On Fri, 3 Apr 2009, guoqing zhu - Sun Microsystems - Beijing China wrote:
>> Illegal Request: 0 Predictive Failure Analysis: 0
>>
>> Could it be the PCI card with the USB connectors that it the problem?
>> the ehci driver?
> I think it should be some hardware problem on the pci USB card. You can
> try to connect some USB1.1 devices to see if they can work.
>
> Guoqing
> Thanks
Sorry for the delayed answer...
I am starting to suspect the PCI USB controller to be the source of the
problem.
prtconf -pv output reveals the following ids:
USB vendor id: 10b9
Device id: three ids with 5237
one id with 5239
Based on http://www.nvidia.com/page/uli_m5273.html this must be the device
I have in my system.
Based on this bugfix
(http://bugs.opensolaris.org/bugdatabase/view_bug.do;jsessionid=1cb7aca10cc59824f69b254df90ad?bug_id=5047712)
and with some minor ajustments I have been able to use the mass storage
device although not as an USB 2.0 hi speed device but as an USB 1.x device
at very low speed...
I changed the ehci_init_workaround() function (see
usr/src/uts/common/io/usb/hcd/ehci_util.c) so the special handling for
PCI_VENDOR_ALI is not active:
...
#if 0
if (ehcip->ehci_vendor_id == PCI_VENDOR_ALI) {
/* Route all Root hub ports to EHCI host controller */
Set_OpReg(ehci_config_flag, EHCI_CONFIG_FLAG_EHCI);
}
#endif
/*
...
This makes the usb mass storage device usable (I have done complete builds
of onnv on this slow usb mass storage device...).
The only annoying problem is the endless numbeer of messages in
/var/adm/messages:
WARNING: /pci at 1f,0/pci at 1/usb at 4,3 (ehci0): No SOF interrupts have
been
received, this USB EHCI hostcontroller is unusable
Regards
Palle
