On Friday 25 July 2014 12:55:59 Mika Westerberg wrote: > On Fri, Jul 25, 2014 at 12:07:14PM +0300, Mika Westerberg wrote: > > On Fri, Jul 25, 2014 at 10:38:36AM +0200, Arnd Bergmann wrote: > > > On Friday 25 July 2014 11:22:49 Mika Westerberg wrote: > > > > > All you need to do is change your filter function to take the > > > > > slave id from pxa_spi_info and stick it in there, e.g. > > > > > > > > > > static bool pxa2xx_spi_dw_dma_filter(struct dma_chan *chan, void *param) > > > > > { > > > > > const struct pxa2xx_spi_master *pdata = param; > > > > > struct dw_dma_chan *dwc = to_dw_dma_chan(chan); > > > > > > > > > > dwc->request_line = fargs->req; > > > > > dwc->src_master = 0; > > > > > dwc->dst_master = 0; > > > > > > > > > > return 1; > > > > > } > > > > > > > > Oh man. That makes pxa2xx_spi dependent on a certain specific DMA engine > > > > driver. > > > > > > I think you can improve this by putting the filter function (and a pointer > > > to it) into the pxa2xx_spi_master data provided by the PCI driver. > > > > Indeed, that looks better. It still makes the PCI part of the driver > > dependent on a particular DMA engine driver but is certainly better than > > the core pxa2xx_spi driver. > > Something like this? > > Hock Leong / Chiaue Ee, are you able to check if this works on your BYT > machines? I think I found a small bug, and one detail that could be improved. > @@ -321,12 +310,14 @@ int pxa2xx_spi_dma_setup(struct driver_data *drv_data) > return -ENOMEM; > > drv_data->tx_chan = dma_request_slave_channel_compat(mask, > - pxa2xx_spi_dma_filter, pdata, dev, "tx"); > + pdata->dma_filter, pdata->dma_filter_param, > + dev, "tx"); > if (!drv_data->tx_chan) > return -ENODEV; If you change this part to pass '(void *)info->tx_slave_id' rather than pdata->dma_filter_param, you can simplify the next code: > +static bool pxa2xx_spi_pci_dma_filter(struct dma_chan *chan, void *param) > +{ > + struct dw_dma_chan *dwc = to_dw_dma_chan(chan); > + const struct pxa_spi_info *info = param; > + > + if (chan->chan_id == info->tx_chan_id) > + dwc->request_line = info->tx_slave_id; > + else if (chan->chan_id == info->rx_chan_id) > + dwc->request_line = info->rx_slave_id; > + else > + return false; > + > + dwc->src_master = 0; > + dwc->dst_master = 0; > + > + return true; > +} to disregard the tx_chan_id, which is really arbitrary as far as I can tell. What I think you got wrong here (by following my bad advice) is the master number. Looking at the code for dw_dma, I think src_master needs to be '1' for your driver. If we do the same change in the drivers/tty/serial/8250/8250_pci.c driver, the dw_dma driver can be simplified a little by removing the special case for the request line setting. diff --git a/drivers/dma/dw/core.c b/drivers/dma/dw/core.c index 1af731b83b3f..674f69fe6662 100644 --- a/drivers/dma/dw/core.c +++ b/drivers/dma/dw/core.c @@ -37,24 +37,6 @@ * support descriptor writeback. */ -static inline bool is_request_line_unset(struct dw_dma_chan *dwc) -{ - return dwc->request_line == (typeof(dwc->request_line))~0; -} - -static inline void dwc_set_masters(struct dw_dma_chan *dwc) -{ - struct dw_dma *dw = to_dw_dma(dwc->chan.device); - struct dw_dma_slave *dws = dwc->chan.private; - unsigned char mmax = dw->nr_masters - 1; - - if (!is_request_line_unset(dwc)) - return; - - dwc->src_master = min_t(unsigned char, mmax, dwc_get_sms(dws)); - dwc->dst_master = min_t(unsigned char, mmax, dwc_get_dms(dws)); -} - #define DWC_DEFAULT_CTLLO(_chan) ({ \ struct dw_dma_chan *_dwc = to_dw_dma_chan(_chan); \ struct dma_slave_config *_sconfig = &_dwc->dma_sconfig; \ @@ -967,10 +949,6 @@ set_runtime_config(struct dma_chan *chan, struct dma_slave_config *sconfig) memcpy(&dwc->dma_sconfig, sconfig, sizeof(*sconfig)); dwc->direction = sconfig->direction; - /* Take the request line from slave_id member */ - if (is_request_line_unset(dwc)) - dwc->request_line = sconfig->slave_id; - convert_burst(&dwc->dma_sconfig.src_maxburst); convert_burst(&dwc->dma_sconfig.dst_maxburst); @@ -1123,8 +1101,6 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan) * doesn't mean what you think it means), and status writeback. */ - dwc_set_masters(dwc); - spin_lock_irqsave(&dwc->lock, flags); i = dwc->descs_allocated; while (dwc->descs_allocated < NR_DESCS_PER_CHANNEL) { Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-spi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html