Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Thomas Gleixner <tglx@linutronix.de>
To:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        linux-kernel@vger.kernel.org
Cc:     Crystal Wood <swood@redhat.com>, John Keeping <john@metanate.com>,
        Boqun Feng <boqun.feng@gmail.com>,
        Ingo Molnar <mingo@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Waiman Long <longman@redhat.com>, Will Deacon <will@kernel.org>
Subject: Re: [PATCH] locking/rtmutex: Do the trylock-slowpath with
 DEBUG_RT_MUTEXES enabled.
In-Reply-To: <20230328165430.9eOXd-55@linutronix.de>
References: <20230322162719.wYG1N0hh@linutronix.de>
 <20230328165430.9eOXd-55@linutronix.de>
Date:   Fri, 21 Apr 2023 19:58:52 +0200
Message-ID: <87zg7115ib.ffs@tglx>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

On Tue, Mar 28 2023 at 18:54, Sebastian Andrzej Siewior wrote:
> On 2023-03-22 17:27:21 [+0100], To Thomas Gleixner wrote:
>> > Aside of that for CONFIG_DEBUG_RT_MUTEXES=y builds it flushes on every
>> > lock operation whether the lock is contended or not.
>> 
>> For mutex & ww_mutex operations. rwsem is not affected by
>> CONFIG_DEBUG_RT_MUTEXES. As for mutex it could be mitigated by invoking
>> try_to_take_rt_mutex() before blk_flush_plug().

> I haven't observed anything in the ww-mutex path so we can ignore it or
> do something similar to this.

Yay for consistency !

I fixed it up to the below.

Thanks,

        tglx
---

--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -219,6 +219,11 @@ static __always_inline bool rt_mutex_cmp
 	return try_cmpxchg_acquire(&lock->owner, &old, new);
 }
 
+static __always_inline bool rt_mutex_try_acquire(struct rt_mutex_base *lock)
+{
+	return rt_mutex_cmpxchg_acquire(lock, old, new);
+}
+
 static __always_inline bool rt_mutex_cmpxchg_release(struct rt_mutex_base *lock,
 						     struct task_struct *old,
 						     struct task_struct *new)
@@ -298,6 +303,20 @@ static __always_inline bool rt_mutex_cmp
 
 }
 
+static __always_inline bool rt_mutex_try_acquire(struct rt_mutex_base *lock)
+{
+	/*
+	 * With debug enabled rt_mutex_cmpxchg trylock() will always fail,
+	 * which will unconditionally invoke blk_flush_plug() in the slow
+	 * path of __rt_mutex_lock() and __ww_rt_mutex_lock() even in the
+	 * non-contended case.
+	 *
+	 * Avoid that by using rt_mutex_slow_trylock() which is fully covered
+	 * by the debug code and can acquire a non-contended rtmutex.
+	 */
+	return rt_mutex_slowtrylock(lock);
+}
+
 static __always_inline bool rt_mutex_cmpxchg_release(struct rt_mutex_base *lock,
 						     struct task_struct *old,
 						     struct task_struct *new)
@@ -1698,9 +1717,8 @@ static int __sched rt_mutex_slowlock(str
 static __always_inline int __rt_mutex_lock(struct rt_mutex_base *lock,
 					   unsigned int state)
 {
-	if (likely(rt_mutex_cmpxchg_acquire(lock, NULL, current)))
+	if (likely(rt_mutex_try_acquire(lock)))
 		return 0;
-
 	/*
 	 * The task is about to sleep. Flush plugged IO as that might
 	 * take locks and corrupt tsk::pi_blocked_on.
--- a/kernel/locking/ww_rt_mutex.c
+++ b/kernel/locking/ww_rt_mutex.c
@@ -62,7 +62,7 @@ static int __sched
 	}
 	mutex_acquire_nest(&rtm->dep_map, 0, 0, nest_lock, ip);
 
-	if (likely(rt_mutex_cmpxchg_acquire(&rtm->rtmutex, NULL, current))) {
+	if (likely(rt_mutex_try_acquire(&rtm->rtmutex, NULL, current))) {
 		if (ww_ctx)
 			ww_mutex_set_context_fastpath(lock, ww_ctx);
 		return 0;