Date: Wed, 16 Aug 95 11:36:16 MDT From: terry@cs.weber.edu (Terry Lambert) To: freebsd-current@FreeBSD.org Subject: Re: procfs problems in -current? Message-ID: <9508161736.AA09871@cs.weber.edu> In-Reply-To: <199508160608.IAA02626@uriah.heep.sax.de> from "J Wunsch" at Aug 16, 95 08:08:35 am
next in thread | previous in thread | raw e-mail | index | archive | help
> As Terry Lambert wrote: > > > > > > ./vnode_if.h: vnode_if.h > > > > > > What's this? > ... > > Why do you ask? 8-). > > Since i've ran into the following trouble: > > My user `bin' used to have checked out copies from config, sys and lkm > (and include, FWIW) below his home directory. This is where > everything happens when i'm playing with new stuff in kernelland. > Once in a while, whenever i think it's necessary to upgrade the > kernel, i `cvs update' those modules and rebuild everything > (including, when required, re-configuring). This used to work all the > time, but recently, all loaded VFS modules yielded garbage and > eventually panicked the system. (The actual reason must have been the > changes in the way the mounts are being tracked, i suppose.) It > turned out that cvs remove'ing the whole lkm module and checking it > out again actually worked, but then this is only an evidence that > something in the dependency handling of the Makefiles (or make(1) > itself :) is broken. The dependencies could be happier, especially for things that need to be rebuilt that aren't standard pieces. I'd prefer a system build rather than a kernel build for changes that affect the semantics of interfaces... I don't know how you'd enforce that if people only updated their kernel. Supposedly, you'd bump the version number in the lkm.h at the same time. Terry Lambert terry@cs.weber.edu --- 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?9508161736.AA09871>