Andreas Schwab wrote: > Paul Eggert <[email protected]> writes: > >> On 05/10/2012 04:04 PM, Nikolaus Rath wrote: >>> But why isn't >>> it du using statvfs instead of statfs? >> >> m4/fsusage.m4 says why: >> >> Do not use statvfs on systems with GNU libc on Linux, because that function >> stats all preceding entries in /proc/mounts, and that makes df hang if even >> one of the corresponding file systems is hard-mounted, but not available. > > This is no longer true if the kernel is at least 2.6.36.
It's tempting to add a configure-time kernel version test, but those (both the configure-time part and the version-number test) can cause trouble. I'd prefer a run-time feature test if it were cheap enough, but even then, the statfs/statvfs choice must currently be made before compiling, due to differing sets of member names. Finally, it's not clear that there is even a problem with our current use of statfs on glibc-based Linux systems.
