Hi Trond,
I was wondering why Frank's patch to bypass the 16 group limitation in
RPC (http://frankvm.xs4all.nl/nfs-ngroups/) has not been incorporated
into mainline yet. Is it because he has not addressed your
concern/recommendation as mentioned in
http://marc.theaimsgroup.com/?l=linux-nfs&m=87941837304355&w=2
If the patch were modified to follow your recommendation in the same
post, would that be acceptable?
Thanks much,
---
Bruce Allan <[email protected]>
Software Engineer, Linux Technology Center
IBM Corporation, Beaverton OR
503-578-4187 IBM Tie-line 775-4187
-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. http://www.ostg.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
P=E5 fr , 30/07/2004 klokka 12:14, skreiv Bruce Allan:
> Hi Trond,
>=20
> I was wondering why Frank's patch to bypass the 16 group limitation in
> RPC (http://frankvm.xs4all.nl/nfs-ngroups/) has not been incorporated
> into mainline yet. Is it because he has not addressed your
> concern/recommendation as mentioned in
> http://marc.theaimsgroup.com/?l=3Dlinux-nfs&m=3D87941837304355&w=3D2
>=20
> If the patch were modified to follow your recommendation in the same
> post, would that be acceptable?
I'm certainly not prepared to merge any new features into the 2.4
kernel, so the patches as they stand are not going in no matter what.
However, I'm in general seriously sceptical about implementing this sort
of thing: when we only had SYS_UNIX as an option, then it might have
been an unsurmountable problem (though even there it was possible to
work around by judicious choices). However now that we have RPCSEC_GSS
with support for strong authentication (such as w/ krb5), it is possible
for the server to know exactly which groups the user is a member of.
Why go for a heuristic algorithm, when there are deterministic ones that
can do better?
Cheers,
Trond
-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. http://www.ostg.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
On Fri, Jul 30, 2004 at 01:40:00PM -0500, Trond Myklebust wrote:
>
> I'm certainly not prepared to merge any new features into the 2.4
> kernel, so the patches as they stand are not going in no matter what.
What about 2.6?
Right now I'm considering porting the patch to 2.6.7 so any advice which
might help getting it merged is welcome.
> However now that we have RPCSEC_GSS
> with support for strong authentication (such as w/ krb5), it is possible
> for the server to know exactly which groups the user is a member of.
That may not be an option.
>
> Why go for a heuristic algorithm, when there are deterministic ones that
> can do better?
The patch is pretty deterministic.
What made me reluctant writing a new version of the patch is that it would
most likely turn out to be more intrusive, further reducing chances of
getting it merged.
--
Frank
-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. http://www.ostg.com
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs