Re: [Qemu-devel] [PATCH v2 4/4] qemu-iotests: Check common namespace for id and node-name

[Kevin Wolf]

Am 21.04.2014 um 09:01 hat Fam Zheng geschrieben:
> On Thu, 04/17 17:48, Kevin Wolf wrote:
> > A name that is taken by an ID can't be taken by a node-name at the same
> > time. Check that conflicts are correctly detected.
> > 
> > Signed-off-by: Kevin Wolf <address@hidden>
> > ---
> >  tests/qemu-iotests/087     | 52 
> > ++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/qemu-iotests/087.out |  5 +++++
> >  2 files changed, 57 insertions(+)
> > 
> >  
> > diff --git a/tests/qemu-iotests/087.out b/tests/qemu-iotests/087.out
> > index 479bf86..7fbee3f 100644
> > --- a/tests/qemu-iotests/087.out
> > +++ b/tests/qemu-iotests/087.out
> > @@ -20,6 +20,11 @@ QMP_VERSION
> >  {"return": {}}
> >  {"return": {}}
> >  {"error": {"class": "GenericError", "desc": "Device with id 'disk' already 
> > exists"}}
> > +{"error": {"class": "GenericError", "desc": "Device with node-name 
> > 'test-node' already exists"}}
> > +main-loop: WARNING: I/O thread spun for 1000 iterations
> 
> Is this warning message deterministic? I think we could avoid the guest code 
> by
> '-machine accel=qtest'.

At least it's deterministic enough that the test case never failed for
me in multiple runs on two different machines with different OS versions
(RHEL 6 and 7).

> BTW I intended to write a patch for this but it would have some overlap with
> Jeff's qemu-iotests common.qemu series:

Then let's commit this as it is and once Jeff's series is in, you can
fix up everything at once?

Kevin