On Wednesday 26 January 2011 15:30:08 Marek Lindner wrote:
> Signed-off-by: Marek Lindner <[email protected]>
> ---
> +     buff_len -= PRISM_HEADER_LEN;
> +     packet_buff += PRISM_HEADER_LEN;
> +
> +     /* we assume a minimum size of 38 bytes
> +      * (802.11 data frame + LLC)
> +      * before we calculate the real size */
> +     if (buff_len <= 38)
> +             return;
> +
> +     wifi_hdr = (struct ieee80211_hdr *)packet_buff;
> +     fc = wifi_hdr->frame_control;
> +
> +     /* not carrying payload */
> +     if ((fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA)
> +             return;

Didn't we say yesterday that we must either use ntohs or only read the first 
byte of the framecontrol? The last option which comes to my mind is to define 
IEEE80211_FCTL_FTYPE and IEEE80211_FTYPE_DATA for big and little endian 
independently.

The same for
IEEE80211_FCTL_TODS 0x0100
IEEE80211_FCTL_FROMDS 0x0200
IEEE80211_FCTL_PROTECTED 0x4000
IEEE80211_STYPE_QOS_DATA 0x0080

Your current version should only work on little endian systems. I would 
suggest following version:

#define IEEE80211_FCTL_FTYPE 0x0c00
#define IEEE80211_FTYPE_DATA 0x0800
#define IEEE80211_FCTL_TODS 0x0001
#define IEEE80211_FCTL_FROMDS 0x0002
#define IEEE80211_FCTL_PROTECTED 0x0040
#define IEEE80211_STYPE_QOS_DATA 0x8000

fc = ntohs(wifi_hdr->frame_control);


> +       if (fc & IEEE80211_STYPE_QOS_DATA)
> +               hdr_len += 2;

Here are you testing only on bit. Are you sure that the other 3 bits aren't 
interesting? At least one combination isn't well defined by the 802.11 
standard from 2007 and some others are null data types.

Best regards,
        Sven

Attachment: signature.asc
Description: This is a digitally signed message part.

Reply via email to