Re: [patch-2.3.40-pre6] kzalloc() (ala kmem_zalloc() of SVR4)

[Jamie Lokier]

Alan Cox wrote:
> > And on a multiprocessor system one idle processor can zero pages while
> > the other is consuming them.
> 
> Nice theory but the CPU->CPU traffic will kill you

I guess it's an open question, whether the CPU->CPU or memory->CPU
traffic on the busy processor takes longer than the local zeroing
followed by eventual CPU->memory flush of the unused portions of the
zero page.

-- Jamie