Re: pcpu allocator on large NUMA machines

From	Tejun Heo <tj@kernel.org>
Newsgroups	linux.kernel
Subject	Re: pcpu allocator on large NUMA machines
Date	2017-07-25 18:50 +0200
Message-ID	<u7bbY-5du-7@gated-at.bofh.it> (permalink)
References	<u6LUe-5yD-15@gated-at.bofh.it> <u6M3T-5Cy-3@gated-at.bofh.it> <u6MwV-655-1@gated-at.bofh.it> <u6WPD-4Ir-3@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

Hello, Michael.

On Tue, Jul 25, 2017 at 11:26:03AM +1000, Michael Ellerman wrote:
> I don't think we want to stop using embed first chunk unless we have to.
> 
> We have code that accesses percpu variables in real mode (with the MMU
> off), and that wouldn't work easily if the first chunk wasn't in the
> linear mapping. So it's not just an optimisation for us.
> 
> We can fairly easily make the vmalloc space 56T, and I'm working on a
> patch to make it ~500T on newer machines.

Yeah, the only constraint is the size of vmalloc area in relation to
the maximum spread across NUMA regions.  If the vmalloc space can be
made bigger, that'd be the best option.  As the area percpu allocator
actually uses is very small comparatively, it doesn't have to be a lot
larger either.

Thanks.

-- 
tejun

Back to linux.kernel | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

pcpu allocator on large NUMA machines Michal Hocko <mhocko@kernel.org> - 2017-07-24 15:50 +0200
  Re: pcpu allocator on large NUMA machines Tejun Heo <tj@kernel.org> - 2017-07-24 16:00 +0200
    Re: pcpu allocator on large NUMA machines Michal Hocko <mhocko@kernel.org> - 2017-07-24 16:30 +0200
      Re: pcpu allocator on large NUMA machines Michael Ellerman <mpe@ellerman.id.au> - 2017-07-25 03:30 +0200
        Re: pcpu allocator on large NUMA machines Tejun Heo <tj@kernel.org> - 2017-07-25 18:50 +0200

csiph-web