| Message ID | CAMgrJ3bU5fPDGq=PUZKwXaOaVJDAeYOzi3rQZej5ZAJ070bfjQ@mail.gmail.com |
|---|---|
| State | Accepted |
| Commit | bff954e9a2ee051c23cf9a6dd576873b5dee8544 |
| Headers | show |
On Sun, Nov 18, 2012 at 11:17:37PM -0500, Robert Shade wrote: > This prevents situations like the following where we roam to a > lesser quality BSS just because the signal level delta is over our > threshold. Thanks, applied. > wlan0: Considering within-ESS reassociation > wlan0: Current BSS: 00:24:6c:74:0a:40 level=-51 > wlan0: Selected BSS: 00:24:6c:74:0a:e0 level=-64 Though, it would be interesting to know how the :e0 BSS got selected in this case.. Do the APs use different security parameters? wpa_scan_result_compar() is supposed to order the BSS entries so that this selection should not have happened in the first place..
diff --git a/wpa_supplicant/events.c b/wpa_supplicant/events.c index 42c4dea..a878735 100644 --- a/wpa_supplicant/events.c +++ b/wpa_supplicant/events.c @@ -1025,6 +1025,11 @@ static int wpa_supplicant_need_to_roam(struct wpa_supplicant *wpa_s, return 1; } + if ((current_bss->level < 0) && (current_bss->level > selected->level)) { + wpa_dbg(wpa_s, MSG_DEBUG, "Skip roam - Current BSS has better signal level"); + return 0; + } + min_diff = 2; if (current_bss->level < 0) {
This prevents situations like the following where we roam to a lesser quality BSS just because the signal level delta is over our threshold. wlan0: Considering within-ESS reassociation wlan0: Current BSS: 00:24:6c:74:0a:40 level=-51 wlan0: Selected BSS: 00:24:6c:74:0a:e0 level=-64 wlan0: Request association: reassociate: 0 selected: 00:24:6c:74:0a:e0 bssid: 00:24:6c:74:0a:40 pending: 00:00:00:00:00:00 wpa_state: COMPLETED Signed-hostap: Robert Shade <robert.shade@gmail.com> --- wpa_supplicant/events.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) if (current_bss->level < -85)