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;
From:   "Eric W. Biederman" <ebiederm@xmission.com>
To:     Mike Christie <michael.christie@oracle.com>
Cc:     Oleg Nesterov <oleg@redhat.com>, linux@leemhuis.info,
        nicolas.dichtel@6wind.com, axboe@kernel.dk,
        torvalds@linux-foundation.org, linux-kernel@vger.kernel.org,
        virtualization@lists.linux-foundation.org, mst@redhat.com,
        sgarzare@redhat.com, jasowang@redhat.com, stefanha@redhat.com,
        brauner@kernel.org
References: <20230522025124.5863-1-michael.christie@oracle.com>
        <20230522025124.5863-4-michael.christie@oracle.com>
        <20230522123029.GA22159@redhat.com>
        <cfca7764-d210-6df9-e182-2c093101c6cf@oracle.com>
        <20230522174757.GC22159@redhat.com> <20230523121506.GA6562@redhat.com>
        <c6e9a5db-798f-fa40-5ae2-a41f2d8ebab5@oracle.com>
Date:   Sat, 27 May 2023 20:41:29 -0500
In-Reply-To: <c6e9a5db-798f-fa40-5ae2-a41f2d8ebab5@oracle.com> (Mike
        Christie's message of "Thu, 25 May 2023 11:15:59 -0500")
Message-ID: <87a5xpkzau.fsf@email.froward.int.ebiederm.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Subject: Re: [PATCH 3/3] fork, vhost: Use CLONE_THREAD to fix freezer/ps
 regression
Precedence: bulk

Mike Christie <michael.christie@oracle.com> writes:

> On 5/23/23 7:15 AM, Oleg Nesterov wrote:
>> 
>> Now the main question. Whatever we do, SIGKILL/SIGSTOP/etc can come right
>> before we call work->fn(). Is it "safe" to run this callback with
>> signal_pending() or fatal_signal_pending() ?
>
> The questions before this one I'll leave for the core vhost devs since
> they know best.

Let me ask a clarifying question:

Is it only the call to schedule() in vhost_worker that you are worried
about not sleeping if signal_pending() or fatal_signal_pending()?

Is there concern that the worker functions aka "work->fn()" will also
have killable or interruptible sleeps that also will misbehave.

We can handle schedule() in vhost_worker without problem.

If a worker function has interruptible or killable sleeps that will turn
into busy waits or worse not sleeping long enough that seems like a
problem.  There is no way to guarantee that the outer loop of
vhost_worker will protect the worker functions from signal_pending()
or fatal_signal_pending() becoming true.


Eric