Re: [lwip-users] SemSignal gets called on a already signaled semaphore

[Jonathan Larmour]

Goldschmidt Simon wrote:
> > > So far I have used Win32, µC/OS-II and embOS and I haven't used 
> > > condidtion variables before (or I don't get your examples...).
> > > Can you give an example under thos systems or linux?
> > pthread_cond_wait on linux/POSIX, and SignalObjectAndWait and 
> > WaitForSingleObject on Win32 (taking mutexes as the object). 
>
> But WaitForSingleObject() is like mutex_lock() (with a timeout), isn't it?
> Where's the condition variable with that? Or did you mean to implement
> counting semaphores using mutexes???

Sorry, I wasn't clear - SignalObjectAndWait is the more important and 
interesting one, and it takes an Event. Signalling the condition variable 
is SetEvent.

Anyway, this is getting rather off-topic. And I think with the approach 
we're now discussing in bug #200021, we won't need to worry about specific 
semaphore implementations any more anyway.

Jifl
--
eCosCentric Limited      http://www.eCosCentric.com/     The eCos experts
Barnwell House, Barnwell Drive, Cambridge, UK.       Tel: +44 1223 245571
Registered in England and Wales: Reg No 4422071.
------["The best things in life aren't things."]------      Opinions==mine