On Thu, 19 May 2005 12:34:08 BST, David Woodhouse said:
On Tue, 2005-05-17 at 12:24 -0400, Valdis.Kletnieks(a)vt.edu wrote:
> [4295584.974000] Debug: sleeping function called from invalid context
> at mm/slab.c:2502
OK, we'll just let audit_log() assemble its own skb and queue it
for a
separate kernel thread to feed up to auditd. We'll fix up the horrid
error handling mess we had before too, where we used to clone the skb
because we know netlink_unicast() would free it even on temporary errors
(like the SO_RCVBUF limit being reached).
This includes one of Steve's earlier patches which ensures that messages
in the skb are NUL-terminated. It should all appear some time soon in
http://www.kernel.org/git/gitweb.cgi?p=linux/kernel/git/dwmw2/audit-2.6.g...
=summary
or more perhaps more usefully (since Thomas allows us to distinguish
between local and merged commits) in
http://www.tglx.de/cgi-bin/gittracker/commit/tracker-linux/audit-2.6.git?...
ect=tracker-linux%2Fdwmw2%2Faudit-2.6.git&pagelen=10&exclude=tracker-linux%2Fto
rvalds%2Flinux-2.6.git&offset=0
Index: kernel/audit.c
===================================================================
--- e45ee43e7af31f847377e8bb3a0a61581732b653/kernel/audit.c (mode:100644)
+++ c1096ff7ae35b77bf8108c3a60b856551c50a9d7/kernel/audit.c (mode:100644)
Patch applies with a few offsets against 12-rc4-mm2, and I'm not
seeing the messages any more...