Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
MIME-Version: 1.0
References: <CAMhUBj=5jD5AjyaF8UmMXAZGFnMEgTruFM2KYL3GGZt-ABcHBQ@mail.gmail.com>
 <CACGkMEvgtzUXw9v=cMN3=Mqm9YTp_2bKmc4--q2+BREvgVdeLg@mail.gmail.com>
In-Reply-To: <CACGkMEvgtzUXw9v=cMN3=Mqm9YTp_2bKmc4--q2+BREvgVdeLg@mail.gmail.com>
From:   Zheyu Ma <zheyuma97@gmail.com>
Date:   Mon, 23 May 2022 12:47:01 +0800
Message-ID: <CAMhUBj=rC8ziy5B8cOroeZQW7f+DhsOkCvLWH-L89zx2LZy-Eg@mail.gmail.com>
Subject: Re: [BUG] vDPA/ifcvf: got a warning when removing the module
To:     Jason Wang <jasowang@redhat.com>
Cc:     mst <mst@redhat.com>, Zhu Lingshan <lingshan.zhu@intel.com>,
        Yongji Xie <xieyongji@bytedance.com>,
        Eli Cohen <elic@nvidia.com>, Parav Pandit <parav@nvidia.com>,
        virtualization <virtualization@lists.linux-foundation.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, May 23, 2022 at 12:15 PM Jason Wang <jasowang@redhat.com> wrote:
>
> On Sat, May 21, 2022 at 10:27 PM Zheyu Ma <zheyuma97@gmail.com> wrote:
> >
> > Hello,
> >
> > I found a bug in the ifcvf driver.
> > When removing the module, I got the following warning:
> >
> > [   14.478123] general protection fault, probably for non-canonical
> > address 0xdffffc0000000005: 0000 [#1] PREEMPT SMP KASAN PTI
> > [   14.478701] KASAN: null-ptr-deref in range
> > [0x0000000000000028-0x000000000000002f]
> > [   14.479922] RIP: 0010:vdpa_mgmtdev_unregister+0x39/0x150
> > [   14.484063] Call Trace:
> > [   14.484186]  <TASK>
> > [   14.484292]  ? _raw_spin_unlock_irqrestore+0x3d/0x60
> > [   14.484536]  ifcvf_remove+0x3a/0x50 [ifcvf]
> > [   14.484743]  pci_device_remove+0x92/0x240
> >
> > The reason is that the 'ifcvf_mgmt_dev' is zero which means that
> > ifcvf_vdpa_dev_add() was not executed.
> > Since I am not familiar with the driver, I cannot find a proper solution for it.
> >
>
> Looks like the drv data needs to be set in probe(). Could you pleas
> try to the attached patch? (compile test only).

This patch works for me, thanks!

Tested-by: Zheyu Ma <zheyuma97@gmail.com>

Regards,
Zheyu Ma