Hello. On 04-02-2011 3:34, Michael Williamson wrote:
> For DMA operation, the davinci spi driver needs an eventq number. > Currently, this number is passed as a IORESOURCE_DMA. This is not > correct, as the eventq is not a DMA channel. Pass the eventq > via the platform data structure instead. > Signed-off-by: Michael Williamson<[email protected]> [...] > diff --git a/arch/arm/mach-davinci/include/mach/spi.h > b/arch/arm/mach-davinci/include/mach/spi.h > index 38f4da5..0853a5f 100644 > --- a/arch/arm/mach-davinci/include/mach/spi.h > +++ b/arch/arm/mach-davinci/include/mach/spi.h [...] > @@ -39,13 +41,16 @@ enum { > * to populate if all chip-selects are internal. > * @cshold_bug: set this to true if the SPI controller on your chip > requires > * a write to CSHOLD bit in between transfers (like in DM355). > + * @dma_event_q: DMA event_q to use if SPI_IO_TYPE_DMA is used for any device s/event_q/event queue/? > + * on the bus. > */ > struct davinci_spi_platform_data { > - u8 version; > - u8 num_chipselect; > - u8 intr_line; > - u8 *chip_sel; > - bool cshold_bug; > + u8 version; > + u8 num_chipselect; > + u8 intr_line; > + u8 *chip_sel; > + bool cshold_bug; > + enum dma_event_q dma_event_q; > }; > > /** > diff --git a/drivers/spi/davinci_spi.c b/drivers/spi/davinci_spi.c > index 6beab99..166a879 100644 > --- a/drivers/spi/davinci_spi.c > +++ b/drivers/spi/davinci_spi.c [...] > @@ -878,17 +877,13 @@ static int davinci_spi_probe(struct platform_device > *pdev) > r = platform_get_resource(pdev, IORESOURCE_DMA, 1); > if (r) > dma_tx_chan = r->start; > - r = platform_get_resource(pdev, IORESOURCE_DMA, 2); > - if (r) > - dma_eventq = r->start; > > dspi->bitbang.txrx_bufs = davinci_spi_bufs; > if (dma_rx_chan != SPI_NO_RESOURCE&& > - dma_tx_chan != SPI_NO_RESOURCE&& > - dma_eventq != SPI_NO_RESOURCE) { > + dma_tx_chan != SPI_NO_RESOURCE) { > dspi->dma.rx_channel = dma_rx_chan; > dspi->dma.tx_channel = dma_tx_chan; > - dspi->dma.eventq = dma_eventq; > + dspi->dma.eventq = pdata->dma_event_q; It makes sense to merge this patch with the patches 2 and 3 as otherwise the bisectability is prevented for DM3[56]5 -- their 'dma_event_q' field will be anonymously initialized to 0 after this patch but should be 1 and 3 instead correspondingly. Although, this doesn't seem to be a fatal defect, so the patches can be kept separate... Don't know what's better. WBR, Sergei ------------------------------------------------------------------------------ The modern datacenter depends on network connectivity to access resources and provide services. The best practices for maximizing a physical server's connectivity to a physical network are well understood - see how these rules translate into the virtual world? http://p.sf.net/sfu/oracle-sfdevnlfb _______________________________________________ spi-devel-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/spi-devel-general
