public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: "André Przywara" <andre.przywara@arm.com>
To: Alexandru Elisei <alexandru.elisei@arm.com>,
	drjones@redhat.com, kvm@vger.kernel.org,
	kvmarm@lists.cs.columbia.edu
Cc: eric.auger@redhat.com, yuzenghui@huawei.com
Subject: Re: [kvm-unit-tests PATCH v2 08/12] arm/arm64: gic: Split check_acked() into two functions
Date: Fri, 18 Dec 2020 15:52:36 +0000	[thread overview]
Message-ID: <3539c229-fd05-2e1c-2159-995e51e2dcc4@arm.com> (raw)
In-Reply-To: <20201217141400.106137-9-alexandru.elisei@arm.com>

On 17/12/2020 14:13, Alexandru Elisei wrote:
> check_acked() has several peculiarities: is the only function among the
> check_* functions which calls report() directly, it does two things
> (waits for interrupts and checks for misfired interrupts) and it also
> mixes printf, report_info and report calls.
> 
> check_acked() also reports a pass and returns as soon all the target CPUs
> have received interrupts, However, a CPU not having received an interrupt
> *now* does not guarantee not receiving an erroneous interrupt if we wait
> long enough.
> 
> Rework the function by splitting it into two separate functions, each with
> a single responsibility: wait_for_interrupts(), which waits for the
> expected interrupts to fire, and check_acked() which checks that interrupts
> have been received as expected.
> 
> wait_for_interrupts() also waits an extra 100 milliseconds after the
> expected interrupts have been received in an effort to make sure we don't
> miss misfiring interrupts.
> 
> Splitting check_acked() into two functions will also allow us to
> customize the behavior of each function in the future more easily
> without using an unnecessarily long list of arguments for check_acked().

Yes, splitting this up looks much better, in general this is a nice
cleanup, thank you!

Some comments below:

> 
> CC: Andre Przywara <andre.przywara@arm.com>
> Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
> ---
>  arm/gic.c | 73 +++++++++++++++++++++++++++++++++++--------------------
>  1 file changed, 47 insertions(+), 26 deletions(-)
> 
> diff --git a/arm/gic.c b/arm/gic.c
> index ec733719c776..a9ef1a5def56 100644
> --- a/arm/gic.c
> +++ b/arm/gic.c
> @@ -62,41 +62,42 @@ static void stats_reset(void)
>  	}
>  }
>  
> -static void check_acked(const char *testname, cpumask_t *mask)
> +static void wait_for_interrupts(cpumask_t *mask)
>  {
> -	int missing = 0, extra = 0, unexpected = 0;
>  	int nr_pass, cpu, i;
> -	bool bad = false;
>  
>  	/* Wait up to 5s for all interrupts to be delivered */
> -	for (i = 0; i < 50; ++i) {
> +	for (i = 0; i < 50; i++) {
>  		mdelay(100);
>  		nr_pass = 0;
>  		for_each_present_cpu(cpu) {
> +			/*
> +			 * A CPU having received more than one interrupts will
> +			 * show up in check_acked(), and no matter how long we
> +			 * wait it cannot un-receive it. Consider at least one
> +			 * interrupt as a pass.
> +			 */
>  			nr_pass += cpumask_test_cpu(cpu, mask) ?
> -				acked[cpu] == 1 : acked[cpu] == 0;
> -			smp_rmb(); /* pairs with smp_wmb in ipi_handler */
> -
> -			if (bad_sender[cpu] != -1) {
> -				printf("cpu%d received IPI from wrong sender %d\n",
> -					cpu, bad_sender[cpu]);
> -				bad = true;
> -			}
> -
> -			if (bad_irq[cpu] != -1) {
> -				printf("cpu%d received wrong irq %d\n",
> -					cpu, bad_irq[cpu]);
> -				bad = true;
> -			}
> +				acked[cpu] >= 1 : acked[cpu] == 0;


I wonder if this logic was already flawed to begin with: For interrupts
we expect to fire, we wait for up to 5 seconds (really that long?), but
for interrupts we expect *not* to fire we are OK if they don't show up
in the first 100 ms. That does not sound consistent.

I am wondering if we should *not* have the initial 100ms wait at all,
since most interrupts will fire immediately (especially in KVM). And
then have *one* extra wait for, say 100ms, to cover latecomers and
spurious interrupts.

But this might be a topic for some extra work/patch?

>  		}
> +
>  		if (nr_pass == nr_cpus) {
> -			report(!bad, "%s", testname);
>  			if (i)
> -				report_info("took more than %d ms", i * 100);
> +				report_info("interrupts took more than %d ms", i * 100);
> +			mdelay(100);

So this is the extra 100ms you mention in the commit message? I am not
convinced this is the right way (see above) or even the right place
(rather at the call site?) to wait. But at least it deserves a comment,
I believe.

>  			return;
>  		}
>  	}
>  
> +	report_info("interrupts timed-out (5s)");
> +}
> +
> +static bool check_acked(cpumask_t *mask)
> +{
> +	int missing = 0, extra = 0, unexpected = 0;
> +	bool pass = true;
> +	int cpu;
> +
>  	for_each_present_cpu(cpu) {
>  		if (cpumask_test_cpu(cpu, mask)) {
>  			if (!acked[cpu])
> @@ -107,11 +108,28 @@ static void check_acked(const char *testname, cpumask_t *mask)
>  			if (acked[cpu])
>  				++unexpected;
>  		}
> +		smp_rmb(); /* pairs with smp_wmb in ipi_handler */
> +
> +		if (bad_sender[cpu] != -1) {
> +			report_info("cpu%d received IPI from wrong sender %d",
> +					cpu, bad_sender[cpu]);
> +			pass = false;
> +		}
> +
> +		if (bad_irq[cpu] != -1) {
> +			report_info("cpu%d received wrong irq %d",
> +					cpu, bad_irq[cpu]);
> +			pass = false;
> +		}
> +	}
> +
> +	if (missing || extra || unexpected) {
> +		report_info("ACKS: missing=%d extra=%d unexpected=%d",
> +				missing, extra, unexpected);
> +		pass = false;

Thanks, that so much easier to read now.

Cheers,
Andre

>  	}
>  
> -	report(false, "%s", testname);
> -	report_info("Timed-out (5s). ACKS: missing=%d extra=%d unexpected=%d",
> -		    missing, extra, unexpected);
> +	return pass;
>  }
>  
>  static void check_spurious(void)
> @@ -303,7 +321,8 @@ static void ipi_test_self(void)
>  	cpumask_clear(&mask);
>  	cpumask_set_cpu(smp_processor_id(), &mask);
>  	gic->ipi.send_self();
> -	check_acked("IPI: self", &mask);
> +	wait_for_interrupts(&mask);
> +	report(check_acked(&mask), "Interrupts received");
>  	report_prefix_pop();
>  }
>  
> @@ -318,7 +337,8 @@ static void ipi_test_smp(void)
>  	for (i = smp_processor_id() & 1; i < nr_cpus; i += 2)
>  		cpumask_clear_cpu(i, &mask);
>  	gic_ipi_send_mask(IPI_IRQ, &mask);
> -	check_acked("IPI: directed", &mask);
> +	wait_for_interrupts(&mask);
> +	report(check_acked(&mask), "Interrupts received");
>  	report_prefix_pop();
>  
>  	report_prefix_push("broadcast");
> @@ -326,7 +346,8 @@ static void ipi_test_smp(void)
>  	cpumask_copy(&mask, &cpu_present_mask);
>  	cpumask_clear_cpu(smp_processor_id(), &mask);
>  	gic->ipi.send_broadcast();
> -	check_acked("IPI: broadcast", &mask);
> +	wait_for_interrupts(&mask);
> +	report(check_acked(&mask), "Interrupts received");
>  	report_prefix_pop();
>  }
>  
> 


  reply	other threads:[~2020-12-18 15:53 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-17 14:13 [kvm-unit-tests PATCH v2 00/12] GIC fixes and improvements Alexandru Elisei
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 01/12] lib: arm/arm64: gicv3: Add missing barrier when sending IPIs Alexandru Elisei
2020-12-18 12:03   ` André Przywara
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 02/12] lib: arm/arm64: gicv2: " Alexandru Elisei
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 03/12] arm/arm64: gic: Remove SMP synchronization from ipi_clear_active_handler() Alexandru Elisei
2020-12-18 12:04   ` André Przywara
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 04/12] arm/arm64: gic: Remove unnecessary synchronization with stats_reset() Alexandru Elisei
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 05/12] arm/arm64: gic: Use correct memory ordering for the IPI test Alexandru Elisei
2020-12-18 12:04   ` André Przywara
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 06/12] arm/arm64: gic: Check spurious and bad_sender in the active test Alexandru Elisei
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 07/12] arm/arm64: gic: Wait for writes to acked or spurious to complete Alexandru Elisei
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 08/12] arm/arm64: gic: Split check_acked() into two functions Alexandru Elisei
2020-12-18 15:52   ` André Przywara [this message]
2021-01-25 17:27     ` Alexandru Elisei
2021-01-27 15:10       ` Andre Przywara
2021-01-27 16:00         ` Alexandru Elisei
2021-02-16 18:04           ` Andre Przywara
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 09/12] arm/arm64: gic: Make check_acked() more generic Alexandru Elisei
2020-12-18 15:52   ` André Przywara
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 10/12] arm64: gic: its-trigger: Don't trigger the LPI while it is pending Alexandru Elisei
2020-12-18 18:15   ` André Przywara
2021-01-25 16:57     ` Alexandru Elisei
2020-12-17 14:13 ` [kvm-unit-tests PATCH v2 11/12] lib: arm64: gic-v3-its: Add wmb() barrier before INT command Alexandru Elisei
2020-12-18 18:36   ` André Przywara
2021-01-25 15:16     ` Alexandru Elisei
2020-12-17 14:14 ` [kvm-unit-tests PATCH v2 12/12] arm64: gic: Use IPI test checking for the LPI tests Alexandru Elisei

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=3539c229-fd05-2e1c-2159-995e51e2dcc4@arm.com \
    --to=andre.przywara@arm.com \
    --cc=alexandru.elisei@arm.com \
    --cc=drjones@redhat.com \
    --cc=eric.auger@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=kvmarm@lists.cs.columbia.edu \
    --cc=yuzenghui@huawei.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox