[PATCH] powerpc: Don't use toc in decrementer_iSeries_masked

[David Gibson]

On Mon, Feb 06, 2006 at 05:34:14PM +1100, Michael Ellerman wrote:
> Since 404849bbd2bfd62e05b36f4753f6e1af6050a824 we've been using
> LOAD_REG_ADDRBASE, which uses the toc pointer, in decrementer_iSeries_masked.
> 
> This can explode if we take the decrementer interrupt while we're in a module,
> because the toc pointer in r2 will be the module's toc pointer.
> 
> Instead do an immediate load. I'm not sure if we really need the trickery in
> here, what do people think?
> 
>  arch/powerpc/kernel/head_64.S |    9 +++++++--
>  1 files changed, 7 insertions(+), 2 deletions(-)
> 
> Index: iseries/arch/powerpc/kernel/head_64.S
> ===================================================================
> --- iseries.orig/arch/powerpc/kernel/head_64.S
> +++ iseries/arch/powerpc/kernel/head_64.S
> @@ -752,8 +752,13 @@ decrementer_iSeries_masked:
>  	li	r11,1
>  	ld	r12,PACALPPACAPTR(r13)
>  	stb	r11,LPPACADECRINT(r12)
> -	LOAD_REG_ADDRBASE(r12,tb_ticks_per_jiffy)
> -	lwz	r12,ADDROFF(tb_ticks_per_jiffy)(r12)
> +	/* We have no TOC pointer in here, so we can't load tb_ticks_per_jiffy
> +	 * using the usual macros. We want to be fast, so we assume the top
> +	 * word of the lppaca pointer is the same as the top word of
> +	 * &tb_ticks_per_jiffy.
> +	 */

you need a
	clrrdi	r12,r12,32
here, because r12's low word may not contain zero.

> +	oris    r12,r12,tb_ticks_per_jiffy at h
> +	lwz	r12,tb_ticks_per_jiffy at l(r12)
>  	mtspr	SPRN_DEC,r12
>  	/* fall through */
>  
> _______________________________________________
> Linuxppc64-dev mailing list
> Linuxppc64-dev at ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc64-dev
> 

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson