xhci_hcd: Runtime PM causes port to inactivate

xhci_hcd: Runtime PM causes port to inactivate

[Patrik Kullman]

Issuing

echo auto > /sys/bus/pci/devices/0000:03:00.0/power/control

on my Asus Zenbook UX31E with a

03:00.0 USB Controller: Fresco Logic Device 1009 (rev 02)

causes its USB3-port to become mostly disabled.

It does have a "Charger+" feature that allows it to charge cellphones
etc even when turned off, and this feature works when power control is
in "auto" mode.

In all other aspects, the port is turned off. It doesn't react to
devices being inserted.

This does not feel like the desired behavior.

Also, suspend does not work without unbinding the xhci_hcd module, not
sure if that is related.
Computer will not suspend/resume unless the following script is used
in /etc/pm/sleep.d:
https://help.ubuntu.com/community/AsusZenbook#Suspend.2C_Hibernate.2C_Shut-down_and_Reboot

This behavior has not changed between 3.0 and 3.2-rc4+ that I've tried today.

Re: xhci_hcd: Runtime PM causes port to inactivate

[Srivatsa S. Bhat]

There is a very similar discussion going on at:
http://thread.gmane.org/gmane.linux.kernel/1222803

Cc'ing all the relevant people and mailing lists.

Regards,
Srivatsa S. Bhat

On 12/04/2011 04:47 PM, Patrik Kullman wrote:

> Issuing
> 
> echo auto > /sys/bus/pci/devices/0000:03:00.0/power/control
> 
> on my Asus Zenbook UX31E with a
> 
> 03:00.0 USB Controller: Fresco Logic Device 1009 (rev 02)
> 
> causes its USB3-port to become mostly disabled.
> 
> It does have a "Charger+" feature that allows it to charge cellphones
> etc even when turned off, and this feature works when power control is
> in "auto" mode.
> 
> In all other aspects, the port is turned off. It doesn't react to
> devices being inserted.
> 
> This does not feel like the desired behavior.
> 
> Also, suspend does not work without unbinding the xhci_hcd module, not
> sure if that is related.
> Computer will not suspend/resume unless the following script is used
> in /etc/pm/sleep.d:
> https://help.ubuntu.com/community/AsusZenbook#Suspend.2C_Hibernate.2C_Shut-down_and_Reboot
> 
> This behavior has not changed between 3.0 and 3.2-rc4+ that I've tried today.

Re: xhci_hcd: Runtime PM causes port to inactivate

[Patrik Kullman]

Actually, I've been digging deeper into the suspend-issue, and the
xhci_hcd module is not to blame in my case. Running 3.2-rc4+, I only
need to use the parts of the script that unbinds/binds the ehci_hcd
module.

However, the original case of setting
/sys/bus/pci/devices/0000:03:00.0/power/control to "auto" causing the
USB-port to be inactivated is still valid.

2011/12/4 Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>:
> There is a very similar discussion going on at:
> http://thread.gmane.org/gmane.linux.kernel/1222803
>
> Cc'ing all the relevant people and mailing lists.
>
> Regards,
> Srivatsa S. Bhat
>
> On 12/04/2011 04:47 PM, Patrik Kullman wrote:
>
>> Issuing
>>
>> echo auto > /sys/bus/pci/devices/0000:03:00.0/power/control
>>
>> on my Asus Zenbook UX31E with a
>>
>> 03:00.0 USB Controller: Fresco Logic Device 1009 (rev 02)
>>
>> causes its USB3-port to become mostly disabled.
>>
>> It does have a "Charger+" feature that allows it to charge cellphones
>> etc even when turned off, and this feature works when power control is
>> in "auto" mode.
>>
>> In all other aspects, the port is turned off. It doesn't react to
>> devices being inserted.
>>
>> This does not feel like the desired behavior.
>>
>> Also, suspend does not work without unbinding the xhci_hcd module, not
>> sure if that is related.
>> Computer will not suspend/resume unless the following script is used
>> in /etc/pm/sleep.d:
>> https://help.ubuntu.com/community/AsusZenbook#Suspend.2C_Hibernate.2C_Shut-down_and_Reboot
>>
>> This behavior has not changed between 3.0 and 3.2-rc4+ that I've tried today.