[K42-discussion] destruction of global kernel objects

[Bryan S Rosenburg]

In messing with XBaseObj, I'd be worried about screwing up the method 
numbers that are embedded in stub-generated code.  Most XBaseObj "methods" 
are never actually invoked via C++ mechanisms.  Instead a method number 
arrives in an incoming protected procedure call and is used as an index 
into the XObj's vtable.  The actual call is programmed out in assembler. 
I'm wondering if you've caused a mismatch in the method numbering between 
what the stub generator expects and what the compiler is actually 
generating for XObjs.

- Bryan




Donour Sizemore <donour at cs.unm.edu> 
Sent by: k42-discussion-bounces+rosnbrg=us.ibm.com at ozlabs.org
09/20/2006 04:43 PM
Please respond to
Discussion about K42 <k42-discussion at ozlabs.org>


To
k42discuss <k42-discussion at ozlabs.org>
cc

Subject
[K42-discussion] destruction of global kernel objects






I've been beating my head against the wall now while trying to add a 
virtual destructor to XBaseObjCore. It causes no problems when building, 
but it causes a failure on boot.

Specifically, during KernelInitPhase2, TypeMgrServer::ClassInit calls 
this function and I cannot figure out why. TypeMgrServer does inherit 
from XBaseObj, I don't think it should be calling the destructor. 
However, this isn't where it's crashing. Instead it's crashing on 
SysFSNode::ClassInit.

I would attach a patch to demonstrate this problem, but all you have to 
do is add an empty virtual destructor to XBaseObjCare, rebuild and run 
regress.

Anyone know what is cause these classinit's to get borked.

donour

_______________________________________________
K42-discussion mailing list
K42-discussion at ozlabs.org
https://ozlabs.org/mailman/listinfo/k42-discussion

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ozlabs.org/pipermail/k42-discussion/attachments/20060920/228fbfda/attachment.htm