tracing: Don't trace kernel thread syscalls
authorHendrik Brueckner <brueckner@linux.vnet.ibm.com>
Tue, 25 Aug 2009 16:02:37 +0000 (18:02 +0200)
committerFrederic Weisbecker <fweisbec@gmail.com>
Wed, 26 Aug 2009 19:29:52 +0000 (21:29 +0200)
Kernel threads don't call syscalls using the sysenter/sysexit
path. Instead they directly call the sys_* or do_* functions
that implement the syscalls inside the kernel.

The current syscall tracepoints only bind the sysenter/sysexit
path, then it has no effect to trace the kernel thread calls
to syscalls in that path.
Setting the TIF_SYSCALL_TRACEPOINT flag is then useless for these.

Actually there is only one case when a kernel thread can reach the
usual syscall exit tracing path: when we create a kernel thread, the
child comes to ret_from_fork and is the fork() return is then traced.
But this information alone is useless, then we don't want to set the
TIF flags for these threads.

Kernel threads have task_struct->mm set to NULL.
(Thanks to Heiko for that hint ;-)
The idea is then to check the mm field in syscall_regfunc() and
set the flag accordingly.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Jiaying Zhang <jiayingz@google.com>
Cc: Martin Bligh <mbligh@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
LKML-Reference: <20090825160237.GG4639@cetus.boeblingen.de.ibm.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
kernel/tracepoint.c

index 1a6a453b7efbbe326c79dbb01271c3e99d92015d..9489a0a9b1be69ed21234d9211df1f05a9857ec5 100644 (file)
@@ -597,7 +597,9 @@ void syscall_regfunc(void)
        if (!sys_tracepoint_refcount) {
                read_lock_irqsave(&tasklist_lock, flags);
                do_each_thread(g, t) {
-                       set_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT);
+                       /* Skip kernel threads. */
+                       if (t->mm)
+                               set_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT);
                } while_each_thread(g, t);
                read_unlock_irqrestore(&tasklist_lock, flags);
        }