Re: [Qemu-devel] [PATCH] qcow2: Fix error handling for immediate backing file read failure

[Markus Armbruster]

Kevin Wolf <address@hidden> writes:

> Requests could return success even though they failed when bdrv_aio_readv
> returned NULL for a backing file read.
>
> Reported-by: Chunqiang Tang <address@hidden>
> Signed-off-by: Kevin Wolf <address@hidden>
> ---
>  block/qcow2.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/block/qcow2.c b/block/qcow2.c
> index 28338bf..647c2a4 100644
> --- a/block/qcow2.c
> +++ b/block/qcow2.c
> @@ -479,8 +479,10 @@ static void qcow2_aio_read_cb(void *opaque, int ret)
>                  BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO);
>                  acb->hd_aiocb = bdrv_aio_readv(bs->backing_hd, 
> acb->sector_num,
>                                      &acb->hd_qiov, n1, qcow2_aio_read_cb, 
> acb);
> -                if (acb->hd_aiocb == NULL)
> +                if (acb->hd_aiocb == NULL) {
> +                    ret = -EIO;
>                      goto done;
> +                }
>              } else {
>                  ret = qcow2_schedule_bh(qcow2_aio_read_bh, acb);
>                  if (ret < 0)

Good catch.

Nearby

        if (qcow2_decompress_cluster(bs, acb->cluster_offset) < 0)
            goto done;

looks suspicious to my ignorant eyes.  Is it okay?