From: Paolo Bonzini <pbonzini@redhat.com>
To: Marc Zyngier <maz@kernel.org>
Cc: Peter Xu <peterx@redhat.com>, Gavin Shan <gshan@redhat.com>,
kvmarm@lists.cs.columbia.edu,
linux-arm-kernel@lists.infradead.org, kvm@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org,
linux-kselftest@vger.kernel.org, corbet@lwn.net,
james.morse@arm.com, alexandru.elisei@arm.com,
suzuki.poulose@arm.com, oliver.upton@linux.dev,
catalin.marinas@arm.com, will@kernel.org, shuah@kernel.org,
seanjc@google.com, dmatlack@google.com, bgardon@google.com,
ricarkol@google.com, zhenyzha@redhat.com, shan.gavin@gmail.com
Subject: Re: [PATCH v1 1/5] KVM: arm64: Enable ring-based dirty memory tracking
Date: Sat, 27 Aug 2022 10:27:08 +0200 [thread overview]
Message-ID: <79cc1418-2448-6a80-e4b8-2041f94c419e@redhat.com> (raw)
In-Reply-To: <8735djvwbu.wl-maz@kernel.org>
On 8/26/22 17:49, Marc Zyngier wrote:
>> Agreed, but that's a problem for userspace to solve. If userspace
>> wants to reset the fields in different CPUs, it has to synchronize
>> with its own invoking of the ioctl.
>
> userspace has no choice. It cannot order on its own the reads that the
> kernel will do to *other* rings.
Those reads will never see KVM_DIRTY_GFN_F_RESET in the flags however,
if userspace has never interacted with the ring. So there will be
exactly one read on those rings, and there's nothing to reorder.
If that's too tricky and you want to add a load-acquire I have no
objection though. It also helps avoiding read-read reordering between
one entry's flags to the next one's, so it's a good idea to have it anyway.
>> The main reason why I preferred a global KVM_RESET_DIRTY_RINGS ioctl
>> was because it takes kvm->slots_lock so the execution would be
>> serialized anyway. Turning slots_lock into an rwsem would be even
>> worse because it also takes kvm->mmu_lock (since slots_lock is a
>> mutex, at least two concurrent invocations won't clash with each other
>> on the mmu_lock).
>
> Whatever the reason, the behaviour should be identical on all
> architectures. As is is, it only really works on x86, and I contend
> this is a bug that needs fixing.
>
> Thankfully, this can be done at zero cost for x86, and at that of a
> set of load-acquires on other architectures.
Yes, the global-ness of the API is orthogonal to the memory ordering
issue. I just wanted to explain why a per-vCPU API probably isn't going
to work great.
Paolo
WARNING: multiple messages have this Message-ID (diff)
From: Paolo Bonzini <pbonzini@redhat.com>
To: Marc Zyngier <maz@kernel.org>
Cc: kvm@vger.kernel.org, linux-doc@vger.kernel.org,
catalin.marinas@arm.com, linux-kselftest@vger.kernel.org,
bgardon@google.com, shuah@kernel.org,
kvmarm@lists.cs.columbia.edu, corbet@lwn.net, will@kernel.org,
shan.gavin@gmail.com, zhenyzha@redhat.com, dmatlack@google.com,
linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v1 1/5] KVM: arm64: Enable ring-based dirty memory tracking
Date: Sat, 27 Aug 2022 10:27:08 +0200 [thread overview]
Message-ID: <79cc1418-2448-6a80-e4b8-2041f94c419e@redhat.com> (raw)
In-Reply-To: <8735djvwbu.wl-maz@kernel.org>
On 8/26/22 17:49, Marc Zyngier wrote:
>> Agreed, but that's a problem for userspace to solve. If userspace
>> wants to reset the fields in different CPUs, it has to synchronize
>> with its own invoking of the ioctl.
>
> userspace has no choice. It cannot order on its own the reads that the
> kernel will do to *other* rings.
Those reads will never see KVM_DIRTY_GFN_F_RESET in the flags however,
if userspace has never interacted with the ring. So there will be
exactly one read on those rings, and there's nothing to reorder.
If that's too tricky and you want to add a load-acquire I have no
objection though. It also helps avoiding read-read reordering between
one entry's flags to the next one's, so it's a good idea to have it anyway.
>> The main reason why I preferred a global KVM_RESET_DIRTY_RINGS ioctl
>> was because it takes kvm->slots_lock so the execution would be
>> serialized anyway. Turning slots_lock into an rwsem would be even
>> worse because it also takes kvm->mmu_lock (since slots_lock is a
>> mutex, at least two concurrent invocations won't clash with each other
>> on the mmu_lock).
>
> Whatever the reason, the behaviour should be identical on all
> architectures. As is is, it only really works on x86, and I contend
> this is a bug that needs fixing.
>
> Thankfully, this can be done at zero cost for x86, and at that of a
> set of load-acquires on other architectures.
Yes, the global-ness of the API is orthogonal to the memory ordering
issue. I just wanted to explain why a per-vCPU API probably isn't going
to work great.
Paolo
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
WARNING: multiple messages have this Message-ID (diff)
From: Paolo Bonzini <pbonzini@redhat.com>
To: Marc Zyngier <maz@kernel.org>
Cc: Peter Xu <peterx@redhat.com>, Gavin Shan <gshan@redhat.com>,
kvmarm@lists.cs.columbia.edu,
linux-arm-kernel@lists.infradead.org, kvm@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-doc@vger.kernel.org,
linux-kselftest@vger.kernel.org, corbet@lwn.net,
james.morse@arm.com, alexandru.elisei@arm.com,
suzuki.poulose@arm.com, oliver.upton@linux.dev,
catalin.marinas@arm.com, will@kernel.org, shuah@kernel.org,
seanjc@google.com, dmatlack@google.com, bgardon@google.com,
ricarkol@google.com, zhenyzha@redhat.com, shan.gavin@gmail.com
Subject: Re: [PATCH v1 1/5] KVM: arm64: Enable ring-based dirty memory tracking
Date: Sat, 27 Aug 2022 10:27:08 +0200 [thread overview]
Message-ID: <79cc1418-2448-6a80-e4b8-2041f94c419e@redhat.com> (raw)
In-Reply-To: <8735djvwbu.wl-maz@kernel.org>
On 8/26/22 17:49, Marc Zyngier wrote:
>> Agreed, but that's a problem for userspace to solve. If userspace
>> wants to reset the fields in different CPUs, it has to synchronize
>> with its own invoking of the ioctl.
>
> userspace has no choice. It cannot order on its own the reads that the
> kernel will do to *other* rings.
Those reads will never see KVM_DIRTY_GFN_F_RESET in the flags however,
if userspace has never interacted with the ring. So there will be
exactly one read on those rings, and there's nothing to reorder.
If that's too tricky and you want to add a load-acquire I have no
objection though. It also helps avoiding read-read reordering between
one entry's flags to the next one's, so it's a good idea to have it anyway.
>> The main reason why I preferred a global KVM_RESET_DIRTY_RINGS ioctl
>> was because it takes kvm->slots_lock so the execution would be
>> serialized anyway. Turning slots_lock into an rwsem would be even
>> worse because it also takes kvm->mmu_lock (since slots_lock is a
>> mutex, at least two concurrent invocations won't clash with each other
>> on the mmu_lock).
>
> Whatever the reason, the behaviour should be identical on all
> architectures. As is is, it only really works on x86, and I contend
> this is a bug that needs fixing.
>
> Thankfully, this can be done at zero cost for x86, and at that of a
> set of load-acquires on other architectures.
Yes, the global-ness of the API is orthogonal to the memory ordering
issue. I just wanted to explain why a per-vCPU API probably isn't going
to work great.
Paolo
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2022-08-27 8:27 UTC|newest]
Thread overview: 98+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-08-19 0:55 [PATCH v1 0/5] KVM: arm64: Enable ring-based dirty memory tracking Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 0:55 ` [PATCH v1 1/5] " Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 8:00 ` Marc Zyngier
2022-08-19 8:00 ` Marc Zyngier
2022-08-19 8:00 ` Marc Zyngier
2022-08-22 1:58 ` Gavin Shan
2022-08-22 1:58 ` Gavin Shan
2022-08-22 18:55 ` Peter Xu
2022-08-22 18:55 ` Peter Xu
2022-08-22 18:55 ` Peter Xu
2022-08-23 3:19 ` Gavin Shan
2022-08-23 3:19 ` Gavin Shan
2022-08-23 3:19 ` Gavin Shan
2022-08-22 21:42 ` Marc Zyngier
2022-08-22 21:42 ` Marc Zyngier
2022-08-22 21:42 ` Marc Zyngier
2022-08-23 5:22 ` Gavin Shan
2022-08-23 5:22 ` Gavin Shan
2022-08-23 5:22 ` Gavin Shan
2022-08-23 13:58 ` Peter Xu
2022-08-23 13:58 ` Peter Xu
2022-08-23 13:58 ` Peter Xu
2022-08-23 19:17 ` Marc Zyngier
2022-08-23 19:17 ` Marc Zyngier
2022-08-23 19:17 ` Marc Zyngier
2022-08-23 21:20 ` Peter Xu
2022-08-23 21:20 ` Peter Xu
2022-08-23 21:20 ` Peter Xu
2022-08-23 22:47 ` Marc Zyngier
2022-08-23 22:47 ` Marc Zyngier
2022-08-23 22:47 ` Marc Zyngier
2022-08-23 23:19 ` Peter Xu
2022-08-23 23:19 ` Peter Xu
2022-08-23 23:19 ` Peter Xu
2022-08-24 14:45 ` Marc Zyngier
2022-08-24 14:45 ` Marc Zyngier
2022-08-24 14:45 ` Marc Zyngier
2022-08-24 16:21 ` Peter Xu
2022-08-24 16:21 ` Peter Xu
2022-08-24 16:21 ` Peter Xu
2022-08-24 20:57 ` Marc Zyngier
2022-08-24 20:57 ` Marc Zyngier
2022-08-24 20:57 ` Marc Zyngier
2022-08-26 6:05 ` Gavin Shan
2022-08-26 6:05 ` Gavin Shan
2022-08-26 6:05 ` Gavin Shan
2022-08-26 10:50 ` Paolo Bonzini
2022-08-26 10:50 ` Paolo Bonzini
2022-08-26 10:50 ` Paolo Bonzini
2022-08-26 15:49 ` Marc Zyngier
2022-08-26 15:49 ` Marc Zyngier
2022-08-26 15:49 ` Marc Zyngier
2022-08-27 8:27 ` Paolo Bonzini [this message]
2022-08-27 8:27 ` Paolo Bonzini
2022-08-27 8:27 ` Paolo Bonzini
2022-08-29 10:27 ` Paolo Bonzini
2022-08-23 14:44 ` Oliver Upton
2022-08-23 14:44 ` Oliver Upton
2022-08-23 14:44 ` Oliver Upton
2022-08-23 20:35 ` Marc Zyngier
2022-08-23 20:35 ` Marc Zyngier
2022-08-23 20:35 ` Marc Zyngier
2022-08-26 10:58 ` Paolo Bonzini
2022-08-26 10:58 ` Paolo Bonzini
2022-08-26 10:58 ` Paolo Bonzini
2022-08-26 15:28 ` Marc Zyngier
2022-08-26 15:28 ` Marc Zyngier
2022-08-26 15:28 ` Marc Zyngier
2022-08-30 14:42 ` Peter Xu
2022-08-30 14:42 ` Peter Xu
2022-08-30 14:42 ` Peter Xu
2022-09-02 0:19 ` Paolo Bonzini
2022-09-02 0:19 ` Paolo Bonzini
2022-09-02 0:19 ` Paolo Bonzini
2022-08-19 0:55 ` [PATCH v1 2/5] KVM: selftests: Use host page size to map ring buffer in dirty_log_test Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 0:55 ` [PATCH v1 3/5] KVM: selftests: Dirty host pages " Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 0:55 ` Gavin Shan
2022-08-19 5:28 ` Andrew Jones
2022-08-19 5:28 ` Andrew Jones
2022-08-19 5:28 ` Andrew Jones
2022-08-22 6:29 ` Gavin Shan
2022-08-22 6:29 ` Gavin Shan
2022-08-23 3:09 ` Gavin Shan
2022-08-23 3:09 ` Gavin Shan
2022-08-23 3:09 ` Gavin Shan
2022-08-19 0:56 ` [PATCH v1 4/5] KVM: selftests: Clear dirty ring states between two modes " Gavin Shan
2022-08-19 0:56 ` Gavin Shan
2022-08-19 0:56 ` Gavin Shan
2022-08-19 0:56 ` [PATCH v1 5/5] KVM: selftests: Automate choosing dirty ring size " Gavin Shan
2022-08-19 0:56 ` Gavin Shan
2022-08-19 0:56 ` Gavin Shan
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=79cc1418-2448-6a80-e4b8-2041f94c419e@redhat.com \
--to=pbonzini@redhat.com \
--cc=alexandru.elisei@arm.com \
--cc=bgardon@google.com \
--cc=catalin.marinas@arm.com \
--cc=corbet@lwn.net \
--cc=dmatlack@google.com \
--cc=gshan@redhat.com \
--cc=james.morse@arm.com \
--cc=kvm@vger.kernel.org \
--cc=kvmarm@lists.cs.columbia.edu \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=maz@kernel.org \
--cc=oliver.upton@linux.dev \
--cc=peterx@redhat.com \
--cc=ricarkol@google.com \
--cc=seanjc@google.com \
--cc=shan.gavin@gmail.com \
--cc=shuah@kernel.org \
--cc=suzuki.poulose@arm.com \
--cc=will@kernel.org \
--cc=zhenyzha@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.