RE: [PATCH 3/7] DMA: PL330: Add DMA capabilities

From: boojin <boojin.kim@samsung.com>
To: 'Jassi Brar' <jassisinghbrar@gmail.com>,
	'Chanho Park' <parkch98@gmail.com>
Cc: 'Kukjin Kim' <kgene.kim@samsung.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-samsung-soc@vger.kernel.org,
	'Vinod Koul' <vinod.koul@intel.com>,
	'Dan Williams' <dan.j.williams@intel.com>,
	'Grant Likely' <grant.likely@secretlab.ca>,
	'Liam Girdwood' <lrg@ti.com>,
	'Mark Brown' <broonie@opensource.wolfsonmicro.com>
Subject: RE: [PATCH 3/7] DMA: PL330: Add DMA capabilities
Date: Thu, 14 Jul 2011 09:57:22 +0900	[thread overview]
Message-ID: <002201cc41c0$fdf79810$f9e6c830$%kim@samsung.com> (raw)
In-Reply-To: <CABb+yY0M659Ob0u+wOBW5i6v_VfpayRkY29-TUNpsp1oVLn-SQ@mail.gmail.com>

Jassi Brar wrote:

> On Tue, Jul 5, 2011 at 12:03 PM, Chanho Park <parkch98@gmail.com> wrote:
> > Kukjin Kim <kgene.kim <at> samsung.com> writes:
> >
> > (snip)
> >
> >> +             if (slave_config->direction == DMA_TO_DEVICE) {
> >> +                     if (slave_config->dst_addr)
> >> +                             peri->fifo_addr = slave_config->dst_addr;
> >> +                     if (slave_config->dst_addr_width) {
> >> +                             i = 0;
> >> +                             while (slave_config->dst_addr_width != (1 <<
> > i))
> >> +                                     i++;
> >> +                             peri->burst_sz = i;
> >> +                     }
> >> +             } else if (slave_config->direction == DMA_FROM_DEVICE) {
> >> +                     if (slave_config->src_addr)
> >> +                             peri->fifo_addr = slave_config->src_addr;
> >> +                     if (slave_config->src_addr_width) {
> >> +                             i = 0;
> >> +                             while (slave_config->src_addr_width != (1 <<
> > i))
> >> +                                     i++;
> >> +                             peri->burst_sz = i;
> >
> > Re-send including cc and mailing lists
> > --
> > pl330 dmac only supports 1/2/4/8/16 bytes burst size.
> > If some bad D/D doesn't use powers of 2 width,
> > dmaengine is going to infinite loop.
> > You'd better check it instead of running loop.
> >
> It might be even better to start with max possible width and
> keep decreasing 'i' in the loop.
I will addressed your comment on next release.