From: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
To: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: David Woodhouse <dwmw2@infradead.org>,
peterz@infradead.org, mhillenb@amazon.de,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Subject: Re: [PATCH v2] kvm/x86: Inform RCU of quiescent state when entering guest mode
Date: Wed, 11 Jul 2018 16:47:20 -0700 [thread overview]
Message-ID: <20180711234720.GJ3593@linux.vnet.ibm.com> (raw)
In-Reply-To: <18288874-e778-4f7f-2fd5-f03624717e9c@de.ibm.com>
On Wed, Jul 11, 2018 at 11:39:10PM +0200, Christian Borntraeger wrote:
>
>
> On 07/11/2018 11:32 PM, Paul E. McKenney wrote:
> > On Wed, Jul 11, 2018 at 11:11:19PM +0200, Christian Borntraeger wrote:
> >>
> >>
> >> On 07/11/2018 10:27 PM, Paul E. McKenney wrote:
> >>> On Wed, Jul 11, 2018 at 08:39:36PM +0200, Christian Borntraeger wrote:
> >>>>
> >>>>
> >>>> On 07/11/2018 08:36 PM, Paul E. McKenney wrote:
> >>>>> On Wed, Jul 11, 2018 at 11:20:53AM -0700, Paul E. McKenney wrote:
> >>>>>> On Wed, Jul 11, 2018 at 07:01:01PM +0100, David Woodhouse wrote:
> >>>>>>> From: David Woodhouse <dwmw@amazon.co.uk>
> >>>>>>>
> >>>>>>> RCU can spend long periods of time waiting for a CPU which is actually in
> >>>>>>> KVM guest mode, entirely pointlessly. Treat it like the idle and userspace
> >>>>>>> modes, and don't wait for it.
> >>>>>>>
> >>>>>>> Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
> >>>>>>
> >>>>>> And idiot here forgot about some of the debugging code in RCU's dyntick-idle
> >>>>>> code. I will reply with a fixed patch.
> >>>>>>
> >>>>>> The code below works just fine as long as you don't enable CONFIG_RCU_EQS_DEBUG,
> >>>>>> so should be OK for testing, just not for mainline.
> >>>>>
> >>>>> And here is the updated code that allegedly avoids splatting when run with
> >>>>> CONFIG_RCU_EQS_DEBUG.
> >>>>>
> >>>>> Thoughts?
> >>>>>
> >>>>> Thanx, Paul
> >>>>>
> >>>>> ------------------------------------------------------------------------
> >>>>>
> >>>>> commit 12cd59e49cf734f907f44b696e2c6e4b46a291c3
> >>>>> Author: David Woodhouse <dwmw@amazon.co.uk>
> >>>>> Date: Wed Jul 11 19:01:01 2018 +0100
> >>>>>
> >>>>> kvm/x86: Inform RCU of quiescent state when entering guest mode
> >>>>>
> >>>>> RCU can spend long periods of time waiting for a CPU which is actually in
> >>>>> KVM guest mode, entirely pointlessly. Treat it like the idle and userspace
> >>>>> modes, and don't wait for it.
> >>>>>
> >>>>> Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
> >>>>> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> >>>>> [ paulmck: Adjust to avoid bad advice I gave to dwmw, avoid WARN_ON()s. ]
> >>>>>
> >>>>> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> >>>>> index 0046aa70205a..b0c82f70afa7 100644
> >>>>> --- a/arch/x86/kvm/x86.c
> >>>>> +++ b/arch/x86/kvm/x86.c
> >>>>> @@ -7458,7 +7458,9 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
> >>>>> vcpu->arch.switch_db_regs &= ~KVM_DEBUGREG_RELOAD;
> >>>>> }
> >>>>>
> >>>>> + rcu_kvm_enter();
> >>>>> kvm_x86_ops->run(vcpu);
> >>>>> + rcu_kvm_exit();
> >>>>
> >>>> As indicated in my other mail. This is supposed to be handled in the guest_enter|exit_ calls around
> >>>> the run function. This would also handle other architectures. So if the guest_enter_irqoff code is
> >>>> not good enough, we should rather fix that instead of adding another rcu hint.
> >>>
> >>> Something like this, on top of the earlier patch? I am not at all
> >>> confident of this patch because there might be other entry/exit
> >>> paths I am missing. Plus there might be RCU uses on the arch-specific
> >>> patch to and from the guest OS.
> >>>
> >>> Thoughts?
> >>>
> >>
> >> If you instrment guest_enter/exit, you should cover all cases and all architectures as far
> >> as I can tell. FWIW, we did this rcu_note thing back then actually handling this particular
> >> case of long running guests blocking rcu for many seconds. And I am pretty sure that
> >> this did help back then.
> >
> > And my second patch on the email you replied to replaced the only call
> > to rcu_virt_note_context_switch(). So maybe it covers what it needs to,
> > but yes, there might well be things I missed. Let's see what David
> > comes up with.
> >
> > What changed was RCU's reactions to longish grace periods. It used to
> > be very aggressive about forcing the scheduler to do otherwise-unneeded
> > context switches, which became a problem somewhere between v4.9 and v4.15.
> > I therefore reduced the number of such context switches, which in turn
> > caused KVM to tell RCU about quiescent states way too infrequently.
>
> You talk about
> commit bcbfdd01dce5556a952fae84ef16fd0f12525e7b
> rcu: Make non-preemptive schedule be Tasks RCU quiescent state
>
> correct? In fact, then whatever (properly sent) patch comes up should contain
> a fixes tag.
Not that one, but this one is at least part of the "team":
28053bc72c0e5 ("rcu: Add long-term CPU kicking"). I might need to use
"git bisect" to find the most relevant commit... :-/
> > The advantage of the rcu_kvm_enter()/rcu_kvm_exit() approach is that
> > it tells RCU of an extended duration in the guest, which means that
> > RCU can ignore the corresponding CPU, which in turn allows the guest
> > to proceed without any RCU-induced interruptions.
> >
> > Does that make sense, or am I missing something? I freely admit to
> > much ignorance of both kvm and s390! ;-)
>
> WIth that explanation it makes perfect sense to replace
> rcu_virt_note_context_switch with rcu_kvm_enter/exit from an rcu performance
> perspective. I assume that rcu_kvm_enter is not much slower than
> rcu_virt_note_context_switch? Because we do call it on every guest entry/exit
> which we might have plenty for ping pong I/O workload.
But is there any way for a guest OS to sneak back out to the hypervisor
without executing one of the rcu_kvm_exit() calls? If there is, RCU
is broken.
Thanx, Paul
next prev parent reply other threads:[~2018-07-11 23:45 UTC|newest]
Thread overview: 93+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-07-06 14:53 [RFC] Make need_resched() return true when rcu_urgent_qs requested David Woodhouse
2018-07-06 16:29 ` Peter Zijlstra
2018-07-06 17:11 ` Paul E. McKenney
2018-07-06 17:14 ` David Woodhouse
2018-07-06 21:12 ` Paul E. McKenney
2018-07-09 8:58 ` Peter Zijlstra
2018-07-09 8:53 ` Peter Zijlstra
2018-07-09 9:18 ` David Woodhouse
2018-07-09 10:44 ` Peter Zijlstra
2018-07-09 10:56 ` David Woodhouse
2018-07-09 11:06 ` Peter Zijlstra
2018-07-09 11:12 ` David Woodhouse
2018-07-09 11:31 ` Peter Zijlstra
2018-07-09 12:34 ` Paul E. McKenney
2018-07-09 12:47 ` David Woodhouse
2018-07-09 14:30 ` Paul E. McKenney
2018-07-09 12:55 ` Peter Zijlstra
2018-07-09 12:57 ` David Woodhouse
2018-07-09 13:02 ` Peter Zijlstra
2018-07-09 14:29 ` Paul E. McKenney
2018-07-09 14:43 ` Peter Zijlstra
2018-07-09 14:54 ` Paul E. McKenney
2018-07-09 15:26 ` Peter Zijlstra
2018-07-09 16:34 ` Paul E. McKenney
2018-07-09 16:44 ` Paul E. McKenney
2018-07-09 18:50 ` David Woodhouse
2018-07-09 20:34 ` Paul E. McKenney
2018-07-09 20:35 ` David Woodhouse
2018-07-09 20:42 ` Paul E. McKenney
2018-07-09 20:45 ` David Woodhouse
2018-07-09 21:05 ` Paul E. McKenney
2018-07-09 22:08 ` Paul E. McKenney
2018-07-11 10:57 ` David Woodhouse
2018-07-11 12:51 ` Paul E. McKenney
2018-07-11 12:58 ` David Woodhouse
2018-07-11 14:25 ` Paul E. McKenney
2018-07-11 14:23 ` David Woodhouse
2018-07-11 14:43 ` Paul E. McKenney
2018-07-11 16:49 ` Paul E. McKenney
2018-07-11 17:03 ` David Woodhouse
2018-07-11 17:48 ` Paul E. McKenney
2018-07-11 18:01 ` [PATCH v2] kvm/x86: Inform RCU of quiescent state when entering guest mode David Woodhouse
2018-07-11 18:20 ` Paul E. McKenney
2018-07-11 18:36 ` Paul E. McKenney
2018-07-11 18:39 ` Christian Borntraeger
2018-07-11 20:27 ` Paul E. McKenney
2018-07-11 20:54 ` David Woodhouse
2018-07-11 21:09 ` Paul E. McKenney
2018-07-11 21:11 ` Christian Borntraeger
2018-07-11 21:32 ` Paul E. McKenney
2018-07-11 21:39 ` Christian Borntraeger
2018-07-11 23:47 ` Paul E. McKenney [this message]
2018-07-12 8:31 ` David Woodhouse
2018-07-12 11:00 ` Christian Borntraeger
2018-07-12 11:10 ` David Woodhouse
2018-07-12 11:58 ` Christian Borntraeger
2018-07-12 12:04 ` Christian Borntraeger
2018-07-11 23:37 ` Paul E. McKenney
2018-07-12 2:15 ` Paul E. McKenney
2018-07-12 6:21 ` Christian Borntraeger
2018-07-12 9:52 ` David Woodhouse
2018-07-11 18:31 ` [RFC] Make need_resched() return true when rcu_urgent_qs requested Christian Borntraeger
2018-07-11 20:17 ` Paul E. McKenney
2018-07-11 20:19 ` David Woodhouse
2018-07-11 21:08 ` Paul E. McKenney
2018-07-12 12:00 ` David Woodhouse
2018-07-12 12:53 ` Paul E. McKenney
2018-07-12 16:17 ` Paul E. McKenney
2018-07-16 15:40 ` Paul E. McKenney
2018-07-17 8:19 ` David Woodhouse
2018-07-17 12:56 ` Paul E. McKenney
2018-07-18 15:36 ` Paul E. McKenney
2018-07-18 16:01 ` David Woodhouse
2018-07-18 16:37 ` Paul E. McKenney
2018-07-18 19:41 ` David Woodhouse
2018-07-18 20:17 ` Paul E. McKenney
2018-07-19 0:26 ` Frederic Weisbecker
2018-07-19 6:45 ` Christian Borntraeger
2018-07-19 7:20 ` David Woodhouse
2018-07-19 10:23 ` Christian Borntraeger
2018-07-19 12:55 ` Paul E. McKenney
2018-07-19 13:14 ` Frederic Weisbecker
2018-07-19 13:36 ` David Woodhouse
2018-07-19 17:09 ` Paul E. McKenney
2018-07-23 8:08 ` David Woodhouse
2018-07-23 12:22 ` Paul E. McKenney
2018-07-19 0:32 ` Frederic Weisbecker
2018-07-19 3:11 ` Paul E. McKenney
2018-07-19 6:16 ` David Woodhouse
2018-07-19 13:17 ` Frederic Weisbecker
2018-07-19 13:15 ` Frederic Weisbecker
2018-07-10 9:24 ` Peter Zijlstra
2018-07-10 16:26 ` Paul E. McKenney
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=20180711234720.GJ3593@linux.vnet.ibm.com \
--to=paulmck@linux.vnet.ibm.com \
--cc=borntraeger@de.ibm.com \
--cc=dwmw2@infradead.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mhillenb@amazon.de \
--cc=peterz@infradead.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).