Hello!
On 03/11/2019 10:43 AM, Geert Uytterhoeven wrote:
>> In case of_get_phy_mode fails, the fix returns NULL to avoid
>> the NULL pointer dereference.
>>
>> Signed-off-by: Kangjie Lu <kjlu@xxxxxxx>
>
> Thanks for your patch!
>
>> --- a/drivers/net/ethernet/renesas/sh_eth.c
>> +++ b/drivers/net/ethernet/renesas/sh_eth.c
>> @@ -3187,6 +3187,8 @@ static struct sh_eth_plat_data *sh_eth_parse_dt(struct device *dev)
>> return NULL;
>>
>> pdata->phy_interface = of_get_phy_mode(np);
>> + if (unlikely(!pdata->phy_interface))
>> + return NULL;
>
> of_get_phy_mode() does not return a pointer, but a phy mode index.
> Hence zero is a valid value.
>
> So I'm wondering where is the potential NULL pointer dereference?
Hm, not seeing NULL deref as well...
> Perhaps we do need to check for negative error codes?
> Or is this already handled later, by {of_,}phy_connect()?
Not seeing such handling too...
The v2 patch seems good though, as the "phy-mode" prop is mandatory anyway.
> Gr{oetje,eeting}s,
>
> Geert
MBR, Sergei
