ath9k: fix erased ieee80211_rx_status.mactime

This patch addresses the erasure of ieee80211_rx_status.mactime.

After the mactime is set in "ath_rx_tasklet" it is immediately nulled in "ath9k_rx_skb_preprocess".

---
 drivers/net/wireless/ath/ath9k/recv.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

b/drivers/net/wireless/ath/ath9k/recv.c
index da0cfe9..ae4bd30 100644

@@ -1140,6 +1140,11 @@ int ath_rx_tasklet(struct ath_softc *sc, int

flush, bool hp)
if (flush)
goto requeue;

+ retval = ath9k_rx_skb_preprocess(common, hw, hdr, &rs,
+ rxs, &decrypt_error);
+ if (retval)
+ goto requeue;
+
rxs->mactime = (tsf & ~0xffffffffULL) | rs.rs_tstamp; if (rs.rs_tstamp > tsf_lower &&
unlikely(rs.rs_tstamp - tsf_lower > 0x10000000))
@@ -1149,11 +1154,6 @@ int ath_rx_tasklet(struct ath_softc *sc, int flush, bool hp)
unlikely(tsf_lower - rs.rs_tstamp > 0x10000000))
rxs->mactime += 0x100000000ULL;

- retval = ath9k_rx_skb_preprocess(common, hw, hdr, &rs,
- rxs, &decrypt_error);
- if (retval)
- goto requeue;
-
/* Ensure we always have an skb to requeue once we are done
* processing the current buffer's skb */
requeue_skb = ath_rxbuf_alloc(common, common->rx_bufsize, GFP_ATOMIC);

What effect does this have? What issue does this fix? Is this a fix which also needs to be propagated to the stable kernels, if so why? etc.

Your patch describes what it does, it doesn't say why its a good fix or what was affected before. Please clarify all this as best as you can so we can help ACK/NACK/propagate the patch to stable if it is appropriate.

Luis

The mactime is always set to 0. I encountered the problem in IBSS mode. In ieee80211_rx_bss_info (net/mac80211/ibss.c) the mactime is used to trigger an IBSS merge. At the moment the merge is always triggered.

I got the "beacon TSF higher than local TSF - IBSS merge with BSSID" log message with every beacon and so recognized this odd behaviour.

I think this has to be fixed and at least influences IBSS mode negatively.

Greetings,
Jan

Think? Does the patch actually fix something or do you think it fixes something? Next time please include all these details on your patch commit log mesage so its easy to understand the reasoning behind the patch.

Luis

It fixes the problem I described. Now the mactime is set correctly and the IBSS merge is done only once and by the correct station. I noticed no negative side effect and tested it with 10+ stations in IBSS mode.

Sorry for the inconvenience. Next time I will provide more information.

Greetings,
Jan

no problem, can you resubmit but add all these details to the commit log? ALso, is this a stable kernel fix, that is, does this fix need to be propagated to the stable kernels, 2.6.35, 2.6.34, 2.6.33, 2.6.32?

Luis