Re: [PATCH 7/7] sparc64: Add function graph tracer support.

From: Frederic Weisbecker <fweisbec@gmail.com>
To: sparclinux@vger.kernel.org
Subject: Re: [PATCH 7/7] sparc64: Add function graph tracer support.
Date: Mon, 19 Apr 2010 20:37:24 +0000	[thread overview]
Message-ID: <20100419203722.GA8811@nowhere> (raw)
In-Reply-To: <20100412.234300.212396783.davem@davemloft.net>

On Mon, Apr 19, 2010 at 12:56:41PM -0700, David Miller wrote:
> From: Frederic Weisbecker <fweisbec@gmail.com>
> Date: Mon, 19 Apr 2010 21:52:47 +0200
> 
> > On Mon, Apr 19, 2010 at 01:15:55AM -0700, David Miller wrote:
> >> From: David Miller <davem@davemloft.net>
> >> Date: Mon, 19 Apr 2010 00:56:13 -0700 (PDT)
> >> 
> >> > Another idea is to decrement the ->disabled counter after the
> >> > local_irq_restore().  Yes, we might lose IRQ handler traces which
> >> > occur between the local_irq_restore() and the counter decrment, but we
> >> > would also be completely immune to recursion problems.
> >> 
> >> So I just gave this a try and I still get the crashes I've
> >> been seeing with the non-graph function tracer, oh well :-)
> > 
> > 
> > Are you using the pause_graph_tracing() thing on the function
> > tracer? If so it won't work as it only applies to the function
> > graph tracer.
> 
> No, I simply moved the atomic_dec() below the local_irq_restore()
> as stated in the first paragraph above :-)

Ah right, read too fast :)