On Thu, 23 Sep 2004, Oliver Neukum wrote:

> > So among all the other features, there needs to be some way to exclude 
> > some set of devices from a system-wide suspend.
> 
> How? If you cut power, you'll cut all power to bus powered devices.
> Other devices will potentially go out of sync. This seems to me a bad
> idea.

Sorry, I didn't express myself very well.  What I meant was, there needs 
to be a way to exempt some devices (such as the swap device) from an 
initial pass during system-wide suspend, then suspend them during a second 
pass (after the memory image has been written).

In fact, the existing PM code already has a second pass.  It's used for
suspending devices which require interrupts to be disabled when they
aren't fully powered.

> IMHO a simple model of suspend() followed by resume() is not enough.
> The order of calls is constrained by the device tree. So you'll switch
> off your swap partition at some point before all devices have saved
> state.

Does the text above answer this point?  It remains true of course that the
swap partition will be shut down while the contents of memory are still in
flux.  That shouldn't matter provided the resume procedure guarantees that
when the memory image is read back, the system/device state is the same as
when the image was written.

Alan Stern



-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to