sched, time: Fix build error with 64 bit cputime_t on 32 bit systems

On 32 bit systems cmpxchg cannot handle 64 bit values, so
some additional magic is required to allow a 32 bit system
with CONFIG_VIRT_CPU_ACCOUNTING_GEN=y enabled to build.

Make sure the correct cmpxchg function is used when doing
an atomic swap of a cputime_t.

Reported-by: Arnd Bergmann <[email protected]>
Signed-off-by: Rik van Riel <[email protected]>
Acked-by: Arnd Bergmann <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: Andrew Morton <[email protected]>
Cc: Benjamin Herrenschmidt <[email protected]>
Cc: Heiko Carstens <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: Martin Schwidefsky <[email protected]>
Cc: Michael Ellerman <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>

arch/powerpc/include/asm/cputime.h

@@ -32,6 +32,8 @@ static inline void setup_cputime_one_jiffy(void) { }
 typedef u64 __nocast cputime_t;
 typedef u64 __nocast cputime64_t;
 
+#define cmpxchg_cputime(ptr, old, new) cmpxchg(ptr, old, new)
+
 #ifdef __KERNEL__
 
 /*

arch/s390/include/asm/cputime.h

@@ -18,6 +18,8 @@
 typedef unsigned long long __nocast cputime_t;
 typedef unsigned long long __nocast cputime64_t;
 
+#define cmpxchg_cputime(ptr, old, new) cmpxchg64(ptr, old, new)
+
 static inline unsigned long __div(unsigned long long n, unsigned long base)
 {
 #ifndef CONFIG_64BIT

include/asm-generic/cputime_jiffies.h

@@ -3,6 +3,8 @@
 
 typedef unsigned long __nocast cputime_t;
 
+#define cmpxchg_cputime(ptr, old, new) cmpxchg(ptr, old, new)
+
 #define cputime_one_jiffy		jiffies_to_cputime(1)
 #define cputime_to_jiffies(__ct)	(__force unsigned long)(__ct)
 #define cputime_to_scaled(__ct)		(__ct)

include/asm-generic/cputime_nsecs.h

@@ -21,6 +21,8 @@
 typedef u64 __nocast cputime_t;
 typedef u64 __nocast cputime64_t;
 
+#define cmpxchg_cputime(ptr, old, new) cmpxchg64(ptr, old, new)
+
 #define cputime_one_jiffy		jiffies_to_cputime(1)
 
 #define cputime_div(__ct, divisor)  div_u64((__force u64)__ct, divisor)

kernel/sched/cputime.c

@@ -554,6 +554,23 @@ drop_precision:
 	return (__force cputime_t) scaled;
 }
 
+/*
+ * Atomically advance counter to the new value. Interrupts, vcpu
+ * scheduling, and scaling inaccuracies can cause cputime_advance
+ * to be occasionally called with a new value smaller than counter.
+ * Let's enforce atomicity.
+ *
+ * Normally a caller will only go through this loop once, or not
+ * at all in case a previous caller updated counter the same jiffy.
+ */
+static void cputime_advance(cputime_t *counter, cputime_t new)
+{
+	cputime_t old;
+
+	while (new > (old = ACCESS_ONCE(*counter)))
+		cmpxchg_cputime(counter, old, new);
+}
+
 /*
  * Adjust tick based cputime random precision against scheduler
  * runtime accounting.
@@ -599,16 +616,8 @@ static void cputime_adjust(struct task_cputime *curr,
 		utime = rtime - stime;
 	}
 
-	/*
-	 * If the tick based count grows faster than the scheduler one,
-	 * the result of the scaling may go backward.
-	 * Let's enforce monotonicity.
-	 * Atomic exchange protects against concurrent cputime_adjust().
-	 */
-	while (stime > (rtime = ACCESS_ONCE(prev->stime)))
-		cmpxchg(&prev->stime, rtime, stime);
-	while (utime > (rtime = ACCESS_ONCE(prev->utime)))
-		cmpxchg(&prev->utime, rtime, utime);
+	cputime_advance(&prev->stime, stime);
+	cputime_advance(&prev->utime, utime);
 
 out:
 	*ut = prev->utime;