On Thursday 19 March 2009 18:43:44 Gonzalo Tornaria wrote: > On Thu, Mar 19, 2009 at 2:35 PM, Jason Moxham <ja...@njkfrudils.plus.com> wrote: > >> Yes, except config.guess may change it (but it shouldn't, as we > >> discussed before this causes compilation failure, e.g. in a kvm > >> virtual cpu which reports wrong cpuid). > > > > My understanding is that config.guess is to "tweek" the exact cpu , the > > ABI is allready decided by then. > > 1. ABI is decided. > 2. config.guess can use ABI to figure out the exact cpu > 3. configure may still change the ABI depending on the cpu reported by > config.guess >
I didn't know configure could still change the ABI at stage 3 > For instance, if config.guess responds "pentium3", the configure > script will lower ABI to 32, even if it was 64 to start with. This is > already fixed (namely, config.guess only returns 64 bit names when ABI > is 64). > > OTOH, this is still "broken" for Athlon, i.e. config.guess may respond > "athlon" even when ABI=64, and this makes configure switch to ABI=32. > The current config.guess (trunk now) will not detect athlon once in 64 mode , so everything should be OK. > I put "broken" in quotes, because this won't happen with a real cpu, > but it does happen with a badly configured virtual cpu (this includes > default out of the box KVM, tested in kvm-72 from debian stable, and > also kvm-84 from the latest upstream source of kvm). > > gonzalo > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-devel@googlegroups.com To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en -~----------~----~----~----~------~----~------~--~---