y2038: __get_old_timespec32() can be static

The kbuild test robot reports two new warnings with the previous
patch:

 kernel/time/time.c:866:5: sparse: symbol '__get_old_timespec32' was not declared. Should it be static?
 kernel/time/time.c:882:5: sparse: symbol '__put_old_timespec32' was not declared. Should it be static?

These are actually older bugs, but came up now after the
symbol got renamed. Fortunately, commit afef05c ("time:
Enable get/put_compat_itimerspec64 always") makes the two functions
(__compat_get_timespec64/__compat_get_timespec64) local to time.c already,
so we can mark them as 'static'.

Fixes: ee16c8f415e4 ("y2038: Globally rename compat_time to old_time32")
Signed-off-by: kbuild test robot <fengguang.wu@intel.com>
[arnd: added changelog text]
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

kernel/time/time.c

@@ -863,7 +863,7 @@ int put_timespec64(const struct timespec64 *ts,
 }
 EXPORT_SYMBOL_GPL(put_timespec64);
 
-int __get_old_timespec32(struct timespec64 *ts64,
+static int __get_old_timespec32(struct timespec64 *ts64,
 				   const struct old_timespec32 __user *cts)
 {
 	struct old_timespec32 ts;
@@ -879,7 +879,7 @@ int __get_old_timespec32(struct timespec64 *ts64,
 	return 0;
 }
 
-int __put_old_timespec32(const struct timespec64 *ts64,
+static int __put_old_timespec32(const struct timespec64 *ts64,
 				   struct old_timespec32 __user *cts)
 {
 	struct old_timespec32 ts = {