On Fri, 22 Nov 2024 08:41:34 GMT, Matthias Baesken <mbaes...@openjdk.org> wrote:
> Hi Julian, thanks for the review ! Any more comments / ideas ? Having the > lto/gcc build fixed, even with a workaround, is much better than a broken > build I think. And even with lto disabled for one single file, I see good > improvements in binary size when using lto. Yes, I agree that getting it working is more important. In the meantime, I am trying to find a solution to the FORBID_C_FUNCTION and LTO problem. It is my hope that I can find something that makes this workaround not needed, and quickly. But if you asked me if I had any ideas now, the only thing I can think of is to NOINLINE os::malloc, since that it where the inlined call to malloc originates from. But I cannot replicate the LTO failure on my WSL install for some reason, so have no way of verifying if it works. Either way, this is probably ok, though might want to ask Kim for the final go ahead, since I don't know if I count as a HotSpot developer ------------- PR Comment: https://git.openjdk.org/jdk/pull/22069#issuecomment-2493222305
