[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230615133415.070156632@goodmis.org>
Date: Thu, 15 Jun 2023 09:05:33 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: linux-kernel@...r.kernel.org
Cc: Masami Hiramatsu <mhiramat@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Andrew Morton <akpm@...ux-foundation.org>,
sunliming <sunliming@...inos.cn>,
Beau Belgrave <beaub@...ux.microsoft.com>
Subject: [for-linus][PATCH 02/15] tracing/user_events: Prevent same name but different args event
From: sunliming <sunliming@...inos.cn>
User processes register name_args for events. If the same name but different
args event are registered. The trace outputs of second event are printed
as the first event. This is incorrect.
Return EADDRINUSE back to the user process if the same name but different args
event has being registered.
Link: https://lore.kernel.org/linux-trace-kernel/20230529032100.286534-1-sunliming@kylinos.cn
Signed-off-by: sunliming <sunliming@...inos.cn>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@...nel.org>
Acked-by: Beau Belgrave <beaub@...ux.microsoft.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@...dmis.org>
---
kernel/trace/trace_events_user.c | 36 +++++++++++++++----
.../selftests/user_events/ftrace_test.c | 6 ++++
2 files changed, 36 insertions(+), 6 deletions(-)
diff --git a/kernel/trace/trace_events_user.c b/kernel/trace/trace_events_user.c
index dbb14705d0d3..37a38496a6be 100644
--- a/kernel/trace/trace_events_user.c
+++ b/kernel/trace/trace_events_user.c
@@ -1786,6 +1786,8 @@ static int user_event_parse(struct user_event_group *group, char *name,
int ret;
u32 key;
struct user_event *user;
+ int argc = 0;
+ char **argv;
/* Prevent dyn_event from racing */
mutex_lock(&event_mutex);
@@ -1793,13 +1795,35 @@ static int user_event_parse(struct user_event_group *group, char *name,
mutex_unlock(&event_mutex);
if (user) {
- *newuser = user;
- /*
- * Name is allocated by caller, free it since it already exists.
- * Caller only worries about failure cases for freeing.
- */
- kfree(name);
+ if (args) {
+ argv = argv_split(GFP_KERNEL, args, &argc);
+ if (!argv) {
+ ret = -ENOMEM;
+ goto error;
+ }
+
+ ret = user_fields_match(user, argc, (const char **)argv);
+ argv_free(argv);
+
+ } else
+ ret = list_empty(&user->fields);
+
+ if (ret) {
+ *newuser = user;
+ /*
+ * Name is allocated by caller, free it since it already exists.
+ * Caller only worries about failure cases for freeing.
+ */
+ kfree(name);
+ } else {
+ ret = -EADDRINUSE;
+ goto error;
+ }
+
return 0;
+error:
+ refcount_dec(&user->refcnt);
+ return ret;
}
user = kzalloc(sizeof(*user), GFP_KERNEL_ACCOUNT);
diff --git a/tools/testing/selftests/user_events/ftrace_test.c b/tools/testing/selftests/user_events/ftrace_test.c
index 7c99cef94a65..6e8c4b47281c 100644
--- a/tools/testing/selftests/user_events/ftrace_test.c
+++ b/tools/testing/selftests/user_events/ftrace_test.c
@@ -228,6 +228,12 @@ TEST_F(user, register_events) {
ASSERT_EQ(0, ioctl(self->data_fd, DIAG_IOCSREG, ®));
ASSERT_EQ(0, reg.write_index);
+ /* Multiple registers to same name but different args should fail */
+ reg.enable_bit = 29;
+ reg.name_args = (__u64)"__test_event u32 field1;";
+ ASSERT_EQ(-1, ioctl(self->data_fd, DIAG_IOCSREG, ®));
+ ASSERT_EQ(EADDRINUSE, errno);
+
/* Ensure disabled */
self->enable_fd = open(enable_file, O_RDWR);
ASSERT_NE(-1, self->enable_fd);
--
2.39.2
Powered by blists - more mailing lists