Web lists-archives.com

[PATCH v7 35/37] tracing: Increase trace_recursive_lock() limit for synthetic events




Synthetic event generation needs to happen while the current event is
still in progress, so add 1 to the trace_recursive_lock() recursion
limit to account for that.

Because we also want to allow for the possibility of a synthetic event
being generated from another synthetic event, add an additional
increment for that as well.

Signed-off-by: Tom Zanussi <tom.zanussi@xxxxxxxxxxxxxxx>
---
 kernel/trace/ring_buffer.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 763b3fb..8c862ea 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -2589,16 +2589,16 @@ static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer,
  *  IRQ context
  *  NMI context
  *
- * If for some reason the ring buffer starts to recurse, we
- * only allow that to happen at most 4 times (one for each
- * context). If it happens 5 times, then we consider this a
- * recusive loop and do not let it go further.
+ * If for some reason the ring buffer starts to recurse, we only allow
+ * that to happen at most 6 times (one for each context, plus possibly
+ * two levels of synthetic event generation). If it happens 7 times,
+ * then we consider this a recusive loop and do not let it go further.
  */
 
 static __always_inline int
 trace_recursive_lock(struct ring_buffer_per_cpu *cpu_buffer)
 {
-	if (cpu_buffer->current_context >= 4)
+	if (cpu_buffer->current_context >= 6)
 		return 1;
 
 	cpu_buffer->current_context++;
-- 
1.9.3