From: Nicolas Saenz Julienne <nsaenzju@redhat.com>
To: Mark Rutland <mark.rutland@arm.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: paulmck <paulmck@kernel.org>, maz <maz@kernel.org>,
frederic <frederic@kernel.org>,
linux-kernel <linux-kernel@vger.kernel.org>,
rcu <rcu@vger.kernel.org>, Thomas Gleixner <tglx@linutronix.de>,
Will Deacon <will@kernel.org>,
kvmarm@lists.cs.columbia.edu,
linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
Anup Patel <Anup.Patel@wdc.com>
Subject: Re: Possible nohz-full/RCU issue in arm64 KVM
Date: Tue, 11 Jan 2022 12:32:38 +0100 [thread overview]
Message-ID: <399d8805ca09f7d3c905b8c653abf60dd7141574.camel@redhat.com> (raw)
In-Reply-To: <YdR4N9QVYOzjowAb@FVFF77S0Q05N>
Hi Mark,
On Tue, 2022-01-04 at 16:39 +0000, Mark Rutland wrote:
> On Fri, Dec 17, 2021 at 04:54:22PM +0100, Paolo Bonzini wrote:
> > On 12/17/21 15:38, Mark Rutland wrote:
> > > For example kvm_guest_enter_irqoff() calls guest_enter_irq_off() which calls
> > > vtime_account_guest_enter(), but kvm_guest_exit_irqoff() doesn't call
> > > guest_exit_irq_off() and the call to vtime_account_guest_exit() is open-coded
> > > elsewhere. Also, guest_enter_irq_off() conditionally calls
> > > rcu_virt_note_context_switch(), but I can't immediately spot anything on the
> > > exit side that corresponded with that, which looks suspicious.
> >
> > rcu_note_context_switch() is a point-in-time notification; it's not strictly
> > necessary, but it may improve performance a bit by avoiding unnecessary IPIs
> > from the RCU subsystem.
> >
> > There's no benefit from doing it when you're back from the guest, because at
> > that point the CPU is just running normal kernel code.
>
> I see.
>
> My main issue here was just that it's really difficult to see how the
> entry/exit logic is balanced, and I reckon we can solve that by splitting
> guest_{enter,exit}_irqoff() into helper functions to handle the vtime
> accounting separately from the context tracking, so that arch code can do
> something like:
>
> guest_timing_enter_irqoff();
>
> guest_eqs_enter_irqoff();
> < actually run vCPU here >
> guest_eqs_exit_irqoff();
>
> < handle pending IRQs here >
>
> guest_timing_exit_irqoff();
>
> ... which I hope should work for RISC-V too.
>
> I've had a go, and I've pushed out a WIP to:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/log/?h=arm64/kvm/rcu
Had a look at the patches and they seeem OK to me.
Thanks!
--
Nicolás Sáenz
WARNING: multiple messages have this Message-ID (diff)
From: Nicolas Saenz Julienne <nsaenzju@redhat.com>
To: Mark Rutland <mark.rutland@arm.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: paulmck <paulmck@kernel.org>, maz <maz@kernel.org>,
frederic <frederic@kernel.org>,
linux-kernel <linux-kernel@vger.kernel.org>,
rcu <rcu@vger.kernel.org>, Thomas Gleixner <tglx@linutronix.de>,
Will Deacon <will@kernel.org>,
kvmarm@lists.cs.columbia.edu,
linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
Anup Patel <Anup.Patel@wdc.com>
Subject: Re: Possible nohz-full/RCU issue in arm64 KVM
Date: Tue, 11 Jan 2022 12:32:38 +0100 [thread overview]
Message-ID: <399d8805ca09f7d3c905b8c653abf60dd7141574.camel@redhat.com> (raw)
In-Reply-To: <YdR4N9QVYOzjowAb@FVFF77S0Q05N>
Hi Mark,
On Tue, 2022-01-04 at 16:39 +0000, Mark Rutland wrote:
> On Fri, Dec 17, 2021 at 04:54:22PM +0100, Paolo Bonzini wrote:
> > On 12/17/21 15:38, Mark Rutland wrote:
> > > For example kvm_guest_enter_irqoff() calls guest_enter_irq_off() which calls
> > > vtime_account_guest_enter(), but kvm_guest_exit_irqoff() doesn't call
> > > guest_exit_irq_off() and the call to vtime_account_guest_exit() is open-coded
> > > elsewhere. Also, guest_enter_irq_off() conditionally calls
> > > rcu_virt_note_context_switch(), but I can't immediately spot anything on the
> > > exit side that corresponded with that, which looks suspicious.
> >
> > rcu_note_context_switch() is a point-in-time notification; it's not strictly
> > necessary, but it may improve performance a bit by avoiding unnecessary IPIs
> > from the RCU subsystem.
> >
> > There's no benefit from doing it when you're back from the guest, because at
> > that point the CPU is just running normal kernel code.
>
> I see.
>
> My main issue here was just that it's really difficult to see how the
> entry/exit logic is balanced, and I reckon we can solve that by splitting
> guest_{enter,exit}_irqoff() into helper functions to handle the vtime
> accounting separately from the context tracking, so that arch code can do
> something like:
>
> guest_timing_enter_irqoff();
>
> guest_eqs_enter_irqoff();
> < actually run vCPU here >
> guest_eqs_exit_irqoff();
>
> < handle pending IRQs here >
>
> guest_timing_exit_irqoff();
>
> ... which I hope should work for RISC-V too.
>
> I've had a go, and I've pushed out a WIP to:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/log/?h=arm64/kvm/rcu
Had a look at the patches and they seeem OK to me.
Thanks!
--
Nicolás Sáenz
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
WARNING: multiple messages have this Message-ID (diff)
From: Nicolas Saenz Julienne <nsaenzju@redhat.com>
To: Mark Rutland <mark.rutland@arm.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: paulmck <paulmck@kernel.org>, maz <maz@kernel.org>,
frederic <frederic@kernel.org>, Anup Patel <Anup.Patel@wdc.com>,
linux-kernel <linux-kernel@vger.kernel.org>,
rcu <rcu@vger.kernel.org>, Thomas Gleixner <tglx@linutronix.de>,
Will Deacon <will@kernel.org>,
kvmarm@lists.cs.columbia.edu,
linux-arm-kernel <linux-arm-kernel@lists.infradead.org>
Subject: Re: Possible nohz-full/RCU issue in arm64 KVM
Date: Tue, 11 Jan 2022 12:32:38 +0100 [thread overview]
Message-ID: <399d8805ca09f7d3c905b8c653abf60dd7141574.camel@redhat.com> (raw)
In-Reply-To: <YdR4N9QVYOzjowAb@FVFF77S0Q05N>
Hi Mark,
On Tue, 2022-01-04 at 16:39 +0000, Mark Rutland wrote:
> On Fri, Dec 17, 2021 at 04:54:22PM +0100, Paolo Bonzini wrote:
> > On 12/17/21 15:38, Mark Rutland wrote:
> > > For example kvm_guest_enter_irqoff() calls guest_enter_irq_off() which calls
> > > vtime_account_guest_enter(), but kvm_guest_exit_irqoff() doesn't call
> > > guest_exit_irq_off() and the call to vtime_account_guest_exit() is open-coded
> > > elsewhere. Also, guest_enter_irq_off() conditionally calls
> > > rcu_virt_note_context_switch(), but I can't immediately spot anything on the
> > > exit side that corresponded with that, which looks suspicious.
> >
> > rcu_note_context_switch() is a point-in-time notification; it's not strictly
> > necessary, but it may improve performance a bit by avoiding unnecessary IPIs
> > from the RCU subsystem.
> >
> > There's no benefit from doing it when you're back from the guest, because at
> > that point the CPU is just running normal kernel code.
>
> I see.
>
> My main issue here was just that it's really difficult to see how the
> entry/exit logic is balanced, and I reckon we can solve that by splitting
> guest_{enter,exit}_irqoff() into helper functions to handle the vtime
> accounting separately from the context tracking, so that arch code can do
> something like:
>
> guest_timing_enter_irqoff();
>
> guest_eqs_enter_irqoff();
> < actually run vCPU here >
> guest_eqs_exit_irqoff();
>
> < handle pending IRQs here >
>
> guest_timing_exit_irqoff();
>
> ... which I hope should work for RISC-V too.
>
> I've had a go, and I've pushed out a WIP to:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/log/?h=arm64/kvm/rcu
Had a look at the patches and they seeem OK to me.
Thanks!
--
Nicolás Sáenz
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
next prev parent reply other threads:[~2022-01-11 11:32 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-17 11:51 Possible nohz-full/RCU issue in arm64 KVM Nicolas Saenz Julienne
2021-12-17 11:51 ` Nicolas Saenz Julienne
2021-12-17 11:51 ` Nicolas Saenz Julienne
2021-12-17 13:21 ` Mark Rutland
2021-12-17 13:21 ` Mark Rutland
2021-12-17 13:21 ` Mark Rutland
2021-12-17 14:15 ` Nicolas Saenz Julienne
2021-12-17 14:15 ` Nicolas Saenz Julienne
2021-12-17 14:15 ` Nicolas Saenz Julienne
2021-12-17 14:38 ` Mark Rutland
2021-12-17 14:38 ` Mark Rutland
2021-12-17 14:38 ` Mark Rutland
2021-12-17 15:54 ` Paolo Bonzini
2021-12-17 15:54 ` Paolo Bonzini
2021-12-17 15:54 ` Paolo Bonzini
2021-12-17 16:07 ` Paul E. McKenney
2021-12-17 16:07 ` Paul E. McKenney
2021-12-17 16:07 ` Paul E. McKenney
2021-12-17 16:20 ` Nicolas Saenz Julienne
2021-12-17 16:20 ` Nicolas Saenz Julienne
2021-12-17 16:20 ` Nicolas Saenz Julienne
2021-12-17 16:43 ` Paul E. McKenney
2021-12-17 16:43 ` Paul E. McKenney
2021-12-17 16:43 ` Paul E. McKenney
2021-12-17 16:34 ` Paolo Bonzini
2021-12-17 16:34 ` Paolo Bonzini
2021-12-17 16:34 ` Paolo Bonzini
2021-12-17 16:45 ` Paul E. McKenney
2021-12-17 16:45 ` Paul E. McKenney
2021-12-17 16:45 ` Paul E. McKenney
2021-12-17 17:02 ` Paolo Bonzini
2021-12-17 17:02 ` Paolo Bonzini
2021-12-17 17:02 ` Paolo Bonzini
2021-12-17 17:12 ` Paul E. McKenney
2021-12-17 17:12 ` Paul E. McKenney
2021-12-17 17:12 ` Paul E. McKenney
2021-12-17 17:23 ` Paolo Bonzini
2021-12-17 17:23 ` Paolo Bonzini
2021-12-17 17:23 ` Paolo Bonzini
2021-12-17 17:47 ` Paul E. McKenney
2021-12-17 17:47 ` Paul E. McKenney
2021-12-17 17:47 ` Paul E. McKenney
2022-01-04 16:39 ` Mark Rutland
2022-01-04 16:39 ` Mark Rutland
2022-01-04 16:39 ` Mark Rutland
2022-01-04 17:07 ` Paolo Bonzini
2022-01-04 17:07 ` Paolo Bonzini
2022-01-04 17:07 ` Paolo Bonzini
2022-01-11 11:32 ` Nicolas Saenz Julienne [this message]
2022-01-11 11:32 ` Nicolas Saenz Julienne
2022-01-11 11:32 ` Nicolas Saenz Julienne
2022-01-11 12:23 ` Mark Rutland
2022-01-11 12:23 ` Mark Rutland
2022-01-11 12:23 ` Mark Rutland
2021-12-17 14:51 ` Paolo Bonzini
2021-12-17 14:51 ` Paolo Bonzini
2021-12-17 14:51 ` Paolo Bonzini
2021-12-20 14:28 ` Marc Zyngier
2021-12-20 14:28 ` Marc Zyngier
2021-12-20 14:28 ` Marc Zyngier
2021-12-20 16:10 ` Frederic Weisbecker
2021-12-20 16:10 ` Frederic Weisbecker
2021-12-20 16:10 ` Frederic Weisbecker
2022-01-04 13:24 ` Mark Rutland
2022-01-04 13:24 ` Mark Rutland
2022-01-04 13:24 ` Mark Rutland
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=399d8805ca09f7d3c905b8c653abf60dd7141574.camel@redhat.com \
--to=nsaenzju@redhat.com \
--cc=Anup.Patel@wdc.com \
--cc=frederic@kernel.org \
--cc=kvmarm@lists.cs.columbia.edu \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=maz@kernel.org \
--cc=paulmck@kernel.org \
--cc=pbonzini@redhat.com \
--cc=rcu@vger.kernel.org \
--cc=tglx@linutronix.de \
--cc=will@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 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.