2.6.17: dmesg flooded with "ohci_hcd 0000:00:02.0: wakeup"

On Tuesday 14 November 2006 12:48 pm, Andrey Borzenkov wrote:
> On Monday 13 November 2006 22:58, Alan Stern wrote:
> > Andrey:
> >
> > Try this patch for 2.6.19-rc5. Although it doesn't make all the changes
> > Dave and I have discussed, it ought to fix your problem.
> >
>
> It did. Thank you

Then this should get merged into 2.6.19-rc ASAP ...

- Dave


> -andrey
>
> > Alan Stern
> >
> >
> > Index: 2.6.19-rc5/drivers/usb/host/ohci-hub.c
> > ===================================================================
> > --- 2.6.19-rc5.orig/drivers/usb/host/ohci-hub.c
> > +++ 2.6.19-rc5/drivers/usb/host/ohci-hub.c
> > @@ -422,7 +422,8 @@ ohci_hub_status_data (struct usb_hcd *hc
> > ohci->autostop = 0;
> > ohci->next_statechange = jiffies +
> > STATECHANGE_DELAY;
> > - } else if (time_after_eq (jiffies,
> > + } else if (device_may_wakeup(&hcd->self.root_hub->dev)
> > + && time_after_eq(jiffies,
> > ohci->next_statechange)
> > && !ohci->ed_rm_list
> > && !(ohci->hc_control &
>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo DeleteThis @vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

On Monday 13 November 2006 9:15 am, Alan Stern wrote:
> On Mon, 13 Nov 2006, David Brownell wrote:
>
> > It's a *driver model* API, which is also accessible from sysfs ... to support
> > per-device policies, for example the (a) workaround. The mechanism exists
> > even on kernels that don't include sysfs ... although on such systems, there
> > is no way for users to do things like say "ignore the fact that this mouse
> > claims to issue wakeup events, its descriptors lie".
>
> Yes, it is separate from sysfs -- but it is _tied_ to the sysfs API.

I can't agree. If you deconfigure sysfs, it still works.
Since it's independent like that, there's no way it's "tied".


> > > and therefore administrative
> > > in nature, but now you say it's also being used to record hardware quirks.
> >
> > No; I'm saying the driver model is used to record that the hardware mechanism
> > isn't available. The fact that it's because of an implementation artifact
> > (bad silicon, or board layout, etc) versus a design artifact (silicon designed
> > without that feature) is immaterial ... in either case, the system can't use
> > the mechanism.
>
> But the information is being recorded in the wrong spot. The correct test
> should use device_can_wakeup, not device_may_wakeup. The can_wakeup flag
> is the one which records whether or not the hardware mechanism is actually
> available.

Go look again. "may" implies (i) can , and (ii) should. So if there's a
hardware quirk registered, (i) always fails. And in the not-uncommon case
where the device misbehavior isn't known to the kernel, userspace has the
option of making (ii) kick in (the workaround mentioned above). This is a
generic approach, it works on all wakeup-capable devices.

So "may" is correct, and "can" is insufficient.



> Okay. I'll write a patch to eliminate autostop and those routines when
> CONFIG_PM is off.
>
> But that doesn't answer the question above: Should autostop check
> device_can_wakeup rather than device_may_wakeup?

See above, and the definition of may_wakeup().


> Also: Does the quirk/bug detection logic clear can_wakeup, as it should?
> Or does it only affect may_wakeup?

See above. Quirks directly recognized by the kernel clear can_wakeup.
Ones that are reported via userspace clear should_wakeup. Either suffices
to ensure that the may_wakeup() predicate fails.

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