RE: deferred cancellation of ipc

[Volkmar Uhlig]

> Marcus Brinkmann <address@hidden> writes:
> 
> > However, rather than finding a very complicated hack I would prefer
> > to hear what the L4 guys think of it.
> 
> Me two. On the one hand, it's probably easy to add to the kernel
> (unless there's some easy way of solving the problem already, which we
> have missed). And on the other hand, if the problem can be solved by
> two dozen lines of user space assembler glue (for each architecture),
> why bother adding a new feature to the kernel?

Seems everybody is kinda busy in KA ;)

The idea is to use delayed preemptions.  When you toggle the delayed
preemption bit the kernel will not preempt the thread except it exceeds
its grace period.  If that happens the kernel will trigger an exception
and you can have some back off code dealing with that (implementation is
up to you).  Delayed preemptions allow to implement _short_ atomic
operations.  
However, you have to be careful with pagers.  If your code or data
disappears you may run into unwanted preemptions.  IIRC we currently do
not trigger an exception on page fault yet but that is just an
implementation issue.

- Volkmar