net: ag71xx: add missing reset_control_put
Subsystems (3)
PATCH v2 1/1 net: ag71xx: add missing reset_control_put
The original downstream driver used devm instead of of. The latter requires reset_control_put to be called in all return paths.
---
v2: don't call after ag71xx_mdio_probe. Already done.
drivers/net/ethernet/atheros/ag71xx.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-)
drivers/net/ethernet/atheros/ag71xx.c +14 -1
@@ -722,8 +722,10 @@ static int ag71xx_mdio_probe(struct ag71xx *ag) mnp =of_get_child_by_name (np,"mdio" ); err =devm_of_mdiobus_register (dev, mii_bus, mnp);of_node_put (mnp); -if (err) +if (err) { +reset_control_put (ag->mdio_reset);return err; + }return 0 ; }
@@ -1924,12 +1926,14 @@ static int ag71xx_probe(struct platform_device *pdev) err =ag71xx_phylink_setup (ag);if (err) {netif_err (ag, probe, ndev,"failed to setup phylink (%d)\n" , err); +reset_control_put (ag->mdio_reset);return err; } err =devm_register_netdev (&pdev->dev, ndev);if (err) {netif_err (ag, probe, ndev,"unable to register net device\n" ); +reset_control_put (ag->mdio_reset);platform_set_drvdata (pdev,NULL );return err; }
@@ -1941,6 +1945,14 @@ static int ag71xx_probe(struct platform_device *pdev)return 0 ; } +static void ag71xx_remove (struct platform_device *pdev) +{ +struct net_device *ndev =platform_get_drvdata (pdev); +struct ag71xx *ag = ag =netdev_priv (ndev); + +reset_control_put (ag->mdio_reset); +} +static const u32 ar71xx_fifo_ar7100[] = {0x0fff0000 ,0x00001fff ,0x00780fff , };
@@ -2025,6 +2037,7 @@ static const struct of_device_id ag71xx_match[] = {static struct platform_driver ag71xx_driver = { .probe = ag71xx_probe, + .remove_new = ag71xx_remove, .driver = { .name ="ag71xx" , .of_match_table = ag71xx_match,
Signature · 1 line
2.46.0
Hi Rosen,
Quoted 2 lines · Rosen Penev
The original downstream driver used devm instead of of. The latter requires
reset_control_putto be called in all return paths.
At the moment of upstreaming this code, the original driver used of_reset_control_get_exclusive() and was fixed by f92bbdcc93 ("ath79:
ag71xx-mdio: get reset control using devm api")
Why not port the original fix?
Regards,
Oleksij
Signature · 4 lines
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 |
Quoted 4 lines · Oleksij Rempel
Hi Rosen,
On Fri, Aug 23, 2024 at 01:04:18PM -0700, Rosen Penev wrote:
Quoted 8 lines
Quoted 2 lines
The original downstream driver used devm instead of of. The latter requires
reset_control_putto be called in all return paths.At the moment of upstreaming this code, the original driver used
of_reset_control_get_exclusive()and was fixed byf92bbdcc93("ath79:
ag71xx-mdio: get reset control using devm api")Why not port the original fix?
On further review, this looks safe to do. Will submit tomorrow.
Quoted 8 lines
Regards,
Oleksij
--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 |