Re: [RFC PATCH 3/6] smp: Run functions concurrently in smp_call_function_many()

From: Nadav Amit <namit@vmware.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>, Andy Lutomirski <luto@kernel.org>,
	Borislav Petkov <bp@alien8.de>,
	LKML <linux-kernel@vger.kernel.org>,
	Rik van Riel <riel@surriel.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Josh Poimboeuf <jpoimboe@redhat.com>
Subject: Re: [RFC PATCH 3/6] smp: Run functions concurrently in smp_call_function_many()
Date: Mon, 27 May 2019 17:39:35 +0000	[thread overview]
Message-ID: <57E94109-CBE8-43BA-98FF-646E8C9EE8A2@vmware.com> (raw)
In-Reply-To: <20190527091534.GS2623@hirez.programming.kicks-ass.net>

> On May 27, 2019, at 2:15 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> 
>> +		/*
>> +		 * Choose the most efficient way to send an IPI. Note that the
>> +		 * number of CPUs might be zero due to concurrent changes to the
>> +		 * provided mask or cpu_online_mask.
>> +		 */
> 
> Since we have preemption disabled here, I don't think online mask can
> shrink, cpu-offline uses stop_machine().

Right. So I’ll update the comment, but IIUC the provided mask might still
change, so I’ll leave the rest of the comment and the code as is.

>> +		if (nr_cpus == 1)
>> +			arch_send_call_function_single_ipi(last_cpu);
>> +		else if (likely(nr_cpus > 1))
>> +			arch_send_call_function_ipi_mask(cfd->cpumask_ipi);
>> +	}