[Bug kprobes/2568] KRetprobe on powerpc - Recursive post-handler - > fault-handler

[jkenisto at us dot ibm dot com]

------- Additional Comments From jkenisto at us dot ibm dot com  2006-04-13 00:35 -------
The user shouldn't have to fuss with the kretprobe's embedded kprobe. 
Typically, this is not a problem because the kretprobe is statically allocated
and implicitly initialized to zeroes.  In this case, the kretprobe was probably
allocated dynamically (and not zeroed).

Brad's probably right that the fix is for register_kretprobe() to NULL out the
embedded kprobe's post-handler.

How about the fault-handler?  It's not clear whether a non-null fault handler
should be called if the user's kretprobe handler faults.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|minor                       |normal


http://sourceware.org/bugzilla/show_bug.cgi?id=2568

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.