| Message ID | 1437998010-12036-1-git-send-email-andriy.shevchenko@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Monday 27 July 2015 14:53:30 you wrote: > The commit ecb89f2f5f3e (mmc: atmel-mci: remove compat for non DT board when > requesting dma chan) removes compat transfer which breaks DMA > support for ATNGW100. This patch returns back that functionality. > > Cc: Ludovic Desroches <ludovic.desroches@atmel.com> > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > How about moving that filter function into arch/avr32 to make the mmc driver independent of the underlying dma engine implementation at the same time? Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, 2015-07-27 at 14:24 +0200, Arnd Bergmann wrote: > On Monday 27 July 2015 14:53:30 you wrote: > > The commit ecb89f2f5f3e (mmc: atmel-mci: remove compat for non DT > > board when > > requesting dma chan) removes compat transfer which breaks DMA > > support for ATNGW100. This patch returns back that functionality. > > > > Cc: Ludovic Desroches <ludovic.desroches@atmel.com> > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > > > > How about moving that filter function into arch/avr32 to make the > mmc driver independent of the underlying dma engine implementation > at the same time? It might be a good idea, but I'm neither an expert in avr32 nor have time right now to do this better. Consider this a bug report with a work around which works for me. I can test any better solution someone proposes. > > Arnd
Hi Andy, Your patch does a bit more than reverting commit ecb89f2f5f3e. I will split it in order to see well which previous patches have been reverted. On Mon, Jul 27, 2015 at 05:45:34PM +0300, Andy Shevchenko wrote: > On Mon, 2015-07-27 at 14:24 +0200, Arnd Bergmann wrote: > > On Monday 27 July 2015 14:53:30 you wrote: > > > The commit ecb89f2f5f3e (mmc: atmel-mci: remove compat for non DT > > > board when > > > requesting dma chan) removes compat transfer which breaks DMA > > > support for ATNGW100. This patch returns back that functionality. > > > > > > Cc: Ludovic Desroches <ludovic.desroches@atmel.com> > > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > > > > > > > How about moving that filter function into arch/avr32 to make the > > mmc driver independent of the underlying dma engine implementation > > at the same time? It would be great to not keep code which becomes architecture dependant. If we can move the filter function somewhere in arch/avr32, we will have to revert commit fda1b26 and use dma_request_slave_channel_compat. I had a look to the arch/avr32 directory but I don't know where to put this kind of stuff. > > It might be a good idea, but I'm neither an expert in avr32 nor have > time right now to do this better. Consider this a bug report with a > work around which works for me. I can test any better solution someone > proposes. Regards Ludovic -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thursday 06 August 2015 09:01:56 Ludovic Desroches wrote: > Hi Andy, > > Your patch does a bit more than reverting commit ecb89f2f5f3e. I will > split it in order to see well which previous patches have been reverted. > > On Mon, Jul 27, 2015 at 05:45:34PM +0300, Andy Shevchenko wrote: > > On Mon, 2015-07-27 at 14:24 +0200, Arnd Bergmann wrote: > > > On Monday 27 July 2015 14:53:30 you wrote: > > > > The commit ecb89f2f5f3e (mmc: atmel-mci: remove compat for non DT > > > > board when > > > > requesting dma chan) removes compat transfer which breaks DMA > > > > support for ATNGW100. This patch returns back that functionality. > > > > > > > > Cc: Ludovic Desroches <ludovic.desroches@atmel.com> > > > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > > > > > > > > > > How about moving that filter function into arch/avr32 to make the > > > mmc driver independent of the underlying dma engine implementation > > > at the same time? > > It would be great to not keep code which becomes architecture dependant. > If we can move the filter function somewhere in arch/avr32, we will have > to revert commit fda1b26 and use dma_request_slave_channel_compat. > > I had a look to the arch/avr32 directory but I don't know where to put > this kind of stuff. Next to the at32_add_device_mci() function, with a pointer to the filter function added to struct mci_platform_data, and the mci_dma_data structure replaced with a void pointer after moving the definition to the same place. Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c index 9a39e0b..70865be 100644 --- a/drivers/mmc/host/atmel-mci.c +++ b/drivers/mmc/host/atmel-mci.c @@ -2277,12 +2277,55 @@ static void atmci_cleanup_slot(struct atmel_mci_slot *slot, mmc_free_host(slot->mmc); } -static int atmci_configure_dma(struct atmel_mci *host) +static bool atmci_filter(struct dma_chan *chan, void *pdata) { + struct mci_platform_data *sl_pdata = pdata; + struct mci_dma_data *sl; + + if (!sl_pdata) + return false; + + sl = sl_pdata->dma_slave; + if (sl && find_slave_dev(sl) == chan->device->dev) { + chan->private = slave_data_ptr(sl); + return true; + } else { + return false; + } +} + +static int atmci_dma_request_channel(struct atmel_mci *host) +{ + struct mci_platform_data *pdata; + dma_cap_mask_t mask; + host->dma.chan = dma_request_slave_channel_reason(&host->pdev->dev, "rxtx"); - if (IS_ERR(host->dma.chan)) - return PTR_ERR(host->dma.chan); + if (!IS_ERR(host->dma.chan)) + return 0; + + if (PTR_ERR(host->dma.chan) == -EPROBE_DEFER) + return -EPROBE_DEFER; + + pdata = host->pdev->dev.platform_data; + + dma_cap_zero(mask); + dma_cap_set(DMA_SLAVE, mask); + + host->dma.chan = dma_request_channel(mask, atmci_filter, pdata); + if (!host->dma.chan) + return -ENODEV; + + return 0; +} + +static int atmci_configure_dma(struct atmel_mci *host) +{ + int ret; + + ret = atmci_dma_request_channel(host); + if (ret) + return ret; dev_info(&host->pdev->dev, "using %s for DMA transfers\n", dma_chan_name(host->dma.chan));
The commit ecb89f2f5f3e (mmc: atmel-mci: remove compat for non DT board when requesting dma chan) removes compat transfer which breaks DMA support for ATNGW100. This patch returns back that functionality. Cc: Ludovic Desroches <ludovic.desroches@atmel.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> --- drivers/mmc/host/atmel-mci.c | 49 +++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 46 insertions(+), 3 deletions(-)