Re: [PATCH net] broadcom: b44: prevent uninitialized value usage

[Jonas Gorski <jonas.gorski@gmail.com>]

Hi,

On 12/4/25 06:22, Alexey Simakov wrote:
> On execution path with raised B44_FLAG_EXTERNAL_PHY, b44_readphy()
> leaves bmcr value uninitialized and it is used later in the code.
> 
> Add check of this flag at the beginning of the b44_nway_reset() and
> exit early of the function if an external PHY is used, that would
> also correspond to other b44_readphy() call sites.
> 
> Found by Linux Verification Center (linuxtesting.org) with Svace.
> 
> Fixes: 753f492093da ("[B44]: port to native ssb support")
> Signed-off-by: Alexey Simakov <bigalex934@gmail.com>
> ---
>  drivers/net/ethernet/broadcom/b44.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/net/ethernet/broadcom/b44.c b/drivers/net/ethernet/broadcom/b44.c
> index 0353359c3fe9..cbfd65881326 100644
> --- a/drivers/net/ethernet/broadcom/b44.c
> +++ b/drivers/net/ethernet/broadcom/b44.c
> @@ -1789,6 +1789,9 @@ static int b44_nway_reset(struct net_device *dev)
>  	u32 bmcr;
>  	int r;
>  
> +	if (bp->flags & B44_FLAG_EXTERNAL_PHY)
> +		return 0;

Wouldn't the right fix here to call phy_ethtool_nway_reset(dev->phydev); instead
of just returning 0? That way it properly restarts auto-negotiation even in this
case.

> +
>  	spin_lock_irq(&bp->lock);
>  	b44_readphy(bp, MII_BMCR, &bmcr);
>  	b44_readphy(bp, MII_BMCR, &bmcr);

Best regards,
Jonas