Re: [nfsv4] re: ACL and Posix Mode bit interaction

[Lisa Week <Lisa.Week at Sun.COM>]

Jim Rees wrote:

>  - How a server uses Mode and ACL for permission checking?
>    - If there is an ACL for the file
>  	- use the ACL and ignore the Mode bits
>      else /* empty ACL, as above */
>  	- use the Mode bits
>
> The Apollo file system tried to keep mode bits and acl in sync, and it was
> horribly confusing.  Afs ignores the mode bits and just uses the acl, and it
> works much better, in my opinion.  So I agree with Rick on this one.
>  
Having the situation where the mode bits and the ACL are out of sync 
would be more confusing.

Consider a client that only supports modes.  What kind of semantics 
would out-of-sync modes and ACLs give them if they are running against a 
server that supports both modes and ACLs?  In that case, an application 
on the client which sets the mode will have no effect on the permissions 
of the file because if there is an ACL then mode bits will get ignored.  
Seems bad.

Even outside of whether or not the client supports ACLs, if an 
application is relying on the setting of a mode to take effect and to 
give reasonable semantics (that it sets the appropriate permission bits 
for owner, owner_group and other), the ACL *and* the mode must be kept 
in sync.


_______________________________________________
nfsv4 mailing list
nfsv4 at ietf.org
https://www1.ietf.org/mailman/listinfo/nfsv4