qcow2: Flush image after creation

Opening the qcow2 image with BDRV_O_NO_FLUSH prevents any flushes during
the image creation. This means that the image has not yet been flushed
to disk when qemu-img create exits. This flush is delayed until the next
operation on the image involving opening it without BDRV_O_NO_FLUSH and
closing (or directly flushing) it. For large images and/or images with a
small cluster size and preallocated metadata, this flush may take a
significant amount of time and may occur unexpectedly.

Reopening the image without BDRV_O_NO_FLUSH right before the end of
qcow2_create2() results in preponing the potentially costly flush into
the image creation, which is expected to take some time (whereas
successive image operations may be not).

Signed-off-by: Max Reitz <mreitz@redhat.com>
---
 block/qcow2.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

Am 23.10.2013 um 21:40 hat Max Reitz geschrieben:
> Opening the qcow2 image with BDRV_O_NO_FLUSH prevents any flushes during
> the image creation. This means that the image has not yet been flushed
> to disk when qemu-img create exits. This flush is delayed until the next
> operation on the image involving opening it without BDRV_O_NO_FLUSH and
> closing (or directly flushing) it. For large images and/or images with a
> small cluster size and preallocated metadata, this flush may take a
> significant amount of time and may occur unexpectedly.
> 
> Reopening the image without BDRV_O_NO_FLUSH right before the end of
> qcow2_create2() results in preponing the potentially costly flush into
> the image creation, which is expected to take some time (whereas
> successive image operations may be not).
> 
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
>  block/qcow2.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/block/qcow2.c b/block/qcow2.c
> index c1abaff..8b98c3a 100644
> --- a/block/qcow2.c
> +++ b/block/qcow2.c
> @@ -1584,7 +1584,15 @@ static int qcow2_create2(const char *filename, int64_t total_size,
>          }
>      }
>  
> -    ret = 0;

I would prefer to keep the explicit ret = 0 there (just like the
unnecessary last 'goto out:', it just makes things more obvious and
consistent)

> +    bdrv_close(bs);
> +
> +    /* Reopen the image without BDRV_O_NO_FLUSH to flush it before returning */
> +    ret = bdrv_open(bs, filename, NULL,
> +        BDRV_O_RDWR | BDRV_O_CACHE_WB, drv, &local_err);
> +    if (error_is_set(&local_err)) {
> +        error_propagate(errp, local_err);

So a goto here wouldn't hurt either. Note how the unnecessary goto in
the block before allowed you to just add your new code without modifying
existing parts.

> +    }
> +
>  out:
>      bdrv_unref(bs);
>      return ret;

Kevin

On 10/23/2013 08:40 PM, Max Reitz wrote:
> Opening the qcow2 image with BDRV_O_NO_FLUSH prevents any flushes during
> the image creation. This means that the image has not yet been flushed
> to disk when qemu-img create exits. This flush is delayed until the next
> operation on the image involving opening it without BDRV_O_NO_FLUSH and
> closing (or directly flushing) it. For large images and/or images with a
> small cluster size and preallocated metadata, this flush may take a
> significant amount of time and may occur unexpectedly.
> 
> Reopening the image without BDRV_O_NO_FLUSH right before the end of
> qcow2_create2() results in preponing the potentially costly flush into

s/preponing/hoisting/

> the image creation, which is expected to take some time (whereas
> successive image operations may be not).
> 
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
>  block/qcow2.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)

Reviewed-by: Eric Blake <eblake@redhat.com>

> -    ret = 0;
> +    bdrv_close(bs);
> +
> +    /* Reopen the image without BDRV_O_NO_FLUSH to flush it before returning */
> +    ret = bdrv_open(bs, filename, NULL,
> +        BDRV_O_RDWR | BDRV_O_CACHE_WB, drv, &local_err);

I would probably have indented as:

ret = bdrv_open(bs, filename, NULL,
                BDRV_O_RDWR | BDRV_O_CACHE_WB,
                drv, &local_err);

but it's trivial enough that I'm also fine with your choice.