lkml.org 
[lkml]   [2011]   [Nov]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 5] PM: Update comments describing device power management callbacks
On Mon, 21 Nov 2011, Rafael J. Wysocki wrote:

> On Monday, November 21, 2011, Alan Stern wrote:
> > On Mon, 21 Nov 2011, Rafael J. Wysocki wrote:
> >
> > > > > I said "Analogous to @suspend()" instead. I'm not sure why this is not
> > > > > sufficient?
> > > >
> > > > Because @suspend() is very different! Its description basically says
> > > > to do three things:
> > > >
> > > > Quiesce the device,
> > > >
> > > > Put it into a low-power state,
> > > >
> > > > And enable wakeup events.
> > >
> > > No, it doesn't any more. It's being changed by the proposed patch too. :-)
> >
> > I must have missed reading that part. Okay... but it seems weird that
> > none of the new descriptions says anything about changing the power
> > state. Shouldn't the description of @suspend say something like "For
> > many platforms and subsystems, the device should be put in a low-power
> > state"?
>
> Hmm. I'm not really sure, actually, because I'd recommend that subsystems
> rather than drivers change power states of devices and this description
> is targeted at driver writers mostly.

The same data structure (dev_pm_ops) is used for both drivers and
subsystems. Therefore the comments should be directed toward both
driver writers and subsystem writers.

You could say: "For many platforms, the subsystem @suspend() or
@suspend_noirq() callback should put the device in a low-power state.
Some subsystems may require the driver to do this instead."

> > > > @freeze() is supposed to do the first but not the second or third.
> > > > This makes it only 33% similar to @suspend(). :-)
> > > >
> > > > Also, the description of @suspend() says nothing about having a
> > > > consistent memory image.
> > >
> > > Because that is irrelevant. The state of the device after the resume
> > > has to be consistent, regardless of whether the resume is from RAM or
> > > from an on-disk image.
> >
> > Sure, the device's state will be consistent. But will the contents of
> > memory image be consistent? Not if the device was doing DMA writes
> > during the time when the image was created.
>
> Well, since .suspend() is also expected to stop DMA, that's rather moot.

It won't be moot if you add the sentences I recommend above. You could
add an additional sentence: "Either way, the @freeze() and
@freeze_noirq() callbacks (both the subsystem's and the driver's)
should always avoid changing the device's state."

Alan Stern



\
 
 \ /
  Last update: 2011-11-22 17:01    [W:0.056 / U:0.664 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site