Re: [PATCH v14 09/12] dmaengine: imx-sdma: remove ERR009165 on i.mx6ul

[Lucas Stach <l.stach@xxxxxxxxxxxxxx>]

Am Montag, dem 12.07.2021 um 04:03 +0000 schrieb Robin Gong:
> On 09/07/21 17:45 Lucas Stach <l.stach@xxxxxxxxxxxxxx> wrote:
> > Am Mittwoch, dem 07.04.2021 um 23:30 +0800 schrieb Robin Gong:
> > > ECSPI issue fixed from i.mx6ul at hardware level, no need
> > > ERR009165 anymore on those chips such as i.mx8mq.
> > > 
> > > Signed-off-by: Robin Gong <yibin.gong@xxxxxxx>
> > > Acked-by: Vinod Koul <vkoul@xxxxxxxxxx>
> > > ---
> > >  drivers/dma/imx-sdma.c | 26 +++++++++++++++++++++++++-
> > >  1 file changed, 25 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c index
> > > 86bd383..af85116 100644
> > > --- a/drivers/dma/imx-sdma.c
> > > +++ b/drivers/dma/imx-sdma.c
> > > @@ -422,6 +422,13 @@ struct sdma_driver_data {
> > >  	int num_events;
> > >  	struct sdma_script_start_addrs	*script_addrs;
> > >  	bool check_ratio;
> > > +	/*
> > > +	 * ecspi ERR009165 fixed should be done in sdma script
> > > +	 * and it has been fixed in soc from i.mx6ul.
> > > +	 * please get more information from the below link:
> > > +	 *
> > https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.n
> > xp.com%2Fdocs%2Fen%2Ferrata%2FIMX6DQCE.pdf&amp;data=04%7C01%7C
> > yibin.gong%40nxp.com%7Cc950b1bdb6544eda369408d942be35d9%7C686ea
> > 1d3bc2b4c6fa92cd99c5c301635%7C0%7C1%7C637614206980361737%7CU
> > nknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6
> > Ik1haWwiLCJXVCI6Mn0%3D%7C2000&amp;sdata=6iT6%2FbzJHyWnkhkDynY
> > wmK9nn2hgCDy1GyzELeEk9K8%3D&amp;reserved=0
> > > +	 */
> > > +	bool ecspi_fixed;
> > >  };
> > > 
> > >  struct sdma_engine {
> > > @@ -542,6 +549,13 @@ static struct sdma_driver_data sdma_imx6q = {
> > >  	.script_addrs = &sdma_script_imx6q,
> > >  };
> > > 
> > > +static struct sdma_driver_data sdma_imx6ul = {
> > > +	.chnenbl0 = SDMA_CHNENBL0_IMX35,
> > > +	.num_events = 48,
> > > +	.script_addrs = &sdma_script_imx6q,
> > > +	.ecspi_fixed = true,
> > > +};
> > > +
> > >  static struct sdma_script_start_addrs sdma_script_imx7d = {
> > >  	.ap_2_ap_addr = 644,
> > >  	.uart_2_mcu_addr = 819,
> > > @@ -575,6 +589,7 @@ static const struct of_device_id sdma_dt_ids[] = {
> > >  	{ .compatible = "fsl,imx31-sdma", .data = &sdma_imx31, },
> > >  	{ .compatible = "fsl,imx25-sdma", .data = &sdma_imx25, },
> > >  	{ .compatible = "fsl,imx7d-sdma", .data = &sdma_imx7d, },
> > > +	{ .compatible = "fsl,imx6ul-sdma", .data = &sdma_imx6ul, },
> > >  	{ .compatible = "fsl,imx8mq-sdma", .data = &sdma_imx8mq, },
> > >  	{ /* sentinel */ }
> > >  };
> > > @@ -1144,8 +1159,17 @@ static int sdma_config_channel(struct dma_chan
> > *chan)
> > >  			if (sdmac->peripheral_type == IMX_DMATYPE_ASRC_SP ||
> > >  			    sdmac->peripheral_type == IMX_DMATYPE_ASRC)
> > >  				sdma_set_watermarklevel_for_p2p(sdmac);
> > > -		} else
> > > +		} else {
> > > +			/*
> > > +			 * ERR009165 fixed from i.mx6ul, no errata need,
> > > +			 * set bit31 to let sdma script skip the errata.
> > > +			 */
> > > +			if (sdmac->peripheral_type == IMX_DMATYPE_CSPI &&
> > > +			    sdmac->direction == DMA_MEM_TO_DEV &&
> > > +			    sdmac->sdma->drvdata->ecspi_fixed)
> > > +				__set_bit(31, &sdmac->watermark_level);
> > 
> > Hm, I don't care much either way, but couldn't we just return the regular
> > mcu_2_app script in sdma_get_pc when ecspi_fixed == true? Seems like this
> > would be a simpler and more targeted code change.
> Yes, return mcu_2_app if ecspi_fixed == true also works, but since sdma firmware
> have already been here to fix ERR009165 on most of legacy i.mx6/7/8 chips, so choosing
> firmware/ram script to do like ROM/mcu_2_app is okay too since both ram script and rom
> script in case of ecspi_fixed are almost same.
> 
Actually, while thinking some more about this: it is preferable to
return mcu_2_app in the ecspi_fixed case, as this allows proper DMA
support on the fixed SoCs without loading the firmware. The way you do
it here still requires the RAM firmware to be loaded in order to get
DMA support at all.

Regards,
Lucas

> > 
> > >  			__set_bit(sdmac->event_id0, sdmac->event_mask);
> > > +		}
> > > 
> > >  		/* Address */
> > >  		sdmac->shp_addr = sdmac->per_address;
> > 
>