Re: [PATCH v3 0/3] net: dsa: mv88e6xxx: add support for VLAN Table Unit

[Vivien Didelot <vivien.didelot@savoirfairelinux.com>]

Hi Andrew, Scott,

On Jul 6, 2015, at 11:46 PM, Scott Feldman sfeldma@gmail.com wrote:

> On Mon, Jul 6, 2015 at 7:00 PM, Andrew Lunn <andrew@lunn.ch> wrote:
>> On Tue, Jul 07, 2015 at 01:38:04AM +0200, Andrew Lunn wrote:
>>> On Sun, Jul 05, 2015 at 10:14:50PM -0400, Vivien Didelot wrote:
>>> > Hi all,
>>> >
>>> > This patchset brings full support for hardware VLANs in DSA, and the Marvell
>>> > 88E6xxx compatible switch chips.
>>>
>>> Hi Vivien
>>>
>>> I just booted these patches on my board, and i'm getting WARNINGS:
>>>
>>> [   61.111302] WARNING: CPU: 0 PID: 2751 at net/switchdev/switchdev.c:265
>>> switchdev_port_obj_add+0xd4/0xdc()
>>
>> Hi Vivien
>>
>> I debugged this a bit.
>>
>> The problem comes from:
>>
>> static int dsa_slave_port_obj_add(struct net_device *dev,
>>                                   struct switchdev_obj *obj)
>> {
>>         int err;
>>
>>         /*
>>          * Skip the prepare phase, since currently the DSA drivers don't need to
>>          * allocate any memory for operations and they will not fail to HW
>>          * (unless something horrible goes wrong on the MDIO bus, in which case
>>          * the prepare phase wouldn't have been able to predict anyway).
>>          */
>>         if (obj->trans != SWITCHDEV_TRANS_COMMIT)
>>                 return 0;
>>
>>         switch (obj->id) {
>>         case SWITCHDEV_OBJ_PORT_VLAN:
>>                 err = dsa_slave_port_vlans_add(dev, obj);
>>                 break;
>>         default:
>>                 err = -EOPNOTSUPP;
>>                 break;
>>         }
>>
>>         return err;
>> }
>>
>> It is being called with obj->id of 2, which is
>> SWITCHDEV_OBJ_IPV4_FIB. This function is called twice. The first time
>> it is with SWITCHDEV_TRANS_PREPARE and we are allowed to return an
>> error. The second time, with SWITCHDEV_TRANS_COMMIT, errors are not
>> allowed.
>>
>> EOPNOTSUPP is considered an error, so since we don't support
>> SWITCHDEV_OBJ_IPV4_FIB we error out the COMMIT phase.
>>
>> Not sure which is cleaner. Test to see if we support the object during
>> the prepare, or allow the commit to accept EOPNOTSUPP as not being an
>> error?
> 
> I think we should return EOPNOTSUPP on PREPARE, so move the trans !=
> COMMIT test inside the case for PORT_VLAN.  That would future-proof
> the func when new objects are added to switchdev (and not supported by
> dsa_slave).

Does this fixup http://ix.io/jxq look good to you?

Thanks,
-v