On 10/7/18 1:45 AM, Tobias Klausmann wrote: > Hi! > > On Sun, 07 Oct 2018, H. Peter Anvin wrote: >>> How much bother is it to keep maintaining alpha as a supported >>> target for glibc? Ultimately, it's a question if people want to >>> put in the time. There will always be users as long as there is >>> a supporting Linux distro, I guess. >>> >> >> The question is about if there is a reason to support a kernel/glibc version >> mismatch. > > Do I understand correctly that the proposed change would break > the combination of Kernel <3.2 and newer glibcs? If so, I think > we should go ahead, 3.2 is pretty old and I would be *very* > surprised if anyone is running that old a kernel but a new glibc. > As you said, Alpha production started almost one and a half > decades ago, hardware compatibility on the kernel side is very > good (aside from weird cases like the UP1500, but even that has a > post-3.2 kernel available). > > I'm sure Matt will correct me if I'm wrong about any of this. > No. That's the case for *all* architectures right now. Alpha as sole architecture is missing: http://lkml.kernel.org/r/20181007033748.224461-2-hpa@xxxxxxxxx ... and it is still not in the kernel :( -hpa
