Re: [PATCH 3/5] kernel/watchdog: adapt the watchdog_hld interface for async model

From: Pingfan Liu <piliu@redhat.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Pingfan Liu <kernelfans@gmail.com>,
	linux-kernel@vger.kernel.org, Sumit Garg <sumit.garg@linaro.org>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>, Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	Mark Rutland <mark.rutland@arm.com>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Jiri Olsa <jolsa@redhat.com>, Namhyung Kim <namhyung@kernel.org>,
	Marc Zyngier <maz@kernel.org>,
	Julien Thierry <jthierry@redhat.com>,
	Kees Cook <keescook@chromium.org>,
	Masahiro Yamada <masahiroy@kernel.org>,
	Sami Tolvanen <samitolvanen@google.com>,
	Petr Mladek <pmladek@suse.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Wang Qing <wangqing@vivo.com>,
	Santosh Sivaraj <santosh@fossix.org>
Subject: Re: [PATCH 3/5] kernel/watchdog: adapt the watchdog_hld interface for async model
Date: Thu, 16 Sep 2021 11:07:22 +0800	[thread overview]
Message-ID: <YUK06iqiCwiIvpRM@piliu.users.ipa.redhat.com> (raw)
In-Reply-To: <YUH89GX1RB8fdcvh@hirez.programming.kicks-ass.net>

On Wed, Sep 15, 2021 at 04:02:28PM +0200, Peter Zijlstra wrote:
> On Wed, Sep 15, 2021 at 11:51:01AM +0800, Pingfan Liu wrote:
> > When lockup_detector_init()->watchdog_nmi_probe(), PMU may be not ready
> > yet. E.g. on arm64, PMU is not ready until
> > device_initcall(armv8_pmu_driver_init).  And it is deeply integrated
> > with the driver model and cpuhp. Hence it is hard to push this
> > initialization before smp_init().
> > 
> > But it is easy to take an opposite approach by enabling watchdog_hld to
> > get the capability of PMU async.
> > 
> > The async model is achieved by introducing an extra parameter notifier
> > of watchdog_nmi_probe().
> > 
> > Note after this patch, the async model, which is utilized by the next
> > patch, does not take effect yet.
> 
> I can't make any sense of what you're trying to do..
> 
Sorry for a bad expression. what I mean is: this patch [3/5] provides an
framework for async model. But since watchdog_nmi_probe() still return 0 or
-ENODEV after this patch, the code's behavior is the same as original.

Does it make sense to you?
> > +static void watchdog_nmi_report_capability(struct watchdog_nmi_status *data)
> > +{
> > +	/* Set status to 1 temporary to block any further access */
> > +	if (atomic_cmpxchg((atomic_t *)&nmi_watchdog_status, -EBUSY, 1)
> > +			== -EBUSY) {
> 
> But this..
> 
Oh, check other codes, for a wrapped condition, blanks should be better choice.
> > +		if (!data->status) {
> > +			nmi_watchdog_status = 0;
> > +			lockup_detector_update_enable();
> > +		} else {
> > +			nmi_watchdog_status = -ENODEV;
> > +			/* turn offf watchdog_enabled forever */
> > +			lockup_detector_update_enable();
> > +			pr_info("Perf NMI watchdog permanently disabled\n");
> > +		}
> > +	}
> > +}
> 
> > @@ -467,7 +494,8 @@ static void watchdog_enable(unsigned int cpu)
> >  	/* Initialize timestamp */
> >  	update_touch_ts();
> >  	/* Enable the perf event */
> > -	if (watchdog_enabled & NMI_WATCHDOG_ENABLED)
> > +	if (watchdog_enabled &
> > +			(NMI_WATCHDOG_ENABLED | NMI_WATCHDOG_UNDETERMINED))
> 
> and this, are horrible indenting.

Ditto.

Thanks for your comment and review.

Regards,

	Pingfan