Re: suspend vs usb and PS/2 ports

[Rafael J. Wysocki]

On Thursday 01 October 2009, Michael Tokarev wrote:
> Rafael J. Wysocki wrote:
> > On Wednesday 30 September 2009, Michael Tokarev wrote:
> >> Hello.
> >>
> >> I've several PCs here which are able to wakeup -
> >> among others - from PS/2 and/or USB keyboard.
> >> Here for testing I'm using PS/2 keyboard and
> >> Asus M3A78-EM motherboard.
> >>
> >> When I do poweroff from linux, the keyboard
> >> stays powered up as it should (according to
> >> the BIOS settings).  But when I do suspend,
> >> keyboard is NOT powered anymore, and hence
> >> the system can't be woken up from it but only
> >> by using the power button.
> >>
> >> This happens consistently for many kernel
> >> versions.  To be fair, I don't even know if
> >> there was any kernel which does not show this
> >> behaviour: old versions was unable to do
> >> suspend/resume cycle on this platform at all.
> >>
> >> Any hints for this please?
> > 
> > With USB, I think keyboard wake-up is off by default as it causes some
> > systems to wake up immediately after suspending.  Alan and Oliver can provide
> > more info about that.
> 
> Actually I wasn't able to get any machine to resume based on
> USB devices - notable keyboard, be it sleep/power button (on
> the keyboards that have it) or any other combination of keys
> configured in the BIOS.  Most motherboards I tried are from
> Asus, one or two from Gigabyte and one more from Biostar.
> 
> That's why I use PS/2=>USB adaptor, to connect an USB keyboard
> to a PS/2 port - this one works just fine, with all the mobos
> I tried (provided the appropriate BIOS settings are turned on
> and ever exists, to start with).
> 
> > I don't know about PS/2.
> > 
> > Can you post the contents of /proc/acpi/wakeup from one of these systems,
> > please?
> 
> Hm.  That's.. curious.  Here we go, my home machine whihc I turn
> on every day from a PS/2-connected keyboard:
> 
> Device	S-state	  Status   Sysfs node
> PCE2	  S4	 disabled
> PCE3	  S4	 disabled
> PCE4	  S4	 disabled
> PCE5	  S4	 disabled
> PCE6	  S4	 disabled  pci:0000:00:06.0
> RLAN	  S4	 disabled  pci:0000:02:00.0
> PCE7	  S4	 disabled
> PCE9	  S4	 disabled
> PCEA	  S4	 disabled
> PCEB	  S4	 disabled
> PCEC	  S4	 disabled
> SBAZ	  S4	 disabled  pci:0000:00:14.2
> PS2K	  S4	 disabled  pnp:00:09
> UAR1	  S4	 disabled  pnp:00:0a
> P0PC	  S4	 disabled  pci:0000:00:14.4
> UHC1	  S4	 disabled  pci:0000:00:12.0
> UHC2	  S4	 disabled  pci:0000:00:12.1
> UHC3	  S4	 disabled  pci:0000:00:12.2
> USB4	  S4	 disabled  pci:0000:00:13.0
> UHC5	  S4	 disabled  pci:0000:00:13.1
> UHC6	  S4	 disabled  pci:0000:00:13.2
> UHC7	  S4	 disabled  pci:0000:00:14.5
> 
> It's all disabled!

Thanks.  In fact, the 'disabled' column doesn't really mean totally disabled.
It means one of the mechanisms one can use is disabled.

For PCI, 'disabled' in this column shouldn't really matter any more. 

> But I *know* it wakes up
> from network and from ps/2 keyboard at least!
> In the corresponding BIOS menu almost everything
> is enabled (excluding RTC alarm, but including
> USB events).  This is an Asus motherboard, M3A-78EM,
> with latest BIOS.
> 
> Here's another one, Gigabyte's GA-MA74GM-S2H,
> which definitely wakes on network:
> 
> Device	S-state	  Status   Sysfs node
> PCI0	  S5	 disabled  no-bus:pci0000:00
> USB0	  S3	 disabled  pci:0000:00:12.0
> USB1	  S3	 disabled  pci:0000:00:12.1
> USB2	  S3	 disabled  pci:0000:00:12.2
> USB3	  S3	 disabled  pci:0000:00:13.0
> USB4	  S3	 disabled  pci:0000:00:13.1
> USB5	  S3	 disabled  pci:0000:00:13.2
> USB6	  S3	 disabled  pci:0000:00:14.5
> SBAZ	  S4	 disabled  pci:0000:00:14.2
> P2P	  S5	 disabled  pci:0000:00:14.4
> PCE2	  S4	 disabled
> PCE3	  S4	 disabled
> PCE4	  S4	 disabled
> PCE5	  S4	 disabled
> PCE6	  S4	 disabled  pci:0000:00:06.0
> PCE7	  S4	 disabled
> PCE8	  S4	 disabled
> 
> hmm.  Blaming BIOS, as usual? :)

Nope.  For the majority of devices we disable wake-up by default.

> But even if that's the case, why there's a difference
> between "just" power-off and power-off when suspending?

These two cases are handled differently by BIOSes.

Can you please check if /sys/devices/.../power/wakeup contains
'enabled' for the PCI devices called USB0 - USB6 above?

> The kernel is 2.6.31 vanilla.

Better use 2.6.32-rc for debugging this.

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