Re: semaphore understanding: sys_semtimedop()

[Manfred Spraul <manfred@colorfullife.com>]

randy_dunlap wrote:

> On Mon, 16 May 2005 22:17:04 +0200 Patrick Plattes wrote:
>
> | The variable decrease will never be used again in this 
> | function, so why this intricate code? Isn't this much easier and works
> | also:
> | 
> |         for (sop = sops; sop < sops + nsops; sop++) {
> |                 if (sop->sem_num >= max)
> |                         max = sop->sem_num;
> |                 if (sop->sem_flg & SEM_UNDO)
> |                         undos++;
> |                 if (sop->sem_op != 0)
> |                         alter = 1;
> |         }
> | 
> | Maybe i'm totally wrong, so please correct me and don't shoot me up,
> | 'cause i'm not a os developer.
>
> Looks like a reasonable and correct optimization to me.
>
>  
It looks correct:
decrease was added during 2.1 development: it's not in 2.0.40, it's in 
2.2.26: I think the idea was that operations with decrease can cause 
other threads to block, therefore a different wakeup strategy was used. 
The wakeup implementation was rewritten, but the loop remained unchanged.
I'll write a patch, thanks Patrick.

--
   Manfred
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/