Date: Fri, 13 Jun 1997 22:55:00 +1000 From: Stephen McKay <syssgm@dtir.qld.gov.au> To: freebsd-hackers@freebsd.org Cc: syssgm@dtir.qld.gov.au Subject: Re: user-mode nfs daemon Message-ID: <199706131255.WAA09027@ogre.dtir.qld.gov.au> In-Reply-To: <199706121703.KAA09145@phaeton.artisoft.com> from Terry Lambert at "Thu, 12 Jun 1997 17:03:44 %2B0000" References: <199706121703.KAA09145@phaeton.artisoft.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[Thanks to all those who pointed me at Alex and the Heidemann paper. The Heidemann paper is 100 pages. Almost big enough to be a weapon. :-) ] On Thursday, 12th June 1997, Terry Lambert wrote: >> Ah, yes, but the user mode one would be so much easier to change. No need >> for continuous build/reboot/login cycles. > >1) Build the NFS server as an LKM. >2) Load it. >3) Test it. >4) Unload it. >5) Change it. >6) Goto 1. Call me old-fashioned, but this doesn't look fun to me. I'm quite likely to incorporate some bugs in my NFS LKM, and down goes my kernel. If I'm unlucky, it takes some files, or a few file systems with it. User mode is really safe and easy for testing. There are even nice source level debuggers. The only thing I use LKM's for right now is running Linux binaries. And if that starts to be a habit, I'll probably compile it in. For me, LKM's are more cute than practical. If they were separate kernel processes, I'd be more interested. But then it probably wouldn't be FreeBSD any more. Stephen.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199706131255.WAA09027>