Re: [RFC 00/15] x86_64: Optimize percpu accesses

[Jeremy Fitzhardinge]

Christoph Lameter wrote:
> Well yes I agree these could be better if the fields would be moved out of the pda
> structure itself but then it wont be mechanical anymore and require more
> review.

Yes, but they'll have more value.  And if you do it as one variable per 
patch, then it should be easy to bisect should any problems arise.

>  But these are an important step because they allow us to get rid of the
> pda operations that do not exist for 32 bit.
>   

No, they don't help at all, because they convert X_pda(Y) (which doesn't 
exist) into x86_X_percpu(pda.Y) (which also doesn't exist).  They don't 
remove any #ifdef CONFIG_X86_64's.  If you're going to tromp all over 
the source, you may as well do the most useful thing in the first step.

   J
--
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/