Date: Fri, 16 Nov 2007 11:13:04 +1100 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Kris Kennaway <kris@FreeBSD.org> Cc: standards@FreeBSD.org Subject: Re: Variant behaviour from truncate over NFS and UFS Message-ID: <20071116110621.U10808@besplex.bde.org> In-Reply-To: <473CC2A0.8020603@FreeBSD.org> References: <473CC2A0.8020603@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 15 Nov 2007, Kris Kennaway wrote: > I came across the following variant behaviour. On NFS: > > hydra1# cd /nfs > hydra1# touch foo > hydra1# chmod a-w foo > hydra1# ls -l foo > -r--r--r-- 1 4294967294 wheel 0 Nov 15 21:57 foo > hydra1# truncate -s 0 foo > truncate: foo: Permission denied > hydra1# chmod u+w foo > hydra1# truncate -s 0 foo > > Compare to UFS: > > hydra1# cd /tmp > hydra1# touch foo > hydra1# chmod a-w foo > hydra1# truncate -s 0 foo > hydra1# ls -l foo > -r--r--r-- 1 root wheel 0 Nov 15 21:57 foo > > Which is correct? Both, I think. truncate(2) requires write permission, and root always has write permission, but root isn't really root for the nfs case. Anyway, the default of maproot=-2 gives much more annoying nonstandard behaviour than this -- it often prevents root doing things that normal users can do. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20071116110621.U10808>