From: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
commit 4b377b4868ef17b040065bd468668c707d2477a5 upstream.
Commit 1a7d0890dd4a ("kprobe/ftrace: bail out if ftrace was killed")
introduced a bad K&R function definition, which we haven't accepted in a
long long time.
Gcc seems to let it slide, but clang notices with the appropriate error:
kernel/kprobes.c:1140:24: error: a function declaration without a prototype is deprecated in all >
1140 | void kprobe_ftrace_kill()
| ^
| void
but this commit was apparently never in linux-next before it was sent
upstream, so it didn't get the appropriate build test coverage.
Fixes: 1a7d0890dd4a kprobe/ftrace: bail out if ftrace was killed
Cc: Stephen Brennan <stephen.s.brennan@xxxxxxxxxx>
Cc: Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>
Cc: Guo Ren <guoren@xxxxxxxxxx>
Cc: Steven Rostedt (Google) <rostedt@xxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
kernel/kprobes.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -1138,7 +1138,7 @@ static int disarm_kprobe_ftrace(struct k
ipmodify ? &kprobe_ipmodify_enabled : &kprobe_ftrace_enabled);
}
-void kprobe_ftrace_kill()
+void kprobe_ftrace_kill(void)
{
kprobe_ftrace_disabled = true;
}
Patches currently in stable-queue which might be from torvalds@xxxxxxxxxxxxxxxxxxxx are
queue-6.9/tty-add-the-option-to-have-a-tty-reject-a-new-ldisc.patch
queue-6.9/kprobe-ftrace-fix-build-error-due-to-bad-function-definition.patch
queue-6.9/rcutorture-fix-rcu_torture_one_read-pipe_count-overf.patch
queue-6.9/revert-mm-mmap-allow-for-the-maximum-number-of-bits-for-randomizing-mmap_base-by-default.patch
