Re: [PATCH 03/15] arm64: Introduce helpers for page table levels

[Marc Zyngier]

On 07/10/15 09:26, Christoffer Dall wrote:
> Hi Suzuki,
> 
> On Tue, Sep 15, 2015 at 04:41:12PM +0100, Suzuki K. Poulose wrote:
>> From: "Suzuki K. Poulose" <suzuki.poulose@xxxxxxx>
>>
>> Introduce helpers for finding the number of page table
>> levels required for a given VA width, shift for a particular
>> page table level.
>>
>> Convert the existing users to the new helpers. More users
>> to follow.
>>
>> Cc: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
>> Cc: Mark Rutland <mark.rutland@xxxxxxx>
>> Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
>> Cc: Will Deacon <will.deacon@xxxxxxx>
>> Signed-off-by: Suzuki K. Poulose <suzuki.poulose@xxxxxxx>
>> Reviewed-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
>> Tested-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
>> ---
>>  arch/arm64/include/asm/pgtable-hwdef.h |   15 ++++++++++++---
>>  1 file changed, 12 insertions(+), 3 deletions(-)
>>
>> diff --git a/arch/arm64/include/asm/pgtable-hwdef.h b/arch/arm64/include/asm/pgtable-hwdef.h
>> index 24154b0..ce18389 100644
>> --- a/arch/arm64/include/asm/pgtable-hwdef.h
>> +++ b/arch/arm64/include/asm/pgtable-hwdef.h
>> @@ -16,13 +16,21 @@
>>  #ifndef __ASM_PGTABLE_HWDEF_H
>>  #define __ASM_PGTABLE_HWDEF_H
>>  
>> +/*
>> + * Number of page-table levels required to address 'va_bits' wide
>> + * address, without section mapping
>> + */
>> +#define ARM64_HW_PGTABLE_LEVELS(va_bits) (((va_bits) - 4) / (PAGE_SHIFT - 3))
> 
> I don't understand the '(va_bits) - 4' here, can you explain it (and add a
> comment to that effect) ?

I just had a chat with Catalin, who did shed some light on this.
It all has to do with rounding up. What you would like to have here is:

#define ARM64_HW_PGTABLE_LEVELS(va_bits) DIV_ROUND_UP(va_bits - PAGE_SHIFT, PAGE_SHIFT - 3)

where (va_bits - PAGE_SHIFT) is the total number of bits we deal
with during a page table walk, and (PAGE_SHIFT - 3) is the number
of bits we deal with per level.

The clue is in how DIV_ROUND_UP is written:

#define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d))

which gives you Suzuki's magic formula.

I'd vote for the DIV_ROUND_UP(), which will make things a lot more readable.

	M.
-- 
Jazz is not dead. It just smells funny...
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/