Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20191029092423.17825-1-haokexin@gmail.com> <CAHk-=wjU9ASiPYFqmGJtOqG-0KtuNtu-aNPPY4M1AbcPdrfz7A@mail.gmail.com>
 <20191029140921.GF23786@pek-khao-d2.corp.ad.wrs.com>
In-Reply-To: <20191029140921.GF23786@pek-khao-d2.corp.ad.wrs.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Tue, 29 Oct 2019 16:23:33 +0100
Message-ID: <CAHk-=wja287VnDKdftnQUYYe=YqX5Lz-iweA-RkcG3EGS2sD6w@mail.gmail.com>
Subject: Re: [PATCH] dump_stack: Avoid the livelock of the dump_lock
To:     Kevin Hao <haokexin@gmail.com>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Oct 29, 2019 at 3:09 PM Kevin Hao <haokexin@gmail.com> wrote:
>
> Do you mean something like this?

Yeah, except I'd make it be something like

                local_irq_restore(flags);
-               cpu_relax();
+               do { cpu_relax(); } while (atomic_read(&dump_lock) != -1);
                goto retry;

instead, ie doing the cpu_relax() inside that loop.

Obviously untested.

          Linus