On Sat, Jun 17, 2017 at 05:01:05PM +0200, Wolfram Sang wrote: > On Fri, Jun 16, 2017 at 11:58:52AM +0200, Simon Horman wrote: > > Add complete to DMA ops to allow DMAC implementation dependent > > handling of DMA completion. > > > > Also implement the operation for SDHI. > > > > Signed-off-by: Simon Horman <[email protected]> > > Codewise, I am all fine with this patch. There is one naming, however: > > > + void (*complete)(struct tmio_mmc_host *host); > > I'd rather call it 'dataend'. For internal DMAC, it is the same. But for > SYS-DMAC there is difference between DATAEND and DMA_COMPLETE. This is > exactly the reason we have commit 52ad9a8e854ca1 ("mmc: tmio: ensure end > of DMA and SD access are in sync"). > > Basically, something like %s/complete/dataend/g would do it for me...
Sure, I will change this. 'compete' seemed like a nice name but I accept your reasoning above.
