Re: RING_HAS_UNCONSUMED_REQUESTS oddness

[Ian Campbell <Ian.Campbell@citrix.com>]

On Thu, 2014-03-06 at 15:47 +0000, Zoltan Kiss wrote:
> Hi,
> 
> I was wondering for a while why this macro looks like this:
> 
> #define RING_HAS_UNCONSUMED_REQUESTS(_r)				\
>     ({									\
> 	unsigned int req = (_r)->sring->req_prod - (_r)->req_cons;	\
> 	unsigned int rsp = RING_SIZE(_r) -				\
> 			   ((_r)->req_cons - (_r)->rsp_prod_pvt);	\
> 	req < rsp ? req : rsp;						\
>     })
> 
> I would expect to check prod - cons, like RING_HAS_UNCONSUMED_RESPONSES does:
> 
> #define RING_HAS_UNCONSUMED_RESPONSES(_r)				\
>     ((_r)->sring->rsp_prod - (_r)->rsp_cons)
> 
> By my understanding, there is no way rsp could be smaller than req, so
> there is no point having this. Am I missing something?

Just looking at this again. All that stuff I said about wraparound was
misleading/irrelevant since req and rsp are not the pointers, but
actually the number of things. Sorry.

Is it the case that this macro considers a request to be unconsumed if
the *response* to a request is outstanding as well as if the request
itself is still on the ring?

I wonder if this apparently weird construction is due to pathological
cases when one or the other end is not picking up requests/responses?
i.e. trying to avoid deadlocking the ring or generating an interrupt
storm when the ring it is full of one or the other or something along
those lines?

Ian.