On 6/3/19 10:08 AM, Nick Gasson wrote: > Hi Andrew, > > On 03/06/2019 16:32, Andrew Haley wrote: >> >> I don't think there's any need to boycott the whole compiler because >> of this bug. It would make more sense to try to get clang fixed. >> >> Maybe for now put the cast in #ifdef CLANG ? >> > > I think I'd rather put the __atomic_fetch_add in an #ifdef CLANG? This > seems a bit neater than casting the integer argument to a pointer.
I guess the solution which uses __atomic_add_fetch instead is best. > Did you have any thoughts about os::current_stack_pointer() from my > earlier email? We need to know what it's used for to know which solution is right. I'm guessing the primary use case is asynchronous runtime stack probes, used by debugging tools. -- Andrew Haley Java Platform Lead Engineer Red Hat UK Ltd. <https://www.redhat.com> https://keybase.io/andrewhaley EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671
