On Tue, Mar 29, 2022 at 03:26:13PM -0700, David Vernet wrote:
> rcu_sync_enter() is an rcu_sync API used by updaters which forces RCU
> readers (e.g. percpu-rwsem) to take a slow-path during an update. It does
> this by setting the gp_state of the rcu_sync object to GP_ENTER (i.e. > 0).
> In this state, readers will take a "slow path", such as having percpu-rwsem
> readers wait on a semaphore rather than just incrementing a reader count.
> When the updater has completed their work, they must invoke rcu_sync_exit()
> to signal to readers that they may again take their fastpaths.
>
> rcu_sync_enter() currently has a comment that stipulates that a later call
> to rcu_sync_exit() (by an updater) will re-enable reader "slowpaths". This
> patch corrects the comment to properly reflect that rcu_sync_exit()
> re-enables reader fastpaths.
>
> Signed-off-by: David Vernet <void@xxxxxxxxxxxxx>
Good eyes, thank you!
As usual, I could not resist the urge to wordsmith. Please take a look
at the following to see if I messed something up.
Thanx, Paul
------------------------------------------------------------------------
commit b89e06a95c05009bcf31949814c42bc420f414a6
Author: David Vernet <void@xxxxxxxxxxxxx>
Date: Tue Mar 29 15:26:13 2022 -0700
rcu_sync: Fix comment to properly reflect rcu_sync_exit() behavior
The rcu_sync_enter() function is used by updaters to force RCU readers
(e.g. percpu-rwsem) to use their slow paths during an update. This is
accomplished by setting the ->gp_state of the rcu_sync structure to
GP_ENTER. In the case of percpu-rwsem, the readers' slow path waits on
a semaphore instead of just incrementing a reader count. Each updater
invokes the rcu_sync_exit() function to signal to readers that they
may again take their fastpaths. The rcu_sync_exit() function sets the
->gp_state of the rcu_sync structure to GP_EXIT, and if all goes well,
after a grace period the ->gp_state reverts back to GP_IDLE.
Unfortunately, the rcu_sync_enter() function currently has a comment
incorrectly stating that rcu_sync_exit() (by an updater) will re-enable
reader "slowpaths". This patch changes the comment to state that this
function re-enables reader fastpaths.
Signed-off-by: David Vernet <void@xxxxxxxxxxxxx>
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
diff --git a/kernel/rcu/sync.c b/kernel/rcu/sync.c
index 33d896d85902..5cefc702158f 100644
--- a/kernel/rcu/sync.c
+++ b/kernel/rcu/sync.c
@@ -111,7 +111,7 @@ static void rcu_sync_func(struct rcu_head *rhp)
* a slowpath during the update. After this function returns, all
* subsequent calls to rcu_sync_is_idle() will return false, which
* tells readers to stay off their fastpaths. A later call to
- * rcu_sync_exit() re-enables reader slowpaths.
+ * rcu_sync_exit() re-enables reader fastpaths.
*
* When called in isolation, rcu_sync_enter() must wait for a grace
* period, however, closely spaced calls to rcu_sync_enter() can
