linux-kernel - Re: [net-next] net: phy: intel-xway: workaround stale LEDs before link-up

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20260118022907.1106701-1-kuba@kernel.org>
Date: Sat, 17 Jan 2026 18:29:07 -0800
From: Jakub Kicinski <kuba@...nel.org>
To: daniel@...rotopia.org
Cc: Jakub Kicinski <kuba@...nel.org>,
	fchan@...linear.com,
	hkallweit1@...il.com,
	jpovazanec@...linear.com,
	yweng@...linear.com,
	davem@...emloft.net,
	andrew@...n.ch,
	linux@...linux.org.uk,
	edumazet@...gle.com,
	ajayaraman@...linear.com,
	john@...ozen.org,
	linux-kernel@...r.kernel.org,
	netdev@...r.kernel.org,
	lrosu@...linear.com,
	bxu@...linear.com,
	pabeni@...hat.com
Subject: Re: [net-next] net: phy: intel-xway: workaround stale LEDs before link-up

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: phy: intel-xway: workaround stale LEDs before link-up

This patch works around a PHY firmware bug where LED manual control and
polarity configuration have no effect until a link is detected. The fix
temporarily clears BMCR_ANENABLE for 100ms to force the LED thread to start.

> diff --git a/drivers/net/phy/intel-xway.c b/drivers/net/phy/intel-xway.c
> index 9766dd99afaa..e06ed7ab0ff3 100644
> --- a/drivers/net/phy/intel-xway.c
> +++ b/drivers/net/phy/intel-xway.c

[ ... ]

> @@ -286,8 +287,33 @@ static int xway_gphy_config_init(struct phy_device *phydev)
>  		return err;
>
>  	/* Use default LED configuration if 'leds' node isn't defined */
> -	if (!of_get_child_by_name(np, "leds"))
> +	if (!of_get_child_by_name(np, "leds")) {
>  		xway_gphy_init_leds(phydev);
> +	} else {

Does this leak the device_node reference returned by of_get_child_by_name()?

The API documentation states that of_get_child_by_name() returns a node
pointer with refcount incremented, requiring of_node_put() when done.
When the "leds" node exists, the returned pointer is never captured and
of_node_put() is never called.

Other PHY drivers handle this correctly, for example vsc85xx_probe_common()
in drivers/net/phy/mscc/mscc_main.c:

    np = of_get_child_by_name(phydev->mdio.dev.of_node, "leds");
    if (np) {
        of_node_put(np);
        ...
    }

And of_phy_leds() in drivers/net/phy/phy_device.c also properly calls
of_node_put(leds) on all exit paths.

Note that this leak also existed in the original code before this patch,
but since this patch is touching this exact code path, it would be a good
opportunity to fix it.
-- 
pw-bot: cr