x86/tsc: Force inlining of cyc2ns bits

Looking at the asm for native_sched_clock() I noticed we don't inline
enough. Mostly caused by sharing code with cyc2ns_read_begin(), which
we didn't used to do. So mark all that __force_inline to make it DTRT.

Fixes: 59eaef78 ("x86/tsc: Remodel cyc2ns to use seqcount_latch()")
Reported-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: hpa@zytor.com
Cc: eric.dumazet@gmail.com
Cc: bp@alien8.de
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20181011104019.695196158@infradead.org

arch/x86/kernel/tsc.c

@@ -58,7 +58,7 @@ struct cyc2ns {
 
 static DEFINE_PER_CPU_ALIGNED(struct cyc2ns, cyc2ns);
 
-void cyc2ns_read_begin(struct cyc2ns_data *data)
+void __always_inline cyc2ns_read_begin(struct cyc2ns_data *data)
 {
 	int seq, idx;
 
@@ -75,7 +75,7 @@ void cyc2ns_read_begin(struct cyc2ns_data *data)
 	} while (unlikely(seq != this_cpu_read(cyc2ns.seq.sequence)));
 }
 
-void cyc2ns_read_end(void)
+void __always_inline cyc2ns_read_end(void)
 {
 	preempt_enable_notrace();
 }
@@ -104,7 +104,7 @@ void cyc2ns_read_end(void)
  *                      -johnstul@us.ibm.com "math is hard, lets go shopping!"
  */
 
-static inline unsigned long long cycles_2_ns(unsigned long long cyc)
+static __always_inline unsigned long long cycles_2_ns(unsigned long long cyc)
 {
 	struct cyc2ns_data data;
 	unsigned long long ns;