It looks as if this issue is related to another USB speed issue, although not caused by or solved by it - bug #296710 "warning: ehci_hcd loaded AFTER uhci_hcd and ohci_hcd"
However, in this particular case it looks rather that the particular system hardware is getting confused and may be related to particular motherboard chipset/port combinations. I'll explain a little as to how the hardware is organised so it is easier to understand. The kernel's USB modules deal with each version of the USB protocol separately. The original USB v1.x low-speed and high-speed (12Mbs) is handled by uhci_hcd. USB v2 is handled by ehci_hcd (480Mbps). The physical hardware of a PC has separate controller devices for each USB version. The PCI [vendor:model] ID determines which kernel module manages a device. The physical external ports of a PC, when they support USB2 *and* USB v1.x devices, are usually *shared* by the host controllers. When the PC powers up (or resets) the first host controller to be initialised (by a kernel module) *claims* ports. Once claimed a port cannot be used by another host controller until a power reset. The kernel modules have no way to control which host controller claims ports. >From the various permutations of tests that have been done so far by users it looks as if this process of claiming external ports partially explains the symptoms, and may help us find a solution. The key thing seems to be that all the reports seem to agree that this *just worked* on Gutsy for the reason that the ports were being handled by the USB v1.x (12Mbps) host controller on the orders of uhci_hcd. In the later kernel releases the ports are being initialised by the USB v2 (480Mbps) host controller on the orders of ehci_hcd. It is possible that this is specific to particular hardware - maybe motherboard make/model, possibly the conjunction of particular pairs of USB host controllers. So, now the explanation is over, what we need is to gather that information to help look for commonalities. Please attach the reports from the following commands as single archives to make it easier to analyse. ---------- sudo -i [ ! -x "$(which dmidecode)" ] && apt-get install dmidecode mkdir /tmp/reports uname -a >/tmp/reports/uname.log lsb_release -a >/tmp/reports/lsb_release.log dmidecode -t system >/tmp/reports/dmidecode.log dmidecode -t baseboard >>/tmp/reports/dmidecode.log lspci --vvnn >/tmp/reports/lspci-vvnn.log lsusb -v >/tmp/reports/lsusb-v.log dmesg >/tmp/reports/dmesg.log tar -czf /tmp/system-report.tar.gz /tmp/reports/* exit ---------- Now attach the resulting file "/tmp/system-report.tar.gz" to this bug report. -- usb connect problems https://bugs.launchpad.net/bugs/231049 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
