δΊ 2013-2-27 22:26, Markus Armbruster ει:
> Wenchao Xia <xiaw...@linux.vnet.ibm.com> writes:
>
>> This patch add function bdrv_query_snapshot_infolist(), which will
>
> adds
>
>> return snapshot info of an image in qmp object format. The implementation
>> code are based on the code moved from qemu-img.c with modification to fit
>> more
>
> implementation is based
>
>> for qmp based block layer API.
>> To help filter out snapshot info not needed, a call back function is
>> added in bdrv_query_snapshot_infolist().
>> bdrv_can_read_snapshot() should be called before call this function,
>> to avoid got *errp set unexpectedly.
>
> "avoid getting"
>
> Not sure about "should". Couldn't a caller safely call this with a
> non-snapshot bs argument, and simply deal with the error?
>
yes, but in some case this error should be avoided. For eg, in
bdrv_query_image_info(), if a image is not inserted, or it can't
take snapshot, other info should also be returned without error,
so if bdrv_can_read_snapshot() is called before, this kind of
error is avoided, and caller can check if there is other
error. I guess it should be documented as "in some case,
bdrv_can_read_snapshot() should be called."
> Long lines, and unusual paragraph formatting. Here's how we usually do
> this:
>
> This patch adds function bdrv_query_snapshot_infolist(), which will
> return 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.
>
> To help filter out snapshot info not needed, a call back function is
> added in bdrv_query_snapshot_infolist().
>
> bdrv_can_read_snapshot() should be called before call this function, to
> avoid getting *errp set unexpectedly.
>
Thanks, going to use this.
>> Signed-off-by: Wenchao Xia <xiaw...@linux.vnet.ibm.com>
>> Reviewed-by: Eric Blake <ebl...@redhat.com>
>> ---
>> block.c | 60
>> ++++++++++++++++++++++++++++++++++++++----------
>> include/block/block.h | 8 +++++-
>> qemu-img.c | 8 +++++-
>> 3 files changed, 61 insertions(+), 15 deletions(-)
>>
>> diff --git a/block.c b/block.c
>> index 368b37c..8d0145a 100644
>> --- a/block.c
>> +++ b/block.c
>> @@ -2813,29 +2813,62 @@ int coroutine_fn
>> bdrv_co_is_allocated_above(BlockDriverState *top,
>> return 0;
>> }
>>
>> -void collect_snapshots(BlockDriverState *bs , ImageInfo *info)
>> +SnapshotInfoList *bdrv_query_snapshot_infolist(BlockDriverState *bs,
>> + SnapshotFilterFunc filter,
>> + void *opaque,
>> + Error **errp)
>> {
>> int i, sn_count;
>> QEMUSnapshotInfo *sn_tab = NULL;
>> - SnapshotInfoList *info_list, *cur_item = NULL;
>> + SnapshotInfoList *info_list, *cur_item = NULL, *head = NULL;
>> + SnapshotInfo *info;
>> +
>> sn_count = bdrv_snapshot_list(bs, &sn_tab);
>> + if (sn_count < 0) {
>> + /* Now bdrv_snapshot_list() use negative value to tip error, so a
>> check
>> + * of it was done here. In future errp can be set by that function
>> + * itself, by changing the call back functions in c files in
>> ./block.
>> + */
>
> I doubt we need this comment.
>
I am OK to remove it.
>> + const char *dev = bdrv_get_device_name(bs);
>> + switch (sn_count) {
>> + case -ENOMEDIUM:
>> + error_setg(errp, "Device '%s' is not inserted.", dev);
>
> We generally don't end error messages with a period. Please remove it
> from this and other messages.
>
OK.
>> + break;
>> + case -ENOTSUP:
>> + error_setg(errp,
>> + "Device '%s' does not support internal snapshot.",
>
> internal snapshots
>
OK.
>> + dev);
>> + break;
>> + default:
>> + error_setg(errp,
>> + "Device '%s' got '%d' for bdrv_snapshot_list(), "
>> + "message '%s'.",
>> + dev, sn_count, strerror(-sn_count));
>
> error_setg(errp, "Can't list snapshots of device '%s': %s",
> dev, sterror(-sn_count));
>
OK.
>> + break;
>> + }
>> + return NULL;
>> + }
>>
>> for (i = 0; i < sn_count; i++) {
>> - info->has_snapshots = true;
>> - info_list = g_new0(SnapshotInfoList, 1);
>> + if (filter && filter(&sn_tab[i], opaque) != 0) {
>> + continue;
>> + }
>
> Is the filter feature really worth it? If yes, should it be added in a
> separate patch?
>
I feel filter make logic more clearer later which can be used later
to filter out inconsistent snapshots. I am OK to form a seperate patch.
>>
>> - info_list->value = g_new0(SnapshotInfo, 1);
>> - info_list->value->id = g_strdup(sn_tab[i].id_str);
>> - info_list->value->name = g_strdup(sn_tab[i].name);
>> - info_list->value->vm_state_size = sn_tab[i].vm_state_size;
>> - info_list->value->date_sec = sn_tab[i].date_sec;
>> - info_list->value->date_nsec = sn_tab[i].date_nsec;
>> - info_list->value->vm_clock_sec = sn_tab[i].vm_clock_nsec /
>> 1000000000;
>> - info_list->value->vm_clock_nsec = sn_tab[i].vm_clock_nsec %
>> 1000000000;
>> + info = g_new0(SnapshotInfo, 1);
>> + info->id = g_strdup(sn_tab[i].id_str);
>> + info->name = g_strdup(sn_tab[i].name);
>> + info->vm_state_size = sn_tab[i].vm_state_size;
>> + info->date_sec = sn_tab[i].date_sec;
>> + info->date_nsec = sn_tab[i].date_nsec;
>> + info->vm_clock_sec = sn_tab[i].vm_clock_nsec / 1000000000;
>> + info->vm_clock_nsec = sn_tab[i].vm_clock_nsec % 1000000000;
>> +
>> + info_list = g_new0(SnapshotInfoList, 1);
>> + info_list->value = info;
>>
>> /* XXX: waiting for the qapi to support qemu-queue.h types */
>> if (!cur_item) {
>> - info->snapshots = cur_item = info_list;
>> + head = cur_item = info_list;
>> } else {
>> cur_item->next = info_list;
>> cur_item = info_list;
>> @@ -2844,6 +2877,7 @@ void collect_snapshots(BlockDriverState *bs ,
>> ImageInfo *info)
>> }
>>
>> g_free(sn_tab);
>> + return head;
>> }
>>
>> void collect_image_info(BlockDriverState *bs,
>> diff --git a/include/block/block.h b/include/block/block.h
>> index e6d915c..51a14f2 100644
>> --- a/include/block/block.h
>> +++ b/include/block/block.h
>> @@ -319,6 +319,13 @@ void bdrv_get_backing_filename(BlockDriverState *bs,
>> char *filename, int filename_size);
>> void bdrv_get_full_backing_filename(BlockDriverState *bs,
>> char *dest, size_t sz);
>> +
>> +typedef int (*SnapshotFilterFunc)(const QEMUSnapshotInfo *sn, void *opaque);
>> +/* assume bs is already opened, use qapi_free_* to free returned value. */
>
> Pretty much all functions assume bs is open, hardly worth a comment.
>
OK, will remove.
>> +SnapshotInfoList *bdrv_query_snapshot_infolist(BlockDriverState *bs,
>> + SnapshotFilterFunc filter,
>> + void *opaque,
>> + Error **errp);
>> BlockInfo *bdrv_query_info(BlockDriverState *s);
>> BlockStats *bdrv_query_stats(const BlockDriverState *bs);
>> bool bdrv_can_read_snapshot(BlockDriverState *bs);
>> @@ -450,7 +457,6 @@ int bdrv_debug_breakpoint(BlockDriverState *bs, const
>> char *event,
>> int bdrv_debug_resume(BlockDriverState *bs, const char *tag);
>> bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag);
>>
>> -void collect_snapshots(BlockDriverState *bs , ImageInfo *info);
>> void collect_image_info(BlockDriverState *bs,
>> ImageInfo *info,
>> const char *filename);
>> diff --git a/qemu-img.c b/qemu-img.c
>> index b650d17..1034cc5 100644
>> --- a/qemu-img.c
>> +++ b/qemu-img.c
>> @@ -1280,7 +1280,13 @@ static ImageInfoList *collect_image_info_list(const
>> char *filename,
>>
>> info = g_new0(ImageInfo, 1);
>> collect_image_info(bs, info, filename);
>> - collect_snapshots(bs, info);
>> + if (bdrv_can_read_snapshot(bs)) {
>> + info->snapshots = bdrv_query_snapshot_infolist(bs, NULL,
>> + NULL, NULL);
>> + if (info->snapshots) {
>> + info->has_snapshots = true;
>> + }
>> + }
>>
>> elem = g_new0(ImageInfoList, 1);
>> elem->value = info;
>
> Silently ignores errors, but the old code does so, too.
>
OK, error will be checked.
> QAPI's has_FOO is pointless when FOO is a pointer. Not your fault.
>
--
Best Regards
Wenchao Xia
