netdev - Re: [PATCH net-next 4/5] net: dsa: mv88e6xxx: Offload bridge learning flag

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]
Date:   Wed, 17 Mar 2021 10:57:45 +0100
From:   Tobias Waldekranz <tobias@...dekranz.com>
To:     Vladimir Oltean <olteanv@...il.com>
Cc:     davem@...emloft.net, kuba@...nel.org, andrew@...n.ch,
        vivien.didelot@...il.com, f.fainelli@...il.com,
        netdev@...r.kernel.org
Subject: Re: [PATCH net-next 4/5] net: dsa: mv88e6xxx: Offload bridge learning flag

On Tue, Mar 16, 2021 at 11:27, Vladimir Oltean <olteanv@...il.com> wrote:
> On Mon, Mar 15, 2021 at 10:13:59PM +0100, Tobias Waldekranz wrote:
>> Allow a user to control automatic learning per port.
>> 
>> Many chips have an explicit "LearningDisable"-bit that can be used for
>> this, but we opt for setting/clearing the PAV instead, as it works on
>> all devices at least as far back as 6083.
>> 
>> Signed-off-by: Tobias Waldekranz <tobias@...dekranz.com>
>> ---
>>  drivers/net/dsa/mv88e6xxx/chip.c | 29 +++++++++++++++++++++--------
>>  drivers/net/dsa/mv88e6xxx/port.c | 21 +++++++++++++++++++++
>>  drivers/net/dsa/mv88e6xxx/port.h |  2 ++
>>  3 files changed, 44 insertions(+), 8 deletions(-)
>> 
>> diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
>> index 01e4ac32d1e5..48e65f22641e 100644
>> --- a/drivers/net/dsa/mv88e6xxx/chip.c
>> +++ b/drivers/net/dsa/mv88e6xxx/chip.c
>> @@ -2689,15 +2689,20 @@ static int mv88e6xxx_setup_port(struct mv88e6xxx_chip *chip, int port)
>>  			return err;
>>  	}
>>  
>> -	/* 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.
>> +	/* Port Association Vector: disable automatic address learning
>> +	 * on all user ports since they start out in standalone
>> +	 * mode. When joining a bridge, learning will be configured to
>> +	 * match the bridge port settings. Enable learning on all
>> +	 * DSA/CPU ports. NOTE: FROM_CPU frames always bypass the
>> +	 * learning process.
>> +	 *
>> +	 * Disable HoldAt1, IntOnAgeOut, LockedPort, IgnoreWrongData,
>> +	 * and RefreshLocked. I.e. setup standard automatic learning.
>>  	 */
>> -	reg = 1 << port;
>> -	/* Disable learning for CPU port */
>> -	if (dsa_is_cpu_port(ds, port))
>> +	if (dsa_is_user_port(ds, port))
>>  		reg = 0;
>> +	else
>> +		reg = 1 << port;
>>  
>>  	err = mv88e6xxx_port_write(chip, port, MV88E6XXX_PORT_ASSOC_VECTOR,
>>  				   reg);
>
> Can this be refactored to use mv88e6xxx_port_set_assoc_vector too?

That was my initial thought as well. But this write also ensures that
the other learning related settings are in a known state. So I settled
for leaving the raw write, but I made sure to document that we depend on
the values of the other flags (second paragraph).

>> @@ -5426,7 +5431,7 @@ static int mv88e6xxx_port_pre_bridge_flags(struct dsa_switch *ds, int port,
>>  	struct mv88e6xxx_chip *chip = ds->priv;
>>  	const struct mv88e6xxx_ops *ops;
>>  
>> -	if (flags.mask & ~(BR_FLOOD | BR_MCAST_FLOOD))
>> +	if (flags.mask & ~(BR_LEARNING | BR_FLOOD | BR_MCAST_FLOOD))
>>  		return -EINVAL;
>>  
>>  	ops = chip->info->ops;
>> @@ -5449,6 +5454,14 @@ static int mv88e6xxx_port_bridge_flags(struct dsa_switch *ds, int port,
>>  
>>  	mv88e6xxx_reg_lock(chip);
>>  
>> +	if (flags.mask & BR_LEARNING) {
>> +		u16 pav = (flags.val & BR_LEARNING) ? (1 << port) : 0;
>> +
>> +		err = mv88e6xxx_port_set_assoc_vector(chip, port, pav);
>> +		if (err)
>> +			goto out;
>> +	}
>> +
>>  	if (flags.mask & BR_FLOOD) {
>>  		bool unicast = !!(flags.val & BR_FLOOD);
>>  
>> diff --git a/drivers/net/dsa/mv88e6xxx/port.c b/drivers/net/dsa/mv88e6xxx/port.c
>> index 4561f289ab76..d716cd61b6c6 100644
>> --- a/drivers/net/dsa/mv88e6xxx/port.c
>> +++ b/drivers/net/dsa/mv88e6xxx/port.c
>> @@ -1171,6 +1171,27 @@ int mv88e6097_port_egress_rate_limiting(struct mv88e6xxx_chip *chip, int port)
>>  				    0x0001);
>>  }
>>  
>> +/* Offset 0x0B: Port Association Vector */
>> +
>> +int mv88e6xxx_port_set_assoc_vector(struct mv88e6xxx_chip *chip, int port,
>> +				    u16 pav)
>> +{
>> +	u16 reg, mask;
>> +	int err;
>> +
>> +	err = mv88e6xxx_port_read(chip, port, MV88E6XXX_PORT_ASSOC_VECTOR,
>> +				  &reg);
>> +	if (err)
>> +		return err;
>> +
>> +	mask = GENMASK(mv88e6xxx_num_ports(chip), 0);
>
> mv88e6xxx_num_ports(chip) - 1, maybe?

Ahh thanks. This made me think that there should be a helper for this;
turns out Vivien added mv88e6xxx_port_mask four years ago :)

>> +	reg &= ~mask;
>> +	reg |= pav & mask;
>> +
>> +	return mv88e6xxx_port_write(chip, port, MV88E6XXX_PORT_ASSOC_VECTOR,
>> +				    reg);
>> +}
>> +
>>  /* Offset 0x0C: Port ATU Control */
>>  
>>  int mv88e6xxx_port_disable_learn_limit(struct mv88e6xxx_chip *chip, int port)
>> diff --git a/drivers/net/dsa/mv88e6xxx/port.h b/drivers/net/dsa/mv88e6xxx/port.h
>> index e6d0eaa6aa1d..635b6571a0e9 100644
>> --- a/drivers/net/dsa/mv88e6xxx/port.h
>> +++ b/drivers/net/dsa/mv88e6xxx/port.h
>> @@ -361,6 +361,8 @@ int mv88e6165_port_set_jumbo_size(struct mv88e6xxx_chip *chip, int port,
>>  				  size_t size);
>>  int mv88e6095_port_egress_rate_limiting(struct mv88e6xxx_chip *chip, int port);
>>  int mv88e6097_port_egress_rate_limiting(struct mv88e6xxx_chip *chip, int port);
>> +int mv88e6xxx_port_set_assoc_vector(struct mv88e6xxx_chip *chip, int port,
>> +				    u16 pav);
>>  int mv88e6097_port_pause_limit(struct mv88e6xxx_chip *chip, int port, u8 in,
>>  			       u8 out);
>>  int mv88e6390_port_pause_limit(struct mv88e6xxx_chip *chip, int port, u8 in,
>> -- 
>> 2.25.1
>>