RE: [PATCH V4 3/7] qapi/net: Add new QMP command for COLO passthrough

[Zhang, Chen]

> -----Original Message-----
> From: Markus Armbruster <armbru@redhat.com>
> Sent: Tuesday, April 6, 2021 4:01 PM
> 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 3/7] qapi/net: Add new QMP command for COLO
> passthrough
> 
> "Zhang, Chen" <chen.zhang@intel.com> writes:
> 
> >> -----Original Message-----
> >> From: Qemu-devel <qemu-devel-
> >> bounces+chen.zhang=intel.com@nongnu.org> On Behalf Of Markus
> >> Armbruster
> >> Sent: Tuesday, March 23, 2021 5:58 PM
> >> 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 3/7] qapi/net: Add new QMP command for COLO
> >> passthrough
> >>
> >> "Zhang, Chen" <chen.zhang@intel.com> writes:
> >>
> >> >> -----Original Message-----
> >> >> From: Markus Armbruster <armbru@redhat.com>
> >> [...]
> >> >> Now let's look at colo-passthrough-del.  I figure it is for
> >> >> deleting the kind of things colo-passthrough-add adds.
> >> >>
> >> >
> >> > Yes.
> >> >
> >> >> What exactly is deleted?  The thing created with the exact same
> >> arguments?
> >> >>
> >> >
> >> > Delete the rule from the module's private bypass list.
> >> > When user input a rule, the colo-passthrough-del will find the
> >> > specific module by the object ID, Then delete the rule.
> >> >
> >> >> This would be unusual.  Commonly, FOO-add and FOO-del both take a
> >> >> string ID argument.  The FOO created by FOO-add remembers its ID,
> >> >> and FOO-del deletes by ID.
> >> >
> >> > The ID not for rules itself, it just logged the modules(ID tagged)
> >> > affected by
> >> the rule.
> >>
> >> I'm not sure I understand.
> >>
> >> If you're pointing out that existing colo-passthrough-del parameter
> >> @id is not suitable for use as unique rule ID: you can always add
> >> another parameter that is suitable.
> >
> > Sorry to missed this mail.
> >
> > For example:
> > The VM running with filter-mirror(object id==0),
> > filter-redirector(object id==1) and colo-compare(object id==2), We use
> colo-passthrough-add/del to add/del a rule with a ID, if the ID==2, the rule
> just affect to colo-compare.
> > The filter-mirror and filter-redirector feel nothing after the add/del.
> 
> I think you're trying to explain existing parameter @id.  The point I was 
> trying
> to make is unrelated to this parameter, except by name collision.
> 
> My point is: our existing "delete" operations select the object to be deleted
> by some unique name that is assigned by the "add" operation.
> The unique name is a property of the object.  The property name is often,
> but not always "id".
> 
> Examples:
> 
>     device_add argument "id" sets the device's unique name.
>     device_del argument "id" selects the device to delete by its name.
> 
>     blockdev-add argument "node-name" sets the block backend device's
>     unique name.
>     blockdev-del argument "node-name" selects the block backend device
>     to delete by its name.
> 
> Is there any particular reason why deletion of your kind of object can't work
> the same way?

Current command can work in this way, It seems that name "ID" can be 
misunderstood.
The id=object0 is OK here. I will change the "id" to "object-name".
Thank you for clear the comments.

Thanks   
Chen