ath5k: add notes about rx timestamp
authorBruno Randolf <bruno@thinktube.com>
Wed, 5 Mar 2008 09:36:05 +0000 (18:36 +0900)
committerJohn W. Linville <linville@tuxdriver.com>
Fri, 7 Mar 2008 21:07:50 +0000 (16:07 -0500)
add comments about the fact that we don't know when exactly the atheros
hardware takes the RX timestamp.

drivers/net/wireless/ath5k/base.c:      Changes-licensed-under: 3-Clause-BSD

Signed-off-by: Bruno Randolf <bruno@thinktube.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
drivers/net/wireless/ath5k/base.c

index 73e7e61..5eacc7c 100644 (file)
@@ -1844,6 +1844,17 @@ accept:
                 * 15bit only. that means TSF extension has to be done within
                 * 32768usec (about 32ms). it might be necessary to move this to
                 * the interrupt handler, like it is done in madwifi.
+                *
+                * Unfortunately we don't know when the hardware takes the rx
+                * timestamp (beginning of phy frame, data frame, end of rx?).
+                * The only thing we know is that it is hardware specific...
+                * On AR5213 it seems the rx timestamp is at the end of the
+                * frame, but i'm not sure.
+                *
+                * NOTE: mac80211 defines mactime at the beginning of the first
+                * data symbol. Since we don't have any time references it's
+                * impossible to comply to that. This affects IBSS merge only
+                * right now, so it's not too bad...
                 */
                rxs.mactime = ath5k_extend_tsf(sc->ah, rs.rs_tstamp);
                rxs.flag |= RX_FLAG_TSFT;