Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 11 Aug 2023 12:14:14 +0100
From:   Will Deacon <will@kernel.org>
To:     Yicong Yang <yangyicong@huawei.com>
Cc:     jonathan.cameron@huawei.com, mark.rutland@arm.com,
        hejunhao3@huawei.com, prime.zeng@hisilicon.com,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        linuxarm@huawei.com, yangyicong@hisilicon.com
Subject: Re: [PATCH] drivers/perf: hisi: Schedule perf session according to
 locality
Message-ID: <20230811111414.GC6993@willie-the-truck>
References: <20230808125147.2080-1-yangyicong@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230808125147.2080-1-yangyicong@huawei.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On Tue, Aug 08, 2023 at 08:51:47PM +0800, Yicong Yang wrote:
> From: Yicong Yang <yangyicong@hisilicon.com>
> 
> The PCIe PMUs locate on different NUMA node but currently we don't
> consider it and likely stack all the sessions on the same CPU:
> 
> [root@localhost tmp]# cat /sys/devices/hisi_pcie*/cpumask
> 0
> 0
> 0
> 0
> 0
> 0
> 
> This can be optimize a bit to use a local CPU for the PMU.
> 
> Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
> ---
>  drivers/perf/hisilicon/hisi_pcie_pmu.c | 15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/perf/hisilicon/hisi_pcie_pmu.c b/drivers/perf/hisilicon/hisi_pcie_pmu.c
> index e10fc7cb9493..60ecf469782b 100644
> --- a/drivers/perf/hisilicon/hisi_pcie_pmu.c
> +++ b/drivers/perf/hisilicon/hisi_pcie_pmu.c
> @@ -665,7 +665,7 @@ static int hisi_pcie_pmu_online_cpu(unsigned int cpu, struct hlist_node *node)
>  	struct hisi_pcie_pmu *pcie_pmu = hlist_entry_safe(node, struct hisi_pcie_pmu, node);
>  
>  	if (pcie_pmu->on_cpu == -1) {
> -		pcie_pmu->on_cpu = cpu;
> +		pcie_pmu->on_cpu = cpumask_local_spread(0, dev_to_node(&pcie_pmu->pdev->dev));
>  		WARN_ON(irq_set_affinity(pcie_pmu->irq, cpumask_of(cpu)));

Hmm, this is a bit weird now, because the interrupt is affine to a different
CPU from the one you've chosen. Are you sure that's ok? When the offline
notifier picks a new target, it moves the irq as well.

Will