Re: [PATCH] wl1251: Fix possible buffer overflow in wl1251_cmd_scan

[Kalle Valo <kvalo@codeaurora.org>]

Lee Gibson <leegib@gmail.com> writes:

> Function wl1251_cmd_scan calls memcpy without checking the length.
> A user could control that length and trigger a buffer overflow.
> Fix by checking the length is within the maximum allowed size.
>
> Signed-off-by: Lee Gibson <leegib@gmail.com>

Please fix the commit log, the user cannot control this length as
cfg80211 checks it before handling it to wl1251. Unless I'm missing
something.

> ---
>  drivers/net/wireless/ti/wl1251/cmd.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ti/wl1251/cmd.c b/drivers/net/wireless/ti/wl1251/cmd.c
> index 498c8db2eb48..e4d028a53d91 100644
> --- a/drivers/net/wireless/ti/wl1251/cmd.c
> +++ b/drivers/net/wireless/ti/wl1251/cmd.c
> @@ -455,8 +455,11 @@ int wl1251_cmd_scan(struct wl1251 *wl, u8 *ssid, size_t ssid_len,
>  	}
>  
>  	cmd->params.ssid_len = ssid_len;

If you are checking the length, you should also check ssid_len here.

> -	if (ssid)
> -		memcpy(cmd->params.ssid, ssid, ssid_len);
> +	if (ssid) {
> +		int len = min_t(int, ssid_len, IEEE80211_MAX_SSID_LEN);
> +
> +		memcpy(cmd->params.ssid, ssid, len);
> +	}

Please use clamp_val().

Also another (and IMHO better) way to cleanup this is to provide a
pointer to struct cfg80211_ssid, which makes it clear that the length
can be trusted and not length checking is not needed. So something like
this:

int wl1251_cmd_scan(struct wl1251 *wl, const struct cfg80211_ssid *ssid,
		    struct ieee80211_channel *channels[],
		    unsigned int n_channels, unsigned int n_probes)

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches