| Message ID | e615351aefba25e990215845e4812e6cb8153b28.1714433716.git.daniel@makrotopia.org (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net] net: dsa: mt7530: fix impossible MDIO address and issue warning | expand |
On 4/30/24 02:45, Daniel Golle wrote: > The MDIO address of the MT7530 and MT7531 switch ICs can be configured > using bootstrap pins. However, there are only 4 possible options for the > switch itself: 7, 15, 23 and 31 (ie. only 3 and 4 can be configured, bit > 0~2 are always 111). Practically all boards known as of today use the > default setting which is to have the switch respond to address 31, while > the built-in switch PHYs respond to address 0~4 in this case. I'm not aware of the MT7531 switch having bootstrap pins to configure the PHY address. Are you sure this is the case? Arınç
On Tue, Apr 30, 2024 at 12:45:46AM +0100, Daniel Golle wrote: > The MDIO address of the MT7530 and MT7531 switch ICs can be configured > using bootstrap pins. However, there are only 4 possible options for the > switch itself: 7, 15, 23 and 31 (ie. only 3 and 4 can be configured, bit > 0~2 are always 111). Practically all boards known as of today use the > default setting which is to have the switch respond to address 31, while > the built-in switch PHYs respond to address 0~4 in this case. > > However, even in MediaTek's SDK the address of the switch is wrongly > stated in the device trees as 0 (while in reality it is 31), so warn the > user about such broken device tree and make a good guess what was > actually intended. > > This is imporant to not break compatibility with older Device Trees as > with commit 868ff5f4944a ("net: dsa: mt7530-mdio: read PHY address of > switch from device tree") the address in device tree will be taken into > account. Doing so instead of assuming the switch is always at > address 31 which was previously hard-coded will obviously break things > for many existing downstream device trees as they contain the wrong > address (0) which previously didn't matter. > > Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch") > Signed-off-by: Daniel Golle <daniel@makrotopia.org> > --- > drivers/net/dsa/mt7530-mdio.c | 13 +++++++++++++ > 1 file changed, 13 insertions(+) > > diff --git a/drivers/net/dsa/mt7530-mdio.c b/drivers/net/dsa/mt7530-mdio.c > index fa3ee85a99c1..119630fd9060 100644 > --- a/drivers/net/dsa/mt7530-mdio.c > +++ b/drivers/net/dsa/mt7530-mdio.c > @@ -193,6 +193,19 @@ mt7530_probe(struct mdio_device *mdiodev) > return PTR_ERR(priv->io_pwr); > } > > + /* Only MDIO bus address 7, 15, 23 and 31 are valid options */ > + if (~(priv->mdiodev->addr & 0x7) & 0x7) { > + /* If the address in DT must be wrong, make a good guess about > + * the most likely intention, and issue a warning. > + */ > + int correct_addr = ((((priv->mdiodev->addr - 7) & ~0x7) % 0x20) + 7) & 0x1f; > + > + dev_warn(&mdiodev->dev, > + "impossible switch MDIO address in device tree: %d, assuming %d\n", > + priv->mdiodev->addr, correct_addr); You could include FW_WARN in this, to indicate this is a firmware issue. It is not used too much with DT, since it was originally intended for ACPI issues, but a few ARM systems use it with DT. Andrew
On Tue, Apr 30, 2024 at 12:45:46AM +0100, Daniel Golle wrote: > The MDIO address of the MT7530 and MT7531 switch ICs can be configured > using bootstrap pins. However, there are only 4 possible options for the > switch itself: 7, 15, 23 and 31 (ie. only 3 and 4 can be configured, bit > 0~2 are always 111). Practically all boards known as of today use the > default setting which is to have the switch respond to address 31, while > the built-in switch PHYs respond to address 0~4 in this case. > > However, even in MediaTek's SDK the address of the switch is wrongly > stated in the device trees as 0 (while in reality it is 31), so warn the > user about such broken device tree and make a good guess what was > actually intended. > > This is imporant to not break compatibility with older Device Trees as > with commit 868ff5f4944a ("net: dsa: mt7530-mdio: read PHY address of > switch from device tree") the address in device tree will be taken into > account. Doing so instead of assuming the switch is always at > address 31 which was previously hard-coded will obviously break things > for many existing downstream device trees as they contain the wrong > address (0) which previously didn't matter. > > Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch") > Signed-off-by: Daniel Golle <daniel@makrotopia.org> The cited commit is present in net-next but not net. So I think this patch should target net-next. ...
diff --git a/drivers/net/dsa/mt7530-mdio.c b/drivers/net/dsa/mt7530-mdio.c index fa3ee85a99c1..119630fd9060 100644 --- a/drivers/net/dsa/mt7530-mdio.c +++ b/drivers/net/dsa/mt7530-mdio.c @@ -193,6 +193,19 @@ mt7530_probe(struct mdio_device *mdiodev) return PTR_ERR(priv->io_pwr); } + /* Only MDIO bus address 7, 15, 23 and 31 are valid options */ + if (~(priv->mdiodev->addr & 0x7) & 0x7) { + /* If the address in DT must be wrong, make a good guess about + * the most likely intention, and issue a warning. + */ + int correct_addr = ((((priv->mdiodev->addr - 7) & ~0x7) % 0x20) + 7) & 0x1f; + + dev_warn(&mdiodev->dev, + "impossible switch MDIO address in device tree: %d, assuming %d\n", + priv->mdiodev->addr, correct_addr); + priv->mdiodev->addr = correct_addr; + } + regmap_config = devm_kzalloc(&mdiodev->dev, sizeof(*regmap_config), GFP_KERNEL); if (!regmap_config)
The MDIO address of the MT7530 and MT7531 switch ICs can be configured using bootstrap pins. However, there are only 4 possible options for the switch itself: 7, 15, 23 and 31 (ie. only 3 and 4 can be configured, bit 0~2 are always 111). Practically all boards known as of today use the default setting which is to have the switch respond to address 31, while the built-in switch PHYs respond to address 0~4 in this case. However, even in MediaTek's SDK the address of the switch is wrongly stated in the device trees as 0 (while in reality it is 31), so warn the user about such broken device tree and make a good guess what was actually intended. This is imporant to not break compatibility with older Device Trees as with commit 868ff5f4944a ("net: dsa: mt7530-mdio: read PHY address of switch from device tree") the address in device tree will be taken into account. Doing so instead of assuming the switch is always at address 31 which was previously hard-coded will obviously break things for many existing downstream device trees as they contain the wrong address (0) which previously didn't matter. Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch") Signed-off-by: Daniel Golle <daniel@makrotopia.org> --- drivers/net/dsa/mt7530-mdio.c | 13 +++++++++++++ 1 file changed, 13 insertions(+)