-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 11/30/2011 08:56 PM, Zach Pfeffer wrote:
On 30 November 2011 11:41, Amit Kucheria amit.kucheria@linaro.org wrote:
Wow - quite a cc-list :)
On Tue, Nov 29, 2011 at 7:47 PM, Zach Pfeffer zach.pfeffer@linaro.org wrote:
Amit/Mounir,
What's your guys plan with cpu_idle for each board? Are you going to try and upstream a solution that will work across all boards? Would you or Mounir be open to filing a BP per board so we can track when cpu_idle will hit each board? Does it make sense to prototype something across each board that we could land in Ubuntu and Android?
Adding other people, leads, etc...
I'll add four more.
The people I've added to cc are the ones looking at upstreaming cpuidle for the various member platforms (the first 3 are Linaro assignees):
Samsung - Amit Kachhap ST-E - Daniel Lezcano Freescale - Rob Lee TI - Kevin Hilman and Santosh Shilimkar
Rob tried to upstream a driver for the imx5 platform. Russell suggested that it is time to have a common cpuidle stub driver for ARM[1].
Daniel Lezcano and Rob are currently working on such a common stub driver. The imx5 cpuidle will get rebased on top of that. Daniel has actively started work on a u8500 cpuidle driver in the meanwhile but is awaiting documentation.
For OMAP4, I was told that there was now a cpuidle driver on its way into mainline. But I didn't find one after a cursory glance at the lists. Perhaps Kevin/Santosh can shed some light there. The last known tree I was aware of for an OMAP4 cpuidle driver was TI maintained one[2].
For exynos, there is a basic cpuidle driver in mainline. Amit Kachhap has been working on enhancing it[3] (adding more states and using more of the common code that was recently added).
In summary, we should have cpuidle working on the member platforms in the coming month. If you want something today, you should look at enabling OMAP4 and Exynos cpuidle configs.
Regards, Amit
Cool, thanks for the overview Amit.
Is there a test that a QA guy can run to verify that cpu_idle is working? A set of commands and a proc node to cat out with some stats perhaps?
This is something we are working on [1]. The test suite [2] has a few trivial tests but we want to add more tests, so it is is in the radar.
The test suite is integrated in LAVA and is run daily. The LAVA script checkout the head git tree and run automatically the tests. Each time we commit a new tests, it is take into account and run in LAVA.
For the moment, all the PM features are not available for all the boards, so most of the tests fail and the results are not very significant but the more we will merge the different PM blocks upstream, the greener the test suite will be :)
[1] https://blueprints.launchpad.net/linaro-power-qa/+spec/qa-cpuidle [2] http://git.linaro.org/gitweb?p=people/dlezcano/pm-qa.git%3Ba=summary
p.s Andy, was the breakage you reported in cpuidle related to the missing export.h patch?
[1] http://thread.gmane.org/gmane.linux.ports.arm.kernel/132859/focus=133279 [2] git://gitorious.org/omap-sw-develoment/linux-omap-dev.git for_3_2/omap4_mpuss_pm-integrated_2 [3] http://thread.gmane.org/gmane.linux.kernel.samsung-soc/7877
- -- http://www.linaro.org/ Linaro.org │ Open source software for ARM SoCs
Follow Linaro: http://www.facebook.com/pages/Linaro Facebook | http://twitter.com/#!/linaroorg Twitter | http://www.linaro.org/linaro-blog/ Blog