I discovered that the preempt_count problem exists in different forms
due to the way faults are handled are different archs when the
kgdb_may_fault and long jump occur to restore the system context. This
means that a general solution was needed vs a per arch solution.
Attached is the patch to fix the kgdb_may_fault memory access issues
when using a kernel that cares about the preempt count. This will avoid
random crashes of threads to the abort to trigger from a bad/unexpected
preempt count change. This patch also fixes a missed case where a
return can happen in an error condition where kgdb_may_fault is still
set. Of course the system will most definitely crash when
kgdb_may_fault is set and a resume occurs.
If there are no objections I will merge this patch soon and make a final
call for contributions to the 2.6.21 uprev. The next step is on to
2.6.22 and updating the kernel.org tree.
Thanks,
Jason.
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Kgdb-bugreport mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/kgdb-bugreport
