Linus Torvalds wrote:
> 
> On Wed, 4 Dec 2002, george anzinger wrote:
> >
> > Once it changes the system call (eax, right), could the new
> > call code then just get the parms from the restart_block.
> 
> Agreed.
> 
> > I think it would be best to keep this as generic as
> > possible, i.e. let the new call code fetch its own
> > paramerers from the restart_block.
> 
> We could even have one _single_ a generic "restart" system call, and have
> the function pointer for that be in the restart block.
> 
> > My question is who sets up these values?  I think you are
> > saying it should be the system call.  Is this right?
> 
> Whatever system call that return -ERESTART_RESTARTBLOCK, yes.
> 
> So it would never get set up at all in the fast path. Only in the error
> case path of a system call that wants to have restarting capabilities.
> 
>                 Linus

Ok, here is a patch to do all this.  This is against
2.5.50-bk4.  (Bk5 did not get posted this evening.)

I think this covers all the bases.  It builds boots and
runs.  I haven't tested nano_sleep to see if it does the
right thing yet...

-- 
George Anzinger   george@mvista.com
High-res-timers: 
http://sourceforge.net/projects/high-res-timers/
Preemption patch:
http://www.kernel.org/pub/linux/kernel/people/rml