[PATCH v2] iommu/rockchip: make clock handling optional

[PATCH v2] iommu/rockchip: make clock handling optional

[Heiko Stuebner]

iommu clocks are optional, so the driver should not fail if they are not
present. Instead just set the number of clocks to 0, which the clk-blk APIs
can handle just fine.

Fixes: f2e3a5f557ad ("iommu/rockchip: Control clocks needed to access the IOMMU")
Signed-off-by: Heiko Stuebner <heiko-4mtYJXux2i+zQB+pC5nmwQ@public.gmane.org>
Reviewed-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
---
changes in v2:
- keep it required in the binding for future variants
  but optional in the code for backwards compatiblity
- follow Robin Murphy's beautification suggestion

As the original changes breaks a lot of display support in 4.17-rc1,
this should abviously go in as fix into 4.17.

 drivers/iommu/rockchip-iommu.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c
index 5fc8656c60f9..422d22990289 100644
--- a/drivers/iommu/rockchip-iommu.c
+++ b/drivers/iommu/rockchip-iommu.c
@@ -1175,8 +1175,15 @@ static int rk_iommu_probe(struct platform_device *pdev)
 	for (i = 0; i < iommu->num_clocks; ++i)
 		iommu->clocks[i].id = rk_iommu_clocks[i];
 
+	/*
+	 * iommu clocks should be present for all new devices and devicetrees
+	 * but there are older devicetrees without clocks out in the wild.
+	 * So clocks as optional for the time being.
+	 */
 	err = devm_clk_bulk_get(iommu->dev, iommu->num_clocks, iommu->clocks);
-	if (err)
+	if (err == -ENOENT)
+		iommu->num_clocks = 0;
+	else if (err)
 		return err;
 
 	err = clk_bulk_prepare(iommu->num_clocks, iommu->clocks);
-- 
2.16.2

Re: [PATCH v2] iommu/rockchip: make clock handling optional

[Enric Balletbo Serra]

2018-04-17 14:09 GMT+02:00 Heiko Stuebner <heiko-4mtYJXux2i+zQB+pC5nmwQ@public.gmane.org>:
> iommu clocks are optional, so the driver should not fail if they are not
> present. Instead just set the number of clocks to 0, which the clk-blk APIs
> can handle just fine.
>
> Fixes: f2e3a5f557ad ("iommu/rockchip: Control clocks needed to access the IOMMU")
> Signed-off-by: Heiko Stuebner <heiko-4mtYJXux2i+zQB+pC5nmwQ@public.gmane.org>
> Reviewed-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> ---
> changes in v2:
> - keep it required in the binding for future variants
>   but optional in the code for backwards compatiblity
> - follow Robin Murphy's beautification suggestion
>
> As the original changes breaks a lot of display support in 4.17-rc1,
> this should abviously go in as fix into 4.17.
>
>  drivers/iommu/rockchip-iommu.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c
> index 5fc8656c60f9..422d22990289 100644
> --- a/drivers/iommu/rockchip-iommu.c
> +++ b/drivers/iommu/rockchip-iommu.c
> @@ -1175,8 +1175,15 @@ static int rk_iommu_probe(struct platform_device *pdev)
>         for (i = 0; i < iommu->num_clocks; ++i)
>                 iommu->clocks[i].id = rk_iommu_clocks[i];
>
> +       /*
> +        * iommu clocks should be present for all new devices and devicetrees
> +        * but there are older devicetrees without clocks out in the wild.
> +        * So clocks as optional for the time being.
> +        */
>         err = devm_clk_bulk_get(iommu->dev, iommu->num_clocks, iommu->clocks);
> -       if (err)
> +       if (err == -ENOENT)
> +               iommu->num_clocks = 0;
> +       else if (err)
>                 return err;
>
>         err = clk_bulk_prepare(iommu->num_clocks, iommu->clocks);
> --
> 2.16.2
>
>
> _______________________________________________
> Linux-rockchip mailing list
> Linux-rockchip-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org
> http://lists.infradead.org/mailman/listinfo/linux-rockchip

Thanks to this patch my display is back :)

Tested-by: Enric Balletbo i Serra <enric.balletbo-ZGY8ohtN/8qB+jHODAdFcQ@public.gmane.org>

Re: [PATCH v2] iommu/rockchip: make clock handling optional

[Heiko Stuebner]

Ho Joerg,

Am Dienstag, 17. April 2018, 14:09:15 CEST schrieb Heiko Stuebner:
> iommu clocks are optional, so the driver should not fail if they are not
> present. Instead just set the number of clocks to 0, which the clk-blk APIs
> can handle just fine.
> 
> Fixes: f2e3a5f557ad ("iommu/rockchip: Control clocks needed to access the IOMMU")
> Signed-off-by: Heiko Stuebner <heiko-4mtYJXux2i+zQB+pC5nmwQ@public.gmane.org>
> Reviewed-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>

could you pick up this patch as fix for 4.17 please?
As without it, we loose display output with old devicetrees.


Thanks
Heiko

Re: [PATCH v2] iommu/rockchip: make clock handling optional

[Joerg Roedel]

On Tue, May 01, 2018 at 11:57:29AM +0200, Heiko Stuebner wrote:
> Ho Joerg,
> 
> Am Dienstag, 17. April 2018, 14:09:15 CEST schrieb Heiko Stuebner:
> > iommu clocks are optional, so the driver should not fail if they are not
> > present. Instead just set the number of clocks to 0, which the clk-blk APIs
> > can handle just fine.
> > 
> > Fixes: f2e3a5f557ad ("iommu/rockchip: Control clocks needed to access the IOMMU")
> > Signed-off-by: Heiko Stuebner <heiko-4mtYJXux2i+zQB+pC5nmwQ@public.gmane.org>
> > Reviewed-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> 
> could you pick up this patch as fix for 4.17 please?
> As without it, we loose display output with old devicetrees.

Applied to the iommu/fixes branch, thanks.