dsa: mv88e6xxx: mv88e6321 rsvd2cpu

dsa: mv88e6xxx: mv88e6321 rsvd2cpu

[Angelo Dureghello]

Hi all,

working on some rstp stuff at userspace level, (kernel stp
disabled), i have seen bpdu packets was forwarded over bridge
ports generating chaos and loops. As far as i know, 802.1d asks
bpdus are not forwarded.

Finally found a solution, adding mv88e6185_g2_mgmt_rsvd2cpu()
for mv88e6321.
Is it a proper solution ? Is there any specific reason why
rsvd2cpu was not implemented for my 6321 ?
I can send the patch i am using actually. in case.

Thanks a lot.

Regards,
-- 
angelo dureghello

Re: dsa: mv88e6xxx: mv88e6321 rsvd2cpu

[Vladimir Oltean]

Hi Angelo,

On Wed, Apr 26, 2023 at 10:12:28AM +0200, Angelo Dureghello wrote:
> Hi all,
> 
> working on some rstp stuff at userspace level, (kernel stp
> disabled), i have seen bpdu packets was forwarded over bridge
> ports generating chaos and loops. As far as i know, 802.1d asks
> bpdus are not forwarded.
> 
> Finally found a solution, adding mv88e6185_g2_mgmt_rsvd2cpu()
> for mv88e6321.
> Is it a proper solution ? Is there any specific reason why
> rsvd2cpu was not implemented for my 6321 ?
> I can send the patch i am using actually. in case.

I don't know why mv88e6321_ops doesn't define a mgmt_rsvd2cpu()
implementation. I would agree that it is necessary, if available.

Re: dsa: mv88e6xxx: mv88e6321 rsvd2cpu

[Andrew Lunn]

On Wed, Apr 26, 2023 at 10:12:28AM +0200, Angelo Dureghello wrote:
> Hi all,
> 
> working on some rstp stuff at userspace level, (kernel stp
> disabled), i have seen bpdu packets was forwarded over bridge
> ports generating chaos and loops. As far as i know, 802.1d asks
> bpdus are not forwarded.
> 
> Finally found a solution, adding mv88e6185_g2_mgmt_rsvd2cpu()
> for mv88e6321.
> Is it a proper solution ? Is there any specific reason why
> rsvd2cpu was not implemented for my 6321 ?
> I can send the patch i am using actually. in case.

Should it be mv88e6185_g2_mgmt_rsvd2cpu() or
mv88e6352_g2_mgmt_rsvd2cpu()? Does the 6321 only support
01:80:c2:00:00:0x or does it have 01:80:c2:00:00:2x as well?

Please do send a patch, and include a Fixes: tag.

       Andrew

Re: dsa: mv88e6xxx: mv88e6321 rsvd2cpu

[Angelo Dureghello]

Hi Andrew,

On 26/04/23 3:35 PM, Andrew Lunn wrote:
> On Wed, Apr 26, 2023 at 10:12:28AM +0200, Angelo Dureghello wrote:
>> Hi all,
>>
>> working on some rstp stuff at userspace level, (kernel stp
>> disabled), i have seen bpdu packets was forwarded over bridge
>> ports generating chaos and loops. As far as i know, 802.1d asks
>> bpdus are not forwarded.
>>
>> Finally found a solution, adding mv88e6185_g2_mgmt_rsvd2cpu()
>> for mv88e6321.
>> Is it a proper solution ? Is there any specific reason why
>> rsvd2cpu was not implemented for my 6321 ?
>> I can send the patch i am using actually. in case.
> 
> Should it be mv88e6185_g2_mgmt_rsvd2cpu() or
> mv88e6352_g2_mgmt_rsvd2cpu()? Does the 6321 only support
> 01:80:c2:00:00:0x or does it have 01:80:c2:00:00:2x as well?
> 
> Please do send a patch, and include a Fixes: tag.
> 

you are right, it supports also the 01:80:c2:00:00:2x,
will test that too, and will send patch v2.

>         Andrew


Thanks,
angelo

Re: dsa: mv88e6xxx: mv88e6321 rsvd2cpu

[Vladimir Oltean]

On Wed, Apr 26, 2023 at 03:59:39PM +0200, Angelo Dureghello wrote:
> will test that too, and will send patch v2.

did you send v1?