linux-kernel - Re: [PATCH net-next v1 2/7] net: dsa: microchip: ksz8: Implement add/del

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: <20230404125002.dv2f4foojhy43dkx@skbuf>
Date:   Tue, 4 Apr 2023 15:50:02 +0300
From:   Vladimir Oltean <olteanv@...il.com>
To:     Oleksij Rempel <o.rempel@...gutronix.de>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Andrew Lunn <andrew@...n.ch>,
        Eric Dumazet <edumazet@...gle.com>,
        Florian Fainelli <f.fainelli@...il.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>,
        Woojung Huh <woojung.huh@...rochip.com>,
        Arun Ramadoss <arun.ramadoss@...rochip.com>,
        kernel@...gutronix.de, linux-kernel@...r.kernel.org,
        netdev@...r.kernel.org, UNGLinuxDriver@...rochip.com
Subject: Re: [PATCH net-next v1 2/7] net: dsa: microchip: ksz8: Implement
 add/del_fdb and use static MAC table operations

On Tue, Apr 04, 2023 at 02:19:11PM +0200, Oleksij Rempel wrote:
> If I compare KSZ879CLX and KSZ8873MLL datasheets, i do not see direct
> answer. The only reason I can imagine is the size of static MAC table.
> All KSZ88xx and KSZ87xx variants have only 8 entries. One is already
> used for STP (even if STP is not enabled, can be optimized). If
> BRIDGE_VLAN compiled, each local address will be configured 2 times.
> So, depending on system configuration the static MAC table will full
> very soon.

Yikes. KSZ8765 has num_statics = 8 and port_cnt = 5 (so 4 user ports I
assume). So if all 4 user ports had their own MAC address, it would
simply not be possible to put them under a VLAN-aware bridge, since that
would consume 2 BR_FDB_LOCAL entries for each port, so the static MAC
table would be full even without taking the bridge's MAC address into
consideration.

Even with CONFIG_BRIDGE_VLAN_FILTERING turned off or with the bridge
option vlan_default_pvid = 0, this would still consume 4 BR_FDB_LOCAL
entries + one for the bridge's MAC address + 1 for STP, leaving only 2
entries usable for *both* bridge fdb, *and* bridge mdb.

I haven't opened the datasheets of these chips. Is it possible to use
the dynamic MAC table to store static(-ish) entries?