mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Ankur Arora <ankur.a.arora@oracle.com>
To: linux-pm@vger.kernel.org, kvm@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org
Cc: catalin.marinas@arm.com, will@kernel.org, tglx@linutronix.de,
	mingo@redhat.com, bp@alien8.de, dave.hansen@linux.intel.com,
	x86@kernel.org, hpa@zytor.com, pbonzini@redhat.com,
	wanpengli@tencent.com, vkuznets@redhat.com, rafael@kernel.org,
	daniel.lezcano@linaro.org, peterz@infradead.org, arnd@arndb.de,
	lenb@kernel.org, mark.rutland@arm.com, harisokn@amazon.com,
	mtosatti@redhat.com, sudeep.holla@arm.com, cl@gentwo.org,
	misono.tomohiro@fujitsu.com, maobibo@loongson.cn,
	joao.m.martins@oracle.com, boris.ostrovsky@oracle.com,
	konrad.wilk@oracle.com
Subject: [PATCH v8 01/11] cpuidle/poll_state: poll via smp_cond_load_relaxed()
Date: Wed, 25 Sep 2024 16:24:15 -0700	[thread overview]
Message-ID: <20240925232425.2763385-2-ankur.a.arora@oracle.com> (raw)
In-Reply-To: <20240925232425.2763385-1-ankur.a.arora@oracle.com>

From: Mihai Carabas <mihai.carabas@oracle.com>

The inner loop in poll_idle() polls up to POLL_IDLE_RELAX_COUNT times,
checking to see if the thread has the TIF_NEED_RESCHED bit set. The
loop exits once the condition is met, or if the poll time limit has
been exceeded.

To minimize the number of instructions executed each iteration, the
time check is done only infrequently (once every POLL_IDLE_RELAX_COUNT
iterations). In addition, each loop iteration executes cpu_relax()
which on certain platforms provides a hint to the pipeline that the
loop is busy-waiting, thus allowing the processor to reduce power
consumption.

However, cpu_relax() is defined optimally only on x86. On arm64, for
instance, it is implemented as a YIELD which only serves a hint to the
CPU that it prioritize a different hardware thread if one is available.
arm64, however, does expose a more optimal polling mechanism via
smp_cond_load_relaxed() which uses LDXR, WFE to wait until a store
to a specified region.

So restructure the loop, folding both checks in smp_cond_load_relaxed().
Also, move the time check to the head of the loop allowing it to exit
straight-away once TIF_NEED_RESCHED is set.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Mihai Carabas <mihai.carabas@oracle.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Misono Tomohiro <misono.tomohiro@fujitsu.com>
Signed-off-by: Ankur Arora <ankur.a.arora@oracle.com>
---
 drivers/cpuidle/poll_state.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/cpuidle/poll_state.c b/drivers/cpuidle/poll_state.c
index 9b6d90a72601..fc1204426158 100644
--- a/drivers/cpuidle/poll_state.c
+++ b/drivers/cpuidle/poll_state.c
@@ -21,21 +21,20 @@ static int __cpuidle poll_idle(struct cpuidle_device *dev,
 
 	raw_local_irq_enable();
 	if (!current_set_polling_and_test()) {
-		unsigned int loop_count = 0;
 		u64 limit;
 
 		limit = cpuidle_poll_time(drv, dev);
 
 		while (!need_resched()) {
-			cpu_relax();
-			if (loop_count++ < POLL_IDLE_RELAX_COUNT)
-				continue;
-
-			loop_count = 0;
+			unsigned int loop_count = 0;
 			if (local_clock_noinstr() - time_start > limit) {
 				dev->poll_time_limit = true;
 				break;
 			}
+
+			smp_cond_load_relaxed(&current_thread_info()->flags,
+					      VAL & _TIF_NEED_RESCHED ||
+					      loop_count++ >= POLL_IDLE_RELAX_COUNT);
 		}
 	}
 	raw_local_irq_disable();
-- 
2.43.5


  reply	other threads:[~2024-09-25 23:25 UTC|newest]

Thread overview: 74+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-09-25 23:24 [PATCH v8 00/11] Enable haltpoll on arm64 Ankur Arora
2024-09-25 23:24 ` Ankur Arora [this message]
2024-10-15 12:04   ` [PATCH v8 01/11] cpuidle/poll_state: poll via smp_cond_load_relaxed() Catalin Marinas
2024-10-15 16:42     ` Christoph Lameter (Ampere)
2024-10-15 16:50       ` Catalin Marinas
2024-10-15 17:17         ` Christoph Lameter (Ampere)
2024-10-15 17:40           ` Catalin Marinas
2024-10-15 21:53             ` Ankur Arora
2024-10-15 22:28               ` Christoph Lameter (Ampere)
2024-10-16  7:06                 ` Ankur Arora
2024-10-17 16:54                   ` Christoph Lameter (Ampere)
2024-10-17 18:36                     ` Ankur Arora
2024-10-15 22:40               ` Christoph Lameter (Ampere)
2024-10-16  9:54                 ` Catalin Marinas
2024-10-17 16:56                   ` Christoph Lameter (Ampere)
2024-10-17 18:15                     ` Catalin Marinas
2024-10-17 19:34                       ` Ankur Arora
2024-10-15 21:32     ` Ankur Arora
2024-10-16  6:20       ` maobibo
2024-10-16 10:06       ` Catalin Marinas
2024-10-16 15:13     ` Okanovic, Haris
2024-10-16 17:04       ` Ankur Arora
2024-10-16 18:04         ` Okanovic, Haris
2024-10-17 14:01       ` Catalin Marinas
2024-10-17 22:47         ` Ankur Arora
2024-10-18 11:05           ` Catalin Marinas
2024-10-18 19:00             ` Ankur Arora
2024-10-21 12:02               ` Catalin Marinas
2024-09-25 23:24 ` [PATCH v8 02/11] cpuidle: rename ARCH_HAS_CPU_RELAX to ARCH_HAS_OPTIMIZED_POLL Ankur Arora
2024-09-25 23:24 ` [PATCH v8 03/11] Kconfig: move ARCH_HAS_OPTIMIZED_POLL to arch/Kconfig Ankur Arora
2024-09-25 23:24 ` [PATCH v8 04/11] cpuidle-haltpoll: define arch_haltpoll_want() Ankur Arora
2024-09-25 23:24 ` [PATCH v8 05/11] governors/haltpoll: drop kvm_para_available() check Ankur Arora
2024-09-25 23:24 ` [PATCH v8 06/11] cpuidle-haltpoll: condition on ARCH_CPUIDLE_HALTPOLL Ankur Arora
2024-09-25 23:24 ` [PATCH v8 07/11] arm64: define TIF_POLLING_NRFLAG Ankur Arora
2024-09-25 23:24 ` [PATCH v8 08/11] arm64: idle: export arch_cpu_idle Ankur Arora
2024-09-25 23:24 ` [PATCH v8 09/11] arm64: select ARCH_HAS_OPTIMIZED_POLL Ankur Arora
2024-09-25 23:24 ` [PATCH v8 10/11] cpuidle/poll_state: limit POLL_IDLE_RELAX_COUNT on arm64 Ankur Arora
2024-09-25 23:24 ` [PATCH v8 11/11] arm64: support cpuidle-haltpoll Ankur Arora
2024-10-02 22:42   ` Okanovic, Haris
2024-10-03  3:29     ` Ankur Arora
2024-10-16 15:13   ` Okanovic, Haris
2024-10-09  2:37 ` [PATCH v8 00/11] Enable haltpoll on arm64 zhenglifeng (A)
2024-10-15  1:53   ` Ankur Arora
2024-10-14 22:54 ` Christoph Lameter (Ampere)
2024-10-15 12:36 ` Marc Zyngier
2024-10-16 21:55   ` Ankur Arora
2024-10-17  8:19     ` Marc Zyngier
2024-10-17 18:35       ` Ankur Arora
2024-10-22 22:01         ` Ankur Arora
2024-11-05 18:30 ` Haris Okanovic
2024-11-05 18:30   ` [PATCH 1/5] asm-generic: add smp_vcond_load_relaxed() Haris Okanovic
2024-11-05 19:36     ` Christoph Lameter (Ampere)
2024-11-06 17:06       ` Okanovic, Haris
2024-11-06 11:08     ` Catalin Marinas
2024-11-06 18:13       ` Okanovic, Haris
2024-11-06 19:55         ` Catalin Marinas
2024-11-06 20:31           ` Okanovic, Haris
2024-11-06 11:39     ` Will Deacon
2024-11-06 17:18       ` Okanovic, Haris
2024-11-05 18:30   ` [PATCH 2/5] arm64: add __READ_ONCE_EX() Haris Okanovic
2024-11-05 19:39     ` Christoph Lameter (Ampere)
2024-11-06 17:37       ` Okanovic, Haris
2024-11-06 11:43     ` Will Deacon
2024-11-06 17:09       ` Okanovic, Haris
2024-11-09  9:49     ` David Laight
2024-11-05 18:30   ` [PATCH 3/5] arm64: refactor delay() to enable polling for value Haris Okanovic
2024-11-05 19:42     ` Christoph Lameter (Ampere)
2024-11-06 17:42       ` Okanovic, Haris
2024-11-06  9:18     ` Catalin Marinas
2024-11-06 17:38       ` Okanovic, Haris
2024-11-05 18:30   ` [PATCH 4/5] arm64: add smp_vcond_load_relaxed() Haris Okanovic
2024-11-05 18:30   ` [PATCH 5/5] cpuidle: implement poll_idle() using smp_vcond_load_relaxed() Haris Okanovic
2024-11-05 19:45     ` Christoph Lameter (Ampere)
2024-11-05 18:49   ` [PATCH v8 00/11] Enable haltpoll on arm64 Ankur Arora

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=20240925232425.2763385-2-ankur.a.arora@oracle.com \
    --to=ankur.a.arora@oracle.com \
    --cc=arnd@arndb.de \
    --cc=boris.ostrovsky@oracle.com \
    --cc=bp@alien8.de \
    --cc=catalin.marinas@arm.com \
    --cc=cl@gentwo.org \
    --cc=daniel.lezcano@linaro.org \
    --cc=dave.hansen@linux.intel.com \
    --cc=harisokn@amazon.com \
    --cc=hpa@zytor.com \
    --cc=joao.m.martins@oracle.com \
    --cc=konrad.wilk@oracle.com \
    --cc=kvm@vger.kernel.org \
    --cc=lenb@kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=maobibo@loongson.cn \
    --cc=mark.rutland@arm.com \
    --cc=mingo@redhat.com \
    --cc=misono.tomohiro@fujitsu.com \
    --cc=mtosatti@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=peterz@infradead.org \
    --cc=rafael@kernel.org \
    --cc=sudeep.holla@arm.com \
    --cc=tglx@linutronix.de \
    --cc=vkuznets@redhat.com \
    --cc=wanpengli@tencent.com \
    --cc=will@kernel.org \
    --cc=x86@kernel.org \
    /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

Powered by JetHome