Re: [PATCH net-next 3/5] net/mlx5e: Compare all fields in IPv6 address

[Simon Horman <simon.horman@corigine.com>]

On Thu, Apr 20, 2023 at 02:05:01PM +0200, Simon Horman wrote:
> On Thu, Apr 20, 2023 at 02:52:43PM +0300, Leon Romanovsky wrote:
> > On Thu, Apr 20, 2023 at 01:09:23PM +0200, Simon Horman wrote:
> > > On Thu, Apr 20, 2023 at 11:02:49AM +0300, Leon Romanovsky wrote:
> > > > From: Leon Romanovsky <leonro@nvidia.com>
> > > > 
> > > > Fix size argument in memcmp to compare whole IPv6 address.
> > > > 
> > > > Fixes: b3beba1fb404 ("net/mlx5e: Allow policies with reqid 0, to support IKE policy holes")
> > > > Reviewed-by: Raed Salem <raeds@nvidia.com>
> > > > Reviewed-by: Emeel Hakim <ehakim@nvidia.com>
> > > > Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> > > > ---
> > > >  drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > 
> > > > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h
> > > > index f7f7c09d2b32..4e9887171508 100644
> > > > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h
> > > > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h
> > > > @@ -287,7 +287,7 @@ static inline bool addr6_all_zero(__be32 *addr6)
> > > >  {
> > > >  	static const __be32 zaddr6[4] = {};
> > > >  
> > > > -	return !memcmp(addr6, zaddr6, sizeof(*zaddr6));
> > > > +	return !memcmp(addr6, zaddr6, sizeof(zaddr6));
> > > 
> > > 1. Perhaps array_size() is appropriate here?
> > 
> > It is overkill here, sizeof(zaddr6) is constant and can't overflow.
> 
> Maybe, but the original code had a bug because using sizeof()
> directly is error prone.

Sorry, just to clarify.
I now realise that ARRAY_SIZE() is what I meant to suggest earlier.

> >   238 /**
> >   239  * array_size() - Calculate size of 2-dimensional array.
> >   240  * @a: dimension one
> >   241  * @b: dimension two
> >   242  *
> >   243  * Calculates size of 2-dimensional array: @a * @b.
> >   244  *
> >   245  * Returns: number of bytes needed to represent the array or SIZE_MAX on
> >   246  * overflow.
> >   247  */
> >   248 #define array_size(a, b)        size_mul(a, b)
> > 
> > > 2. It's a shame that ipv6_addr_any() or some other common helper
> > >    can't be used.
> > 
> > I didn't use ipv6_addr_any() as it required from me to cast "__be32 *addr6"
> > to be "struct in6_addr *" just to replace one line memcmp to another one
> > line function.
> > 
> > Do you want me to post this code instead?
> 
> No :)
> 
> I don't have a strong desire for churn.
> Just for correct code.
> 
> As your patch is correct, it is fine by me in the current form.
> 
> Reviewed-by: Simon Horman <simon.horman@corigine.com>
>