Re: [PATCH] wifi: mt76: mt7915: add support for MT7981

[Simon Horman]

On Mon, May 15, 2023 at 02:08:22PM +0300, Kalle Valo wrote:
> Daniel Golle <daniel@xxxxxxxxxxxxxx> writes:
> 
> > On Sun, May 14, 2023 at 04:53:43PM +0200, Simon Horman wrote:
> >> On Sat, May 13, 2023 at 03:35:51PM +0200, Daniel Golle wrote:
> >> > From: Alexander Couzens <lynxis@xxxxxxx>
> >> > 
> >> > Add support for the MediaTek MT7981 SoC which is similar to the MT7986
> >> > but with a newer IP cores and only 2x ARM Cortex-A53 instead of 4x.
> >> > Unlike MT7986 the MT7981 can only connect a single wireless frontend,
> >> > usually MT7976 is used for DBDC.
> >> > 
> >> > Signed-off-by: Alexander Couzens <lynxis@xxxxxxx>
> >> > Signed-off-by: Daniel Golle <daniel@xxxxxxxxxxxxxx>
> >> 
> >> ...
> >> 
> >> > @@ -489,7 +516,10 @@ static int mt7986_wmac_adie_patch_7976(struct mt7915_dev *dev, u8 adie)
> >> >  		rg_xo_01 = 0x1d59080f;
> >> >  		rg_xo_03 = 0x34c00fe0;
> >> >  	} else {
> >> > -		rg_xo_01 = 0x1959f80f;
> >> > +		if (is_mt7981(&dev->mt76))
> >> > +			rg_xo_01 = 0x1959c80f;
> >> > +		else if (is_mt7986(&dev->mt76))
> >> > +			rg_xo_01 = 0x1959f80f;
> >> 
> >> Hi Daniel,
> >> 
> >> 		rg_xo_01 will be used uninitialised below if we get here
> >> 		and neither of the conditions above are true.
> >> 
> >> 		Can this occur?
> >
> > No, it cannot occur. Either of is_mt7981() or is_mt7986() will return
> > true, as the driver is bound via one of the two compatibles
> > 'mediatek,mt7986-wmac' or newly added 'mediatek,mt7981-wmac'.
> > Based on that the match_data is either 0x7986 or 0x7981, which is then
> > used as chip_id, which is used by the is_mt7981() and is_mt7986()
> > functions.
> 
> But what if later more changes are made, for example a third compatible
> is added? It would be good to add a warning or something else to protect
> that.
> 
> And I would not be a surpised if a compiler or static analyser would
> even warn about the uninitialised variable.

FWIIW, gcc-12 [-Wmaybe-uninitialized], clang-16 [-Wsometimes-uninitialized]
and smatch warn about this.