Merge branch 'ethtool-hdrsplit'

Jakub Kicinski says:

====================
ethtool: add header/data split indication

TCP ZC Rx requires data to be placed neatly into pages, separate
from the networking headers. This is not supported by most devices
so to make deployment easy this set adds a way for the driver to
report support for this feature thru ethtool.

The larger scope of configuring splitting headers and data, or DMA
scatter seems dauntingly broad, so this set focuses specifically
on the question "is this device usable with TCP ZC Rx?".

The aim is to avoid a litany of conditions on HW platforms, features,
and firmware versions in orchestration systems when the drivers can
easily tell their SG config.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

5 files changed, 30 insertions, 5 deletions

diff --git a/Documentation/networking/ethtool-netlink.rst b/Documentation/networking/ethtool-netlink.rst
index 9d98e0511249..cae28af7a476 100644
--- a/Documentation/networking/ethtool-netlink.rst
+++ b/Documentation/networking/ethtool-netlink.rst
@@ -860,8 +860,16 @@ Kernel response contents:
   ``ETHTOOL_A_RINGS_RX_JUMBO``          u32     size of RX jumbo ring
   ``ETHTOOL_A_RINGS_TX``                u32     size of TX ring
   ``ETHTOOL_A_RINGS_RX_BUF_LEN``        u32     size of buffers on the ring
+  ``ETHTOOL_A_RINGS_TCP_DATA_SPLIT``    u8      TCP header / data split
   ====================================  ======  ===========================
 
+``ETHTOOL_A_RINGS_TCP_DATA_SPLIT`` indicates whether the device is usable with
+page-flipping TCP zero-copy receive (``getsockopt(TCP_ZEROCOPY_RECEIVE)``).
+If enabled the device is configured to place frame headers and data into
+separate buffers. The device configuration must make it possible to receive
+full memory pages of data, for example because MTU is high enough or through
+HW-GRO.
+
 
 RINGS_SET
 =========
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index 003330e8cd58..5edbee92f5c4 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -11,6 +11,7 @@
 #include <linux/ctype.h>
 #include <linux/stringify.h>
 #include <linux/ethtool.h>
+#include <linux/ethtool_netlink.h>
 #include <linux/linkmode.h>
 #include <linux/interrupt.h>
 #include <linux/pci.h>
@@ -802,9 +803,11 @@ static void bnxt_get_ringparam(struct net_device *dev,
 	if (bp->flags & BNXT_FLAG_AGG_RINGS) {
 		ering->rx_max_pending = BNXT_MAX_RX_DESC_CNT_JUM_ENA;
 		ering->rx_jumbo_max_pending = BNXT_MAX_RX_JUM_DESC_CNT;
+		kernel_ering->tcp_data_split = ETHTOOL_TCP_DATA_SPLIT_ENABLED;
 	} else {
 		ering->rx_max_pending = BNXT_MAX_RX_DESC_CNT;
 		ering->rx_jumbo_max_pending = 0;
+		kernel_ering->tcp_data_split = ETHTOOL_TCP_DATA_SPLIT_DISABLED;
 	}
 	ering->tx_max_pending = BNXT_MAX_TX_DESC_CNT;
 
diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h
index 11efc45de66a..e0853f48b75e 100644
--- a/include/linux/ethtool.h
+++ b/include/linux/ethtool.h
@@ -70,9 +70,11 @@ enum {
 /**
  * struct kernel_ethtool_ringparam - RX/TX ring configuration
  * @rx_buf_len: Current length of buffers on the rx ring.
+ * @tcp_data_split: Scatter packet headers and data to separate buffers
  */
 struct kernel_ethtool_ringparam {
 	u32	rx_buf_len;
+	u8	tcp_data_split;
 };
 
 /**
diff --git a/include/uapi/linux/ethtool_netlink.h b/include/uapi/linux/ethtool_netlink.h
index cca6e474a085..417d4280d7b5 100644
--- a/include/uapi/linux/ethtool_netlink.h
+++ b/include/uapi/linux/ethtool_netlink.h
@@ -319,6 +319,12 @@ enum {
 /* RINGS */
 
 enum {
+	ETHTOOL_TCP_DATA_SPLIT_UNKNOWN = 0,
+	ETHTOOL_TCP_DATA_SPLIT_DISABLED,
+	ETHTOOL_TCP_DATA_SPLIT_ENABLED,
+};
+
+enum {
 	ETHTOOL_A_RINGS_UNSPEC,
 	ETHTOOL_A_RINGS_HEADER,				/* nest - _A_HEADER_* */
 	ETHTOOL_A_RINGS_RX_MAX,				/* u32 */
@@ -330,6 +336,7 @@ enum {
 	ETHTOOL_A_RINGS_RX_JUMBO,			/* u32 */
 	ETHTOOL_A_RINGS_TX,				/* u32 */
 	ETHTOOL_A_RINGS_RX_BUF_LEN,                     /* u32 */
+	ETHTOOL_A_RINGS_TCP_DATA_SPLIT,			/* u8 */
 
 	/* add new constants above here */
 	__ETHTOOL_A_RINGS_CNT,
diff --git a/net/ethtool/rings.c b/net/ethtool/rings.c
index c1d5f5e0fdc9..18a5035d3bee 100644
--- a/net/ethtool/rings.c
+++ b/net/ethtool/rings.c
@@ -53,7 +53,8 @@ static int rings_reply_size(const struct ethnl_req_info *req_base,
 	       nla_total_size(sizeof(u32)) +	/* _RINGS_RX_MINI */
 	       nla_total_size(sizeof(u32)) +	/* _RINGS_RX_JUMBO */
 	       nla_total_size(sizeof(u32)) +	/* _RINGS_TX */
-	       nla_total_size(sizeof(u32));     /* _RINGS_RX_BUF_LEN */
+	       nla_total_size(sizeof(u32)) +	/* _RINGS_RX_BUF_LEN */
+	       nla_total_size(sizeof(u8));	/* _RINGS_TCP_DATA_SPLIT */
 }
 
 static int rings_fill_reply(struct sk_buff *skb,
@@ -61,9 +62,11 @@ static int rings_fill_reply(struct sk_buff *skb,
 			    const struct ethnl_reply_data *reply_base)
 {
 	const struct rings_reply_data *data = RINGS_REPDATA(reply_base);
-	const struct kernel_ethtool_ringparam *kernel_ringparam = &data->kernel_ringparam;
+	const struct kernel_ethtool_ringparam *kr = &data->kernel_ringparam;
 	const struct ethtool_ringparam *ringparam = &data->ringparam;
 
+	WARN_ON(kr->tcp_data_split > ETHTOOL_TCP_DATA_SPLIT_ENABLED);
+
 	if ((ringparam->rx_max_pending &&
 	     (nla_put_u32(skb, ETHTOOL_A_RINGS_RX_MAX,
 			  ringparam->rx_max_pending) ||
@@ -84,9 +87,11 @@ static int rings_fill_reply(struct sk_buff *skb,
 			  ringparam->tx_max_pending) ||
 	      nla_put_u32(skb, ETHTOOL_A_RINGS_TX,
 			  ringparam->tx_pending)))  ||
-	    (kernel_ringparam->rx_buf_len &&
-	     (nla_put_u32(skb, ETHTOOL_A_RINGS_RX_BUF_LEN,
-			  kernel_ringparam->rx_buf_len))))
+	    (kr->rx_buf_len &&
+	     (nla_put_u32(skb, ETHTOOL_A_RINGS_RX_BUF_LEN, kr->rx_buf_len))) ||
+	    (kr->tcp_data_split &&
+	     (nla_put_u8(skb, ETHTOOL_A_RINGS_TCP_DATA_SPLIT,
+			 kr->tcp_data_split))))
 		return -EMSGSIZE;
 
 	return 0;