Re: touch and utimens troubles on new/old software combinations

[Mike Frysinger]

On Monday 02 June 2008, Daniel Jacobowitz wrote:
> On Mon, Jun 02, 2008 at 01:50:13PM -0600, Bob Proulx wrote:
> > Jim Meyering wrote:
> > > Mike Frysinger wrote:
> > > > for example, if you're running a recent version of glibc (say 2.7)
> > > > compiled against recent kernel headers (say 2.6.25) but execute on
> > > > an older kernel (say 2.6.18), then the resulting touch binary will
> > > > attempt to use utimensat() which fails with ENOSYS.
> >
> > Most common systems only support backward compatibility.  I have not
> > heard of a system which supported forward compatibility.
> >
> > In other words, compiling on a platform usually results in an
> > executable that only runs on that version or later of components of
> > the system and not usually older versions.  That has always been true
> > of unix-like systems.  Sometimes it happens to work anyway simply by
> > the happy chance that nothing in the call graph changed.  But it isn't
> > generally possible when new interfaces are added to the system.
>
> FYI, this is not how glibc handles kernel headers, and that's worked
> fine for many years.  You compile against the newest version you would
> like to take advantage of, and specify the oldest version to be
> supported separately.

exactly.  the default kernel target for Gentoo's glibc atm is 2.6.9, but we 
almost always build against the latest kernel headers possible.

> Mike, I thought the *at wrappers fell back to emulation if the
> syscalls were missing.  Is that impossible for utimensat?

i'm looking at sysdeps/unix/sysv/linux/utimensat.c and does not appear to have 
any fallback emulation support.  i'm not about to try suggesting adding such 
fallback code to glibc due to drepper, but i think it's a good idea for all 
of the *at functions.

also after reading it, i dont think gnu/stubs.h would help in this particular 
case.  gnulib would need a runtime test to detect that utimensat() is 
actually not available.
-mike

signature.asc
Description: This is a digitally signed message part.