Date: Thu, 15 Aug 1996 04:31:18 -0700 (PDT) From: asami@cs.berkeley.edu (Satoshi Asami) To: doc@freebsd.org Subject: kernelconfig.sgml Message-ID: <199608151131.EAA18345@silvia.HIP.Berkeley.EDU>
index | next in thread | raw e-mail
I read kernelconfig.sgml and noticed a few things. I just committed
fixes that are straightforward enough. It still leaves a couple of
questions, all in the last section ("If something goes wrong"):
(1) Can I just type "kernel.old" to boot the old kernel? I was always
taught to type "/kernel.old". I can try it at next reboot, but it
will be a while (hopefully :).
(2) I thought ps(1) and friends use "sysctl -n kern.bootfile" (or the
equivalent system call) to find the name of the kernel, in which
case the suggestion to move the working kernel to "/kernel" is
outdated, and should probably be changed to "move working kernel
to /kernel, also do sysctl -w kern.bootfile=/kernel so that ps(1)
et al. will continue to work properly".
(3) You may want to add a section "Kernel works, but I lost all
network services!". It happened once on a -current kernel. It
was fixed by recompiling ifconfig. It's caused by the same reason
as ps failures (libkvm mismatch), but looks much more chaotic (the
system spews all sorts of error messages starting from loopback
device not being available), so it may warrant a separate section,
or at least a special mention, as the user will panic (I sure
did, and I've run -current for more than a year ;).
Satoshi
help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199608151131.EAA18345>