oprofile/x86: Add regs->ip to oprofile trace
authorJosh Poimboeuf <jpoimboe@redhat.com>
Wed, 24 Aug 2016 16:50:15 +0000 (11:50 -0500)
committerIngo Molnar <mingo@kernel.org>
Thu, 8 Sep 2016 06:58:40 +0000 (08:58 +0200)
dump_trace() doesn't add the interrupted instruction's address to the
trace, so add it manually.  This makes the profile more useful, and also
makes it more consistent with what perf profiling does.

Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Acked-by: Robert Richter <rric@kernel.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Byungchul Park <byungchul.park@lge.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Nilay Vaish <nilayvaish@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/6c745a83dbd69fc6857ef9b2f8be0f011d775936.1472057064.git.jpoimboe@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/oprofile/backtrace.c

index cb31a4440e588552f5fd046ae787c107aa09ec94..2ef6c8b56311cb575bbfd0bd879486f08d85b365 100644 (file)
@@ -114,9 +114,16 @@ x86_backtrace(struct pt_regs * const regs, unsigned int depth)
 
        if (!user_mode(regs)) {
                unsigned long stack = kernel_stack_pointer(regs);
-               if (depth)
-                       dump_trace(NULL, regs, (unsigned long *)stack, 0,
-                                  &backtrace_ops, &depth);
+
+               if (!depth)
+                       return;
+
+               oprofile_add_trace(regs->ip);
+               if (!--depth)
+                       return;
+
+               dump_trace(NULL, regs, (unsigned long *)stack, 0,
+                          &backtrace_ops, &depth);
                return;
        }