linux-kernel - Re: Re: [PATCH] net: dwmac-rk: MAC clock should be truned off

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-next>] [day] [month] [year] [list]

Message-ID: <112fa3c4-908d-4e31-9288-b3a2949555b0@lunn.ch>
Date: Sat, 24 May 2025 16:48:15 +0200
From: Andrew Lunn <andrew@...n.ch>
To: lizhe <sensor1010@....com>
Cc: andrew+netdev@...n.ch, davem@...emloft.net, edumazet@...gle.com,
	kuba@...nel.org, pabeni@...hat.com, mcoquelin.stm32@...il.com,
	alexandre.torgue@...s.st.com, jonas@...boo.se,
	rmk+kernel@...linux.org.uk, david.wu@...k-chips.com, wens@...e.org,
	u.kleine-koenig@...libre.com, an.petrous@....nxp.com,
	netdev@...r.kernel.org, linux-stm32@...md-mailman.stormreply.com,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: Re: [PATCH] net: dwmac-rk: MAC clock should be truned off

On Sat, May 24, 2025 at 10:05:47PM +0800, lizhe wrote:
> Hi， Anerdw
> The following is the logic for calling this function： 
> 
> 
> rk_gmac_powerup() {
> 
> ret = phy_power_on(bsp_priv, true);      // here.
> 
> if (ret) {
> 
> gmac_clk_enable(bsp_priv, false);
> 
> return ret;
> 
> }
> 
> }

Ah, there is something funny with your patch. Look at the diff:

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index 700858ff6f7c..036e45be5828 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -1648,7 +1648,7 @@  static int gmac_clk_enable(struct rk_priv_data *bsp_priv, bool enable)

This line tells you where in the file you are patching, and the
function to be patched. This is what i looked at,
gmac_clk_enable(). And gmac_clk_enable() has a similar structure, ret
declared at the beginning, return 0 at the end. But the only way to
that return 0 is without error.

But patch is actually for:

 static int phy_power_on(struct rk_priv_data *bsp_priv, bool enable)
 {
 	struct regulator *ldo = bsp_priv->regulator;
-	int ret;
+	int ret = 0;
 	struct device *dev = &bsp_priv->pdev->dev;

 	if (enable) {
@@ -1661,7 +1661,7 @@  static int phy_power_on(struct rk_priv_data *bsp_priv, bool enable)
 			dev_err(dev, "fail to disable phy-supply\n");
 	}

-	return 0;
+	return ret;
 }

And agree, the error codes are ignored in phy_power_on().

But i have a few questions:

How did you generate this diff? This is the first time i've made this
mistake, as far as i know. I trust the context information when
reviewing patches. Yet here it is wrong. Why? Is this actually normal?
I know diff gets it wrong for python, i don't trust it at all with
that language, but i've not noticed such problems with C.

Did you look at the history of phy_power_on()? It looks pretty
deliberate ignoring errors. Maybe there is a reason why this happens?
git blame and git log might explain why it is like this.

	Andrew