RE: syscall convention

Previous message: richard offer: "RE: syscall convention"
Maybe in reply to: David Wheeler: "syscall convention"
Next in thread: David Wheeler: "Re: syscall convention"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

steven_kramerat_private

> -----Original Message-----
> From: richard offer [mailto:offerat_private]
> * I took the bait.  Realizing it's only a convention, I think 
> it may be
> * better to use fewer bytes than 8 and leave some room for 
> versioning of the
> * modules for when semantic changes are made to the interface.
> 
> While still a convention :-) I suggest that its up to the policy to do
> their own versioning.

Yes.  No disagreement here.  However, since the suggested method is proposed
to be in a comment, I wanted that comment to reflect that the ID may
include not only the name, but versioning info.

> 
> If I do CAPP v1 and then release CAPP v2. Its in my best interest that
> those are compatable, I would do that by careful selection of 
> cmd values
> rather than using a new modid.

Usually that is the case.  But let's assume CAPP v3, v4 ... v7 comes about,
and then some outside influence or <DING> bright idea <DING> hits you and
you have to radically change the module, affecting the interface.  (It's
not always in the best interest to keep things the same.  If so, we'd be
very proficient COBOL or FORTRAN programmers.)  You still may
want (for marketing reasons or whatever) to call the module CAPP, and have 
cmds be cognizant of which of the
(now 2) interfaces of CAPP they will interact with.  

> richard.
> 
--steve kramer

_______________________________________________
linux-security-module mailing list
linux-security-moduleat_private
http://mail.wirex.com/mailman/listinfo/linux-security-module