Hi,
Josh Triplett <josh@xxxxxxxxxxxxxxxx> writes:
> On Mon, Dec 07, 2015 at 03:03:40PM -0600, Felipe Balbi wrote:
>> Josh Triplett <josh@xxxxxxxxxxxxxxxx> writes:
>> >> > Looking at the sysfs path for the input device as shown in dmesg,
>> >> > /sys/devices/pci0000\:00/0000\:00\:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2\:1.1/supports_autosuspend
>> >> > did say "1". However, the device didn't actually appear to have
>> >> > autosuspend *enabled*.
>> >> >
>> >> > I'd appreciate any suggestions for how to debug this issue.
>> >>
>> >> what does 'control' on the power directory show ?
>> > I don't have a file "control" in the power directory when plugged in through
>> > the dock. The contents of all the files under power when plugged in via dock:
>>
>> huh ???
>>
>> > ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-2/2-2:1.1/0003:17EF:6009.002D/power/control <==
>> > auto
>>
>> here's one enabling autosuspend
>>
>> > ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-2/2-2:1.1/0003:17EF:6009.002D/power/control <==
>> > auto
>>
>> and another, also enabling runtime pm. Change both of these from 'auto'
>> to 'on' and see if that helps.
>
> Copy/paste error; notice that both of those have the same path.
okay.
> The following files appear under the power directory when plugged in via
> the dock:
if those are not enabled, then the device doesn't have pm_runtime
callbacks, see:
int dpm_sysfs_add(struct device *dev)
{
int rc;
rc = sysfs_create_group(&dev->kobj, &pm_attr_group);
if (rc)
return rc;
if (pm_runtime_callbacks_present(dev)) {
rc = sysfs_merge_group(&dev->kobj, &pm_runtime_attr_group);
if (rc)
goto err_out;
}
if (device_can_wakeup(dev)) {
rc = sysfs_merge_group(&dev->kobj, &pm_wakeup_attr_group);
if (rc)
goto err_runtime;
}
if (dev->power.set_latency_tolerance) {
rc = sysfs_merge_group(&dev->kobj,
&pm_qos_latency_tolerance_attr_group);
if (rc)
goto err_wakeup;
}
return 0;
err_wakeup:
sysfs_unmerge_group(&dev->kobj, &pm_wakeup_attr_group);
err_runtime:
sysfs_unmerge_group(&dev->kobj, &pm_runtime_attr_group);
err_out:
sysfs_remove_group(&dev->kobj, &pm_attr_group);
return rc;
}
It seems like usbcore always calls pm_runtime_no_callbacks() (which
makes pm_runtime_callbacks_present() above fail) for all interfaces, but
I can't seem to find where that's undone. Alan, any hints ?
> ~$ head /sys/devices/pci0000:00/0000:00:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2:1.1/power/*
> ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2:1.1/power/async <==
> enabled
enabled
> ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2:1.1/power/runtime_active_kids <==
> 0
>
> ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2:1.1/power/runtime_enabled <==
> enabled
>
> ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2:1.1/power/runtime_status <==
> suspended
>
> ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-3/2-3.1/2-3.1.2/2-3.1.2:1.1/power/runtime_usage <==
> 0
>
>
> Notice that no file named "control" exists.
>
> By contrast, the following files appear under the power directory when
> plugged in directly:
>
> ~$ head /sys/devices/pci0000:00/0000:00:14.0/usb2/2-2/2-2:1.1/0003:17EF:6009.0032/power/*
> ==> /sys/devices/pci0000:00/0000:00:14.0/usb2/2-2/2-2:1.1/0003:17EF:6009.0032/power/async <==
> disabled
disabled :-)
see if setting the other case to disabled changes anything.
--
balbi
Attachment:
signature.asc
Description: PGP signature
