* [PATCH net-next v3] net: phy: et1011c: don't write an error code to the config register
@ 2026-11-18 15:07 Elias Pukari
2026-11-18 15:18 ` Andrew Lunn
0 siblings, 1 reply; 2+ messages in thread
From: Elias Pukari @ 2026-11-18 15:07 UTC (permalink / raw)
To: netdev
Cc: Elias Pukari, Andrew Lunn, Heiner Kallweit, Russell King,
David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni,
linux-kernel
et1011c_read_status() can write a read error to the config register.
Return MDIO errors and cache each PHY's speed only after configuration
succeeds.
LLM helped
Found by code inspection. Tested with MDIO error injection in KUnit
under QEMU/KASAN. No hardware or kselftest run.
Signed-off-by: Elias Pukari <elias.pukari@gmail.com>
---
Full builds unverified due to warnings on the unpatched base.
https://lore.kernel.org/netdev/20261115180735.649299-1-elias.pukari@gmail.com/
drivers/net/phy/et1011c.c | 37 ++++++++++++++++++++++++++-----------
1 file changed, 26 insertions(+), 11 deletions(-)
diff --git a/drivers/net/phy/et1011c.c b/drivers/net/phy/et1011c.c
index 6cd8d7758..77a6e8dcf 100644
--- a/drivers/net/phy/et1011c.c
+++ b/drivers/net/phy/et1011c.c
@@ -44,6 +44,12 @@ MODULE_DESCRIPTION("LSI ET1011C PHY driver");
MODULE_AUTHOR("Chaithrika U S");
MODULE_LICENSE("GPL");
+static int et1011c_probe(struct phy_device *phydev)
+{
+ phydev->priv = devm_kzalloc_obj(&phydev->mdio.dev, int);
+ return phydev->priv ? 0 : -ENOMEM;
+}
+
static int et1011c_config_aneg(struct phy_device *phydev)
{
int ctl = phy_read(phydev, MII_BMCR);
@@ -60,25 +66,33 @@ static int et1011c_config_aneg(struct phy_device *phydev)
static int et1011c_read_status(struct phy_device *phydev)
{
- static int speed;
+ int *configured_speed = phydev->priv;
+ bool was_up = phydev->link;
+ int val;
int ret;
- u32 val;
ret = genphy_read_status(phydev);
+ if (ret < 0 || !phydev->link) {
+ *configured_speed = SPEED_UNKNOWN;
+ return ret;
+ }
- if (speed != phydev->speed) {
- speed = phydev->speed;
+ if (!was_up || *configured_speed <= 0 ||
+ *configured_speed != phydev->speed) {
+ *configured_speed = SPEED_UNKNOWN;
val = phy_read(phydev, ET1011C_STATUS_REG);
+ if (val < 0)
+ return val;
+
if ((val & ET1011C_SPEED_MASK) ==
ET1011C_GIGABIT_SPEED) {
- val = phy_read(phydev, ET1011C_CONFIG_REG);
- val &= ~ET1011C_TX_FIFO_MASK;
- phy_write(phydev, ET1011C_CONFIG_REG, val |
- ET1011C_GMII_INTERFACE |
- ET1011C_SYS_CLK_EN |
- ET1011C_TX_FIFO_DEPTH_16);
-
+ ret = phy_modify(phydev, ET1011C_CONFIG_REG,
+ ET1011C_TX_FIFO_MASK, ET1011C_GMII_INTERFACE |
+ ET1011C_SYS_CLK_EN | ET1011C_TX_FIFO_DEPTH_16);
+ if (ret < 0)
+ return ret;
}
+ *configured_speed = phydev->speed;
}
return ret;
}
@@ -88,6 +102,7 @@ static struct phy_driver et1011c_driver[] = { {
.name = "ET1011C",
.phy_id_mask = 0xfffffff0,
/* PHY_GBIT_FEATURES */
+ .probe = et1011c_probe,
.config_aneg = et1011c_config_aneg,
.read_status = et1011c_read_status,
} };
base-commit: d6816d9a93b66a12bf01e10037e377f3333ad6cc
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH net-next v3] net: phy: et1011c: don't write an error code to the config register
2026-11-18 15:07 [PATCH net-next v3] net: phy: et1011c: don't write an error code to the config register Elias Pukari
@ 2026-11-18 15:18 ` Andrew Lunn
0 siblings, 0 replies; 2+ messages in thread
From: Andrew Lunn @ 2026-11-18 15:18 UTC (permalink / raw)
To: Elias Pukari
Cc: netdev, Heiner Kallweit, Russell King, David S. Miller,
Eric Dumazet, Jakub Kicinski, Paolo Abeni, linux-kernel
> +static int et1011c_probe(struct phy_device *phydev)
> +{
> + phydev->priv = devm_kzalloc_obj(&phydev->mdio.dev, int);
> + return phydev->priv ? 0 : -ENOMEM;
Look at other kernel code. And in look mostly at phy drivers. How
often do you see this sort of return code?
You want your code to look like all the other code.
Also, how often do you see a priv which is not structure?
> static int et1011c_config_aneg(struct phy_device *phydev)
> {
> int ctl = phy_read(phydev, MII_BMCR);
> @@ -60,25 +66,33 @@ static int et1011c_config_aneg(struct phy_device *phydev)
>
> static int et1011c_read_status(struct phy_device *phydev)
> {
> - static int speed;
> + int *configured_speed = phydev->priv;
> + bool was_up = phydev->link;
> + int val;
> int ret;
> - u32 val;
>
> ret = genphy_read_status(phydev);
> + if (ret < 0 || !phydev->link) {
> + *configured_speed = SPEED_UNKNOWN;
> + return ret;
> + }
Look at other users of genphy_read_status(). How many have code like
this? If you are doing something which no other driver does, it is
likely to be wrong.
Andrew
---
pw-bot: cr
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-18 15:18 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 15:07 [PATCH net-next v3] net: phy: et1011c: don't write an error code to the config register Elias Pukari
2026-11-18 15:18 ` Andrew Lunn
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox