Re: [Qemu-devel] [RFC] monitor: turn on Out-Of-Band by default again

[Peter Xu]

On Wed, May 30, 2018 at 04:04:58PM +0800, Peter Xu wrote:
> On Tue, May 22, 2018 at 02:40:26PM -0400, John Snow wrote:
> > 
> > 
> > On 05/21/2018 10:13 AM, Eric Blake wrote:
> > > On 05/21/2018 03:42 AM, Peter Xu wrote:
> > >> We turned Out-Of-Band feature of monitors off for 2.12 release.  Now we
> > >> try to turn that on again.
> > > 
> > > "try to turn" sounds weak, like you aren't sure of this patch.  If you
> > > aren't sure, then why should we feel safe in applying it?  This text is
> > > going in the permanent git history, so sound bold, rather than hesitant!
> > > 
> > > "We have resolved the issues from last time (commit 3fd2457d reverted by
> > > commit a4f90923):
> > > - issue 1 ...
> > > - issue 2 ...
> > > So now we are ready to enable advertisement of the feature by default"
> > > 
> > > with better descriptions of the issues that you fixed (I can think of at
> > > least the fixes adding thread-safety to the current monitor, and fixing
> > > early use of the monitor before qmp_capabilities completes; there may
> > > also be other issues that you want to call out).
> > > 
> > >>
> > >> Signed-off-by: Peter Xu <address@hidden>
> > >> -- 
> > >> Now OOB should be okay with all known tests (except iotest qcow2, since
> > >> it is still broken on master),
> > > 
> > > Which tests are still failing for you?  Ideally, you can still
> > > demonstrate that the tests not failing without this patch continue to
> > > pass with this patch, even if you call out the tests that have known
> > > issues to still be resolved.
> > > 
> > 
> > Probably 91 and 169. If any others fail that's news to me.
> 
> I just gave it a shot on my workstation too (./check -qcow2):
> 
> Not run: 045 059 064 070 075 076 077 078 081 083 084 088 092 093 094 101 106 
> 109 113 116 119 123 128 131 135 136 146 148 149 160 162 171 173 175 199 207 
> 210 3
> Failures: 087 188 189 198 206
> Failed 5 of 167 tests
> 
> I'm testing against master, e609fa7.

Hmm... I ran again the same test and the same master commit but this
time it passed all 167 tests on my laptop.  So I assume the previous 5
failures are unreproducable at least every time (or there might be
something wrong with my testbed).  I'll rerun some more times, and
when I post the OOB patch I'll cover all correct qcow2 tests.

Regards,

-- 
Peter Xu