Re: [PATCH RFC 1/2] net: dsa: integrate with SWITCHDEV for HW bridging

[Guenter Roeck <linux@roeck-us.net>]

On 02/17/2015 12:24 PM, Florian Fainelli wrote:
> On 17/02/15 12:13, Guenter Roeck wrote:
>> On 02/17/2015 11:26 AM, Florian Fainelli wrote:
>>> In order to support bridging offloads in DSA switch drivers, select
>>> NET_SWITCHDEV to get access to the port_stp_update and parent_get_id
>>> NDOs that we are required to implement.
>>>
>>> To facilitate the integratation at the DSA driver level, we implement 3
>>> types of operations:
>>>
>>> - port_join_bridge
>>> - port_leave_bridge
>>> - port_stp_update
>>>
>>> DSA will resolve which switch ports that are currently bridge port
>>> members as some Switch hardware/drivers need to know about that to limit
>>> the register programming to just the relevant registers (especially for
>>> slow MDIO buses).
>>>
>>> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
>>> ---
>>>    include/net/dsa.h  |  10 +++++
>>>    net/dsa/Kconfig    |   1 +
>>>    net/dsa/dsa.c      |   7 ++++
>>>    net/dsa/dsa_priv.h |   2 +
>>>    net/dsa/slave.c    | 117
>>> +++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>    5 files changed, 137 insertions(+)
>>>
>>> diff --git a/include/net/dsa.h b/include/net/dsa.h
>>> index ed3c34bbb67a..92be34791963 100644
>>> --- a/include/net/dsa.h
>>> +++ b/include/net/dsa.h
>>> @@ -275,6 +275,16 @@ struct dsa_switch_driver {
>>>        int    (*get_regs_len)(struct dsa_switch *ds, int port);
>>>        void    (*get_regs)(struct dsa_switch *ds, int port,
>>>                    struct ethtool_regs *regs, void *p);
>>> +
>>> +    /*
>>> +     * Bridge integration
>>> +     */
>>> +    int    (*port_join_bridge)(struct dsa_switch *ds, int port,
>>> +                    u32 br_port_mask);
>>> +    int    (*port_leave_bridge)(struct dsa_switch *ds, int port,
>>> +                     u32 br_port_mask);
>>> +    int    (*port_stp_update)(struct dsa_switch *ds, int port,
>>> +                   u8 state);
>>>    };
>>>
>>>    void register_switch_driver(struct dsa_switch_driver *type);
>>> diff --git a/net/dsa/Kconfig b/net/dsa/Kconfig
>>> index 5f8ac404535b..b45206e8dd3e 100644
>>> --- a/net/dsa/Kconfig
>>> +++ b/net/dsa/Kconfig
>>> @@ -8,6 +8,7 @@ config NET_DSA
>>>        tristate
>>>        depends on HAVE_NET_DSA
>>>        select PHYLIB
>>> +    select NET_SWITCHDEV
>>
>> Should this be "select" or "depends on" ?
>>
>> Downside of depends is that we'll need some ifdefs in the code,
>> but on the other side it would let people disable it if it is
>> not needed.
>
> The code overhead is not huge, and I would think that by enforcing
> NET_SWITCHDEV we encourage better DSA driver practices and promote HW
> bridging, if you think this should be made conditional, I guess we can
> do that.
>

For sure not me ... I am happy forcing it. In my use case it would
always be enabled. I just don't want to go too far along that route
just to have to add a bunch of ifdefs later on.

Guenter