From: Andrew Lunn <andrew@lunn.ch>
To: Mengyuan Lou <mengyuanlou@net-swift.com>
Cc: netdev@vger.kernel.org, jiawenwu@net-swift.com
Subject: Re: [PATCH net-next v6] net: ngbe: Add ngbe mdio bus driver.
Date: Sun, 8 Jan 2023 17:00:42 +0100 [thread overview]
Message-ID: <Y7roqgyjDN91hSIH@lunn.ch> (raw)
In-Reply-To: <20230108093903.27054-1-mengyuanlou@net-swift.com>
> +static int ngbe_phy_read_reg_internal(struct mii_bus *bus, int phy_addr, int regnum)
> +{
> + struct wx *wx = bus->priv;
> +
> + if (regnum & MII_ADDR_C45)
> + return -EOPNOTSUPP;
> + return (u16)rd32(wx, NGBE_PHY_CONFIG(regnum));
You ignore phy_addr. Which suggests you only allow one internal
PHY. Best practice here is to put the internal PHY on phy_addr 0, and
return 0xffff for all other phy_addr values. If phylib probes the full
range, or userspace tries to access the full range, it will look like
there is no PHY at these other addresses.
> +}
> +
> +static int ngbe_phy_write_reg_internal(struct mii_bus *bus, int phy_addr, int regnum, u16 value)
> +{
> + struct wx *wx = bus->priv;
> +
> + if (regnum & MII_ADDR_C45)
> + return -EOPNOTSUPP;
> + wr32(wx, NGBE_PHY_CONFIG(regnum), value);
> + return 0;
Here, silently ignore writes to phy_addr != 0.
> + /* wait to complete */
> + ret = read_poll_timeout(rd32, val, !(val & NGBE_MSCC_BUSY), 1000,
> + 100000, false, wx, NGBE_MSCC);
> + if (ret) {
> + wx_err(wx, "PHY address command did not complete.\n");
> + return ret;
> + }
> +
> + return (u16)rd32(wx, NGBE_MSCC);
> +}
> +
> + /* wait to complete */
> + ret = read_poll_timeout(rd32, val, !(val & NGBE_MSCC_BUSY), 1000,
> + 100000, false, wx, NGBE_MSCC);
> + if (ret)
> + wx_err(wx, "PHY address command did not complete.\n");
You have the exact same error message. When you see such an error in
the log, it can sometimes be useful to know was it a read or a write
which failed. So i would suggest you put read/write into the message.
> +static void ngbe_phy_fixup(struct wx *wx)
> +{
> + struct phy_device *phydev = wx->phydev;
> + struct ethtool_eee eee;
> +
> + if (wx->mac_type != em_mac_type_mdi)
> + return;
Does this mean that if using the internal PHY the MAC does support EEE
and half duplex?
> + /* disable EEE, EEE not supported by mac */
> + memset(&eee, 0, sizeof(eee));
> + phy_ethtool_set_eee(phydev, &eee);
> +
> + phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_10baseT_Half_BIT);
> + phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_100baseT_Half_BIT);
> + phy_remove_link_mode(phydev, ETHTOOL_LINK_MODE_1000baseT_Half_BIT);
> +}
> +
> +int ngbe_mdio_init(struct wx *wx)
> +{
> + struct pci_dev *pdev = wx->pdev;
> + struct mii_bus *mii_bus;
> + int ret;
> +
> + mii_bus = devm_mdiobus_alloc(&pdev->dev);
> + if (!mii_bus)
> + return -ENOMEM;
> +
> + mii_bus->name = "ngbe_mii_bus";
> + mii_bus->read = ngbe_phy_read_reg;
> + mii_bus->write = ngbe_phy_write_reg;
> + mii_bus->phy_mask = GENMASK(31, 4);
> + mii_bus->probe_capabilities = MDIOBUS_C22_C45;
That is not strictly true. The internal MDIO bus does not suport
C45. In practice, it probably does not matter.
Andrew
next prev parent reply other threads:[~2023-01-08 16:00 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-08 9:39 [PATCH net-next v6] net: ngbe: Add ngbe mdio bus driver Mengyuan Lou
2023-01-08 13:22 ` Heiner Kallweit
2023-01-08 16:00 ` Andrew Lunn [this message]
2023-01-09 2:40 ` mengyuanlou
2023-01-09 13:11 ` Andrew Lunn
2023-01-09 14:38 ` mengyuanlou
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Y7roqgyjDN91hSIH@lunn.ch \
--to=andrew@lunn.ch \
--cc=jiawenwu@net-swift.com \
--cc=mengyuanlou@net-swift.com \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).