Re: [tip: timers/core] hrtimer: Consolidate reprogramming code

From: Thomas Gleixner <tglx@linutronix.de>
To: Mike Galbraith <efault@gmx.de>,
	linux-kernel@vger.kernel.org, linux-tip-commits@vger.kernel.org
Cc: Peter Zijlstra <peterz@infradead.org>, x86@kernel.org
Subject: Re: [tip: timers/core] hrtimer: Consolidate reprogramming code
Date: Thu, 12 Aug 2021 16:32:41 +0200	[thread overview]
Message-ID: <877dgqivhy.ffs@tglx> (raw)
In-Reply-To: <87a6lmiwi0.ffs@tglx>

On Thu, Aug 12 2021 at 16:11, Thomas Gleixner wrote:

> On Thu, Aug 12 2021 at 09:19, Mike Galbraith wrote:
>> Greetings Peter, may your day get off to a better start than my box's
>> did :)
>>
>> On Tue, 2021-08-10 at 16:02 +0000, tip-bot2 for Peter Zijlstra wrote:
>>> The following commit has been merged into the timers/core branch of tip:
>>>
>>> Commit-ID:     b14bca97c9f5c3e3f133445b01c723e95490d843
>>> Gitweb:        https://git.kernel.org/tip/b14bca97c9f5c3e3f133445b01c723e95490d843
>>> Author:        Peter Zijlstra <peterz@infradead.org>
>>> AuthorDate:    Tue, 13 Jul 2021 15:39:47 +02:00
>>> Committer:     Thomas Gleixner <tglx@linutronix.de>
>>> CommitterDate: Tue, 10 Aug 2021 17:57:22 +02:00
>>>
>>> hrtimer: Consolidate reprogramming code
>>
>> Per git-bisect, this is the tip.today commit that's bricking my box
>> early in boot.
>
> Let me stare at that.

Can you please test whether the below fixes it for you?

I have yet to find a machine which reproduces it as I really want to
understand which particular part is causing this.

Thanks,

        tglx
---

--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -652,24 +652,10 @@ static inline int hrtimer_hres_active(vo
 	return __hrtimer_hres_active(this_cpu_ptr(&hrtimer_bases));
 }
 
-static void
-__hrtimer_reprogram(struct hrtimer_cpu_base *cpu_base, int skip_equal,
-		    struct hrtimer *next_timer, ktime_t expires_next)
+static void __hrtimer_reprogram(struct hrtimer_cpu_base *cpu_base,
+				struct hrtimer *next_timer,
+				ktime_t expires_next)
 {
-	/*
-	 * If the hrtimer interrupt is running, then it will reevaluate the
-	 * clock bases and reprogram the clock event device.
-	 */
-	if (cpu_base->in_hrtirq)
-		return;
-
-	if (expires_next > cpu_base->expires_next)
-		return;
-
-	if (skip_equal && expires_next == cpu_base->expires_next)
-		return;
-
-	cpu_base->next_timer = next_timer;
 	cpu_base->expires_next = expires_next;
 
 	/*
@@ -707,8 +693,10 @@ hrtimer_force_reprogram(struct hrtimer_c
 
 	expires_next = hrtimer_update_next_event(cpu_base);
 
-	__hrtimer_reprogram(cpu_base, skip_equal, cpu_base->next_timer,
-			    expires_next);
+	if (skip_equal && expires_next == cpu_base->expires_next)
+		return;
+
+	__hrtimer_reprogram(cpu_base, cpu_base->next_timer, expires_next);
 }
 
 /* High resolution timer related functions */
@@ -863,7 +851,19 @@ static void hrtimer_reprogram(struct hrt
 	if (base->cpu_base != cpu_base)
 		return;
 
-	__hrtimer_reprogram(cpu_base, true, timer, expires);
+	if (expires >= cpu_base->expires_next)
+		return;
+
+	/*
+	 * If the hrtimer interrupt is running, then it will reevaluate the
+	 * clock bases and reprogram the clock event device.
+	 */
+	if (cpu_base->in_hrtirq)
+		return;
+
+	cpu_base->next_timer = timer;
+
+	__hrtimer_reprogram(cpu_base, timer, expires);
 }
 
 static bool update_needs_ipi(struct hrtimer_cpu_base *cpu_base,
