[RFC: which stable kernels?] b43: fix frequency reported on G-PHY with /new/ firmware

["Rafał Miłecki" <zajec5@gmail.com>]

---
This bug was there for years, however didn't cause any problems. It was
exposed recently by the 3.15 release by the:

commit 3afc2167f60a327a2c1e1e2600ef209a3c2b75b7
Author: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Date:   Tue Mar 4 16:50:13 2014 +0200

    cfg80211/mac80211: ignore signal if the frame was heard on wrong channel

Should we mark it for stable for 3.15 only? Or all olders kernels as well?
---
 drivers/net/wireless/b43/xmit.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/drivers/net/wireless/b43/xmit.c b/drivers/net/wireless/b43/xmit.c
index 4f38f19..6e6ef3f 100644
--- a/drivers/net/wireless/b43/xmit.c
+++ b/drivers/net/wireless/b43/xmit.c
@@ -811,9 +811,13 @@ void b43_rx(struct b43_wldev *dev, struct sk_buff *skb, const void *_rxhdr)
 		break;
 	case B43_PHYTYPE_G:
 		status.band = IEEE80211_BAND_2GHZ;
-		/* chanid is the radio channel cookie value as used
-		 * to tune the radio. */
-		status.freq = chanid + 2400;
+		/* Somewhere between 478.104 and 508.1084 firmware for G-PHY
+		 * has been modified to be compatible with N-PHY and others.
+		 */
+		if (dev->fw.rev >= 508)
+			status.freq = ieee80211_channel_to_frequency(chanid, status.band);
+		else
+			status.freq = chanid + 2400;
 		break;
 	case B43_PHYTYPE_N:
 	case B43_PHYTYPE_LP:
-- 
1.8.4.5