kvmclock: fix clock_is_reliable on migration from QEMU < 2.9
commitc2b01cfec1f1426d95c27abacc3ea5b5fbb9a57e
authorMichael Chapman <mike@very.puzzling.org>
Fri, 6 Apr 2018 05:34:06 +0000 (6 15:34 +1000)
committerPaolo Bonzini <pbonzini@redhat.com>
Mon, 9 Apr 2018 14:36:39 +0000 (9 16:36 +0200)
tree8b70046af10290f8a0d5a4e9290ad2684795dd16
parent4968a2c6edb7b46b127c19a8426575d23b55ab1f
kvmclock: fix clock_is_reliable on migration from QEMU < 2.9

When migrating from a pre-2.9 QEMU, no clock_is_reliable flag is
transferred. We should assume that the source host has an unreliable
KVM_GET_CLOCK, rather than using whatever was determined locally, to
ensure that any drift from the TSC-based value calculated by the guest
is corrected.

Signed-off-by: Michael Chapman <mike@very.puzzling.org>
Message-Id: <20180406053406.774-1-mike@very.puzzling.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
hw/i386/kvm/clock.c