Re: [PATCH] Add per-process flag to control thp

[Kees Cook]

On Fri, Aug 2, 2013 at 1:34 PM, Alex Thorlton <athorlton@xxxxxxx> wrote:
>> What kind of workloads are you talking about?
>
> Our benchmarking team has a list of several of the SPEC OMP benchmarks
> that perform significantly better when THP is disabled. I tried to get
> the list but one of our servers is acting up and I can't get to it
> right now :/
>
>> What's wrong with madvise? Could you elaborate?
>
> The main issue with using madvise is that it's not an option with static
> binaries, but there are also some users who have legacy Fortran code
> that they're not willing/able to change.
>
>> And I think thp_disabled should be reset to 0 on exec.
>
> The main purpose for this getting carried down from the parent process
> is that we'd like to be able to have a userland program set this flag on
> itself, and then spawn off children who will also carry the flag.
> This allows us to set the flag for programs where we're unable to modify
> the code, thus resolving the issues detailed above.

This could be done with LD_PRELOAD for uncontrolled binaries. Though
it does seem sensible to make it act like most personality flags do
(i.e. inherited).

-Kees

-- 
Kees Cook
Chrome OS Security
--
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/