Re: [Qemu-devel] [PATCH 1/5] qcow2: Fix warnings in check_refcount()

[Kevin Wolf]

Am Freitag, 17. April 2009 23:03 schrieb Anthony Liguori:
> Kevin Wolf wrote:
> > Hi Anthony,
> >
> > Am Freitag, 17. April 2009 22:39 schrieb Anthony Liguori:
> >> Do you have a qemu-io script handy that can be used to stress something
> >> like this patch set?  After the last qcow2 regression, I'm wary of
> >> additional cleanups that we can't validate with a strong stress test.
> >
> > This patch series is harmless in that respect. You can tell alone from
> > looking at the patches that it can't cause regressions in normal
> > operation, because it only touches code which was previosuly not even
> > built and is only called by qemu-img (after patch 3) and when DEBUG_ALLOC
> > is defined.
>
> I'm basically at the point of not wanting to touch qcow2 without serious
> testing.  That said, I can do enough on my own to satisfy me so I'll
> commit this series later today or tomorrow.

I perfectly understand that you don't want to break it again. But then, the 
only way to avoid new bugs is to stop development completely. This isn't a 
solution either.

This is even more true for changes which are actually made for testing and 
debugging purposes like these. This series is what helped me to find the 
corruption bug.

What we should do is to make sure that qcow2 patches (especially those 
touching the core) are given a thorough review before committing.

> > But you would better apply the corruption fix I sent on Wednesday. ;-)
>
> Yes, I just checked that in.  Very good catch!
>
> > And even though I think that this series can't break anything, we
> > definitely could use a strong test suite. I'm almost sure that there is
> > at least one bug left (the one Jamie Lokier saw from 5006 on, but nobody
> > ever found it).
>
> You don't think that was Nolan's fix?

Hm, I haven't look very much in detail at it. But according to the commit log 
only qcow_is_allocated() was affected, and I can't see how booting Jamie's 
Windows guest would call this function.

Kevin