[3.19.y-ckt stable] Patch "mac80211: don't use napi_gro_receive() outside NAPI context" has been added to staging queue

Kamal Mostafa kamal at canonical.com
Tue Jul 7 00:10:37 UTC 2015


This is a note to let you know that I have just added a patch titled

    mac80211: don't use napi_gro_receive() outside NAPI context

to the linux-3.19.y-queue branch of the 3.19.y-ckt extended stable tree 
which can be found at:

    http://kernel.ubuntu.com/git/ubuntu/linux.git/log/?h=linux-3.19.y-queue

This patch is scheduled to be released in version 3.19.8-ckt3.

If you, or anyone else, feels it should not be added to this tree, please 
reply to this email.

For more information about the 3.19.y-ckt tree, see
https://wiki.ubuntu.com/Kernel/Dev/ExtendedStable

Thanks.
-Kamal

------

>From 717937c0ead730963b589f986f5229d6ec3c6731 Mon Sep 17 00:00:00 2001
From: Johannes Berg <johannes.berg at intel.com>
Date: Tue, 19 May 2015 15:40:21 +0200
Subject: mac80211: don't use napi_gro_receive() outside NAPI context

commit 22d3a3c829fa9ecdb493d1f1f2838d543f8d86a3 upstream.

No matter how the driver manages its NAPI context, there's no way
sending frames to it from a timer can be correct, since it would
corrupt the internal GRO lists.

To avoid that, always use the non-NAPI path when releasing frames
from the timer.

Reported-by: Jean Trivelly <jean.trivelly at intel.com>
Signed-off-by: Johannes Berg <johannes.berg at intel.com>
Signed-off-by: Kamal Mostafa <kamal at canonical.com>
---
 net/mac80211/ieee80211_i.h | 3 +++
 net/mac80211/rx.c          | 5 +++--
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h
index fa7568c..5ab2b63 100644
--- a/net/mac80211/ieee80211_i.h
+++ b/net/mac80211/ieee80211_i.h
@@ -204,6 +204,8 @@ enum ieee80211_packet_rx_flags {
  * @IEEE80211_RX_CMNTR: received on cooked monitor already
  * @IEEE80211_RX_BEACON_REPORTED: This frame was already reported
  *	to cfg80211_report_obss_beacon().
+ * @IEEE80211_RX_REORDER_TIMER: this frame is released by the
+ *	reorder buffer timeout timer, not the normal RX path
  *
  * These flags are used across handling multiple interfaces
  * for a single frame.
@@ -211,6 +213,7 @@ enum ieee80211_packet_rx_flags {
 enum ieee80211_rx_flags {
 	IEEE80211_RX_CMNTR		= BIT(0),
 	IEEE80211_RX_BEACON_REPORTED	= BIT(1),
+	IEEE80211_RX_REORDER_TIMER	= BIT(2),
 };

 struct ieee80211_rx_data {
diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c
index 909913d..c218688 100644
--- a/net/mac80211/rx.c
+++ b/net/mac80211/rx.c
@@ -2083,7 +2083,8 @@ ieee80211_deliver_skb(struct ieee80211_rx_data *rx)
 		/* deliver to local stack */
 		skb->protocol = eth_type_trans(skb, dev);
 		memset(skb->cb, 0, sizeof(skb->cb));
-		if (rx->local->napi)
+		if (!(rx->flags & IEEE80211_RX_REORDER_TIMER) &&
+		    rx->local->napi)
 			napi_gro_receive(rx->local->napi, skb);
 		else
 			netif_receive_skb(skb);
@@ -3178,7 +3179,7 @@ void ieee80211_release_reorder_timeout(struct sta_info *sta, int tid)
 		/* This is OK -- must be QoS data frame */
 		.security_idx = tid,
 		.seqno_idx = tid,
-		.flags = 0,
+		.flags = IEEE80211_RX_REORDER_TIMER,
 	};
 	struct tid_ampdu_rx *tid_agg_rx;

--
1.9.1





More information about the kernel-team mailing list