From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Johannes Berg , Lorenzo Bianconi , Sasha Levin Subject: [PATCH 4.14 21/94] mac80211: always account for A-MSDU header changes Date: Mon, 8 Oct 2018 20:31:02 +0200 Message-Id: <20181008175606.048800057@linuxfoundation.org> In-Reply-To: <20181008175605.067676667@linuxfoundation.org> References: <20181008175605.067676667@linuxfoundation.org> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: 4.14-stable review patch. If anyone has any objections, please let me know. ------------------ From: Johannes Berg [ Upstream commit aa58acf325b4aadeecae2bfc90658273b47dbace ] In the error path of changing the SKB headroom of the second A-MSDU subframe, we would not account for the already-changed length of the first frame that just got converted to be in A-MSDU format and thus is a bit longer now. Fix this by doing the necessary accounting. It would be possible to reorder the operations, but that would make the code more complex (to calculate the necessary pad), and the headroom expansion should not fail frequently enough to make that worthwhile. Fixes: 6e0456b54545 ("mac80211: add A-MSDU tx support") Signed-off-by: Johannes Berg Acked-by: Lorenzo Bianconi Signed-off-by: Johannes Berg Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman --- net/mac80211/tx.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -3188,7 +3188,7 @@ static bool ieee80211_amsdu_aggregate(st if (!ieee80211_amsdu_realloc_pad(local, skb, sizeof(rfc1042_header) + 2 + pad)) - goto out; + goto out_recalc; ret = true; data = skb_push(skb, ETH_ALEN + 2); @@ -3205,11 +3205,13 @@ static bool ieee80211_amsdu_aggregate(st head->data_len += skb->len; *frag_tail = skb; - flow->backlog += head->len - orig_len; - tin->backlog_bytes += head->len - orig_len; - - fq_recalc_backlog(fq, tin, flow); +out_recalc: + if (head->len != orig_len) { + flow->backlog += head->len - orig_len; + tin->backlog_bytes += head->len - orig_len; + fq_recalc_backlog(fq, tin, flow); + } out: spin_unlock_bh(&fq->lock);