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;
Subject: Re: 5.13-rt1 + KVM = WARNING: at fs/eventfd.c:74 eventfd_signal()
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        Al Viro <viro@zeniv.linux.org.uk>,
        He Zhe <zhe.he@windriver.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Juri Lelli <jlelli@redhat.com>,
        "Michael S. Tsirkin" <mst@redhat.com>,
        Jason Wang <jasowang@redhat.com>,
        Sebastian Andrzej Siewior <bigeasy@linutronix.de>
References: <df278db6-1fc0-3d42-9c0e-f5a085c6351e@redhat.com>
 <8dfc0ee9-b97a-8ca8-d057-31c8cad3f5b6@redhat.com>
 <f0254740-944d-201b-9a66-9db1fe480ca6@redhat.com>
 <475f84e2-78ee-1a24-ef57-b16c1f2651ed@redhat.com>
 <a56ddd50-2cd1-f16e-5180-5232c449fbd0@redhat.com>
 <9951ef78-587a-34e1-7b5b-280285e37098@redhat.com>
From:   Daniel Bristot de Oliveira <bristot@redhat.com>
Message-ID: <82c20a95-4616-eeef-2459-b2decdd7f3bd@redhat.com>
Date:   Thu, 15 Jul 2021 14:34:25 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.11.0
MIME-Version: 1.0
In-Reply-To: <9951ef78-587a-34e1-7b5b-280285e37098@redhat.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 7/15/21 11:46 AM, Paolo Bonzini wrote:
> On 15/07/21 10:22, Daniel Bristot de Oliveira wrote:
>> On 7/14/21 12:35 PM, Paolo Bonzini wrote:
>>> On 14/07/21 11:23, Jason Wang wrote:
>>>> I think this can be probably fixed here:
>>>>
>>>> https://lore.kernel.org/lkml/20210618084412.18257-1-zhe.he@windriver.com/
>>>
>>> That seems wrong; in particular it wouldn't protect against AB/BA deadlocks.
>>> In fact, the bug is with the locking; the code assumes that
>>> spin_lock_irqsave/spin_unlock_irqrestore is non-preemptable and therefore
>>> increments and decrements the percpu variable inside the critical section.
>>>
>>> This obviously does not fly with PREEMPT_RT; the right fix should be
>>> using a local_lock.  Something like this (untested!!):
>>
>> the lock needs to be per-pcu...
> 
> Great, thanks for testing and fixing the patch!  Are you going to post
> it again once you've confirmed it works?

I can do that... unless you want to send it yourself...

-- Daniel

> Paolo