[wdmaudiodev] Re: Msvad stalling after resume from sleep

  • From: "Jolly, DavidX M" <davidx.m.jolly@xxxxxxxxx>
  • To: ambrish dantrey <a4ambrish@xxxxxxxxx>, "wdmaudiodev@xxxxxxxxxxxxx" <wdmaudiodev@xxxxxxxxxxxxx>
  • Date: Thu, 28 Mar 2013 19:01:29 +0000

Hi Ambrish,

Thanks for the quick response.

I have implemented the IPowerNotify interface and attempted to pause the stream 
manually by calling SetState(KSSTATE_PAUSE). However, I'm still seeing the same 
problem. I was under the impression that the kernel automatically handled 
streams by placing them into a paused state prior to sleeping. Is this not true?



From: ambrish dantrey [mailto:a4ambrish@xxxxxxxxx]
Sent: Thursday, March 28, 2013 11:43 AM
To: wdmaudiodev@xxxxxxxxxxxxx; Jolly, DavidX M
Subject: Re: [wdmaudiodev] Msvad stalling after resume from sleep

You may want to consider implementing IPowerNotify for your miniport driver:


This way you wouldn't have to rely on application to PAUSE the stream before 
going to sleep. You can do this in driver itself.

On Thu, Mar 28, 2013 at 9:26 PM, Jolly, DavidX M 
<davidx.m.jolly@xxxxxxxxx<mailto:davidx.m.jolly@xxxxxxxxx>> wrote:

I have implemented a virtual audio driver, based on the MSVAD simple example in 
the SDK.

Everything seems to be working well, unless I go into a sleep state while 
streaming audio in any player other than WMP. In this case, when the computer 
resumes from sleep, the audio panel volume control will be frozen, and the 
entire driver appears to be stalled. The computer has to be restarted in order 
to fix this problem.

After looking into the problem more, I'm seeing that the computer is going to 
sleep without first placing the stream into a pause or stop state. This causes 
the stream to resume from sleep in a running state. The reason WMP doesn't 
cause this issue is because it proactively pauses and stops the stream. 
Furthermore, in this bad state, the computer does attempt to pause the stream 
immediately after we wake up. For example, a good sleep transition will look 
like this: RUNNING -> PAUSE -> [SLEEP/WAKE] -> PAUSE -> RUNNING; while a bad 
sleep will look like this: RUNNING -> [SLEEP/WAKE] -> RUNNING -> PAUSE (from 
before the sleep) -> FROZEN.

It seems like either the portcls or the miniport are preventing the OS from 
setting the proper stream state. I've tried everything I can think of to 
resolve this issue, with no luck. If anyone could lend me some advice, I'd 
greatly appreciate it.



Other related posts: