Re: [RFC PATCH 0/4] block layer: split block APIs in graph and I/O

[Stefan Hajnoczi]

On Wed, Sep 15, 2021 at 02:11:41PM +0200, Paolo Bonzini wrote:
> On 13/09/21 15:10, Stefan Hajnoczi wrote:
> > On Wed, Sep 08, 2021 at 09:10:17AM -0400, Emanuele Giuseppe Esposito wrote:
> > > Currently, block layer APIs like block-backend.h contain a mix of
> > > functions that are either running in the main loop and under the
> > > BQL, or are thread-safe functions and run in iothreads performing I/O.
> > > The functions running under BQL also take care of modifying the
> > > block graph, by using drain and/or aio_context_acquire/release.
> > > This makes it very confusing to understand where each function
> > > runs, and what assumptions it provided with regards to thread
> > > safety.
> > > 
> > > We call the functions running under BQL "graph API", and
> > > distinguish them from the thread-safe "I/O API".
> > 
> > Maybe "BQL" is clearer than "graph" because not all functions classified
> > as "graph" need to traverse/modify the graph.
> 
> Bikeshedding, I like it! :)
> 
> ... on the other hand, qemu-storage-daemon does not have a BQL (see patch
> 1); "graph API" functions run from the main (monitor) thread.
> 
> The characteristic of the "graph API" is that they affect global state, so
> another possibility could be "global state API".  But is there any global
> state apart from the BlockDriverState graph and the associated
> BlockBackends?

I would be happy with that name too.

Stefan

signature.asc
Description: PGP signature