Re: race in usbnet.c in full RT

["Eugeny S. Mints" <emints@ru.mvista.com>]

Steven Rostedt wrote:
> On Wed, 2005-06-08 at 12:34 +0200, Ingo Molnar wrote:
> 
>>* Eugeny S. Mints <emints@ru.mvista.com> wrote:
>>
>>
>>>seems there is a race in drivers/net/usbnet.c in full RT mode. To be 
>>>honest I haven't hardly checked this on the latest kernel and latest 
>>>RT patch but just took a look at usbnet.c and latest RT patch and 
>>>haven't observed any related changes.
>>
>>thanks, i've applied your patch to my tree. Note that your patch is 
>>specific to the -RT kernel (both in terms of semantics and in term of 
>>API dependence), so it does not make any sense to apply it upstream.  
>>David, please ignore it.
>>
> 
> 
> Is this action only take place on the same CPU, or is this also an SMP
> problem?  I would think if this is a race with full RT, that this may
> also be a race with SMP, unless the race is guaranteed to always happen
> on the same CPU. Then this is only a RT problem.
thanks, good point. looks like it could be SMP problem but probably 
David is able to say it for sure as usb host code expert.
David?
	Eugeny
> 
> -- Steve
> 
> 
> 
>