Re: [PATCH 1/6] block: add bitmap-populate job

[John Snow]

On 2/27/20 1:11 AM, Vladimir Sementsov-Ogievskiy wrote:
> 
> Still, if user pass disabled bitmap, it will be invalid immediately after
> job finish.

Yes ... In truth, I really want to augment this job to provide a defined
point-in-time semantic so it can be fully useful in all circumstances.

At the moment, the point-in-time it provides is "at finalize", which may
or may not be the single most useful semantic. (It matches mirror --
because that was easier.)

The other option is "at start" which matches backup, but will require a
new filter and some changes to permissions. That was more invasive, so I
avoided it for now.

(In the interest of a speedy resolution for libvirt, can we add a
critical_moment=[START|FINALIZE] option to this job *later*?)