[MPTCP] Re: [PATCH v4 7/9] mptcp: add netlink based PM

[MPTCP] Re: [PATCH v4 7/9] mptcp: add netlink based PM

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 622 bytes --]

On Mon, 2020-02-24 at 16:51 +0100, Paolo Abeni wrote:
> +enum {
> +	MPTCP_PM_CMD_UNSPEC,
> +
> +	MPTCP_PM_CMD_ADD_ADDR,
> +	MPTCP_PM_CMD_DEL_ADDR,
> +	MPTCP_PM_CMD_GET_ADDR,
> +	MPTCP_PM_CMD_FLUSH_ADDRS,
> +	MPTCP_PM_CMD_SET_RCV_ADD_ADDRS,
> +	MPTCP_PM_CMD_GET_RCV_ADD_ADDRS,

To enforce the server subflow limits, the idea would be to change (in a
separated series) the above 2 commands to:
	MPTCP_PM_CMD_SET_LIMITS,
	MPTCP_PM_CMD_GET_LIMITS,

and add a new attribute for the max subflow limits

Plus the core infrastructure/hooks to let the PM really enforce such
limit.

WDYT?

Cheers,

Paolo

[MPTCP] Re: [PATCH v4 7/9] mptcp: add netlink based PM

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1862 bytes --]

Hi Paolo,

On 26/02/2020 13:05, Paolo Abeni wrote:
> On Wed, 2020-02-26 at 12:44 +0100, Matthieu Baerts wrote:
>> Hi Paolo,
>>
>> On 26/02/2020 12:36, Paolo Abeni wrote:
>>> On Mon, 2020-02-24 at 16:51 +0100, Paolo Abeni wrote:
>>>> +enum {
>>>> +	MPTCP_PM_CMD_UNSPEC,
>>>> +
>>>> +	MPTCP_PM_CMD_ADD_ADDR,
>>>> +	MPTCP_PM_CMD_DEL_ADDR,
>>>> +	MPTCP_PM_CMD_GET_ADDR,
>>>> +	MPTCP_PM_CMD_FLUSH_ADDRS,
>>>> +	MPTCP_PM_CMD_SET_RCV_ADD_ADDRS,
>>>> +	MPTCP_PM_CMD_GET_RCV_ADD_ADDRS,
>>>
>>> To enforce the server subflow limits, the idea would be to change (in a
>>> separated series) the above 2 commands to:
>>> 	MPTCP_PM_CMD_SET_LIMITS,
>>> 	MPTCP_PM_CMD_GET_LIMITS,
>>>
>>> and add a new attribute for the max subflow limits
>>>
>>> Plus the core infrastructure/hooks to let the PM really enforce such
>>> limit.
>>>
>>> WDYT?
>>
>> That sounds good to me and more "natural". Thank you for looking at this!
>>
>> Just one question: would they be generic commands to add/get limits? Or
>> should we set a more specific name, e.g.:
>>
>>     MPTCP_PM_CMD_SET_MAX_SUBFLOWS
>>     MPTCP_PM_CMD_GET_MAX_SUBFLOWS
> 
> Before doing the actual coding, I think the first option would lead to
> smaller/simpler implementation (no duplicate boilerplates) and should
> be also more easily extendible for the same reason.

I guess you will have MPTCP_PM_CMD_SET_LIMITS only accepting 
MPTCP_PM_LIMITS_ATTR_MAX_SUBFLOW for the moment.

It can indeed be extended later.

> But I guess is mostly a matter of taste, and I know I'm moslty wrong on
> that kind of topics - except for important subjects like pizza and
> coffee ;)

🍍

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium

[MPTCP] Re: [PATCH v4 7/9] mptcp: add netlink based PM

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1449 bytes --]

On Wed, 2020-02-26 at 12:44 +0100, Matthieu Baerts wrote:
> Hi Paolo,
> 
> On 26/02/2020 12:36, Paolo Abeni wrote:
> > On Mon, 2020-02-24 at 16:51 +0100, Paolo Abeni wrote:
> > > +enum {
> > > +	MPTCP_PM_CMD_UNSPEC,
> > > +
> > > +	MPTCP_PM_CMD_ADD_ADDR,
> > > +	MPTCP_PM_CMD_DEL_ADDR,
> > > +	MPTCP_PM_CMD_GET_ADDR,
> > > +	MPTCP_PM_CMD_FLUSH_ADDRS,
> > > +	MPTCP_PM_CMD_SET_RCV_ADD_ADDRS,
> > > +	MPTCP_PM_CMD_GET_RCV_ADD_ADDRS,
> > 
> > To enforce the server subflow limits, the idea would be to change (in a
> > separated series) the above 2 commands to:
> > 	MPTCP_PM_CMD_SET_LIMITS,
> > 	MPTCP_PM_CMD_GET_LIMITS,
> > 
> > and add a new attribute for the max subflow limits
> > 
> > Plus the core infrastructure/hooks to let the PM really enforce such
> > limit.
> > 
> > WDYT?
> 
> That sounds good to me and more "natural". Thank you for looking at this!
> 
> Just one question: would they be generic commands to add/get limits? Or 
> should we set a more specific name, e.g.:
> 
>    MPTCP_PM_CMD_SET_MAX_SUBFLOWS
>    MPTCP_PM_CMD_GET_MAX_SUBFLOWS

Before doing the actual coding, I think the first option would lead to
smaller/simpler implementation (no duplicate boilerplates) and should
be also more easily extendible for the same reason.

But I guess is mostly a matter of taste, and I know I'm moslty wrong on
that kind of topics - except for important subjects like pizza and
coffee ;)

/P

[MPTCP] Re: [PATCH v4 7/9] mptcp: add netlink based PM

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 1152 bytes --]

Hi Paolo,

On 26/02/2020 12:36, Paolo Abeni wrote:
> On Mon, 2020-02-24 at 16:51 +0100, Paolo Abeni wrote:
>> +enum {
>> +	MPTCP_PM_CMD_UNSPEC,
>> +
>> +	MPTCP_PM_CMD_ADD_ADDR,
>> +	MPTCP_PM_CMD_DEL_ADDR,
>> +	MPTCP_PM_CMD_GET_ADDR,
>> +	MPTCP_PM_CMD_FLUSH_ADDRS,
>> +	MPTCP_PM_CMD_SET_RCV_ADD_ADDRS,
>> +	MPTCP_PM_CMD_GET_RCV_ADD_ADDRS,
> 
> To enforce the server subflow limits, the idea would be to change (in a
> separated series) the above 2 commands to:
> 	MPTCP_PM_CMD_SET_LIMITS,
> 	MPTCP_PM_CMD_GET_LIMITS,
> 
> and add a new attribute for the max subflow limits
> 
> Plus the core infrastructure/hooks to let the PM really enforce such
> limit.
> 
> WDYT?

That sounds good to me and more "natural". Thank you for looking at this!

Just one question: would they be generic commands to add/get limits? Or 
should we set a more specific name, e.g.:

   MPTCP_PM_CMD_SET_MAX_SUBFLOWS
   MPTCP_PM_CMD_GET_MAX_SUBFLOWS

Cheers,
Matt
-- 
Matthieu Baerts | R&D Engineer
matthieu.baerts(a)tessares.net
Tessares SA | Hybrid Access Solutions
www.tessares.net
1 Avenue Jean Monnet, 1348 Louvain-la-Neuve, Belgium