Re: [Qemu-devel] [PATCH V11 05/17] block: add snapshot info query function bdrv_query_snapshot_info_list()

[Markus Armbruster]

Eric Blake <address@hidden> writes:

> On 04/02/2013 05:47 AM, Wenchao Xia wrote:
>>   This patch adds function bdrv_query_snapshot_info_list(), which will
>> retrieve snapshot info of an image in qmp object format. The implementation
>> is based on the code moved from qemu-img.c with modification to fit more
>> for qmp based block layer API.
>> 
>> Signed-off-by: Wenchao Xia <address@hidden>
>> ---
>>  block/qapi.c         |   55 
>> ++++++++++++++++++++++++++++++++++++++-----------
>>  include/block/qapi.h |    4 ++-
>>  qemu-img.c           |    5 +++-
>>  3 files changed, 49 insertions(+), 15 deletions(-)
>
>> +/*
>> + * return 0 on success, @p_list will be set only on success, and caller 
>> need to
>
> s/need/needs/
>
>> + * check *p_list on success.
>
> I wonder if this wording would be any better:
>
> Returns 0 on success, with *p_list either set to describe snapshot
> information, or NULL because there are no snapshots.  Returns -1 on
> error, with *p_list untouched.

It actually returns -errno then, doesn't it?

>
>> + */
>> +int bdrv_query_snapshot_info_list(BlockDriverState *bs,
>> +                                  SnapshotInfoList **p_list,
>> +                                  Error **errp)
>>  {
>
> At any rate, my only commentary was on grammar and a possible wording
> for a comment, while the code itself is fine from my viewpoint; so feel
> free to add:
>
> Reviewed-by: Eric Blake <address@hidden>
>
>> +++ b/qemu-img.c
>> @@ -1735,7 +1735,10 @@ static ImageInfoList
>> *collect_image_info_list(const char *filename,
>>  
>>          info = g_new0(ImageInfo, 1);
>>          bdrv_collect_image_info(bs, info, filename);
>> -        bdrv_collect_snapshots(bs, info);
>> +        if (!bdrv_query_snapshot_info_list(bs, &info->snapshots, NULL) &&
>> +            info->snapshots) {
>> +            info->has_snapshots = true;
>> +        }
>
> Hmm.  info->snapshots starts life as NULL (thanks to g_new0), and is
> untouched on error.  Since you are ignoring any errors, you technically
> could write:
>
> bdrv_query_snapshot_info_list(bs, &info->snapshots, NULL);
> if (info->snapshots) {
>     info->has_snapshots = true;
> }
>
> for the same semantics.  That means that as of this commit, no caller
> cares about the return value of bdrv_query_snapshot_info_list (they only
> care about whether info->snapshots was changed to non-null), so it could
> return void for a slightly simpler implementation.

Return the list, or NULL.

> But I don't know if any later patches in the series start to care about
> which error was returned.

Me neither :)