If we ever fail to properly set jobinfo->statsType,
qemuDomainJobInfoToParams would return -1 without setting an error.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
---
 src/qemu/qemu_domain.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index fee44812c1..9afe705929 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -730,6 +730,12 @@ qemuDomainJobInfoToParams(qemuDomainJobInfoPtr jobInfo,
         return qemuDomainDumpJobInfoToParams(jobInfo, type, params, nparams);
 
     case QEMU_DOMAIN_JOB_STATS_TYPE_NONE:
+        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                       _("invalid job statistics type"));
+        break;
+
+    default:
+        virReportEnumRangeError(qemuDomainJobStatsType, jobInfo->statsType);
         break;
     }
 
-- 
2.18.0

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On 06/28/2018 03:29 PM, Jiri Denemark wrote:
> If we ever fail to properly set jobinfo->statsType,
> qemuDomainJobInfoToParams would return -1 without setting an error.

Have you actually seen such error? Looks to me like both callers handle
this case properly.

> 
> Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
> ---
>  src/qemu/qemu_domain.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
> index fee44812c1..9afe705929 100644
> --- a/src/qemu/qemu_domain.c
> +++ b/src/qemu/qemu_domain.c
> @@ -730,6 +730,12 @@ qemuDomainJobInfoToParams(qemuDomainJobInfoPtr jobInfo,
>          return qemuDomainDumpJobInfoToParams(jobInfo, type, params, nparams);
>  
>      case QEMU_DOMAIN_JOB_STATS_TYPE_NONE:
> +        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
> +                       _("invalid job statistics type"));
> +        break;
> +
> +    default:
> +        virReportEnumRangeError(qemuDomainJobStatsType, jobInfo->statsType);
>          break;
>      }
>  
> 

Can't we join TYPE_NONE and default together and just report enum range
error? What is the point in differentiating the two?

Michal

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On Thu, Jun 28, 2018 at 15:45:12 +0200, Michal Privoznik wrote:
> On 06/28/2018 03:29 PM, Jiri Denemark wrote:
> > If we ever fail to properly set jobinfo->statsType,
> > qemuDomainJobInfoToParams would return -1 without setting an error.
> 
> Have you actually seen such error? Looks to me like both callers handle
> this case properly.

Yes, when the code for destination migration filled in migration stats,
but failed to set the stats type. As a result no stats were reported
and the programmer's error was silently ignored.

> > Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
> > ---
> >  src/qemu/qemu_domain.c | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
> > index fee44812c1..9afe705929 100644
> > --- a/src/qemu/qemu_domain.c
> > +++ b/src/qemu/qemu_domain.c
> > @@ -730,6 +730,12 @@ qemuDomainJobInfoToParams(qemuDomainJobInfoPtr jobInfo,
> >          return qemuDomainDumpJobInfoToParams(jobInfo, type, params, nparams);
> >  
> >      case QEMU_DOMAIN_JOB_STATS_TYPE_NONE:
> > +        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
> > +                       _("invalid job statistics type"));
> > +        break;
> > +
> > +    default:
> > +        virReportEnumRangeError(qemuDomainJobStatsType, jobInfo->statsType);
> >          break;
> >      }
> >  
> > 
> 
> Can't we join TYPE_NONE and default together and just report enum range
> error? What is the point in differentiating the two?

I did it in the first version of this patch (a long time ago), but
Daniel explained that virReportEnumRangeError is supposed to be used
only for unknown enum values. A different error should be used for known
values which do not make sense in a specific place in the code.

Jirka

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On 06/28/2018 04:30 PM, Jiri Denemark wrote:
> On Thu, Jun 28, 2018 at 15:45:12 +0200, Michal Privoznik wrote:
>> On 06/28/2018 03:29 PM, Jiri Denemark wrote:
>>> If we ever fail to properly set jobinfo->statsType,
>>> qemuDomainJobInfoToParams would return -1 without setting an error.
>>
>> Have you actually seen such error? Looks to me like both callers handle
>> this case properly.
> 
> Yes, when the code for destination migration filled in migration stats,
> but failed to set the stats type. As a result no stats were reported
> and the programmer's error was silently ignored.
> 
>>> Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
>>> ---
>>>  src/qemu/qemu_domain.c | 6 ++++++
>>>  1 file changed, 6 insertions(+)
>>>
>>> diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
>>> index fee44812c1..9afe705929 100644
>>> --- a/src/qemu/qemu_domain.c
>>> +++ b/src/qemu/qemu_domain.c
>>> @@ -730,6 +730,12 @@ qemuDomainJobInfoToParams(qemuDomainJobInfoPtr jobInfo,
>>>          return qemuDomainDumpJobInfoToParams(jobInfo, type, params, nparams);
>>>  
>>>      case QEMU_DOMAIN_JOB_STATS_TYPE_NONE:
>>> +        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
>>> +                       _("invalid job statistics type"));
>>> +        break;
>>> +
>>> +    default:
>>> +        virReportEnumRangeError(qemuDomainJobStatsType, jobInfo->statsType);
>>>          break;
>>>      }
>>>  
>>>
>>
>> Can't we join TYPE_NONE and default together and just report enum range
>> error? What is the point in differentiating the two?
> 
> I did it in the first version of this patch (a long time ago), but
> Daniel explained that virReportEnumRangeError is supposed to be used
> only for unknown enum values. A different error should be used for known
> values which do not make sense in a specific place in the code.

Ah, okay.

ACK then.

Michal

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list