Re: [PATCH V4 2/7] qapi/net.json: Add L4_Connection definition

From: Markus Armbruster <armbru@redhat.com>
To: "Zhang, Chen" <chen.zhang@intel.com>
Cc: Lukas Straub <lukasstraub2@web.de>,
	Li Zhijian <lizhijian@cn.fujitsu.com>,
	Jason Wang <jasowang@redhat.com>,
	qemu-dev <qemu-devel@nongnu.org>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Zhang Chen <zhangckid@gmail.com>
Subject: Re: [PATCH V4 2/7] qapi/net.json: Add L4_Connection definition
Date: Mon, 22 Mar 2021 13:31:03 +0100	[thread overview]
Message-ID: <87k0pz4bg8.fsf@dusky.pond.sub.org> (raw)
In-Reply-To: <5b75057ecc784296aa271f5f6692906a@intel.com> (Chen Zhang's message of "Mon, 22 Mar 2021 10:00:06 +0000")

"Zhang, Chen" <chen.zhang@intel.com> writes:

>> -----Original Message-----
>> From: Markus Armbruster <armbru@redhat.com>
>> Sent: Friday, March 19, 2021 11:48 PM
>> To: Zhang, Chen <chen.zhang@intel.com>
>> Cc: Jason Wang <jasowang@redhat.com>; qemu-dev <qemu-
>> devel@nongnu.org>; Eric Blake <eblake@redhat.com>; Dr. David Alan
>> Gilbert <dgilbert@redhat.com>; Markus Armbruster <armbru@redhat.com>;
>> Li Zhijian <lizhijian@cn.fujitsu.com>; Lukas Straub <lukasstraub2@web.de>;
>> Zhang Chen <zhangckid@gmail.com>
>> Subject: Re: [PATCH V4 2/7] qapi/net.json: Add L4_Connection definition
>> 
>> Zhang Chen <chen.zhang@intel.com> writes:
>> 
>> > Add L4_Connection struct for other QMP commands.
>> > Except protocol field is necessary, other fields are optional.
>> >
>> > Signed-off-by: Zhang Chen <chen.zhang@intel.com>
>> > ---
>> >  qapi/net.json | 26 ++++++++++++++++++++++++++
>> >  1 file changed, 26 insertions(+)
>> >
>> > diff --git a/qapi/net.json b/qapi/net.json index
>> > 498ea7aa72..cd4a8ed95e 100644
>> > --- a/qapi/net.json
>> > +++ b/qapi/net.json
>> > @@ -825,3 +825,29 @@
>> >  { 'enum': 'IP_PROTOCOL', 'data': [ 'tcp', 'udp', 'dccp', 'sctp', 'udplite',
>> >      'icmp', 'igmp', 'ipv6' ] }
>> >
>> > +##
>> > +# @L4_Connection:
>> > +#
>> > +# Layer 4 network connection.
>> > +#
>> > +# Just for IPv4.
>> > +#
>> > +# @protocol: Transport layer protocol like TCP/UDP...
>> > +#
>> > +# @id: For specific module with Qemu object ID, If there is no such part,
>> > +#      it means global rules.
>> 
>> Clear as mud.
>
> Sorry, let me re-clear it.
> If I understand correctly, The ID shouldn't be here, but I found the 'boxed' flag just can add only one 'data' like this:
> +##
> +{ 'command': 'colo-passthrough-add', 'boxed': true,
> +     'data': 'L4_Connection' }
>
> I original want to this:
> +##
> +{ 'command': 'colo-passthrough-add', 
> +     'data': { 'id': 'str', 'boxed': false, 'conn': 'L4_Connection', 'boxed': true  }
>
> So, I add the @id as an optional argument here.
>
> rewrite the comments:
> +# @id: Assign the rule to Qemu network handle module object ID. Like colo-compare, net-filter. 
>
> Please see the ID details in patch3 too. 

So, colo-passthrough-add takes an @id argument (to be tacked onto
packets to help with further processing, I understand), and arguments to
match packets.

Naming the argument type L4_Connection is misleading.

Even naming the match arguments L4_Connection would be misleading.
"Connection" has a specific meaning in networking.  There are TCP
connections.  There is no such thing as an UDP connection.

A TCP connection is uniquely identified by a pair of endpoints, i.e. by
source address, source port, destination address, destination port.
Same for other connection-oriented protocols.  The protocol is not part
of the connection.  Thus, L4_Connection would be misleading even for the
connection-oriented case.

You need a named type for colo-passthrough-add's argument because you
share it with colo-passthrough-del.  I'm not sure that's what we want
(I'm going to write more on that in a moment).  If it is what we want,
then please pick a another, descriptive name.