Skip to contentSkip to conversation

net: ethernet: ixgbe: don't propagate -ENODEV from ixgbe_mii_bus_init()

PATCH In review Bartosz Golaszewski 3 participants 4 messages Sep 28, 2020 – Oct 12, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net: ethernet: ixgbe: don't propagate -ENODEV from ixgbe_mii_bus_init()

From: Bartosz Golaszewski <[email protected]>

It's a valid use-case for ixgbe_mii_bus_init() to return -ENODEV - we still want to finalize the registration of the ixgbe device. Check the error code and don't bail out if err == -ENODEV.

This fixes an issue on C3000 family of SoCs where four ixgbe devices share a single MDIO bus and ixgbe_mii_bus_init() returns -ENODEV for three of them but we still want to register them.

Fixes 09ef193fef7e ("net: ethernet: ixgbe: check the return value of ixgbe_mii_bus_init()")
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +1 -1
index 2f8a4cfc5fa1..d1623af30125 100644
@@ -11032,7 +11032,7 @@ static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 			true);
 
 	err = ixgbe_mii_bus_init(hw);
-	if (err)
+	if (err && err != -ENODEV)
 		goto err_netdev;
 
 	return 0;
Signature · 1 line
2.26.1
#1 lore raw
Quoted 32 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

It's a valid use-case for ixgbe_mii_bus_init() to return -ENODEV - we still want to finalize the registration of the ixgbe device. Check the error code and don't bail out if err == -ENODEV.

This fixes an issue on C3000 family of SoCs where four ixgbe devices share a single MDIO bus and ixgbe_mii_bus_init() returns -ENODEV for three of them but we still want to register them.

Fixes 09ef193fef7e ("net: ethernet: ixgbe: check the return value of ixgbe_mii_bus_init()")
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +1 -2
index 2f8a4cfc5fa1..d1623af30125 100644
@@ -11032,7 +11032,7 @@ static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
                        true);
 
        err = ixgbe_mii_bus_init(hw);
-       if (err)
+       if (err && err != -ENODEV)
                goto err_netdev;
 
        return 0;
--
Quoted 2 lines

2.26.1

Hi!

Gentle ping for this patch. Who's picking up networking patches now that David is OoO? Should I Cc someone else?

Bartosz

maintainer
#2 lore raw
Quoted 31 lines · On Mon, 12 Oct 2020 14:20:16 +0200 Bartosz Golaszewski wrote:
Quoted 30 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

It's a valid use-case for ixgbe_mii_bus_init() to return -ENODEV - we still want to finalize the registration of the ixgbe device. Check the error code and don't bail out if err == -ENODEV.

This fixes an issue on C3000 family of SoCs where four ixgbe devices share a single MDIO bus and ixgbe_mii_bus_init() returns -ENODEV for three of them but we still want to register them.

Fixes 09ef193fef7e ("net: ethernet: ixgbe: check the return value of ixgbe_mii_bus_init()")
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +1 -1
index 2f8a4cfc5fa1..d1623af30125 100644
@@ -11032,7 +11032,7 @@ static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
                        true);
 
        err = ixgbe_mii_bus_init(hw);
-       if (err)
+       if (err && err != -ENODEV)
                goto err_netdev;
 
Quoted 4 lines

drivers/net/ethernet/intel/ixgbe/ixgbe_main.c static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent) · line 11038

Quoted 1 line
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +0 -0
        return 0;

Gentle ping for this patch. Who's picking up networking patches now that David is OoO? Should I Cc someone else?

Intel went through a maintainer change of its own, and they usually pick up their patches and send a PR.

Tony, do you want me to apply this directly?

#3 lore raw
Quoted 45 lines · Jakub Kicinski
Quoted 39 lines · On Mon, 12 Oct 2020 14:20:16 +0200 Bartosz Golaszewski wrote:
Quoted 37 lines · Bartosz Golaszewski

From: Bartosz Golaszewski <[email protected]>

It's a valid use-case for ixgbe_mii_bus_init() to return -ENODEV
- we
still want to finalize the registration of the ixgbe device. Check the
error code and don't bail out if err == -ENODEV.

This fixes an issue on C3000 family of SoCs where four ixgbe devices
share a single MDIO bus and ixgbe_mii_bus_init() returns -ENODEV for
three of them but we still want to register them.

Fixes 09ef193fef7e ("net: ethernet: ixgbe: check the return

value of ixgbe_mii_bus_init()")

---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +1 -1
index 2f8a4cfc5fa1..d1623af30125 100644
@@ -11032,7 +11032,7 @@ static int ixgbe_probe(struct pci_dev
 *pdev, const struct pci_device_id *ent)
                        true);
 
        err = ixgbe_mii_bus_init(hw);
-       if (err)
+       if (err && err != -ENODEV)
                goto err_netdev;
 
Quoted 4 lines
Quoted 1 line
        return 0;

Gentle ping for this patch. Who's picking up networking patches now that David is OoO? Should I Cc someone else?

Quoted 4 lines

Intel went through a maintainer change of its own, and they usually pick up their patches and send a PR.

Tony, do you want me to apply this directly?

Hi Jakub,

I can take it. That way we can get some testing done on it.

Thanks,
Tony