[coyotos-dev] Need for revised invocation

Sandro Magi smagi at naasking.homeip.net
Sun Feb 5 22:02:35 EST 2006


Charles Landau wrote:
> At 8:47 PM -0500 2/2/06, Jonathan S. Shapiro wrote:
> 
>> There is another set of patterns that motivates a new look at the
>> invocation mechanism: POSIX.
>>
>> ...
>>
>> This cannot be done in KeyKOS/EROS without a very large number of kernel
>> threads
> 
> 
> These new objects could replace these helper threads also.
> 
> None of the above requires preemption. Preemption is evil and should be 
> avoided at almost any cost.

The discussions and proposals thus far have been very interesting, but I 
haven't seen any commentary on why preemption is considered inherently evil.

Asynchronous IPC used to be "evil" primarily because the kernel was 
vulnerable to DoS and the buffering carried hidden performance costs.

This asynchronous design does not suffer from these design flaws and it 
appears to be more flexible than purely synchronous IPC designs given 
current hardware constraints.

So what are the disadvantages of this design, asynchrony or preemption 
in general? Is it because this preemption may necessitate more locking 
within an object to ensure that the activation handler does not break 
something?

Sandro


More information about the coyotos-dev mailing list