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;
Subject: Re: WARNING in percpu_ref_kill_and_confirm
From:   Jens Axboe <axboe@kernel.dk>
To:     Linus Torvalds <torvalds@linux-foundation.org>
Cc:     syzbot <syzbot+10d25e23199614b7721f@syzkaller.appspotmail.com>,
        Arnd Bergmann <arnd@arndb.de>, Borislav Petkov <bp@alien8.de>,
        "Darrick J. Wong" <darrick.wong@oracle.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Peter Anvin <hpa@zytor.com>,
        Linux API <linux-api@vger.kernel.org>,
        linux-arch <linux-arch@vger.kernel.org>,
        linux-block <linux-block@vger.kernel.org>,
        linux-fsdevel <linux-fsdevel@vger.kernel.org>,
        Linux List Kernel Mailing <linux-kernel@vger.kernel.org>,
        Andrew Lutomirski <luto@kernel.org>,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        Ingo Molnar <mingo@redhat.com>,
        Michael Ellerman <mpe@ellerman.id.au>,
        syzkaller-bugs <syzkaller-bugs@googlegroups.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Al Viro <viro@zeniv.linux.org.uk>,
        the arch/x86 maintainers <x86@kernel.org>
References: <00000000000043fe9c058720a5d3@google.com>
 <ad884b18-0e03-d6be-02b0-4e5273f068ce@kernel.dk>
 <CAHk-=wgG=RqK8T5kWiZfH7da6rBuqp3kAesP=ETgceb6RzvZng@mail.gmail.com>
 <53a17444-9539-5810-82a0-ceeefa742508@kernel.dk>
Message-ID: <cfa6be6d-4125-6d1e-8993-e9ec9dbaa7bb@kernel.dk>
Date:   Mon, 22 Apr 2019 10:38:32 -0600
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.6.1
MIME-Version: 1.0
In-Reply-To: <53a17444-9539-5810-82a0-ceeefa742508@kernel.dk>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 4/22/19 10:32 AM, Jens Axboe wrote:
> On 4/22/19 10:27 AM, Linus Torvalds wrote:
>> [ Crossed emails ]
>>
>> On Mon, Apr 22, 2019 at 9:23 AM Jens Axboe <axboe@kernel.dk> wrote:
>>>
>>> I think the below should fix this. Very early versions of io_uring didn't
>>> have this issue, since we did the percpu ref tryget for io_uring_register().
>>
>> Ok, so I like your patch better than mine, but note how syzbot
>> bisected this to the initial merge of the io_uring code.
> 
> Yes, I did think about that too...
> 
>> I agree that code shouldn't have had this particular issue, but it
>> looks like it does.
>>
>> Is there some way to race with io_ring_ctx_wait_and_kill(), which
>> _also_ does that ref_kill() thing? I'm not seeing how that could
>> happen, but maybe if the file ref counts get screwed up you have
>> ->release() called early..
> 
> I just tried on the current code and it triggers easily, but that's
> with that mutex patch in there. I agree it should not trigger before
> that, unless something is wonky. I'll try and play around with it a bit
> and see what is going on (or if I can trigger it at all with the mutex
> change reverted).

With the mutex change in, I can trigger it in a second or so. Just ran
the reproducer with that change reverted, and I'm not seeing any badness.
So I do wonder if the bisect results are accurate?

I think the dying check should cover it, and then marked with fixing
that mutex commit.

-- 
Jens Axboe