On Wed, Aug 05, 2015 at 12:15:05PM +0800, Jun Nie wrote: > 2015-08-05 11:38 GMT+08:00 Vinod Koul <vinod.koul@xxxxxxxxx>: > > On Wed, Aug 05, 2015 at 11:32:00AM +0800, Jun Nie wrote: > >> 2015-08-05 11:24 GMT+08:00 Vinod Koul <vinod.koul@xxxxxxxxx>: > >> > On Tue, Jul 21, 2015 at 11:01:05AM +0800, Jun Nie wrote: > >> >> Align src and dst width to fix data alignment issue. Hardware > >> >> burst length limitation can be addressed well too. > >> > What is the data alignment issue you have? Why not document that in code and > >> > changelog? > > > > ?? > > > When buffer size is not multiple of burst length * data width, src/dst > data width should be identical for these residue data. Will doc it in > commit log and code in next version patch. > > > >> >> > >> >> Signed-off-by: Jun Nie <jun.nie@xxxxxxxxxx> > >> >> --- > >> >> drivers/dma/zx296702_dma.c | 8 ++++---- > >> >> 1 file changed, 4 insertions(+), 4 deletions(-) > >> >> > >> >> diff --git a/drivers/dma/zx296702_dma.c b/drivers/dma/zx296702_dma.c > >> >> index ec470bc..4757f74 100644 > >> >> --- a/drivers/dma/zx296702_dma.c > >> >> +++ b/drivers/dma/zx296702_dma.c > >> >> @@ -143,6 +143,7 @@ static void zx_dma_terminate_chan(struct zx_dma_phy *phy, struct zx_dma_dev *d) > >> >> > >> >> val = readl_relaxed(phy->base + REG_ZX_CTRL); > >> >> val &= ~ZX_CH_ENABLE; > >> >> + val |= ZX_FORCE_CLOSE; > >> > how is this related to data width? > >> > >> Burst length has limitation and destination burst length is calculated > >> from other src/dst parameters. So it is safe to align src/dst data > >> width to avoid to run into limitation. > > > > thats fine but what does it have to do with terminate and ZX_FORCE_CLOSE? > > It fix other bug that's really do not relate to data width. Hardware > will not stop till all transmit ion is done when enable bit is > cleared. Cyclic DMA does not stop without the change as a result of > ring DMA chain. Seems it should be changed in cyclic dma patch. A new > patch should be prepare now, right? Yes please.. I have applied cyclic patch so would prefer updates... -- ~Vinod -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
