| Message ID | 20231117100958.425354-1-robimarko@gmail.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next,1/2] net: phy: aquantia: validate PHY mode on AQR107 | expand |
On Fri, Nov 17, 2023 at 11:09:48AM +0100, Robert Marko wrote: > The Aquantia driver is not setting the PHY mode itself, but it does however > still check if the PHY mode set in DTS is one of the supported modes. > > However, the set PHY mode does not have to match the actual one, so lets > add update the PHY mode during .config_init and warn if they differ. This looks completely wrong to me. These PHYs can be configured to change their MAC-facing interface mode according to the media negotiated speed, but you are only checking that _if_ the media is up, then the interface that has resulted from that negotiation matches what is in DTS. That could be dependent on the link partner, so what works for a platform when connected to one link partner may issue your "info"-level warning when connected to a different link partner. So no, this to me looks completely wrong. You need to check the VEND1_GLOBAL_CFG_* registers, and determine from those what interface mode(s) will be used, and then use that to validate the mode. It just so happens that... http://git.armlinux.org.uk/cgit/linux-arm.git/commit/?h=net-queue&id=f7b531ee8855f81d267a8a42c44da51576f48daf http://git.armlinux.org.uk/cgit/linux-arm.git/commit/?h=net-queue&id=f55389aa5d11da8a32dfd65a1b98049878ce09f0 builds a bitmap that can then be tested to check this. Whether the above is entirely correct or not, I can't really say, I don't have enough information on this PHY.
On Fri, 17 Nov 2023 at 13:46, Russell King (Oracle) <linux@armlinux.org.uk> wrote: > > On Fri, Nov 17, 2023 at 11:09:48AM +0100, Robert Marko wrote: > > The Aquantia driver is not setting the PHY mode itself, but it does however > > still check if the PHY mode set in DTS is one of the supported modes. > > > > However, the set PHY mode does not have to match the actual one, so lets > > add update the PHY mode during .config_init and warn if they differ. > > This looks completely wrong to me. These PHYs can be configured to > change their MAC-facing interface mode according to the media negotiated > speed, but you are only checking that _if_ the media is up, then the > interface that has resulted from that negotiation matches what is in > DTS. That could be dependent on the link partner, so what works for a > platform when connected to one link partner may issue your "info"-level > warning when connected to a different link partner. > > So no, this to me looks completely wrong. > > You need to check the VEND1_GLOBAL_CFG_* registers, and determine from > those what interface mode(s) will be used, and then use that to validate > the mode. > > It just so happens that... > > http://git.armlinux.org.uk/cgit/linux-arm.git/commit/?h=net-queue&id=f7b531ee8855f81d267a8a42c44da51576f48daf > http://git.armlinux.org.uk/cgit/linux-arm.git/commit/?h=net-queue&id=f55389aa5d11da8a32dfd65a1b98049878ce09f0 > > builds a bitmap that can then be tested to check this. Whether the > above is entirely correct or not, I can't really say, I don't have > enough information on this PHY. Hi, Yeah, I get the issue now. Nice, those got merged into net-next, so I will iterate by using those. Regards, Robert > > -- > RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ > FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!
diff --git a/drivers/net/phy/aquantia/aquantia_main.c b/drivers/net/phy/aquantia/aquantia_main.c index cc4a97741c4a..7711e052e737 100644 --- a/drivers/net/phy/aquantia/aquantia_main.c +++ b/drivers/net/phy/aquantia/aquantia_main.c @@ -505,6 +505,21 @@ static void aqr107_chip_info(struct phy_device *phydev) fw_major, fw_minor, build_id, prov_id); } +static void aqr107_validate_mode(struct phy_device *phydev, + phy_interface_t dts_mode) +{ + int ret; + + /* Get the actual PHY mode */ + ret = aqr107_read_status(phydev); + if (ret) + return; + + if (dts_mode != phydev->interface) + phydev_info(phydev, "%s mode is set in DTS while %s mode is actual. Please update your devicetree.\n", + phy_modes(dts_mode), phy_modes(phydev->interface)); +} + static int aqr107_config_init(struct phy_device *phydev) { int ret; @@ -528,6 +543,8 @@ static int aqr107_config_init(struct phy_device *phydev) if (!ret) aqr107_chip_info(phydev); + aqr107_validate_mode(phydev, phydev->interface); + return aqr107_set_downshift(phydev, MDIO_AN_VEND_PROV_DOWNSHIFT_DFLT); }
The Aquantia driver is not setting the PHY mode itself, but it does however still check if the PHY mode set in DTS is one of the supported modes. However, the set PHY mode does not have to match the actual one, so lets add update the PHY mode during .config_init and warn if they differ. Signed-off-by: Robert Marko <robimarko@gmail.com> --- drivers/net/phy/aquantia/aquantia_main.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)