On Fri, Dec 6, 2013 at 1:14 PM, Dave Allan <[email protected]> wrote: > Not sure if my earlier message got eaten by spam filters. This > behavior is specified by POSIX and has been the source of extended > debate in the Linux kernel community many of whose members find it as > inexplicable as we do. See, e.g.: > > http://lwn.net/Articles/244829/ > http://thread.gmane.org/gmane.linux.kernel/565148 > http://en.wikipedia.org/wiki/Stat_%28system_call%29#Criticism_of_atime > > I may also be missing the point you're trying to make. :)
I don't see the issue being the performance aspects of atime (which as far as I can tell is what is being discussed in those links). The issue is how should atime be treated if a filesystem is mounted read-only. Linux has a "noatime" mount option which one would hope would result in atimes never changing on local native filesystems mounted with otherwise default options. One might also desire that using the "ro" mount option would result in the same thing. Whether those same desired results will occur when the situation involves NFS and the server isn't Linux is clearly going to be more problematic. As for your POSIX comment, it may be just my poor search skills, but I can't find either the mount command or mount() call as being part of POSIX in the resources that I checked. Maybe we should just assume that using "ro", "noatime", or even NFS takes us outside the scope of POSIX and anything could happen. Given that some commercial unices claim to be fully POSIX compliance, I wonder if there any notes in their test documents on what filesystems/mount options you have to use to have a fully compliant system. Bill Bogstad _______________________________________________ bblisa mailing list [email protected] http://www.bblisa.org/mailman/listinfo/bblisa
