arm64: Optimize ttwu IPI

hulk inclusion
category: feature
bugzilla: NA
CVE: NA

When it is to wake up a task in a remote cpu shared LLC , we can
simply set need_resched flag, waking up a cpu that is in polling
idle. This wakeup action does not require an IPI.

But the premise is that it need to support _TIF_POLLING_NRFLAG
Signed-off-by: NXiangyou Xie <xiexiangyou@huawei.com>
Reviewed-by: NHanjun Guo <guohanjun@huawei.com>
Reviewed-by: Nzhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

arch/arm64/include/asm/thread_info.h

@@ -69,6 +69,7 @@ void arch_release_task_struct(struct task_struct *tsk);
  *  TIF_NEED_RESCHED	- rescheduling necessary
  *  TIF_NOTIFY_RESUME	- callback before returning to user
  *  TIF_USEDFPU		- FPU was used by this task this quantum (SMP)
+ *  TIF_POLLING_NRFLAG - idle is polling for TIF_NEED_RESCHED
  */
 #define TIF_SIGPENDING		0
 #define TIF_NEED_RESCHED	1
@@ -82,6 +83,7 @@ void arch_release_task_struct(struct task_struct *tsk);
 #define TIF_SYSCALL_AUDIT	9
 #define TIF_SYSCALL_TRACEPOINT	10
 #define TIF_SECCOMP		11
+#define TIF_POLLING_NRFLAG	16
 #define TIF_MEMDIE		18	/* is terminating due to OOM killer */
 #define TIF_FREEZE		19
 #define TIF_RESTORE_SIGMASK	20
@@ -108,6 +110,7 @@ void arch_release_task_struct(struct task_struct *tsk);
 #define _TIF_32BIT		(1 << TIF_32BIT)
 #define _TIF_SVE		(1 << TIF_SVE)
 #define _TIF_32BIT_AARCH64	(1 << TIF_32BIT_AARCH64)
+#define _TIF_POLLING_NRFLAG	(1 << TIF_POLLING_NRFLAG)
 
 #define _TIF_WORK_MASK		(_TIF_NEED_RESCHED | _TIF_SIGPENDING | \
 				 _TIF_NOTIFY_RESUME | _TIF_FOREIGN_FPSTATE | \