qcacld-3.0: Enable the TX orphan by ini key gEnableTxOrphan
Enable the TX orphan by ini key gEnableTxOrphan, which is protected by
TX flow control.
Change-Id: Ib8e6d88ea0c7fda62f2d13ae3c592866ddc82521
CRs-Fixed: 2115621
diff --git a/core/hdd/src/wlan_hdd_softap_tx_rx.c b/core/hdd/src/wlan_hdd_softap_tx_rx.c
index 85a7bc2..d8b90fd 100644
--- a/core/hdd/src/wlan_hdd_softap_tx_rx.c
+++ b/core/hdd/src/wlan_hdd_softap_tx_rx.c
@@ -199,7 +199,34 @@
static inline struct sk_buff *hdd_skb_orphan(struct hdd_adapter *pAdapter,
struct sk_buff *skb)
{
- if (pAdapter->tx_flow_low_watermark > 0)
+#if (LINUX_VERSION_CODE <= KERNEL_VERSION(3, 19, 0))
+ struct hdd_context *hdd_ctx = WLAN_HDD_GET_CTX(pAdapter);
+#endif
+ int need_orphan = 0;
+
+ if (pAdapter->tx_flow_low_watermark > 0) {
+#if (LINUX_VERSION_CODE > KERNEL_VERSION(3, 19, 0))
+ /*
+ * The TCP TX throttling logic is changed a little after
+ * 3.19-rc1 kernel, the TCP sending limit will be smaller,
+ * which will throttle the TCP packets to the host driver.
+ * The TCP UP LINK throughput will drop heavily. In order to
+ * fix this issue, need to orphan the socket buffer asap, which
+ * will call skb's destructor to notify the TCP stack that the
+ * SKB buffer is unowned. And then the TCP stack will pump more
+ * packets to host driver.
+ *
+ * The TX packets might be dropped for UDP case in the iperf
+ * testing. So need to be protected by follow control.
+ */
+ need_orphan = 1;
+#else
+ if (hdd_ctx->config->tx_orphan_enable)
+ need_orphan = 1;
+#endif
+ }
+
+ if (need_orphan)
skb_orphan(skb);
else
skb = skb_unshare(skb, GFP_ATOMIC);
@@ -219,9 +246,12 @@
struct sk_buff *skb) {
struct sk_buff *nskb;
+#if (LINUX_VERSION_CODE > KERNEL_VERSION(3, 19, 0))
struct hdd_context *hdd_ctx = WLAN_HDD_GET_CTX(pAdapter);
+#endif
nskb = skb_unshare(skb, GFP_ATOMIC);
+#if (LINUX_VERSION_CODE > KERNEL_VERSION(3, 19, 0))
if (unlikely(hdd_ctx->config->tx_orphan_enable) && (nskb == skb)) {
/*
* For UDP packets we want to orphan the packet to allow the app
@@ -231,6 +261,7 @@
++pAdapter->hdd_stats.hddTxRxStats.txXmitOrphaned;
skb_orphan(skb);
}
+#endif
return nskb;
}
#endif /* QCA_LL_LEGACY_TX_FLOW_CONTROL */
@@ -362,26 +393,7 @@
if (!qdf_nbuf_ipa_owned_get(skb)) {
#endif
-#if (LINUX_VERSION_CODE > KERNEL_VERSION(3, 19, 0))
- /*
- * The TCP TX throttling logic is changed a little after
- * 3.19-rc1 kernel, the TCP sending limit will be smaller,
- * which will throttle the TCP packets to the host driver.
- * The TCP UP LINK throughput will drop heavily. In order to
- * fix this issue, need to orphan the socket buffer asap, which
- * will call skb's destructor to notify the TCP stack that the
- * SKB buffer is unowned. And then the TCP stack will pump more
- * packets to host driver.
- *
- * The TX packets might be dropped for UDP case in the iperf
- * testing. So need to be protected by follow control.
- */
skb = hdd_skb_orphan(pAdapter, skb);
-#else
- /* Check if the buffer has enough header room */
- skb = skb_unshare(skb, GFP_ATOMIC);
-#endif
-
if (!skb)
goto drop_pkt_accounting;