cannot resume using USB keyboard/mouse
When I suspend my desktop box, I cannot resume again by pressing a key on my USB keyboard or clicking a button on my USB mouse. I have to press the chassis power button to wake things up again. (I'd expect this to be necessary after hibernation, but not after suspend.) The same machine resumes from suspend using keyboard or mouse buttons under Vista, so I know this is something that the hardware is capable of.
Is this a bug? Or is there some local configuration tweak I should be making to tell the system to wake up in response to USB key presses & mouse clicks?
I'm running Fedora 8, with a Fedora-provided kernel version 2.6.23.1-49.fc8. Attached below is a pruned-down hardware description (from lshw) showing the mouse and keyboard, plus other buses and bridges by which they are connected. Are there any other useful details I can provide?
Thanks!
-- Ben
H/W path Device Class Description
===========================================================
system Nobilis /0 bus DG965RY /0/100 bridge 82P965/G965 Memory Controller Hub /0/100/1d bus 82801H (ICH8 Family) USB UHCI Controller #1 /0/100/1d/1 usb5 bus UHCI Host Controller /0/100/1d/1/2 bus TUSB2040/2070 Hub /0/100/1d/1/2/1 generic Microsoft Natural Keyboard Pro /0/100/1d/1/2/2 generic Microsoft IntelliMouse
Quoted 17 lines · On Saturday 24 November 2007 08:15:05 Ben Liblit wrote:
When I suspend my desktop box, I cannot resume again by pressing a key on my USB keyboard or clicking a button on my USB mouse. I have to press the chassis power button to wake things up again. (I'd expect this to be necessary after hibernation, but not after suspend.) The same machine resumes from suspend using keyboard or mouse buttons under Vista, so I know this is something that the hardware is capable of.
Is this a bug? Or is there some local configuration tweak I should be making to tell the system to wake up in response to USB key presses & mouse clicks?
I'm running Fedora 8, with a Fedora-provided kernel version 2.6.23.1-49.fc8. Attached below is a pruned-down hardware description (from lshw) showing the mouse and keyboard, plus other buses and bridges by which they are connected. Are there any other useful details I can provide?
Quoted 4 lines
Thanks!
-- Ben
You probably need to tell the system which devices can do the wakeup. To do so first do a "cat /proc/acpi/wakeup".
and then write corresponding device name (the first column) to that file (under root)
For example, mine /proc/acpi/wakeup:
maxim@MAIN:~$ cat /proc/acpi/wakeup Device S-state Status Sysfs node SLPB S4 *enabled P32 S4 enabled pci:0000:00:1e.0 UAR1 S4 enabled pnp:00:0a ILAN S4 enabled pci:0000:00:19.0 PEGP S4 enabled pci:0000:00:01.0 PEX0 S4 enabled pci:0000:00:1c.0 PEX1 S4 enabled pci:0000:00:1c.1 PEX2 S4 enabled pci:0000:00:1c.2 PEX3 S4 enabled pci:0000:00:1c.3 PEX4 S4 enabled pci:0000:00:1c.4 PEX5 S4 enabled UHC1 S3 enabled pci:0000:00:1d.0 UHC2 S3 enabled pci:0000:00:1d.1 UHC3 S3 enabled pci:0000:00:1d.2 UHC4 S3 enabled EHCI S3 enabled pci:0000:00:1d.7 EHC2 S3 enabled pci:0000:00:1a.7 UH42 S3 enabled pci:0000:00:1a.0 UHC5 S3 enabled pci:0000:00:1a.1 AZAL S3 enabled pci:0000:00:1b.0 maxim@MAIN:~$
the UHC1-5, EHCI and EHC2 are the usb1.0 resume devices (ports) so I can do
echo "UHC1" > /proc/acpi/wakeup
echo "UHC2" > /proc/acpi/wakeup
....
And then check whenever the "status" column" shows "enabled"
Note: some resume devices are connected together:
For example doing 'echo "ILAN" > /proc/acpi/wakeup' enables the LAN , EHCI , AZAL.
So my script that enables all wakeup devices is:
#! /bin/bash
echo "UHC1" > /proc/acpi/wakeup # ICH8 UHCI1
echo "UHC2" > /proc/acpi/wakeup # ICH8 UHCI2
echo "UHC3" > /proc/acpi/wakeup # ICH8 UHCI3
echo "UHC4" > /proc/acpi/wakeup # ICH8 UHCI4
echo "UHC5" > /proc/acpi/wakeup # ICH8 UHCI5
echo "ILAN" > /proc/acpi/wakeup # ICH8 internal (lan/HDA/EHCI)
echo "P32" > /proc/acpi/wakeup # ICH8 external (PCI/PCI express)
echo "UAR1" > /proc/acpi/wakeup # SUPERIO (UART/Sleep button)
And last thing, writing to /proc/acpi/wakeup a device name that is already enabled,
disables it, so if I for example run the above script twice (with wakeup devices disabled), all wakeup devices will be disabled again.
Best regards,
Maxim Levitsky
Quoted 7 lines · Maxim Levitsky
the UHC1-5, EHCI and EHC2 are the usb1.0 resume devices (ports) so I can do
echo "UHC1" >
/proc/acpi/wakeup
echo "UHC2" >/proc/acpi/wakeup....
Ah hah!
Echoing EHCI and EHC2 into that file has no affect on my system; they remain disabled. But enabling UHC[1-5] works, and my system now resumes on USB key presses and mouse clicks as desired.
Thank you very much for your help, Maxim!
Quoted 11 lines · On Saturday 24 November 2007 20:29:06 Ben Liblit wrote:
Quoted 7 lines · Maxim Levitsky
the UHC1-5, EHCI and EHC2 are the usb1.0 resume devices (ports) so I can do
echo "UHC1" >
/proc/acpi/wakeup
echo "UHC2" >/proc/acpi/wakeup....
Ah hah!
You are welcome.
Note that the names like UHC1 are system depedent,
those are names that are present in ACPI tables,
so you need to read this file first, and then guess which one is which
You can also check the "Sysfs node" column, and then using sysfs or lspci figure out the device function. But I guess that in your case the USB controllers have same "UHC[1-5]" name.
And btw since most(if not all) keyboard/mice/... are USB 1 devices, the EHCI setting won't work anyway.
Best regards,
Maxim Levitsky