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;
MIME-Version: 1.0
References: <20220305095525.5145-1-mail@anirudhrb.com>
In-Reply-To: <20220305095525.5145-1-mail@anirudhrb.com>
From:   Jason Wang <jasowang@redhat.com>
Date:   Mon, 7 Mar 2022 12:28:10 +0800
Message-ID: <CACGkMEtb6qSq2=WXWeaDZknw77C7pQwSgxP0-KxCCVhTyM-HwQ@mail.gmail.com>
Subject: Re: [PATCH v3] vhost: fix hung thread due to erroneous iotlb entries
To:     Anirudh Rayabharam <mail@anirudhrb.com>
Cc:     "Michael S. Tsirkin" <mst@redhat.com>,
        syzbot+0abd373e2e50d704db87@syzkaller.appspotmail.com,
        kvm <kvm@vger.kernel.org>,
        virtualization <virtualization@lists.linux-foundation.org>,
        netdev <netdev@vger.kernel.org>,
        linux-kernel <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sat, Mar 5, 2022 at 5:56 PM Anirudh Rayabharam <mail@anirudhrb.com> wrote:
>
> In vhost_iotlb_add_range_ctx(), range size can overflow to 0 when
> start is 0 and last is ULONG_MAX. One instance where it can happen
> is when userspace sends an IOTLB message with iova=size=uaddr=0
> (vhost_process_iotlb_msg). So, an entry with size = 0, start = 0,
> last = ULONG_MAX ends up in the iotlb. Next time a packet is sent,
> iotlb_access_ok() loops indefinitely due to that erroneous entry.
>
>         Call Trace:
>          <TASK>
>          iotlb_access_ok+0x21b/0x3e0 drivers/vhost/vhost.c:1340
>          vq_meta_prefetch+0xbc/0x280 drivers/vhost/vhost.c:1366
>          vhost_transport_do_send_pkt+0xe0/0xfd0 drivers/vhost/vsock.c:104
>          vhost_worker+0x23d/0x3d0 drivers/vhost/vhost.c:372
>          kthread+0x2e9/0x3a0 kernel/kthread.c:377
>          ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:295
>          </TASK>
>
> Reported by syzbot at:
>         https://syzkaller.appspot.com/bug?extid=0abd373e2e50d704db87
>
> To fix this, do two things:
>
> 1. Return -EINVAL in vhost_chr_write_iter() when userspace asks to map
>    a range with size 0.
> 2. Fix vhost_iotlb_add_range_ctx() to handle the range [0, ULONG_MAX]
>    by splitting it into two entries.
>
> Fixes: 0bbe30668d89e ("vhost: factor out IOTLB")
> Reported-by: syzbot+0abd373e2e50d704db87@syzkaller.appspotmail.com
> Tested-by: syzbot+0abd373e2e50d704db87@syzkaller.appspotmail.com
> Signed-off-by: Anirudh Rayabharam <mail@anirudhrb.com>
> ---
> Changes in v3:
> 1. Simplify expression since start is always 0
> 2. Fix checkpatch issue
> 3. Add Fixes tag
>
> v2: https://lore.kernel.org/kvm/20220224143320.3751-1-mail@anirudhrb.com/
> Changes in v2:
> 1. Don't reject range [0, ULONG_MAX], split it instead.
> 2. Validate msg.size in vhost_chr_write_iter().
>
> v1: https://lore.kernel.org/lkml/20220221195303.13560-1-mail@anirudhrb.com/
>
> ---
>  drivers/vhost/iotlb.c | 11 +++++++++++
>  drivers/vhost/vhost.c |  5 +++++
>  2 files changed, 16 insertions(+)
>
> diff --git a/drivers/vhost/iotlb.c b/drivers/vhost/iotlb.c
> index 670d56c879e5..40b098320b2a 100644
> --- a/drivers/vhost/iotlb.c
> +++ b/drivers/vhost/iotlb.c
> @@ -57,6 +57,17 @@ int vhost_iotlb_add_range_ctx(struct vhost_iotlb *iotlb,
>         if (last < start)
>                 return -EFAULT;
>
> +       /* If the range being mapped is [0, ULONG_MAX], split it into two entries
> +        * otherwise its size would overflow u64.
> +        */
> +       if (start == 0 && last == ULONG_MAX) {
> +               u64 mid = last / 2;
> +
> +               vhost_iotlb_add_range_ctx(iotlb, start, mid, addr, perm, opaque);

Do we need to check the errors and fail?

Others look good.

Thanks

> +               addr += mid + 1;
> +               start = mid + 1;
> +       }
> +
>         if (iotlb->limit &&
>             iotlb->nmaps == iotlb->limit &&
>             iotlb->flags & VHOST_IOTLB_FLAG_RETIRE) {
> diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
> index 59edb5a1ffe2..55475fd59fb7 100644
> --- a/drivers/vhost/vhost.c
> +++ b/drivers/vhost/vhost.c
> @@ -1170,6 +1170,11 @@ ssize_t vhost_chr_write_iter(struct vhost_dev *dev,
>                 goto done;
>         }
>
> +       if (msg.size == 0) {
> +               ret = -EINVAL;
> +               goto done;
> +       }
> +
>         if (dev->msg_handler)
>                 ret = dev->msg_handler(dev, &msg);
>         else
> --
> 2.35.1
>