[PATCH v2 03/23] cpu/hotplug: Avoid busy-polling on archs where cpu_relax() is a no-op

[Will Deacon]

On some architectures (such as arm64), cpu_relax() is effectively a NOP
and so isn't particularly efficient when used in a tight polling loop
such as the CPU state synchronisation in cpuhp_ap_sync_alive().

Once an incoming CPU has reached the SYNC_STATE_ALIVE state, we know
that it is executing within the kernel and so we can use the more
efficient polling mechanism provided by the atomic_cond_read* API,
allowing the use of WFE on arm64 and Zawrs on RISC-V.

Change cpuhp_ap_sync_alive() to poll using atomic_cond_read_relaxed()
instead of cpu_relax(). No change on x86.

Reviewed-by: Jinjie Ruan <ruanjinjie@xxxxxxxxxx>
Signed-off-by: Will Deacon <will@xxxxxxxxxx>
---
 kernel/cpu.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index 97a9bfe4edad..786c3191431c 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -395,8 +395,7 @@ void cpuhp_ap_sync_alive(void)
 	cpuhp_ap_update_sync_state(SYNC_STATE_ALIVE);
 
 	/* Wait for the control CPU to release it. */
-	while (atomic_read(st) != SYNC_STATE_SHOULD_ONLINE)
-		cpu_relax();
+	atomic_cond_read_acquire(st, VAL == SYNC_STATE_SHOULD_ONLINE);
 }
 
 static bool cpuhp_can_boot_ap(unsigned int cpu)
-- 
2.56.0.385.gd3acb90ef8-goog