From: Pierre Ossman <drzeus-list@drzeus.cx>
To: "LKML" <linux-kernel@vger.kernel.org>,
linux-pm@lists.linux-foundation.org, "Pallipadi,
Venkatesh" <venkatesh.pallipadi@intel.com>
Cc: "Dave Jones" <davej@codemonkey.org.uk>,
"Andi Kleen" <andi@firstfloor.org>,
"Alan Stern" <stern@rowland.harvard.edu>,
"Adam Belay" <abelay@novell.com>,
"Lee Revell" <rlrevell@joe-job.com>,
"Pavel Machek" <pavel@ucw.cz>
Subject: Re: [linux-pm] [PATCH] cpuidle: avoid singing capacitors
Date: Tue, 11 Mar 2008 08:51:45 +0100 [thread overview]
Message-ID: <20080311085145.5fcf3186@mjolnir.drzeus.cx> (raw)
In-Reply-To: <20080309151659.63d54f38@mjolnir.drzeus.cx>
I had a sudden surge of inspiration and decided to follow up on this
one a bit:
>
> 5. Second core
>
> Disabling the second core makes the noise go away. This might be a subset of 1., as I've been told that a stopped core enters C1.
>
I have now found a very hacky workaround that is slightly better than
disabling C3 altogether; making C3 exclusive to one core at a time. It
seems to kill the noise and the system now spends 50% in C3, instead of
0%.
I read somewhere that the Intel Duo:s have an extra low power mode that
is entered if both cores are suspended at the same time (disabling the
shared cache or something like that I guess). So perhaps that mode is
the culprit. I'm hoping there's a way to disable just that feature. Do
any of you know a way? Venkatesh perhaps?
Here's the patch I'm currently running at least:
diff --git a/drivers/cpuidle/governors/menu.c b/drivers/cpuidle/governors/menu.c
index 78d77c5..9a859fb 100644
--- a/drivers/cpuidle/governors/menu.c
+++ b/drivers/cpuidle/governors/menu.c
@@ -16,6 +16,8 @@
#define BREAK_FUZZ 4 /* 4 us */
+static atomic_t cur_deep_cpu = ATOMIC_INIT(-1);
+
struct menu_device {
int last_state_idx;
@@ -50,6 +52,13 @@ static int menu_select(struct cpuidle_device *dev)
break;
if (s->exit_latency > pm_qos_requirement(PM_QOS_CPU_DMA_LATENCY))
break;
+
+ if (s->flags & CPUIDLE_FLAG_DEEP) {
+ rmb();
+ if (atomic_cmpxchg(&cur_deep_cpu, -1, dev->cpu) != -1)
+ break;
+ wmb();
+ }
}
data->last_state_idx = i - 1;
@@ -80,6 +89,11 @@ static void menu_reflect(struct cpuidle_device *dev)
if (!(target->flags & CPUIDLE_FLAG_TIME_VALID))
measured_us = USEC_PER_SEC / HZ;
+ if (atomic_read(&cur_deep_cpu) == dev->cpu) {
+ atomic_set(&cur_deep_cpu, -1);
+ wmb();
+ }
+
/* Predict time remaining until next break event */
if (measured_us + BREAK_FUZZ < data->expected_us - target->exit_latency) {
data->predicted_us = max(measured_us, data->last_measured_us);
--
-- Pierre Ossman
Linux kernel, MMC maintainer http://www.kernel.org
PulseAudio, core developer http://pulseaudio.org
rdesktop, core developer http://www.rdesktop.org
next prev parent reply other threads:[~2008-03-11 7:52 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-02-29 18:38 [RFC][PATCH] cpuidle: avoid singing capacitors Pierre Ossman
2008-02-29 21:44 ` Lennart Sorensen
2008-03-01 12:31 ` Pierre Ossman
2008-03-01 13:40 ` Pierre Ossman
2008-03-02 2:27 ` Lee Revell
2008-03-02 14:17 ` Pierre Ossman
2008-03-03 12:36 ` Andi Kleen
2008-03-03 20:18 ` [PATCH] " Pierre Ossman
2008-03-03 20:46 ` Pavel Machek
2008-03-03 21:03 ` Pierre Ossman
2008-03-03 21:08 ` Pavel Machek
2008-03-03 21:14 ` Pallipadi, Venkatesh
2008-03-03 21:17 ` Pierre Ossman
2008-03-03 22:04 ` Pallipadi, Venkatesh
2008-03-03 23:05 ` [linux-pm] " Alan Stern
2008-03-03 23:10 ` Andi Kleen
2008-03-04 4:00 ` Dave Jones
2008-03-04 6:14 ` Pierre Ossman
2008-03-04 17:19 ` Pierre Ossman
2008-03-04 17:29 ` Andi Kleen
2008-03-04 17:30 ` Pierre Ossman
2008-03-04 17:43 ` Andi Kleen
2008-03-04 18:04 ` Pierre Ossman
2008-03-04 18:34 ` Andi Kleen
2008-03-05 6:04 ` Pierre Ossman
2008-03-05 15:48 ` Andi Kleen
2008-03-05 16:53 ` Pierre Ossman
2008-03-05 17:32 ` Andi Kleen
2008-03-04 19:01 ` Pallipadi, Venkatesh
2008-03-05 6:02 ` Pierre Ossman
2008-03-05 8:40 ` Pierre Ossman
2008-03-05 9:03 ` Pavel Machek
2008-03-05 13:42 ` Pierre Ossman
2008-03-05 13:47 ` Pavel Machek
2008-03-05 13:52 ` Pierre Ossman
2008-03-06 8:27 ` Pierre Ossman
2008-03-09 14:16 ` Pierre Ossman
2008-03-09 18:19 ` Rafael J. Wysocki
2008-03-09 18:50 ` Alan Stern
2008-03-09 19:30 ` Henrique de Moraes Holschuh
2008-03-09 20:14 ` Pierre Ossman
2008-03-09 20:41 ` Henrique de Moraes Holschuh
2008-03-09 20:54 ` Henrique de Moraes Holschuh
2008-03-10 10:00 ` Pavel Machek
2008-03-10 12:49 ` Pierre Ossman
2008-03-10 13:04 ` Andi Kleen
2008-03-10 13:29 ` Pierre Ossman
2008-03-12 19:11 ` Len Brown
2008-03-13 8:10 ` Pavel Machek
2008-03-13 10:42 ` Andi Kleen
2008-03-14 4:13 ` Len Brown
2008-03-13 16:34 ` Pierre Ossman
2008-03-13 16:47 ` Pallipadi, Venkatesh
2008-03-13 17:44 ` Pierre Ossman
2008-03-13 17:49 ` Pierre Ossman
2008-03-14 19:40 ` Pierre Ossman
2008-03-14 21:15 ` Pallipadi, Venkatesh
2008-03-15 0:41 ` Pierre Ossman
2008-03-11 7:51 ` Pierre Ossman [this message]
2008-03-11 10:48 ` Andi Kleen
2008-03-11 15:20 ` Pierre Ossman
2008-03-11 17:31 ` Pierre Ossman
2008-03-12 19:17 ` Len Brown
2008-03-12 20:31 ` Len Brown
2008-03-04 9:40 ` Andi Kleen
2008-03-03 23:09 ` Andi Kleen
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=20080311085145.5fcf3186@mjolnir.drzeus.cx \
--to=drzeus-list@drzeus.cx \
--cc=abelay@novell.com \
--cc=andi@firstfloor.org \
--cc=davej@codemonkey.org.uk \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@lists.linux-foundation.org \
--cc=pavel@ucw.cz \
--cc=rlrevell@joe-job.com \
--cc=stern@rowland.harvard.edu \
--cc=venkatesh.pallipadi@intel.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;
as well as URLs for NNTP newsgroup(s).