Re: [PATCH] ipc,sem: move restart loop to do_smart_update

[Manfred Spraul]

Hi Rik,

On 05/20/2013 12:32 AM, Rik van Riel wrote:
> @@ -751,17 +744,19 @@ static int update_queue(struct sem_array *sma, int semnum, struct list_head *pt)
>   static void do_smart_update(struct sem_array *sma, struct sembuf *sops, int nsops,
>   			int otime, struct list_head *pt)
>   {
> -	int i;
> +	int i, restart;
>   
> + again:
> +	restart = 0;
>   	if (sma->complex_count || sops == NULL) {
> -		if (update_queue(sma, -1, pt))
> +		if (update_queue(sma, -1, pt, &restart))
>   			otime = 1;
>   	}
I didn't notice it immediately, this is insufficient:
Complex operations may affect all semaphores in the array.
Thus it is not sufficient to scan the per-semaphore queues of the 
semaphores that were touched by *sops, all must be scanned.

Perhaps something like:
> +        if (update_queue(sma, -1, pt, &restart))
> +            sops = NULL;
>              otime = 1;
>      }
(untested!)

Test case:
1: semop(<key>, {{1, -1, 0}, {2, 1, 0}}, 2 <unfinished ...>
2: semop(<key>, {{2, -1, 0}}, 1 <unfinished ...>
3: semop(<key>, {{1, 1, 0}}, 1)     = 0

[3] is able to run, calls do_smart_update().
The global queue is scanned and [1] is released
Without the "sops = NULL", [2] sleeps forever.

--
    Manfred

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