Re: [patch - regression]unkill cxacru atm driver

On Thu, 16 Aug 2007, Oliver Neukum wrote:

> Hi Greg,
> 
> it seems like you overdid it a bit in your quest to clean up the
> use of urb->status. In this driver you read it the first thing, which
> means that you are in a race against URB completion you'll
> usually lose, returning -EINPROGRESS. This kills the driver.
> 
> 	Regards
> 		Oliver
> 
> Signed-off-by: Oliver Neukum <[email protected]>
> ---
> 
> --- a/drivers/usb/atm/cxacru.c	2007-08-14 14:58:14.000000000 +0200
> +++ b/drivers/usb/atm/cxacru.c	2007-08-16 15:53:55.000000000 +0200
> @@ -456,7 +456,6 @@ static int cxacru_start_wait_urb(struct 
>  				 int* actual_length)
>  {
>  	struct timer_list timer;
> -	int status = urb->status;
>  
>  	init_timer(&timer);
>  	timer.expires = jiffies + msecs_to_jiffies(CMD_TIMEOUT);
> @@ -468,7 +467,7 @@ static int cxacru_start_wait_urb(struct 
>  
>  	if (actual_length)
>  		*actual_length = urb->actual_length;
> -	return status;
> +	return urb->status; /* must read status after completion */
>  }

This should be changed to resemble the usb_start_wait_urb routine in 
messages.c.  The idea is that urb->status is available _only_ in the 
completion handler, nowhere else.  Greg's mistake was to think that the 
start_wait_urb routines were completion handlers... but they aren't.

Alan Stern


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel