Re: [Qemu-devel] Loadable block drivers?

[Fam Zheng <famz@redhat.com>]

On Thu, 04/05 11:40, Lindsay Mathieson wrote:
> On 4 April 2018 at 23:41, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> 
> > On Tue, Apr 03, 2018 at 11:30:33AM +0800, Fam Zheng wrote:
> > > On Tue, 04/03 13:17, Lindsay Mathieson wrote:
> > > > On 3 April 2018 at 13:11, Fam Zheng <famz@redhat.com> wrote:
> > > >
> > > > > On Tue, 04/03 12:59, Lindsay Mathieson wrote:
> > > > > > Hi all, was looking at developing a block driver for qemu - have
> > examined
> > > > > > the drivers at:
> > > > > >
> > > > > >   https://github.com/qemu/qemu/tree/master/block
> > > > > >
> > > > > > And it seems straightforward enough.
> > > > > >
> > > > > > One thing that is unclear - all the drivers appear to be compiled
> > > > > directly
> > > > > > into qemu. Is there no way to load them dynamically as .so modules?
> > > > >
> > > > > './configure --enable-modules' will enable building block drivers as
> > .so
> > > > > objects, and they are loaded dynamically. These are in-tree .so
> > modules;
> > > > > out-of-tree modules like in Linux kernel are intentionally forbidden.
> > > > >
> > > > > Fam
> > > > >
> > > >
> > > >
> > > >
> > > > Rats, I take it that means I can't develop a testing block module and
> > load
> > > > it with an pre-existing qemu install.
> > >
> > > No, that's not possible.
> >
> > Depending on what you are trying to do, you could use the blkdebug,
> > null-co, NBD, or iSCSI drivers to perform your testing.
> >
> > blkdebug does fault injection (e.g. you can test what happens when
> > certain I/O requests fail).
> >
> > null-co is a nop block driver useful for some types of performance
> > testing and it also supports introducing an artificial delays.
> >
> > NBD and iSCSI can be used to forward I/O requests to an external server
> > where you can implement any behavior you want.
> >
> > We can discuss it more if you can explain what you're trying to do.
> >
> > Stefan
> >
> 
> 
> 
> Thanks Stefan, looking to develop a lizardfs block driver. A process that
> only involved building a module rather than the entire qemu tree would mike
> life easier, especially if I could test it on a live system (proxmox
> cluster). A custom qemu install is not an option for that.

I cannot think of a way to hot plugging a block driver to a running QEMU, but
perhaps you can live migrate the VM from the stock QEMU to a custom built one to
achieve similar.

Fam