sh: oprofile: Make sure the backtrace op is available for timer-fallback.
authorPaul Mundt <lethal@linux-sh.org>
Tue, 12 Oct 2010 22:43:50 +0000 (07:43 +0900)
committerPaul Mundt <lethal@linux-sh.org>
Wed, 27 Oct 2010 08:03:50 +0000 (17:03 +0900)
Presently with hardware counter support disabled the backtrace op never
gets initialized. This is a regression over the previous behaviour, so
simply add it back in.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
arch/sh/oprofile/common.c

index 84533142da9b1e84e1aa05fa6d82d02208223f84..b4c2d2b946ddc084bfe48b2d58128d47e2f724aa 100644 (file)
@@ -21,9 +21,9 @@
 #include <linux/slab.h>
 #include <asm/processor.h>
 
-#ifdef CONFIG_HW_PERF_EVENTS
 extern void sh_backtrace(struct pt_regs * const regs, unsigned int depth);
 
+#ifdef CONFIG_HW_PERF_EVENTS
 /*
  * This will need to be reworked when multiple PMUs are supported.
  */
@@ -57,7 +57,7 @@ void __exit oprofile_arch_exit(void)
 #else
 int __init oprofile_arch_init(struct oprofile_operations *ops)
 {
-       pr_info("oprofile: hardware counters not available\n");
+       ops->backtrace = sh_backtrace;
        return -ENODEV;
 }
 void __exit oprofile_arch_exit(void) {}