Re: [PATCH v2 for-next] RDMA/hns: Fix memory corruption when allocating XRCDN

[Jason Gunthorpe <jgg@nvidia.com>]

On Mon, Mar 22, 2021 at 10:44:29AM +0800, Weihang Li wrote:
> It's incorrect to cast the type of pointer to xrcdn from (u32 *) to
> (unsigned long *), then pass it into hns_roce_bitmap_alloc(), this will
> lead to a memory corruption.
> 
> Fixes: 32548870d438 ("RDMA/hns: Add support for XRC on HIP09")
> Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
> Signed-off-by: Weihang Li <liweihang@huawei.com>
> Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
> ---
> Changes since v1:
> - xrcdn won't be set if hns_roce_bitmap_alloc() fails.
> - Link: https://patchwork.kernel.org/project/linux-rdma/patch/1616143536-24874-1-git-send-email-liweihang@huawei.com/
> 
>  drivers/infiniband/hw/hns/hns_roce_pd.c | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)

Applied to for-next

 
> diff --git a/drivers/infiniband/hw/hns/hns_roce_pd.c b/drivers/infiniband/hw/hns/hns_roce_pd.c
> index 3ca51ce..68a59ff 100644
> --- a/drivers/infiniband/hw/hns/hns_roce_pd.c
> +++ b/drivers/infiniband/hw/hns/hns_roce_pd.c
> @@ -140,8 +140,16 @@ void hns_roce_cleanup_uar_table(struct hns_roce_dev *hr_dev)
>  
>  static int hns_roce_xrcd_alloc(struct hns_roce_dev *hr_dev, u32 *xrcdn)
>  {
> -	return hns_roce_bitmap_alloc(&hr_dev->xrcd_bitmap,
> -				     (unsigned long *)xrcdn);
> +	unsigned long obj;
> +	int ret;
> +
> +	ret = hns_roce_bitmap_alloc(&hr_dev->xrcd_bitmap, &obj);
> +	if (ret)
> +		return ret;
> +
> +	*xrcdn = (u32)obj;

Though this cast is useless, I removed it

Thanks,
Jason