Thread (17 messages) flat view 17 messages, 2 authors, 2026-02-04

Re: [net-next,v3,08/15] net: ftgmac100: Move NCSI probe code into a helper

From: Jakub Kicinski <kuba@kernel.org>
Date: 2026-02-04 03:34:23
Also in: lkml

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.

For local reproduction steps see https://netdev-ai.bots.linux.dev/ai-local.html
---
net: ftgmac100: Move NCSI probe code into a helper

This commit refactors ftgmac100_probe() by extracting NCSI-specific probe
code into a new helper function ftgmac100_probe_ncsi(). The commit message
states "No functional change intended."
quoted hunk ↗ jump to hunk
diff --git a/drivers/net/ethernet/faraday/ftgmac100.c b/drivers/net/ethernet/faraday/ftgmac100.c
index e511d6d5f7c2..9d8256d4d368 100644
--- a/drivers/net/ethernet/faraday/ftgmac100.c
+++ b/drivers/net/ethernet/faraday/ftgmac100.c
@@ -1839,6 +1839,39 @@ static bool ftgmac100_has_child_node(struct device_node *np, const char *name)
	return ret;
 }

+static int ftgmac100_probe_ncsi(struct net_device *netdev,
+				struct ftgmac100 *priv,
+				struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct phy_device *phydev;
+	int err;
+
+	if (!IS_ENABLED(CONFIG_NET_NCSI)) {
+		dev_err(&pdev->dev, "NCSI stack not enabled\n");
+		return -EINVAL;
+	}
+
+	dev_info(&pdev->dev, "Using NCSI interface\n");
+	priv->use_ncsi = true;
+	priv->ndev = ncsi_register_dev(netdev, ftgmac100_ncsi_handler);
+	if (!priv->ndev)
+		return -EINVAL;
+
+	phydev = fixed_phy_register(&ncsi_phy_status, np);
+	if (IS_ERR(phydev)) {
+		dev_err(&pdev->dev, "failed to register fixed PHY device\n");
+		return PTR_ERR(phydev);
If fixed_phy_register() fails here, doesn't priv->ndev leak? The NCSI device
was registered via ncsi_register_dev() above, but this error path returns
without calling ncsi_unregister_dev().
+	}
+	err = phy_connect_direct(netdev, phydev, ftgmac100_adjust_link,
+				 PHY_INTERFACE_MODE_RMII);
+	if (err) {
+		dev_err(&pdev->dev, "Connecting PHY failed\n");
+		fixed_phy_unregister(phydev);
+	}
+	return err;
Similarly, if phy_connect_direct() fails, fixed_phy_unregister() is called
but ncsi_unregister_dev(priv->ndev) is not.

[ ... ]
quoted hunk ↗ jump to hunk
@@ -1928,32 +1960,9 @@ static int ftgmac100_probe(struct platform_device *pdev)
 	}

 	if (np && of_get_property(np, "use-ncsi", NULL)) {
-		if (!IS_ENABLED(CONFIG_NET_NCSI)) {
-			dev_err(&pdev->dev, "NCSI stack not enabled\n");
-			err = -EINVAL;
-			goto err_phy_connect;
[ ... ]
-		phydev = fixed_phy_register(&ncsi_phy_status, np);
-		if (IS_ERR(phydev)) {
-			dev_err(&pdev->dev, "failed to register fixed PHY device\n");
-			err = PTR_ERR(phydev);
-			goto err_phy_connect;
The old code jumped to err_phy_connect on failure, which falls through to
err_ncsi_dev where ncsi_unregister_dev(priv->ndev) is called.
+		err = ftgmac100_probe_ncsi(netdev, priv, pdev);
+		if (err)
+			goto err_setup_mdio;
The new code jumps to err_setup_mdio, which skips the err_ncsi_dev label
entirely. Combined with the helper not cleaning up priv->ndev on its error
paths, this appears to be a behavioral change from the original code.
-- 
pw-bot: cr
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help