[Qemu-devel] [PATCH] backup: allow target without .bdrv_get_info

[Qemu-devel] [PATCH] backup: allow target without .bdrv_get_info

[Vladimir Sementsov-Ogievskiy]

Currently backup to nbd target is broken, as nbd doesn't have
.bdrv_get_info realization.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
---

Hi all!

Since commit

commit 4c9bca7e39a6e07ad02c1dcde3478363344ec60b
Author: John Snow <jsnow@redhat.com>
Date:   Thu Feb 25 15:58:30 2016 -0500

    block/backup: avoid copying less than full target clusters

backup to nbd target is broken, we have "Couldn't determine the cluster size of
the target image".

Proposed NBD protocol extension - NBD_OPT_INFO should finally solve this problem.
But until it is not realized, we need allow backup to nbd target due to backward
compatibility.

Furthermore, is it entirely ok to disallow backup if bds lacks .bdrv_get_info?
Which behavior should be default: to fail backup or to use default cluster size?

 block/backup.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/block/backup.c b/block/backup.c
index ea38733..095a390 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -638,7 +638,10 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
      * backup cluster size is smaller than the target cluster size. Even for
      * targets with a backing file, try to avoid COW if possible. */
     ret = bdrv_get_info(target, &bdi);
-    if (ret < 0 && !target->backing) {
+    if (ret == -ENOTSUP) {
+        /* Cluster size is not defined */
+        job->cluster_size = BACKUP_CLUSTER_SIZE_DEFAULT;
+    } else if (ret < 0 && !target->backing) {
         error_setg_errno(errp, -ret,
             "Couldn't determine the cluster size of the target image, "
             "which has no backing file");
-- 
1.8.3.1

Re: [Qemu-devel] [PATCH] backup: allow target without .bdrv_get_info

[Fam Zheng]

On Tue, 02/14 17:48, Vladimir Sementsov-Ogievskiy wrote:
> Currently backup to nbd target is broken, as nbd doesn't have
> .bdrv_get_info realization.
> 
> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> ---
> 
> Hi all!
> 
> Since commit
> 
> commit 4c9bca7e39a6e07ad02c1dcde3478363344ec60b
> Author: John Snow <jsnow@redhat.com>
> Date:   Thu Feb 25 15:58:30 2016 -0500
> 
>     block/backup: avoid copying less than full target clusters
> 
> backup to nbd target is broken, we have "Couldn't determine the cluster size of
> the target image".
> 
> Proposed NBD protocol extension - NBD_OPT_INFO should finally solve this problem.
> But until it is not realized, we need allow backup to nbd target due to backward
> compatibility.
> 
> Furthermore, is it entirely ok to disallow backup if bds lacks .bdrv_get_info?
> Which behavior should be default: to fail backup or to use default cluster size?
> 
>  block/backup.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/block/backup.c b/block/backup.c
> index ea38733..095a390 100644
> --- a/block/backup.c
> +++ b/block/backup.c
> @@ -638,7 +638,10 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
>       * backup cluster size is smaller than the target cluster size. Even for
>       * targets with a backing file, try to avoid COW if possible. */
>      ret = bdrv_get_info(target, &bdi);
> -    if (ret < 0 && !target->backing) {
> +    if (ret == -ENOTSUP) {
> +        /* Cluster size is not defined */
> +        job->cluster_size = BACKUP_CLUSTER_SIZE_DEFAULT;
> +    } else if (ret < 0 && !target->backing) {
>          error_setg_errno(errp, -ret,
>              "Couldn't determine the cluster size of the target image, "
>              "which has no backing file");
> -- 
> 1.8.3.1
> 

Target not having .bdrv_get_info means 'some sane assumptions are okay' to me, so

Reviewed-by: Fam Zheng <famz@redhat.com>

(But we are going to have a .bdrv_get_info after NBD_OPT_INFO, right?)

Re: [Qemu-devel] [PATCH] backup: allow target without .bdrv_get_info

[Denis V. Lunev]

On 02/14/2017 06:05 PM, Fam Zheng wrote:
> On Tue, 02/14 17:48, Vladimir Sementsov-Ogievskiy wrote:
>> Currently backup to nbd target is broken, as nbd doesn't have
>> .bdrv_get_info realization.
>>
>> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
>> ---
>>
>> Hi all!
>>
>> Since commit
>>
>> commit 4c9bca7e39a6e07ad02c1dcde3478363344ec60b
>> Author: John Snow <jsnow@redhat.com>
>> Date:   Thu Feb 25 15:58:30 2016 -0500
>>
>>     block/backup: avoid copying less than full target clusters
>>
>> backup to nbd target is broken, we have "Couldn't determine the cluster size of
>> the target image".
>>
>> Proposed NBD protocol extension - NBD_OPT_INFO should finally solve this problem.
>> But until it is not realized, we need allow backup to nbd target due to backward
>> compatibility.
>>
>> Furthermore, is it entirely ok to disallow backup if bds lacks .bdrv_get_info?
>> Which behavior should be default: to fail backup or to use default cluster size?
>>
>>  block/backup.c | 5 ++++-
>>  1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/block/backup.c b/block/backup.c
>> index ea38733..095a390 100644
>> --- a/block/backup.c
>> +++ b/block/backup.c
>> @@ -638,7 +638,10 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
>>       * backup cluster size is smaller than the target cluster size. Even for
>>       * targets with a backing file, try to avoid COW if possible. */
>>      ret = bdrv_get_info(target, &bdi);
>> -    if (ret < 0 && !target->backing) {
>> +    if (ret == -ENOTSUP) {
>> +        /* Cluster size is not defined */
>> +        job->cluster_size = BACKUP_CLUSTER_SIZE_DEFAULT;
>> +    } else if (ret < 0 && !target->backing) {
>>          error_setg_errno(errp, -ret,
>>              "Couldn't determine the cluster size of the target image, "
>>              "which has no backing file");
>> -- 
>> 1.8.3.1
>>
> Target not having .bdrv_get_info means 'some sane assumptions are okay' to me, so
>
> Reviewed-by: Fam Zheng <famz@redhat.com>
>
> (But we are going to have a .bdrv_get_info after NBD_OPT_INFO, right?)
>
yes, sooner or later :)

Re: [Qemu-devel] [Qemu-block] [PATCH] backup: allow target without .bdrv_get_info

[Eric Blake]

[-- Attachment #1: Type: text/plain, Size: 777 bytes --]

On 02/14/2017 09:05 AM, Fam Zheng wrote:
>> Proposed NBD protocol extension - NBD_OPT_INFO should finally solve this problem.
>> But until it is not realized, we need allow backup to nbd target due to backward
>> compatibility.
>>
>> Furthermore, is it entirely ok to disallow backup if bds lacks .bdrv_get_info?
>> Which behavior should be default: to fail backup or to use default cluster size?
>>

> 
> Target not having .bdrv_get_info means 'some sane assumptions are okay' to me, so
> 
> Reviewed-by: Fam Zheng <famz@redhat.com>
> 
> (But we are going to have a .bdrv_get_info after NBD_OPT_INFO, right?)

Yes. I'm reposting that series today.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]

Re: [Qemu-devel] [PATCH] backup: allow target without .bdrv_get_info

[John Snow]

On 02/14/2017 09:48 AM, Vladimir Sementsov-Ogievskiy wrote:
> Currently backup to nbd target is broken, as nbd doesn't have
> .bdrv_get_info realization.
> 
> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> ---
> 
> Hi all!
> 
> Since commit
> 
> commit 4c9bca7e39a6e07ad02c1dcde3478363344ec60b
> Author: John Snow <jsnow@redhat.com>
> Date:   Thu Feb 25 15:58:30 2016 -0500
> 
>     block/backup: avoid copying less than full target clusters
> 

Tch.

> backup to nbd target is broken, we have "Couldn't determine the cluster size of
> the target image".
> 
> Proposed NBD protocol extension - NBD_OPT_INFO should finally solve this problem.
> But until it is not realized, we need allow backup to nbd target due to backward
> compatibility.
> 
> Furthermore, is it entirely ok to disallow backup if bds lacks .bdrv_get_info?
> Which behavior should be default: to fail backup or to use default cluster size?
> 
>  block/backup.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/block/backup.c b/block/backup.c
> index ea38733..095a390 100644
> --- a/block/backup.c
> +++ b/block/backup.c
> @@ -638,7 +638,10 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
>       * backup cluster size is smaller than the target cluster size. Even for
>       * targets with a backing file, try to avoid COW if possible. */
>      ret = bdrv_get_info(target, &bdi);
> -    if (ret < 0 && !target->backing) {
> +    if (ret == -ENOTSUP) {
> +        /* Cluster size is not defined */
> +        job->cluster_size = BACKUP_CLUSTER_SIZE_DEFAULT;

I realize this is necessary for backwards compatibility, but perhaps we
need a warning message that explains why this backup *may* not be usable
in the incremental/top cases.

(I think none/full should be ok?)

> +    } else if (ret < 0 && !target->backing) {
>          error_setg_errno(errp, -ret,
>              "Couldn't determine the cluster size of the target image, "
>              "which has no backing file");
> 

-- 
—js