Any comments?

On Thu, 28 Jul 2016 13:32:27 +0200
David Sterba [off-list ref] wrote:

I'll comment on the overall approach and skip code-specific comments.

The changelog does not explain why there's a need for a new blockgroup
type and what's the relation to the existing types. It seems that it
extends the data/metadata/system group, but I think this is totally
wrong.  

I agree in principle, but I did not want to modify the existing balance
targets, but, instead, piggyback on the existing balance implementation
to re-balance the data.
This approach was recommended to be by an experienced BTrFS developer
on the IRC as the right way to implement the change. My previous
implementation before asking on the IRC used a new ioctl call to
change the hard coded values and then re-write the data (not a good
approach in hindsight.)
 

The proposed changes modify part of the on-disk format, that would
require a incompat bit and brings the usual load of unpleasant issues
with backward compatibility. The current data structures should be
enough for configurable stripe size.  If you want to make stripe size
configurable, then replace all hardcoded values of BTRFS_STRIPE_LEN.  

No re-balance required after passing the stripe size change command?
What about the on-disk metadata, that relies on the "stripesize" and
"stripe_len" as variables for calculations and the basis of pre-set
metadata?

Thanks
Sanidhya