On Wed, Sep 10, 2003 at 03:20:31AM +0100, John Levon wrote:
> 
> I'm probably labouring under some fundamental misunderstanding here,
> but I'll plow on regardless.
> 
>      83 /* needs to be called with BKL held */
>      84 int usb_probe_interface(struct device *dev)
> 
> is exposed to the driver model. Whilst I've not yet dis-entangled the
> call stack that gets here, I can't see where the BKL gets taken for the
> store_new_id() case. What is the magical BKL protecting ?
> 
> Is this a problem ?

Hm, I thought we didn't need to do that anymore.  And we don't hold the
BKL during that call too...

I wonder if Oliver remembers...

thanks,

greg k-h


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to