Re: [PATCH 1/2] i2c: virtio: disable timeout handling

[Viresh Kumar <viresh.kumar@linaro.org>]

+Greg.

On 19-10-21, 09:46, Vincent Whitchurch wrote:
> If a timeout is hit, it can result is incorrect data on the I2C bus
> and/or memory corruptions in the guest since the device can still be
> operating on the buffers it was given while the guest has freed them.
> 
> Here is, for example, the start of a slub_debug splat which was
> triggered on the next transfer after one transfer was forced to timeout
> by setting a breakpoint in the backend (rust-vmm/vhost-device):
> 
>  BUG kmalloc-1k (Not tainted): Poison overwritten
>  First byte 0x1 instead of 0x6b
>  Allocated in virtio_i2c_xfer+0x65/0x35c age=350 cpu=0 pid=29
>  	__kmalloc+0xc2/0x1c9
>  	virtio_i2c_xfer+0x65/0x35c
>  	__i2c_transfer+0x429/0x57d
>  	i2c_transfer+0x115/0x134
>  	i2cdev_ioctl_rdwr+0x16a/0x1de
>  	i2cdev_ioctl+0x247/0x2ed
>  	vfs_ioctl+0x21/0x30
>  	sys_ioctl+0xb18/0xb41
>  Freed in virtio_i2c_xfer+0x32e/0x35c age=244 cpu=0 pid=29
>  	kfree+0x1bd/0x1cc
>  	virtio_i2c_xfer+0x32e/0x35c
>  	__i2c_transfer+0x429/0x57d
>  	i2c_transfer+0x115/0x134
>  	i2cdev_ioctl_rdwr+0x16a/0x1de
>  	i2cdev_ioctl+0x247/0x2ed
>  	vfs_ioctl+0x21/0x30
>  	sys_ioctl+0xb18/0xb41
> 
> There is no simple fix for this (the driver would have to always create
> bounce buffers and hold on to them until the device eventually returns
> the buffers), so just disable the timeout support for now.

That is a very valid problem, and I have faced it too when my QEMU
setup is very slow :)

> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
> ---
>  drivers/i2c/busses/i2c-virtio.c | 14 +++++---------
>  1 file changed, 5 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-virtio.c b/drivers/i2c/busses/i2c-virtio.c
> index f10a603b13fb..7b2474e6876f 100644
> --- a/drivers/i2c/busses/i2c-virtio.c
> +++ b/drivers/i2c/busses/i2c-virtio.c
> @@ -106,11 +106,10 @@ static int virtio_i2c_prepare_reqs(struct virtqueue *vq,
>  
>  static int virtio_i2c_complete_reqs(struct virtqueue *vq,
>  				    struct virtio_i2c_req *reqs,
> -				    struct i2c_msg *msgs, int num,
> -				    bool timedout)
> +				    struct i2c_msg *msgs, int num)
>  {
>  	struct virtio_i2c_req *req;
> -	bool failed = timedout;
> +	bool failed = false;
>  	unsigned int len;
>  	int i, j = 0;
>  
> @@ -132,7 +131,7 @@ static int virtio_i2c_complete_reqs(struct virtqueue *vq,
>  			j++;
>  	}
>  
> -	return timedout ? -ETIMEDOUT : j;
> +	return j;
>  }
>  
>  static int virtio_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs,
> @@ -141,7 +140,6 @@ static int virtio_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs,
>  	struct virtio_i2c *vi = i2c_get_adapdata(adap);
>  	struct virtqueue *vq = vi->vq;
>  	struct virtio_i2c_req *reqs;
> -	unsigned long time_left;
>  	int count;
>  
>  	reqs = kcalloc(num, sizeof(*reqs), GFP_KERNEL);
> @@ -164,11 +162,9 @@ static int virtio_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs,
>  	reinit_completion(&vi->completion);
>  	virtqueue_kick(vq);
>  
> -	time_left = wait_for_completion_timeout(&vi->completion, adap->timeout);
> -	if (!time_left)
> -		dev_err(&adap->dev, "virtio i2c backend timeout.\n");
> +	wait_for_completion(&vi->completion);

Doing this may not be a good thing based on the kernel rules I have
understood until now. Maybe Greg and Wolfram can clarify on this.

We are waiting here for an external entity (Host kernel) or a firmware
that uses virtio for transport. If the other side is hacked, it can
make the kernel hang here for ever. I thought that is something that
the kernel should never do.

-- 
viresh