Re: [PATCH]mdc800: proper memory barriers

On Fri, Mar 30, 2007 at 12:30:11PM +0200, Oliver Neukum wrote:
> Hi,
> 
> some drivers were indeed written before the days of out of order
> execution and it shows. This adds the barriers that make sure that
> the flags are set in the proper order.
> 
> 	Regards
> 		Oliver
> Signed-off-by: Oliver Neukum <[email protected]>
> -- 
> --- a/drivers/usb/image/mdc800.c	2007-03-30 12:17:06.000000000 +0200
> +++ b/drivers/usb/image/mdc800.c	2007-03-30 12:17:11.000000000 +0200
> @@ -280,7 +280,7 @@
>  /*
>   * USB IRQ Handler for InputLine
>   */
> -static void mdc800_usb_irq (struct urb *urb)
> +static void mdc800_usb_irq(struct urb *urb)
>  {
>  	int data_received=0, wake_up;
>  	unsigned char* b=urb->transfer_buffer;
> @@ -330,7 +330,9 @@
>  
>  	if (wake_up)
>  	{
> +		smp_mb();
>  		mdc800->camera_request_ready=0;
> +		smp_wmb();

Ick, is there any other way to achieve this, other than useing smp_mb()?
A lock perhaps?

I _really_ don't want to see that macro/function be used in drivers,
it shows something is wrong with the logic if it is needed.

thanks,

greg k-h

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel