From: Sean Christopherson <seanjc@google.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Wanpeng Li <wanpengli@tencent.com>,
Jim Mattson <jmattson@google.com>, Joerg Roedel <joro@8bytes.org>,
David Hildenbrand <david@redhat.com>,
David Matlack <dmatlack@google.com>,
Ben Gardon <bgardon@google.com>,
Mingwei Zhang <mizhang@google.com>
Subject: Re: [PATCH v4 21/30] KVM: x86/mmu: Zap invalidated roots via asynchronous worker
Date: Sat, 5 Mar 2022 00:34:31 +0000 [thread overview]
Message-ID: <YiKwFznqqiB9VRyn@google.com> (raw)
In-Reply-To: <8b8c28cf-cf54-f889-be7d-afc9f5430ecd@redhat.com>
On Fri, Mar 04, 2022, Paolo Bonzini wrote:
> On 3/4/22 17:02, Sean Christopherson wrote:
> > On Fri, Mar 04, 2022, Paolo Bonzini wrote:
> > > On 3/3/22 22:32, Sean Christopherson wrote:
> > > I didn't remove the paragraph from the commit message, but I think it's
> > > unnecessary now. The workqueue is flushed in kvm_mmu_zap_all_fast() and
> > > kvm_mmu_uninit_tdp_mmu(), unlike the buggy patch, so it doesn't need to take
> > > a reference to the VM.
> > >
> > > I think I don't even need to check kvm->users_count in the defunct root
> > > case, as long as kvm_mmu_uninit_tdp_mmu() flushes and destroys the workqueue
> > > before it checks that the lists are empty.
> >
> > Yes, that should work. IIRC, the WARN_ONs will tell us/you quite quickly if
> > we're wrong :-) mmu_notifier_unregister() will call the "slow" kvm_mmu_zap_all()
> > and thus ensure all non-root pages zapped, but "leaking" a worker will trigger
> > the WARN_ON that there are no roots on the list.
>
> Good, for the record these are the commit messages I have:
>
> KVM: x86/mmu: Zap invalidated roots via asynchronous worker
> Use the system worker threads to zap the roots invalidated
> by the TDP MMU's "fast zap" mechanism, implemented by
> kvm_tdp_mmu_invalidate_all_roots().
> At this point, apart from allowing some parallelism in the zapping of
> roots, the workqueue is a glorified linked list: work items are added and
> flushed entirely within a single kvm->slots_lock critical section. However,
> the workqueue fixes a latent issue where kvm_mmu_zap_all_invalidated_roots()
> assumes that it owns a reference to all invalid roots; therefore, no
> one can set the invalid bit outside kvm_mmu_zap_all_fast(). Putting the
> invalidated roots on a linked list... erm, on a workqueue ensures that
> tdp_mmu_zap_root_work() only puts back those extra references that
> kvm_mmu_zap_all_invalidated_roots() had gifted to it.
>
> and
>
> KVM: x86/mmu: Zap defunct roots via asynchronous worker
> Zap defunct roots, a.k.a. roots that have been invalidated after their
> last reference was initially dropped, asynchronously via the existing work
> queue instead of forcing the work upon the unfortunate task that happened
> to drop the last reference.
> If a vCPU task drops the last reference, the vCPU is effectively blocked
> by the host for the entire duration of the zap. If the root being zapped
> happens be fully populated with 4kb leaf SPTEs, e.g. due to dirty logging
> being active, the zap can take several hundred seconds. Unsurprisingly,
> most guests are unhappy if a vCPU disappears for hundreds of seconds.
> E.g. running a synthetic selftest that triggers a vCPU root zap with
> ~64tb of guest memory and 4kb SPTEs blocks the vCPU for 900+ seconds.
> Offloading the zap to a worker drops the block time to <100ms.
> There is an important nuance to this change. If the same work item
> was queued twice before the work function has run, it would only
> execute once and one reference would be leaked. Therefore, now that
> queueing items is not anymore protected by write_lock(&kvm->mmu_lock),
> kvm_tdp_mmu_invalidate_all_roots() has to check root->role.invalid and
> skip already invalid roots. On the other hand, kvm_mmu_zap_all_fast()
> must return only after those skipped roots have been zapped as well.
> These two requirements can be satisfied only if _all_ places that
> change invalid to true now schedule the worker before releasing the
> mmu_lock. There are just two, kvm_tdp_mmu_put_root() and
> kvm_tdp_mmu_invalidate_all_roots().
Very nice!
next prev parent reply other threads:[~2022-03-05 0:34 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-03 19:38 [PATCH v4 00/30] KVM: x86/mmu: Overhaul TDP MMU zapping and flushing Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 01/30] KVM: x86/mmu: Check for present SPTE when clearing dirty bit in TDP MMU Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 02/30] KVM: x86/mmu: Fix wrong/misleading comments in TDP MMU fast zap Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 03/30] KVM: x86/mmu: Formalize TDP MMU's (unintended?) deferred TLB flush logic Paolo Bonzini
2022-03-03 23:39 ` Mingwei Zhang
2022-03-03 19:38 ` [PATCH v4 04/30] KVM: x86/mmu: Document that zapping invalidated roots doesn't need to flush Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 05/30] KVM: x86/mmu: Require mmu_lock be held for write in unyielding root iter Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 06/30] KVM: x86/mmu: only perform eager page splitting on valid roots Paolo Bonzini
2022-03-03 20:03 ` Sean Christopherson
2022-03-03 19:38 ` [PATCH v4 07/30] KVM: x86/mmu: do not allow readers to acquire references to invalid roots Paolo Bonzini
2022-03-03 20:12 ` Sean Christopherson
2022-03-03 19:38 ` [PATCH v4 08/30] KVM: x86/mmu: Check for !leaf=>leaf, not PFN change, in TDP MMU SP removal Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 09/30] KVM: x86/mmu: Batch TLB flushes from TDP MMU for MMU notifier change_spte Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 10/30] KVM: x86/mmu: Drop RCU after processing each root in MMU notifier hooks Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 11/30] KVM: x86/mmu: Add helpers to read/write TDP MMU SPTEs and document RCU Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 12/30] KVM: x86/mmu: WARN if old _or_ new SPTE is REMOVED in non-atomic path Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 13/30] KVM: x86/mmu: Refactor low-level TDP MMU set SPTE helper to take raw values Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 14/30] KVM: x86/mmu: Zap only the target TDP MMU shadow page in NX recovery Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 15/30] KVM: x86/mmu: Skip remote TLB flush when zapping all of TDP MMU Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 16/30] KVM: x86/mmu: Add dedicated helper to zap TDP MMU root shadow page Paolo Bonzini
2022-03-04 0:07 ` Mingwei Zhang
2022-03-03 19:38 ` [PATCH v4 17/30] KVM: x86/mmu: Require mmu_lock be held for write to zap TDP MMU range Paolo Bonzini
2022-03-04 0:14 ` Mingwei Zhang
2022-03-03 19:38 ` [PATCH v4 18/30] KVM: x86/mmu: Zap only TDP MMU leafs in kvm_zap_gfn_range() Paolo Bonzini
2022-03-04 1:16 ` Mingwei Zhang
2022-03-04 16:11 ` Sean Christopherson
2022-03-04 18:00 ` Mingwei Zhang
2022-03-04 18:42 ` Sean Christopherson
2022-03-11 15:09 ` Vitaly Kuznetsov
2022-03-13 18:40 ` Mingwei Zhang
2022-03-25 15:13 ` Sean Christopherson
2022-03-26 18:10 ` Mingwei Zhang
2022-03-28 15:06 ` Sean Christopherson
2022-03-03 19:38 ` [PATCH v4 19/30] KVM: x86/mmu: Do remote TLB flush before dropping RCU in TDP MMU resched Paolo Bonzini
2022-03-04 1:19 ` Mingwei Zhang
2022-03-03 19:38 ` [PATCH v4 20/30] KVM: x86/mmu: Defer TLB flush to caller when freeing TDP MMU shadow pages Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 21/30] KVM: x86/mmu: Zap invalidated roots via asynchronous worker Paolo Bonzini
2022-03-03 20:54 ` Sean Christopherson
2022-03-03 21:06 ` Sean Christopherson
2022-03-03 21:20 ` Sean Christopherson
2022-03-03 21:32 ` Sean Christopherson
2022-03-04 6:48 ` Paolo Bonzini
2022-03-04 16:02 ` Sean Christopherson
2022-03-04 18:11 ` Paolo Bonzini
2022-03-05 0:34 ` Sean Christopherson [this message]
2022-03-05 19:53 ` Paolo Bonzini
2022-03-08 21:29 ` Sean Christopherson
2022-03-11 17:50 ` Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 22/30] KVM: x86/mmu: Allow yielding when zapping GFNs for defunct TDP MMU root Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 23/30] KVM: x86/mmu: Zap roots in two passes to avoid inducing RCU stalls Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 24/30] KVM: x86/mmu: Zap defunct roots via asynchronous worker Paolo Bonzini
2022-03-03 22:08 ` Sean Christopherson
2022-03-03 19:38 ` [PATCH v4 25/30] KVM: x86/mmu: Check for a REMOVED leaf SPTE before making the SPTE Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 26/30] KVM: x86/mmu: WARN on any attempt to atomically update REMOVED SPTE Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 27/30] KVM: selftests: Move raw KVM_SET_USER_MEMORY_REGION helper to utils Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 28/30] KVM: selftests: Split out helper to allocate guest mem via memfd Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 29/30] KVM: selftests: Define cpu_relax() helpers for s390 and x86 Paolo Bonzini
2022-03-03 19:38 ` [PATCH v4 30/30] KVM: selftests: Add test to populate a VM with the max possible guest mem Paolo Bonzini
2022-03-08 14:47 ` Paolo Bonzini
2022-03-08 15:36 ` Christian Borntraeger
2022-03-08 21:09 ` Sean Christopherson
2022-03-08 17:25 ` [PATCH v4 00/30] KVM: x86/mmu: Overhaul TDP MMU zapping and flushing Paolo Bonzini
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=YiKwFznqqiB9VRyn@google.com \
--to=seanjc@google.com \
--cc=bgardon@google.com \
--cc=david@redhat.com \
--cc=dmatlack@google.com \
--cc=jmattson@google.com \
--cc=joro@8bytes.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mizhang@google.com \
--cc=pbonzini@redhat.com \
--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).