Re: [PATCH 10/12] block.c: add subtree_drains where needed

[Emanuele Giuseppe Esposito]

On 02/02/2022 18:38, Paolo Bonzini wrote:
> On 2/2/22 16:37, Emanuele Giuseppe Esposito wrote:
>> So we have disk B with backing file C, and new disk A that wants to have
>> backing file C.
>>
>> I think I understand what you mean, so in theory the operation would be
>> - create new child
>> - add child to A->children list
>> - add child to C->parents list
>>
>> So in theory we need to
>> * drain A (without subtree), because it can't happen that child nodes of
>>    A have in-flight requests that look at A status (children list),
>> right?
>>    In other words, if A has another node X, can a request in X inspect
>>    A->children

I am not sure if this can happen. It doesn't seem so, though. All
functions inspecting ->parents are either GS or don't call recursive
function that go down on children list again.

>> * drain C, as parents can inspect C status (like B). Same assumption
>>    here, C->children[x]->bs cannot have requests inspecting C->parents
>>    list?

What if C->children[x]->bs drains? we would have a child inspecting
C->parents.

> 
> In that case (i.e. if parents have to be drained, but children need not)
> bdrv_drained_begin_unlocked would be enough, right?

Should be, if that is the case.

> 
> That would mean that ->children is I/O state but ->parents is global
> state.  I think it's quite a bit more complicated to analyze and to
> understand.

Not sure I follow this one, why is ->parents GS? it is also used by the
drain API...

Emanuele