mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: linux-kernel@vger.kernel.org, mingo@kernel.org, rjw@rjwysocki.net
Cc: tglx@linutronix.de, peterz@infradead.org,
	Len Brown <lenb@kernel.org>,
	linux-acpi@vger.kernel.org
Subject: [PATCH 01/35] ACPI/acpi_pad: Remove the local apic nonsense
Date: Mon, 16 Feb 2015 13:14:36 +0100	[thread overview]
Message-ID: <20150216122412.023909722@infradead.org> (raw)
In-Reply-To: <20150216121435.203983131@infradead.org>

[-- Attachment #1: acpi-pad-kill-lapic-nonsense.patch --]
[-- Type: text/plain, Size: 3882 bytes --]

From: Thomas Gleixner <tglx@linutronix.de>

While looking through the (ab)use of the clockevents_notify() function
I stumbled over the following gem in the acpi_pad code:

  if (lapic_detected_unstable && !lapic_marked_unstable) {
     /* LAPIC could halt in idle, so notify users */
     for_each_online_cpu(i)
       clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ON, &i);
     lapic_marked_unstable = 1;
  }

This code calls on the cpu which detects the lapic unstable condition
first clockevents_notify() to tell the core code that the broadcast
should be enabled on all online cpus. Brilliant stuff that as it
notifies the core code a num_online_cpus() times that the broadcast
should be enabled on the current cpu.

This probably has never been noticed because that code got never
tested with NOHZ=n and HIGHRES_TIMER=n or it just worked by chance
because one of the other mechanisms told the core in the right way
that the local apic timer is wreckaged.

Sigh, this is:

 - The 4th incarnation of idle drivers which has their own mechanism
   to detect and deal with X86_FEATURE_ARAT.

 - The 2nd incarnation of fake idle mechanisms with a different set of
   brainmelting bugs.

 - Has been merged against an explicit NAK of the scheduler
   maintainer with the promise to improve it over time.

 - Another example of featuritis driven trainwreck engineering.

 - Another pointless waste of my time.

Fix this nonsense by removing that lapic detection and notification
logic and simply call into the clockevents code unconditonally. The
ARAT feature is marked in the lapic clockevent already so the core
code will just ignore the requests and return.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Len Brown <lenb@kernel.org>
Cc: linux-acpi@vger.kernel.org
Cc: Peter Zijlstra <peterz@infradead.org>
---
 drivers/acpi/acpi_pad.c |   26 +++++---------------------
 1 file changed, 5 insertions(+), 21 deletions(-)

Index: linux/drivers/acpi/acpi_pad.c
===================================================================
--- linux.orig/drivers/acpi/acpi_pad.c
+++ linux/drivers/acpi/acpi_pad.c
@@ -41,8 +41,6 @@ static unsigned long power_saving_mwait_
 
 static unsigned char tsc_detected_unstable;
 static unsigned char tsc_marked_unstable;
-static unsigned char lapic_detected_unstable;
-static unsigned char lapic_marked_unstable;
 
 static void power_saving_mwait_init(void)
 {
@@ -82,13 +80,10 @@ static void power_saving_mwait_init(void
 		 */
 		if (!boot_cpu_has(X86_FEATURE_NONSTOP_TSC))
 			tsc_detected_unstable = 1;
-		if (!boot_cpu_has(X86_FEATURE_ARAT))
-			lapic_detected_unstable = 1;
 		break;
 	default:
-		/* TSC & LAPIC could halt in idle */
+		/* TSC could halt in idle */
 		tsc_detected_unstable = 1;
-		lapic_detected_unstable = 1;
 	}
 #endif
 }
@@ -177,28 +172,17 @@ static int power_saving_thread(void *dat
 				mark_tsc_unstable("TSC halts in idle");
 				tsc_marked_unstable = 1;
 			}
-			if (lapic_detected_unstable && !lapic_marked_unstable) {
-				int i;
-				/* LAPIC could halt in idle, so notify users */
-				for_each_online_cpu(i)
-					clockevents_notify(
-						CLOCK_EVT_NOTIFY_BROADCAST_ON,
-						&i);
-				lapic_marked_unstable = 1;
-			}
+			clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ON, &cpu);
+
 			local_irq_disable();
 			cpu = smp_processor_id();
-			if (lapic_marked_unstable)
-				clockevents_notify(
-					CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &cpu);
+			clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &cpu);
 			stop_critical_timings();
 
 			mwait_idle_with_hints(power_saving_mwait_eax, 1);
 
 			start_critical_timings();
-			if (lapic_marked_unstable)
-				clockevents_notify(
-					CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &cpu);
+			clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &cpu);
 			local_irq_enable();
 
 			if (time_before(expire_time, jiffies)) {



  reply	other threads:[~2015-02-16 13:16 UTC|newest]

Thread overview: 106+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-02-16 12:14 [PATCH 00/35] Some kernel/time cleanups Peter Zijlstra
2015-02-16 12:14 ` Peter Zijlstra [this message]
2015-02-16 18:54   ` [PATCH 01/35] ACPI/acpi_pad: Remove the local apic nonsense Rafael J. Wysocki
2015-02-18  6:16     ` Rafael J. Wysocki
2015-02-16 12:14 ` [PATCH 02/35] clockevents: Remove CONFIG_GENERIC_CLOCKEVENTS_BUILD Peter Zijlstra
2015-02-16 12:14 ` [PATCH 03/35] tick: Cleanup internal header files Peter Zijlstra
2015-02-16 12:14 ` [PATCH 04/35] tick: Cleanup internal header file Peter Zijlstra
2015-02-16 12:14 ` [PATCH 05/35] tick: Move core only declarations and functions to core Peter Zijlstra
2015-02-16 12:14 ` [PATCH 06/35] clockevents: Remove extra local_irq_save in clockevents_exchange_device() Peter Zijlstra
2015-02-16 12:14 ` [PATCH 07/35] clockevents: Make suspend/resume calls explicit Peter Zijlstra
2015-02-16 19:07   ` Rafael J. Wysocki
2015-02-16 12:14 ` [PATCH 08/35] tick: Make tick_resume_broadcast_oneshot static Peter Zijlstra
2015-02-16 12:14 ` [PATCH 09/35] tick/xen: Provide and use tick_resume_local() Peter Zijlstra
2015-02-16 16:37   ` David Vrabel
2015-02-16 12:14 ` [PATCH 10/35] clockevents: Make tick handover explicit Peter Zijlstra
2015-02-16 12:14 ` [PATCH 11/35] clockevents: Cleanup dead cpu explicitely Peter Zijlstra
2015-02-16 12:14 ` [PATCH 12/35] clockevents: Provide explicit broadcast control function Peter Zijlstra
2015-02-23 18:08   ` Daniel Lezcano
2015-02-24  0:38   ` Tony Lindgren
2015-02-16 12:14 ` [PATCH 13/35] x86, amd_idle: Use explicit broadcast control functions Peter Zijlstra
2015-02-16 12:14 ` [PATCH 14/35] ACPI / PAD: Use explicit broadcast control function Peter Zijlstra
2015-02-16 12:14 ` [PATCH 15/35] ACPI/ processor_idle: Use explicit broadcast controll function Peter Zijlstra
2015-02-16 12:14 ` [PATCH 16/35] cpuidle: Use explicit broadcast control function Peter Zijlstra
2015-02-23 18:11   ` Daniel Lezcano
2015-02-16 12:14 ` [PATCH 17/35] intel_idle: " Peter Zijlstra
2015-02-16 12:14 ` [PATCH 18/35] ARM: OMAP: " Peter Zijlstra
2015-02-24  0:40   ` Tony Lindgren
2015-02-16 12:14 ` [PATCH 19/35] clockevents: Remove the broadcast control leftovers Peter Zijlstra
2015-02-16 12:14 ` [PATCH 20/35] clockevents: Provide explicit broadcast oneshot control function Peter Zijlstra
2015-02-24  0:41   ` Tony Lindgren
2015-02-16 12:14 ` [PATCH 21/35] x86, amd_idle: Use explicit broadcast oneshot control functions Peter Zijlstra
2015-02-16 12:14 ` [PATCH 22/35] ACPI / PAD: Use explicit broadcast oneshot control function Peter Zijlstra
2015-02-16 12:14 ` [PATCH 23/35] ACPI/ processor_idle: Use explicit broadcast oneshot controll function Peter Zijlstra
2015-02-16 12:14 ` [PATCH 24/35] intel_idle: Use explicit broadcast oneshot control function Peter Zijlstra
2015-02-16 12:15 ` [PATCH 25/35] ARM: OMAP: " Peter Zijlstra
2015-02-24  0:41   ` Tony Lindgren
2015-02-16 12:15 ` [PATCH 26/35] ARM: tegra: " Peter Zijlstra
2015-02-16 12:15 ` [PATCH 27/35] sched/idle: " Peter Zijlstra
2015-02-21  0:56   ` Nicolas Pitre
2015-02-21 11:19     ` Peter Zijlstra
2015-02-16 12:15 ` [PATCH 28/35] clockevents: Remove the broadcast oneshot control leftovers Peter Zijlstra
2015-02-16 12:15 ` [PATCH 29/35] tick/timekeeping: Make PM_SLEEP related code conditional Peter Zijlstra
2015-02-16 12:15 ` [PATCH 30/35] timekeeping: Get rid of stale comment Peter Zijlstra
2015-02-16 12:15 ` [PATCH 31/35] clockevents: Consolidate tick_shutdown() Peter Zijlstra
2015-02-16 12:15 ` [PATCH 32/35] clockevents: Fix cpu down race for hrtimer based broadcasting Peter Zijlstra
2015-02-17  4:03   ` Preeti U Murthy
2015-02-17 10:39     ` Peter Zijlstra
2015-02-18  3:10       ` Preeti U Murthy
2015-02-18 13:06         ` Peter Zijlstra
2015-02-19  7:01           ` Preeti U Murthy
2015-02-19  9:53             ` Peter Zijlstra
2015-02-19 17:51   ` Nicolas Pitre
2015-02-21 12:46     ` Peter Zijlstra
2015-02-21 17:45       ` Nicolas Pitre
2015-02-23 16:14         ` Peter Zijlstra
2015-02-23 16:32           ` Nicolas Pitre
2015-02-23 17:33             ` Nicolas Pitre
2015-02-26  5:31               ` Preeti U Murthy
2015-02-27  8:49                 ` Preeti U Murthy
2015-03-02 14:56                   ` Peter Zijlstra
2015-03-05  4:20                     ` Preeti U Murthy
2015-03-29  6:45                       ` Ingo Molnar
2015-03-16  4:52                     ` Preeti U Murthy
2015-03-25  5:32                       ` Preeti U Murthy
2015-02-16 12:15 ` [PATCH 33/35] tick: Make hrtimer broadcasting opt-in Peter Zijlstra
2015-02-17  3:06   ` Preeti U Murthy
2015-02-17 10:04   ` Lorenzo Pieralisi
2015-02-16 12:15 ` [PATCH 34/35] tick: Provide tick_suspend_local() Peter Zijlstra
2015-02-17 14:31   ` Lorenzo Pieralisi
2015-02-17 17:55     ` Nicolas Pitre
2015-02-17 21:29       ` Nicolas Pitre
2015-02-18 13:35     ` Peter Zijlstra
2015-02-18 16:22       ` Rafael J. Wysocki
2015-02-18 16:38       ` Lorenzo Pieralisi
2015-02-16 12:15 ` [PATCH 35/35] arm: bl_switcher: Kill tick suspend hackery Peter Zijlstra
2015-02-16 19:14 ` [PATCH 00/35] Some kernel/time cleanups Rafael J. Wysocki
2015-03-25 12:04 ` [PATCH 00/09] Some kernel/time cleanups (related to timer_suspend/resume) Rafael J. Wysocki
2015-03-25 12:05   ` [PATCH 01/09] clockevents: Remove CONFIG_GENERIC_CLOCKEVENTS_BUILD Rafael J. Wysocki
2015-04-02 12:28     ` [tip:timers/core] " tip-bot for Thomas Gleixner
2015-03-25 12:06   ` [PATCH 02/09] tick: Move clocksource related stuff to timekeeping.h Rafael J. Wysocki
2015-04-02 12:28     ` [tip:timers/core] " tip-bot for Thomas Gleixner
2015-03-25 12:06   ` [PATCH 03/09] tick: Simplify tick-internal.h Rafael J. Wysocki
2015-04-02 12:29     ` [tip:timers/core] " tip-bot for Thomas Gleixner
2015-03-25 12:07   ` [PATCH 04/09] tick: Move core only declarations and functions to core Rafael J. Wysocki
2015-04-02 12:29     ` [tip:timers/core] " tip-bot for Thomas Gleixner
2015-03-25 12:08   ` [PATCH 05/09] clockevents: Remove extra local_irq_save in clockevents_exchange_device() Rafael J. Wysocki
2015-04-02 12:29     ` [tip:timers/core] clockevents: Remove extra local_irq_save() " tip-bot for Thomas Gleixner
2015-03-25 12:09   ` [PATCH 06/09] clockevents: Make suspend/resume calls explicit Rafael J. Wysocki
2015-04-02 12:30     ` [tip:timers/core] " tip-bot for Thomas Gleixner
2015-03-25 12:09   ` [PATCH 07/09] tick: Make tick_resume_broadcast_oneshot static Rafael J. Wysocki
2015-04-02 12:30     ` [tip:timers/core] tick: Make tick_resume_broadcast_oneshot() static tip-bot for Thomas Gleixner
2015-03-25 12:11   ` [PATCH 08/09] tick/xen: Provide and use tick_suspend_local() and tick_resume_local() Rafael J. Wysocki
2015-03-25 20:13     ` Konrad Rzeszutek Wilk
2015-03-25 20:57       ` Rafael J. Wysocki
2015-03-25 20:44         ` Konrad Rzeszutek Wilk
2015-03-25 21:36           ` Rafael J. Wysocki
2015-04-02 12:30     ` [tip:timers/core] " tip-bot for Thomas Gleixner
2015-04-03 13:21       ` [PATCH][fix for tip/timers/core] timers / PM: Fix up tick_unfreeze() Rafael J. Wysocki
2015-04-03 13:18         ` [tip:timers/core] timers/PM: " tip-bot for Rafael J. Wysocki
2015-04-03 13:31         ` [PATCH][cleanup for tip/timers/core] timers / PM: Drop unnecessary braces from tick_freeze() Rafael J. Wysocki
2015-04-03 13:18           ` [tip:timers/core] timers/PM: " tip-bot for Rafael J. Wysocki
2015-03-25 12:11   ` [PATCH 09/09] arm: bl_switcher: Kill tick suspend hackery Rafael J. Wysocki
2015-03-25 12:34     ` Nicolas Pitre
2015-04-02 12:30     ` [tip:timers/core] arm/bL_switcher: " tip-bot for Thomas Gleixner
2015-03-25 12:13   ` [PATCH 00/09] Some kernel/time cleanups (related to timer_suspend/resume) Rafael J. Wysocki
2015-03-31 22:45   ` [GIT PULL] kernel/time cleanups (related to tick_suspend/resume) Rafael J. Wysocki

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=20150216122412.023909722@infradead.org \
    --to=peterz@infradead.org \
    --cc=lenb@kernel.org \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=rjw@rjwysocki.net \
    --cc=tglx@linutronix.de \
    /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

all inboxes | Powered by JetHome®