FILS: Fix maximum NSS calculation for FD frame
Maximum NSS calculation assumed the host to be little endian while retrieving MCS values from HE capabilities which is incorrect. Use WPA_GET_LE16() instead. Add a check for HE as the current NSS calculation assumes HE support. Signed-off-by: Aloka Dixit <quic_alokad@quicinc.com>
This commit is contained in:
parent
ecae45ff66
commit
903e3a1e62
1 changed files with 26 additions and 3 deletions
|
@ -1410,14 +1410,37 @@ static u16 hostapd_fils_discovery_cap(struct hostapd_data *hapd)
|
|||
cap_info |= phy_index << FD_CAP_PHY_INDEX_SHIFT;
|
||||
cap_info |= chwidth << FD_CAP_BSS_CHWIDTH_SHIFT;
|
||||
|
||||
if (mode) {
|
||||
u16 *mcs = (u16 *) mode->he_capab[IEEE80211_MODE_AP].mcs;
|
||||
if (mode && phy_index == FD_CAP_PHY_INDEX_HE) {
|
||||
const u8 *he_mcs = mode->he_capab[IEEE80211_MODE_AP].mcs;
|
||||
int i;
|
||||
u16 nss = 0;
|
||||
u16 nss = 0, mcs[6];
|
||||
|
||||
os_memset(mcs, 0xffff, 6 * sizeof(u16));
|
||||
|
||||
if (mcs_nss_size == 4) {
|
||||
mcs[0] = WPA_GET_LE16(&he_mcs[0]);
|
||||
mcs[1] = WPA_GET_LE16(&he_mcs[2]);
|
||||
}
|
||||
|
||||
if (mcs_nss_size == 8) {
|
||||
mcs[2] = WPA_GET_LE16(&he_mcs[4]);
|
||||
mcs[3] = WPA_GET_LE16(&he_mcs[6]);
|
||||
}
|
||||
|
||||
if (mcs_nss_size == 12) {
|
||||
mcs[4] = WPA_GET_LE16(&he_mcs[8]);
|
||||
mcs[5] = WPA_GET_LE16(&he_mcs[10]);
|
||||
}
|
||||
|
||||
for (i = 0; i < HE_NSS_MAX_STREAMS; i++) {
|
||||
u16 nss_mask = 0x3 << (i * 2);
|
||||
|
||||
/*
|
||||
* If NSS values supported by RX and TX are different
|
||||
* then choose the smaller of the two as the maximum
|
||||
* supported NSS as that is the value supported by
|
||||
* both RX and TX.
|
||||
*/
|
||||
if (mcs_nss_size == 4 &&
|
||||
(((mcs[0] & nss_mask) == nss_mask) ||
|
||||
((mcs[1] & nss_mask) == nss_mask)))
|
||||
|
|
Loading…
Reference in a new issue