Date: Wed, 6 Nov 1996 11:14:34 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: ponds!ponds!rivers (Thomas David Rivers) Cc: ponds!lambert.org!terry, ponds!Artisoft.COM!ponds!rivers, ponds!Artisoft.COM!ponds!freebsd.org!dyson, ponds!Artisoft.COM!ponds!freebsd.org!freebsd-hackers, ponds!Artisoft.COM!ponds!lambert.org!terry Subject: Re: More info on the daily panics... Message-ID: <199611061814.LAA08484@phaeton.artisoft.com> In-Reply-To: <199611061300.IAA02838@lakes.water.net> from "Thomas David Rivers" at Nov 6, 96 08:00:35 am
next in thread | previous in thread | raw e-mail | index | archive | help
> Although this isn't the complete patch you discuss below, > I believe it to be the proper fix for 2.1.5-STABLE.... I'm > providing to a) Let others acquire it, b) Get this into 2.1.6, > c) Ensure I understand the intent of your change. > > This patch doesn't use the nice macro, but it changes getnewvnode() > to allocate a totally new node when we've wrapped around the end of > the list... > > After Terry signs off on this, can someone get it committed to > 2.1.6 (nee 2.1.5-STABLE)... > > - Thanks - > - Dave Rivers - [ ... patch elided ... ] Yes. This is an equivalent workaround. I suppose the big question is: did this fix the problem for you? -- One caution: if you do not use the macro, and the queue type is changed yet again at some later time, this will introduce a "sneaky" bug that will not be obvious. What's worse, if you are already familiar with the code, it will "look right" because it will have been right the last time you looked at it. Remember the queue type change on the mounted fs list? There was a similar bug that "looked right" (I think that's where the macro came from, FWIW). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199611061814.LAA08484>