Re: [PATCH v8 net-next 2/2] net: dsa: mv88e6xxx: mac-auth/MAB implementation

[netdev]

On 2022-11-15 23:23, Vladimir Oltean wrote:
> Is it beneficial in any way to pass the violation type to
> mv88e6xxx_handle_violation(), considering that we only call it from the
> "miss" code path, and if we were to call it with something else 
> ("member"),
> it would return a strange error code (1)?
>
> I don't necessarily see any way in which we'll need to handle the
> "member" (migration, right?) violation any different in the future,
> except ignore it, either.

MV88E6XXX_G1_ATU_OP_AGE_OUT_VIOLATION will also be handled, and it could 
be
that MV88E6XXX_G1_ATU_OP_FULL_VIOLATION would want handling, though I 
don't
know of plans for that.

The MV88E6XXX_G1_ATU_OP_MEMBER_VIOLATION interrupt can be suppressed if 
we
want.

I think a switch on the type is the most readable code form.


p.s. I have changed it, so that global1_atu.c reads:

        if (val & MV88E6XXX_G1_ATU_OP_MISS_VIOLATION) {
                dev_err_ratelimited(chip->dev,
                                    "ATU miss violation for %pM portvec 
%x spid %d\n",
                                    entry.mac, entry.portvec, spid);
                chip->ports[spid].atu_miss_violation++;

                if (!fid) {
                        err = -EINVAL;
                        goto out;
                }

                if (chip->ports[spid].mab)
                        err = mv88e6xxx_handle_violation(chip, spid, 
&entry,
                                                         fid, 
MV88E6XXX_G1_ATU_OP_MISS_VIOLATION);
                if (err)
                        goto out;
        }

with the use of out_unlock in the chip mutex locked case.