lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:	Tue, 08 May 2012 09:44:18 +0800
From:	Huang Ying <ying.huang@...el.com>
To:	"Rafael J. Wysocki" <rjw@...k.pl>
Cc:	huang ying <huang.ying.caritas@...il.com>,
	Bjorn Helgaas <bhelgaas@...gle.com>, ming.m.lin@...el.com,
	linux-kernel@...r.kernel.org, linux-pm@...r.kernel.org,
	Zheng Yan <zheng.z.yan@...el.com>,
	Lan Tianyu <tianyu.lan@...el.com>
Subject: Re: [RFC v2 2/5] PM, Add sysfs file power_off to control device
 power off policy

On Mon, 2012-05-07 at 22:53 +0200, Rafael J. Wysocki wrote:
> On Saturday, May 05, 2012, huang ying wrote:
> > On Sat, May 5, 2012 at 3:33 AM, Rafael J. Wysocki <rjw@...k.pl> wrote:
> > > On Friday, May 04, 2012, Huang Ying wrote:
> > >> From: Lan Tianyu <tianyu.lan@...el.com>
> > >>
> > >> Some devices can be powered off to save more power via some platform
> > >> mechanism, e.g., ACPI.  But that may not work as expected for some
> > >> device or platform.  So, this patch adds a sysfs file named power_off
> > >> under <device>/power directory to provide a mechanism for user to control
> > >> whether to allow the device to be power off.
> > >>
> > >> power_off => "enabled" means allowing the device to be powered off if
> > >> possible.
> > >>
> > >> power_off => "disabled" means the device must be power on anytime.
> > >>
> > >> Also add flag power_off_user to struct dev_pm_info to record users'
> > >> choice. The bus layer can use this field to determine whether to
> > >> power off the device.
> > >
> > > It looks like the new attribute is added for all devices regardless of whether
> > > or not they actually can be powered off?  If so, then please don't do that,
> > > it's _extremely_ confusing.
> > 
> > Yes.  You are right.
> > 
> > > If you need user space to be able to control that functionality (and I can
> > > think of a couple of cases in which you do), there need to be 2 flags,
> > > can_power_off and may_power_off, where the first one is set by the kernel
> > > if it is known that power can be removed from the device - the attribute
> > > should be created when this flag is set and removed when it is unset.
> > >
> > > Then, the setting of the second flag will be controlled by the new attribute.
> > >
> > > And you'll need to patch quite a few places where devices actually have that
> > > capability, like where power domains are in use, so that's quire a lot of
> > > work.
> > 
> > If so, I think maybe we need 3 flags:
> > 
> > - can_power_off, set by kernel when it is possible to power off the device
> 
> Well, on a second thought, it may be difficult to determine that in some
> cases (eg. for devices belonging to power domains with additional constraints
> related to the other devices in the same domain etc.).
> 
> In other cases power may be removed from devices indirectly, like for example
> by putting a device's parent into a low power state.
> 
> So I guess the can_power_off flag may not be practical after all.
> 
> > - may_power_off_user, set by user via sysfs attribute
> 
> I'd call that one power_off_allowed, meaning "allowed by user space".
> 
> > - may_power_off, set by kernel according to may_power_off_user, power
> > QoS and some other conditions
> 
> And I'd call that one power_must_be_on, meaning "don't power off even if
> allowed by user space".
> 
> > Sysfs attribute for may_power_off_user is only created if can_power_off is true.
> > 
> > I think we still can do that step by step.  For example, when we add
> > power off support to PCI devices, we set can_power_off to true for PCI
> > devices that is possible to be powered off;  when we add power domain
> > support, we set can_power_off to true for devices in power domain.  Do
> > you agree?
> 
> I think we may add helpers for exporting/unexporting power_off_allowed
> like for the PM QoS latency attribute.  Then, whoever wants to support
> power_off_allowed and use it will export it through that helper.

That sounds good!

> Still, I'm afraid we're trying to special case something that really ins't
> a special case.  Namely, we may want to restrict devices from using some
> other low-power states as well, not only power off (eg. we may want to
> prevent devices' clocks from being stopped).

One step towards generalization is to provide a way for user to specify
lowest power state allowed.  For example, for PCI devices, they can
specify D1, D2, D3hot or D3cold.  But it is hard to generalize a set of
low power states for all kind of devices.  Maybe we should keep this
user space interface bus specific?  For example, we can have a sysfs
file like "lowest_pm_state_allowed" for each PCI devices.

BTW:  I wonder that are there standard low power states defined for
devices on platform bus.

Best Regards,
Huang Ying


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

Powered by blists - more mailing lists