On Sat, 2006-10-21 at 09:53 +0100, Roy Bamford wrote: > On 2006.10.21 09:02, Mike Frysinger wrote: > > we're going to have to cut off support for i386 targets starting with > > glibc-2.6 ... the upstream plans are to require TLS and i386 does not > > have > > the atomic instructions required to support it > > > > some other implications ... the glibc-compat20 people will also be > > stuck with > > glibc-2.5 (as that implies no TLS support in linuxthreads) > > -mike > > > > Mike, > > "cut off" as in nothing for i386 or leave i386 at glibc-2.5? > Not that its matters much, there can't be many 80386 boxes running > Gentoo now.
*sigh* Dropping support for "i386" means not being able to use the "i386-*" CHOST. This causes trouble for building generic stages. Release Engineering will still be using CHOST="i386-*" for the "x86" stages. It just means the stages won't get glibc updates. Unfortunately, at some point, we'll get hit with some nasty exploit and we'll have to stop offering them, altogether. I'm sure at that point, we'll get the users coming out of the woodworks complaining about how we've killed the spirit of Gentoo or some other nonsense. > What is the lowest IA32 arch that will be supported ? What's your idea of "supported"? If you mean by glibc, any CHOST i486-* or above will work, though i486-* and i586-* require much more be done in software than silicon. > I'm looking for clarification because very often i386 is used to mean > anything < i686. I have a personal interest too. My server is a k6-2. Here, i386 means i386. For Release Engineering, i386 also means "x86" support. -- Chris Gianelloni Release Engineering Strategic Lead Alpha/AMD64/x86 Architecture Teams Games Developer/Council Member/Foundation Trustee Gentoo Foundation
signature.asc
Description: This is a digitally signed message part
