Re: [PATCH V7 1/3] MIPS: Rearrange PTE bits into fixed positions.

[David Daney <ddaney.cavm@gmail.com>]

On 02/26/2015 07:38 PM, Steven J. Hill wrote:
> On 02/26/2015 06:51 PM, David Daney wrote:
>>
>> That's not really what I meant in my previous response on the subject.
>> When I said:
>>
>>      Why not just use RI for everything, instead of taking up two bits
>>      to represent a single binary concept?
>>
>>      For the case where there is no RI hardware active, it is a purely
>>      software bit and you can easily invert the meaning and just have a
>>      _PAGE_NO_READ bit.
>>
>> I envisioned something like:
>>
>>      64-bit, all revisions:    CCC D V G RI XI [S H] M A W P
>>      32-bit, all revisions:    CCC D V G RI XI M A W P
>>
> Which is what I implemented.

I think there is still misunderstanding.

Your patches leave us with definitions for *both* _PAGE_READ *and* 
_PAGE_NO_READ defined in the source code.  My suggestion was to 
eliminate all vestiges of _PAGE_READ and _PAGE_READ_SHIFT, and unify all 
variants to use _PAGE_NO_READ

> I now only use one bit that functions
> either as _PAGE_READ or _PAGE_READ_ONLY depending on the RI/XI
> functionality present. Did you bother to read the code and understand
> it, or just look at the commit message?

I did read it, see above.