Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
User-Agent: Microsoft-MacOutlook/16.42.20101102
Date:   Sun, 04 Jul 2021 18:18:19 +0800
Subject: Re: [PATCH] kretprobe scalability improvement
From:   Matt Wu <wuqiang.matt@bytedance.com>
To:     Christoph Hellwig <hch@infradead.org>
CC:     <naveen.n.rao@linux.ibm.com>, <anil.s.keshavamurthy@intel.com>,
        <davem@davemloft.net>, <mhiramat@kernel.org>, <mingo@kernel.org>,
        <peterz@infradead.org>, <linux-kernel@vger.kernel.org>,
        <mattwu@163.com>
Message-ID: <C46C9460-7FD6-4F6F-9603-84E2A8BA8D8F@bytedance.com>
Thread-Topic: [PATCH] kretprobe scalability improvement
Mime-version: 1.0
Content-type: text/plain;
        charset="UTF-8"
Content-transfer-encoding: 7bit
Precedence: bulk

On 2021/7/4, 5:17 PM, "Christoph Hellwig" <hch@infradead.org> wrote:

    Would it make sense to just reuse kernel/bpf/percpu_freelist.c for
    kretprobes?

Thanks for the info. bpf/percpu_freelist does meet the requirements of kretprobe_instances management and is more compact in memory. But  the spinlock usage (through locally) looks a bit heavy since it must disable interrupts on local cpu.

I will perform a test to collect some data for comparison.

Regards,
Matt Wu