On Thu, 28 Apr 2005, Oliver Neukum wrote: > > > Please consider scsi. It has no idea about what is going on. > > > > In principle this shouldn't matter. If a device is autosuspended then it > > should autoresume whenever a new request (such as a new SCSI command) > > arrives. > > Provided that you know when you can safely autosuspend. And provided > you can restore the full state. I doubt you can with scsi because the driver > doesn't understand the commands.
Oliver, I don't understand your point. Which devices and drivers are you referring to? "Provided that you know when you can safely autosuspend." If the driver doesn't know when it can suspend its device, who else would know? "And provided you can restore the full state." Well yes, clearly, if suspending a device loses some state unavoidably then the driver should never autosuspend. Not unless someone tells it that it's okay to go ahead and lose the state. "I doubt you can with scsi because the driver doesn't understand the commands." Well, for example, usb-storage is a SCSI driver of sorts, and it certainly understands suspend/resume. (Or more accurately, it can be made to understand with a relatively small patch.) I'm not sure what the state of suspend/resume support is in drivers like sd or sr, but that shouldn't make much difference to usb-storage. It doesn't have to tell the SCSI layer when the USB transport is suspended. All it has to do is resume the USB connection when the next SCSI command comes along. Alan Stern ------------------------------------------------------- SF.Net email is sponsored by: Tell us your software development plans! Take this survey and enter to win a one-year sub to SourceForge.net Plus IDC's 2005 look-ahead and a copy of this survey Click here to start! http://www.idcswdc.com/cgi-bin/survey?id=105hix _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
