Re: [Patch] Do not crash upon remote-remote migration

On Sat, Apr 01, 2006 at 12:49:45PM +0200, Vincent Hanquez wrote:
>
> NACK.
> 
> I think that's better to implement the R2R feature than end up putting
> workaround at development stage. It's not like that's the only feature
> missing anyway, and some stuff are more critical ...
> 
> The crash should be fix by the protection I put in task_local_send,
> and there should be something like "task_local_send_move_request" called
> instead of task_local_send when the process is DDEPUTY.
> 
> it's just ending up doing nothing for the moment.

After thinking about it for a while, I tend to agree. 
Doing a remote->home->remote' in kernel space isn't any faster than
doing the same thing in user space, so as long as it really does
nothing (as opposed to crashing), I'm o.k. with waiting for a working
remote-remote migration.
I have dropped the patch from my local tree as well.


A.

-- 
Ansgar Esztermann
Researcher & Sysadmin
http://www2.thphy.uni-duesseldorf.de/~ansgar

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQFEMR3AwAkt8h1/xRgRAsVuAKCNhkOw9SYk9fXq7CpaB+s84ZigXgCfddqX
D783vD+tGN7IVAbEH2b4OQM=
=qZ1X
-----END PGP SIGNATURE-----