[libvirt] [PATCH] qemu: Adjust size for qcow2/qed if not on sector boundary

[John Ferlan]

On 04/08/2014 12:26 PM, John Ferlan wrote:
> https://bugzilla.redhat.com/show_bug.cgi?id=1002813
> 
> If qemuDomainBlockResize() is passed a size not on a KiB boundary - that
> is passed a size based in bytes (VIR_DOMAIN_BLOCK_RESIZE_BYTES), then
> depending on the source format (qcow2 or qed), the value passed must
> be on a sector (or 512 byte) boundary. Since other libvirt code quietly
> adjusts the capacity values, then do so here as well - of course ensuring
> that adjustment still fits.
> 
> Signed-off-by: John Ferlan <jferlan at redhat.com>
> ---
>  src/qemu/qemu_driver.c | 22 ++++++++++++++++++++++
>  1 file changed, 22 insertions(+)
> 

Although discussion was taken off this list - the changes here were
ACK'd and pushed today...

Essentially the following API's will round up the value as well:

    virStorageBackendCreateQcowCreate()
    virStorageBackendLogicalCreateVol()
    virStorageBackendCreateQemuImgCmd()

For libvirt created volumes, virStorageBackendCreateQemuImgCmd() or
virStorageBackendCreateQcowCreate() is called - both will take the
capacity value and VIR_DIV_UP using 1024. For the vol-resize path (e.g.
non running vm case), virStorageBackendFilesystemResizeQemuImg() will
use ROUND_UP on 512 byte value because it knows (and comments) that
qemu-img will fail to resize on non sector boundaries.

Additionally, it was noted that using "K" and "KiB" would produce 1024
based results, it's libvirt's allowance of "KB" for sizes that results
in the nuance. Being strict KB shouldn't be used for storage, but rather
than penalize for not knowing the difference between KiB and KB the code
just assumes KiB should have been used.

John

> diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
> index 4bb4819..3e407d7 100644
> --- a/src/qemu/qemu_driver.c
> +++ b/src/qemu/qemu_driver.c
> @@ -9421,6 +9421,7 @@ qemuDomainBlockResize(virDomainPtr dom,
>      virDomainObjPtr vm;
>      qemuDomainObjPrivatePtr priv;
>      int ret = -1, idx;
> +    unsigned long long size_up;
>      char *device = NULL;
>      virDomainDiskDefPtr disk = NULL;
>  
> @@ -9441,6 +9442,12 @@ qemuDomainBlockResize(virDomainPtr dom,
>              return -1;
>          }
>          size *= 1024;
> +        size_up = size;
> +    } else {
> +        /* For 'qcow2' and 'qed', qemu resize blocks expects values
> +         * on sector boundary, so round our value up to prepare
> +         */
> +        size_up = VIR_ROUND_UP(size, 512);
>      }
>  
>      if (!(vm = qemuDomObjFromDomain(dom)))
> @@ -9467,6 +9474,21 @@ qemuDomainBlockResize(virDomainPtr dom,
>      }
>      disk = vm->def->disks[idx];
>  
> +    /* qcow2 and qed must be sized appropriately, so be sure our value
> +     * is sized appropriately and will fit
> +     */
> +    if (size != size_up &&
> +        (disk->src.format == VIR_STORAGE_FILE_QCOW2 ||
> +         disk->src.format == VIR_STORAGE_FILE_QED)) {
> +        if (size_up > ULLONG_MAX) {
> +            virReportError(VIR_ERR_OVERFLOW,
> +                           _("size must be less than %llu KiB"),
> +                           ULLONG_MAX / 1024);
> +            goto endjob;
> +        }
> +        size = size_up;
> +    }
> +
>      if (virAsprintf(&device, "%s%s", QEMU_DRIVE_HOST_PREFIX,
>                      disk->info.alias) < 0)
>          goto endjob;
>