[libvirt] [PATCH v2 4/4] storage: Report error from VolOpen if proper flag is passed

[Cole Robinson]

On 03/31/2014 03:08 PM, Daniel P. Berrange wrote:
> On Mon, Mar 31, 2014 at 02:50:48PM -0400, Cole Robinson wrote:
>> diff --git a/src/storage/storage_backend_fs.c b/src/storage/storage_backend_fs.c
>> index e02d17f..4d44897 100644
>> --- a/src/storage/storage_backend_fs.c
>> +++ b/src/storage/storage_backend_fs.c
>> @@ -904,8 +907,9 @@ virStorageBackendFileSystemRefresh(virConnectPtr conn ATTRIBUTE_UNUSED,
>>                   * have logged a similar message for the same problem, but only
>>                   * if AUTO format detection was used. */
>>                  virReportError(VIR_ERR_INTERNAL_ERROR,
>> -                               _("cannot probe backing volume info: %s"),
>> -                               vol->backingStore.path);
>> +                               _("cannot probe backing volume path '%s': %s"),
>> +                               vol->backingStore.path,
>> +                               virGetLastErrorMessage());
> 
> [snip]
> 
>> diff --git a/src/storage/storage_backend_scsi.c b/src/storage/storage_backend_scsi.c
>> index 4c2484d..51404ff 100644
>> --- a/src/storage/storage_backend_scsi.c
>> +++ b/src/storage/storage_backend_scsi.c
>> @@ -202,8 +202,8 @@ virStorageBackendSCSINewLun(virStoragePoolObjPtr pool,
>>      if (virStorageBackendUpdateVolInfo(vol, true, true,
>>                                         VIR_STORAGE_VOL_OPEN_DEFAULT) < 0) {
>>          virReportError(VIR_ERR_INTERNAL_ERROR,
>> -                       _("Failed to update volume for '%s'"),
>> -                       devpath);
>> +                       _("Failed to update volume for '%s': %s"),
>> +                       devpath, virGetLastErrorMessage());
>>          retval = -1;
>>          goto free_vol;
> 
> The fact that you can call virGetLastErrorMessage() here is showing that
> we're overwriting earlier errors. IMHO use of virGetLastErrorMessage()
> is exclusivly for client applicatons / test suites.  We should just not
> overwrite the original errors here.
> 
> Regards,
> Daniel
> 

Indeed in this case it's not that useful, I was just preserving original
behavior. I'll drop it in the next round.

Thanks,
Cole