linux-kernel - [PATCH 0/3] x86/kvm/hyper-v: More fixes for TSC page clocksource for Hyper-V on KVM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 28 Feb 2018 14:43:58 +0100
From:   Vitaly Kuznetsov <vkuznets@...hat.com>
To:     kvm@...r.kernel.org
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org,
        Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>,
        "K. Y. Srinivasan" <kys@...rosoft.com>,
        "Michael Kelley (EOSG)" <Michael.H.Kelley@...rosoft.com>,
        Andrey Smetanin <asmetanin@...tuozzo.com>,
        Roman Kagan <rkagan@...tuozzo.com>,
        "Denis V . Lunev" <den@...nvz.org>
Subject: [PATCH 0/3] x86/kvm/hyper-v: More fixes for TSC page clocksource for Hyper-V on KVM

Previously (see commit 72c139bacf) it was found that Hyper-V frequency
MSRs are required to make Hyper-V on KVM pass through TSC page as stable
clocksource to its guests. However, to make things work this is not
enough. Hyper-V tries to prove that TSC frequency underneath it won't
change (e.g. when it's migrated), there are two mechanisms:

1) Invariant TSC (CPUID.80000007H:EDX[8]). If Hyper-V sees this bit it will
 treat TSC as stable. We, however, don't want to pass it as it makes
 migration hard (e.g. Qemu adds a migration blocker when 'invtsc' flag is
 passed. Genuine Hyper-V running in L0 doesn't pass it either.

2) Hyper-V Reenlightenment (CPUID.40000003H:EAX[13]).

This patch series add rudimentary support for Hyper-V reenlightenment
notifications to KVM ('producer' part; we already implemented
reenlightenment 'consumer' for KVM-on Hyper-V some time ago) and fixes
bugs I found during testing. Fully fledged reenlightenment  implementation
will be added later when we learn to migrate nested workloads in KVM making
testing possible.

Qemu patches are also required, I'll post them separately.

Vitaly Kuznetsov (3):
  x86/kvm/hyper-v: add reenlightenment MSRs support
  x86/kvm/hyper-v: remove stale entries from vec_bitmap/auto_eoi_bitmap
    on vector change
  x86/kvm/hyper-v: inject #GP only when invalid SINTx vector is unmasked

 arch/x86/include/asm/kvm_host.h    |  4 +++
 arch/x86/include/uapi/asm/hyperv.h |  3 ++
 arch/x86/kvm/hyperv.c              | 58 +++++++++++++++++++++++++++++++-------
 arch/x86/kvm/x86.c                 | 12 +++++++-
 4 files changed, 66 insertions(+), 11 deletions(-)

-- 
2.14.3