<div dir="ltr"><div class="gmail_extra">Eventually, I can imagine only one possibility of why initial problem can happen. If child don't write to memory at all, and then dump happens, and criu will see dirty bits in child as they were in parent, becouse there were no PF's.<br clear="all">
<div><div dir="ltr"><br>Best Regards, Tikhomirov Pavel.</div></div>
<br><br><div class="gmail_quote">2014-04-15 23:27 GMT+04:00 Cyrill Gorcunov <span dir="ltr"><<a href="mailto:gorcunov@gmail.com" target="_blank">gorcunov@gmail.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div class="">On Tue, Apr 15, 2014 at 11:15:19PM +0400, Cyrill Gorcunov wrote:<br>
> On Tue, Apr 15, 2014 at 11:07:28PM +0400, Pavel Tikhomirov wrote:<br>
> > Please read this comment.<br>
> > Best Regards, Tikhomirov Pavel.<br>
> ><br>
> > 2014-04-15 17:43 GMT+04:00 Pavel Tikhomirov <<a href="mailto:snorcht@gmail.com">snorcht@gmail.com</a>>:<br>
> ><br>
> > Ok, but what if parent writes to memory first and gets COW'ed,<br>
> > and child gets writable memory when access it.<br>
> > It can result in no PF.<br>
> > Just an Idea, need to understand COW mechanism more precisely.<br>
><br>
> OK, as far as I remember, once the process is forked, its pte/pmd structures<br>
> become allocated but not present, once child attempt to read something, the<br>
<br>
</div>Sorry, for COW there is copy_one_pte which would WP both parent and child.<br>
</blockquote></div><br></div></div>