| Message ID | 20240129224948.1531452-1-andrew@lunn.ch (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 585b40e25dc9ff3d2b03d1495150540849009e5b |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net,v2] net: dsa: mv88e6xxx: Fix failed probe due to unsupported C45 reads | expand |
Hello: This patch was applied to netdev/net.git (main) by Jakub Kicinski <kuba@kernel.org>: On Mon, 29 Jan 2024 23:49:48 +0100 you wrote: > Not all mv88e6xxx device support C45 read/write operations. Those > which do not return -EOPNOTSUPP. However, when phylib scans the bus, > it considers this fatal, and the probe of the MDIO bus fails, which in > term causes the mv88e6xxx probe as a whole to fail. > > When there is no device on the bus for a given address, the pull up > resistor on the data line results in the read returning 0xffff. The > phylib core code understands this when scanning for devices on the > bus. C45 allows multiple devices to be supported at one address, so > phylib will perform a few reads at each address, so although thought > not the most efficient solution, it is a way to avoid fatal > errors. Make use of this as a minimal fix for stable to fix the > probing problems. > > [...] Here is the summary with links: - [net,v2] net: dsa: mv88e6xxx: Fix failed probe due to unsupported C45 reads https://git.kernel.org/netdev/net/c/585b40e25dc9 You are awesome, thank you!
diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c index 383b3c4d6f59..614cabb5c1b0 100644 --- a/drivers/net/dsa/mv88e6xxx/chip.c +++ b/drivers/net/dsa/mv88e6xxx/chip.c @@ -3659,7 +3659,7 @@ static int mv88e6xxx_mdio_read_c45(struct mii_bus *bus, int phy, int devad, int err; if (!chip->info->ops->phy_read_c45) - return -EOPNOTSUPP; + return 0xffff; mv88e6xxx_reg_lock(chip); err = chip->info->ops->phy_read_c45(chip, bus, phy, devad, reg, &val);
Not all mv88e6xxx device support C45 read/write operations. Those which do not return -EOPNOTSUPP. However, when phylib scans the bus, it considers this fatal, and the probe of the MDIO bus fails, which in term causes the mv88e6xxx probe as a whole to fail. When there is no device on the bus for a given address, the pull up resistor on the data line results in the read returning 0xffff. The phylib core code understands this when scanning for devices on the bus. C45 allows multiple devices to be supported at one address, so phylib will perform a few reads at each address, so although thought not the most efficient solution, it is a way to avoid fatal errors. Make use of this as a minimal fix for stable to fix the probing problems. Follow up patches will rework how C45 operates to make it similar to C22 which considers -ENODEV as a none-fatal, and swap mv88e6xxx to using this. Cc: stable@vger.kernel.org Fixes: 743a19e38d02 ("net: dsa: mv88e6xxx: Separate C22 and C45 transactions") Reported-by: Tim Menninger <tmenninger@purestorage.com> Signed-off-by: Andrew Lunn <andrew@lunn.ch> --- drivers/net/dsa/mv88e6xxx/chip.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)