Re: [RFC net-next v2 3/5] net: phy: nxp-c45-tja11xx add MACsec support

[Russell King (Oracle)]

On Fri, Aug 25, 2023 at 03:50:06PM +0200, Andrew Lunn wrote:
> > > > > +static bool nxp_c45_rx_sc_valid(struct nxp_c45_secy *phy_secy,
> > > > > +				struct macsec_rx_sc *rx_sc)
> > > > > +{
> > > > > +	u16 port =  (__force u64)rx_sc->sci >> (ETH_ALEN * 8);
> > > > 
> > > > u64 sci = be64_to_cpu((__force __be64)rx_sc->sci);
> > > 
> > > why is the __force needed? What happens with a normal cast?
> > > 
> > 
> > Sparse will print warnings if __force is missing.
> 
> What is the warning? I just want to make sure __force is the correct
> solution, not that something has the wrong type and we should be
> fixing a design issue.

Hi Andrew,

rx_sc->sci is sci_t, which is defined as:

typedef u64 __bitwise sci_t;

Sparse documentation (Documentation/dev-tools/sparse.rst) states that:

  "__bitwise" is a type attribute, so you have to do something like this::
...
  which makes PM_SUSPEND and PM_RESUME "bitwise" integers (the "__force"
  is there because sparse will complain about casting to/from a bitwise
  type, but in this case we really _do_ want to force the conversion).

So basically, sci is a bitwise type, which means sparse gives it
special properties that ensures it can only be operated on using
similarly typed integers.

So, those __force casts are needed to convert sci to something else.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!