On 11/04/2018 18:39, Kevin Wolf wrote:
> The much easier and more obviously correct way is to fundamentally
> change the way the functions work: Iterate over all BlockDriverStates,
> no matter who owns them, and drain them individually. Compensation is
> only necessary when a new BDS is created inside a drain_all section.
> Removal of a BDS doesn't require any action because it's gone afterwards
Ok, now I see (I think) why you chose the recursive check for in-flight
requests. The higher quiesce_count is not a problem, but I am still not
convinced about the recursion.
> This change means that some nodes get a higher bs->quiesce_count now
> because each node propagates its individual drain to all of its parents.
> In the old subtree drain, propagation back to the parent that made the
> recursive drain request was avoided. While this isn't perfectly
> beautiful, accepting the higher counts seems preferable to adding drain
> code to multiple other places that modify the graph.