[PATCH 0/5] NET: PHY names vs MDIO bus names regression fixes

[PATCH 0/5] NET: PHY names vs MDIO bus names regression fixes

[Florian Fainelli]

Hello David,

This patch serie aims at fixing the regressions introduced with the
"use an unique MDIO bus name" patchset. Some drivers still formatted
the PHY id to use the old bus name (e.g: "0" or "1") and therefore
the PHY probing logic gets broken because their underlying MDIO bus
name was changed too from "0" to <foo>-<id>.

The offending patches got introduced during 3.3-rc1.

Thanks!

Florian Fainelli (5):
  cpmac: fix PHY name to match MDIO bus name
  bcm63xx-enet: fix PHY name to match MDIO bus name
  fec: fix PHY name to match fixed MDIO bus name
  octeon: fix PHY name to match MDIO bus name
  ixp4xx-eth: fix PHY name to match MDIO bus name

 drivers/net/ethernet/broadcom/bcm63xx_enet.c |    2 +-
 drivers/net/ethernet/freescale/fec.c         |    2 +-
 drivers/net/ethernet/octeon/octeon_mgmt.c    |    4 ++--
 drivers/net/ethernet/ti/cpmac.c              |    4 ++--
 drivers/net/ethernet/xscale/ixp4xx_eth.c     |    3 ++-
 5 files changed, 8 insertions(+), 7 deletions(-)

-- 
1.7.5.4

[PATCH 1/5] cpmac: fix PHY name to match MDIO bus name

[Florian Fainelli]

Commit d1733f07: cpmac: use an unique MDIO bus name changed the MDIO bus
name from "1" to "cpmac-1", this breaks the PHY connection logic because
the PHY name still uses the old bus names "0" and "1", fix that to
always use the mdio bus id instead.

CC: stable@vger.kernel.org
Signed-off-by: Florian Fainelli <florian@openwrt.org>
---
 drivers/net/ethernet/ti/cpmac.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/ti/cpmac.c b/drivers/net/ethernet/ti/cpmac.c
index 97e1df3..860c252 100644
--- a/drivers/net/ethernet/ti/cpmac.c
+++ b/drivers/net/ethernet/ti/cpmac.c
@@ -1122,7 +1122,7 @@ static int __devinit cpmac_probe(struct platform_device *pdev)
 	pdata = pdev->dev.platform_data;
 
 	if (external_switch || dumb_switch) {
-		strncpy(mdio_bus_id, "0", MII_BUS_ID_SIZE); /* fixed phys bus */
+		strncpy(mdio_bus_id, "fixed-0", MII_BUS_ID_SIZE); /* fixed phys bus */
 		phy_id = pdev->id;
 	} else {
 		for (phy_id = 0; phy_id < PHY_MAX_ADDR; phy_id++) {
@@ -1138,7 +1138,7 @@ static int __devinit cpmac_probe(struct platform_device *pdev)
 	if (phy_id == PHY_MAX_ADDR) {
 		dev_err(&pdev->dev, "no PHY present, falling back "
 					"to switch on MDIO bus 0\n");
-		strncpy(mdio_bus_id, "0", MII_BUS_ID_SIZE); /* fixed phys bus */
+		strncpy(mdio_bus_id, "fixed-0", MII_BUS_ID_SIZE); /* fixed phys bus */
 		phy_id = pdev->id;
 	}
 
-- 
1.7.5.4

[PATCH 2/5] bcm63xx-enet: fix PHY name to match MDIO bus name

[Florian Fainelli]

Commit 3e617506: bcm63xx_enet: use an unique MDIO bus name introduced
a regression in the PHY connection logic, since the PHY name was formatted
to expect the bus name to be "0" or "1", whereas it is now "bcm63xx-enet-0"
or "bcm63xx-enet-1".

CC: stable@vger.kernel.org
Reported-by: Joel EJC <joel_ejc@yahoofr>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
---
 drivers/net/ethernet/broadcom/bcm63xx_enet.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bcm63xx_enet.c b/drivers/net/ethernet/broadcom/bcm63xx_enet.c
index 986019b..c7ca7ec 100644
--- a/drivers/net/ethernet/broadcom/bcm63xx_enet.c
+++ b/drivers/net/ethernet/broadcom/bcm63xx_enet.c
@@ -797,7 +797,7 @@ static int bcm_enet_open(struct net_device *dev)
 	if (priv->has_phy) {
 		/* connect to PHY */
 		snprintf(phy_id, sizeof(phy_id), PHY_ID_FMT,
-			 priv->mac_id ? "1" : "0", priv->phy_id);
+			 priv->mii_bus->id, priv->phy_id);
 
 		phydev = phy_connect(dev, phy_id, bcm_enet_adjust_phy_link, 0,
 				     PHY_INTERFACE_MODE_MII);
-- 
1.7.5.4

[PATCH 3/5] fec: fix PHY name to match fixed MDIO bus name

[Florian Fainelli]

Commit "391420f7: fec: use an unique MDIO bus name" first modified
the MDIO bus name to include the platform name, then in commit
"a7ed07d5: net: fec: correct phy_name buffer length when init phy_name"
the PHY name formatting was fixed in the case the PHY matches a PHY
driver.

The FEC driver however, also handles the case where we want to attach
to the fixed MDIO bus name, which was previously named "0", and now
"fixed-0". Change the PHY formatting logic to account for that.

CC: stable@vger.kernel.org
Signed-off-by: Florian Fainelli <florian@openwrt.org>
---
 drivers/net/ethernet/freescale/fec.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/freescale/fec.c b/drivers/net/ethernet/freescale/fec.c
index 0ee2ca7..a12b3f5 100644
--- a/drivers/net/ethernet/freescale/fec.c
+++ b/drivers/net/ethernet/freescale/fec.c
@@ -986,7 +986,7 @@ static int fec_enet_mii_probe(struct net_device *ndev)
 		printk(KERN_INFO
 			"%s: no PHY, assuming direct connection to switch\n",
 			ndev->name);
-		strncpy(mdio_bus_id, "0", MII_BUS_ID_SIZE);
+		strncpy(mdio_bus_id, "fixed-0", MII_BUS_ID_SIZE);
 		phy_id = 0;
 	}
 
-- 
1.7.5.4

[PATCH 4/5] octeon: fix PHY name to match MDIO bus name

[Florian Fainelli]

Commit "d6c25be: mdio-octeon: use an unique MDIO bus name" changed the
octeon MDIO bus name from "0" to "mdio-octeon-0", change the PHY
formatting logic to account for that name change, so that PHY connection
on this bus succeeds.

CC: stable@vger.kernel.org
Signed-off-by: Florian Fainelli <florian@openwrt.org>
---
 drivers/net/ethernet/octeon/octeon_mgmt.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/octeon/octeon_mgmt.c b/drivers/net/ethernet/octeon/octeon_mgmt.c
index 212f43b..cd827ff 100644
--- a/drivers/net/ethernet/octeon/octeon_mgmt.c
+++ b/drivers/net/ethernet/octeon/octeon_mgmt.c
@@ -670,7 +670,7 @@ static void octeon_mgmt_adjust_link(struct net_device *netdev)
 static int octeon_mgmt_init_phy(struct net_device *netdev)
 {
 	struct octeon_mgmt *p = netdev_priv(netdev);
-	char phy_id[20];
+	char phy_id[MII_BUS_ID_SIZE + 3];
 
 	if (octeon_is_simulation()) {
 		/* No PHYs in the simulator. */
@@ -678,7 +678,7 @@ static int octeon_mgmt_init_phy(struct net_device *netdev)
 		return 0;
 	}
 
-	snprintf(phy_id, sizeof(phy_id), PHY_ID_FMT, "0", p->port);
+	snprintf(phy_id, sizeof(phy_id), PHY_ID_FMT, "mdio-octeon-0", p->port);
 
 	p->phydev = phy_connect(netdev, phy_id, octeon_mgmt_adjust_link, 0,
 				PHY_INTERFACE_MODE_MII);
-- 
1.7.5.4

[PATCH 5/5] ixp4xx-eth: fix PHY name to match MDIO bus name

[Florian Fainelli]

Commit 0869b3a4: ixp4xx-eth: use an unique MDIO bus name changed
the MDIO bus name from "0" to "ixp4xx-eth-0", as a result the PHY
name is not longer appropriate and will not match the MDIO bus name
so PHY connection will not succeed, fix that.

CC: stable@vger.kernel.org
Signed-off-by: Florian Fainelli <florian@openwrt.org>
---
 drivers/net/ethernet/xscale/ixp4xx_eth.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/xscale/ixp4xx_eth.c b/drivers/net/ethernet/xscale/ixp4xx_eth.c
index 72a854f..41a8b5a 100644
--- a/drivers/net/ethernet/xscale/ixp4xx_eth.c
+++ b/drivers/net/ethernet/xscale/ixp4xx_eth.c
@@ -1416,7 +1416,8 @@ static int __devinit eth_init_one(struct platform_device *pdev)
 	__raw_writel(DEFAULT_CORE_CNTRL, &port->regs->core_control);
 	udelay(50);
 
-	snprintf(phy_id, MII_BUS_ID_SIZE + 3, PHY_ID_FMT, "0", plat->phy);
+	snprintf(phy_id, MII_BUS_ID_SIZE + 3, PHY_ID_FMT,
+		mdio_bus->id, plat->phy);
 	port->phydev = phy_connect(dev, phy_id, &ixp4xx_adjust_link, 0,
 				   PHY_INTERFACE_MODE_MII);
 	if (IS_ERR(port->phydev)) {
-- 
1.7.5.4

Re: [PATCH 0/5] NET: PHY names vs MDIO bus names regression fixes

[David Miller]

From: Florian Fainelli <florian@openwrt.org>
Date: Mon, 13 Feb 2012 12:23:19 +0100

> This patch serie aims at fixing the regressions introduced with the
> "use an unique MDIO bus name" patchset. Some drivers still formatted
> the PHY id to use the old bus name (e.g: "0" or "1") and therefore
> the PHY probing logic gets broken because their underlying MDIO bus
> name was changed too from "0" to <foo>-<id>.
> 
> The offending patches got introduced during 3.3-rc1.

Then why the heck are you CC:'ing -stable on these patches?

We're still in the 3.3-rcX phase, so there is no other tree in
the world these patches should be applied to.

In any event, for networking stuff, you shouldn't be adding
the stable CC: anyways, I take care of all networking stable
submissions myself.

Re: [PATCH 0/5] NET: PHY names vs MDIO bus names regression fixes

[Florian Fainelli]

Le lundi 13 février 2012 22:02:04, David Miller a écrit :
> From: Florian Fainelli <florian@openwrt.org>
> Date: Mon, 13 Feb 2012 12:23:19 +0100
> 
> > This patch serie aims at fixing the regressions introduced with the
> > "use an unique MDIO bus name" patchset. Some drivers still formatted
> > the PHY id to use the old bus name (e.g: "0" or "1") and therefore
> > the PHY probing logic gets broken because their underlying MDIO bus
> > name was changed too from "0" to <foo>-<id>.
> > 
> > The offending patches got introduced during 3.3-rc1.
> 
> Then why the heck are you CC:'ing -stable on these patches?

There was no reason, just some bad habit.

> 
> We're still in the 3.3-rcX phase, so there is no other tree in
> the world these patches should be applied to.
> 
> In any event, for networking stuff, you shouldn't be adding
> the stable CC: anyways, I take care of all networking stable
> submissions myself.

Allright, do you want me to resend these without the CC or are they okay as 
is?
-- 
Florian

Re: [PATCH 0/5] NET: PHY names vs MDIO bus names regression fixes

[David Miller]

From: Florian Fainelli <florian@openwrt.org>
Date: Mon, 13 Feb 2012 23:00:27 +0100

> Allright, do you want me to resend these without the CC or are they okay as 
> is?

If you actually took a look at my tree you'd see all of your
changes there, and therefore you wouldn't have to waste time
asking silly questions like this.