[PATCH 3.16 126/328] uprobes: Use synchronize_rcu() not synchronize_sched()

From: Ben Hutchings
Date: Sun Dec 09 2018 - 17:13:33 EST

3.16.62-rc1 review patch. If anyone has any objections, please let me know.


From: "Steven Rostedt (VMware)" <rostedt@xxxxxxxxxxx>

commit 016f8ffc48cb01d1e7701649c728c5d2e737d295 upstream.

While debugging another bug, I was looking at all the synchronize*()
functions being used in kernel/trace, and noticed that trace_uprobes was
using synchronize_sched(), with a comment to synchronize with
{u,ret}_probe_trace_func(). When looking at those functions, the data is
protected with "rcu_read_lock()" and not with "rcu_read_lock_sched()". This
is using the wrong synchronize_*() function.

Link: http://lkml.kernel.org/r/20180809160553.469e1e32@xxxxxxxxxxxxxxxxxx

Fixes: 70ed91c6ec7f8 ("tracing/uprobes: Support ftrace_event_file base multibuffer")
Acked-by: Oleg Nesterov <oleg@xxxxxxxxxx>
Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
kernel/trace/trace_uprobe.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -949,7 +949,7 @@ probe_event_disable(struct trace_uprobe

/* synchronize with u{,ret}probe_trace_func */
- synchronize_sched();
+ synchronize_rcu();

if (!list_empty(&tu->tp.files))