Re: xenhid power state transitions

[Troy Crosley <troycrosley@xxxxxxxxx>]

I found that those ioctls weren't getting called in my testing. I implemented the power state transition boilerplate found in the other drivers to get xenhid to release and re-acquire its interfaces for S3. I'll make a PR in case there's any interest.

On Fri, Oct 9, 2020 at 11:18 AM Paul Durrant <xadimgnik@xxxxxxxxx> wrote:

I’m no hid expert but all the interfaces should be released on transition to D3. It may well be that power transitions are handled by the class driver. I don’t see handling of IOCTL_HID_ACTIVATE_DEVICE or IOCTL_HID_DEACTIVATE_DEVICE in FdoDispatchInternal() so that may be all that is necessary.

 

  Paul

 

From: win-pv-devel <win-pv-devel-bounces@xxxxxxxxxxxxxxxxxxxx> On Behalf Of Troy Crosley
Sent: 09 October 2020 16:03
To: win-pv-devel@xxxxxxxxxxxxxxxxxxxx
Subject: xenhid power state transitions

 

Hi all. I noticed that xenhid doesn't have the same power state transition boilerplate as the rest of the drivers. As a result, xenhid doesn't release its interfaces when the system is entering S3, causing an error when xenbus tries to enter D3 due to interfaces still being unexpectedly active. Should xenhid have the same power state transition boilerplate as the other drivers? Is there any consideration to be made due to its use of the hid class driver?

Regards,
Troy Crosley