From: Sean Christopherson <seanjc@google.com>
To: Like Xu <like.xu.linux@gmail.com>
Cc: David Dunn <daviddunn@google.com>,
Jim Mattson <jmattson@google.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Wanpeng Li <wanpengli@tencent.com>,
Joerg Roedel <joro@8bytes.org>,
kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
Stephane Eranian <eranian@google.com>,
Peter Zijlstra <peterz@infradead.org>
Subject: Re: KVM: x86: Reconsider the current approach of vPMU
Date: Thu, 10 Feb 2022 17:12:11 +0000 [thread overview]
Message-ID: <YgVHawnQWuSAk+C1@google.com> (raw)
In-Reply-To: <39b64c56-bc8d-272d-da92-5aa29e54cdaf@gmail.com>
On Thu, Feb 10, 2022, Like Xu wrote:
> On 10/2/2022 5:00 am, Sean Christopherson wrote:
> > On Wed, Feb 09, 2022, Peter Zijlstra wrote:
> > > Guests must not unilaterally steal the PMU.
> >
> > The proposal is to add an option to allow userspace to gift the PMU to the guest,
>
> Please define the verb "gift" in more details.
Add a knob that allows host userspace to control toggle between host perf having
sole ownership of the PMU, versus ownership of the PMU being "gifted" to KVM guests
upon VM-Entry and returned back to the host at VM-Exit.
IIUC, it's the same idea as PT's PT_MODE_HOST_GUEST mode, just applied to the PMU.
By default, the host would have sole ownership, and access to the knob would be
restricted appropriately. KVM would disallow creation any VM that requires
joint ownership, e.g. launching a TDX guest would require the knob to be enabled.
> How do we balance the performance data collection needs of the
> 'hypervisor user space' and the 'system-wide profiler user space' ?
If host userspace enables the knob and transitions into a joint ownership mode,
then host userspace is explicitly acknowledging that it will no longer be able
to profile KVM guests.
Balancing between host and guest then gets factored into VM placement, e.g. VMs
that need or are paying for access to the PMU can only land on systems that are
configured for joint ownership. If profiling the guest from the host is important,
then place those guests only on hosts with sole ownership.
next prev parent reply other threads:[~2022-02-10 17:12 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-01-17 8:53 [PATCH kvm/queue v2 0/3] KVM: x86/pmu: Fix out-of-date AMD amd_event_mapping[] Like Xu
2022-01-17 8:53 ` [PATCH kvm/queue v2 1/3] KVM: x86/pmu: Replace pmu->available_event_types with a new BITMAP Like Xu
2022-02-01 12:26 ` Paolo Bonzini
2022-01-17 8:53 ` [PATCH kvm/queue v2 2/3] perf: x86/core: Add interface to query perfmon_event_map[] directly Like Xu
2022-02-01 12:27 ` Paolo Bonzini
2022-02-02 14:43 ` Peter Zijlstra
2022-02-02 22:35 ` Jim Mattson
2022-02-03 17:33 ` David Dunn
2022-02-09 8:10 ` KVM: x86: Reconsider the current approach of vPMU Like Xu
2022-02-09 13:33 ` Peter Zijlstra
2022-02-09 21:00 ` Sean Christopherson
2022-02-10 12:08 ` Like Xu
2022-02-10 17:12 ` Sean Christopherson [this message]
2022-02-16 3:33 ` Like Xu
2022-02-16 17:53 ` Jim Mattson
2022-02-09 13:21 ` [PATCH kvm/queue v2 2/3] perf: x86/core: Add interface to query perfmon_event_map[] directly Peter Zijlstra
2022-02-09 15:40 ` Dave Hansen
2022-02-09 18:47 ` Jim Mattson
2022-02-09 18:57 ` Dave Hansen
2022-02-09 19:24 ` David Dunn
2022-02-10 13:29 ` Like Xu
2022-02-10 15:34 ` Liang, Kan
2022-02-10 16:34 ` Jim Mattson
2022-02-10 18:30 ` Liang, Kan
2022-02-10 19:16 ` Jim Mattson
2022-02-10 19:46 ` Liang, Kan
2022-02-10 19:55 ` David Dunn
2022-02-11 14:11 ` Liang, Kan
2022-02-11 18:08 ` Jim Mattson
2022-02-11 21:47 ` Liang, Kan
2022-02-12 23:31 ` Jim Mattson
2022-02-14 21:55 ` Liang, Kan
2022-02-14 22:55 ` Jim Mattson
2022-02-16 7:36 ` Like Xu
2022-02-16 18:10 ` Jim Mattson
2022-02-16 7:30 ` Like Xu
2022-02-16 5:08 ` Like Xu
2022-02-10 12:55 ` Like Xu
2022-02-12 23:32 ` Jim Mattson
2022-02-08 11:52 ` Like Xu
2022-01-17 8:53 ` [PATCH kvm/queue v2 3/3] KVM: x86/pmu: Setup the {inte|amd}_event_mapping[] when hardware_setup Like Xu
2022-02-01 12:28 ` Paolo Bonzini
2022-02-08 10:10 ` Like Xu
2022-01-26 11:22 ` [PATCH kvm/queue v2 0/3] KVM: x86/pmu: Fix out-of-date AMD amd_event_mapping[] Like Xu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YgVHawnQWuSAk+C1@google.com \
--to=seanjc@google.com \
--cc=daviddunn@google.com \
--cc=eranian@google.com \
--cc=jmattson@google.com \
--cc=joro@8bytes.org \
--cc=kvm@vger.kernel.org \
--cc=like.xu.linux@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=vkuznets@redhat.com \
--cc=wanpengli@tencent.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).