[v4] net: ag71xx: get reset control using devm api

[PATCHv4 net-next] net: ag71xx: get reset control using devm api

Posted by Rosen Penev 1 year, 3 months ago

Currently, the of variant is missing reset_control_put in error paths.
The devm variant does not require it.

Allows removing mdio_reset from the struct as it is not used outside the
function.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 v2: don't call after ag71xx_mdio_probe. Already done.
 v3: use devm instead.
 v4: resend to bump the CI
 drivers/net/ethernet/atheros/ag71xx.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/atheros/ag71xx.c b/drivers/net/ethernet/atheros/ag71xx.c
index 89cd001b385f..d81aa0ccd572 100644
--- a/drivers/net/ethernet/atheros/ag71xx.c
+++ b/drivers/net/ethernet/atheros/ag71xx.c
@@ -379,7 +379,6 @@ struct ag71xx {
 	u32 fifodata[3];
 	int mac_idx;
 
-	struct reset_control *mdio_reset;
 	struct clk *clk_mdio;
 };
 
@@ -683,6 +682,7 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
 	struct device *dev = &ag->pdev->dev;
 	struct net_device *ndev = ag->ndev;
 	static struct mii_bus *mii_bus;
+	struct reset_control *mdio_reset;
 	struct device_node *np, *mnp;
 	int err;
 
@@ -698,10 +698,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
 	if (!mii_bus)
 		return -ENOMEM;
 
-	ag->mdio_reset = of_reset_control_get_exclusive(np, "mdio");
-	if (IS_ERR(ag->mdio_reset)) {
+	mdio_reset = devm_reset_control_get_exclusive(dev, "mdio");
+	if (IS_ERR(mdio_reset)) {
 		netif_err(ag, probe, ndev, "Failed to get reset mdio.\n");
-		return PTR_ERR(ag->mdio_reset);
+		return PTR_ERR(mdio_reset);
 	}
 
 	mii_bus->name = "ag71xx_mdio";
@@ -712,10 +712,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
 	mii_bus->parent = dev;
 	snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s.%d", np->name, ag->mac_idx);
 
-	if (!IS_ERR(ag->mdio_reset)) {
-		reset_control_assert(ag->mdio_reset);
+	if (!IS_ERR(mdio_reset)) {
+		reset_control_assert(mdio_reset);
 		msleep(100);
-		reset_control_deassert(ag->mdio_reset);
+		reset_control_deassert(mdio_reset);
 		msleep(200);
 	}
 
-- 
2.46.0

Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api

Posted by Jakub Kicinski 1 year, 3 months ago

On Mon, 26 Aug 2024 14:21:57 -0700 Rosen Penev wrote:
> Currently, the of variant is missing reset_control_put in error paths.
> The devm variant does not require it.
> 
> Allows removing mdio_reset from the struct as it is not used outside the
> function.

> @@ -683,6 +682,7 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
>  	struct device *dev = &ag->pdev->dev;
>  	struct net_device *ndev = ag->ndev;
>  	static struct mii_bus *mii_bus;
> +	struct reset_control *mdio_reset;

nit: maintain the longest to shortest ordering of the variables
(sorted by line length not type length)

>  	struct device_node *np, *mnp;
>  	int err;
>  
> @@ -698,10 +698,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
>  	if (!mii_bus)
>  		return -ENOMEM;
>  
> -	ag->mdio_reset = of_reset_control_get_exclusive(np, "mdio");
> -	if (IS_ERR(ag->mdio_reset)) {
> +	mdio_reset = devm_reset_control_get_exclusive(dev, "mdio");
> +	if (IS_ERR(mdio_reset)) {
>  		netif_err(ag, probe, ndev, "Failed to get reset mdio.\n");
> -		return PTR_ERR(ag->mdio_reset);
> +		return PTR_ERR(mdio_reset);
>  	}
>  
>  	mii_bus->name = "ag71xx_mdio";
> @@ -712,10 +712,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
>  	mii_bus->parent = dev;
>  	snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s.%d", np->name, ag->mac_idx);
>  
> -	if (!IS_ERR(ag->mdio_reset)) {
> -		reset_control_assert(ag->mdio_reset);
> +	if (!IS_ERR(mdio_reset)) {

Are you planning to follow up to remove this check?
Would be nice to do that as second patch in the same series

> +		reset_control_assert(mdio_reset);
>  		msleep(100);
> -		reset_control_deassert(ag->mdio_reset);
> +		reset_control_deassert(mdio_reset);
>  		msleep(200);
>  	}
>

Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api

Posted by Rosen Penev 1 year, 3 months ago

On Tue, Aug 27, 2024 at 4:13 PM Jakub Kicinski <kuba@kernel.org> wrote:
>
> On Mon, 26 Aug 2024 14:21:57 -0700 Rosen Penev wrote:
> > Currently, the of variant is missing reset_control_put in error paths.
> > The devm variant does not require it.
> >
> > Allows removing mdio_reset from the struct as it is not used outside the
> > function.
>
> > @@ -683,6 +682,7 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> >       struct device *dev = &ag->pdev->dev;
> >       struct net_device *ndev = ag->ndev;
> >       static struct mii_bus *mii_bus;
> > +     struct reset_control *mdio_reset;
>
> nit: maintain the longest to shortest ordering of the variables
> (sorted by line length not type length)
>
> >       struct device_node *np, *mnp;
> >       int err;
> >
> > @@ -698,10 +698,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> >       if (!mii_bus)
> >               return -ENOMEM;
> >
> > -     ag->mdio_reset = of_reset_control_get_exclusive(np, "mdio");
> > -     if (IS_ERR(ag->mdio_reset)) {
> > +     mdio_reset = devm_reset_control_get_exclusive(dev, "mdio");
> > +     if (IS_ERR(mdio_reset)) {
> >               netif_err(ag, probe, ndev, "Failed to get reset mdio.\n");
> > -             return PTR_ERR(ag->mdio_reset);
> > +             return PTR_ERR(mdio_reset);
> >       }
> >
> >       mii_bus->name = "ag71xx_mdio";
> > @@ -712,10 +712,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> >       mii_bus->parent = dev;
> >       snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s.%d", np->name, ag->mac_idx);
> >
> > -     if (!IS_ERR(ag->mdio_reset)) {
> > -             reset_control_assert(ag->mdio_reset);
> > +     if (!IS_ERR(mdio_reset)) {
>
> Are you planning to follow up to remove this check?
> Would be nice to do that as second patch in the same series
I actually have no idea why this is here. I assume it's some mistake.
I don't think it's meant to be optional...
>
> > +             reset_control_assert(mdio_reset);
> >               msleep(100);
> > -             reset_control_deassert(ag->mdio_reset);
> > +             reset_control_deassert(mdio_reset);
> >               msleep(200);
> >       }
> >
>

Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api

Posted by Luiz Angelo Daros de Luca 1 year, 3 months ago

> On Tue, Aug 27, 2024 at 4:13 PM Jakub Kicinski <kuba@kernel.org> wrote:
> >
> > On Mon, 26 Aug 2024 14:21:57 -0700 Rosen Penev wrote:
> > > Currently, the of variant is missing reset_control_put in error paths.
> > > The devm variant does not require it.
> > >
> > > Allows removing mdio_reset from the struct as it is not used outside the
> > > function.
> >
> > > @@ -683,6 +682,7 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> > >       struct device *dev = &ag->pdev->dev;
> > >       struct net_device *ndev = ag->ndev;
> > >       static struct mii_bus *mii_bus;
> > > +     struct reset_control *mdio_reset;
> >
> > nit: maintain the longest to shortest ordering of the variables
> > (sorted by line length not type length)
> >
> > >       struct device_node *np, *mnp;
> > >       int err;
> > >
> > > @@ -698,10 +698,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> > >       if (!mii_bus)
> > >               return -ENOMEM;
> > >
> > > -     ag->mdio_reset = of_reset_control_get_exclusive(np, "mdio");
> > > -     if (IS_ERR(ag->mdio_reset)) {
> > > +     mdio_reset = devm_reset_control_get_exclusive(dev, "mdio");



> > > +     if (IS_ERR(mdio_reset)) {
> > >               netif_err(ag, probe, ndev, "Failed to get reset mdio.\n");
> > > -             return PTR_ERR(ag->mdio_reset);
> > > +             return PTR_ERR(mdio_reset);
> > >       }
> > >
> > >       mii_bus->name = "ag71xx_mdio";
> > > @@ -712,10 +712,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> > >       mii_bus->parent = dev;
> > >       snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s.%d", np->name, ag->mac_idx);
> > >
> > > -     if (!IS_ERR(ag->mdio_reset)) {
> > > -             reset_control_assert(ag->mdio_reset);
> > > +     if (!IS_ERR(mdio_reset)) {
> >
> > Are you planning to follow up to remove this check?
> > Would be nice to do that as second patch in the same series
> I actually have no idea why this is here. I assume it's some mistake.
> I don't think it's meant to be optional...

{devm,of}_reset_control_get_exclusive() will return an error if the OF
node is missing. If it should be optional, it should be
devm_reset_control_get_optional_exclusive(), which would return NULL
if it is missing.

The equivalent driver used in OpenWrt does explicitly make it
optional. https://github.com/openwrt/openwrt/blob/4646aa169986036772b9f75393c08508d20ddf8b/target/linux/ath79/files/drivers/net/ethernet/atheros/ag71xx/ag71xx_main.c#L1532,
while the mac_reset is mandatory. They might have a reason for that or
maybe only heritage from Atheros AG7100 driver.

> >
> > > +             reset_control_assert(mdio_reset);
> > >               msleep(100);
> > > -             reset_control_deassert(ag->mdio_reset);
> > > +             reset_control_deassert(mdio_reset);
> > >               msleep(200);
> > >       }

Regards,

Luiz

Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api

Posted by Rosen Penev 1 year, 3 months ago

On Thu, Aug 29, 2024 at 3:20 PM Luiz Angelo Daros de Luca
<luizluca@gmail.com> wrote:
>
> > On Tue, Aug 27, 2024 at 4:13 PM Jakub Kicinski <kuba@kernel.org> wrote:
> > >
> > > On Mon, 26 Aug 2024 14:21:57 -0700 Rosen Penev wrote:
> > > > Currently, the of variant is missing reset_control_put in error paths.
> > > > The devm variant does not require it.
> > > >
> > > > Allows removing mdio_reset from the struct as it is not used outside the
> > > > function.
> > >
> > > > @@ -683,6 +682,7 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> > > >       struct device *dev = &ag->pdev->dev;
> > > >       struct net_device *ndev = ag->ndev;
> > > >       static struct mii_bus *mii_bus;
> > > > +     struct reset_control *mdio_reset;
> > >
> > > nit: maintain the longest to shortest ordering of the variables
> > > (sorted by line length not type length)
> > >
> > > >       struct device_node *np, *mnp;
> > > >       int err;
> > > >
> > > > @@ -698,10 +698,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> > > >       if (!mii_bus)
> > > >               return -ENOMEM;
> > > >
> > > > -     ag->mdio_reset = of_reset_control_get_exclusive(np, "mdio");
> > > > -     if (IS_ERR(ag->mdio_reset)) {
> > > > +     mdio_reset = devm_reset_control_get_exclusive(dev, "mdio");
>
>
>
> > > > +     if (IS_ERR(mdio_reset)) {
> > > >               netif_err(ag, probe, ndev, "Failed to get reset mdio.\n");
> > > > -             return PTR_ERR(ag->mdio_reset);
> > > > +             return PTR_ERR(mdio_reset);
> > > >       }
> > > >
> > > >       mii_bus->name = "ag71xx_mdio";
> > > > @@ -712,10 +712,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag)
> > > >       mii_bus->parent = dev;
> > > >       snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s.%d", np->name, ag->mac_idx);
> > > >
> > > > -     if (!IS_ERR(ag->mdio_reset)) {
> > > > -             reset_control_assert(ag->mdio_reset);
> > > > +     if (!IS_ERR(mdio_reset)) {
> > >
> > > Are you planning to follow up to remove this check?
> > > Would be nice to do that as second patch in the same series
> > I actually have no idea why this is here. I assume it's some mistake.
> > I don't think it's meant to be optional...
>
> {devm,of}_reset_control_get_exclusive() will return an error if the OF
> node is missing. If it should be optional, it should be
> devm_reset_control_get_optional_exclusive(), which would return NULL
> if it is missing.
I assume during upstreaming it was suggested to remove optional.
>
> The equivalent driver used in OpenWrt does explicitly make it
> optional. https://github.com/openwrt/openwrt/blob/4646aa169986036772b9f75393c08508d20ddf8b/target/linux/ath79/files/drivers/net/ethernet/atheros/ag71xx/ag71xx_main.c#L1532,
And also the wrong check (it should be NULL).

> while the mac_reset is mandatory. They might have a reason for that or
> maybe only heritage from Atheros AG7100 driver.
>
> > >
> > > > +             reset_control_assert(mdio_reset);
> > > >               msleep(100);
> > > > -             reset_control_deassert(ag->mdio_reset);
> > > > +             reset_control_deassert(mdio_reset);
> > > >               msleep(200);
> > > >       }
>
> Regards,
>
> Luiz