Date: Fri, 12 Jul 2002 19:23:27 -0400 (EDT) From: Andrew Gallatin <gallatin@cs.duke.edu> To: Ian Dowse <iedowse@maths.tcd.ie> Cc: current@freebsd.org Subject: Re: bdwrite: buffer is not busy Message-ID: <15663.25839.352038.649456@grasshopper.cs.duke.edu> In-Reply-To: <200207122218.aa20866@salmon.maths.tcd.ie> References: <20020712182534.GA336@nagual.pp.ru> <200207122218.aa20866@salmon.maths.tcd.ie>
next in thread | previous in thread | raw e-mail | index | archive | help
Ian Dowse writes: > In message <20020712182534.GA336@nagual.pp.ru>, "Andrey A. Chernov" writes: > >I see this panic constantly during last month or two, UP machine, no > >softupdates. Anybody else saw it too? Any ideas? > > The "buffer is not busy" panic is usually a secondary panic that > occurs while trying to sync the disks after a different panic. If > possible, try to get the first panic message, or ideally a stack > trace. > Why do we dump after sync'ing fs's? It seems like we should dump first, then sync later. Also, with interrupt threads in -current, panic's are totally bizzare. Alphas can't dump because they get into a situation where some system thread wakes up and can never go to sleep again (like bufdaemon). I think the whole set of panicstr hacks need to be re-thought.. I'd like to do something that (as Julian says) puts a big stick in the works just after we panic which allows only interrupt threads and the panic'ing thread to run. What's the best way to do this? Should I add to the panicstr hacks, or is there a more elegent way to achieve this? Thanks, Drew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?15663.25839.352038.649456>