Rquestiing a power IRP

Is there a way to let the PNP subsystem know that we need a device to power
up or power down if we are not the power policy owner? We need to make sure
that a device is powered up before we ask the device something. There are
multiple API's to tell the OS that we need a specific system power state.
But I do not see any way to notify a specific device. Sending a power IRP
directly to the device will result in a WDF bugcheck when the verifier is
running. The better approach would be to communicate our requirements to
the PNP subsystem and let the OS do what it knows best.

Regards,

Rossetoecioccolato.

"RossettoeCioccolato" <> wrote in message
news:eg3BGoV$...
> Is there a way to let the PNP subsystem know that we need a device to
> power up or power down if we are not the power policy owner? We need to
> make sure that a device is powered up before we ask the device something.
> There are multiple API's to tell the OS that we need a specific system
> power state. But I do not see any way to notify a specific device.
> Sending a power IRP directly to the device will result in a WDF bugcheck
> when the verifier is running. The better approach would be to communicate
> our requirements to the PNP subsystem and let the OS do what it knows
> best.
>
> Regards,
>
> Rossetoecioccolato.

In other words, you ask why there is a PPO?
Everyone should be allowed to manage power of any device?

-- pa

Pavel,

> In other words, you ask why there is a PPO?
> Everyone should be allowed to manage power of any device?
>

My question does not even remotely suggest that. If I ask the PPO owner,
either directly or through the PnP subsystem, to power up the device and the
PPO owner determines that it is able to comply, how does that violate PPO
ownership.

I must say that this is the least helpful of all the Microsoft newsgroups.
Responses to posts, if they occur at all, rarely contain any useful
information and almost always display a high level of attitude. What is the
justification for this lack of professionalism?

Regards,

Rossetoecioccolato.

just send the io request through the PPO driver. the PPO driver will then
power on the device before sending the request down the stack/to the
hardware. if you are sitting below the PPO in the stack, you are a little
stuck because you probably are sending io requests that should not go
through the PPO

d

"RossettoeCioccolato" wrote in message
news:uTJb$eb$...

Pavel,

> In other words, you ask why there is a PPO?
> Everyone should be allowed to manage power of any device?
>

My question does not even remotely suggest that. If I ask the PPO owner,
either directly or through the PnP subsystem, to power up the device and the
PPO owner determines that it is able to comply, how does that violate PPO
ownership.

I must say that this is the least helpful of all the Microsoft newsgroups.
Responses to posts, if they occur at all, rarely contain any useful
information and almost always display a high level of attitude. What is the
justification for this lack of professionalism?

Regards,

Rossetoecioccolato.

Doron,

Thanks. I can just rely on the PPO to do the right thing then.

Regards,

Rossetoecioccolato.

"Doron Holan [MSFT]" <> wrote in message
news:eFvcv%23e$...
> just send the io request through the PPO driver. the PPO driver will then
> power on the device before sending the request down the stack/to the
> hardware. if you are sitting below the PPO in the stack, you are a little
> stuck because you probably are sending io requests that should not go
> through the PPO
>
> d
>

> Is there a way to let the PNP subsystem know that we need a device to power
> up or power down if we are not the power policy owner?

No.

>We need to make sure that a device is powered up before we ask the device something.

Then you must be a PPO.

Why is it hard? For most devices, the functional driver is also a PPO.

--
Maxim S. Shatskih
Windows DDK MVP

http://www.storagecraft.com

"Maxim S. Shatskih" <> wrote in message
news:udYcE4m$...

> Then you must be a PPO.
>
> Why is it hard? For most devices, the functional driver is also a PPO.

In many interesting cases, it isn't.
Example: the PPO for NDIS miniports is the ndis.sys.
You want to put your USB netcard to sleep (SS), but cannot persuade ndis.sys
to issue the power request.

A solution proposed by Doron H. was to set a very short idle timeout.

Wake up usually is trivial; when a request comes from top of the stack, the
PPO just does the right thing.
But again, the miniport or device may want to wake up for some internal
reason, without ndis knows it...

Regards,
-- pa