[PATCH 1/2] v2 476: Set CCR2[DSTI] to prevent isync from flushing shadow TLB

[Dave Kleikamp]

Josh,
Please pull this patch.  I just found a bone-headed mistake that makes
the whole patch a no-op.  I'll need to fix it and put it through a bit
of testing before I can re-submit it.

The other patch in this series should be okay.

Thanks,
Shaggy

On Mon, 2010-09-27 at 16:56 -0500, Dave Kleikamp wrote:
> When the DSTI (Disable Shadow TLB Invalidate) bit is set in the CCR2
> register, the isync command does not flush the shadow TLB (iTLB & dTLB).
> 
> However, since the shadow TLB does not contain context information, we
> want the shadow TLB flushed in situations where we are switching context.
> In those situations, we explicitly clear the DSTI bit before performing
> isync, and set it again afterward.  We also need to do the same when we
> perform isync after explicitly flushing the TLB.
> 
> Signed-off-by: Dave Kleikamp <shaggy at linux.vnet.ibm.com>
> ---
>  arch/powerpc/include/asm/reg_booke.h  |    4 ++++
>  arch/powerpc/kernel/head_44x.S        |   25 +++++++++++++++++++++++++
>  arch/powerpc/mm/tlb_nohash_low.S      |   14 +++++++++++++-
>  arch/powerpc/platforms/44x/misc_44x.S |   26 ++++++++++++++++++++++++++
>  4 files changed, 68 insertions(+), 1 deletions(-)
> 

   --- snip ---

> --- a/arch/powerpc/kernel/head_44x.S
> +++ b/arch/powerpc/kernel/head_44x.S

Not only is this in the wrong place (non-47x initialization) but ...

> @@ -861,6 +877,15 @@ skpinv:	addi	r4,r4,1				/* Increment */
>  	isync
>  #endif /* CONFIG_PPC_EARLY_DEBUG_44x */
> 
> +BEGIN_MMU_FTR_SECTION
> +	mfspr   r3,SPRN_CCR2_476
> +	/* With CCR2(DSTI) set, isync does not invalidate the shadow TLB */
> +	oris	r3,r3,CCR2_476_DSTI at h
> +	rlwinm	r3,r3,0,~CCR2_476_DSTI

^^^ This instruction doesn't belong at all.  It clears the bit right
after setting it.  This one was just introduced removing the config
option, but it was in the wrong place all along.

> +	mtspr   SPRN_CCR2_476,r3
> +	isync
> +END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_47x)
> +
>  	/* Establish the interrupt vector offsets */
>  	SET_IVOR(0,  CriticalInput);
>  	SET_IVOR(1,  MachineCheck);

I wasn't diligent enough checking a year-old patch that I got back to
work on.  The code is very similar in two places and the patch applied
to the wrong section.

Thanks,
Shaggy
-- 
Dave Kleikamp
IBM Linux Technology Center