| Message ID | 8485d9976f327cf23e051f64c3a3c5cbff4214a7.1458311540.git.moinejf@free.fr (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, Mar 18, 2016 at 03:12:26PM +0100, Jean-Francois Moine wrote: > The H3 accepts 4 as a burst value. > > Signed-off-by: Jean-Francois Moine <moinejf@free.fr> > --- > drivers/dma/sun6i-dma.c | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/drivers/dma/sun6i-dma.c b/drivers/dma/sun6i-dma.c > index ce1e4d6..9378fda 100644 > --- a/drivers/dma/sun6i-dma.c > +++ b/drivers/dma/sun6i-dma.c > @@ -101,6 +101,7 @@ struct sun6i_dma_config { > u32 nr_max_channels; > u32 nr_max_requests; > u32 nr_max_vchans; > + u32 burst_4; > }; > > /* > @@ -238,6 +239,8 @@ static inline s8 convert_burst(u32 maxburst) > switch (maxburst) { > case 1: > return 0; > + case 4: > + return 1; > case 8: > return 2; > default: > @@ -470,6 +473,10 @@ static int set_config(struct sun6i_dma_dev *sdev, > { > s8 src_width, dst_width, src_burst, dst_burst; > > + if (!sdev->cfg->burst_4 && > + (sconfig->src_maxburst == 4 || sconfig->dst_maxburst == 4)) > + return -EINVAL; > + I still believe this should be dealt with at the framework level. Maxime
On Mon, 21 Mar 2016 08:29:18 +0100 Maxime Ripard <maxime.ripard@free-electrons.com> wrote: > > @@ -470,6 +473,10 @@ static int set_config(struct sun6i_dma_dev *sdev, > > { > > s8 src_width, dst_width, src_burst, dst_burst; > > > > + if (!sdev->cfg->burst_4 && > > + (sconfig->src_maxburst == 4 || sconfig->dst_maxburst == 4)) > > + return -EINVAL; > > + > > I still believe this should be dealt with at the framework level. What do you mean? In the DMA clients?
On Mon, Mar 21, 2016 at 09:35:33AM +0100, Jean-Francois Moine wrote: > On Mon, 21 Mar 2016 08:29:18 +0100 > Maxime Ripard <maxime.ripard@free-electrons.com> wrote: > > > > @@ -470,6 +473,10 @@ static int set_config(struct sun6i_dma_dev *sdev, > > > { > > > s8 src_width, dst_width, src_burst, dst_burst; > > > > > > + if (!sdev->cfg->burst_4 && > > > + (sconfig->src_maxburst == 4 || sconfig->dst_maxburst == 4)) > > > + return -EINVAL; > > > + > > > > I still believe this should be dealt with at the framework level. > > What do you mean? In the DMA clients? No, in the dmaengine framework. A driver should register the burst size they support, and dmaengine_slave_config would reject any invalid burst size based on that. Maxime
On Tue, 29 Mar 2016 08:46:50 +0200 Maxime Ripard <maxime.ripard@free-electrons.com> wrote: > On Mon, Mar 21, 2016 at 09:35:33AM +0100, Jean-Francois Moine wrote: > > On Mon, 21 Mar 2016 08:29:18 +0100 > > Maxime Ripard <maxime.ripard@free-electrons.com> wrote: > > > > > > @@ -470,6 +473,10 @@ static int set_config(struct sun6i_dma_dev *sdev, > > > > { > > > > s8 src_width, dst_width, src_burst, dst_burst; > > > > > > > > + if (!sdev->cfg->burst_4 && > > > > + (sconfig->src_maxburst == 4 || sconfig->dst_maxburst == 4)) > > > > + return -EINVAL; > > > > + > > > > > > I still believe this should be dealt with at the framework level. > > > > What do you mean? In the DMA clients? > > No, in the dmaengine framework. > > A driver should register the burst size they support, and > dmaengine_slave_config would reject any invalid burst size based on > that. Maybe you are right, but this does not seem to be very easy: the max burst size is on a u32 value, and, as a bit map (256Mb) is forbidden, the DMA drivers should declare a list of (min, max) authorized values. Well, I will not start such a development for just testing if 4 is a good burst size when calling the sun6i DMA driver...
diff --git a/drivers/dma/sun6i-dma.c b/drivers/dma/sun6i-dma.c index ce1e4d6..9378fda 100644 --- a/drivers/dma/sun6i-dma.c +++ b/drivers/dma/sun6i-dma.c @@ -101,6 +101,7 @@ struct sun6i_dma_config { u32 nr_max_channels; u32 nr_max_requests; u32 nr_max_vchans; + u32 burst_4; }; /* @@ -238,6 +239,8 @@ static inline s8 convert_burst(u32 maxburst) switch (maxburst) { case 1: return 0; + case 4: + return 1; case 8: return 2; default: @@ -470,6 +473,10 @@ static int set_config(struct sun6i_dma_dev *sdev, { s8 src_width, dst_width, src_burst, dst_burst; + if (!sdev->cfg->burst_4 && + (sconfig->src_maxburst == 4 || sconfig->dst_maxburst == 4)) + return -EINVAL; + if (direction == DMA_MEM_TO_DEV) { src_burst = convert_burst(sconfig->src_maxburst ? sconfig->src_maxburst : 8); @@ -900,12 +907,14 @@ static struct sun6i_dma_config sun8i_a23_dma_cfg = { /* * The H3 has 12 physical channels, a maximum DRQ port id of 27, * and a total of 34 usable source and destination endpoints. + * Setting the maxburst to '4' is possible. */ static struct sun6i_dma_config sun8i_h3_dma_cfg = { .nr_max_channels = 12, .nr_max_requests = 27, .nr_max_vchans = 34, + .burst_4 = 1, }; static const struct of_device_id sun6i_dma_match[] = {
The H3 accepts 4 as a burst value. Signed-off-by: Jean-Francois Moine <moinejf@free.fr> --- drivers/dma/sun6i-dma.c | 9 +++++++++ 1 file changed, 9 insertions(+)