Re: [PATCH 01/13] rcu/kvfree: Add kvfree_rcu_mightsleep() and kfree_rcu_mightsleep()

[Paul E. McKenney]

On Thu, Feb 02, 2023 at 07:54:31AM +0000, Zhuo, Qiuxu wrote:
> > From: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
> > Sent: Wednesday, February 1, 2023 11:08 PM
> > To: LKML <linux-kernel@xxxxxxxxxxxxxxx>; RCU <rcu@xxxxxxxxxxxxxxx>; Paul E .
> > McKenney <paulmck@xxxxxxxxxx>
> > Cc: Uladzislau Rezki <urezki@xxxxxxxxx>; Oleksiy Avramchenko
> > <oleksiy.avramchenko@xxxxxxxx>; Jens Axboe <axboe@xxxxxxxxx>; Philipp
> > Reisner <philipp.reisner@xxxxxxxxxx>; Bryan Tan <bryantan@xxxxxxxxxx>;
> > Steven Rostedt <rostedt@xxxxxxxxxxx>; Eric Dumazet
> > <edumazet@xxxxxxxxxx>; Bob Pearson <rpearsonhpe@xxxxxxxxx>; Ariel
> > Levkovich <lariel@xxxxxxxxxx>; Theodore Ts'o <tytso@xxxxxxx>; Julian
> > Anastasov <ja@xxxxxx>
> > Subject: [PATCH 01/13] rcu/kvfree: Add kvfree_rcu_mightsleep() and
> > kfree_rcu_mightsleep()
> > 
> > These two macroses will replace single-argument k[v]free_rcu() ones.
> > By adding an extra _mightsleep prefix we can avoid of situations when someone
> 
> s/prefix/suffix

Good eyes, thank you!

Please see below for the version currently queued in the -rcu tree.

							Thanx, Paul

> > intended to give a second argument but forgot to do it in a code where sleeping
> > is illegal.
> > 
> > Signed-off-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
> > ...

------------------------------------------------------------------------

commit d6b54a44de23c83944d042b6f6fd9d19ccd3f1b8
Author: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
Date:   Wed Feb 1 16:08:07 2023 +0100

    rcu/kvfree: Add kvfree_rcu_mightsleep() and kfree_rcu_mightsleep()
    
    The kvfree_rcu() and kfree_rcu() APIs are hazardous in that if you forget
    the second argument, it works, but might sleep.  This sleeping can be a
    correctness bug from atomic contexts, and even in non-atomic contexts
    it might introduce unacceptable latencies.  This commit therefore adds
    kvfree_rcu_mightsleep() and kfree_rcu_mightsleep(), which will replace
    the single-argument kvfree_rcu() and kfree_rcu(), respectively.
    
    This commit enables a series of commits that switch from single-argument
    kvfree_rcu() and kfree_rcu() to their _mightsleep() counterparts.  Once
    all of these commits land, the single-argument versions will be removed.
    
    Signed-off-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
    Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>

diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
index 822ff7b4bb1ed..094321c17e48a 100644
--- a/include/linux/rcupdate.h
+++ b/include/linux/rcupdate.h
@@ -1013,6 +1013,9 @@ static inline notrace void rcu_read_unlock_sched_notrace(void)
 #define kvfree_rcu(...) KVFREE_GET_MACRO(__VA_ARGS__,		\
 	kvfree_rcu_arg_2, kvfree_rcu_arg_1)(__VA_ARGS__)
 
+#define kvfree_rcu_mightsleep(ptr) kvfree_rcu_arg_1(ptr)
+#define kfree_rcu_mightsleep(ptr) kvfree_rcu_mightsleep(ptr)
+
 #define KVFREE_GET_MACRO(_1, _2, NAME, ...) NAME
 #define kvfree_rcu_arg_2(ptr, rhf)					\
 do {									\