* Re: [PATCH v6] staging: r8188eu: fix a potential memory leak in rtw_init_cmd_priv()
[not found] <tencent_1B6AAE10471D4556788892F8FF3E4812F306@qq.com>
@ 2022-09-13 19:14 ` philipp hortmann
0 siblings, 0 replies; only message in thread
From: philipp hortmann @ 2022-09-13 19:14 UTC (permalink / raw)
To: xkernel.wang; +Cc: Larry.Finger, phil, gregkh, linux-staging, linux-kernel
On Fri, Sep 09, 2022 at 07:45:57PM +0800, xkernel.wang@foxmail.com wrote:
> From: Xiaoke Wang <xkernel.wang@foxmail.com>
>
> In rtw_init_cmd_priv(), if `pcmdpriv->rsp_allocated_buf` is allocated
> in failure, then `pcmdpriv->cmd_allocated_buf` will not be properly
> released. Besides, considering there are only two error paths and the
> first one can directly return, we do not need to implicitly jump to the
> `exit` tag to execute the error handling code.
>
> So this patch added `kfree(pcmdpriv->cmd_allocated_buf);` on the error
> path to release the resource and simplified the return logic of
> rtw_init_cmd_priv(). As there is no proper device to test with, no
> runtime testing was performed.
>
> Signed-off-by: Xiaoke Wang <xkernel.wang@foxmail.com>
> ---
> ChangeLog:
> v1->v2: simplify the function.
> v2->v3: rebase and recover some whitespace cleanups.
> v3->v4: update the description.
> v4->v5: rebase...
> v5->v6: update the description.
> drivers/staging/r8188eu/core/rtw_cmd.c | 15 +++++----------
> 1 file changed, 5 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/staging/r8188eu/core/rtw_cmd.c b/drivers/staging/r8188eu/core/rtw_cmd.c
> index 9c700ff..e38a826 100644
> --- a/drivers/staging/r8188eu/core/rtw_cmd.c
> +++ b/drivers/staging/r8188eu/core/rtw_cmd.c
> @@ -58,8 +58,6 @@ static int _rtw_enqueue_cmd(struct __queue *queue, struct cmd_obj *obj)
>
> u32 rtw_init_cmd_priv(struct cmd_priv *pcmdpriv)
> {
> - u32 res = _SUCCESS;
> -
> init_completion(&pcmdpriv->enqueue_cmd);
> /* sema_init(&(pcmdpriv->cmd_done_sema), 0); */
> init_completion(&pcmdpriv->start_cmd_thread);
> @@ -74,27 +72,24 @@ u32 rtw_init_cmd_priv(struct cmd_priv *pcmdpriv)
> pcmdpriv->cmd_allocated_buf = kzalloc(MAX_CMDSZ + CMDBUFF_ALIGN_SZ,
> GFP_KERNEL);
>
> - if (!pcmdpriv->cmd_allocated_buf) {
> - res = _FAIL;
> - goto exit;
> - }
> + if (!pcmdpriv->cmd_allocated_buf)
> + return _FAIL;
>
> pcmdpriv->cmd_buf = pcmdpriv->cmd_allocated_buf + CMDBUFF_ALIGN_SZ - ((size_t)(pcmdpriv->cmd_allocated_buf) & (CMDBUFF_ALIGN_SZ - 1));
>
> pcmdpriv->rsp_allocated_buf = kzalloc(MAX_RSPSZ + 4, GFP_KERNEL);
>
> if (!pcmdpriv->rsp_allocated_buf) {
> - res = _FAIL;
> - goto exit;
> + kfree(pcmdpriv->cmd_allocated_buf);
> + return _FAIL;
> }
>
> pcmdpriv->rsp_buf = pcmdpriv->rsp_allocated_buf + 4 - ((size_t)(pcmdpriv->rsp_allocated_buf) & 3);
>
> pcmdpriv->cmd_done_cnt = 0;
> pcmdpriv->rsp_cnt = 0;
> -exit:
>
> - return res;
> + return _SUCCESS;
> }
>
> u32 rtw_init_evt_priv(struct evt_priv *pevtpriv)
> --
Tested-by: Philipp Hortmann <philipp.g.hortmann@gmail.com> # Edimax N150
^ permalink raw reply [flat|nested] only message in thread