How to interpret PM_TRACE output

How to interpret PM_TRACE output

[Tino Keitel]

Hi folks,

I tried PM_TRACE to find the driver that breaks resume from suspend.
I got working resume until I switched to the sk98lin driver
(because sky2 doesn't support wake on LAN). That's why I was quite sure that
sk98lin is the culprit, but I tried PM_TRACE anymay.

Here is the PM_TRACE output in dmesg:

  Magic number: 0:150:255
  hash matches drivers/base/power/resume.c:28
  hash matches device 0000:00:1d.3

$ lspci | grep 1d.3
00:1d.3 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #4

/proc/interrupts:

 17:      52387          0   IO-APIC-level  uhci_hcd:usb5, eth0, i915@pci:0000:00:02.0
 20:    1223105    1222776   IO-APIC-level  ehci_hcd:usb1, uhci_hcd:usb2

Since UHCI #4 (usb5, as ehci is usb1) and eth0 (sk98lin) use the same
interrupt, is it right to assume that the sk98lin driver does bad
interrupt handling and therefore breaks the usb5 device on resume?

Regards,
Tino

Re: How to interpret PM_TRACE output

[Pavel Machek]

On Wed 13-12-06 22:22:59, Tino Keitel wrote:
> Hi folks,
> 
> I tried PM_TRACE to find the driver that breaks resume from suspend.
> I got working resume until I switched to the sk98lin driver
> (because sky2 doesn't support wake on LAN). That's why I was quite sure that
> sk98lin is the culprit, but I tried PM_TRACE anymay.

See Doc*/power/*.
							Pavel
-- 
Thanks for all the (sleeping) penguins.

Re: How to interpret PM_TRACE output

[Tino Keitel]

On Sat, Dec 16, 2006 at 08:57:48 +0000, Pavel Machek wrote:
> On Wed 13-12-06 22:22:59, Tino Keitel wrote:
> > Hi folks,
> > 
> > I tried PM_TRACE to find the driver that breaks resume from suspend.
> > I got working resume until I switched to the sk98lin driver
> > (because sky2 doesn't support wake on LAN). That's why I was quite sure that
> > sk98lin is the culprit, but I tried PM_TRACE anymay.
> 
> See Doc*/power/*.

There is a nice mixture of documentation about swusp, video stuff,
developer documentation, and one short paragraph about PM_TRACE that
tells me nothing new. Could you point me to the documentation part that
you are referring to, and that tells me what to do if PM_TRACE shows
the usb device but the failure only occurs when I load the sk98lin
driver?

Thanks and regards,
Tino

Re: How to interpret PM_TRACE output

[Pavel Machek]

Hi!

> > > I tried PM_TRACE to find the driver that breaks resume from suspend.
> > > I got working resume until I switched to the sk98lin driver
> > > (because sky2 doesn't support wake on LAN). That's why I was quite sure that
> > > sk98lin is the culprit, but I tried PM_TRACE anymay.
> > 
> > See Doc*/power/*.
> 
> There is a nice mixture of documentation about swusp, video stuff,
> developer documentation, and one short paragraph about PM_TRACE that
> tells me nothing new. Could you point me to the documentation part that
> you are referring to, and that tells me what to do if PM_TRACE shows
> the usb device but the failure only occurs when I load the sk98lin
> driver?

Hmmm, so it fails somewhere in usb only if sk98lin is loaded? If you
unload it again, resume works? Are usb interrupts shared? Where
exactly in the usb does it fail?
-- 
Thanks for all the (sleeping) penguins.

Re: How to interpret PM_TRACE output

[Tino Keitel]

On Wed, Dec 20, 2006 at 16:19:04 +0000, Pavel Machek wrote:
> Hi!
> 
> > > > I tried PM_TRACE to find the driver that breaks resume from suspend.
> > > > I got working resume until I switched to the sk98lin driver
> > > > (because sky2 doesn't support wake on LAN). That's why I was quite sure that
> > > > sk98lin is the culprit, but I tried PM_TRACE anymay.
> > > 
> > > See Doc*/power/*.
> > 
> > There is a nice mixture of documentation about swusp, video stuff,
> > developer documentation, and one short paragraph about PM_TRACE that
> > tells me nothing new. Could you point me to the documentation part that
> > you are referring to, and that tells me what to do if PM_TRACE shows
> > the usb device but the failure only occurs when I load the sk98lin
> > driver?
> 
> Hmmm, so it fails somewhere in usb only if sk98lin is loaded? If you
> unload it again, resume works? Are usb interrupts shared? Where

Yes, it works with sky2. Yes, the USB device that is reported to fail
by PM_TRACE shares the interrupt with eth0, which is sk98lin (see my
original posting in this thread).

> exactly in the usb does it fail?

I don't know, all I have is the PM_TRACE output.

Meanwhile, tried to remove uhci_hcd before suspend, and wakeup works
then. However, my DVB-T box is dead after resume (reloading the driver
doesn't work, only unplug/replug the device helps). It works with
suspend to disk, though.

Regards,
Tino

Re: How to interpret PM_TRACE output

[Pavel Machek]

Hi!

> > > developer documentation, and one short paragraph about PM_TRACE that
> > > tells me nothing new. Could you point me to the documentation part that
> > > you are referring to, and that tells me what to do if PM_TRACE shows
> > > the usb device but the failure only occurs when I load the sk98lin
> > > driver?
> > 
> > Hmmm, so it fails somewhere in usb only if sk98lin is loaded? If you
> > unload it again, resume works? Are usb interrupts shared? Where
> 
> Yes, it works with sky2. Yes, the USB device that is reported to fail
> by PM_TRACE shares the interrupt with eth0, which is sk98lin (see my
> original posting in this thread).
> 
> > exactly in the usb does it fail?
> 
> I don't know, all I have is the PM_TRACE output.
> 
> Meanwhile, tried to remove uhci_hcd before suspend, and wakeup works
> then.

Send a nice bug report to usb people, then...
							Pavel
-- 
Thanks for all the (sleeping) penguins.