* [PATCHv4 net-next] net: ag71xx: get reset control using devm api
@ 2024-08-26 21:21 Rosen Penev
2024-08-27 23:12 ` Jakub Kicinski
0 siblings, 1 reply; 9+ messages in thread
From: Rosen Penev @ 2024-08-26 21:21 UTC (permalink / raw)
To: netdev
Cc: davem, edumazet, kuba, pabeni, linux, linux-kernel, o.rempel,
p.zabel
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
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-26 21:21 [PATCHv4 net-next] net: ag71xx: get reset control using devm api Rosen Penev
@ 2024-08-27 23:12 ` Jakub Kicinski
2024-08-27 23:36 ` Rosen Penev
0 siblings, 1 reply; 9+ messages in thread
From: Jakub Kicinski @ 2024-08-27 23:12 UTC (permalink / raw)
To: Rosen Penev
Cc: netdev, davem, edumazet, pabeni, linux, linux-kernel, o.rempel,
p.zabel
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);
> }
>
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-27 23:12 ` Jakub Kicinski
@ 2024-08-27 23:36 ` Rosen Penev
2024-08-29 22:20 ` Luiz Angelo Daros de Luca
0 siblings, 1 reply; 9+ messages in thread
From: Rosen Penev @ 2024-08-27 23:36 UTC (permalink / raw)
To: Jakub Kicinski
Cc: netdev, davem, edumazet, pabeni, linux, linux-kernel, o.rempel,
p.zabel
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);
> > }
> >
>
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-27 23:36 ` Rosen Penev
@ 2024-08-29 22:20 ` Luiz Angelo Daros de Luca
2024-08-30 3:54 ` Rosen Penev
0 siblings, 1 reply; 9+ messages in thread
From: Luiz Angelo Daros de Luca @ 2024-08-29 22:20 UTC (permalink / raw)
To: Rosen Penev
Cc: Jakub Kicinski, netdev, davem, edumazet, pabeni, linux,
linux-kernel, o.rempel, p.zabel
> 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
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-29 22:20 ` Luiz Angelo Daros de Luca
@ 2024-08-30 3:54 ` Rosen Penev
0 siblings, 0 replies; 9+ messages in thread
From: Rosen Penev @ 2024-08-30 3:54 UTC (permalink / raw)
To: Luiz Angelo Daros de Luca
Cc: Jakub Kicinski, netdev, davem, edumazet, pabeni, linux,
linux-kernel, o.rempel, p.zabel
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
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCHv4 net-next] net: ag71xx: get reset control using devm api
@ 2024-08-26 19:28 Rosen Penev
2024-08-28 21:09 ` Jacob Keller
2024-08-29 4:49 ` Oleksij Rempel
0 siblings, 2 replies; 9+ messages in thread
From: Rosen Penev @ 2024-08-26 19:28 UTC (permalink / raw)
To: netdev
Cc: davem, edumazet, kuba, pabeni, linux, linux-kernel, o.rempel,
p.zabel
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
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
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-26 19:28 Rosen Penev
@ 2024-08-28 21:09 ` Jacob Keller
2024-08-29 4:49 ` Oleksij Rempel
1 sibling, 0 replies; 9+ messages in thread
From: Jacob Keller @ 2024-08-28 21:09 UTC (permalink / raw)
To: Rosen Penev, netdev
Cc: davem, edumazet, kuba, pabeni, linux, linux-kernel, o.rempel,
p.zabel
On 8/26/2024 12:28 PM, 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.
>
> Signed-off-by: Rosen Penev <rosenp@gmail.com>
> ---
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-26 19:28 Rosen Penev
2024-08-28 21:09 ` Jacob Keller
@ 2024-08-29 4:49 ` Oleksij Rempel
2024-08-29 5:56 ` Rosen Penev
1 sibling, 1 reply; 9+ messages in thread
From: Oleksij Rempel @ 2024-08-29 4:49 UTC (permalink / raw)
To: Rosen Penev
Cc: netdev, davem, edumazet, kuba, pabeni, linux, linux-kernel,
p.zabel
On Mon, Aug 26, 2024 at 12:28:45PM -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.
>
> Signed-off-by: Rosen Penev <rosenp@gmail.com>
You forgot to include my Reviewed-by tags from v3
--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCHv4 net-next] net: ag71xx: get reset control using devm api
2024-08-29 4:49 ` Oleksij Rempel
@ 2024-08-29 5:56 ` Rosen Penev
0 siblings, 0 replies; 9+ messages in thread
From: Rosen Penev @ 2024-08-29 5:56 UTC (permalink / raw)
To: Oleksij Rempel
Cc: netdev, davem, edumazet, kuba, pabeni, linux, linux-kernel,
p.zabel
On Wed, Aug 28, 2024 at 9:49 PM Oleksij Rempel <o.rempel@pengutronix.de> wrote:
>
> On Mon, Aug 26, 2024 at 12:28:45PM -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.
> >
> > Signed-off-by: Rosen Penev <rosenp@gmail.com>
>
> You forgot to include my Reviewed-by tags from v3
Correct. Although Jakub wants me to respin this removing that always
true branch.
>
> --
> Pengutronix e.K. | |
> Steuerwalder Str. 21 | http://www.pengutronix.de/ |
> 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
> Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2024-08-30 3:54 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-08-26 21:21 [PATCHv4 net-next] net: ag71xx: get reset control using devm api Rosen Penev
2024-08-27 23:12 ` Jakub Kicinski
2024-08-27 23:36 ` Rosen Penev
2024-08-29 22:20 ` Luiz Angelo Daros de Luca
2024-08-30 3:54 ` Rosen Penev
-- strict thread matches above, loose matches on Subject: below --
2024-08-26 19:28 Rosen Penev
2024-08-28 21:09 ` Jacob Keller
2024-08-29 4:49 ` Oleksij Rempel
2024-08-29 5:56 ` Rosen Penev
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).