> That behavior was seen on the Synopsys host, not the Intel host, > correct? Yes. Looks like the L1 transitions that are not fatal on the Intel host are much longer in my trace, usually above 100ms. This would be another indication that in the Synopsys case the L1 resume is host-triggered. > Ok. Mathias has a patch to enable it for internal devices and BESL > devices. I'll send out the updated patchset shortly. Can you run lsusb > on your Intel system, and see if the webcam supports Link PM? If so, it > would be great if you could add the patches, disable auto-suspend, and > double check that the latest version of lsusb shows that the device goes > into L1. I don't have access to a system today or tomorrow, but I can > check next week if necessary. Unfortunately, none of the internal devices on our current Haswell system support LPM. Our firmware also doesn't set the right ACPI table entries to mark ports as removable. I think the easiest way to verify your patch would be a small kernel hack that would force an inserted device to look non-removable. -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
