Date: Mon, 6 Apr 2009 11:15:04 +0100 From: Doug Rabson <dfr@rabson.org> To: Rick Macklem <rmacklem@uoguelph.ca> Cc: freebsd-fs@freebsd.org Subject: Re: nfsv4 sharing nfssvc() with the regular nfsd Message-ID: <74607C8A-226C-47FB-BFA5-E99AF535AD01@rabson.org> In-Reply-To: <Pine.GSO.4.63.0904021748300.26606@muncher.cs.uoguelph.ca> References: <Pine.GSO.4.63.0904021748300.26606@muncher.cs.uoguelph.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2 Apr 2009, at 23:01, Rick Macklem wrote:
> For nfsv4 to live side-by-side with the regular nfsd, they must either
> share the nfssvc() system call or a new one must be allocated for
> nfsv4.
>
> As such, I've cobbled some code to-gether to allow the nfssvc()
> syscall
> to be shared. It basically consists of a small module called nfssvc
> with
> only the nfssvc() syscall function in it, where nfsserver and nfsv4
> "register" with it by setting the appropriate function pointer non-
> null. These functions are then called, based on the NFSSVC_xxx flag
> value. (I've coalesced the NFSSVC_xxx flags into a separate .h file,
> to avoid confusion.)
This sounds about right.
>
> I also deleted the following, since I believe that it is just cruft.
> (sysproto.h is included in all of these files.)
> #ifndef _SYS_SYSPROTO_H_
> struct nfssvc_args {
> int flag;
> caddr_t argp;
> };
> #endif
> Is there a reason for the above?
I can't think of one so I'm going to go with 'historical reasons',
>
>
> I've attached the "diff -u" in case anyone would be willing to
> review it, rick.
The patch looks ok. The only thing I would change is to change the
names of the various call_foo variables so that they start with "nfs_"
for consistency.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?74607C8A-226C-47FB-BFA5-E99AF535AD01>