On 16/12/2021 15:57, Hanna Reitz wrote:
On 24.11.21 07:43, Emanuele Giuseppe Esposito wrote:
According to the assertions put in the previous patch, we should
first drain and then modify the ->children list. In this way
we prevent other iothreads to read the list while it is being
updated.
In this case, moving the drain won't cause any harm, because
child is a parameter of the drain function so it will still be
included in the operation, despite not being in the list.
Sounds good.
Uhm.. I don't think this is useful at all. I was thinking to drop this
patch in v6.
My plans on subtree drains, ->children and ->parent lists are
explained in "Removal of Aiocontext lock and usage of subtree drains
in aborted transactions"
20211213104014.69858-1-eesposit@redhat.com/">https://patchew.org/QEMU/20211213104014.69858-1-eesposit@redhat.com/
And as I say there, commit d736f119da makes sure that it is safe to
modify the graph even side a bdrv_subtree_drained_begin/end() section.
So this should be unnecessary.