On Thu, 14 Nov 2002, Duncan Haldane wrote: | OK. This time I gfot something from ksymoops.... | ... | | any ideas?
Hi, I think that the usb_bus_list_lock semaphore just isn't initialized yet. This can happen because we don't have much control over the order of initcalls in 2.4. There is more control over them in 2.5. So I think that the usb_cpia_init() is happening before usb_init(). One thing that we could do is have usb_register() return an error code in this case, but that probably wouldn't help in the long term. If it were a module, the module could be reloaded, but in this case, it's built-into-kernel, so it would just be dead. Another thing that we could do is try to do some init inside usb_register(), but that looks unsafe (racy). Another thing that we might do is try to force (control) the init order. This might be the best fix overall. Other ideas out there? -- ~Randy ------------------------------------------------------- This sf.net email is sponsored by: To learn the basics of securing your web site with SSL, click here to get a FREE TRIAL of a Thawte Server Certificate: http://www.gothawte.com/rd524.html _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel