glibc 2.0.94, Linux, chown/lchown and non-i386 architectures
Geoff Keating
geoffk@ozemail.com.au
Sat Jul 25 06:32:00 GMT 1998
> Date: Fri, 24 Jul 1998 07:26:33 -0700
> From: Joel Klecker <jk@espy.org>
> It isn't program breakage, it's libc breakage that is the problem. If glibc
> is compiled with for example, 2.1.109 kernel headers, it is impossible to
> run that glibc with a kernel version that has the pre-2.1.8x chown/lchown
> behavior (real examples: 2.0.32 and 2.1.24).
It's not impossible to run that glibc with early kernels, it's just
that chown() returns ENOSYS.
I claim that this is correct behaviour; chown() is really not
supported on those kernels.
I know this is bad. I will probably end up using a user-space
emulation of chown() (which is also bad, but in a different and far
more interesting way). Then libc will behave the same on all kernel
versions.
--
Geoffrey Keating <geoffk@ozemail.com.au>
More information about the Libc-alpha
mailing list