<div dir="ltr">hello andrei and dimitry,<div><br></div><div>(dimitry, thats a cool email address).</div><div><br></div><div>so i started creating a VM for you guys to reproduce the bug, and found out it works great on t2.large aws instances - but crashes consistently on c5.large aws instances, which i think explains why you couldn't reproduce it...</div><div>dimitry's test fails on the c5.large, and succeeds on the t2.large (rest of the tests you requested, pass on both). output attached. (i had to add sudo and change protobuf-python -> python protobuf).</div><div><br></div><div>pypy crashes after restore on my personal laptop as well, but unfortunatly the tests didn't run on my laptop, output exception and cpuinfo also attached.</div><div><br></div><div>i'm creating a VM for you guys on AWS with a c5.large instance type to work on, will send connection details later on a private email.</div><div>i'm using the following ami, at eu-west-1 aws region<br></div><div>"ami-0741d47e" # ubuntu/images/hvm-ssd/ubuntu-artful-17.10-amd64-server-20180102<br></div><div><br></div><div>i'll also finish writing a script to reproduce the issue, and will send you the instructions in that email.</div><div><br></div><div>thanks you very much,</div><div>Shlomi.</div><div><br></div><div><br></div><div><div><br></div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Sun, Jan 28, 2018 at 9:01 PM, Andrei Vagin <span dir="ltr"><<a href="mailto:avagin@virtuozzo.com" target="_blank">avagin@virtuozzo.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">On Sat, Jan 27, 2018 at 12:28:31PM -0800, Andrei Vagin wrote:<br>
> On Fri, Jan 26, 2018 at 02:02:35PM +0200, Shlomi Matichin wrote:<br>
> > 0x00007fc88f5c7788: jmpq 0x7fc88f5c7706<br>
> > 0x00007fc88f5c778d: nopl (%rax)<br>
> > 0x00007fc88f5c7790: movl $0x0,0x4(%r13)<br>
> > 0x00007fc88f5c7798: jmpq 0x7fc88f5c76fc<br>
> > 0x00007fc88f5c779d: nopl (%rax)<br>
> > 0x00007fc88f5c77a0: movl $0x0,0x0(%r13)<br>
> > 0x00007fc88f5c77a8: jmpq 0x7fc88f5c76ee<br>
> > 0x00007fc88f5c77ad: pxor %xmm3,%xmm3<br>
> > => 0x00007fc88f5c77b1: cvtsd2ss %xmm0,%xmm3<br>
><br>
> Looks like we incorrectly save fpu state...<br>
><br>
> I tried to reproduce this issue on my laptop, in digital ocean with the same<br>
> ubuntu version, and I even created an aws account. But in all my attempts,<br>
> everything works as expected.<br>
><br>
> Could you write a detailed instruction how you reproduce this bug on a<br>
> ec2 instance? What image do you use? Maybe you can create there an<br>
> instance and give us access to it?<br>
><br>
> Could you show registers before and after dump/restore?<br>
><br>
> For that you need to attach to a target process by gdb and call " info<br>
> all-registers", then you need to detach from a process, do dump/restore<br>
> and attach again to get registers. Pls, don't forget to check that the<br>
> issue is reproduced in this case.<br>
<br>
</span>We have an idea that mxcsr isn't dumped properlly in your case. And Dima<br>
wrote a test to check this. Could you try to run it?<br>
<br>
1. git clone <a href="https://github.com/avagin/criu" rel="noreferrer" target="_blank">https://github.com/avagin/criu</a> -b fpu2<br>
2. cd criu<br>
3. apt-get update && apt-get install -y \<br>
build-essential \<br>
protobuf-c-compiler \<br>
libprotobuf-c0-dev \<br>
libprotobuf-dev \<br>
bsdmainutils \<br>
protobuf-compiler \<br>
python-minimal \<br>
libaio-dev \<br>
libcap-dev \<br>
iptables \<br>
libnl-3-dev \<br>
libnl-route-3-dev \<br>
libselinux-dev \<br>
pkg-config \<br>
git-core \<br>
libnet-dev \<br>
gcc \<br>
python-yaml protobuf-python<br>
4. make<br>
5. make -C test/zdtm/lib<br>
6. python test/zdtm.py run -t zdtm/static/fpu02<br>
<br>
In addition, you can try to run other fpu tests:<br>
7. python test/zdtm.py run -t zdtm/static/sse00 -t zdtm/static/sse20 -t zdtm/static/mmx00 -t zdtm/static/fpu00<br>
<br>
Thanks,<br>
Andrei<br>
<br>
><br>
> Thanks,<br>
> Andrei<br>
</blockquote></div><br></div>