[OpenAFS-devel] Re: [OpenAFS] 2.6 kernel support anytime soon?Workarounds?

[Derek Atkins]

"Douglas E. Engert" <deengert@anl.gov> writes:

> That could be a problem if they then tried to access AFS the PAG 
> could not be found. But is this only in some error situation?
> How often do you see process A get a PAG, then start process B
> then process A ends, and Process B then tries to access AFS?
> Or A with a PAG starts B that starts C, B ends then C tries
> to access AFS. 
>
> A process has to have some parent.What is the parent of C 
> at this time? Is it 1 or is it A. If its A then we mightr need
> to require the proces that got the PAG to stay active. I know it ugly. 

Ahh, but you've now opened up a hole.  Process A has a PAG (p1) and
forks process B which obtains a new PAG (p2).  Process B forks yet
another process C (which doesn't obtain a new PAG, just uses
inheritence).  If C can kill B before it accesses AFS, it's possible
that C can escape the PAG jail and reclaim p1 (assuming it can get
itself reparented to A).

This is not the behavior we want.

> But the situation looks desperate, This calls for new thinking.  

Well, we could try to convince people to stop using Linux and use
a "real os"  ;)

Seriously, in over 15 years of AFS usage on over 20 OSes/Versions that
I've used, Linux 2.6 is the FIRST to go BACKWARDS from working with
AFS to NOT working with AFS due to blocking these entry points.

Personally I'm quite happy to stay with Linux 2.4 for a while longer
and consider alternatives.

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord@MIT.EDU                        PGP key available