Re: Can a Kext Unload Itself?
Re: Can a Kext Unload Itself?
- Subject: Re: Can a Kext Unload Itself?
- From: Dean Reece <email@hidden>
- Date: Sat, 3 Feb 2007 23:52:21 -0800
On Feb 3, 2007, at 10:46 PM, Jones Curtis wrote:
Is there an acceptable way for a kext to unload itself without any
external help (like a socket to a user process that actually calls
kextunload)?
IOKit KEXTs can arrange to be unloaded by simply having the instance
counts of all their classes go to zero. A short time after this
happens, the kext will be unloaded.
For non-IOKit KEXTs there isn't a good way to do this today. The
problem is that your kext can't call into a function to cause itself
to unload, because the kext will still need to be around to execute
the return path. I suppose we could invent a non-returning unload
function that terminated the thread it was called on. Frankly, the
question has never come up as far as I know. I guess if you need a
helper to load, having a helper to unload seems natural enough that
nobody has asked before.
Go ahead and write up an enhancement request if you think this is
something you'd like to see added for a future release.
Cheers,
- Dean
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden