Re: [PATCH] stop_machine: pass curstate to ack_state()

From: Rong Tao
Date: Tue Oct 24 2023 - 20:54:05 EST



On 10/25/23 8:50 AM, Rong Tao wrote:

On 10/25/23 8:43 AM, Rong Tao wrote:
The multi_cpu_stop() state machine uses multi_stop_data::state to hold
the current state, and this is read and written atomically except in
ack_state(), which performs a non-atomic read.

As ack_state() only performs this non-atomic read when there is a single
writer, this is benign, but it makes reasoning about the state machine a
little harder.

Remove the non-atomic read and pass the (atomically read) curstate in
instead. This makes it clear that we do not expect any racy writes, and
avoids a redundant load.

I'd prefer if we make this comment:

stop_machine: pass curstate to ack_state()
OK, thanks a lot, i'll try that.


Acked-by: Mark Rutland <mark.rutland@xxxxxxx>
Signed-off-by: Rong Tao <rongtao@xxxxxxxx>
---
 kernel/stop_machine.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index cedb17ba158a..268c2e581698 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -188,10 +188,11 @@ static void set_state(struct multi_stop_data *msdata,
 }

 /* Last one to ack a state moves to the next state. */
-static void ack_state(struct multi_stop_data *msdata)
+static void ack_state(struct multi_stop_data *msdata,
+              enum multi_stop_state curstate)
 {
     if (atomic_dec_and_test(&msdata->thread_ack))
-        set_state(msdata, msdata->state + 1);
+        set_state(msdata, curstate + 1);
 }

 notrace void __weak stop_machine_yield(const struct cpumask *cpumask)
@@ -242,7 +243,7 @@ static int multi_cpu_stop(void *data)
             default:
                 break;
             }
-            ack_state(msdata);
+            ack_state(msdata, curstate);
         } else if (curstate > MULTI_STOP_PREPARE) {
             /*
              * At this stage all other CPUs we depend on must spin