Re: [PATCH] IPI performance benchmark

[Yury Norov]

On Mon, Dec 11, 2017 at 03:35:02PM +0100, Christian Borntraeger wrote:
> 
> 
> On 12/11/2017 03:16 PM, Yury Norov wrote:
> > This benchmark sends many IPIs in different modes and measures
> > time for IPI delivery (first column), and total time, ie including
> > time to acknowledge the receive by sender (second column).
> > 
> > The scenarios are:
> > Dry-run:	do everything except actually sending IPI. Useful
> > 		to estimate system overhead.
> > Self-IPI:	Send IPI to self CPU.
> > Normal IPI:	Send IPI to some other CPU.
> > Broadcast IPI:	Send broadcast IPI to all online CPUs.
> > 
> > For virtualized guests, sending and reveiving IPIs causes guest exit.
> > I used this test to measure performance impact on KVM subsystem of
> > Christoffer Dall's series "Optimize KVM/ARM for VHE systems".
> > 
> > https://www.spinics.net/lists/kvm/msg156755.html
> > 
> > Test machine is ThunderX2, 112 online CPUs. Below the results normalized
> > to host dry-run time. Smaller - better.
> > 
> > Host, v4.14:
> > Dry-run:	  0	    1
> > Self-IPI:         9	   18
> > Normal IPI:      81	  110
> > Broadcast IPI:    0	 2106
> > 
> > Guest, v4.14:
> > Dry-run:          0	    1
> > Self-IPI:        10	   18
> > Normal IPI:     305	  525
> > Broadcast IPI:    0    	 9729
> > 
> > Guest, v4.14 + VHE:
> > Dry-run:          0	    1
> > Self-IPI:         9	   18
> > Normal IPI:     176	  343
> > Broadcast IPI:    0	 9885
> > 
> > CC: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> > CC: Ashish Kalra <Ashish.Kalra@xxxxxxxxxx>
> > CC: Christoffer Dall <christoffer.dall@xxxxxxxxxx>
> > CC: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
> > CC: Linu Cherian <Linu.Cherian@xxxxxxxxxx>
> > CC: Sunil Goutham <Sunil.Goutham@xxxxxxxxxx>
> > Signed-off-by: Yury Norov <ynorov@xxxxxxxxxxxxxxxxxx>
> > ---
> >  arch/Kconfig           |  10 ++++
> >  kernel/Makefile        |   1 +
> >  kernel/ipi_benchmark.c | 134 +++++++++++++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 145 insertions(+)
> >  create mode 100644 kernel/ipi_benchmark.c
> > diff --git a/arch/Kconfig b/arch/Kconfig
> > index 057370a0ac4e..80d6ef439199 100644
> > --- a/arch/Kconfig
> > +++ b/arch/Kconfig
> > @@ -82,6 +82,16 @@ config JUMP_LABEL
> >  	 ( On 32-bit x86, the necessary options added to the compiler
> >  	   flags may increase the size of the kernel slightly. )
> > 
> > +config IPI_BENCHMARK
> > +	tristate "Test IPI performance on SMP systems"
> > +	depends on SMP
> > +	help
> > +	  Test IPI performance on SMP systems. If system has only one online
> > +	  CPU, sending IPI to other CPU is obviously not possible, and ENOENT
> > +	  is returned for corresponding test.
> > +
> > +	  If unsure, say N.
> > +
> >  config STATIC_KEYS_SELFTEST
> >  	bool "Static key selftest"
> >  	depends on JUMP_LABEL
> > diff --git a/kernel/Makefile b/kernel/Makefile
> > index 172d151d429c..04e550e1990c 100644
> > --- a/kernel/Makefile
> > +++ b/kernel/Makefile
> > @@ -101,6 +101,7 @@ obj-$(CONFIG_TRACEPOINTS) += trace/
> >  obj-$(CONFIG_IRQ_WORK) += irq_work.o
> >  obj-$(CONFIG_CPU_PM) += cpu_pm.o
> >  obj-$(CONFIG_BPF) += bpf/
> > +obj-$(CONFIG_IPI_BENCHMARK) += ipi_benchmark.o
> > 
> >  obj-$(CONFIG_PERF_EVENTS) += events/
> > 
> > diff --git a/kernel/ipi_benchmark.c b/kernel/ipi_benchmark.c
> > new file mode 100644
> > index 000000000000..35f1f7598c36
> > --- /dev/null
> > +++ b/kernel/ipi_benchmark.c
> > @@ -0,0 +1,134 @@
> > +/*
> > + * Performance test for IPI on SMP machines.
> > + *
> > + * Copyright (c) 2017 Cavium Networks.
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of version 2 of the GNU General Public
> > + * License as published by the Free Software Foundation.
> > + *
> > + * This program is distributed in the hope that it will be useful, but
> > + * WITHOUT ANY WARRANTY; without even the implied warranty of
> > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> > + * General Public License for more details.
> > + */
> > +
> > +#include <linux/module.h>
> > +#include <linux/kernel.h>
> > +#include <linux/init.h>
> > +#include <linux/ktime.h>
> > +
> > +#define NTIMES 100000
> > +
> > +#define POKE_ANY	0
> > +#define DRY_RUN		1
> > +#define POKE_SELF	2
> > +#define POKE_ALL	3
> > +
> > +static void __init handle_ipi(void *t)
> > +{
> > +	ktime_t *time = (ktime_t *) t;
> > +
> > +	if (time)
> > +		*time = ktime_get() - *time;
> > +}
> > +
> > +static ktime_t __init send_ipi(int flags)
> > +{
> > +	ktime_t time;
> > +	unsigned int cpu = get_cpu();
> > +
> > +	switch (flags) {
> > +	case POKE_ALL:
> > +		/* If broadcasting, don't force all CPUs to update time. */
> > +		smp_call_function_many(cpu_online_mask, handle_ipi, NULL, 1);
> > +		/* Fall thru */
> > +	case DRY_RUN:
> > +		/* Do everything except actually sending IPI. */
> > +		time = 0;
> > +		break;
> > +	case POKE_ANY:
> > +		cpu = cpumask_any_but(cpu_online_mask, cpu);
> > +		if (cpu >= nr_cpu_ids) {
> > +			time = -ENOENT;
> > +			break;
> > +		}
> > +		/* Fall thru */
> > +	case POKE_SELF:
> > +		time = ktime_get();
> > +		smp_call_function_single(cpu, handle_ipi, &time, 1);
> > +		break;
> > +	default:
> > +		time = -EINVAL;
> > +	}
> > +
> > +	put_cpu();
> > +	return time;
> > +}
> > +
> > +static int __init __bench_ipi(unsigned long i, ktime_t *time, int flags)
> > +{
> > +	ktime_t t;
> > +
> > +	*time = 0;
> > +	while (i--) {
> > +		t = send_ipi(flags);
> > +		if ((int) t < 0)
> > +			return (int) t;
> > +
> > +		*time += t;
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> > +static int __init bench_ipi(unsigned long times, int flags,
> > +				ktime_t *ipi, ktime_t *total)
> > +{
> > +	int ret;
> > +
> > +	*total = ktime_get();
> > +	ret = __bench_ipi(times, ipi, flags);
> > +	if (unlikely(ret))
> > +		return ret;
> > +
> > +	*total = ktime_get() - *total;
> > +
> > +	return 0;
> > +}
> > +
> > +static int __init init_bench_ipi(void)
> > +{
> > +	ktime_t ipi, total;
> > +	int ret;
> > +
> > +	ret = bench_ipi(NTIMES, DRY_RUN, &ipi, &total);
> > +	if (ret)
> > +		pr_err("Dry-run FAILED: %d\n", ret);
> > +	else
> > +		pr_err("Dry-run:       %18llu, %18llu ns\n", ipi, total);
> 
> you do not use NTIMES here to calculate the average value. Is that intended?

I think, it's more visually to represent all results in number of dry-run
times, like I did in patch description. So on kernel side I expose raw data
and calculate final values after finishing tests.

If you think that average values are preferable, I can do that in v2.

Yury