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:   Tue, 16 May 2023 15:18:11 +0200
From:   Oleg Nesterov <oleg@redhat.com>
To:     Mike Christie <michael.christie@oracle.com>
Cc:     Linus Torvalds <torvalds@linux-foundation.org>,
        Christian Brauner <brauner@kernel.org>,
        Thorsten Leemhuis <linux@leemhuis.info>,
        nicolas.dichtel@6wind.com,
        Linux kernel regressions list <regressions@lists.linux.dev>,
        hch@infradead.org, stefanha@redhat.com, jasowang@redhat.com,
        mst@redhat.com, sgarzare@redhat.com,
        virtualization@lists.linux-foundation.org, ebiederm@xmission.com,
        konrad.wilk@oracle.com, linux-kernel@vger.kernel.org,
        Jens Axboe <axboe@kernel.dk>
Subject: Re: [PATCH v11 8/8] vhost: use vhost_tasks for worker threads
Message-ID: <20230516131810.GA28890@redhat.com>
References: <aba6cca4-e66c-768f-375c-b38c8ba5e8a8@6wind.com>
 <CAHk-=wgadfsCnKHLON7op=Qs5t3w3PVz5ZDbvbKsfb=yBg=yjQ@mail.gmail.com>
 <78c5e150-26cf-7724-74ee-4a0b16b944b1@oracle.com>
 <48842e92-835e-bc3f-7118-48b8f415f532@leemhuis.info>
 <CAHk-=wicXCNR1bBioaAiBqEDgv-CoQD9z2RsM2VOTPzkK0xr8g@mail.gmail.com>
 <20230515-vollrausch-liebgeworden-2765f3ca3540@brauner>
 <CAHk-=wgXJ5VS1iBkfsG=HDjsyhn5XYDKt5xhQcNuz-e7VKyg8A@mail.gmail.com>
 <122b597e-a5fa-daf7-27bb-6f04fa98d496@oracle.com>
 <CAHk-=wgTpRDwS+F0Gd6+bM+TEh+x4Aiz8EBDTnTm3Q4TFdWOww@mail.gmail.com>
 <8bfb7d1d-f7d0-94ca-4777-e31a2003027a@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <8bfb7d1d-f7d0-94ca-4777-e31a2003027a@oracle.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Precedence: bulk

On 05/15, Mike Christie wrote:
>
> Oleg and Christian,
>
>
> Below is an updated patch that doesn't check for PF_USER_WORKER in the
> signal.c code and instead will check for if we have blocked the signal.

Looks like I need to read the whole series... will try tomorrow.

> --- a/kernel/fork.c
> +++ b/kernel/fork.c
> @@ -2336,8 +2336,15 @@ __latent_entropy struct task_struct *copy_process(
>  	p->flags &= ~PF_KTHREAD;
>  	if (args->kthread)
>  		p->flags |= PF_KTHREAD;
> -	if (args->user_worker)
> +	if (args->user_worker) {
> +		/*
> +		 * User worker are similar to io_threads but they do not
> +		 * support signals and cleanup is driven via another kernel
> +		 * interface so even SIGKILL is blocked.
> +		 */
>  		p->flags |= PF_USER_WORKER;
> +		siginitsetinv(&p->blocked, 0);

I never liked the fact that io-threads block the signals, this adds
another precedent... OK, this needs another discussion.

> +static void try_set_pending_sigkill(struct task_struct *t)
> +{
> +	/*
> +	 * User workers don't support signals and their exit is driven through
> +	 * their kernel layer, so by default block even SIGKILL.
> +	 */
> +	if (sigismember(&t->blocked, SIGKILL))
> +		return;
> +
> +	sigaddset(&t->pending.signal, SIGKILL);
> +	signal_wake_up(t, 1);
> +}

so why do you need this? to avoid fatal_signal_pending() or signal_pending() ?

In the latter case this change is not enough.

Oleg.