Re: [PATCH] staging:r8188eu: avoid skb_clone for amsdu to msdu conversion
From: Greg Kroah-Hartman
Date: Thu Apr 23 2020 - 07:29:16 EST
On Sat, Apr 18, 2020 at 11:41:12AM +0300, Ivan Safonov wrote:
> skb clones use same data buffer, so tail of one skb is corrupted by beginning of next skb.
Please properly wrap your changelog text at the correct column (72).
Also, your subject: line does not have the correct driver name :(
>
> Signed-off-by: Ivan Safonov <insafonov@xxxxxxxxx>
> ---
> drivers/staging/rtl8188eu/core/rtw_recv.c | 19 ++++++-------------
> 1 file changed, 6 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/staging/rtl8188eu/core/rtw_recv.c b/drivers/staging/rtl8188eu/core/rtw_recv.c
> index d4278361e002..a036ef104198 100644
> --- a/drivers/staging/rtl8188eu/core/rtw_recv.c
> +++ b/drivers/staging/rtl8188eu/core/rtw_recv.c
> @@ -1525,21 +1525,14 @@ static int amsdu_to_msdu(struct adapter *padapter, struct recv_frame *prframe)
>
> /* Allocate new skb for releasing to upper layer */
> sub_skb = dev_alloc_skb(nSubframe_Length + 12);
> - if (sub_skb) {
> - skb_reserve(sub_skb, 12);
> - skb_put_data(sub_skb, pdata, nSubframe_Length);
> - } else {
> - sub_skb = skb_clone(prframe->pkt, GFP_ATOMIC);
> - if (sub_skb) {
> - sub_skb->data = pdata;
> - sub_skb->len = nSubframe_Length;
> - skb_set_tail_pointer(sub_skb, nSubframe_Length);
> - } else {
> - DBG_88E("skb_clone() Fail!!! , nr_subframes=%d\n", nr_subframes);
> - break;
> - }
> + if (!sub_skb) {
> + DBG_88E("dev_alloc_skb() Fail!!! , nr_subframes=%d\n", nr_subframes);
> + break;
> }
>
> + skb_reserve(sub_skb, 12);
> + skb_put_data(sub_skb, pdata, nSubframe_Length);
> +
Have you tested this?
thanks,
greg k-h