svn commit: r243645 - head/usr.sbin/nfsd

[Rick Macklem]

Mateusz Guzik wrote:
> On Wed, Nov 28, 2012 at 02:47:32AM +0000, Alfred Perlstein wrote:
> > Author: alfred
> > Date: Wed Nov 28 02:47:31 2012
> > New Revision: 243645
> > URL: http://svnweb.freebsd.org/changeset/base/243645
> >
> > Log:
> >   Don't allow minthreads > maxthreads.
> >
> >   Suggested by: rmacklem
> >
> > Modified:
> >   head/usr.sbin/nfsd/nfsd.c
> >
> > Modified: head/usr.sbin/nfsd/nfsd.c
> > ==============================================================================
> > --- head/usr.sbin/nfsd/nfsd.c Wed Nov 28 02:23:59 2012 (r243644)
> > +++ head/usr.sbin/nfsd/nfsd.c Wed Nov 28 02:47:31 2012 (r243645)
> > @@ -224,6 +224,10 @@ main(int argc, char **argv)
> >  		udpflag = 1;
> >  	argv += optind;
> >  	argc -= optind;
> > + if (minthreads_set && maxthreads_set && minthreads > maxthreads)
> > + errx(EX_USAGE,
> > + "error: minthreads(%d) can't be greater than "
> > + "maxthreads(%d)", minthreads, maxthreads);
> >
> >  	/*
> >  	 * XXX
> 
> Should not this be also checked in the kernel? Looks like nfssvc_nfsd
> is
> trustful:
> [..]
> if (args) {
> nfsrv_pool->sp_minthreads = args->minthreads;
> nfsrv_pool->sp_maxthreads = args->maxthreads;
> } else {
> nfsrv_pool->sp_minthreads = 4;
> nfsrv_pool->sp_maxthreads = 4;
> }
> [..]
> 
Well, since only root can do this and I can't think of why a sysadmin
would use anything other than nfsd, I'm not sure it matters much?
(But I don't see a problem with adding a sanity check in the kernel code.)

rick

> --
> Mateusz Guzik <mjguzik gmail.com>