Re: [Qemu-devel] [PATCH V4 4/7] qmp: Allow to change password on names block driver states.

[Fam Zheng <famz@redhat.com>]

On 2013年12月10日 23:16, Luiz Capitulino wrote:
> On Tue, 10 Dec 2013 15:25:07 +0100
> Kevin Wolf <kwolf@redhat.com> wrote:
>
>> My objection to your approach is strong because Benoît already sent an
>> alternative which I believe is less worse because with it, arguments
>> actually mean what their names tell instead of having additional bools
>> for "oh, and I said A, but I didn't mean it, I really want B".
>
> Current proposal:
>
> { 'command': 'block_passwd', 'data': {'*device': 'str',
>                                        '*node-name': 'str', 'password': 'str'} }
>

I vote for this.

> When I look at it, I ask myself:
>
>   - What happens when device=NULL?
>
>   - What happens when node-name=NULL?
>
>   - What happens when device=NULL and node-name=NULL?
>
>   - What happens when device != NULL and node-node != NULL?
>
>   - What happens when device != NULL but node-node=NULL?
>
>   - What happens when device=NULL but node-node != NULL?
>
> My proposal:
>
> { 'command': 'block_passwd', 'data': {'device': 'str',
>                                        '*device-is-node': 'bool', 'password': 'str'} }
>
>   - What happens when device-is-node=NULL?
>
>   - What happens when device-is-node != NULL?
>

I think our starting point is that device names and node names are 
separate name spaces. For this reason we should avoid connecting them by 
reusing one field. From a user's view, "device-is-node", or anything 
meaning this, just doesn't sound interesting, because we also told them 
"device-is-not-node".

Fam