Re: [PATCH][RFC] AIO: always reinitialize iocb->ki_run_list at the end of aio_run_iocb()

[Andrew Morton <akpm@linux-foundation.org>]

On Wed, 28 Apr 2010 02:51:43 +0400
Sergey Temerkhanov <temerkhanov@yandex.ru> wrote:

> This patch makes aio_run_iocb() to always reinitialize iocb->ki_run_list (not 
> only when iocb->ki_retry() function returns -EIOCBRETRY) so that subsequent 
> call of kick_iocb() will succeed.
> 
> Regards, Sergey Temerkhanov,
> Cifronic ZAO.
> 
> 
> [reinit-ki_run_list.patch  text/x-patch (657B)]
> diff -r 97344a0f62c9 fs/aio.c
> --- a/fs/aio.c	Tue Apr 27 21:18:14 2010 +0400
> +++ b/fs/aio.c	Tue Apr 27 21:30:23 2010 +0400
> @@ -748,6 +748,9 @@
>  out:
>  	spin_lock_irq(&ctx->ctx_lock);
>  
> +	/* will make __queue_kicked_iocb succeed from here on */
> +	INIT_LIST_HEAD(&iocb->ki_run_list);
> +
>  	if (-EIOCBRETRY == ret) {
>  		/*
>  		 * OK, now that we are done with this iteration
> @@ -756,8 +759,6 @@
>  		 * "kick" can start the next iteration
>  		 */
>  
> -		/* will make __queue_kicked_iocb succeed from here on */
> -		INIT_LIST_HEAD(&iocb->ki_run_list);
>  		/* we must queue the next iteration ourselves, if it
>  		 * has already been kicked */
>  		if (kiocbIsKicked(iocb)) {

I assume that this fixes some runtime problem which you observed?

Can you please describe that problem?  This code is pretty old - what
was your application doing that nobody else's application has thus far
done?

Also, please send your Signed-off-by: for this patch, as per
Documentation/Submittingpatches, thanks.