This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
[Bug kprobes/2637] skipped probes in FC6
- From: "anil dot s dot keshavamurthy at intel dot com" <sourceware-bugzilla at sourceware dot org>
- To: systemtap at sources dot redhat dot com
- Date: 5 May 2006 21:33:45 -0000
- Subject: [Bug kprobes/2637] skipped probes in FC6
- References: <20060502201312.2637.hunt@redhat.com>
- Reply-to: sourceware-bugzilla at sourceware dot org
------- Additional Comments From anil dot s dot keshavamurthy at intel dot com 2006-05-05 21:33 -------
> True. But such situations have in the past been extremely rare. What is
> it about the new kprobes that makes it now so common and predictable? We
> need to understand this better.
In the past, we did not had the fixup_exception() call from
kprobes_exception_handler() code path and due to which any page_fault
generated by accessing user_data from pre/post handler used to
fall back to do_page_fault() function and hence you were
not seeing failure that often.
Again when we fall back to do_page_fault() function
we will technically be in preempt_disable() state which
is also wrong and potentially we could see a hang the system
on UP system.
Hence the right thing to do here is to catch that page_fault exceptions and
try fixing up the exceptions without falling back on do_page_fault() code path.
With this simple fix, you would now experience any copy_from_user() calls from
pre/post hanlder to fail if that page is not in memory.
--
http://sourceware.org/bugzilla/show_bug.cgi?id=2637
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.