On Thu, Oct 17, 2024 at 10:36 PM Anthonin Bonnefoy <anthonin.bonne...@datadoghq.com> wrote: > I've run some additional tests, mostly pgbench with > options=-cjit_above_cost=0 for an extended duration on an instance > that was impacted. I haven't seen any issues nor performance > regressions compared to the unpatched version. > > I will switch the commitfest entry to Ready for Committer if there's > no objection.
Thanks! I'm going to go ahead and commit this. I asked Andres if he wanted to object to this plan (as author of our LLVM stuff) and he did not. I tried, for several frustrating days, to figure out how to solve our problem using JITLink and stay on a more "supported" path, but it all seems a bit unready, as various things don't work or aren't available in various versions in our support range. At least I now have the bones of a patch to prepare for JITLink in LLVM 20 or whenever they force our hand... I'll write about that in a new thread soon.
