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;
Date:   Wed, 18 Jan 2023 10:22:26 +0100
From:   Johan Hovold <johan@kernel.org>
To:     Thomas Gleixner <tglx@linutronix.de>
Cc:     Johan Hovold <johan+linaro@kernel.org>,
        Marc Zyngier <maz@kernel.org>, x86@kernel.org,
        platform-driver-x86@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, linux-mips@vger.kernel.org,
        linux-kernel@vger.kernel.org, Hsin-Yi Wang <hsinyi@chromium.org>,
        Mark-PK Tsai <mark-pk.tsai@mediatek.com>
Subject: Re: [PATCH v4 06/19] irqdomain: Drop revmap mutex
Message-ID: <Y8e6Us0Qgt0p5S4R@hovoldconsulting.com>
References: <20230116135044.14998-1-johan+linaro@kernel.org>
 <20230116135044.14998-7-johan+linaro@kernel.org>
 <871qnslut3.ffs@tglx>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <871qnslut3.ffs@tglx>
Precedence: bulk

On Tue, Jan 17, 2023 at 10:23:20PM +0100, Thomas Gleixner wrote:
> On Mon, Jan 16 2023 at 14:50, Johan Hovold wrote:
> > The global irq_domain_mutex is now held in all paths that update the
> > revmap structures so there is no longer any need for the revmap mutex.
> 
> This can also go after the 3rd race fix, but ...
> 
> >  static void irq_domain_clear_mapping(struct irq_domain *domain,
> >  				     irq_hw_number_t hwirq)
> >  {
> > +	lockdep_assert_held(&irq_domain_mutex);
> 
> these lockdep asserts want to be part of the [dis]association race
> fixes. They are completely unrelated to the removal of the revmap_mutex.

No, they are very much related to the removal of the revmap_mutex. These
functions deal with the revmap structures which before this patch were
clearly only modified with the revmap_mutex held.

The lockdep assert is here to guarantee that my claim that all current
(and future) paths that end up modifying these structures do so under
the irq_domain_mutex instead.

> Your race fixes change the locking and you want to ensure that all
> callers comply right there, no?

I want to make sure that all callers of these function comply, yes.
That's why the asserts belong in this patch.

Johan