site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Thread-index: AclZUW96hDfD1iBNSiCLABNuYVcqKg== Thread-topic: Stall on shutdown/restart after auto kext-unload Earlier this year I wrote a kernel extension which includes several subclasses of storage/scsi classes. I can go into more detail if needed, but I doubt it's relevant for this problem. On some machines, if my driver is installed, the system won't shutdown - it sits spinning that circular multi-segment thing. The thing is, this only happens if there are no devices present which match my driver, or if any such devices are removed >1 minute before shutdown. This raised my suspicions about the kernel unloading my extension - I know it does that if no matching devices exist, and indeed I found that if I shut down while the kext is loaded, all is OK. If I wait until kextstat shows nothing, the system won't shutdown. However, if I kextunload manually before the shutdown, all is OK. So, very little of my code actually runs - only init(), probe() and free() in one of my classes. My driver seems to unload OK, yet the system won't shutdown. Any help would be appreciated! I've got two-machine debug up and running, but am a bit clueless how to find out why it's hanging. Any ideas? Thanks all, Andy PS problem has only been observed on OX 10.4.11 and 10.3.9 after a security update late summer/early fall. No symptoms observed on 10.5.5. Also, this doesn't seem to affect Xeon-based machines, only:- PowerMac G4 PowerMac G5 MacbookPro Core Duo MacMini Core Solo iMac Core2 Duo _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... This email sent to site_archiver@lists.apple.com