From: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>
commit 372e0d01da71c84dcecf7028598a33813b0d5256 upstream.
The race between adding a function probe and reading the probes that exist
is very subtle. It needs a comment. Also, the issue can also happen if the
probe has has the EMPTY_HASH as its func_hash.
Cc: stable@xxxxxxxxxxxxxxx
Fixes: 7b60f3d876156 ("ftrace: Dynamically create the probe ftrace_ops for the trace_array")
Signed-off-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
kernel/trace/ftrace.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -3113,7 +3113,11 @@ t_probe_next(struct seq_file *m, loff_t
hash = iter->probe->ops.func_hash->filter_hash;
- if (!hash)
+ /*
+ * A probe being registered may temporarily have an empty hash
+ * and it's at the end of the func_probes list.
+ */
+ if (!hash || hash == EMPTY_HASH)
return NULL;
size = 1 << hash->size_bits;
@@ -4311,6 +4315,10 @@ register_ftrace_function_probe(char *glo
mutex_unlock(&ftrace_lock);
+ /*
+ * Note, there's a small window here that the func_hash->filter_hash
+ * may be NULL or empty. Need to be carefule when reading the loop.
+ */
mutex_lock(&probe->ops.func_hash->regex_lock);
orig_hash = &probe->ops.func_hash->filter_hash;
