Patch "iwlwifi: mvm: fix RX SKB header size and align it properly" has been added to the 4.9-stable tree

This is a note to let you know that I've just added the patch titled

iwlwifi: mvm: fix RX SKB header size and align it properly

to the 4.9-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
iwlwifi-mvm-fix-rx-skb-header-size-and-align-it-properly.patch
and it can be found in the queue-4.9 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

From: Johannes Berg <[email protected]>
Date: Wed, 14 Dec 2016 13:48:04 +0100
Subject: iwlwifi: mvm: fix RX SKB header size and align it properly

From: Johannes Berg <[email protected]>

[ Upstream commit 5cddd05c9cbe420436799716d009bc0372ef8268 ]

When receiving a frame, we currently pull in sizeof(*hdr) plus some extra (crypto/snap), which is too much, most headers aren't actually sizeof(*hdr) since that takes into account the 4-address format but doesn't take into account QoS. As a result, a typical frame will have 4 bytes of the payload in the SKB header already.

Fix this by calculating the correct header length, and now that we have that, align the end of the SKB header to a multiple of 4 so that the IP header will be aligned properly when pulled in.

---
 drivers/net/wireless/intel/iwlwifi/mvm/rx.c |   18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

@@ -104,7 +104,20 @@ static void iwl_mvm_pass_packet_to_mac80
 					    u8 crypt_len,
 					    struct iwl_rx_cmd_buffer *rxb)
 {
-	unsigned int hdrlen, fraglen;
+	unsigned int hdrlen = ieee80211_hdrlen(hdr->frame_control);
+	unsigned int fraglen;
+
+	/*
+	 * The 'hdrlen' (plus the 8 bytes for the SNAP and the crypt_len,
+	 * but those are all multiples of 4 long) all goes away, but we
+	 * want the *end* of it, which is going to be the start of the IP
+	 * header, to be aligned when it gets pulled in.
+	 * The beginning of the skb->data is aligned on at least a 4-byte
+	 * boundary after allocation. Everything here is aligned at least
+	 * on a 2-byte boundary so we can just take hdrlen & 3 and pad by
+	 * the result.
+	 */
+	skb_reserve(skb, hdrlen & 3);
 
 	/* If frame is small enough to fit in skb->head, pull it completely.
 	 * If not, only pull ieee80211_hdr (including crypto if present, and

@@ -118,8 +131,7 @@ static void iwl_mvm_pass_packet_to_mac80
 	 * If the latter changes (there are efforts in the standards group
 	 * to do so) we should revisit this and ieee80211_data_to_8023().
 	 */
-	hdrlen = (len <= skb_tailroom(skb)) ? len :
-					      sizeof(*hdr) + crypt_len + 8;
+	hdrlen = (len <= skb_tailroom(skb)) ? len : hdrlen + crypt_len + 8;
 
 	memcpy(skb_put(skb, hdrlen), hdr, hdrlen);
 	fraglen = len - hdrlen;

Patches currently in stable-queue which might be from [email protected] are

queue-4.9/mac80211-remove-bug-when-interface-type-is-invalid.patch queue-4.9/mac80211_hwsim-enforce-ps_manual_poll-to-be-set-after-ps_enabled.patch queue-4.9/mac80211_hwsim-use-per-interface-power-level.patch queue-4.9/iwlwifi-mvm-fix-rx-skb-header-size-and-align-it-properly.patch