Messages in this thread |  | | | Subject | Re: RFC: remove __read_mostly | | From | Harvey Harrison <> | | Date | Thu, 13 Dec 2007 14:44:15 -0800 | |
On Thu, 2007-12-13 at 14:32 -0800, David Miller wrote:
> From: Adrian Bunk <bunk@kernel.org>
> Date: Thu, 13 Dec 2007 23:20:44 +0100
>
> > My question is:
> > Is there anywhere in the kernel a case where __read_mostly brings a
> > measurable improvement or can it be removed?
>
> Yes, on SMP when read-mostly objects share cache lines
> with other objects which are frequently written to.
>
> That is the whole reason we created __read_mostly
I'm curious if anyone has been looking into replacing the __read_mostly
approach with Mathieu's immediate values patchset. Wouldn't they solve
the cacheline sharing as well (perhaps more eficiently even with trading
some icache for dcache)?
Harvey
|  |