> > When OHCI is in the "UsbSuspend" state, it shouldn't
> > be accessing memory. And that's the state it enters
> > when it gets the suspend request ... from ACPI, via PCI.
> > Right? Those PCI calls mean entering S3 or S4 states
> > it seems ... it's not clear to me whether there are enough
> > APIs to let device drivers (PCI or USB!) support C3,
> > but I might well have missed some important ones.
>
> C3 is a processor state, so device drivers needn't worry about it, except to
> realize that constant bus mastering is causing the CPU to not enter as deep
> a sleep as it potentially could. (IIRC a cpu in C2 takes more than half a
> watt more than a cpu in C3.)
There's a powerpoint presentation in the developer area of www.usb.org
from last October, talking about power management and USB. One
thing it says is that there ** IS ** a need for device drivers support for
entering the C3 state ... which makes sense, since the USB device drivers
are what's making the USB Host Controller do the constant mastering,
and the host controller itself would need a hint to become less active.
There's more to this problem than that, but it does seem like maybe some
API support is needed to handle C3 on Linux. (Pages 12-13 of that ppt
presentation.)
- Dave
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
