Date: Fri, 16 Jan 2004 18:24:24 -0500 (EST) From: Robert Watson <rwatson@FreeBSD.org> To: Dag-Erling =?iso-8859-1?q?Sm=F8rgrav?= <des@des.no> Cc: cvs-src@FreeBSD.org Subject: Re: cvs commit: src/sys/kern kern_descrip.c Message-ID: <Pine.NEB.3.96L.1040116182253.81408I-100000@fledge.watson.org> In-Reply-To: <xzpsmifr0ln.fsf@dwp.des.no>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 16 Jan 2004, Dag-Erling Smørgrav wrote: > "Brian F. Feldman" <green@FreeBSD.org> writes: > > I don't see a reason "newfdp" needs to be locked, since it is not > > referenced by anything yet. If "fdp" alone is locked, that is > > sufficient to ensure they will both be consistent. > > It needs to be locked because it is passed to functions which assert > that it is locked. It sounds like this is an API problem, and is probably what we should fix. I've found WITNESS an invaluable debugging tool for locking, and when programming on systems without it, it's a very painful experience (i.e., debugging lock orders on Darwin). You've also pointed out that the extra locking work being done is actually unnecessary, so maybe we just need an _unlocked() version of the API, or changes elsewhere? Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert@fledge.watson.org Senior Research Scientist, McAfee Research
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1040116182253.81408I-100000>