cvs commit: src/sys/kern vfs_mount.c

Yar Tikhiy yar at comp.chem.msu.su
Mon Feb 18 06:26:30 PST 2008


On Mon, Feb 18, 2008 at 01:50:34PM +0100, Dag-Erling Sm??rgrav wrote:
> Yar Tikhiy <yar at FreeBSD.org> writes:
> >   Log:
> >   In the new order of things dictated by nmount(2), a read-only mount
> >   is to be requested via a "ro" option.  At the same time, MNT_RDONLY
> >   is gradually becoming an indicator of the current state of the FS
> >   instead of a command flag.  Today passing MNT_RDONLY alone to the
> >   kernel's mount machinery will lead to various glitches.  (See the
> >   PRs for examples.)
> >   
> >   Therefore mount the root FS with a "ro" option instead of the
> >   MNT_RDONLY flag.  (Note that MNT_RDONLY still is added to the mount
> >   flags internally, by vfs_donmount(), if "ro" was specified.)
> 
> Can you guarantee that this will not f*** up the bootp / dhcp + nfsroot
> case?  There are dragons in that code which were decidedly not funny to
> track down and fix.

Indeed; or, as a Russian proverb has it, "Just don't touch the sh*t,
and it won't stink". :-) But, fortunately, I've just had a chance
to test this WRT the NFS_ROOT case.  Thank Daemon, that code path
is much clearer now than it used to be in the 4.4BSD days.

-- 
Yar


More information about the cvs-src mailing list