Re: BUG: spinlock trylock failure on UP, i.MX28 3.12.15-rt25

From: Peter Zijlstra <peterz@infradead.org>
To: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>,
	Stanislav Meduna <stano@meduna.org>,
	"linux-rt-users@vger.kernel.org" <linux-rt-users@vger.kernel.org>,
	Linux ARM Kernel <linux-arm-kernel@lists.infradead.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: Re: BUG: spinlock trylock failure on UP, i.MX28 3.12.15-rt25
Date: Wed, 23 Apr 2014 10:50:31 +0200	[thread overview]
Message-ID: <20140423085031.GI11096@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <53577559.9090501@linutronix.de>

On Wed, Apr 23, 2014 at 10:10:01AM +0200, Sebastian Andrzej Siewior wrote:
> On 04/22/2014 04:09 PM, Peter Zijlstra wrote:
> > On Tue, Apr 22, 2014 at 09:46:57AM -0400, Steven Rostedt wrote:
> >>>  According to lockdep a trylock should not fail on UP.
> > 
> > Oh!? Where does it say that? A trylock can fail at all times.
> 
> kernel/locking/spinlock_debug.c:
> 
> int do_raw_spin_trylock(raw_spinlock_t *lock)
>  {
>          int ret = arch_spin_trylock(&lock->raw_lock);
> 
>          if (ret)
>                  debug_spin_lock_after(lock);
>  #ifndef CONFIG_SMP
>          /*
>           * Must not happen on UP:
>           */
>          SPIN_BUG_ON(!ret, lock, "trylock failure on UP");
>  #endif
>          return ret;
>  }
> 
> How can a trylock (spinlock, not mutex) fail on UP? That would mean the
> lock is not interrupt safe.
> Unless, you attempt to take the lock from interrupt context via trylock
> while in general you take the spinlock in process context with
> interrupts enabled.

But that's not lockdep. That's the spinlock debugging code, entirely
different beasts.