[PATCH 1/2] vhost-vdpa: Fix integer overflow in vhost_vdpa_process_iotlb_update()

[PATCH 1/2] vhost-vdpa: Fix integer overflow in vhost_vdpa_process_iotlb_update()

[Xie Yongji]

The "msg->iova + msg->size" addition can have an integer overflow
if the iotlb message is from a malicious user space application.
So let's fix it.

Fixes: 1b48dc03e575 ("vhost: vdpa: report iova range")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
---
 drivers/vhost/vdpa.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c
index 210ab35a7ebf..8e3c8790d493 100644
--- a/drivers/vhost/vdpa.c
+++ b/drivers/vhost/vdpa.c
@@ -615,6 +615,7 @@ static int vhost_vdpa_process_iotlb_update(struct vhost_vdpa *v,
 	int ret = 0;
 
 	if (msg->iova < v->range.first ||
+	    msg->iova - 1 > U64_MAX - msg->size ||
 	    msg->iova + msg->size - 1 > v->range.last)
 		return -EINVAL;
 
-- 
2.11.0

[PATCH 2/2] vhost: Fix the calculation in vhost_overflow()

[Xie Yongji]

This fixes the incorrect calculation for integer overflow
when the last address of iova range is 0xffffffff.

Fixes: ec33d031a14b ("vhost: detect 32 bit integer wrap around“)
Reported-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
---
 drivers/vhost/vhost.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
index b9e853e6094d..a9fd1b311d2f 100644
--- a/drivers/vhost/vhost.c
+++ b/drivers/vhost/vhost.c
@@ -738,7 +738,8 @@ static bool log_access_ok(void __user *log_base, u64 addr, unsigned long sz)
 static bool vhost_overflow(u64 uaddr, u64 size)
 {
 	/* Make sure 64 bit math will not overflow. */
-	return uaddr > ULONG_MAX || size > ULONG_MAX || uaddr > ULONG_MAX - size;
+	return uaddr > ULONG_MAX || size > ULONG_MAX ||
+	       uaddr - 1 > ULONG_MAX - size;
 }
 
 /* Caller should have vq mutex and device mutex. */
-- 
2.11.0

Re: [PATCH 1/2] vhost-vdpa: Fix integer overflow in vhost_vdpa_process_iotlb_update()

[Jason Wang]

在 2021/7/16 下午6:22, Xie Yongji 写道:
> The "msg->iova + msg->size" addition can have an integer overflow
> if the iotlb message is from a malicious user space application.
> So let's fix it.
>
> Fixes: 1b48dc03e575 ("vhost: vdpa: report iova range")
> Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>


Acked-by: Jason Wang <jasowang@redhat.com>


> ---
>   drivers/vhost/vdpa.c | 1 +
>   1 file changed, 1 insertion(+)
>
> diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c
> index 210ab35a7ebf..8e3c8790d493 100644
> --- a/drivers/vhost/vdpa.c
> +++ b/drivers/vhost/vdpa.c
> @@ -615,6 +615,7 @@ static int vhost_vdpa_process_iotlb_update(struct vhost_vdpa *v,
>   	int ret = 0;
>   
>   	if (msg->iova < v->range.first ||
> +	    msg->iova - 1 > U64_MAX - msg->size ||
>   	    msg->iova + msg->size - 1 > v->range.last)
>   		return -EINVAL;
>   

Re: [PATCH 2/2] vhost: Fix the calculation in vhost_overflow()

[Jason Wang]

在 2021/7/16 下午6:22, Xie Yongji 写道:
> This fixes the incorrect calculation for integer overflow
> when the last address of iova range is 0xffffffff.
>
> Fixes: ec33d031a14b ("vhost: detect 32 bit integer wrap around“)
> Reported-by: Jason Wang <jasowang@redhat.com>
> Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
> ---
>   drivers/vhost/vhost.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
> index b9e853e6094d..a9fd1b311d2f 100644
> --- a/drivers/vhost/vhost.c
> +++ b/drivers/vhost/vhost.c
> @@ -738,7 +738,8 @@ static bool log_access_ok(void __user *log_base, u64 addr, unsigned long sz)
>   static bool vhost_overflow(u64 uaddr, u64 size)
>   {
>   	/* Make sure 64 bit math will not overflow. */
> -	return uaddr > ULONG_MAX || size > ULONG_MAX || uaddr > ULONG_MAX - size;
> +	return uaddr > ULONG_MAX || size > ULONG_MAX ||
> +	       uaddr - 1 > ULONG_MAX - size;
>   }


Acked-by: Jason Wang <jasowang@redhat.com>


>   
>   /* Caller should have vq mutex and device mutex. */