[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [RFC 08/13] block: drop bdrv_get_aio_context()
From: |
Stefan Hajnoczi |
Subject: |
Re: [Qemu-devel] [RFC 08/13] block: drop bdrv_get_aio_context() |
Date: |
Mon, 17 Jun 2013 16:57:57 +0200 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Fri, Jun 14, 2013 at 10:12:00AM -0400, Paolo Bonzini wrote:
> Il 14/06/2013 05:48, Stefan Hajnoczi ha scritto:
> > Associating a BlockDriverState with a single AioContext is not flexible
> > enough. Once we make BlockDriverState thread-safe, it will be possible
> > to call bdrv_*() functions from multiple event loops.
>
> I'm afraid that this is trading some pain now (converting
> qemu_bh_new/qemu_set_fd_handler to aio_bh_new/aio_set_fd_handler) for
> more pain later (having to make BDS thread-safe). There aren't that
> many (~40) in block layer code.
>
> Making BlockDriverState thread-safe is hard, it is much simpler to run
> all the BlockDriverState code in the AioContext thread itself.
>
> There are some things that cannot (basically monitor commands and other
> places that are currently using bdrv_drain_all) but they can simply take
> a "big AioContext lock".
I don't like a big AioContext lock that stops the event loop because we
need to support a M:N device:iothread model where stopping an event loop
means artificially stopping other devices too.
Maybe it's workable though since the only commands that would take an
AioContext lock are rare and shouldn't impact performance. I guess then
it comes down to robustness if a hung NFS mount can affect the entire VM
instead of just hanging an emulated disk device.
Still, I have tried the lock approach and agree it is hard. I'll
experiment some more and hopefully come up with something that handles
block jobs and the run-time NBD server.
Stefan
- [Qemu-devel] [RFC 02/13] dataplane: sync virtio.c and vring.c virtqueue state, (continued)
- [Qemu-devel] [RFC 02/13] dataplane: sync virtio.c and vring.c virtqueue state, Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 03/13] block: add BlockDevOps->drain_threads_cb(), Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 04/13] virtio-blk: implement BlockDevOps->drain_threads_cb(), Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 06/13] qemu-thread: add TLS wrappers, Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 07/13] block: add thread_aio_context TLS variable, Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 09/13] main-loop: use thread-local AioContext, Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 10/13] block: disable I/O throttling outside main loop, Stefan Hajnoczi, 2013/06/14
- [Qemu-devel] [RFC 08/13] block: drop bdrv_get_aio_context(), Stefan Hajnoczi, 2013/06/14
[Qemu-devel] [RFC 12/13] dataplane: drop ioq Linux AIO request queue, Stefan Hajnoczi, 2013/06/14
[Qemu-devel] [RFC 11/13] dataplane: use block layer for I/O, Stefan Hajnoczi, 2013/06/14
[Qemu-devel] [RFC 13/13] block: drop raw_get_aio_fd(), Stefan Hajnoczi, 2013/06/14
[Qemu-devel] [RFC 05/13] exec: do not use qemu/tls.h, Stefan Hajnoczi, 2013/06/14