On 27-07-19, 17:30, Jia-Ju Bai wrote:
> In mv_xor_prep_dma_xor(), there is an if statement on line 577 to check
> whether sw_desc is NULL:
> if (sw_desc)
>
> When sw_desc is NULL, it is used on line 594:
> dev_dbg(..., sw_desc, &sw_desc->async_tx);
>
> Thus, a possible null-pointer dereference may occur.
>
> To fix this bug, sw_desc is checked before being used.
>
> This bug is found by a static analysis tool STCheck written by us.
>
> Signed-off-by: Jia-Ju Bai <baijiaju1...@gmail.com>
> ---
> drivers/dma/mv_xor.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/dma/mv_xor.c b/drivers/dma/mv_xor.c
> index 0ac8e7b34e12..08c0b2a9eb32 100644
> --- a/drivers/dma/mv_xor.c
> +++ b/drivers/dma/mv_xor.c
> @@ -589,9 +589,11 @@ mv_xor_prep_dma_xor(struct dma_chan *chan, dma_addr_t
> dest, dma_addr_t *src,
> }
> }
>
> - dev_dbg(mv_chan_to_devp(mv_chan),
> - "%s sw_desc %p async_tx %p \n",
> - __func__, sw_desc, &sw_desc->async_tx);
> + if (sw_desc) {
> + dev_dbg(mv_chan_to_devp(mv_chan),
> + "%s sw_desc %p async_tx %p \n",
> + __func__, sw_desc, &sw_desc->async_tx);
> + }
why not move this into the preceeding if condition?
> return sw_desc ? &sw_desc->async_tx : NULL;
> }
>
> --
> 2.17.0
--
~Vinod
