pcmcia usb high-speed fails/root on usb device so can't rmmod ehci_hcd
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux (Ubuntu) |
Expired
|
Undecided
|
Unassigned |
Bug Description
This is similar to:
#286575 usb-automount not working - workaround "modprobe -r ehci_hcd"
https:/
which became:
http://
BUT it's on an i386 with root on a usb flash and affects usb ports on an SIIG JU-PCM422-S2 PCMCIA 4 Port USB 2.0 adapter; other pcmcia usb adapters give similar results on this machine, so I tend to suspect the Cardbus Controller.
I got the SIIG to sort of work at high speed on an HP zt3000, but on the machine it's intended for, Compaq V2552US,
there's something about the way it works with the CardBus bridge (Texas Instruments PCIxx21/x515 Cardbus Controller)
that prevents it from working at high speed.
Searches suggested that rmmod ehci_hcd resolves this, but with root on a usb device, it's not an option.
rmmod ohci_hcd followed by reinserting a usb device to the SIIG doesn't let ehci_hcd work, but the device does get
recognized immediately as "full speed" if modprobe ohci_hcd is then given.
I've tried booting with
"pci=assign-busses modprobe=
Boot parameters haven't changed anything other than the bus assignment to:
Bus: primary=00, secondary=01, subordinate=01, sec-latency=64
Bus: primary=00, secondary=02, subordinate=06, sec-latency=64
Bus: primary=02, secondary=03, subordinate=06, sec-latency=176
but this didn't fix it.
This is all with the pcmcia adapter plugged in, root on a kingston usb-drive, and another kingston usb-drive plugged in to the SIIG pcmcia USB hub JU-PCM422-S2.
After the attachment was assembled, I tried adding
"options scsi_mod inq_timeout=20"
to /etc/modprobe.
and updating the initrd accordingly, and rebooting with:
pci=assign-busses module yenta_socket.
yielding the only apparent difference:
Jan 22 17:29:59 xprime kernel: [ 63.730952] yenta_socket: Unknown symbol pccard_
Jan 22 17:29:59 xprime kernel: [ 63.736608] yenta_socket: Unknown symbol pccard_
reinserting still yields the usual:
[ 1128.904071] usb 6-4: device descriptor read/64, error -110
[ 1159.448301] usb 6-4: device descriptor read/64, error -110
[ 1159.664066] usb 6-4: reset high speed USB device using ehci_hcd and address 7
[ 1170.072070] usb 6-4: device not accepting address 7, error -110
[ 1170.184080] usb 6-4: reset high speed USB device using ehci_hcd and address 7
[ 1180.592060] usb 6-4: device not accepting address 7, error -110
[ 1180.593818] usb 6-4: USB disconnect, address 7
[ 1180.598271] scsi 5:0:0:0: Device offlined - not ready after error recovery
[ 1180.708053] usb 6-4: new high speed USB device using ehci_hcd and address 8
[...etc...]
and eventual recognition as "full speed"
attached are:
var/log/messages lspci-vvmm lsusb proc/ioports dmidecode lsusb-v lspci-kv proc/version lsb_release-rd
Tried a significantly modified different distribution with a 2.6.27.4 kernel I compiled, (but not
above module parameters)
Same sort of results including "error -110" and "full speed" operation after failure to assign address at high speed.