From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarkko Nikula Subject: Re: cpuidle status in mainline for Beagleboard xM Date: Fri, 02 Sep 2011 09:05:53 +0300 Message-ID: <4E607241.8000704@bitmer.com> References: <87vctcjo2w.fsf@ti.com> <87ei00f94i.fsf@ti.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from bitmer.com ([213.157.87.50]:48596 "EHLO bitmer.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758094Ab1IBGGL (ORCPT ); Fri, 2 Sep 2011 02:06:11 -0400 In-Reply-To: <87ei00f94i.fsf@ti.com> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Kevin Hilman Cc: javier Martin , linux-omap@vger.kernel.org, Koen Kooi On 09/01/2011 08:46 PM, Kevin Hilman wrote: > javier Martin writes: > >> Hi Kevin, >> thanks for your help. >> >>> CPU is staying in C0 probably because UARTs are not being idled, so SoC >>> cannot hit deeper idle states. Try the following at the command line to >>> to enable UART idle timeouts, so the SoC can attempt idle after the >>> timeout period >>> >>> # UART timeouts: omap-serial (4th UART only on OMAP36xx and OMAP4) >>> echo 5> /sys/devices/platform/omap/omap_uart.0/sleep_timeout >>> echo 5> /sys/devices/platform/omap/omap_uart.1/sleep_timeout >>> echo 5> /sys/devices/platform/omap/omap_uart.2/sleep_timeout >>> echo 5> /sys/devices/platform/omap/omap_uart.3/sleep_timeout >>> >>> After 5 seconds of inactivity on the UARTs, you should see the SoC >>> hitting deeper C-states. >> I've tried that but it still doesn't hit any C-state deeper than 0. >> I'll try the same test using your pm branch you pointed me out and >> post the results. >> > The CPUidle stuff works in mainline (after allowing UARTs to idle). Other usual things to check that display is off (echo 1 > /sys/class/graphics/fb0/blank) and no cable to musb/otg port. Haven't tried myself with recent kernel but does EHCI and hub on XM let to idle cpu at all? At least on one board having on-board hub I had to disable or unload ehci module in order to hit the retention. -- Jarkko