2024-05-08 00:52:06

by Daniel Golle

Subject: [PATCH net] net: phy: air_en8811h: reset netdev rules when LED is set manually

Setting LED_OFF via the brightness_set should deactivate hw control,
so make sure netdev trigger rules also get cleared in that case.
This matches the behaviour when using the 'netdev' trigger without
any hardware offloading and fixes unwanted memory of the default
netdev trigger rules when another trigger (or no trigger) had been
selected meanwhile.

Fixes: 71e79430117d ("net: phy: air_en8811h: Add the Airoha EN8811H PHY driver")
Signed-off-by: Daniel Golle <[email protected]>
---
This is basically a stop-gap measure until unified LED handling has
been implemented accross all MediaTek and Airoha PHYs.
See also
https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/

drivers/net/phy/air_en8811h.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/drivers/net/phy/air_en8811h.c b/drivers/net/phy/air_en8811h.c
index 4c9a1c9c805e..3cdc8c6b30b6 100644
--- a/drivers/net/phy/air_en8811h.c
+++ b/drivers/net/phy/air_en8811h.c
@@ -544,6 +544,10 @@ static int air_hw_led_on_set(struct phy_device *phydev, u8 index, bool on)

changed |= (priv->led[index].rules != 0);

+ /* clear netdev trigger rules in case LED_OFF has been set */
+ if (!on)
+ priv->led[index].rules = 0;
+
if (changed)
return phy_modify_mmd(phydev, MDIO_MMD_VEND2,
AIR_PHY_LED_ON(index),
--
2.45.0

2024-05-09 09:18:23

by Paolo Abeni

Subject: Re: [PATCH net] net: phy: air_en8811h: reset netdev rules when LED is set manually

Hi,

On Wed, 2024-05-08 at 01:51 +0100, Daniel Golle wrote:
> Setting LED_OFF via the brightness_set should deactivate hw control,
> so make sure netdev trigger rules also get cleared in that case.
> This matches the behaviour when using the 'netdev' trigger without
> any hardware offloading and fixes unwanted memory of the default
> netdev trigger rules when another trigger (or no trigger) had been
> selected meanwhile.
>
> Fixes: 71e79430117d ("net: phy: air_en8811h: Add the Airoha EN8811H PHY driver")
> Signed-off-by: Daniel Golle <[email protected]>

The patch is marked for 'net' but it's actually targeting net-next,
since the air_en8811h is only there. Does not apply to 'net' so our CI
did not trigger. Could you please repost with a the correct target
tree?

Thanks,

Paolo