On Mon, Nov 04, 2024 at 03:43:29PM +0000, Kyrylo Tkachov wrote: > Here’s a rework of the original patch which now doesn’t call simplify_rtx on > its operands in the new simplify_rotate_op. > It wasn’t necessary to do the necessary optimization and it was blowing up > the recursion depth needlessly. > With this version the sha1.c case compiles as fast as before and x86 > bootstrap completes fine. > AArch64 bootstrap and test passes fine as well. > Ok for mainline? (There are some aarch64 optimization tests in later applied > patches in the series that depend on this going in)
LGTM, thanks. Jakub