[PATCH] hid: enable hid device to suspend/resume asynchronously

[PATCH] hid: enable hid device to suspend/resume asynchronously

[Fu, Zhonghui]

Enable hid device to suspend/resume asynchronously. This can improve
system suspend/resume speed.

Signed-off-by: Zhonghui Fu <zhonghui.fu@linux.intel.com>
---
 drivers/hid/hid-core.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c
index e6fce23..27ac160 100644
--- a/drivers/hid/hid-core.c
+++ b/drivers/hid/hid-core.c
@@ -2627,6 +2627,7 @@ struct hid_device *hid_allocate_device(void)
 	device_initialize(&hdev->dev);
 	hdev->dev.release = hid_device_release;
 	hdev->dev.bus = &hid_bus_type;
+	device_enable_async_suspend(&hdev->dev);
 
 	hid_close_report(hdev);
 
-- 1.7.1

Re: [PATCH] hid: enable hid device to suspend/resume asynchronously

[Fu, Zhonghui]

Hi,

Any comments are welcome.


Thanks,
Zhonghui



On 2015/8/17 23:55, Fu, Zhonghui wrote:
> Enable hid device to suspend/resume asynchronously. This can improve
> system suspend/resume speed.
>
> Signed-off-by: Zhonghui Fu <zhonghui.fu@linux.intel.com>
> ---
>  drivers/hid/hid-core.c |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c
> index e6fce23..27ac160 100644
> --- a/drivers/hid/hid-core.c
> +++ b/drivers/hid/hid-core.c
> @@ -2627,6 +2627,7 @@ struct hid_device *hid_allocate_device(void)
>  	device_initialize(&hdev->dev);
>  	hdev->dev.release = hid_device_release;
>  	hdev->dev.bus = &hid_bus_type;
> +	device_enable_async_suspend(&hdev->dev);
>  
>  	hid_close_report(hdev);
>  
> -- 1.7.1
>

Re: [PATCH] hid: enable hid device to suspend/resume asynchronously

[Jiri Kosina]

On Mon, 17 Aug 2015, Fu, Zhonghui wrote:

> Enable hid device to suspend/resume asynchronously. This can improve
> system suspend/resume speed.

How well was this tested?

Power management is notorious for not being really in excellent shape on 
many HID devices.

So I'd like to be careful.

Thanks,

-- 
Jiri Kosina
SUSE Labs

Re: [PATCH] hid: enable hid device to suspend/resume asynchronously

[Fu, Zhonghui]

On 2015/9/4 21:10, Jiri Kosina wrote:
> On Mon, 17 Aug 2015, Fu, Zhonghui wrote:
>
>> Enable hid device to suspend/resume asynchronously. This can improve
>> system suspend/resume speed.
> How well was this tested?
>
> Power management is notorious for not being really in excellent shape on 
> many HID devices.
>
> So I'd like to be careful.
I were on leave these days, so sorry for late reply.

This can reduce system suspend/resume time about 20ms, from 1030ms to 1010ms on ASUS T100TA machine. Although the improvement is not very significant, but this can let your parent/child device suspend/resume more asynchronously and take advantage of multicore to improve overall system suspend/resume speed.

I have resent this patch with updated commit message - "[PATCH v2] hid: enable hid device to suspend/resume asynchronously".


Thanks,
Zhonghui

>
> Thanks,
>

Re: [PATCH] hid: enable hid device to suspend/resume asynchronously

[Jiri Kosina]

On Thu, 24 Sep 2015, Fu, Zhonghui wrote:

> >> Enable hid device to suspend/resume asynchronously. This can improve
> >> system suspend/resume speed.
> > How well was this tested?
> >
> > Power management is notorious for not being really in excellent shape on 
> > many HID devices.
> >
> > So I'd like to be careful.
> I were on leave these days, so sorry for late reply.
> 
> This can reduce system suspend/resume time about 20ms, from 1030ms to 
> 1010ms on ASUS T100TA machine. Although the improvement is not very 
> significant, but this can let your parent/child device suspend/resume 
> more asynchronously and take advantage of multicore to improve overall 
> system suspend/resume speed.
> 
> I have resent this patch with updated commit message - "[PATCH v2] hid: 
> enable hid device to suspend/resume asynchronously".

Is there any reason why not enable this from userspace via sysfs instead, 
based on whitelist on known-to-be-well-behaved devices, from udev?

I am really careful when it comes to enabling features like this 
(especially given the state of PM support on some low-end devices) 
globally.

Thanks,

-- 
Jiri Kosina
SUSE Labs