Re: [Qemu-devel] [PATCH v4 00/11] convert savevm to use qapi and introduce qmp command

[Luiz Capitulino]

On Wed, 10 Apr 2013 15:22:49 +0200
Pavel Hrdina <address@hidden> wrote:

> On 10.4.2013 14:49, Eric Blake wrote:
> > On 04/10/2013 06:40 AM, Luiz Capitulino wrote:
> >> On Wed, 10 Apr 2013 06:24:11 -0600
> >> Eric Blake <address@hidden> wrote:
> >>
> >>>>    - If you want to overwrite an existing snapshot, you could specify
> >>>>      the 'id' or the 'name' argument or both of them and also you will
> >>>>      have to use the 'force' argument
> >>>
> >>> But the argument made in this thread is that QMP should _not_ have a
> >>> force argument.  It should be a flat-out error in QMP to try to create a
> >>> snapshot with a conflicting name or tag; preferably with a distinct
> >>> error type.  Higher-level apps (HMP savevm -f) would try to create; if
> >>> -f is not specified, the error is good enough; if -f is specified and
> >>> that particular error is returned, then HMP calls delete and then
> >>> re-tries the create.  No 'force' argument needed at the QMP layer.
> >>
> >> To avoid adding a new error class, the HMP command could query for the
> >> snapshot name and delete it if it exists before creating the snapshot.
> >
> > Atomic collision detection is nicer than having to pre-query - fewer QMP
> > calls in the common case.  But you're right that none of the existing
> > ErrorClass categories fit the idea of "creation refused because name
> > would collide".

We can add a new error class if it's important to be atomic. I don't think
this is hugely important though, first because we don't support multiple
QMP connections (it should work, but we don't advertise it as supported) and
also because I don't think we have ever designed commands with this in mind.

> We still could have the force parameter for HMP because the QMP command 
> will return an error message that the snapshot already exist. The only 
> drawback will be that the error message will not contain an information 
> that you could use a '-f' parameter for override.

You're right, but we're discussing how the HMP command could detect that
an existing snapshot should be deleted.