On Thu, 2024-07-04 at 11:23 +0800, Xi Ruoyao wrote: > On Thu, 2024-07-04 at 10:38 +0800, Huacai Chen wrote: > > > So if a LoongArch Glibc is built with Linux kernel headers >= 6.11, > > > it'll use fstatat **even configured --with-kernel=5.19** and fail to run > > > on Linux kernel <= 6.10. This will immediately blow up building Linux > > > From Scratch on a host distro with an "old" kernel. > > The patch which adds newstat back will CC the stable list and be > > backported to old kernels. > > AFAIK in Glibc --enable-kernel=x.y (not with, I was too sleepy > yesterday) means it'll work with even x.y.0. And even if we "re- > purpose" x.y to mean "the latest x.y patch release" people can still > explicitly spell the patch level, like --enable-kernel=5.19.0. > > Thus we still need to handle this in Glibc. > > And the backport will raise another question: assume 6.6.40 gets the > backport, what should we do with --enable-kernel=6.6.40? Maybe we > should we assume newfstatat is available but then people will start to > complain "hey 6.9.7 > 6.6.40 but my Glibc configured with --enable- > kernel=6.6.40 does not work on 6.9.7"... > > To me the only rational way seems only assuming 6.11 or later Or the first 6.10.x which will get the backport. > has newfstatat on LoongArch. -- Xi Ruoyao <xry111@xxxxxxxxxxx> School of Aerospace Science and Technology, Xidian University
