On 8/15/07, Pete Zaitcev <zaitcev@xxxxxxxxxx> wrote: > On Tue, 14 Aug 2007 19:52:39 -0600, Lamont Peterson <lamont@xxxxxxxxxxxx> wrote: > > > I've wondered for a long time why the ehci driver wasn't first. I've seen > > many a system (especially notebooks) where the USB 2.0 ports were running as > > USB 1.1. I hope there's no good reasons to stop this change from going in. > > In theory there should be no issues with loading ehci_hcd after its > companion (either uhci_hcd on Intel and VIA or ohci_hcd elsewhere). > The problem I am trying to address is how the discovery and initialization > continues on the companion bus when we start initializing EHCI in parallel, > which happens because khubd continues to run when modprobe has exited. > This effects a sudden disconnect as EHCI takes over the port; if the > takeover fails (usually because the device does not support High-Speed) > EHCI returns the port and the interrupted initialization repeats. > > If EHCI fails to take over a root hub port despite a device being High-Speed > capable, it is a bug somewhere and initialization order is not likely > to fix it (although it may, mostly due to interactions with firmware). > File a bug or let me know directly. > This related to http://bugzilla.kernel.org/show_bug.cgi?id=8535 (USB freeze/system death)? tom -- Tom London -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
