On Mon, Apr 19, 2021 at 10:55:35AM +0200, Emanuele Giuseppe Esposito wrote:
> For simplicity, use bdrv_drained_begin/end to avoid concurrent
> writes to the write threshold, or reading it while it is being set.
> qmp_block_set_write_threshold is protected by the BQL.
> 
> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
> Co-developed-by: Paolo Bonzini <pbonzini@redhat.com>
> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
> ---
>  block/write-threshold.c | 9 +++------
>  1 file changed, 3 insertions(+), 6 deletions(-)
> 
> diff --git a/block/write-threshold.c b/block/write-threshold.c
> index 63040fa4f2..77c74bdaa7 100644
> --- a/block/write-threshold.c
> +++ b/block/write-threshold.c
> @@ -111,7 +111,6 @@ void qmp_block_set_write_threshold(const char *node_name,
>                                     Error **errp)
>  {
>      BlockDriverState *bs;
> -    AioContext *aio_context;
>  
>      bs = bdrv_find_node(node_name);
>      if (!bs) {
> @@ -119,10 +118,8 @@ void qmp_block_set_write_threshold(const char *node_name,
>          return;
>      }
>  
> -    aio_context = bdrv_get_aio_context(bs);
> -    aio_context_acquire(aio_context);
> -
> +    /* Avoid a concurrent write_threshold_disable.  */
> +    bdrv_drained_begin(bs);

Is there documentation that says it's safe to call
bdrv_drained_begin(bs) without the AioContext acquired? AIO_WAIT_WHILE()
contradicts this:

  The caller's thread must be the IOThread that owns @ctx or the main loop
  thread (with @ctx acquired exactly once).
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^