linux-kernel - Re: [PATCH] clockevents/drivers/cs5535: Un-break driver with 'set-state' interface

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Mon, 2 Jan 2017 15:04:54 +0530
From:   Viresh Kumar <viresh.kumar@...aro.org>
To:     David Gstir <david@...ma-star.at>
Cc:     linux-kernel@...r.kernel.org, daniel.lezcano@...aro.org,
        tglx@...utronix.de, richard@....at, stable@...r.kernel.org,
        Andres Salomon <dilinger@...ued.net>
Subject: Re: [PATCH] clockevents/drivers/cs5535: Un-break driver with
 'set-state' interface

On 28-12-16, 22:56, David Gstir wrote:
> Since migrating to the 'set-state' interface the cs5535 driver causes a
> crash after loading: Right after initialization, the IRQ handler
> (mfgpt_tick) is triggered with clock event device in detached state. This
> state not properly handled and causes a crash through NULL pointer
> dereference upon calling the clockevent's event_handler.
> 
> This patch fixes this by handling the detached state the same way the
> shutdown state is handled.
> 
> Fixes: 8f9327cbb6e8 ("clockevents/drivers/cs5535: Migrate to new 'set-state' interface")
> Cc: stable@...r.kernel.org
> Cc: Andres Salomon <dilinger@...ued.net>
> Cc: Viresh Kumar <viresh.kumar@...aro.org>
> Signed-off-by: David Gstir <david@...ma-star.at>
> ---
>  drivers/clocksource/cs5535-clockevt.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/clocksource/cs5535-clockevt.c b/drivers/clocksource/cs5535-clockevt.c
> index 9a7e37cf56b0..649e0cd90805 100644
> --- a/drivers/clocksource/cs5535-clockevt.c
> +++ b/drivers/clocksource/cs5535-clockevt.c
> @@ -117,7 +117,8 @@ static irqreturn_t mfgpt_tick(int irq, void *dev_id)
>  	/* Turn off the clock (and clear the event) */
>  	disable_timer(cs5535_event_clock);
>  
> -	if (clockevent_state_shutdown(&cs5535_clockevent))
> +	if (clockevent_state_shutdown(&cs5535_clockevent) ||
> +			clockevent_state_detached(&cs5535_clockevent))
>  		return IRQ_HANDLED;
>  
>  	/* Clear the counter */

Sorry for breaking it, but it looks we have unearthed a bug because of
my patch.

I wouldn't enable the IRQ unless the clockevents core has asked for
it, i.e. by calling set_state_periodic() or set_state_oneshot().

The driver is currently enabling the IRQ from its init code and that's
where the problem is IMHO.

-- 
viresh