Re: [PATCH] media: renesas: vsp1: blacklist r8a7795 ES1.*

[Laurent Pinchart]

Hi Wolfram,

Thank you for the patch.

On Wed, Jan 18, 2023 at 01:20:02PM +0100, Wolfram Sang wrote:
> The earliest revision of these SoC may hang when underrunning. Later
> revisions have that fixed. Bail out when we detect a problematic
> version.
> 
> Signed-off-by: Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx>
> ---
> 
> The BSP tries to work around the issue, yet this is neither upstreamable
> nor are we sure the solution is complete. Because the early SoC revision
> is hardly in use, we simply "document" the problem upstream.

The workaround isn't upstreamable as-is, but I think it could be
upstreamed after being cleaned up.

Overall, how much support do we still have upstream for H3 ES1.x, and do
we need to keep it ? H3 ES.1x is relatively old, does someone still rely
on it ?

>  drivers/media/platform/renesas/vsp1/vsp1_drv.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/drivers/media/platform/renesas/vsp1/vsp1_drv.c b/drivers/media/platform/renesas/vsp1/vsp1_drv.c
> index c260d318d298..b395e8eb0af9 100644
> --- a/drivers/media/platform/renesas/vsp1/vsp1_drv.c
> +++ b/drivers/media/platform/renesas/vsp1/vsp1_drv.c
> @@ -17,6 +17,7 @@
>  #include <linux/platform_device.h>
>  #include <linux/pm_runtime.h>
>  #include <linux/reset.h>
> +#include <linux/sys_soc.h>
>  #include <linux/videodev2.h>
>  
>  #include <media/rcar-fcp.h>
> @@ -875,13 +876,24 @@ static const struct vsp1_device_info *vsp1_lookup_info(struct vsp1_device *vsp1)
>  	return NULL;
>  }
>  
> +static const struct soc_device_attribute vsp1_blacklist[]  = {
> +	/* H3 ES1.* has underrun hang issues */
> +	{ .soc_id = "r8a7795", .revision = "ES1.*" },
> +	{ /* Sentinel */ }
> +};
> +
>  static int vsp1_probe(struct platform_device *pdev)
>  {
> +	const struct soc_device_attribute *attr;
>  	struct vsp1_device *vsp1;
>  	struct device_node *fcp_node;
>  	int ret;
>  	int irq;
>  
> +	attr = soc_device_match(vsp1_blacklist);
> +	if (attr)
> +		return -ENOTSUPP;
> +
>  	vsp1 = devm_kzalloc(&pdev->dev, sizeof(*vsp1), GFP_KERNEL);
>  	if (vsp1 == NULL)
>  		return -ENOMEM;

-- 
Regards,

Laurent Pinchart