Currently, memdup_user() is utilized at two positions to copy userspace
arrays. This is done without overflow checks.

Use the new wrapper memdup_array_user() to copy the arrays more safely.

Suggested-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
---
 drivers/infiniband/hw/hfi1/user_exp_rcv.c | 4 ++--
 drivers/infiniband/hw/hfi1/user_sdma.c    | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/infiniband/hw/hfi1/user_exp_rcv.c b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
index 96058baf36ed..53a623892d9d 100644
--- a/drivers/infiniband/hw/hfi1/user_exp_rcv.c
+++ b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
@@ -491,8 +491,8 @@ int hfi1_user_exp_rcv_clear(struct hfi1_filedata *fd,
 	if (unlikely(tinfo->tidcnt > fd->tid_used))
 		return -EINVAL;
 
-	tidinfo = memdup_user(u64_to_user_ptr(tinfo->tidlist),
-			      sizeof(tidinfo[0]) * tinfo->tidcnt);
+	tidinfo = memdup_array_user(u64_to_user_ptr(tinfo->tidlist),
+				    tinfo->tidcnt, sizeof(tidinfo[0]));
 	if (IS_ERR(tidinfo))
 		return PTR_ERR(tidinfo);
 
diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c
index 29ae7beb9b03..f7fa8e699a78 100644
--- a/drivers/infiniband/hw/hfi1/user_sdma.c
+++ b/drivers/infiniband/hw/hfi1/user_sdma.c
@@ -494,8 +494,8 @@ int hfi1_user_sdma_process_request(struct hfi1_filedata *fd,
 		 * equal to the pkt count. However, there is no way to
 		 * tell at this point.
 		 */
-		tmp = memdup_user(iovec[idx].iov_base,
-				  ntids * sizeof(*req->tids));
+		tmp = memdup_array_user(iovec[idx].iov_base,
+					ntids, sizeof(*req->tids));
 		if (IS_ERR(tmp)) {
 			ret = PTR_ERR(tmp);
 			SDMA_DBG(req, "Failed to copy %d TIDs (%d)",
-- 
2.41.0

On Thu, 02 Nov 2023 20:13:09 +0100, Philipp Stanner wrote:
> Currently, memdup_user() is utilized at two positions to copy userspace
> arrays. This is done without overflow checks.
> 
> Use the new wrapper memdup_array_user() to copy the arrays more safely.
> 
> 

Applied, thanks!

[1/1] drivers/infiniband: copy userspace arrays safely
      https://git.kernel.org/rdma/rdma/c/c170d4ff21a8a5

Best regards,
-- 
Leon Romanovsky <leon@kernel.org>

On 11/2/23 3:13 PM, Philipp Stanner wrote:
> Currently, memdup_user() is utilized at two positions to copy userspace
> arrays. This is done without overflow checks.
> 
> Use the new wrapper memdup_array_user() to copy the arrays more safely.
> 
> Suggested-by: Dave Airlie <airlied@redhat.com>
> Signed-off-by: Philipp Stanner <pstanner@redhat.com>
> ---
>  drivers/infiniband/hw/hfi1/user_exp_rcv.c | 4 ++--
>  drivers/infiniband/hw/hfi1/user_sdma.c    | 4 ++--
>  2 files changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/infiniband/hw/hfi1/user_exp_rcv.c b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
> index 96058baf36ed..53a623892d9d 100644
> --- a/drivers/infiniband/hw/hfi1/user_exp_rcv.c
> +++ b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
> @@ -491,8 +491,8 @@ int hfi1_user_exp_rcv_clear(struct hfi1_filedata *fd,
>  	if (unlikely(tinfo->tidcnt > fd->tid_used))
>  		return -EINVAL;
>  
> -	tidinfo = memdup_user(u64_to_user_ptr(tinfo->tidlist),
> -			      sizeof(tidinfo[0]) * tinfo->tidcnt);
> +	tidinfo = memdup_array_user(u64_to_user_ptr(tinfo->tidlist),
> +				    tinfo->tidcnt, sizeof(tidinfo[0]));
>  	if (IS_ERR(tidinfo))
>  		return PTR_ERR(tidinfo);
>  
> diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c
> index 29ae7beb9b03..f7fa8e699a78 100644
> --- a/drivers/infiniband/hw/hfi1/user_sdma.c
> +++ b/drivers/infiniband/hw/hfi1/user_sdma.c
> @@ -494,8 +494,8 @@ int hfi1_user_sdma_process_request(struct hfi1_filedata *fd,
>  		 * equal to the pkt count. However, there is no way to
>  		 * tell at this point.
>  		 */
> -		tmp = memdup_user(iovec[idx].iov_base,
> -				  ntids * sizeof(*req->tids));
> +		tmp = memdup_array_user(iovec[idx].iov_base,
> +					ntids, sizeof(*req->tids));
>  		if (IS_ERR(tmp)) {
>  			ret = PTR_ERR(tmp);
>  			SDMA_DBG(req, "Failed to copy %d TIDs (%d)",

Acked-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>