openwrt/package/kernel/mac80211/patches/358-mac80211-add-NEED_ALIGNED4_SKBS-hw-flag.patch

From: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Date: Fri, 19 Feb 2016 11:01:50 +0100
Subject: [PATCH] mac80211: add NEED_ALIGNED4_SKBS hw flag

HW/driver should set NEED_ALIGNED4_SKBS flag in case
require aligned skbs to four-byte boundaries.
This affect only TX direction.

Padding is added after ieee80211_hdr, before IV/LLC.

Before we have to do memmove(hdrlen) twice in the
dirver. Once before we pass this to HW and next
in tx completion (to be sure monitor will report
this tx frame correctly).

With this patch we can skip this memmove() and save CPU.

Currently this was tested with ath9k, both hw/sw crypt for
wep/tkip/ccmp.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
---

--- a/include/net/mac80211.h
+++ b/include/net/mac80211.h
@@ -2059,6 +2059,9 @@ struct ieee80211_txq {
  * @IEEE80211_HW_SUPPORTS_TDLS_BUFFER_STA: Hardware supports buffer STA on
  *	TDLS links.
  *
+ * @IEEE80211_HW_NEEDS_ALIGNED4_SKBS: Driver need aligned skbs to four-byte.
+ *	Padding will be added after ieee80211_hdr, before IV/LLC.
+ *
  * @NUM_IEEE80211_HW_FLAGS: number of hardware flags, used for sizing arrays
  */
 enum ieee80211_hw_flags {
@@ -2102,6 +2105,7 @@ enum ieee80211_hw_flags {
 	IEEE80211_HW_REPORTS_LOW_ACK,
 	IEEE80211_HW_SUPPORTS_TX_FRAG,
 	IEEE80211_HW_SUPPORTS_TDLS_BUFFER_STA,
+	IEEE80211_HW_NEEDS_ALIGNED4_SKBS,
 
 	/* keep last, obviously */
 	NUM_IEEE80211_HW_FLAGS
--- a/net/mac80211/debugfs.c
+++ b/net/mac80211/debugfs.c
@@ -212,6 +212,7 @@ static const char *hw_flag_names[] = {
 	FLAG(REPORTS_LOW_ACK),
 	FLAG(SUPPORTS_TX_FRAG),
 	FLAG(SUPPORTS_TDLS_BUFFER_STA),
+	FLAG(NEEDS_ALIGNED4_SKBS),
 #undef FLAG
 };
 
--- a/net/mac80211/ieee80211_i.h
+++ b/net/mac80211/ieee80211_i.h
@@ -1548,6 +1548,29 @@ ieee80211_vif_get_num_mcast_if(struct ie
 	return -1;
 }
 
+static inline unsigned int
+ieee80211_hdr_padsize(struct ieee80211_hw *hw, unsigned int hdrlen)
+{
+	/*
+	 * While hdrlen is already aligned to two-byte boundaries,
+	 * simple check with & 2 will return correct padsize.
+	 */
+	if (ieee80211_hw_check(hw, NEEDS_ALIGNED4_SKBS))
+		return hdrlen & 2;
+	return 0;
+}
+
+static inline unsigned int
+ieee80211_padded_hdrlen(struct ieee80211_hw *hw, __le16 fc)
+{
+	unsigned int hdrlen;
+
+	hdrlen = ieee80211_hdrlen(fc);
+	hdrlen += ieee80211_hdr_padsize(hw, hdrlen);
+
+	return hdrlen;
+}
+
 u64 ieee80211_calculate_rx_timestamp(struct ieee80211_local *local,
 				     struct ieee80211_rx_status *status,
 				     unsigned int mpdu_len,
--- a/net/mac80211/sta_info.h
+++ b/net/mac80211/sta_info.h
@@ -300,7 +300,7 @@ struct ieee80211_fast_tx {
 	u8 hdr_len;
 	u8 sa_offs, da_offs, pn_offs;
 	u8 band;
-	u8 hdr[30 + 2 + IEEE80211_FAST_XMIT_MAX_IV +
+	u8 hdr[30 + 2 + 2 + IEEE80211_FAST_XMIT_MAX_IV +
 	       sizeof(rfc1042_header)] __aligned(2);
 
 	struct rcu_head rcu_head;
--- a/net/mac80211/status.c
+++ b/net/mac80211/status.c
@@ -642,9 +642,22 @@ void ieee80211_tx_monitor(struct ieee802
 	struct sk_buff *skb2;
 	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
 	struct ieee80211_sub_if_data *sdata;
+	struct ieee80211_hdr *hdr = (void *)skb->data;
 	struct net_device *prev_dev = NULL;
+	unsigned int hdrlen, padsize;
 	int rtap_len;
 
+	/* Remove padding if was added */
+	if (ieee80211_hw_check(&local->hw, NEEDS_ALIGNED4_SKBS)) {
+		hdrlen = ieee80211_hdrlen(hdr->frame_control);
+		padsize = ieee80211_hdr_padsize(&local->hw, hdrlen);
+
+		if (padsize && skb->len > hdrlen + padsize) {
+			memmove(skb->data + padsize, skb->data, hdrlen);
+			skb_pull(skb, padsize);
+		}
+	}
+
 	/* send frame to monitor interfaces now */
 	rtap_len = ieee80211_tx_radiotap_len(info);
 	if (WARN_ON_ONCE(skb_headroom(skb) < rtap_len)) {
--- a/net/mac80211/tkip.c
+++ b/net/mac80211/tkip.c
@@ -201,10 +201,12 @@ void ieee80211_get_tkip_p2k(struct ieee8
 {
 	struct ieee80211_key *key = (struct ieee80211_key *)
 			container_of(keyconf, struct ieee80211_key, conf);
+	struct ieee80211_hw *hw = &key->local->hw;
 	const u8 *tk = &key->conf.key[NL80211_TKIP_DATA_OFFSET_ENCR_KEY];
 	struct tkip_ctx *ctx = &key->u.tkip.tx;
 	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
-	const u8 *data = (u8 *)hdr + ieee80211_hdrlen(hdr->frame_control);
+	const u8 *data = (u8 *)hdr + ieee80211_padded_hdrlen(hw,
+							hdr->frame_control);
 	u32 iv32 = get_unaligned_le32(&data[4]);
 	u16 iv16 = data[2] | (data[0] << 8);
 
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -1172,8 +1172,7 @@ ieee80211_tx_prepare(struct ieee80211_su
 	info->flags &= ~IEEE80211_TX_INTFL_NEED_TXPROCESSING;
 
 	hdr = (struct ieee80211_hdr *) skb->data;
-
-	tx->hdrlen = ieee80211_hdrlen(hdr->frame_control);
+	tx->hdrlen = ieee80211_padded_hdrlen(&local->hw, hdr->frame_control);
 
 	if (likely(sta)) {
 		if (!IS_ERR(sta))
@@ -2184,7 +2183,7 @@ netdev_tx_t ieee80211_monitor_start_xmit
 		goto fail;
 
 	hdr = (struct ieee80211_hdr *)(skb->data + len_rthdr);
-	hdrlen = ieee80211_hdrlen(hdr->frame_control);
+	hdrlen = ieee80211_padded_hdrlen(&local->hw, hdr->frame_control);
 
 	if (skb->len < len_rthdr + hdrlen)
 		goto fail;
@@ -2402,7 +2401,7 @@ static struct sk_buff *ieee80211_build_h
 	struct ieee80211_chanctx_conf *chanctx_conf;
 	struct ieee80211_sub_if_data *ap_sdata;
 	enum nl80211_band band;
-	int ret;
+	int padsize, ret;
 
 	if (IS_ERR(sta))
 		sta = NULL;
@@ -2622,6 +2621,9 @@ static struct sk_buff *ieee80211_build_h
 		hdrlen += 2;
 	}
 
+	/* Check aligned4 skb required */
+	padsize = ieee80211_hdr_padsize(&sdata->local->hw, hdrlen);
+
 	/*
 	 * Drop unicast frames to unauthorised stations unless they are
 	 * EAPOL frames from the local station.
@@ -2702,6 +2704,7 @@ static struct sk_buff *ieee80211_build_h
 
 	skb_pull(skb, skip_header_bytes);
 	head_need = hdrlen + encaps_len + meshhdrlen - skb_headroom(skb);
+	head_need += padsize;
 
 	/*
 	 * So we need to modify the skb header and hence need a copy of
@@ -2734,6 +2737,9 @@ static struct sk_buff *ieee80211_build_h
 		memcpy(skb_push(skb, meshhdrlen), &mesh_hdr, meshhdrlen);
 #endif
 
+	if (padsize)
+		memset(skb_push(skb, padsize), 0, padsize);
+
 	if (ieee80211_is_data_qos(fc)) {
 		__le16 *qos_control;
 
@@ -2909,6 +2915,9 @@ void ieee80211_check_fast_xmit(struct st
 		fc |= cpu_to_le16(IEEE80211_STYPE_QOS_DATA);
 	}
 
+	/* Check aligned4 skb required */
+	build.hdr_len += ieee80211_hdr_padsize(&local->hw, build.hdr_len);
+
 	/* We store the key here so there's no point in using rcu_dereference()
 	 * but that's fine because the code that changes the pointers will call
 	 * this function after doing so. For a single CPU that would be enough,
@@ -3495,7 +3504,7 @@ begin:
 
 		if (tx.key &&
 		    (tx.key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV))
-			pn_offs = ieee80211_hdrlen(hdr->frame_control);
+			pn_offs = tx.hdrlen;
 
 		ieee80211_xmit_fast_finish(sta->sdata, sta, pn_offs,
 					   tx.key, skb);
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -1230,6 +1230,7 @@ void ieee80211_send_auth(struct ieee8021
 			 u32 tx_flags)
 {
 	struct ieee80211_local *local = sdata->local;
+	struct ieee80211_hw *hw = &local->hw;
 	struct sk_buff *skb;
 	struct ieee80211_mgmt *mgmt;
 	unsigned int hdrlen;
@@ -1256,7 +1257,7 @@ void ieee80211_send_auth(struct ieee8021
 		skb_put_data(skb, extra, extra_len);
 
 	if (auth_alg == WLAN_AUTH_SHARED_KEY && transaction == 3) {
-		hdrlen = ieee80211_hdrlen(mgmt->frame_control);
+		hdrlen = ieee80211_padded_hdrlen(hw, mgmt->frame_control);
 		mgmt->frame_control |= cpu_to_le16(IEEE80211_FCTL_PROTECTED);
 		err = ieee80211_wep_encrypt(local, skb, hdrlen, key,
 					    key_len, key_idx);