ftrace: Properly unset FTRACE_HASH_FL_MOD

commit 0ce0638edf5ec83343302b884fa208179580700a upstream.

When executing following commands like what document said, but the log
"#### all functions enabled ####" was not shown as expect:
  1. Set a 'mod' filter:
    $ echo 'write*:mod:ext3' > /sys/kernel/tracing/set_ftrace_filter
  2. Invert above filter:
    $ echo '!write*:mod:ext3' >> /sys/kernel/tracing/set_ftrace_filter
  3. Read the file:
    $ cat /sys/kernel/tracing/set_ftrace_filter

By some debugging, I found that flag FTRACE_HASH_FL_MOD was not unset
after inversion like above step 2 and then result of ftrace_hash_empty()
is incorrect.

Link: https://lkml.kernel.org/r/20220926152008.2239274-1-zhengyejian1@huawei.com
Cc: <mingo@redhat.com>
Cc: stable@vger.kernel.org
Fixes: 8c08f0d5c6fb ("ftrace: Have cached module filters be an active filter")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index e10cf1b..37db5bf 100644 (file)
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -5654,8 +5654,12 @@ int ftrace_regex_release(struct inode *inode, struct file *file)
 
                if (filter_hash) {
                        orig_hash = &iter->ops->func_hash->filter_hash;
-                       if (iter->tr && !list_empty(&iter->tr->mod_trace))
-                               iter->hash->flags |= FTRACE_HASH_FL_MOD;
+                       if (iter->tr) {
+                               if (list_empty(&iter->tr->mod_trace))
+                                       iter->hash->flags &= ~FTRACE_HASH_FL_MOD;
+                               else
+                                       iter->hash->flags |= FTRACE_HASH_FL_MOD;
+                       }
                } else
                        orig_hash = &iter->ops->func_hash->notrace_hash;