Virtualization - Feb 2021 - [PATCH RFC v2 08/10] vdpa: add vdpa simulator for block device

On Thu, Jan 28, 2021 at 03:41:25PM +0100, Stefano Garzarella
wrote:
> +static void vdpasim_blk_work(struct work_struct *work)
> +{
> +	struct vdpasim *vdpasim = container_of(work, struct vdpasim, work);
> +	u8 status = VIRTIO_BLK_S_OK;
> +	int i;
> +
> +	spin_lock(&vdpasim->lock);
> +
> +	if (!(vdpasim->status & VIRTIO_CONFIG_S_DRIVER_OK))
> +		goto out;
> +
> +	for (i = 0; i < VDPASIM_BLK_VQ_NUM; i++) {
> +		struct vdpasim_virtqueue *vq = &vdpasim->vqs[i];
> +
> +		if (!vq->ready)
> +			continue;
> +
> +		while (vringh_getdesc_iotlb(&vq->vring, &vq->out_iov,
> +					    &vq->in_iov, &vq->head,
> +					    GFP_ATOMIC) > 0) {
> +			int write;
> +
> +			vq->in_iov.i = vq->in_iov.used - 1;
> +			write = vringh_iov_push_iotlb(&vq->vring, &vq->in_iov,
> +						      &status, 1);
> +			if (write <= 0)
> +				break;
This code looks fragile:

1. Relying on unsigned underflow and the while loop in
   vringh_iov_push_iotlb() to handle the case where in_iov.used == 0 is
   risky and could break.

2. Does this assume that the last in_iov element has size 1? For
   example, the guest driver may send a single "in" iovec with size
513
   when reading 512 bytes (with an extra byte for the request status).

Please validate inputs fully, even in test/development code, because
it's likely to be copied by others when writing production code (or
deployed in production by unsuspecting users) :).
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL:
<http://lists.linuxfoundation.org/pipermail/virtualization/attachments/20210202/29d012b1/attachment.sig>

On Tue, Feb 02, 2021 at 09:34:12AM +0000, Stefan Hajnoczi
wrote:
>On Thu, Jan 28, 2021 at 03:41:25PM +0100, Stefano Garzarella wrote:
>> +static void vdpasim_blk_work(struct work_struct *work)
>> +{
>> +	struct vdpasim *vdpasim = container_of(work, struct vdpasim, work);
>> +	u8 status = VIRTIO_BLK_S_OK;
>> +	int i;
>> +
>> +	spin_lock(&vdpasim->lock);
>> +
>> +	if (!(vdpasim->status & VIRTIO_CONFIG_S_DRIVER_OK))
>> +		goto out;
>> +
>> +	for (i = 0; i < VDPASIM_BLK_VQ_NUM; i++) {
>> +		struct vdpasim_virtqueue *vq = &vdpasim->vqs[i];
>> +
>> +		if (!vq->ready)
>> +			continue;
>> +
>> +		while (vringh_getdesc_iotlb(&vq->vring, &vq->out_iov,
>> +					    &vq->in_iov, &vq->head,
>> +					    GFP_ATOMIC) > 0) {
>> +			int write;
>> +
>> +			vq->in_iov.i = vq->in_iov.used - 1;
>> +			write = vringh_iov_push_iotlb(&vq->vring,
&vq->in_iov,
>> +						      &status, 1);
>> +			if (write <= 0)
>> +				break;
>
>This code looks fragile:
>
>1. Relying on unsigned underflow and the while loop in
>   vringh_iov_push_iotlb() to handle the case where in_iov.used == 0 is
>   risky and could break.
>
>2. Does this assume that the last in_iov element has size 1? For
>   example, the guest driver may send a single "in" iovec with
size 513
>   when reading 512 bytes (with an extra byte for the request status).
>
>Please validate inputs fully, even in test/development code, because
>it's likely to be copied by others when writing production code (or
>deployed in production by unsuspecting users) :).
Perfectly agree on that, so I addressed these things, also following 
your review on the previous version, on the next patch of this series:
"vdpa_sim_blk: implement ramdisk behaviour".

Do you think should I move these checks in this patch?

I did this to leave Max credit for this patch and add more code to 
emulate a ramdisk in later patches.

Thanks,
Stefano