Re: [PATCH net-next 07/15] net: dsa: mt7530: do not run mt7530_setup_port5() if port 5 is disabled

[Dan Carpenter <dan.carpenter@linaro.org>]

On Sun, Dec 17, 2023 at 03:22:27PM +0300, Arınç ÜNAL wrote:
> On 8.12.2023 21:46, Vladimir Oltean wrote:
> > Hmm, maybe the problem, all along, was that we let the -ENODEV return
> > code from of_get_phy_mode() pass through? "interface" will really be
> > uninitialized in that case. It's not a false positive.
> > 
> > Instead of:
> > 
> > 	ret = of_get_phy_mode(mac_np, &interface);
> > 	if (ret && ret != -ENODEV) {
> > 		...
> > 		return ret;
> > 	}
> > 
> > it should have been like this, to not complain:
> > 
> > 	ret = of_get_phy_mode(mac_np, &interface);
> > 	if (ret) {
> > 		...
> > 		return ret;
> > 	}
> > 
> 
> I just tried this, smatch still reports "interface" as uninitialised.
> 
> $ export ARCH=mips CROSS_COMPILE=mips-linux-gnu-
> $ ../smatch/smatch_scripts/kchecker --spammy drivers/net/dsa/mt7530.c
> 
>   UPD     include/config/kernel.release
>   UPD     include/generated/utsrelease.h
>   CHECK   scripts/mod/empty.c
>   CALL    scripts/checksyscalls.sh
>   CC      drivers/net/dsa/mt7530.o
>   CHECK   drivers/net/dsa/mt7530.c
> drivers/net/dsa/mt7530.c:217 mt7530_mii_read() warn: call of 'warn_slowpath_fmt' with non-constant format argument
> drivers/net/dsa/mt7530.c:454 mt7530_setup_port6() error: uninitialized symbol 'ncpo1'.
> drivers/net/dsa/mt7530.c:868 mt7530_set_ageing_time() error: uninitialized symbol 'age_count'.
> drivers/net/dsa/mt7530.c:868 mt7530_set_ageing_time() error: uninitialized symbol 'age_unit'.
> drivers/net/dsa/mt7530.c:2324 mt7530_setup() error: uninitialized symbol 'interface'.

That's so strange.

Vladimir was right that I was misreading what he said and also hadn't
noticed the break.

For me, his approach silences the warning with or without the cross
function DB.  Also of_get_phy_mode() initializes interface on all paths
so checking for -EINVAL doesn't matter as far as this warning is
concerned.

regards,
dan carpenter