Re: [PATCH 03/10] smp,cpumask: introduce on_each_cpu_cond_mask

From: Rik van Riel <riel@surriel.com>
To: Andy Lutomirski <luto@amacapital.net>
Cc: Andy Lutomirski <luto@kernel.org>,
	LKML <linux-kernel@vger.kernel.org>,
	kernel-team <kernel-team@fb.com>,
	Peter Zijlstra <peterz@infradead.org>, X86 ML <x86@kernel.org>,
	Vitaly Kuznetsov <vkuznets@redhat.com>,
	Ingo Molnar <mingo@kernel.org>, Mike Galbraith <efault@gmx.de>,
	Dave Hansen <dave.hansen@intel.com>,
	will.daecon@arm.com, Catalin Marinas <catalin.marinas@arm.com>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>
Subject: Re: [PATCH 03/10] smp,cpumask: introduce on_each_cpu_cond_mask
Date: Sun, 29 Jul 2018 13:39:09 -0400	[thread overview]
Message-ID: <1532885949.28585.20.camel@surriel.com> (raw)
In-Reply-To: <E710FBA5-CC5E-4941-ACBF-4AB3424F1F68@amacapital.net>

[-- Attachment #1: Type: text/plain, Size: 1785 bytes --]

On Sun, 2018-07-29 at 08:36 -0700, Andy Lutomirski wrote:
> On Jul 29, 2018, at 5:00 AM, Rik van Riel <riel@surriel.com> wrote:
> 
> > On Sat, 2018-07-28 at 19:57 -0700, Andy Lutomirski wrote:
> > > On Sat, Jul 28, 2018 at 2:53 PM, Rik van Riel <riel@surriel.com>
> > > wrote:
> > > > Introduce a variant of on_each_cpu_cond that iterates only over
> > > > the
> > > > CPUs in a cpumask, in order to avoid making callbacks for every
> > > > single
> > > > CPU in the system when we only need to test a subset.
> > > Nice.
> > > Although, if you want to be really fancy, you could optimize this
> > > (or
> > > add a variant) that does the callback on the local CPU in
> > > parallel
> > > with the remote ones.  That would give a small boost to TLB
> > > flushes.
> > 
> > The test_func callbacks are not run remotely, but on
> > the local CPU, before deciding who to send callbacks
> > to.
> > 
> > The actual IPIs are sent in parallel, if the cpumask
> > allocation succeeds (it always should in many kernel
> > configurations, and almost always in the rest).
> > 
> 
> What I meant is that on_each_cpu_mask does:
> 
> smp_call_function_many(mask, func, info, wait);
> if (cpumask_test_cpu(cpu, mask)) {
>    unsigned long flags;
>    local_irq_save(flags); func(info);
>    local_irq_restore(flags);
> }
> 
> So it IPIs all the remote CPUs in parallel, then waits, then does the
> local work.  In principle, the local flush could be done after
> triggering the IPIs but before they all finish.

Sure, moving the function call for the local CPU
into smp_call_function_many might be a nice optimization.

A quick grep suggests it touch stuff all over the tree,
so it could be a nice Outreachy intern project :)

-- 
All Rights Reversed.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]