System calls or Per process namespaces?

[Anant Narayanan]

Which do you think will be more useful to us right now:

Implementation of the remaining system calls + #e/#c devices (this
will definitely get rc working)
- or -
Implementation of Plan 9 style per-process namespaces and union mounts
(long term glendix goal)

Feedback is appreciated :-)

--
Anant

[J.R. Mauro]

On Saturday, April 11, 2009, Anant Narayanan <an...@kix.in> wrote:
>
> Which do you think will be more useful to us right now:
>
> Implementation of the remaining system calls + #e/#c devices (this
> will definitely get rc working)

My vote goes here. We can iron out the details later, but for now
making things work would be nice

[Chirag Anand]

>> Implementation of the remaining system calls + #e/#c devices (this
>> will definitely get rc working)

My vote goes here. We can iron out the details later, but for now
making things work would be nice

I agree with Mauro here. Making things work for now would be a better idea. We can extend these to implement per-process namespaces later. As it will also take a lot of time and we need to get many things working before we can actually start with this (and for it's testing etc.).

 

--
Regards

Chirag Anand
4th Year, B.Tech
Computer Science Department, JUIT Solan

Blog :http://techfreaks4u.com/blog
Linux User: 476783

anything weird is worth a try...