Re: [PATCH 3/3] context_tracking,x86: remove extraneous irq disable & enable from context tracking on syscall entry

From: Mike Galbraith <umgwanakikbuti@gmail.com>
To: Rik van Riel <riel@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>,
	Andy Lutomirski <luto@amacapital.net>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	X86 ML <x86@kernel.org>,
	williams@redhat.com, Andrew Lutomirski <luto@kernel.org>,
	fweisbec@redhat.com, Peter Zijlstra <peterz@infradead.org>,
	Heiko Carstens <heiko.carstens@de.ibm.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>,
	Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH 3/3] context_tracking,x86: remove extraneous irq disable & enable from context tracking on syscall entry
Date: Sat, 02 May 2015 06:06:21 +0200	[thread overview]
Message-ID: <1430539581.3137.61.camel@gmail.com> (raw)
In-Reply-To: <5543C05E.9040209@redhat.com>

On Fri, 2015-05-01 at 14:05 -0400, Rik van Riel wrote:
> On 05/01/2015 12:34 PM, Ingo Molnar wrote:
> > 
> > * Rik van Riel <riel@redhat.com> wrote:
> > 
> >>> I can understand people running hard-RT workloads not wanting to 
> >>> see the overhead of a timer tick or a scheduler tick with variable 
> >>> (and occasionally heavy) work done in IRQ context, but the jitter 
> >>> caused by a single trivial IPI with constant work should be very, 
> >>> very low and constant.
> >>
> >> Not if the realtime workload is running inside a KVM guest.
> > 
> > I don't buy this:
> > 
> >> At that point an IPI, either on the host or in the guest, involves a 
> >> full VMEXIT & VMENTER cycle.
> > 
> > So a full VMEXIT/VMENTER costs how much, 2000 cycles? That's around 1 
> > usec on recent hardware, and I bet it will get better with time.
> > 
> > I'm not aware of any hard-RT workload that cannot take 1 usec 
> > latencies.
> 
> Now think about doing this kind of IPI from inside a guest,
> to another VCPU on the same guest.
> 
> Now you are looking at VMEXIT/VMENTER on the first VCPU,
> plus the cost of the IPI on the host, plus the cost of
> the emulation layer, plus VMEXIT/VMENTER on the second
> VCPU to trigger the IPI work, and possibly a second
> VMEXIT/VMENTER for IPI completion.
> 
> I suspect it would be better to do RCU callback offload
> in some other way.

I don't get it.  How the heck do people manage to talk about realtime in
virtual boxen, and not at least crack a smile.  Real, virtual, real,
virtual... what's wrong with this picture?

Why is virtual realtime not an oxymoron?

I did that for grins once, and it was either really funny, or really
sad, not sure which... but it did not look really really useful.

	-Mike