Re: [Qemu-devel] [PATCH v3 0/2] Fix slow startup with many disks

[Stefan Hajnoczi]

On Wed, Jun 17, 2015 at 01:37:18PM +0300, Alexander Yarygin wrote:
> Changes in v3:
>    - Added aio_context_acquire/aio_context_release around blk_drain() in
>      "virtio-blk: Use blk_drain() to drain IO requests" + updated commit
>      description
> 
> Please update Cc: qemu-stable@ if it necessarily.
> 
> Changes in v2:
>     - Patch "block-backend: Introduce blk_drain() and replace blk_drain_all()"
>       splitted to two.
>     - blk_drain() moved to above virtio_blk_data_plane_stop().
> 
> Previous thread:
> http://lists.gnu.org/archive/html/qemu-devel/2015-06/msg02786.html
> 
> During reset the aio_poll() function is called at least amount_of_disks^2 
> times:
> 
> for_each disk
>     virtio_blk_reset()
>         bdrv_drain_all()
>                 for_each disk
>                          aio_poll()
> 
> For example, startup with 1000 disks takes over 13 minutes.
> 
> Patches 1 and 2 removes inner loop by using bdrv_drain() instead
> of bdrv_drain_all(). bdrv_drain() works on one disk at time.
> 
> Since bdrv_drain_all() is still called in other places, patch 3 optimizes
> it for cases, where there are more disks than iothreads.
> 
> Thanks.
> 
> Alexander Yarygin (2):
>   block-backend: Introduce blk_drain()
>   virtio-blk: Use blk_drain() to drain IO requests
> 
>  block/block-backend.c          |  5 +++++
>  hw/block/virtio-blk.c          | 15 ++++++++++-----
>  include/sysemu/block-backend.h |  1 +
>  3 files changed, 16 insertions(+), 5 deletions(-)
> 
> -- 
> 1.9.1
> 

Thanks, applied to my block tree:
https://github.com/stefanha/qemu/commits/block

Stefan

pgpAMQAUMB1_U.pgp
Description: PGP signature