Merge branch '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
1GbE Intel Wired LAN Driver Updates 2019-10-21

This series contains updates to e1000e and igc only.

Sasha adds stream control transmission protocol (SCTP) CRC checksum
support for igc.  Also added S0ix support to the e1000e driver.  Then
added multicast support by adding the address list to the MTA table and
providing the option for IPv6 address for igc.  In addition, added
receive checksum support to igc as well.  Lastly, cleaned up some code
that was not fully implemented yet for the VLAN filter table array.

v2: Dropped patch 1 & 2 from the original series.  Patch 1 is being sent
    to 'net' tree as a fix and patch 2 implementation needs to be
    re-worked.  Updated the patch to add support for S0ix to fix the
    reverse Xmas tree issues and made the entry/exit functions void
    since they constantly returned success.  All based on community
    feedback.
v3: Cleaned up patch 4 of the series based on feedback from the
    community.  Cleaned up a stray comma in a code comment and removed
    the 'inline' of a function that would be inlined by the compiler
    anyways.
====================

Signed-off-by: Jakub Kicinski <[email protected]>

Author: Jakub Kicinski

drivers/net/ethernet/intel/e1000e/netdev.c

@@ -6294,6 +6294,174 @@ static void e1000e_flush_lpic(struct pci_dev *pdev)
 	pm_runtime_put_sync(netdev->dev.parent);
 }
 
+/* S0ix implementation */
+static void e1000e_s0ix_entry_flow(struct e1000_adapter *adapter)
+{
+	struct e1000_hw *hw = &adapter->hw;
+	u32 mac_data;
+	u16 phy_data;
+
+	/* Disable the periodic inband message,
+	 * don't request PCIe clock in K1 page770_17[10:9] = 10b
+	 */
+	e1e_rphy(hw, HV_PM_CTRL, &phy_data);
+	phy_data &= ~HV_PM_CTRL_K1_CLK_REQ;
+	phy_data |= BIT(10);
+	e1e_wphy(hw, HV_PM_CTRL, phy_data);
+
+	/* Make sure we don't exit K1 every time a new packet arrives
+	 * 772_29[5] = 1 CS_Mode_Stay_In_K1
+	 */
+	e1e_rphy(hw, I217_CGFREG, &phy_data);
+	phy_data |= BIT(5);
+	e1e_wphy(hw, I217_CGFREG, phy_data);
+
+	/* Change the MAC/PHY interface to SMBus
+	 * Force the SMBus in PHY page769_23[0] = 1
+	 * Force the SMBus in MAC CTRL_EXT[11] = 1
+	 */
+	e1e_rphy(hw, CV_SMB_CTRL, &phy_data);
+	phy_data |= CV_SMB_CTRL_FORCE_SMBUS;
+	e1e_wphy(hw, CV_SMB_CTRL, phy_data);
+	mac_data = er32(CTRL_EXT);
+	mac_data |= E1000_CTRL_EXT_FORCE_SMBUS;
+	ew32(CTRL_EXT, mac_data);
+
+	/* DFT control: PHY bit: page769_20[0] = 1
+	 * Gate PPW via EXTCNF_CTRL - set 0x0F00[7] = 1
+	 */
+	e1e_rphy(hw, I82579_DFT_CTRL, &phy_data);
+	phy_data |= BIT(0);
+	e1e_wphy(hw, I82579_DFT_CTRL, phy_data);
+
+	mac_data = er32(EXTCNF_CTRL);
+	mac_data |= E1000_EXTCNF_CTRL_GATE_PHY_CFG;
+	ew32(EXTCNF_CTRL, mac_data);
+
+	/* Check MAC Tx/Rx packet buffer pointers.
+	 * Reset MAC Tx/Rx packet buffer pointers to suppress any
+	 * pending traffic indication that would prevent power gating.
+	 */
+	mac_data = er32(TDFH);
+	if (mac_data)
+		ew32(TDFH, 0);
+	mac_data = er32(TDFT);
+	if (mac_data)
+		ew32(TDFT, 0);
+	mac_data = er32(TDFHS);
+	if (mac_data)
+		ew32(TDFHS, 0);
+	mac_data = er32(TDFTS);
+	if (mac_data)
+		ew32(TDFTS, 0);
+	mac_data = er32(TDFPC);
+	if (mac_data)
+		ew32(TDFPC, 0);
+	mac_data = er32(RDFH);
+	if (mac_data)
+		ew32(RDFH, 0);
+	mac_data = er32(RDFT);
+	if (mac_data)
+		ew32(RDFT, 0);
+	mac_data = er32(RDFHS);
+	if (mac_data)
+		ew32(RDFHS, 0);
+	mac_data = er32(RDFTS);
+	if (mac_data)
+		ew32(RDFTS, 0);
+	mac_data = er32(RDFPC);
+	if (mac_data)
+		ew32(RDFPC, 0);
+
+	/* Enable the Dynamic Power Gating in the MAC */
+	mac_data = er32(FEXTNVM7);
+	mac_data |= BIT(22);
+	ew32(FEXTNVM7, mac_data);
+
+	/* Disable the time synchronization clock */
+	mac_data = er32(FEXTNVM7);
+	mac_data |= BIT(31);
+	mac_data &= ~BIT(0);
+	ew32(FEXTNVM7, mac_data);
+
+	/* Dynamic Power Gating Enable */
+	mac_data = er32(CTRL_EXT);
+	mac_data |= BIT(3);
+	ew32(CTRL_EXT, mac_data);
+
+	/* Enable the Dynamic Clock Gating in the DMA and MAC */
+	mac_data = er32(CTRL_EXT);
+	mac_data |= E1000_CTRL_EXT_DMA_DYN_CLK_EN;
+	ew32(CTRL_EXT, mac_data);
+
+	/* No MAC DPG gating SLP_S0 in modern standby
+	 * Switch the logic of the lanphypc to use PMC counter
+	 */
+	mac_data = er32(FEXTNVM5);
+	mac_data |= BIT(7);
+	ew32(FEXTNVM5, mac_data);
+}
+
+static void e1000e_s0ix_exit_flow(struct e1000_adapter *adapter)
+{
+	struct e1000_hw *hw = &adapter->hw;
+	u32 mac_data;
+	u16 phy_data;
+
+	/* Disable the Dynamic Power Gating in the MAC */
+	mac_data = er32(FEXTNVM7);
+	mac_data &= 0xFFBFFFFF;
+	ew32(FEXTNVM7, mac_data);
+
+	/* Enable the time synchronization clock */
+	mac_data = er32(FEXTNVM7);
+	mac_data |= BIT(0);
+	ew32(FEXTNVM7, mac_data);
+
+	/* Disable Dynamic Power Gating */
+	mac_data = er32(CTRL_EXT);
+	mac_data &= 0xFFFFFFF7;
+	ew32(CTRL_EXT, mac_data);
+
+	/* Disable the Dynamic Clock Gating in the DMA and MAC */
+	mac_data = er32(CTRL_EXT);
+	mac_data &= 0xFFF7FFFF;
+	ew32(CTRL_EXT, mac_data);
+
+	/* Revert the lanphypc logic to use the internal Gbe counter
+	 * and not the PMC counter
+	 */
+	mac_data = er32(FEXTNVM5);
+	mac_data &= 0xFFFFFF7F;
+	ew32(FEXTNVM5, mac_data);
+
+	/* Enable the periodic inband message,
+	 * Request PCIe clock in K1 page770_17[10:9] =01b
+	 */
+	e1e_rphy(hw, HV_PM_CTRL, &phy_data);
+	phy_data &= 0xFBFF;
+	phy_data |= HV_PM_CTRL_K1_CLK_REQ;
+	e1e_wphy(hw, HV_PM_CTRL, phy_data);
+
+	/* Return back configuration
+	 * 772_29[5] = 0 CS_Mode_Stay_In_K1
+	 */
+	e1e_rphy(hw, I217_CGFREG, &phy_data);
+	phy_data &= 0xFFDF;
+	e1e_wphy(hw, I217_CGFREG, phy_data);
+
+	/* Change the MAC/PHY interface to Kumeran
+	 * Unforce the SMBus in PHY page769_23[0] = 0
+	 * Unforce the SMBus in MAC CTRL_EXT[11] = 0
+	 */
+	e1e_rphy(hw, CV_SMB_CTRL, &phy_data);
+	phy_data &= ~CV_SMB_CTRL_FORCE_SMBUS;
+	e1e_wphy(hw, CV_SMB_CTRL, phy_data);
+	mac_data = er32(CTRL_EXT);
+	mac_data &= ~E1000_CTRL_EXT_FORCE_SMBUS;
+	ew32(CTRL_EXT, mac_data);
+}
+
 static int e1000e_pm_freeze(struct device *dev)
 {
 	struct net_device *netdev = dev_get_drvdata(dev);
@@ -6649,7 +6817,10 @@ static int e1000e_pm_thaw(struct device *dev)
 
 static int e1000e_pm_suspend(struct device *dev)
 {
+	struct net_device *netdev = pci_get_drvdata(to_pci_dev(dev));
+	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct pci_dev *pdev = to_pci_dev(dev);
+	struct e1000_hw *hw = &adapter->hw;
 	int rc;
 
 	e1000e_flush_lpic(pdev);
@@ -6660,14 +6831,25 @@ static int e1000e_pm_suspend(struct device *dev)
 	if (rc)
 		e1000e_pm_thaw(dev);
 
+	/* Introduce S0ix implementation */
+	if (hw->mac.type >= e1000_pch_cnp)
+		e1000e_s0ix_entry_flow(adapter);
+
 	return rc;
 }
 
 static int e1000e_pm_resume(struct device *dev)
 {
+	struct net_device *netdev = pci_get_drvdata(to_pci_dev(dev));
+	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct pci_dev *pdev = to_pci_dev(dev);
+	struct e1000_hw *hw = &adapter->hw;
 	int rc;
 
+	/* Introduce S0ix implementation */
+	if (hw->mac.type >= e1000_pch_cnp)
+		e1000e_s0ix_exit_flow(adapter);
+
 	rc = __e1000_resume(pdev);
 	if (rc)
 		return rc;

drivers/net/ethernet/intel/e1000e/regs.h

@@ -18,6 +18,7 @@
 #define E1000_FEXTNVM	0x00028	/* Future Extended NVM - RW */
 #define E1000_FEXTNVM3	0x0003C	/* Future Extended NVM 3 - RW */
 #define E1000_FEXTNVM4	0x00024	/* Future Extended NVM 4 - RW */
+#define E1000_FEXTNVM5	0x00014	/* Future Extended NVM 5 - RW */
 #define E1000_FEXTNVM6	0x00010	/* Future Extended NVM 6 - RW */
 #define E1000_FEXTNVM7	0x000E4	/* Future Extended NVM 7 - RW */
 #define E1000_FEXTNVM9	0x5BB4	/* Future Extended NVM 9 - RW */
@@ -234,4 +235,7 @@
 #define E1000_RXMTRL	0x0B634	/* Time sync Rx EtherType and Msg Type - RW */
 #define E1000_RXUDP	0x0B638	/* Time Sync Rx UDP Port - RW */
 
+/* PHY registers */
+#define I82579_DFT_CTRL	PHY_REG(769, 20)
+
 #endif

drivers/net/ethernet/intel/igc/igc.h

@@ -411,7 +411,6 @@ struct igc_adapter {
 	u32 tx_hwtstamp_timeouts;
 	u32 tx_hwtstamp_skipped;
 	u32 rx_hwtstamp_cleared;
-	u32 *shadow_vfta;
 
 	u32 rss_queues;
 	u32 rss_indir_tbl_init;

drivers/net/ethernet/intel/igc/igc_defines.h

@@ -282,7 +282,10 @@
 #define IGC_RCTL_BAM		0x00008000 /* broadcast enable */
 
 /* Receive Descriptor bit definitions */
-#define IGC_RXD_STAT_EOP	0x02    /* End of Packet */
+#define IGC_RXD_STAT_EOP	0x02	/* End of Packet */
+#define IGC_RXD_STAT_IXSM	0x04	/* Ignore checksum */
+#define IGC_RXD_STAT_UDPCS	0x10	/* UDP xsum calculated */
+#define IGC_RXD_STAT_TCPCS	0x20	/* TCP xsum calculated */
 
 #define IGC_RXDEXT_STATERR_CE		0x01000000
 #define IGC_RXDEXT_STATERR_SE		0x02000000
@@ -402,4 +405,7 @@
 #define IGC_ADVTXD_TUCMD_L4T_TCP	0x00000800  /* L4 Packet Type of TCP */
 #define IGC_ADVTXD_TUCMD_L4T_SCTP	0x00001000 /* L4 packet TYPE of SCTP */
 
+/* Maximum size of the MTA register table in all supported adapters */
+#define MAX_MTA_REG			128
+
 #endif /* _IGC_DEFINES_H_ */

drivers/net/ethernet/intel/igc/igc_hw.h

@@ -91,6 +91,7 @@ struct igc_mac_info {
 	u16 mta_reg_count;
 	u16 uta_reg_count;
 
+	u32 mta_shadow[MAX_MTA_REG];
 	u16 rar_entry_count;
 
 	u8 forced_speed_duplex;

drivers/net/ethernet/intel/igc/igc_mac.c

@@ -784,3 +784,107 @@ bool igc_enable_mng_pass_thru(struct igc_hw *hw)
 out:
 	return ret_val;
 }
+
+/**
+ *  igc_hash_mc_addr - Generate a multicast hash value
+ *  @hw: pointer to the HW structure
+ *  @mc_addr: pointer to a multicast address
+ *
+ *  Generates a multicast address hash value which is used to determine
+ *  the multicast filter table array address and new table value.  See
+ *  igc_mta_set()
+ **/
+static u32 igc_hash_mc_addr(struct igc_hw *hw, u8 *mc_addr)
+{
+	u32 hash_value, hash_mask;
+	u8 bit_shift = 0;
+
+	/* Register count multiplied by bits per register */
+	hash_mask = (hw->mac.mta_reg_count * 32) - 1;
+
+	/* For a mc_filter_type of 0, bit_shift is the number of left-shifts
+	 * where 0xFF would still fall within the hash mask.
+	 */
+	while (hash_mask >> bit_shift != 0xFF)
+		bit_shift++;
+
+	/* The portion of the address that is used for the hash table
+	 * is determined by the mc_filter_type setting.
+	 * The algorithm is such that there is a total of 8 bits of shifting.
+	 * The bit_shift for a mc_filter_type of 0 represents the number of
+	 * left-shifts where the MSB of mc_addr[5] would still fall within
+	 * the hash_mask.  Case 0 does this exactly.  Since there are a total
+	 * of 8 bits of shifting, then mc_addr[4] will shift right the
+	 * remaining number of bits. Thus 8 - bit_shift.  The rest of the
+	 * cases are a variation of this algorithm...essentially raising the
+	 * number of bits to shift mc_addr[5] left, while still keeping the
+	 * 8-bit shifting total.
+	 *
+	 * For example, given the following Destination MAC Address and an
+	 * MTA register count of 128 (thus a 4096-bit vector and 0xFFF mask),
+	 * we can see that the bit_shift for case 0 is 4.  These are the hash
+	 * values resulting from each mc_filter_type...
+	 * [0] [1] [2] [3] [4] [5]
+	 * 01  AA  00  12  34  56
+	 * LSB                 MSB
+	 *
+	 * case 0: hash_value = ((0x34 >> 4) | (0x56 << 4)) & 0xFFF = 0x563
+	 * case 1: hash_value = ((0x34 >> 3) | (0x56 << 5)) & 0xFFF = 0xAC6
+	 * case 2: hash_value = ((0x34 >> 2) | (0x56 << 6)) & 0xFFF = 0x163
+	 * case 3: hash_value = ((0x34 >> 0) | (0x56 << 8)) & 0xFFF = 0x634
+	 */
+	switch (hw->mac.mc_filter_type) {
+	default:
+	case 0:
+		break;
+	case 1:
+		bit_shift += 1;
+		break;
+	case 2:
+		bit_shift += 2;
+		break;
+	case 3:
+		bit_shift += 4;
+		break;
+	}
+
+	hash_value = hash_mask & (((mc_addr[4] >> (8 - bit_shift)) |
+				  (((u16)mc_addr[5]) << bit_shift)));
+
+	return hash_value;
+}
+
+/**
+ *  igc_update_mc_addr_list - Update Multicast addresses
+ *  @hw: pointer to the HW structure
+ *  @mc_addr_list: array of multicast addresses to program
+ *  @mc_addr_count: number of multicast addresses to program
+ *
+ *  Updates entire Multicast Table Array.
+ *  The caller must have a packed mc_addr_list of multicast addresses.
+ **/
+void igc_update_mc_addr_list(struct igc_hw *hw,
+			     u8 *mc_addr_list, u32 mc_addr_count)
+{
+	u32 hash_value, hash_bit, hash_reg;
+	int i;
+
+	/* clear mta_shadow */
+	memset(&hw->mac.mta_shadow, 0, sizeof(hw->mac.mta_shadow));
+
+	/* update mta_shadow from mc_addr_list */
+	for (i = 0; (u32)i < mc_addr_count; i++) {
+		hash_value = igc_hash_mc_addr(hw, mc_addr_list);
+
+		hash_reg = (hash_value >> 5) & (hw->mac.mta_reg_count - 1);
+		hash_bit = hash_value & 0x1F;
+
+		hw->mac.mta_shadow[hash_reg] |= BIT(hash_bit);
+		mc_addr_list += ETH_ALEN;
+	}
+
+	/* replace the entire MTA table */
+	for (i = hw->mac.mta_reg_count - 1; i >= 0; i--)
+		array_wr32(IGC_MTA, i, hw->mac.mta_shadow[i]);
+	wrfl();
+}

drivers/net/ethernet/intel/igc/igc_mac.h

@@ -29,6 +29,8 @@ s32 igc_get_speed_and_duplex_copper(struct igc_hw *hw, u16 *speed,
 				    u16 *duplex);
 
 bool igc_enable_mng_pass_thru(struct igc_hw *hw);
+void igc_update_mc_addr_list(struct igc_hw *hw,
+			     u8 *mc_addr_list, u32 mc_addr_count);
 
 enum igc_mng_mode {
 	igc_mng_mode_none = 0,