[coyotos-dev] Sleep, wakeup, and persistence - wall clock

[Charles Landau]

At 12:26 PM -0400 9/28/07, Jonathan S. Shapiro wrote:
>On Tue, 2007-09-25 at 21:49 +0200, Valerio Bellizzomi wrote:
>>  On 25/09/2007, at 11.29, Charles Landau wrote:
>
>>  >At 2:30 PM +0200 9/15/07, Valerio Bellizzomi wrote:
>>  >>Yes, I agree that [cancelling sleeps on restart] seems to be the
>>  >>only sensible thing to do. If the
>>  >>system had a power loss, probably there is some alarm to service ASAP on
>>  >>restart, and you don't want you application to be sleeping at that time.
>>  >
>>  >I don't necessarily defend that suggestion. In any case, any need for
>>  >interface 2 can be met by two processes, one waiting for a wall clock
>>  >time and one waiting for a restart.
>
>There is something fundamentally wrong with a design that requires 32
>kilobytes worth of state (two processes, their code, their stacks) in
>order to go to sleep.

Again, my knowledge of Coyotos IPC is limited, but perhaps you can 
have just one process waiting on an endpoint, which will be woken up 
by both the wall clock and by the occurrence of restart.