Currently kernel crashes when of_phy_connect has issues:
Unable to handle kernel access to user memory outside uaccess routines at virtual address
0000000000000308
...
pc : phy_attached_print+0x28/0x1b0
lr : phy_attached_info+0x14/0x20
...
Call trace:
phy_attached_print+0x28/0x1b0
phy_attached_info+0x14/0x20
nss_dp_adjust_link+0x544/0x6c4 [qca_nss_dp]
of_phy_connect returns either pointer or NULL, so can't be checked with
IS_ERR macro.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit
38c7cf0e69aeefdec44d513307732e4daf7d9794)
- dp_priv->phydev = phy_connect(netdev, phy_id,
- &nss_dp_adjust_link,
- dp_priv->phy_mii_type);
+- if (IS_ERR(dp_priv->phydev)) {
+- netdev_dbg(netdev, "failed to connect to phy device\n");
+ dp_priv->phydev = of_phy_connect(netdev, dp_priv->phy_node,
+ &nss_dp_adjust_link, 0,
+ dp_priv->phy_mii_type);
- if (IS_ERR(dp_priv->phydev)) {
-- netdev_dbg(netdev, "failed to connect to phy device\n");
++ if (!(dp_priv->phydev)) {
+ dev_err(&pdev->dev, "Could not attach to PHY\n");
goto phy_setup_fail;
}