[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20181228090630.GB31460@lunn.ch>
Date: Fri, 28 Dec 2018 10:06:30 +0100
From: Andrew Lunn <andrew@...n.ch>
To: Pavel Machek <pavel@....cz>
Cc: Anderson Luiz Alves <alacn1@...il.com>,
vivien.didelot@...oirfairelinux.com, f.fainelli@...il.com,
davem@...emloft.net, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 net] mv88e6060: disable hardware level MAC learning
On Fri, Dec 28, 2018 at 09:22:55AM +0100, Pavel Machek wrote:
> On Fri 2018-11-30 21:58:36, Anderson Luiz Alves wrote:
> > Disable hardware level MAC learning because it breaks station roaming.
> > When enabled it drops all frames that arrive from a MAC address
> > that is on a different port at learning table.
> >
> > Signed-off-by: Anderson Luiz Alves <alacn1@...il.com>
>
> Will not this effectively disable the hardware switching leading to
> serious performance implications?
Hi Pavel
This switch is not doing hardware switching to start with. See
mv88e6060_setup_port():
/* Port based VLAN map: give each port its own address
* database, allow the CPU port to talk to each of the 'real'
* ports, and allow each of the 'real' ports to only talk to
* the CPU port.
*/
REG_WRITE(addr, PORT_VLAN_MAP,
((p & 0xf) << PORT_VLAN_MAP_DBNUM_SHIFT) |
(dsa_is_cpu_port(ds, p) ? dsa_user_ports(ds) :
BIT(dsa_to_port(ds, p)->cpu_dp->index)));
/* Port Association Vector: when learning source addresses
* of packets, add the address to the address database using
* a port bitmap that has only the bit for this port set and
* the other bits clear.
*/
REG_WRITE(addr, PORT_ASSOC_VECTOR, BIT(p));
This is a very old driver, added way before offload support was added.
Andrew
Powered by blists - more mailing lists