From: Sean Christopherson <seanjc@google.com>
To: David Woodhouse <dwmw2@infradead.org>
Cc: Paul Durrant <paul@xen.org>, Paolo Bonzini <pbonzini@redhat.com>,
Jonathan Corbet <corbet@lwn.net>,
Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
Dave Hansen <dave.hansen@linux.intel.com>,
x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
Shuah Khan <shuah@kernel.org>,
kvm@vger.kernel.org, linux-doc@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org
Subject: Re: [PATCH v12 18/20] KVM: pfncache: check the need for invalidation under read lock first
Date: Wed, 7 Feb 2024 07:10:28 -0800 [thread overview]
Message-ID: <ZcOdZKmmYz3kMgwp@google.com> (raw)
In-Reply-To: <a817d64f3fe7b935a02e78df02dc0c6281e61af3.camel@infradead.org>
On Tue, Feb 06, 2024, David Woodhouse wrote:
> On Tue, 2024-02-06 at 20:47 -0800, Sean Christopherson wrote:
> >
> > I'm saying this:
> >
> > When processing mmu_notifier invalidations for gpc caches, pre-check for
> > overlap with the invalidation event while holding gpc->lock for read, and
> > only take gpc->lock for write if the cache needs to be invalidated. Doing
> > a pre-check without taking gpc->lock for write avoids unnecessarily
> > contending the lock for unrelated invalidations, which is very beneficial
> > for caches that are heavily used (but rarely subjected to mmu_notifier
> > invalidations).
> >
> > is much friendlier to readers than this:
> >
> > Taking a write lock on a pfncache will be disruptive if the cache is
> > heavily used (which only requires a read lock). Hence, in the MMU notifier
> > callback, take read locks on caches to check for a match; only taking a
> > write lock to actually perform an invalidation (after a another check).
>
> That's a somewhat subjective observation. I actually find the latter to
> be far more succinct and obvious.
>
> Actually... maybe I find yours harder because it isn't actually stating
> the situation as I understand it. You said "unrelated invalidation" in
> your first email, and "overlap with the invalidation event" in this
> one... neither of which makes sense to me because there is no *other*
> invalidation here.
I am referring to the "mmu_notifier invalidation event". While a particular GPC
may not be affected by the invalidation, it's entirely possible that a different
GPC and/or some chunk of guest memory does need to be invalidated/zapped.
> We're only talking about the MMU notifier gratuitously taking the write
It's not "the MMU notifier" though, it's KVM that unnecessarily takes a lock. I
know I'm being somewhat pedantic, but the distinction does matter. E.g. with
guest_memfd, there will be invalidations that get routed through this code, but
that do not originate in the mmu_notifier.
And I think it's important to make it clear to readers that an mmu_notifier really
just is a notification from the primary MMU, albeit a notification that comes with
a rather strict contract.
> lock on a GPC that it *isn't* going to invalidate (the common case),
> and that disrupting users which are trying to take the read lock on
> that GPC.
next prev parent reply other threads:[~2024-02-07 15:10 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-15 12:56 [PATCH v12 00/20] KVM: xen: update shared_info and vcpu_info handling Paul Durrant
2024-01-15 12:56 ` [PATCH v12 01/20] KVM: pfncache: Add a map helper function Paul Durrant
2024-01-15 12:56 ` [PATCH v12 02/20] KVM: pfncache: remove unnecessary exports Paul Durrant
2024-01-15 12:56 ` [PATCH v12 03/20] KVM: xen: mark guest pages dirty with the pfncache lock held Paul Durrant
2024-02-07 3:17 ` Sean Christopherson
2024-02-07 3:26 ` David Woodhouse
2024-02-07 15:15 ` Sean Christopherson
2024-02-07 8:48 ` Paul Durrant
2024-01-15 12:56 ` [PATCH v12 04/20] KVM: pfncache: add a mark-dirty helper Paul Durrant
2024-02-07 3:20 ` Sean Christopherson
2024-02-07 8:47 ` Paul Durrant
2024-02-09 15:58 ` Sean Christopherson
2024-02-09 16:05 ` Paul Durrant
2024-01-15 12:56 ` [PATCH v12 05/20] KVM: pfncache: remove KVM_GUEST_USES_PFN usage Paul Durrant
2024-01-15 12:56 ` [PATCH v12 06/20] KVM: pfncache: stop open-coding offset_in_page() Paul Durrant
2024-01-15 12:56 ` [PATCH v12 07/20] KVM: pfncache: include page offset in uhva and use it consistently Paul Durrant
2024-01-15 12:56 ` [PATCH v12 08/20] KVM: pfncache: allow a cache to be activated with a fixed (userspace) HVA Paul Durrant
2024-02-07 4:03 ` Sean Christopherson
2024-02-07 4:13 ` David Woodhouse
2024-02-14 16:01 ` Sean Christopherson
2024-02-14 16:09 ` Paul Durrant
2024-02-14 15:21 ` Paul Durrant
2024-02-14 16:20 ` Sean Christopherson
2024-02-14 16:33 ` Paul Durrant
2024-01-15 12:56 ` [PATCH v12 09/20] KVM: xen: separate initialization of shared_info cache and content Paul Durrant
2024-01-15 12:56 ` [PATCH v12 10/20] KVM: xen: re-initialize shared_info if guest (32/64-bit) mode is set Paul Durrant
2024-01-15 12:56 ` [PATCH v12 11/20] KVM: xen: allow shared_info to be mapped by fixed HVA Paul Durrant
2024-02-07 4:10 ` Sean Christopherson
2024-02-07 8:53 ` Paul Durrant
2024-02-08 8:52 ` Paul Durrant
2024-02-08 16:48 ` Sean Christopherson
2024-02-08 16:51 ` Paul Durrant
2024-02-08 17:26 ` David Woodhouse
2024-02-09 16:01 ` Sean Christopherson
2024-01-15 12:56 ` [PATCH v12 12/20] KVM: xen: allow vcpu_info " Paul Durrant
2024-01-15 12:57 ` [PATCH v12 13/20] KVM: selftests / xen: map shared_info using HVA rather than GFN Paul Durrant
2024-02-07 4:14 ` Sean Christopherson
2024-02-07 8:54 ` Paul Durrant
2024-02-07 14:58 ` Sean Christopherson
2024-01-15 12:57 ` [PATCH v12 14/20] KVM: selftests / xen: re-map vcpu_info using HVA rather than GPA Paul Durrant
2024-01-15 12:57 ` [PATCH v12 15/20] KVM: xen: advertize the KVM_XEN_HVM_CONFIG_SHARED_INFO_HVA capability Paul Durrant
2024-01-15 12:57 ` [PATCH v12 16/20] KVM: xen: split up kvm_xen_set_evtchn_fast() Paul Durrant
2024-01-15 12:57 ` [PATCH v12 17/20] KVM: xen: don't block on pfncache locks in kvm_xen_set_evtchn_fast() Paul Durrant
2024-02-07 4:17 ` Sean Christopherson
2024-02-07 4:21 ` David Woodhouse
2024-01-15 12:57 ` [PATCH v12 18/20] KVM: pfncache: check the need for invalidation under read lock first Paul Durrant
2024-02-07 4:22 ` Sean Christopherson
2024-02-07 4:27 ` David Woodhouse
2024-02-07 4:47 ` Sean Christopherson
2024-02-07 4:59 ` David Woodhouse
2024-02-07 15:10 ` Sean Christopherson [this message]
2024-01-15 12:57 ` [PATCH v12 19/20] KVM: xen: allow vcpu_info content to be 'safely' copied Paul Durrant
2024-01-15 12:57 ` [PATCH v12 20/20] KVM: pfncache: rework __kvm_gpc_refresh() to fix locking issues Paul Durrant
2024-01-25 15:03 ` [PATCH v12 00/20] KVM: xen: update shared_info and vcpu_info handling Paul Durrant
2024-01-25 20:07 ` David Woodhouse
2024-01-26 1:19 ` Sean Christopherson
2024-02-02 17:37 ` Paul Durrant
2024-02-02 22:03 ` Sean Christopherson
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=ZcOdZKmmYz3kMgwp@google.com \
--to=seanjc@google.com \
--cc=bp@alien8.de \
--cc=corbet@lwn.net \
--cc=dave.hansen@linux.intel.com \
--cc=dwmw2@infradead.org \
--cc=hpa@zytor.com \
--cc=kvm@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=paul@xen.org \
--cc=pbonzini@redhat.com \
--cc=shuah@kernel.org \
--cc=tglx@linutronix.de \
--cc=x86@kernel.org \
/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).