Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.80.23.2.2\))
Subject: Re: KCSAN + KVM = host reset
From:   Qian Cai <cai@lca.pw>
In-Reply-To: <CANpmjNMEgc=+bLU472jy37hYPYo5_c+Kbyti8-mubPsEGBrm3A@mail.gmail.com>
Date:   Thu, 9 Apr 2020 11:30:14 -0400
Cc:     Paolo Bonzini <pbonzini@redhat.com>,
        "paul E. McKenney" <paulmck@kernel.org>,
        kasan-dev <kasan-dev@googlegroups.com>,
        LKML <linux-kernel@vger.kernel.org>, kvm@vger.kernel.org
Content-Transfer-Encoding: quoted-printable
Message-Id: <2730C0CC-B8B5-4A65-A4ED-9DFAAE158AA6@lca.pw>
References: <E180B225-BF1E-4153-B399-1DBF8C577A82@lca.pw>
 <fb39d3d2-063e-b828-af1c-01f91d9be31c@redhat.com>
 <017E692B-4791-46AD-B9ED-25B887ECB56B@lca.pw>
 <CANpmjNMiHNVh3BVxZUqNo4jW3DPjoQPrn-KEmAJRtSYORuryEA@mail.gmail.com>
 <B7F7F73E-EE27-48F4-A5D0-EBB29292913E@lca.pw>
 <CANpmjNMEgc=+bLU472jy37hYPYo5_c+Kbyti8-mubPsEGBrm3A@mail.gmail.com>
To:     Marco Elver <elver@google.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On Apr 9, 2020, at 11:22 AM, Marco Elver <elver@google.com> wrote:
>=20
> On Thu, 9 Apr 2020 at 17:10, Qian Cai <cai@lca.pw> wrote:
>>=20
>>=20
>>=20
>>> On Apr 9, 2020, at 3:03 AM, Marco Elver <elver@google.com> wrote:
>>>=20
>>> On Wed, 8 Apr 2020 at 23:29, Qian Cai <cai@lca.pw> wrote:
>>>>=20
>>>>=20
>>>>=20
>>>>> On Apr 8, 2020, at 5:25 PM, Paolo Bonzini <pbonzini@redhat.com> =
wrote:
>>>>>=20
>>>>> On 08/04/20 22:59, Qian Cai wrote:
>>>>>> Running a simple thing on this AMD host would trigger a reset =
right away.
>>>>>> Unselect KCSAN kconfig makes everything work fine (the host would =
also
>>>>>> reset If only "echo off > /sys/kernel/debug/kcsan=E2=80=9D before =
running qemu-kvm).
>>>>>=20
>>>>> Is this a regression or something you've just started to play =
with?  (If
>>>>> anything, the assembly language conversion of the AMD world switch =
that
>>>>> is in linux-next could have reduced the likelihood of such a =
failure,
>>>>> not increased it).
>>>>=20
>>>> I don=E2=80=99t remember I had tried this combination before, so =
don=E2=80=99t know if it is a
>>>> regression or not.
>>>=20
>>> What happens with KASAN? My guess is that, since it also happens =
with
>>> "off", something that should not be instrumented is being
>>> instrumented.
>>=20
>> No, KASAN + KVM works fine.
>>=20
>>>=20
>>> What happens if you put a 'KCSAN_SANITIZE :=3D n' into
>>> arch/x86/kvm/Makefile? Since it's hard for me to reproduce on this
>>=20
>> Yes, that works, but this below alone does not work,
>>=20
>> KCSAN_SANITIZE_kvm-amd.o :=3D n
>=20
> There are some other files as well, that you could try until you hit
> the right one.
>=20
> But since this is in arch, 'KCSAN_SANITIZE :=3D n' wouldn't be too bad
> for now. If you can't narrow it down further, do you want to send a
> patch?

No, that would be pretty bad because it will disable KCSAN for Intel
KVM as well which is working perfectly fine right now. It is only AMD
is broken.

>=20
> Thanks,
> -- Marco
>=20
>> I have been able to reproduce this on a few AMD hosts.
>>=20
>>> exact system, I'd ask you to narrow it down by placing =
'KCSAN_SANITIZE
>>> :=3D n' into suspect subsystems' Makefiles. Once you get it to work =
with
>>> that, we can refine the solution.
>>>=20
>>> Thanks,
>>> -- Marco