On Jun 20, 2008, at 10:12 AM, Guillaume Rousse wrote:
> --
> Guillaume Rousse
> Moyens Informatiques - INRIA Futurs
> Tel: 01 69 35 69 62
> Signed-off-by: Christiaan Welvaart <[email protected]>
> --
> --- nfs-utils-1.1.1/utils/mount/mount.c.bak 2007-10-19
> 05:07:28.000000000 +0200
> +++ nfs-utils-1.1.1/utils/mount/mount.c 2007-12-12
> 20:27:45.054651566 +0100
> @@ -394,7 +394,7 @@
> if (!fake)
> print_one(spec, mount_point, fs_type, mount_opts);
>
> - if (!nomtab)
> + if (!nomtab && mtab_is_writable())
> ret = add_mtab(spec, mount_point, fs_type, flags, *extra_opts,
> 0, 0 /* these are always zero for NFS */ );
> return ret;
I think the result of this change would be that if the mtab isn't
writable (for example, if it's a symlink to /proc/mounts) then
add_mtab() wouldn't print an error message saying it couldn't write.
That seems harmless enough.
Looking at the shared routine update_mtab(), the first thing it does
is check if the mtab exists and is writable. That check should
probably be copied to utils/mount/mount.c:add_mtab() instead of using
this patch.
I might even advocate moving add_mtab to support/nfs/fstab.c.
--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com