> Now, back to OHCI. In 2.5.21, things are somewhat different > than in 2.5.20, and it fails even before the IrDA driver enter in > action. The full log is below. > I also tries 2.5.20 on the same box with a non-SMP kernel, and > that didn't make any difference. > Yes, I know about Opti bugs. It's just that it used to work...
I've spent some time chasing this on some hardware that previously worked fine with ohci-hcd ... no luck quite yet, but I thought I'd mention that this isn't being ignored. What I'm seeing feels like a silicon bug is getting triggered. The host controller does what it always did (even as "usb-ohci"): stick an empty ED onto the control queue (ed_controlhead), enable that queue (CLE), adds some TDs to that ED, then tells it the queue was filled (CLF). That sequence works fine on most OHCI hardware. Then the controller does nothing ... except that if I look carefully, I can see that it's actually generating a stream of schedule overrun events (illegal here, given the empty periodic schdule), suggesting that somehow it got so wedged (because of something that slighty resembles a periodic?) that it can't do real work. Here's where a PCI analyser trace would be a huge help... :) Such symptoms have been seen before with the usb-ohci driver, so it's likely that something has just made them more likely. Am I right to conclude that Christopher Hoover was able to get that very same card working in a different system? - Dave > ohci-pci.c: 2002-Jun-01 USB 1.1 'Open' Host Controller (OHCI) Driver (PCI) > ohci-pci.c: block sizes: ed 48 td 64 > hcd-pci.c: ohci-hcd @ 00:0d.0, OPTi Inc. 82C861 > hcd-pci.c: irq 11, pci mem c88bf000 > ... > hub.c: new USB device 00:0d.0-2, assigned address 2 > usb_control/bulk_msg: timeout > ohci-dbg.c: UNLINK:[ e91] dev:0,ep=0-O,CTRL,flags: 0,len:0/0,stat:-2 > hcd.c: 00:0d.0: wait for giveback urb c65409e0 > ... _______________________________________________________________ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
