site_archiver@lists.apple.com Delivered-To: darwin-kernel@lists.apple.com The odds are that someone is watching the root of your filesystem with a kqueue. How are you cleaning up that vnode? A vflush() without SKIPROOT, even without FORCECLOSE, should reclaim it. Cheers, Dan On Dec 17, 2009, at 12:18 PM, Alain Renaud wrote:

Hello all,

I am working on a filesystem implementation for macosx that is working fine with macosx 10.5 now for the 10.6 port I have an issue as I cannot seem to release the mountpoint vnode.

My problem is that with 10.6 when VFS_UNMOUNT is called my vfs rootvnode as the usecount and kusecount set to 1 which prevent me from releasing this vnode. The same code with 10.5 the usecount/kusecount are set to 0 so I am then able to release the vnode and delete all reference to the filesystem before returning to the dounmount code.

The Question: Is this an expected new behavior with the 10.6 kernel or do I have a bug elsewhere in the code that increase the usecount/kusecount without reducing it after.

Hope this make sense and thanks in advance for any help.

-- Alain Renaud -- ---------------------------------------- Always willing to share my ignorance...

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/daniel.h%40apple.com

This email sent to daniel.h@apple.com

_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.a... This email sent to site_archiver@lists.apple.com