[U-Boot] [PATCH 3/6] armv8/mmu: Clean up TCR programming

[Marc Zyngier]

Hi Albert,

On 02/07/15 22:06, Albert ARIBAUD wrote:
> Hello Marc,
> 
> On Fri, 20 Mar 2015 18:13:26 +0000, Marc Zyngier <marc.zyngier at arm.com>
> wrote:
>> On 20/03/15 11:47, Thierry Reding wrote:
>>> From: Thierry Reding <treding at nvidia.com>
>>>
>>> Use the inner shareable attribute for memory, which makes more sense
>>> considering that this code is called when caches are being enabled.
>>>
>>> While at it, fix the values for the shareability attribute field to
>>> match the documentation.
>>>
>>> Cc: Albert Aribaud <albert.u.boot at aribaud.net>
>>> Cc: Marc Zyngier <marc.zyngier at arm.com>
>>> Signed-off-by: Thierry Reding <treding at nvidia.com>
>>> ---
>>>  arch/arm/include/asm/armv8/mmu.h | 8 ++++----
>>>  1 file changed, 4 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/arch/arm/include/asm/armv8/mmu.h b/arch/arm/include/asm/armv8/mmu.h
>>> index 4b9cb5296572..6d42f5533a74 100644
>>> --- a/arch/arm/include/asm/armv8/mmu.h
>>> +++ b/arch/arm/include/asm/armv8/mmu.h
>>> @@ -93,8 +93,8 @@
>>>  #define TCR_ORGN_WBNWA		(3 << 10)
>>>  #define TCR_ORGN_MASK		(3 << 10)
>>>  #define TCR_SHARED_NON		(0 << 12)
>>> -#define TCR_SHARED_OUTER	(1 << 12)
>>> -#define TCR_SHARED_INNER	(2 << 12)
>>> +#define TCR_SHARED_OUTER	(2 << 12)
>>> +#define TCR_SHARED_INNER	(3 << 12)
>>>  #define TCR_TG0_4K		(0 << 14)
>>>  #define TCR_TG0_64K		(1 << 14)
>>>  #define TCR_TG0_16K		(2 << 14)
>>> @@ -102,9 +102,9 @@
>>>  #define TCR_EL2_IPS_BITS	(3 << 16)	/* 42 bits physical address */
>>>  #define TCR_EL3_IPS_BITS	(3 << 16)	/* 42 bits physical address */
>>>  
>>> -/* PTWs cacheable, inner/outer WBWA and non-shareable */
>>> +/* PTWs cacheable, inner/outer WBWA and inner shareable */
>>>  #define TCR_FLAGS		(TCR_TG0_64K |		\
>>> -				TCR_SHARED_NON |	\
>>> +				TCR_SHARED_INNER |	\
>>>  				TCR_ORGN_WBWA |		\
>>>  				TCR_IRGN_WBWA |		\
>>>  				TCR_T0SZ(VA_BITS))
>>>
>>
>> Acked-by: Marc Zyngier <marc.zyngier at arm.com>
>>
>> One thing though: the architecture doesn't mandate 64k pages to be
>> implemented by the HW. Actually, it doesn't mandate any particular page
>> size, you just have to implement at least one (4k, 16k or 64k).
>>
>> It would be good to test if 64k pages are implemented (by testing
>> ID_AA64MMFR0_EL1) and not try to enable caches if not, possibly
>> displaying a warning for the unsuspecting u-boot hacker.
> 
> So Marc, is this a request for a change, or is the patch applicable as
> it is?

No, the patch is extremely valuable as it is, and should be applied. My
remark is probably not applicable for generally available CPUs, and
could be implemented as a later improvement.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...