netdev - Re: [PATCH net-next] net: dsa: mv88e6xxx: handle multiple ports in ATU

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <87oa3jo1l7.fsf@ketchup.i-did-not-set--mail-host-address--so-tickle-me>
Date:   Mon, 19 Sep 2016 20:29:40 -0400
From:   Vivien Didelot <vivien.didelot@...oirfairelinux.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        kernel@...oirfairelinux.com,
        "David S. Miller" <davem@...emloft.net>,
        Florian Fainelli <f.fainelli@...il.com>
Subject: Re: [PATCH net-next] net: dsa: mv88e6xxx: handle multiple ports in ATU

Hi Andrew,

Andrew Lunn <andrew@...n.ch> writes:

> Hi Vivien
>
>> +	do {
>> +		err = _mv88e6xxx_atu_getnext(chip, fid, &next);
>> +		if (err)
>> +			return err;
>> +
>> +		if (next.state == GLOBAL_ATU_DATA_STATE_UNUSED)
>> +			break;
>> +
>> +		if (ether_addr_equal(next.mac, addr)) {
>> +			*entry = next;
>> +			return 0;
>> +		}
>> +	} while (!is_broadcast_ether_addr(next.mac));
>
> This is correct, but i wonder how well it scales? When we have a lot
> of entries in the ATU, this is going to take time. At some point in
> the future, we might want to keep a shadow copy of static entries of
> the ATU in RAM. We then don't need to search for them.

There won't be any issue about time here, because we are searching a
precise FID. What takes time is dumping the whole 4095 FIDs at once.

> But that is for later, once we have some complaints this is too slow.

Yes, there are severals things we can cache in this driver, but to be
honest I'd prefer not to cache anything for the moment, until the driver
is robust. Caching data must be the very last point IMO.

Thanks,

        Vivien