<p dir="ltr">On May 28, 2015 8:12 PM, "Stanislav German-Evtushenko" <<a href="mailto:ginermail@gmail.com">ginermail@</a><a href="mailto:ginermail@gmail.com">gmail.com</a>> wrote:<br>
><br>
> > Your test code was a really interesting example, and lead to<br>
> > really unexpected results (at least for me). But as Eneko already<br>
> > mentioned, nobody would write such code. It is simply not thread save,<br>
> > and I think qemu does it correctly.<br>
><br>
> I have written that code only because nobody wanted to test real-life<br>
> test cases I had provided (VM with low memory and Windows 7 inside).<br>
> However now I think it is going to be even easier to reproduce in a<br>
> VM. My guess if you install a VM with virtual drive on DRBD or MD RAID<br>
> and cache=none, create ext3 or ext4 partition inside this VM and run<br>
> my code inside then you will get inconsistence. May be you need to run<br>
> it not once but 10-20 times but finally it should happen.<br>
><br>
> Stanislav</p>
<p dir="ltr">Also I want to mentioned that I started to write that C-code because of the following message in logs which clearly stated what happened:</p>
<p dir="ltr">Jun 14 08:50:12 kvm5 kernel: block drbd0: Digest mismatch, buffer modified by upper layers during write: 21158352s +4096</p>