Re: [PATCH v2] i386: Make migration fail when Hyper-V reenlightenment was enabled but 'user_tsc_khz' is unset

[Vitaly Kuznetsov <vkuznets@redhat.com>]

"Dr. David Alan Gilbert" <dgilbert@redhat.com> writes:

> * Vitaly Kuznetsov (vkuznets@redhat.com) wrote:
>> "Dr. David Alan Gilbert" <dgilbert@redhat.com> writes:
>> 
>> > * Vitaly Kuznetsov (vkuznets@redhat.com) wrote:
>> >> KVM doesn't fully support Hyper-V reenlightenment notifications on
>> >> migration. In particular, it doesn't support emulating TSC frequency
>> >> of the source host by trapping all TSC accesses so unless TSC scaling
>> >> is supported on the destination host and KVM_SET_TSC_KHZ succeeds, it
>> >> is unsafe to proceed with migration.
>> >> 
>> >> KVM_SET_TSC_KHZ is called from two sites: kvm_arch_init_vcpu() and
>> >> kvm_arch_put_registers(). The later (intentionally) doesn't propagate
>> >> errors allowing migrations to succeed even when TSC scaling is not
>> >> supported on the destination. This doesn't suit 're-enlightenment'
>> >> use-case as we have to guarantee that TSC frequency stays constant.
>> >> 
>> >> Require 'tsc-frequency=' command line option to be specified for successful
>> >> migration when re-enlightenment was enabled by the guest.
>> >> 
>> >> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>> >> ---
>> >> This patch is a successor of "[PATCH 3/3] i386: Make sure
>> >> kvm_arch_set_tsc_khz() succeeds on migration when 'hv-reenlightenment'
>> >> was exposed" taking a different approach suggested by Paolo.
>> >> ---
>> >>  docs/hyperv.txt                |  5 +++++
>> >>  target/i386/kvm/hyperv-proto.h |  1 +
>> >>  target/i386/machine.c          | 20 ++++++++++++++++++++
>> >>  3 files changed, 26 insertions(+)
>> >> 
>> >> diff --git a/docs/hyperv.txt b/docs/hyperv.txt
>> >> index 5df00da54fc4..e53c581f4586 100644
>> >> --- a/docs/hyperv.txt
>> >> +++ b/docs/hyperv.txt
>> >> @@ -160,6 +160,11 @@ the hypervisor) until it is ready to switch to the new one. This, in conjunction
>> >>  with hv-frequencies, allows Hyper-V on KVM to pass stable clocksource (Reference
>> >>  TSC page) to its own guests.
>> >>  
>> >> +Note, KVM doesn't fully support re-enlightenment notifications and doesn't
>> >> +emulate TSC accesses after migration so 'tsc-frequency=' CPU option also has to
>> >> +be specified to make migration succeed. The destination host has to either have
>> >> +the same TSC frequency or support TSC scaling CPU feature.
>> >> +
>> >>  Recommended: hv-frequencies
>> >>  
>> >>  3.16. hv-evmcs
>> >> diff --git a/target/i386/kvm/hyperv-proto.h b/target/i386/kvm/hyperv-proto.h
>> >> index 056a305be38c..e30d64b4ade4 100644
>> >> --- a/target/i386/kvm/hyperv-proto.h
>> >> +++ b/target/i386/kvm/hyperv-proto.h
>> >> @@ -139,6 +139,7 @@
>> >>   * Reenlightenment notification MSRs
>> >>   */
>> >>  #define HV_X64_MSR_REENLIGHTENMENT_CONTROL      0x40000106
>> >> +#define HV_REENLIGHTENMENT_ENABLE_BIT           (1u << 16)
>> >>  #define HV_X64_MSR_TSC_EMULATION_CONTROL        0x40000107
>> >>  #define HV_X64_MSR_TSC_EMULATION_STATUS         0x40000108
>> >>  
>> >> diff --git a/target/i386/machine.c b/target/i386/machine.c
>> >> index 7259fe6868c6..137604ddb898 100644
>> >> --- a/target/i386/machine.c
>> >> +++ b/target/i386/machine.c
>> >> @@ -883,11 +883,31 @@ static bool hyperv_reenlightenment_enable_needed(void *opaque)
>> >>          env->msr_hv_tsc_emulation_status != 0;
>> >>  }
>> >>  
>> >> +static int hyperv_reenlightenment_post_load(void *opaque, int version_id)
>> >> +{
>> >> +    X86CPU *cpu = opaque;
>> >> +    CPUX86State *env = &cpu->env;
>> >> +
>> >> +    /*
>> >> +     * KVM doesn't fully support re-enlightenment notifications so we need to
>> >> +     * make sure TSC frequency doesn't change upon migration.
>> >> +     */
>> >> +    if ((env->msr_hv_reenlightenment_control & HV_REENLIGHTENMENT_ENABLE_BIT) &&
>> >> +        !env->user_tsc_khz) {
>> >> +        error_report("Guest enabled re-enlightenment notifications, "
>> >> +                     "'tsc-frequency=' has to be specified");
>> >
>> > It's unusual to fail on the destination for a valid configuration but
>> > guest state;  wouldn't it be better to always insist on tsc-frequency if
>> > that hv feature is exposed; failing early before reeiving the state?
>> >
>> 
>> Doing so would make a number of currently existing configurations
>> invalid, even when re-enlightenment is not to be used by the
>> guest. AFAIR Windows without Hyper-V doesn't enable it. Generally, we
>> just advise people to 'enable all currently supported hyper-v
>> enlightenments' to make things easier so reenlightenment may end up
>> being added for no particular reason.
>
> Ouch, that's difficult - the problem with testing this late is that the
> migration fails right at the end so it's an unpleasent surprise.
>
> Could you disallow re-enlightenment without tsc-frequency on new machine
> types?
>

Will do. I'm not exactly sure if I should target 6.0 or 6.1 atm, let's
try the former first.

-- 
Vitaly