From: "Rafael J. Wysocki" <rjw@sisk.pl>
To: Jiri Slaby <jirislaby@gmail.com>
Subject: Re: Fan runs full speed after S3 sometimes (regression)
Date: Sat, 21 Aug 2010 01:30:03 +0200
User-Agent: KMail/1.13.5 (Linux/2.6.36-rc1-rjw+; KDE/4.4.4; x86_64; ; )
Cc: Pavel Machek <pavel@ucw.cz>, linux-acpi@vger.kernel.org,
        "Linux-pm mailing list" <linux-pm@lists.linux-foundation.org>,
        LKML <linux-kernel@vger.kernel.org>
References: <4C613619.30309@gmail.com> <20100820063107.GB1607@ucw.cz> <4C6E233A.5000209@gmail.com>
In-Reply-To: <4C6E233A.5000209@gmail.com>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Message-Id: <201008210130.03280.rjw@sisk.pl>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1650
Lines: 40

On Friday, August 20, 2010, Jiri Slaby wrote:
> On 08/20/2010 08:31 AM, Pavel Machek wrote:
> > On Tue 2010-08-10 20:39:44, Jiri Slaby wrote:
> >> On 08/10/2010 01:20 PM, Jiri Slaby wrote:
> >>> Hi,
> >>>
> >>> till I was running on 2.6.34-rc5-mm1 (mmotm) and earlier everything was
> >>> fine. Now, I'm running 2.6.35-rc6-mm1 (.35-rc5-mm1 was the same) and
> >>> sometimes when the computer (desktop) returns from S3, the CPU fan is
> >>> running at full speed making a big noise.
> >>
> >> Just to add, full power cycle is needed to get rid of that. Warm reboot
> >> is not enough.
> > 
> > Seems like bios problem... Update bios?
> 
> I don't think so, it had been working for more than 2 years before this
> issue appeared.
> 
> >>> I don't know how to debug that. Bisection is out of question -- it
> >>> happens once in 10 or 20 cases. There are no fan entries in /sys.
> > 
> > Well, doing 20 suspends in a row from a script should not be too bad.
> 
> 20 is just a guess. I haven't seen it for pretty few days until the day
> before yesterday. And it looks like CPU1 is brought up a half way. When
> the fan was at full speed I tried:
> echo 0 > /sys/devices/system/cpu/cpu1/online
> and whole system freezed. Otherwise this commands succeeds.
> 
> I'll load a crash kernel prior turning the CPU off next time.

You can also try to use RTC wakealarm to run suspend/resume in a tight loop.

Rafael
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/