Hi Roman,

We may have to back it out of the next patch-release
since I just sent it to Linus.

Are you returning to linux-usb contributory status?
Can you and David work thru this together?


> I don't think this is the right way to handle suspend/resume.
> So, I don't like that patch.
> 
> As far as I can remember there was a patch made by Alan which
> disables the HC-Interrupt and solves that problem, right.
> (But Linus don't like it that way.)

It could be a shared interrupt that was being disabled/enabled
at the interrupt controller.  Linus wants this handled by power
management, not in each driver, IIRC.

> Now, why we just disable the interrupts by setting the MIE-Bit in 
> the HcInterruptDisable Register on suspend and activate the interrupts
> by setting the MIE-Bit in the HcInterruptEnable Register on 
> resume, again?
> 
> 
> Roman
> 
> 
> 
> David Brownell wrote:
> > 
> > Here's the patch I mentioned earlier, which avoids the Oops by
> > making suspend/resume work again.  Against pre4-2 with the patch
> > I submitted for fewer magic #s and better debugging (that matters).
> > 
> > For me, it's the difference between minor or major 
> reinitialization of
> > the USB subsystem over suspend/resume, and just having it work.
> > 
> > The notable fixes:
> > 
> >         - On power management suspend, disconnect all devices on
> >           the controller's bus.  Controller left in reset state.
> > 
> >         - On power management resume, reset and restart the 
> controller.
> >           Otherwise it didn't seem to want to restart.
> > 
> > There are a few other changes in here, and any relating to 
> whitespace
> > or debugging messages could of course be discarded ... 
> though I'd leave
> > the change splitting the "#endif break;" line into two, 
> unless someone
> > believes the code could really have been right as-was.
> > 
> > I'd like to know how this works for other non-PMAC OHCI laptops.
> > (PMAC has its own suspend/resume code, untouched by this patch.)
> > 
> > - Dave


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to