Re: [Qemu-devel] [PATCH v2 2/2] thread-pool: avoid deadlock in nested aio_poll() calls

[Paolo Bonzini <pbonzini@redhat.com>]

Il 15/07/2014 16:44, Stefan Hajnoczi ha scritto:
> The thread pool has a race condition if two elements complete before
> thread_pool_completion_bh() runs:
>
>   If element A's callback waits for element B using aio_poll() it will
>   deadlock since pool->completion_bh is not marked scheduled when the
>   nested aio_poll() runs.
>
> Fix this by marking the BH scheduled while thread_pool_completion_bh()
> is executing.  This way any nested aio_poll() loops will enter
> thread_pool_completion_bh() and complete the remaining elements.
>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
>  thread-pool.c | 6 ++++++
>  1 file changed, 6 insertions(+)
>
> diff --git a/thread-pool.c b/thread-pool.c
> index 4cfd078..23888dc 100644
> --- a/thread-pool.c
> +++ b/thread-pool.c
> @@ -185,6 +185,12 @@ restart:
>              QLIST_REMOVE(elem, all);
>              /* Read state before ret.  */
>              smp_rmb();
> +
> +            /* Schedule ourselves in case elem->common.cb() calls aio_poll() to
> +             * wait for another request that completed at the same time.
> +             */
> +            qemu_bh_schedule(pool->completion_bh);
> +
>              elem->common.cb(elem->common.opaque, elem->ret);
>              qemu_aio_release(elem);
>              goto restart;

Please make this "return" instead, so that it is clearer that the next 
invocation of the BH is a continuation of this one.

Paolo