Re: [RFC 0/4] mirror: implement incremental and bitmap modes

[Fiona Ebner]

Am 01.03.24 um 15:14 schrieb Vladimir Sementsov-Ogievskiy:
> 
> As we already understood, (block-)job-api needs some spring-cleaning.
> Unfortunately I don't have much time on it, but still I decided to start
> from finally depreacting block-job-* API and moving to job-*.. Probably
> bitmap/bitmap-mode/sync APIs also need some optimization, keeping in
> mind new block-dirty-bitmap-merge api.
> 
> So, what I could advice in this situation for newc interfaces:
> 
> 1. be minimalistic
> 2. add `x-` prefix when unsure
> 
> So, following these two rules, what about x-bitmap field, which may be
> combined only with 'full' mode, and do what you need?
> 

AFAIU, it should rather be marked as @unstable in QAPI [0]? Then it
doesn't need to be renamed if it becomes stable later.

> About documentation: actually, I never liked that we use for backup job
> "MirrorSyncMode". Now it looks more like "BackupSyncMode", having two
> values supported only by backup.
> 
> I'm also unsure how mode=full&bitmap=some_bitmap differs from
> mode=bitmap&bitmap=some_bitmap..
> 

With the current patches, it was an error to specify @bitmap for other
modes than 'incremental' and 'bitmap'.

> So, I'd suggest simply rename MirrorSyncMode to BackupSyncMode, and add
> separate MirrorSyncMode with only "full", "top" and "none" values.
> 

Sounds good to me!

[0]:
https://gitlab.com/qemu-project/qemu/-/commit/a3c45b3e62962f99338716b1347cfb0d427cea44

Best Regards,
Fiona