Re(6): Baffled
Re(6): Baffled
- Subject: Re(6): Baffled
- From: Peter Lovell <email@hidden>
- Date: Fri, 14 Feb 2003 14:38:47 -0500
Woops. You had mentioned IOKit and I hadn't paid enough attention.
IOKit is not my area of expertise (I deal mostly with non-IOKit NKE's)
but I do seem to remember that startup and stop routines for IOKit are
handled differently. I thought that start and stop entry points for
IOKit modules were actually in IOKit code, and the IOKit start/stop
then called your methods.
Regards.....Peter
On Friday, February 14, 2003, at 01:33 PM, Eric Long wrote:
Did you make sure that the entry-point names are defined as extern "C"
?? They mustn't be mangled so if your source file is C++ then you must
define them with regular C linkage.
The module start/stop function definitions reside exclusively in a .c
file.
I think that particular message refers to the IOKit related code, which
descends from IOKit classes.
If the OS is looking for my start/stop function names and can't find
them, I
would expect to see some kind of error noted. It doesn't look to me
like it
is even trying to find them. It's like it doesn't think there's
anything
besides the IOKit related code in the kext.
The problem occurs even when the kext is owned by root, in group
wheel. In
remote debugging, I can see all the symbols I expect. If I set
breakpoints
in my start/stop functions, they never get hit. I've looked at the
data in
the executable. It has the start/stop function names, unmangled,
within it.
Very strange.
Eric
_______________________________________________
darwin-kernel mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/darwin-kernel
Do not post admin requests to the list. They will be ignored.
_______________________________________________
darwin-kernel mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/darwin-kernel
Do not post admin requests to the list. They will be ignored.