[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <b98fcfc6-2759-7293-b3b5-0282830c9379@redhat.com>
Date: Tue, 15 Oct 2019 10:59:05 +0200
From: Paolo Bonzini <pbonzini@...hat.com>
To: Suleiman Souhlal <suleiman@...gle.com>
Cc: rkrcmar@...hat.com, Thomas Gleixner <tglx@...utronix.de>,
John Stultz <john.stultz@...aro.org>, sboyd@...nel.org,
Linux Kernel <linux-kernel@...r.kernel.org>,
kvm@...r.kernel.org, Suleiman Souhlal <ssouhlal@...ebsd.org>,
tfiga@...omium.org, Vitaly Kuznetsov <vkuznets@...hat.com>
Subject: Re: [RFC v2 2/2] x86/kvmclock: Introduce kvm-hostclock clocksource.
On 15/10/19 10:39, Suleiman Souhlal wrote:
> I think we have that already (pvtk->flags).
> I'll change the if statement above to use pvtk instead of pv_timekeeper.
Of course, thanks.
>>> +kvm_hostclock_init(void)
>>> +{
>>> + unsigned long pa;
>>> +
>>> + pa = __pa(&pv_timekeeper);
>>> + wrmsrl(MSR_KVM_TIMEKEEPER_EN, pa);
>>
>>
>> As Vitaly said, a new CPUID bit must be defined in
>> Documentation/virt/kvm/cpuid.txt, and used here. Also please make bit 0
>> an enable bit.
>
> I think I might not be able to move the enable bit to 0 because we
> need the generation count (pvclock_timekeeper.gen) to be the first
> word of the struct due to the way we copy the data to userspace,
> similarly to how kvm_setup_pvclock_page() does it.
I mean bit 0 of the MSR.
Thanks,
Paolo
Powered by blists - more mailing lists