David: It turns out that the scsi error-handler threads aren't getting removed because of a problem in the driver model core caused by a recent patch. I believe the patch below will fix it. (I haven't tried this patch, but one very similar to it worked okay.)
I've been trying to convince Greg that this is a real bug. In the meantime, see if this helps. Alan Stern ===== class.c 1.30 vs edited ===== --- 1.30/drivers/base/class.c Tue Jun 3 15:15:17 2003 +++ edited/drivers/base/class.c Wed Jun 11 14:50:39 2003 @@ -282,8 +282,12 @@ class_device_driver_link(class_dev); register_done: - if (error && parent) - class_put(parent); + if (error) { + if (class_dev->dev) + put_device(class_dev->dev); + if (parent) + class_put(parent); + } class_device_put(class_dev); return error; } @@ -316,6 +320,8 @@ kobject_del(&class_dev->kobj); + if (class_dev->dev) + put_device(class_dev->dev); if (parent) class_put(parent); } ------------------------------------------------------- This SF.NET email is sponsored by: eBay Great deals on office technology -- on eBay now! Click here: http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5 _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel