On Mon, Dec 05, 2022 at 06:14:41AM +0000, Yuan Can wrote:
> The coex_cb needs to be freed when rsi_create_kthread() failed in
> rsi_coex_attach().
>
> Fixes: 2108df3c4b18 ("rsi: add coex support")
> Signed-off-by: Yuan Can <yuancan@xxxxxxxxxx>
Reviewed-by: Simon Horman <simon.horman@xxxxxxxxxxxx>
> ---
> drivers/net/wireless/rsi/rsi_91x_coex.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/net/wireless/rsi/rsi_91x_coex.c b/drivers/net/wireless/rsi/rsi_91x_coex.c
> index 8a3d86897ea8..45ac9371f262 100644
> --- a/drivers/net/wireless/rsi/rsi_91x_coex.c
> +++ b/drivers/net/wireless/rsi/rsi_91x_coex.c
> @@ -160,6 +160,7 @@ int rsi_coex_attach(struct rsi_common *common)
> rsi_coex_scheduler_thread,
> "Coex-Tx-Thread")) {
> rsi_dbg(ERR_ZONE, "%s: Unable to init tx thrd\n", __func__);
> + kfree(coex_cb);
I was going to ask if the assignment of coex_cb to common->coex_cb
also needs to be cleaned up. But I see that the caller frees
common on error, so I think this is ok.
> return -EINVAL;
> }
> return 0;
> --
> 2.17.1
>
