From: Andrew Lunn <andrew@lunn.ch>
To: Frank Sae <Frank.Sae@motor-comm.com>
Cc: davem@davemloft.net, edumazet@google.com, kuba@kernel.org,
pabeni@redhat.com, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org, xiaogang.fan@motor-comm.com,
fei.zhang@motor-comm.com, hua.sun@motor-comm.com
Subject: Re: [PATCH net-next v2 09/21] motorcomm:yt6801: Implement some hw_ops function
Date: Sat, 23 Nov 2024 02:03:08 +0100 [thread overview]
Message-ID: <46206a81-e230-411c-8a78-d461d238b171@lunn.ch> (raw)
In-Reply-To: <20241120105625.22508-10-Frank.Sae@motor-comm.com>
It took a lot of effort to find your MDIO code. And MDIO bus driver
makes a good patch on its own.
> +static int mdio_loop_wait(struct fxgmac_pdata *pdata, u32 max_cnt)
> +{
> + u32 val, i;
> +
> + for (i = 0; i < max_cnt; i++) {
> + val = rd32_mac(pdata, MAC_MDIO_ADDRESS);
> + if ((val & MAC_MDIO_ADDR_BUSY) == 0)
> + break;
> +
> + fsleep(10);
> + }
> +
> + if (i >= max_cnt) {
> + WARN_ON(1);
> + yt_err(pdata, "%s timeout. used cnt:%d, reg_val=%x.\n",
> + __func__, i + 1, val);
> +
> + return -ETIMEDOUT;
> + }
Please replace this using one of the helpers in
include/linux/iopoll.h.
> +#define PHY_WR_CONFIG(reg_offset) (0x8000205 + ((reg_offset) * 0x10000))
> +static int fxgmac_phy_write_reg(struct fxgmac_pdata *pdata, u32 reg_id, u32 data)
> +{
> + int ret;
> +
> + wr32_mac(pdata, data, MAC_MDIO_DATA);
> + wr32_mac(pdata, PHY_WR_CONFIG(reg_id), MAC_MDIO_ADDRESS);
> + ret = mdio_loop_wait(pdata, PHY_MDIO_MAX_TRY);
> + if (ret < 0)
> + return ret;
> +
> + yt_dbg(pdata, "%s, id:%x %s, ctrl:0x%08x, data:0x%08x\n", __func__,
> + reg_id, (ret == 0) ? "ok" : "err", PHY_WR_CONFIG(reg_id), data);
> +
> + return ret;
> +}
> +
> +#define PHY_RD_CONFIG(reg_offset) (0x800020d + ((reg_offset) * 0x10000))
> +static int fxgmac_phy_read_reg(struct fxgmac_pdata *pdata, u32 reg_id)
> +{
> + u32 val;
> + int ret;
> +
> + wr32_mac(pdata, PHY_RD_CONFIG(reg_id), MAC_MDIO_ADDRESS);
> + ret = mdio_loop_wait(pdata, PHY_MDIO_MAX_TRY);
> + if (ret < 0)
> + return ret;
> +
> + val = rd32_mac(pdata, MAC_MDIO_DATA); /* Read data */
> + yt_dbg(pdata, "%s, id:%x ok, ctrl:0x%08x, val:0x%08x.\n", __func__,
> + reg_id, PHY_RD_CONFIG(reg_id), val);
> +
> + return val;
> +}
And where is the rest of the MDIO bus driver?
> +static int fxgmac_config_flow_control(struct fxgmac_pdata *pdata)
> +{
> + u32 val = 0;
> + int ret;
> +
> + fxgmac_config_tx_flow_control(pdata);
> + fxgmac_config_rx_flow_control(pdata);
> +
> + /* Set auto negotiation advertisement pause ability */
> + if (pdata->tx_pause || pdata->rx_pause)
> + val |= ADVERTISE_PAUSE_CAP | ADVERTISE_PAUSE_ASYM;
> +
> + ret = phy_modify(pdata->phydev, MII_ADVERTISE,
> + ADVERTISE_PAUSE_CAP | ADVERTISE_PAUSE_ASYM, val);
> + if (ret < 0)
> + return ret;
> +
> + return phy_modify(pdata->phydev, MII_BMCR, BMCR_RESET, BMCR_RESET);
> +}
Yet more code messing with the PHY. This all needs to go.
> +static int fxgmac_phy_clear_interrupt(struct fxgmac_pdata *pdata)
> +{
> + u32 stats_pre, stats;
> +
> + if (mutex_trylock(&pdata->phydev->mdio.bus->mdio_lock) == 0) {
> + yt_dbg(pdata, "lock not ready!\n");
> + return 0;
> + }
> +
> + stats_pre = fxgmac_phy_read_reg(pdata, PHY_INT_STATUS);
> + if (stats_pre < 0)
> + goto unlock;
> +
> + stats = fxgmac_phy_read_reg(pdata, PHY_INT_STATUS);
> + if (stats < 0)
> + goto unlock;
> +
> + phy_unlock_mdio_bus(pdata->phydev);
> +
> +#define LINK_DOWN 0x800
> +#define LINK_UP 0x400
> +#define LINK_CHANGE (LINK_DOWN | LINK_UP)
> + if ((stats_pre & LINK_CHANGE) != (stats & LINK_CHANGE)) {
> + yt_dbg(pdata, "phy link change\n");
> + return 1;
> + }
> +
> + return 0;
> +unlock:
> + phy_unlock_mdio_bus(pdata->phydev);
> + yt_err(pdata, "fxgmac_phy_read_reg err!\n");
> + return -ETIMEDOUT;
> +}
You need to rework your PHY interrupt handling. The PHY driver is
responsible for handing the interrupt registers in the PHY. Ideally
you just want to export an interrupt to phylib, so it can do all the
work.
Andrew
next prev parent reply other threads:[~2024-11-23 1:03 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-20 10:56 [PATCH net-next v2 00/21] net:yt6801: Add Motorcomm yt6801 PCIe driver Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 03/21] motorcomm:yt6801: Implement the fxgmac_drv_probe function Frank Sae
2024-11-22 17:30 ` Vadim Fedorenko
2024-11-20 10:56 ` [PATCH net-next v2 04/21] motorcomm:yt6801: Implement the .ndo_open function Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 05/21] motorcomm:yt6801: Implement the fxgmac_start function Frank Sae
2024-11-22 23:16 ` Andrew Lunn
2024-11-25 9:31 ` Frank Sae
2024-11-25 14:18 ` Andrew Lunn
2024-11-26 3:15 ` Frank Sae
2024-11-26 9:28 ` Frank Sae
2024-11-26 13:48 ` Andrew Lunn
2024-11-20 10:56 ` [PATCH net-next v2 08/21] motorcomm:yt6801: Implement the fxgmac_read_mac_addr function Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 12/21] motorcomm:yt6801: Implement .ndo_tx_timeout and .ndo_change_mtu functions Frank Sae
2024-11-22 23:19 ` Andrew Lunn
2024-11-20 10:56 ` [PATCH net-next v2 13/21] motorcomm:yt6801: Implement some ethtool_ops function Frank Sae
2024-11-22 22:34 ` Andrew Lunn
2024-11-20 10:56 ` [PATCH net-next v2 14/21] motorcomm:yt6801: Implement the WOL function of ethtool_ops Frank Sae
2024-11-22 23:04 ` Andrew Lunn
2024-11-20 10:56 ` [PATCH net-next v2 15/21] motorcomm:yt6801: Implement pci_driver suspend and resume Frank Sae
2024-11-22 23:10 ` Andrew Lunn
2024-11-20 10:56 ` [PATCH net-next v2 16/21] motorcomm:yt6801: Add a Makefile in the motorcomm folder Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 17/21] motorcomm:yt6801: Update the Makefile and Kconfig in the motorcomm Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 18/21] motorcomm:yt6801: Update the Makefile and Kconfig in the ethernet Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 19/21] ethernet: Update the index.rst in the ethernet documentation folder Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 20/21] motorcomm:yt6801: Add a yt6801.rst " Frank Sae
2024-11-20 10:56 ` [PATCH net-next v2 21/21] MAINTAINERS:Add the motorcomm ethernet driver entry Frank Sae
2024-11-20 11:14 ` [PATCH net-next v2 11/21] motorcomm:yt6801: Implement some net_device_ops function Frank Sae
2024-11-20 11:14 ` [PATCH net-next v2 09/21] motorcomm:yt6801: Implement some hw_ops function Frank Sae
2024-11-23 1:03 ` Andrew Lunn [this message]
2024-11-25 9:49 ` Frank Sae
2024-11-25 14:39 ` Andrew Lunn
2024-11-26 3:04 ` Frank Sae
2024-11-21 1:52 ` [PATCH net-next v2 07/21] motorcomm:yt6801: Implement the fxgmac_init function Frank Sae
2024-11-21 1:53 ` [PATCH net-next v2 02/21] motorcomm:yt6801: Implement pci_driver shutdown Frank Sae
2024-11-21 1:53 ` [PATCH net-next v2 01/21] motorcomm:yt6801: Add support for a pci table in this module Frank Sae
2024-11-21 1:53 ` [PATCH net-next v2 06/21] motorcomm:yt6801: Implement the poll functions Frank Sae
2024-11-21 6:18 ` [PATCH net-next v2 10/21] motorcomm:yt6801: Implement .ndo_start_xmit function Frank Sae
2024-11-21 9:00 ` [PATCH net-next v2 00/21] net:yt6801: Add Motorcomm yt6801 PCIe driver Paolo Abeni
2024-11-22 22:14 ` Andrew Lunn
2024-11-23 0:45 ` Andrew Lunn
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=46206a81-e230-411c-8a78-d461d238b171@lunn.ch \
--to=andrew@lunn.ch \
--cc=Frank.Sae@motor-comm.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=fei.zhang@motor-comm.com \
--cc=hua.sun@motor-comm.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=xiaogang.fan@motor-comm.com \
/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
all inboxes | Powered by JetHome®