RE: [PATCH v3 3/3] dmaengine: imx-sdma: allocate max 20 bds for one transfer

From: Robin Gong
Date: Tue Jul 24 2018 - 21:24:54 EST


> -----Original Message-----
> From: Lucas Stach [mailto:l.stach@xxxxxxxxxxxxxx]
> Sent: 2018å7æ24æ 17:22
> To: Robin Gong <yibin.gong@xxxxxxx>; vkoul@xxxxxxxxxx;
> dan.j.williams@xxxxxxxxx; s.hauer@xxxxxxxxxxxxxx; linux@xxxxxxxxxxxxxxx
> Cc: dmaengine@xxxxxxxxxxxxxxx; dl-linux-imx <linux-imx@xxxxxxx>;
> kernel@xxxxxxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH v3 3/3] dmaengine: imx-sdma: allocate max 20 bds for one
> transfer
>
> Am Montag, den 23.07.2018, 13:55 +0000 schrieb Robin Gong:
> > > -----Original Message-----
> > > From: Lucas Stach [mailto:l.stach@xxxxxxxxxxxxxx]
> > > Sent: 2018å7æ23æ 18:54
> > > To: Robin Gong <yibin.gong@xxxxxxx>; vkoul@xxxxxxxxxx;
> > > dan.j.williams@xxxxxxxxx; s.hauer@xxxxxxxxxxxxxx; linux@xxxxxxxxxxx
> > > g.uk
> > > Cc: dmaengine@xxxxxxxxxxxxxxx; dl-linux-imx <linux-imx@xxxxxxx>;
> > > kernel@xxxxxxxxxxxxxx; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx;
> > > linux-kernel@xxxxxxxxxxxxxxx
> > > Subject: Re: [PATCH v3 3/3] dmaengine: imx-sdma: allocate max 20 bds
> > > for one transfer
> > >
> > > Am Dienstag, den 24.07.2018, 01:46 +0800 schrieb Robin Gong:
> > > > If multi-bds used in one transfer, all bds should be consisten
> > > > memory.To easily follow it, enlarge the dma pool size into 20 bds,
> > > > and it will report error if the number of bds is over than 20. For
> > > > dmatest, the max count for single transfer is NUM_BD *
> > >
> > > SDMA_BD_MAX_CNT
> > > > = 20 * 65535 = ~1.28MB.
> > >
> > > Both the commit message and the comment need a lot more care to
> > > actually tell what this commit is trying to achieve. Currently I
> > > don't follow at all. What does "consisten" mean? Do you mean BDs
> > > should be contiguous in memory?
> >
> > Yes, BDs should be contiguousÂÂone by one in memory.
>
> Okay, but this isn't what the code change does. By increasing the size
> parameter of the dma pool you just allocate 20 times as much memory as
> needed for each BD. So actually the BDs end up being very non- contiguous in
> memory as there are now holes of 19 BD sizes between the start of each BD.
Please notice only allocate bds memory from dma pool one time even in multi bds.
That's different with the common use case that allocate memory from dma pool everytime
for every bd. Why do this is to make sure all bd memory is contiguous for single transfer
whatever single bd or multi-bds, since two call dma_pool_alloc() can't promise the address
is contiguous especially for multi thread case such as dmatest 'threads_per_chan = 5'. You
can change to ' norandom=true ' and ' test_buf_size = 163840' in dmatest.c to look what
issue coming without this patch.
>
> So something isn't right with this change.
I think this patch is the easy way to resolve the bd contiguous issue, but the cost is to
allocate more dma pool memory which may not used.
>
> Regards,
> Lucas
>
> > >
> > > What do you gain by over-allocating each BD by a factor of 20?
> >
> > I guess dma_pool_alloc will return error in such case, and then cause
> > dma setup transfer failure.
> > >
> > > Regards,
> > > Lucas
> > >
> > > > Signed-off-by: Robin Gong <yibin.gong@xxxxxxx>
> > > > ---
> > > > Âdrivers/dma/imx-sdma.c | 17 ++++++++++++++++-
> > > > Â1 file changed, 16 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c index
> > > > b4ec2d2..5973489 100644
> > > > --- a/drivers/dma/imx-sdma.c
> > > > +++ b/drivers/dma/imx-sdma.c
> > > > @@ -298,6 +298,15 @@ struct sdma_context_data {
> > > > > Â u32ÂÂscratch7;
> > > >
> > > > Â} __attribute__ ((packed));
> > > >
> > > > +/*
> > > > + * All bds in one transfer should be consitent on SDMA. To
> > > > easily
> > > > +follow it,just
> > > > + * set the dma pool size as the enough bds. For example, in
> > > > dmatest
> > > > +case, the
> > > > + * max 20 bds means the max for single transfer is NUM_BD *
> > > > +SDMA_BD_MAX_CNT = 20
> > > > + * * 65535 = ~1.28MB. 20 bds supposed to be enough basically.If
> > > > it's
> > > > +still not
> > > > + * enough in some specific cases, enlarge it here.Warning
> > > > message
> > > > +would also
> > > > + * appear if the bd numbers is over than 20.
> > > > + */
> > > > +#define NUM_BD 20
> > > >
> > > > Âstruct sdma_engine;
> > > >
> > > > @@ -1273,7 +1282,7 @@ static int sdma_alloc_chan_resources(struct
> > > > dma_chan *chan)
> > > > > Â goto disable_clk_ahb;
> > > > > Â sdmac->bd_pool = dma_pool_create("bd_pool", chan-
> > > > > >device->dev,
> > > > > - sizeof(struct
> > > > > sdma_buffer_descriptor),
> > > > > + NUM_BD * sizeof(struct
> > > > > sdma_buffer_descriptor),
> > > > > Â 32, 0);
> > > > > Â return 0;
> > > >
> > > > @@ -1314,6 +1323,12 @@ static struct sdma_desc
> > > > *sdma_transfer_init(struct sdma_channel *sdmac,
> > > > Â{
> > > > > Â struct sdma_desc *desc;
> > > > > + if (bds > NUM_BD) {
> > > > > + dev_err(sdmac->sdma->dev, "%d bds exceed the
> > > > > max %d\n",
> > > > > + bds, NUM_BD);
> > > > > + goto err_out;
> > > > > + }
> > > >
> > > > +
> > > > > Â desc = kzalloc((sizeof(*desc)), GFP_NOWAIT);
> > > > > Â if (!desc)
> > > > > Â goto err_out;