Date:   Wed, 22 Feb 2023 19:53:59 -0800
From:   Davidlohr Bueso <dave@stgolabs.net>
To:     "Paul E. McKenney" <paulmck@kernel.org>
Cc:     "Zhang, Qiang1" <qiang1.zhang@intel.com>,
        "josh@joshtriplett.org" <josh@joshtriplett.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] locktorture: Add raw_spinlock* torture tests for
 PREEMPT_RT kernels
Message-ID: <20230223035359.6jgserikqtc3vnra@offworld>
References: <20230215061035.1534950-1-qiang1.zhang@intel.com>
 <20230218193438.GE2948950@paulmck-ThinkPad-P17-Gen-1>
 <PH0PR11MB5880CD867814DFAAF8930E5BDAA79@PH0PR11MB5880.namprd11.prod.outlook.com>
 <20230222225755.GW2948950@paulmck-ThinkPad-P17-Gen-1>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Disposition: inline
In-Reply-To: <20230222225755.GW2948950@paulmck-ThinkPad-P17-Gen-1>
User-Agent: NeoMutt/20220429
Precedence: bulk

On Wed, 22 Feb 2023, Paul E. McKenney wrote:

>commit edc9d419ee8c22821ffd664466a5cf19208c3f02
>Author: Zqiang <qiang1.zhang@intel.com>
>Date:   Wed Feb 15 14:10:35 2023 +0800
>
>    locktorture: Add raw_spinlock* torture tests for PREEMPT_RT kernels
>
>    In PREEMPT_RT kernels, both spin_lock() and spin_lock_irq() are converted
>    to sleepable rt_spin_lock().  This means that the interrupt related suffix
>    for spin_lock/unlock(_irq, irqsave/irqrestore) do not affect the CPU's
>    interrupt state. This commit therefore adds raw spin-lock torture tests.
>    This in turn permits pure spin locks to be tested in PREEMPT_RT kernels.
>
>    Signed-off-by: Zqiang <qiang1.zhang@intel.com>
>    Signed-off-by: Paul E. McKenney <paulmck@kernel.org>

This is a nice addition, thanks. Just one comment below.

>diff --git a/kernel/locking/locktorture.c b/kernel/locking/locktorture.c
>index 9425aff089365..ed8e5baafe49f 100644
>--- a/kernel/locking/locktorture.c
>+++ b/kernel/locking/locktorture.c
>@@ -257,6 +257,61 @@ static struct lock_torture_ops spin_lock_irq_ops = {
>	.name		= "spin_lock_irq"
> };
>
>+#ifdef CONFIG_PREEMPT_RT
>+static DEFINE_RAW_SPINLOCK(torture_raw_spinlock);

How about leaving raw spinlocks regardless of preempt-rt, and instead
change the default lock (which is spin_lock) based on CONFIG_PREEMPT_RT
and use the raw one in that case?

Thanks,
Davidlohr