Re: [PATCH 3/6][next] ath6kl: wmi: Replace one-element array with flexible-array member in struct wmi_channel_list_reply

[Jeff Johnson]

On 2/22/2022 6:38 PM, Gustavo A. R. Silva wrote:
> Replace one-element array with flexible-array member in struct
> wmi_channel_list_reply.
>
> It's also worth noting that due to the flexible array transformation,
> the size of struct wmi_channel_list_reply changed, see below.
>
> Before flex-array transformation:
>
> struct wmi_channel_list_reply {
> 	u8                         reserved;             /*     0     1 */
> 	u8                         num_ch;               /*     1     1 */
> 	__le16                     ch_list[1];           /*     2     2 */
>
> 	/* size: 4, cachelines: 1, members: 3 */
> 	/* last cacheline: 4 bytes */
> };
>
> After flex-array transformation:
>
> struct wmi_channel_list_reply {
> 	u8                         reserved;             /*     0     1 */
> 	u8                         num_ch;               /*     1     1 */
> 	__le16                     ch_list[];            /*     2     0 */
>
> 	/* size: 2, cachelines: 1, members: 3 */
> 	/* last cacheline: 2 bytes */
> };
>
> So, the following change preserves the logic that if _len_ is at least
> 4 bytes in size, this is the existence of at least one channel in
> ch_list[] is being considered, then the execution jumps to call
> ath6kl_wakeup_event(wmi->parent_dev);, otherwise _len_ is 2 bytes or
> less and the code returns -EINVAL:
>
> 	-       if (len < sizeof(struct wmi_channel_list_reply))
> 	+       if (len <= sizeof(struct wmi_channel_list_reply))
>
> This issue was found with the help of Coccinelle and audited and fixed,
> manually.
>
> Link: https://www.kernel.org/doc/html/v5.16/process/deprecated.html#zero-length-and-one-element-arrays
> Link: https://github.com/KSPP/linux/issues/79
> Signed-off-by: Gustavo A. R. Silva <gustavoars@xxxxxxxxxx>
> ---
> Hi!
>
> It'd be great if someone can confirm or comment on the following
> changes described in the changelog text:
>
> 	-       if (len < sizeof(struct wmi_channel_list_reply))
> 	+       if (len <= sizeof(struct wmi_channel_list_reply))

My opinion is this can remain unchanged since being unchanged would 
correctly handle a channel list with no channels whereas the original 
code required that at least one channel be present.

The test is really there just to make sure the entirety of the "fixed" 
portion of the message is present.

Ultimately it doesn't matter since no actual processing of the channel 
list takes place.

If actual processing took place, then it would make sense to have an 
additional test to verify the len is large enough to handle num_ch entries.


> Thanks
>
>   drivers/net/wireless/ath/ath6kl/wmi.c | 2 +-
>   drivers/net/wireless/ath/ath6kl/wmi.h | 2 +-
>   2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath6kl/wmi.c b/drivers/net/wireless/ath/ath6kl/wmi.c
> index bdfc057c5a82..049d75f31f3c 100644
> --- a/drivers/net/wireless/ath/ath6kl/wmi.c
> +++ b/drivers/net/wireless/ath/ath6kl/wmi.c
> @@ -1240,7 +1240,7 @@ static int ath6kl_wmi_ratemask_reply_rx(struct wmi *wmi, u8 *datap, int len)
>   
>   static int ath6kl_wmi_ch_list_reply_rx(struct wmi *wmi, u8 *datap, int len)
>   {
> -	if (len < sizeof(struct wmi_channel_list_reply))
> +	if (len <= sizeof(struct wmi_channel_list_reply))
>   		return -EINVAL;
>   
>   	ath6kl_wakeup_event(wmi->parent_dev);
> diff --git a/drivers/net/wireless/ath/ath6kl/wmi.h b/drivers/net/wireless/ath/ath6kl/wmi.h
> index 9e168752bec2..432e4f428a4a 100644
> --- a/drivers/net/wireless/ath/ath6kl/wmi.h
> +++ b/drivers/net/wireless/ath/ath6kl/wmi.h
> @@ -1373,7 +1373,7 @@ struct wmi_channel_list_reply {
>   	u8 num_ch;
>   
>   	/* channel in Mhz */
> -	__le16 ch_list[1];
> +	__le16 ch_list[];
>   } __packed;
>   
>   /* List of Events (target to host) */

whether or not you modify the length check consider this:
Reviewed-by: Jeff Johnson <quic_jjohnson@xxxxxxxxxxx>