linux-kernel - Re: [PATCH] Remove CLOCK_TICK

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]

Message-ID: <1312539672.2764.41.camel@work-vm>
Date:	Fri, 05 Aug 2011 03:21:12 -0700
From:	john stultz <johnstul@...ibm.com>
To:	dsaxena@...xity.net
Cc:	Thomas Gleixner <tglx@...utronix.de>, linux-kernel@...r.kernel.org,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	patches@...aro.org
Subject: Re: [PATCH] Remove CLOCK_TICK_RATE from ACTHZ definition

On Wed, 2011-08-03 at 17:00 -0700, Deepak Saxena wrote:
> As per https://lkml.org/lkml/2011/2/21/323, we remove
> CLOCK_TICK_RATE from the definition of ACTHZ as the
> majority of CLOCK_TICK_RATE values are no longer valid
> or needed for proper operation of the time code.
> 
> Signed-off-by: Deepak Saxena <dsaxena@...aro.org>
> ---
>  include/linux/jiffies.h |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/include/linux/jiffies.h b/include/linux/jiffies.h
> index f97672a..d232f00 100644
> --- a/include/linux/jiffies.h
> +++ b/include/linux/jiffies.h
> @@ -55,7 +55,7 @@
>                               + ((((NOM) % (DEN)) << (LSH)) + (DEN) / 2) / (DEN))
> 
>  /* HZ is the requested value. ACTHZ is actual HZ ("<< 8" is for accuracy) */
> -#define ACTHZ (SH_DIV (CLOCK_TICK_RATE, LATCH, 8))
> +#define ACTHZ (HZ << 8)
> 
>  /* TICK_NSEC is the time between ticks in nsec assuming real ACTHZ */
>  #define TICK_NSEC (SH_DIV (1000000UL * 1000, ACTHZ, 8))

So I'm hesitant about this patch, since the whole point of ACTHZ is to
provide a higher precision value for the actual irq frequency (as
hardware won't be able to do exactly 100/250/300/1000HZ).

One could argue that with highres timers, jiffies is usually updated by
the timekeeping code, which for the most part isn't affected by irq freq
error. So we could reasonably forgo this minor error correction, only
hurting systems that use the jiffies clocksource, or require
ARCH_USES_GETTIMEOFFSET.

However, we would probably want to kill off all the ACTHZ usage as well
since we're at that point gaining nothing from the extra complexity.

thanks
-john




--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/