ixgbe: DCB set PFC high and low water marks per data sheet specs
Currently the high and low water marks for PFC are being set
conservatively for jumbo frames. This means the RX buffers
are being underutilized in the default 1500 MTU. This patch
fixes this so that the water marks are set as described in
the data sheet considering the MTU size.
The equation used is,
RTT * 1.44 + MTU * 1.44 + MTU
Where RTT is the round trip time and MTU is the max frame size
in KB. To avoid floating point arithmetic FC_HIGH_WATER is
defined
((((RTT + MTU) * 144) + 99) / 100) + MTU
This changes how the hardware field fc.low_water and
fc.high_water are used. With this change they are no longer
storing the actual low water and high water markers but are
storing the required head room in the buffer. This simplifies
the logic and we do not need to account for the size of the
buffer when setting the thresholds.
Testing with iperf and 16 threads showed a slight uptick in
throughput over a single traffic class .1-.2Gbps and a reduction
in pause frames. Without the patch a 30 second run would show
~10-15 pause frames being transmitted with the patch ~2-5 are
seen. Test were run back to back with 82599.
Note RXPBSIZE is in KB and low and high water marks fields are
also in KB. However the FCRT* registers are 32B granularity and
right shifted 5 into the register,
(((rx_pbsize - water_mark) * 1024) / 32) << 5
is the most explicit conversion here we simplify
(rx_pbsize - water_mark) * 32 << 5 = (rx_pbsize - water_mark) << 10
This patch updates the PFC thresholds and legacy FC thresholds.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
diff --git a/drivers/net/ixgbe/ixgbe.h b/drivers/net/ixgbe/ixgbe.h
index 018e143..4f98486 100644
--- a/drivers/net/ixgbe/ixgbe.h
+++ b/drivers/net/ixgbe/ixgbe.h
@@ -61,10 +61,8 @@
#define IXGBE_MIN_RXD 64
/* flow control */
-#define IXGBE_DEFAULT_FCRTL 0x10000
#define IXGBE_MIN_FCRTL 0x40
#define IXGBE_MAX_FCRTL 0x7FF80
-#define IXGBE_DEFAULT_FCRTH 0x20000
#define IXGBE_MIN_FCRTH 0x600
#define IXGBE_MAX_FCRTH 0x7FFF0
#define IXGBE_DEFAULT_FCPAUSE 0xFFFF
diff --git a/drivers/net/ixgbe/ixgbe_82598.c b/drivers/net/ixgbe/ixgbe_82598.c
index 9c02d60..25b20f9 100644
--- a/drivers/net/ixgbe/ixgbe_82598.c
+++ b/drivers/net/ixgbe/ixgbe_82598.c
@@ -357,6 +357,7 @@
u32 fctrl_reg;
u32 rmcs_reg;
u32 reg;
+ u32 rx_pba_size;
u32 link_speed = 0;
bool link_up;
@@ -459,16 +460,18 @@
/* Set up and enable Rx high/low water mark thresholds, enable XON. */
if (hw->fc.current_mode & ixgbe_fc_tx_pause) {
- if (hw->fc.send_xon) {
- IXGBE_WRITE_REG(hw, IXGBE_FCRTL(packetbuf_num),
- (hw->fc.low_water | IXGBE_FCRTL_XONE));
- } else {
- IXGBE_WRITE_REG(hw, IXGBE_FCRTL(packetbuf_num),
- hw->fc.low_water);
- }
+ rx_pba_size = IXGBE_READ_REG(hw, IXGBE_RXPBSIZE(packetbuf_num));
+ rx_pba_size >>= IXGBE_RXPBSIZE_SHIFT;
- IXGBE_WRITE_REG(hw, IXGBE_FCRTH(packetbuf_num),
- (hw->fc.high_water | IXGBE_FCRTH_FCEN));
+ reg = (rx_pba_size - hw->fc.low_water) << 6;
+ if (hw->fc.send_xon)
+ reg |= IXGBE_FCRTL_XONE;
+ IXGBE_WRITE_REG(hw, IXGBE_FCRTL(packetbuf_num), reg);
+
+ reg = (rx_pba_size - hw->fc.high_water) << 10;
+ reg |= IXGBE_FCRTH_FCEN;
+
+ IXGBE_WRITE_REG(hw, IXGBE_FCRTH(packetbuf_num), reg);
}
/* Configure pause time (2 TCs per register) */
diff --git a/drivers/net/ixgbe/ixgbe_common.c b/drivers/net/ixgbe/ixgbe_common.c
index e3eca13..62aa2be 100644
--- a/drivers/net/ixgbe/ixgbe_common.c
+++ b/drivers/net/ixgbe/ixgbe_common.c
@@ -1595,6 +1595,7 @@
u32 mflcn_reg, fccfg_reg;
u32 reg;
u32 rx_pba_size;
+ u32 fcrtl, fcrth;
#ifdef CONFIG_DCB
if (hw->fc.requested_mode == ixgbe_fc_pfc)
@@ -1671,41 +1672,21 @@
IXGBE_WRITE_REG(hw, IXGBE_MFLCN, mflcn_reg);
IXGBE_WRITE_REG(hw, IXGBE_FCCFG, fccfg_reg);
- reg = IXGBE_READ_REG(hw, IXGBE_MTQC);
- /* Thresholds are different for link flow control when in DCB mode */
- if (reg & IXGBE_MTQC_RT_ENA) {
- rx_pba_size = IXGBE_READ_REG(hw, IXGBE_RXPBSIZE(packetbuf_num));
+ rx_pba_size = IXGBE_READ_REG(hw, IXGBE_RXPBSIZE(packetbuf_num));
+ rx_pba_size >>= IXGBE_RXPBSIZE_SHIFT;
- /* Always disable XON for LFC when in DCB mode */
- reg = (rx_pba_size >> 5) & 0xFFE0;
- IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(packetbuf_num), reg);
+ fcrth = (rx_pba_size - hw->fc.high_water) << 10;
+ fcrtl = (rx_pba_size - hw->fc.low_water) << 10;
- reg = (rx_pba_size >> 2) & 0xFFE0;
- if (hw->fc.current_mode & ixgbe_fc_tx_pause)
- reg |= IXGBE_FCRTH_FCEN;
- IXGBE_WRITE_REG(hw, IXGBE_FCRTH_82599(packetbuf_num), reg);
- } else {
- /*
- * Set up and enable Rx high/low water mark thresholds,
- * enable XON.
- */
- if (hw->fc.current_mode & ixgbe_fc_tx_pause) {
- if (hw->fc.send_xon) {
- IXGBE_WRITE_REG(hw,
- IXGBE_FCRTL_82599(packetbuf_num),
- (hw->fc.low_water |
- IXGBE_FCRTL_XONE));
- } else {
- IXGBE_WRITE_REG(hw,
- IXGBE_FCRTL_82599(packetbuf_num),
- hw->fc.low_water);
- }
-
- IXGBE_WRITE_REG(hw, IXGBE_FCRTH_82599(packetbuf_num),
- (hw->fc.high_water | IXGBE_FCRTH_FCEN));
- }
+ if (hw->fc.current_mode & ixgbe_fc_tx_pause) {
+ fcrth |= IXGBE_FCRTH_FCEN;
+ if (hw->fc.send_xon)
+ fcrtl |= IXGBE_FCRTL_XONE;
}
+ IXGBE_WRITE_REG(hw, IXGBE_FCRTH_82599(packetbuf_num), fcrth);
+ IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(packetbuf_num), fcrtl);
+
/* Configure pause time (2 TCs per register) */
reg = IXGBE_READ_REG(hw, IXGBE_FCTTV(packetbuf_num / 2));
if ((packetbuf_num & 1) == 0)
diff --git a/drivers/net/ixgbe/ixgbe_dcb_82598.c b/drivers/net/ixgbe/ixgbe_dcb_82598.c
index 50288bc..9a5e89c 100644
--- a/drivers/net/ixgbe/ixgbe_dcb_82598.c
+++ b/drivers/net/ixgbe/ixgbe_dcb_82598.c
@@ -256,21 +256,17 @@
* for each traffic class.
*/
for (i = 0; i < MAX_TRAFFIC_CLASS; i++) {
- if (dcb_config->rx_pba_cfg == pba_equal) {
- rx_pba_size = IXGBE_RXPBSIZE_64KB;
- } else {
- rx_pba_size = (i < 4) ? IXGBE_RXPBSIZE_80KB
- : IXGBE_RXPBSIZE_48KB;
- }
+ rx_pba_size = IXGBE_READ_REG(hw, IXGBE_RXPBSIZE(i));
+ rx_pba_size >>= IXGBE_RXPBSIZE_SHIFT;
+ reg = (rx_pba_size - hw->fc.low_water) << 10;
- reg = ((rx_pba_size >> 5) & 0xFFF0);
if (dcb_config->tc_config[i].dcb_pfc == pfc_enabled_tx ||
dcb_config->tc_config[i].dcb_pfc == pfc_enabled_full)
reg |= IXGBE_FCRTL_XONE;
IXGBE_WRITE_REG(hw, IXGBE_FCRTL(i), reg);
- reg = ((rx_pba_size >> 2) & 0xFFF0);
+ reg = (rx_pba_size - hw->fc.high_water) << 10;
if (dcb_config->tc_config[i].dcb_pfc == pfc_enabled_tx ||
dcb_config->tc_config[i].dcb_pfc == pfc_enabled_full)
reg |= IXGBE_FCRTH_FCEN;
diff --git a/drivers/net/ixgbe/ixgbe_dcb_82599.c b/drivers/net/ixgbe/ixgbe_dcb_82599.c
index 05f2247..374e1f7 100644
--- a/drivers/net/ixgbe/ixgbe_dcb_82599.c
+++ b/drivers/net/ixgbe/ixgbe_dcb_82599.c
@@ -251,19 +251,17 @@
/* Configure PFC Tx thresholds per TC */
for (i = 0; i < MAX_TRAFFIC_CLASS; i++) {
- if (dcb_config->rx_pba_cfg == pba_equal)
- rx_pba_size = IXGBE_RXPBSIZE_64KB;
- else
- rx_pba_size = (i < 4) ? IXGBE_RXPBSIZE_80KB
- : IXGBE_RXPBSIZE_48KB;
+ rx_pba_size = IXGBE_READ_REG(hw, IXGBE_RXPBSIZE(i));
+ rx_pba_size >>= IXGBE_RXPBSIZE_SHIFT;
- reg = ((rx_pba_size >> 5) & 0xFFE0);
+ reg = (rx_pba_size - hw->fc.low_water) << 10;
+
if (dcb_config->tc_config[i].dcb_pfc == pfc_enabled_full ||
dcb_config->tc_config[i].dcb_pfc == pfc_enabled_tx)
reg |= IXGBE_FCRTL_XONE;
IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(i), reg);
- reg = ((rx_pba_size >> 2) & 0xFFE0);
+ reg = (rx_pba_size - hw->fc.high_water) << 10;
if (dcb_config->tc_config[i].dcb_pfc == pfc_enabled_full ||
dcb_config->tc_config[i].dcb_pfc == pfc_enabled_tx)
reg |= IXGBE_FCRTH_FCEN;
diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c
index a137f9d..f374207 100644
--- a/drivers/net/ixgbe/ixgbe_main.c
+++ b/drivers/net/ixgbe/ixgbe_main.c
@@ -4854,6 +4854,7 @@
int j;
struct tc_configuration *tc;
#endif
+ int max_frame = dev->mtu + ETH_HLEN + ETH_FCS_LEN;
/* PCI config space info */
@@ -4930,8 +4931,8 @@
#ifdef CONFIG_DCB
adapter->last_lfc_mode = hw->fc.current_mode;
#endif
- hw->fc.high_water = IXGBE_DEFAULT_FCRTH;
- hw->fc.low_water = IXGBE_DEFAULT_FCRTL;
+ hw->fc.high_water = FC_HIGH_WATER(max_frame);
+ hw->fc.low_water = FC_LOW_WATER(max_frame);
hw->fc.pause_time = IXGBE_DEFAULT_FCPAUSE;
hw->fc.send_xon = true;
hw->fc.disable_fc_autoneg = false;
@@ -5193,6 +5194,7 @@
static int ixgbe_change_mtu(struct net_device *netdev, int new_mtu)
{
struct ixgbe_adapter *adapter = netdev_priv(netdev);
+ struct ixgbe_hw *hw = &adapter->hw;
int max_frame = new_mtu + ETH_HLEN + ETH_FCS_LEN;
/* MTU < 68 is an error and causes problems on some kernels */
@@ -5203,6 +5205,9 @@
/* must set new MTU before calling down or up */
netdev->mtu = new_mtu;
+ hw->fc.high_water = FC_HIGH_WATER(max_frame);
+ hw->fc.low_water = FC_LOW_WATER(max_frame);
+
if (netif_running(netdev))
ixgbe_reinit_locked(adapter);
diff --git a/drivers/net/ixgbe/ixgbe_type.h b/drivers/net/ixgbe/ixgbe_type.h
index d3cc6ce..96dea77 100644
--- a/drivers/net/ixgbe/ixgbe_type.h
+++ b/drivers/net/ixgbe/ixgbe_type.h
@@ -2113,6 +2113,14 @@
#define IXGBE_PHYSICAL_LAYER_10GBASE_XAUI 0x1000
#define IXGBE_PHYSICAL_LAYER_SFP_ACTIVE_DA 0x2000
+/* Flow Control Macros */
+#define PAUSE_RTT 8
+#define PAUSE_MTU(MTU) ((MTU + 1024 - 1) / 1024)
+
+#define FC_HIGH_WATER(MTU) ((((PAUSE_RTT + PAUSE_MTU(MTU)) * 144) + 99) / 100 +\
+ PAUSE_MTU(MTU))
+#define FC_LOW_WATER(MTU) (2 * (2 * PAUSE_MTU(MTU) + PAUSE_RTT))
+
/* Software ATR hash keys */
#define IXGBE_ATR_BUCKET_HASH_KEY 0xE214AD3D
#define IXGBE_ATR_SIGNATURE_HASH_KEY 0x14364D17