linux-kernel - BUG: Re: [PATCH v2 3/3] clocksource: use clockevents_config_and

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

Date:	Tue, 23 Apr 2013 15:07:55 +0200
From:	Marc Kleine-Budde <mkl@...gutronix.de>
To:	Shawn Guo <shawn.guo@...aro.org>
CC:	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	Andres Salomon <dilinger@...ued.net>, arm@...nel.org,
	John Stultz <john.stultz@...aro.org>,
	Nicolas Ferre <nicolas.ferre@...el.com>,
	"kernel@...gutronix.de" <kernel@...gutronix.de>
Subject: BUG: Re: [PATCH v2 3/3] clocksource: use clockevents_config_and_register()
 where possible

Hello,

On 01/12/2013 12:50 PM, Shawn Guo wrote:
> The clockevent core is able to figure out the best mult and shift,
> calculate min_delta_ns and max_delta_ns, with the necessary info passed
> into clockevents_config_and_register().  Use this combined configure
> and register function where possible to make the codes less error prone
> and gain some positive diff stat.
> 
> Signed-off-by: Shawn Guo <shawn.guo@...aro.org>
> Cc: Andres Salomon <dilinger@...ued.net>
> Cc: Nicolas Ferre <nicolas.ferre@...el.com>
> Acked-by: Maxime Ripard <maxime.ripard@...e-electrons.com>

[...]

I just bisected a timer problem down to this commit.

> diff --git a/drivers/clocksource/tcb_clksrc.c b/drivers/clocksource/tcb_clksrc.c
> index 32cb929..8a61872 100644
> --- a/drivers/clocksource/tcb_clksrc.c
> +++ b/drivers/clocksource/tcb_clksrc.c
> @@ -157,7 +157,6 @@ static struct tc_clkevt_device clkevt = {
>  		.name		= "tc_clkevt",
>  		.features	= CLOCK_EVT_FEAT_PERIODIC
>  					| CLOCK_EVT_FEAT_ONESHOT,
> -		.shift		= 32,
>  		/* Should be lower than at91rm9200's system timer */
>  		.rating		= 125,
>  		.set_next_event	= tc_next_event,
> @@ -196,13 +195,9 @@ static void __init setup_clkevents(struct atmel_tc *tc, int clk32k_divisor_idx)
>  
>  	timer_clock = clk32k_divisor_idx;
>  
> -	clkevt.clkevt.mult = div_sc(32768, NSEC_PER_SEC, clkevt.clkevt.shift);
> -	clkevt.clkevt.max_delta_ns
> -		= clockevent_delta2ns(0xffff, &clkevt.clkevt);
> -	clkevt.clkevt.min_delta_ns = clockevent_delta2ns(1, &clkevt.clkevt) + 1;
                                                                           ^^^^^^

With this patch, the "+ 1" gets lost, resulting in a different
"min_delta_ns".

good:
setup_clkevents: shift=32 mult=140737 min=30518 max=1999976422

bad:
setup_clkevents: shift=32 mult=140737 min=30517 max=1999976422

This leads to problems with schedule_delayed_work() with a delay of
"1". These worksqueues seem to be hanging for several seconds. I'll
send a RFC patch.

>  	clkevt.clkevt.cpumask = cpumask_of(0);
>  
> -	clockevents_register_device(&clkevt.clkevt);
> +	clockevents_config_and_register(&clkevt.clkevt, 32768, 1, 0xffff);
>  
>  	setup_irq(irq, &tc_irqaction);
>  }
> 

Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


Download attachment "signature.asc" of type "application/pgp-signature" (264 bytes)