Re: [PATCH net-next 03/11] ethernet: use eth_hw_addr_set()

[Ido Schimmel <idosch@idosch.org>]

On Sat, Oct 02, 2021 at 05:12:55PM -0700, Jakub Kicinski wrote:
> On Sat, 2 Oct 2021 19:32:53 +0300 Ido Schimmel wrote:
> > On Fri, Oct 01, 2021 at 02:32:20PM -0700, Jakub Kicinski wrote:
> > > Convert all Ethernet drivers from memcpy(... ETH_ADDR)
> > > to eth_hw_addr_set():
> > > 
> > >   @@
> > >   expression dev, np;
> > >   @@
> > >   - memcpy(dev->dev_addr, np, ETH_ALEN)
> > >   + eth_hw_addr_set(dev, np)  
> > 
> > Some use:
> > 
> > memcpy(dev->dev_addr, np, dev->addr_len)
> > 
> > Not sure if you missed it or if it's going to be in part 2. I assume the
> > latter, but thought I would ask.
> 
> Yup, still
> 
>  417 files changed, 1239 insertions(+), 960 deletions(-)
> 
> to go. I thought I'd start upstreaming from the most obvious /
> mechanical changes.
> 
> For the memcpy(..., dev->addr_len) I'm thinking of using
> eth_hw_addr_set() in appropriate sections of the tree (drivers/ethernet,
> driver/wireless) and convert the rest to this helper:
> 
> static inline void dev_addr_set(struct net_device *dev, const u8 *addr)
> {
> 	memcpy(dev->dev_addr, addr, dev->addr_len);
> }
> 
> dev_addr_set() everywhere would be more obviously correct, but using
> eth_hw_addr_set() seems cleaner for Ethernet drivers. Second opinion
> on this would be good if you have a preference.

Yes, I agree with using eth_hw_addr_set() for Ethernet drivers as it
will make the drivers in part 1 and part 2 consistent.