [PATCH] lra: initialize allocated_hard_reg_p[] for hard regs referenced in RTL [PR118533] ...which you commented on. She is working on them. I disagree with H.J.'s comment. I have said before at the Cauldron and in some bugzilla's, that Surya's fix is a correct fix. The issues encountered here seem to be latent issues exposed by Surya's fix (read also Matz's reply) and as such, this patch should stay. The correct path here is to track down those latent issues and fix those. I've asked Surya to continue to work on the fallout, but any help from other's is greatly appreciated! Reverting now would also cause performance regressions on Power, RISC-V and ARM. Peter
