linux-kernel - Re: [BUG,2.6.28,s390] Fails to boot in Hercules S/390 emulator

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]

Message-ID: <alpine.LFD.2.00.0903120201040.29264@localhost.localdomain>
Date:	Thu, 12 Mar 2009 02:30:50 +0100 (CET)
From:	Thomas Gleixner <tglx@...utronix.de>
To:	john stultz <johnstul@...ibm.com>
cc:	Frans Pop <elendil@...net.nl>, linux-s390@...r.kernel.org,
	Roman Zippel <zippel@...ux-m68k.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [BUG,2.6.28,s390] Fails to boot in Hercules S/390 emulator

On Wed, 11 Mar 2009, john stultz wrote:
> For a cleaner version, could you try the following, against 2.6.29-git
> with no other modification?

cleaner ?

> xtime_nsec is expected at times to be negative. Instead of trying to
> handle all the shifting properly via casts, define it as a s64 instead
> of a u64.
> 
> NOT FOR INCLUSION
> Signed-off-by: John Stultz <johnstul@...ibm.com>
> 
> diff --git a/include/linux/clocksource.h b/include/linux/clocksource.h
> index f88d32f..e217000 100644
> --- a/include/linux/clocksource.h
> +++ b/include/linux/clocksource.h
> @@ -86,7 +86,7 @@ struct clocksource {
>  	 * more than one cache line.
>  	 */
>  	cycle_t cycle_last ____cacheline_aligned_in_smp;
> -	u64 xtime_nsec;
> +	s64 xtime_nsec;
>  	s64 error;
>  	struct timespec raw_time;
>  
> diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c
> index 900f1b6..387be3c 100644
> --- a/kernel/time/timekeeping.c
> +++ b/kernel/time/timekeeping.c
> @@ -546,7 +546,7 @@ void update_wall_time(void)
>  	/* store full nanoseconds into xtime after rounding it up and
>  	 * add the remainder to the error difference.
>  	 */
> -	xtime.tv_nsec = ((s64)clock->xtime_nsec >> clock->shift) + 1;
> +	xtime.tv_nsec = (clock->xtime_nsec >> clock->shift) + 1;
>  	clock->xtime_nsec -= (s64)xtime.tv_nsec << clock->shift;
>  	clock->error += clock->xtime_nsec << (NTP_SCALE_SHIFT - clock->shift);

This code sequence does:

   xtime.tv_nsec = (clock->xtime_nsec >> clock->shift) + 1;
   clock->xtime_nsec -= xtime.tv_nsec << clock->shift;

Lets substitute a bit:

   a = xtime.tv_nsec
   b = clock->xtime_nsec
   c = clock->shift
   r = result (which ends up in b aka clock->xtime_nsec again for the next round)

=> a = (b >> c) + 1
   r = b - (a << c)

   b >> c = b / 2^c
   a << c = a * 2^c

=> a = (b / 2^c) + 1
   r = b - (a * 2^c)

=> r = b - ((b / 2^c) + 1) * 2^c
   r = b - ((2^c * b / 2^c) + 2^c)
   r = b - (2^c * b / 2^c) - 2^c
   r = b - b - 2^c
   r = -2^c
   
=> r = -(1 << c)

So the whole business boils down to:

   clock->xtime_nsec = -(1 << clock->shift);

Famous last words (see also arch/x86/kernel/tsc.c and
arch/x86/include/asm/timer.h):

  -johnstul@...ibm.com "math is hard, lets go shopping!"

Thanks,

	tglx
--
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/