Re: [PATCH 3/6] dma: Add a jz4740 dmaengine driver

[Vinod Koul]

On Fri, May 24, 2013 at 08:41:05AM +0200, Lars-Peter Clausen wrote:
> On 05/24/2013 07:54 AM, Vinod Koul wrote:
> > On Fri, May 24, 2013 at 07:58:04AM +0200, Lars-Peter Clausen wrote:
> >> This one needs both.
> >>
> >>>> +	jzcfg.mode = JZ4740_DMA_MODE_SINGLE;
> >>>> +	jzcfg.request_type = config->slave_id;
> >>>> +
> >>>> +	chan->config = *config;
> >>>> +
> >>>> +	jz4740_dma_configure(chan->jz_chan, &jzcfg);
> >>>> +
> >>>> +	return 0;
> >>> You are NOT use src_addr/dstn_addr? How else are you passing the periphral
> >>> address?
> >> I'm saving the whole config, which will later be used to retrieve the source or
> >> dest address.
> > well I missed that and it is a bad idea. You dont know when client has
> > freed/thrown the pointer so copy this instead..
> 
> I do copy the full config, not just the pointer to the config. Although
> src_addr and dest_addr are the only two fields which are used later on at this
> point. So I could change it to just copy src_addr and dest_addr, or well just
> one of them depending on the direction.
One of them based on direction would be right

> 
> > 
> >>
> >>>> +}
> >> [...]
> >>>> +static int jz4740_dma_alloc_chan_resources(struct dma_chan *c)
> >>>> +{
> >>>> +	struct jz4740_dmaengine_chan *chan = to_jz4740_dma_chan(c);
> >>>> +
> >>>> +	chan->jz_chan = jz4740_dma_request(chan, NULL);
> >>>> +	if (!chan->jz_chan)
> >>>> +		return -EBUSY;
> >>>> +
> >>>> +	jz4740_dma_set_complete_cb(chan->jz_chan, jz4740_dma_complete_cb);
> >>>> +
> >>>> +	return 0;
> >>> Zero is not expected value, you need to return the descriptors allocated
> >>> sucessfully.
> >>
> >> Well, zero descriptors have been allocated. As far as I can see only a negative
> >> return value is treated as an error. Also the core doesn't seem to use the
> >> return value for anything else but checking if it is an error.
> > This is the API defination
> > * @device_alloc_chan_resources: allocate resources and return the
> > *      number of allocated descriptors
> > 
> 
> But 0 is still the number of descriptors that have been pre-allocated.
and that should change, typically the driver will preallocate a pool of
descriptors. These are to be used later for .device_prep_xxx calls.

--
~Vinod
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/