summaryrefslogtreecommitdiffstats
path: root/drivers/net/ixgb/ixgb_hw.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/net/ixgb/ixgb_hw.c')
-rw-r--r--drivers/net/ixgb/ixgb_hw.c1202
1 files changed, 1202 insertions, 0 deletions
diff --git a/drivers/net/ixgb/ixgb_hw.c b/drivers/net/ixgb/ixgb_hw.c
new file mode 100644
index 000000000000..69329c73095a
--- /dev/null
+++ b/drivers/net/ixgb/ixgb_hw.c
@@ -0,0 +1,1202 @@
+/*******************************************************************************
+
+
+ Copyright(c) 1999 - 2005 Intel Corporation. All rights reserved.
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the Free
+ Software Foundation; either version 2 of the License, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ more details.
+
+ You should have received a copy of the GNU General Public License along with
+ this program; if not, write to the Free Software Foundation, Inc., 59
+ Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+ The full GNU General Public License is included in this distribution in the
+ file called LICENSE.
+
+ Contact Information:
+ Linux NICS <linux.nics@intel.com>
+ Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+
+*******************************************************************************/
+
+/* ixgb_hw.c
+ * Shared functions for accessing and configuring the adapter
+ */
+
+#include "ixgb_hw.h"
+#include "ixgb_ids.h"
+
+/* Local function prototypes */
+
+static uint32_t ixgb_hash_mc_addr(struct ixgb_hw *hw, uint8_t * mc_addr);
+
+static void ixgb_mta_set(struct ixgb_hw *hw, uint32_t hash_value);
+
+static void ixgb_get_bus_info(struct ixgb_hw *hw);
+
+static boolean_t ixgb_link_reset(struct ixgb_hw *hw);
+
+static void ixgb_optics_reset(struct ixgb_hw *hw);
+
+static ixgb_phy_type ixgb_identify_phy(struct ixgb_hw *hw);
+
+uint32_t ixgb_mac_reset(struct ixgb_hw *hw);
+
+uint32_t ixgb_mac_reset(struct ixgb_hw *hw)
+{
+ uint32_t ctrl_reg;
+
+ ctrl_reg = IXGB_CTRL0_RST |
+ IXGB_CTRL0_SDP3_DIR | /* All pins are Output=1 */
+ IXGB_CTRL0_SDP2_DIR |
+ IXGB_CTRL0_SDP1_DIR |
+ IXGB_CTRL0_SDP0_DIR |
+ IXGB_CTRL0_SDP3 | /* Initial value 1101 */
+ IXGB_CTRL0_SDP2 |
+ IXGB_CTRL0_SDP0;
+
+#ifdef HP_ZX1
+ /* Workaround for 82597EX reset errata */
+ IXGB_WRITE_REG_IO(hw, CTRL0, ctrl_reg);
+#else
+ IXGB_WRITE_REG(hw, CTRL0, ctrl_reg);
+#endif
+
+ /* Delay a few ms just to allow the reset to complete */
+ msec_delay(IXGB_DELAY_AFTER_RESET);
+ ctrl_reg = IXGB_READ_REG(hw, CTRL0);
+#ifdef DBG
+ /* Make sure the self-clearing global reset bit did self clear */
+ ASSERT(!(ctrl_reg & IXGB_CTRL0_RST));
+#endif
+
+ if (hw->phy_type == ixgb_phy_type_txn17401) {
+ ixgb_optics_reset(hw);
+ }
+
+ return ctrl_reg;
+}
+
+/******************************************************************************
+ * Reset the transmit and receive units; mask and clear all interrupts.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+boolean_t
+ixgb_adapter_stop(struct ixgb_hw *hw)
+{
+ uint32_t ctrl_reg;
+ uint32_t icr_reg;
+
+ DEBUGFUNC("ixgb_adapter_stop");
+
+ /* If we are stopped or resetting exit gracefully and wait to be
+ * started again before accessing the hardware.
+ */
+ if(hw->adapter_stopped) {
+ DEBUGOUT("Exiting because the adapter is already stopped!!!\n");
+ return FALSE;
+ }
+
+ /* Set the Adapter Stopped flag so other driver functions stop
+ * touching the Hardware.
+ */
+ hw->adapter_stopped = TRUE;
+
+ /* Clear interrupt mask to stop board from generating interrupts */
+ DEBUGOUT("Masking off all interrupts\n");
+ IXGB_WRITE_REG(hw, IMC, 0xFFFFFFFF);
+
+ /* Disable the Transmit and Receive units. Then delay to allow
+ * any pending transactions to complete before we hit the MAC with
+ * the global reset.
+ */
+ IXGB_WRITE_REG(hw, RCTL, IXGB_READ_REG(hw, RCTL) & ~IXGB_RCTL_RXEN);
+ IXGB_WRITE_REG(hw, TCTL, IXGB_READ_REG(hw, TCTL) & ~IXGB_TCTL_TXEN);
+ msec_delay(IXGB_DELAY_BEFORE_RESET);
+
+ /* Issue a global reset to the MAC. This will reset the chip's
+ * transmit, receive, DMA, and link units. It will not effect
+ * the current PCI configuration. The global reset bit is self-
+ * clearing, and should clear within a microsecond.
+ */
+ DEBUGOUT("Issuing a global reset to MAC\n");
+
+ ctrl_reg = ixgb_mac_reset(hw);
+
+ /* Clear interrupt mask to stop board from generating interrupts */
+ DEBUGOUT("Masking off all interrupts\n");
+ IXGB_WRITE_REG(hw, IMC, 0xffffffff);
+
+ /* Clear any pending interrupt events. */
+ icr_reg = IXGB_READ_REG(hw, ICR);
+
+ return (ctrl_reg & IXGB_CTRL0_RST);
+}
+
+
+/******************************************************************************
+ * Identifies the vendor of the optics module on the adapter. The SR adapters
+ * support two different types of XPAK optics, so it is necessary to determine
+ * which optics are present before applying any optics-specific workarounds.
+ *
+ * hw - Struct containing variables accessed by shared code.
+ *
+ * Returns: the vendor of the XPAK optics module.
+ *****************************************************************************/
+static ixgb_xpak_vendor
+ixgb_identify_xpak_vendor(struct ixgb_hw *hw)
+{
+ uint32_t i;
+ uint16_t vendor_name[5];
+ ixgb_xpak_vendor xpak_vendor;
+
+ DEBUGFUNC("ixgb_identify_xpak_vendor");
+
+ /* Read the first few bytes of the vendor string from the XPAK NVR
+ * registers. These are standard XENPAK/XPAK registers, so all XPAK
+ * devices should implement them. */
+ for (i = 0; i < 5; i++) {
+ vendor_name[i] = ixgb_read_phy_reg(hw,
+ MDIO_PMA_PMD_XPAK_VENDOR_NAME
+ + i, IXGB_PHY_ADDRESS,
+ MDIO_PMA_PMD_DID);
+ }
+
+ /* Determine the actual vendor */
+ if (vendor_name[0] == 'I' &&
+ vendor_name[1] == 'N' &&
+ vendor_name[2] == 'T' &&
+ vendor_name[3] == 'E' && vendor_name[4] == 'L') {
+ xpak_vendor = ixgb_xpak_vendor_intel;
+ } else {
+ xpak_vendor = ixgb_xpak_vendor_infineon;
+ }
+
+ return (xpak_vendor);
+}
+
+/******************************************************************************
+ * Determine the physical layer module on the adapter.
+ *
+ * hw - Struct containing variables accessed by shared code. The device_id
+ * field must be (correctly) populated before calling this routine.
+ *
+ * Returns: the phy type of the adapter.
+ *****************************************************************************/
+static ixgb_phy_type
+ixgb_identify_phy(struct ixgb_hw *hw)
+{
+ ixgb_phy_type phy_type;
+ ixgb_xpak_vendor xpak_vendor;
+
+ DEBUGFUNC("ixgb_identify_phy");
+
+ /* Infer the transceiver/phy type from the device id */
+ switch (hw->device_id) {
+ case IXGB_DEVICE_ID_82597EX:
+ DEBUGOUT("Identified TXN17401 optics\n");
+ phy_type = ixgb_phy_type_txn17401;
+ break;
+
+ case IXGB_DEVICE_ID_82597EX_SR:
+ /* The SR adapters carry two different types of XPAK optics
+ * modules; read the vendor identifier to determine the exact
+ * type of optics. */
+ xpak_vendor = ixgb_identify_xpak_vendor(hw);
+ if (xpak_vendor == ixgb_xpak_vendor_intel) {
+ DEBUGOUT("Identified TXN17201 optics\n");
+ phy_type = ixgb_phy_type_txn17201;
+ } else {
+ DEBUGOUT("Identified G6005 optics\n");
+ phy_type = ixgb_phy_type_g6005;
+ }
+ break;
+ case IXGB_DEVICE_ID_82597EX_LR:
+ DEBUGOUT("Identified G6104 optics\n");
+ phy_type = ixgb_phy_type_g6104;
+ break;
+ default:
+ DEBUGOUT("Unknown physical layer module\n");
+ phy_type = ixgb_phy_type_unknown;
+ break;
+ }
+
+ return (phy_type);
+}
+
+/******************************************************************************
+ * Performs basic configuration of the adapter.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *
+ * Resets the controller.
+ * Reads and validates the EEPROM.
+ * Initializes the receive address registers.
+ * Initializes the multicast table.
+ * Clears all on-chip counters.
+ * Calls routine to setup flow control settings.
+ * Leaves the transmit and receive units disabled and uninitialized.
+ *
+ * Returns:
+ * TRUE if successful,
+ * FALSE if unrecoverable problems were encountered.
+ *****************************************************************************/
+boolean_t
+ixgb_init_hw(struct ixgb_hw *hw)
+{
+ uint32_t i;
+ uint32_t ctrl_reg;
+ boolean_t status;
+
+ DEBUGFUNC("ixgb_init_hw");
+
+ /* Issue a global reset to the MAC. This will reset the chip's
+ * transmit, receive, DMA, and link units. It will not effect
+ * the current PCI configuration. The global reset bit is self-
+ * clearing, and should clear within a microsecond.
+ */
+ DEBUGOUT("Issuing a global reset to MAC\n");
+
+ ctrl_reg = ixgb_mac_reset(hw);
+
+ DEBUGOUT("Issuing an EE reset to MAC\n");
+#ifdef HP_ZX1
+ /* Workaround for 82597EX reset errata */
+ IXGB_WRITE_REG_IO(hw, CTRL1, IXGB_CTRL1_EE_RST);
+#else
+ IXGB_WRITE_REG(hw, CTRL1, IXGB_CTRL1_EE_RST);
+#endif
+
+ /* Delay a few ms just to allow the reset to complete */
+ msec_delay(IXGB_DELAY_AFTER_EE_RESET);
+
+ if (ixgb_get_eeprom_data(hw) == FALSE) {
+ return(FALSE);
+ }
+
+ /* Use the device id to determine the type of phy/transceiver. */
+ hw->device_id = ixgb_get_ee_device_id(hw);
+ hw->phy_type = ixgb_identify_phy(hw);
+
+ /* Setup the receive addresses.
+ * Receive Address Registers (RARs 0 - 15).
+ */
+ ixgb_init_rx_addrs(hw);
+
+ /*
+ * Check that a valid MAC address has been set.
+ * If it is not valid, we fail hardware init.
+ */
+ if (!mac_addr_valid(hw->curr_mac_addr)) {
+ DEBUGOUT("MAC address invalid after ixgb_init_rx_addrs\n");
+ return(FALSE);
+ }
+
+ /* tell the routines in this file they can access hardware again */
+ hw->adapter_stopped = FALSE;
+
+ /* Fill in the bus_info structure */
+ ixgb_get_bus_info(hw);
+
+ /* Zero out the Multicast HASH table */
+ DEBUGOUT("Zeroing the MTA\n");
+ for(i = 0; i < IXGB_MC_TBL_SIZE; i++)
+ IXGB_WRITE_REG_ARRAY(hw, MTA, i, 0);
+
+ /* Zero out the VLAN Filter Table Array */
+ ixgb_clear_vfta(hw);
+
+ /* Zero all of the hardware counters */
+ ixgb_clear_hw_cntrs(hw);
+
+ /* Call a subroutine to setup flow control. */
+ status = ixgb_setup_fc(hw);
+
+ /* 82597EX errata: Call check-for-link in case lane deskew is locked */
+ ixgb_check_for_link(hw);
+
+ return (status);
+}
+
+/******************************************************************************
+ * Initializes receive address filters.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *
+ * Places the MAC address in receive address register 0 and clears the rest
+ * of the receive addresss registers. Clears the multicast table. Assumes
+ * the receiver is in reset when the routine is called.
+ *****************************************************************************/
+void
+ixgb_init_rx_addrs(struct ixgb_hw *hw)
+{
+ uint32_t i;
+
+ DEBUGFUNC("ixgb_init_rx_addrs");
+
+ /*
+ * If the current mac address is valid, assume it is a software override
+ * to the permanent address.
+ * Otherwise, use the permanent address from the eeprom.
+ */
+ if (!mac_addr_valid(hw->curr_mac_addr)) {
+
+ /* Get the MAC address from the eeprom for later reference */
+ ixgb_get_ee_mac_addr(hw, hw->curr_mac_addr);
+
+ DEBUGOUT3(" Keeping Permanent MAC Addr =%.2X %.2X %.2X ",
+ hw->curr_mac_addr[0],
+ hw->curr_mac_addr[1], hw->curr_mac_addr[2]);
+ DEBUGOUT3("%.2X %.2X %.2X\n",
+ hw->curr_mac_addr[3],
+ hw->curr_mac_addr[4], hw->curr_mac_addr[5]);
+ } else {
+
+ /* Setup the receive address. */
+ DEBUGOUT("Overriding MAC Address in RAR[0]\n");
+ DEBUGOUT3(" New MAC Addr =%.2X %.2X %.2X ",
+ hw->curr_mac_addr[0],
+ hw->curr_mac_addr[1], hw->curr_mac_addr[2]);
+ DEBUGOUT3("%.2X %.2X %.2X\n",
+ hw->curr_mac_addr[3],
+ hw->curr_mac_addr[4], hw->curr_mac_addr[5]);
+
+ ixgb_rar_set(hw, hw->curr_mac_addr, 0);
+ }
+
+ /* Zero out the other 15 receive addresses. */
+ DEBUGOUT("Clearing RAR[1-15]\n");
+ for(i = 1; i < IXGB_RAR_ENTRIES; i++) {
+ IXGB_WRITE_REG_ARRAY(hw, RA, (i << 1), 0);
+ IXGB_WRITE_REG_ARRAY(hw, RA, ((i << 1) + 1), 0);
+ }
+
+ return;
+}
+
+/******************************************************************************
+ * Updates the MAC's list of multicast addresses.
+ *
+ * hw - Struct containing variables accessed by shared code
+ * mc_addr_list - the list of new multicast addresses
+ * mc_addr_count - number of addresses
+ * pad - number of bytes between addresses in the list
+ *
+ * The given list replaces any existing list. Clears the last 15 receive
+ * address registers and the multicast table. Uses receive address registers
+ * for the first 15 multicast addresses, and hashes the rest into the
+ * multicast table.
+ *****************************************************************************/
+void
+ixgb_mc_addr_list_update(struct ixgb_hw *hw,
+ uint8_t *mc_addr_list,
+ uint32_t mc_addr_count,
+ uint32_t pad)
+{
+ uint32_t hash_value;
+ uint32_t i;
+ uint32_t rar_used_count = 1; /* RAR[0] is used for our MAC address */
+
+ DEBUGFUNC("ixgb_mc_addr_list_update");
+
+ /* Set the new number of MC addresses that we are being requested to use. */
+ hw->num_mc_addrs = mc_addr_count;
+
+ /* Clear RAR[1-15] */
+ DEBUGOUT(" Clearing RAR[1-15]\n");
+ for(i = rar_used_count; i < IXGB_RAR_ENTRIES; i++) {
+ IXGB_WRITE_REG_ARRAY(hw, RA, (i << 1), 0);
+ IXGB_WRITE_REG_ARRAY(hw, RA, ((i << 1) + 1), 0);
+ }
+
+ /* Clear the MTA */
+ DEBUGOUT(" Clearing MTA\n");
+ for(i = 0; i < IXGB_MC_TBL_SIZE; i++) {
+ IXGB_WRITE_REG_ARRAY(hw, MTA, i, 0);
+ }
+
+ /* Add the new addresses */
+ for(i = 0; i < mc_addr_count; i++) {
+ DEBUGOUT(" Adding the multicast addresses:\n");
+ DEBUGOUT7(" MC Addr #%d =%.2X %.2X %.2X %.2X %.2X %.2X\n", i,
+ mc_addr_list[i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad)],
+ mc_addr_list[i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad) +
+ 1],
+ mc_addr_list[i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad) +
+ 2],
+ mc_addr_list[i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad) +
+ 3],
+ mc_addr_list[i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad) +
+ 4],
+ mc_addr_list[i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad) +
+ 5]);
+
+ /* Place this multicast address in the RAR if there is room, *
+ * else put it in the MTA
+ */
+ if(rar_used_count < IXGB_RAR_ENTRIES) {
+ ixgb_rar_set(hw,
+ mc_addr_list +
+ (i * (IXGB_ETH_LENGTH_OF_ADDRESS + pad)),
+ rar_used_count);
+ DEBUGOUT1("Added a multicast address to RAR[%d]\n", i);
+ rar_used_count++;
+ } else {
+ hash_value = ixgb_hash_mc_addr(hw,
+ mc_addr_list +
+ (i *
+ (IXGB_ETH_LENGTH_OF_ADDRESS
+ + pad)));
+
+ DEBUGOUT1(" Hash value = 0x%03X\n", hash_value);
+
+ ixgb_mta_set(hw, hash_value);
+ }
+ }
+
+ DEBUGOUT("MC Update Complete\n");
+ return;
+}
+
+/******************************************************************************
+ * Hashes an address to determine its location in the multicast table
+ *
+ * hw - Struct containing variables accessed by shared code
+ * mc_addr - the multicast address to hash
+ *
+ * Returns:
+ * The hash value
+ *****************************************************************************/
+static uint32_t
+ixgb_hash_mc_addr(struct ixgb_hw *hw,
+ uint8_t *mc_addr)
+{
+ uint32_t hash_value = 0;
+
+ DEBUGFUNC("ixgb_hash_mc_addr");
+
+ /* The portion of the address that is used for the hash table is
+ * determined by the mc_filter_type setting.
+ */
+ switch (hw->mc_filter_type) {
+ /* [0] [1] [2] [3] [4] [5]
+ * 01 AA 00 12 34 56
+ * LSB MSB - According to H/W docs */
+ case 0:
+ /* [47:36] i.e. 0x563 for above example address */
+ hash_value =
+ ((mc_addr[4] >> 4) | (((uint16_t) mc_addr[5]) << 4));
+ break;
+ case 1: /* [46:35] i.e. 0xAC6 for above example address */
+ hash_value =
+ ((mc_addr[4] >> 3) | (((uint16_t) mc_addr[5]) << 5));
+ break;
+ case 2: /* [45:34] i.e. 0x5D8 for above example address */
+ hash_value =
+ ((mc_addr[4] >> 2) | (((uint16_t) mc_addr[5]) << 6));
+ break;
+ case 3: /* [43:32] i.e. 0x634 for above example address */
+ hash_value = ((mc_addr[4]) | (((uint16_t) mc_addr[5]) << 8));
+ break;
+ default:
+ /* Invalid mc_filter_type, what should we do? */
+ DEBUGOUT("MC filter type param set incorrectly\n");
+ ASSERT(0);
+ break;
+ }
+
+ hash_value &= 0xFFF;
+ return (hash_value);
+}
+
+/******************************************************************************
+ * Sets the bit in the multicast table corresponding to the hash value.
+ *
+ * hw - Struct containing variables accessed by shared code
+ * hash_value - Multicast address hash value
+ *****************************************************************************/
+static void
+ixgb_mta_set(struct ixgb_hw *hw,
+ uint32_t hash_value)
+{
+ uint32_t hash_bit, hash_reg;
+ uint32_t mta_reg;
+
+ /* The MTA is a register array of 128 32-bit registers.
+ * It is treated like an array of 4096 bits. We want to set
+ * bit BitArray[hash_value]. So we figure out what register
+ * the bit is in, read it, OR in the new bit, then write
+ * back the new value. The register is determined by the
+ * upper 7 bits of the hash value and the bit within that
+ * register are determined by the lower 5 bits of the value.
+ */
+ hash_reg = (hash_value >> 5) & 0x7F;
+ hash_bit = hash_value & 0x1F;
+
+ mta_reg = IXGB_READ_REG_ARRAY(hw, MTA, hash_reg);
+
+ mta_reg |= (1 << hash_bit);
+
+ IXGB_WRITE_REG_ARRAY(hw, MTA, hash_reg, mta_reg);
+
+ return;
+}
+
+/******************************************************************************
+ * Puts an ethernet address into a receive address register.
+ *
+ * hw - Struct containing variables accessed by shared code
+ * addr - Address to put into receive address register
+ * index - Receive address register to write
+ *****************************************************************************/
+void
+ixgb_rar_set(struct ixgb_hw *hw,
+ uint8_t *addr,
+ uint32_t index)
+{
+ uint32_t rar_low, rar_high;
+
+ DEBUGFUNC("ixgb_rar_set");
+
+ /* HW expects these in little endian so we reverse the byte order
+ * from network order (big endian) to little endian
+ */
+ rar_low = ((uint32_t) addr[0] |
+ ((uint32_t)addr[1] << 8) |
+ ((uint32_t)addr[2] << 16) |
+ ((uint32_t)addr[3] << 24));
+
+ rar_high = ((uint32_t) addr[4] |
+ ((uint32_t)addr[5] << 8) |
+ IXGB_RAH_AV);
+
+ IXGB_WRITE_REG_ARRAY(hw, RA, (index << 1), rar_low);
+ IXGB_WRITE_REG_ARRAY(hw, RA, ((index << 1) + 1), rar_high);
+ return;
+}
+
+/******************************************************************************
+ * Writes a value to the specified offset in the VLAN filter table.
+ *
+ * hw - Struct containing variables accessed by shared code
+ * offset - Offset in VLAN filer table to write
+ * value - Value to write into VLAN filter table
+ *****************************************************************************/
+void
+ixgb_write_vfta(struct ixgb_hw *hw,
+ uint32_t offset,
+ uint32_t value)
+{
+ IXGB_WRITE_REG_ARRAY(hw, VFTA, offset, value);
+ return;
+}
+
+/******************************************************************************
+ * Clears the VLAN filer table
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+void
+ixgb_clear_vfta(struct ixgb_hw *hw)
+{
+ uint32_t offset;
+
+ for(offset = 0; offset < IXGB_VLAN_FILTER_TBL_SIZE; offset++)
+ IXGB_WRITE_REG_ARRAY(hw, VFTA, offset, 0);
+ return;
+}
+
+/******************************************************************************
+ * Configures the flow control settings based on SW configuration.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+
+boolean_t
+ixgb_setup_fc(struct ixgb_hw *hw)
+{
+ uint32_t ctrl_reg;
+ uint32_t pap_reg = 0; /* by default, assume no pause time */
+ boolean_t status = TRUE;
+
+ DEBUGFUNC("ixgb_setup_fc");
+
+ /* Get the current control reg 0 settings */
+ ctrl_reg = IXGB_READ_REG(hw, CTRL0);
+
+ /* Clear the Receive Pause Enable and Transmit Pause Enable bits */
+ ctrl_reg &= ~(IXGB_CTRL0_RPE | IXGB_CTRL0_TPE);
+
+ /* The possible values of the "flow_control" parameter are:
+ * 0: Flow control is completely disabled
+ * 1: Rx flow control is enabled (we can receive pause frames
+ * but not send pause frames).
+ * 2: Tx flow control is enabled (we can send pause frames
+ * but we do not support receiving pause frames).
+ * 3: Both Rx and TX flow control (symmetric) are enabled.
+ * other: Invalid.
+ */
+ switch (hw->fc.type) {
+ case ixgb_fc_none: /* 0 */
+ /* Set CMDC bit to disable Rx Flow control */
+ ctrl_reg |= (IXGB_CTRL0_CMDC);
+ break;
+ case ixgb_fc_rx_pause: /* 1 */
+ /* RX Flow control is enabled, and TX Flow control is
+ * disabled.
+ */
+ ctrl_reg |= (IXGB_CTRL0_RPE);
+ break;
+ case ixgb_fc_tx_pause: /* 2 */
+ /* TX Flow control is enabled, and RX Flow control is
+ * disabled, by a software over-ride.
+ */
+ ctrl_reg |= (IXGB_CTRL0_TPE);
+ pap_reg = hw->fc.pause_time;
+ break;
+ case ixgb_fc_full: /* 3 */
+ /* Flow control (both RX and TX) is enabled by a software
+ * over-ride.
+ */
+ ctrl_reg |= (IXGB_CTRL0_RPE | IXGB_CTRL0_TPE);
+ pap_reg = hw->fc.pause_time;
+ break;
+ default:
+ /* We should never get here. The value should be 0-3. */
+ DEBUGOUT("Flow control param set incorrectly\n");
+ ASSERT(0);
+ break;
+ }
+
+ /* Write the new settings */
+ IXGB_WRITE_REG(hw, CTRL0, ctrl_reg);
+
+ if (pap_reg != 0) {
+ IXGB_WRITE_REG(hw, PAP, pap_reg);
+ }
+
+ /* Set the flow control receive threshold registers. Normally,
+ * these registers will be set to a default threshold that may be
+ * adjusted later by the driver's runtime code. However, if the
+ * ability to transmit pause frames in not enabled, then these
+ * registers will be set to 0.
+ */
+ if(!(hw->fc.type & ixgb_fc_tx_pause)) {
+ IXGB_WRITE_REG(hw, FCRTL, 0);
+ IXGB_WRITE_REG(hw, FCRTH, 0);
+ } else {
+ /* We need to set up the Receive Threshold high and low water
+ * marks as well as (optionally) enabling the transmission of XON
+ * frames. */
+ if(hw->fc.send_xon) {
+ IXGB_WRITE_REG(hw, FCRTL,
+ (hw->fc.low_water | IXGB_FCRTL_XONE));
+ } else {
+ IXGB_WRITE_REG(hw, FCRTL, hw->fc.low_water);
+ }
+ IXGB_WRITE_REG(hw, FCRTH, hw->fc.high_water);
+ }
+ return (status);
+}
+
+/******************************************************************************
+ * Reads a word from a device over the Management Data Interface (MDI) bus.
+ * This interface is used to manage Physical layer devices.
+ *
+ * hw - Struct containing variables accessed by hw code
+ * reg_address - Offset of device register being read.
+ * phy_address - Address of device on MDI.
+ *
+ * Returns: Data word (16 bits) from MDI device.
+ *
+ * The 82597EX has support for several MDI access methods. This routine
+ * uses the new protocol MDI Single Command and Address Operation.
+ * This requires that first an address cycle command is sent, followed by a
+ * read command.
+ *****************************************************************************/
+uint16_t
+ixgb_read_phy_reg(struct ixgb_hw *hw,
+ uint32_t reg_address,
+ uint32_t phy_address,
+ uint32_t device_type)
+{
+ uint32_t i;
+ uint32_t data;
+ uint32_t command = 0;
+
+ ASSERT(reg_address <= IXGB_MAX_PHY_REG_ADDRESS);
+ ASSERT(phy_address <= IXGB_MAX_PHY_ADDRESS);
+ ASSERT(device_type <= IXGB_MAX_PHY_DEV_TYPE);
+
+ /* Setup and write the address cycle command */
+ command = ((reg_address << IXGB_MSCA_NP_ADDR_SHIFT) |
+ (device_type << IXGB_MSCA_DEV_TYPE_SHIFT) |
+ (phy_address << IXGB_MSCA_PHY_ADDR_SHIFT) |
+ (IXGB_MSCA_ADDR_CYCLE | IXGB_MSCA_MDI_COMMAND));
+
+ IXGB_WRITE_REG(hw, MSCA, command);
+
+ /**************************************************************
+ ** Check every 10 usec to see if the address cycle completed
+ ** The COMMAND bit will clear when the operation is complete.
+ ** This may take as long as 64 usecs (we'll wait 100 usecs max)
+ ** from the CPU Write to the Ready bit assertion.
+ **************************************************************/
+
+ for(i = 0; i < 10; i++)
+ {
+ udelay(10);
+
+ command = IXGB_READ_REG(hw, MSCA);
+
+ if ((command & IXGB_MSCA_MDI_COMMAND) == 0)
+ break;
+ }
+
+ ASSERT((command & IXGB_MSCA_MDI_COMMAND) == 0);
+
+ /* Address cycle complete, setup and write the read command */
+ command = ((reg_address << IXGB_MSCA_NP_ADDR_SHIFT) |
+ (device_type << IXGB_MSCA_DEV_TYPE_SHIFT) |
+ (phy_address << IXGB_MSCA_PHY_ADDR_SHIFT) |
+ (IXGB_MSCA_READ | IXGB_MSCA_MDI_COMMAND));
+
+ IXGB_WRITE_REG(hw, MSCA, command);
+
+ /**************************************************************
+ ** Check every 10 usec to see if the read command completed
+ ** The COMMAND bit will clear when the operation is complete.
+ ** The read may take as long as 64 usecs (we'll wait 100 usecs max)
+ ** from the CPU Write to the Ready bit assertion.
+ **************************************************************/
+
+ for(i = 0; i < 10; i++)
+ {
+ udelay(10);
+
+ command = IXGB_READ_REG(hw, MSCA);
+
+ if ((command & IXGB_MSCA_MDI_COMMAND) == 0)
+ break;
+ }
+
+ ASSERT((command & IXGB_MSCA_MDI_COMMAND) == 0);
+
+ /* Operation is complete, get the data from the MDIO Read/Write Data
+ * register and return.
+ */
+ data = IXGB_READ_REG(hw, MSRWD);
+ data >>= IXGB_MSRWD_READ_DATA_SHIFT;
+ return((uint16_t) data);
+}
+
+/******************************************************************************
+ * Writes a word to a device over the Management Data Interface (MDI) bus.
+ * This interface is used to manage Physical layer devices.
+ *
+ * hw - Struct containing variables accessed by hw code
+ * reg_address - Offset of device register being read.
+ * phy_address - Address of device on MDI.
+ * device_type - Also known as the Device ID or DID.
+ * data - 16-bit value to be written
+ *
+ * Returns: void.
+ *
+ * The 82597EX has support for several MDI access methods. This routine
+ * uses the new protocol MDI Single Command and Address Operation.
+ * This requires that first an address cycle command is sent, followed by a
+ * write command.
+ *****************************************************************************/
+void
+ixgb_write_phy_reg(struct ixgb_hw *hw,
+ uint32_t reg_address,
+ uint32_t phy_address,
+ uint32_t device_type,
+ uint16_t data)
+{
+ uint32_t i;
+ uint32_t command = 0;
+
+ ASSERT(reg_address <= IXGB_MAX_PHY_REG_ADDRESS);
+ ASSERT(phy_address <= IXGB_MAX_PHY_ADDRESS);
+ ASSERT(device_type <= IXGB_MAX_PHY_DEV_TYPE);
+
+ /* Put the data in the MDIO Read/Write Data register */
+ IXGB_WRITE_REG(hw, MSRWD, (uint32_t)data);
+
+ /* Setup and write the address cycle command */
+ command = ((reg_address << IXGB_MSCA_NP_ADDR_SHIFT) |
+ (device_type << IXGB_MSCA_DEV_TYPE_SHIFT) |
+ (phy_address << IXGB_MSCA_PHY_ADDR_SHIFT) |
+ (IXGB_MSCA_ADDR_CYCLE | IXGB_MSCA_MDI_COMMAND));
+
+ IXGB_WRITE_REG(hw, MSCA, command);
+
+ /**************************************************************
+ ** Check every 10 usec to see if the address cycle completed
+ ** The COMMAND bit will clear when the operation is complete.
+ ** This may take as long as 64 usecs (we'll wait 100 usecs max)
+ ** from the CPU Write to the Ready bit assertion.
+ **************************************************************/
+
+ for(i = 0; i < 10; i++)
+ {
+ udelay(10);
+
+ command = IXGB_READ_REG(hw, MSCA);
+
+ if ((command & IXGB_MSCA_MDI_COMMAND) == 0)
+ break;
+ }
+
+ ASSERT((command & IXGB_MSCA_MDI_COMMAND) == 0);
+
+ /* Address cycle complete, setup and write the write command */
+ command = ((reg_address << IXGB_MSCA_NP_ADDR_SHIFT) |
+ (device_type << IXGB_MSCA_DEV_TYPE_SHIFT) |
+ (phy_address << IXGB_MSCA_PHY_ADDR_SHIFT) |
+ (IXGB_MSCA_WRITE | IXGB_MSCA_MDI_COMMAND));
+
+ IXGB_WRITE_REG(hw, MSCA, command);
+
+ /**************************************************************
+ ** Check every 10 usec to see if the read command completed
+ ** The COMMAND bit will clear when the operation is complete.
+ ** The write may take as long as 64 usecs (we'll wait 100 usecs max)
+ ** from the CPU Write to the Ready bit assertion.
+ **************************************************************/
+
+ for(i = 0; i < 10; i++)
+ {
+ udelay(10);
+
+ command = IXGB_READ_REG(hw, MSCA);
+
+ if ((command & IXGB_MSCA_MDI_COMMAND) == 0)
+ break;
+ }
+
+ ASSERT((command & IXGB_MSCA_MDI_COMMAND) == 0);
+
+ /* Operation is complete, return. */
+}
+
+/******************************************************************************
+ * Checks to see if the link status of the hardware has changed.
+ *
+ * hw - Struct containing variables accessed by hw code
+ *
+ * Called by any function that needs to check the link status of the adapter.
+ *****************************************************************************/
+void
+ixgb_check_for_link(struct ixgb_hw *hw)
+{
+ uint32_t status_reg;
+ uint32_t xpcss_reg;
+
+ DEBUGFUNC("ixgb_check_for_link");
+
+ xpcss_reg = IXGB_READ_REG(hw, XPCSS);
+ status_reg = IXGB_READ_REG(hw, STATUS);
+
+ if ((xpcss_reg & IXGB_XPCSS_ALIGN_STATUS) &&
+ (status_reg & IXGB_STATUS_LU)) {
+ hw->link_up = TRUE;
+ } else if (!(xpcss_reg & IXGB_XPCSS_ALIGN_STATUS) &&
+ (status_reg & IXGB_STATUS_LU)) {
+ DEBUGOUT("XPCSS Not Aligned while Status:LU is set.\n");
+ hw->link_up = ixgb_link_reset(hw);
+ } else {
+ /*
+ * 82597EX errata. Since the lane deskew problem may prevent
+ * link, reset the link before reporting link down.
+ */
+ hw->link_up = ixgb_link_reset(hw);
+ }
+ /* Anything else for 10 Gig?? */
+}
+
+/******************************************************************************
+ * Check for a bad link condition that may have occured.
+ * The indication is that the RFC / LFC registers may be incrementing
+ * continually. A full adapter reset is required to recover.
+ *
+ * hw - Struct containing variables accessed by hw code
+ *
+ * Called by any function that needs to check the link status of the adapter.
+ *****************************************************************************/
+boolean_t ixgb_check_for_bad_link(struct ixgb_hw *hw)
+{
+ uint32_t newLFC, newRFC;
+ boolean_t bad_link_returncode = FALSE;
+
+ if (hw->phy_type == ixgb_phy_type_txn17401) {
+ newLFC = IXGB_READ_REG(hw, LFC);
+ newRFC = IXGB_READ_REG(hw, RFC);
+ if ((hw->lastLFC + 250 < newLFC)
+ || (hw->lastRFC + 250 < newRFC)) {
+ DEBUGOUT
+ ("BAD LINK! too many LFC/RFC since last check\n");
+ bad_link_returncode = TRUE;
+ }
+ hw->lastLFC = newLFC;
+ hw->lastRFC = newRFC;
+ }
+
+ return bad_link_returncode;
+}
+
+/******************************************************************************
+ * Clears all hardware statistics counters.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+void
+ixgb_clear_hw_cntrs(struct ixgb_hw *hw)
+{
+ volatile uint32_t temp_reg;
+
+ DEBUGFUNC("ixgb_clear_hw_cntrs");
+
+ /* if we are stopped or resetting exit gracefully */
+ if(hw->adapter_stopped) {
+ DEBUGOUT("Exiting because the adapter is stopped!!!\n");
+ return;
+ }
+
+ temp_reg = IXGB_READ_REG(hw, TPRL);
+ temp_reg = IXGB_READ_REG(hw, TPRH);
+ temp_reg = IXGB_READ_REG(hw, GPRCL);
+ temp_reg = IXGB_READ_REG(hw, GPRCH);
+ temp_reg = IXGB_READ_REG(hw, BPRCL);
+ temp_reg = IXGB_READ_REG(hw, BPRCH);
+ temp_reg = IXGB_READ_REG(hw, MPRCL);
+ temp_reg = IXGB_READ_REG(hw, MPRCH);
+ temp_reg = IXGB_READ_REG(hw, UPRCL);
+ temp_reg = IXGB_READ_REG(hw, UPRCH);
+ temp_reg = IXGB_READ_REG(hw, VPRCL);
+ temp_reg = IXGB_READ_REG(hw, VPRCH);
+ temp_reg = IXGB_READ_REG(hw, JPRCL);
+ temp_reg = IXGB_READ_REG(hw, JPRCH);
+ temp_reg = IXGB_READ_REG(hw, GORCL);
+ temp_reg = IXGB_READ_REG(hw, GORCH);
+ temp_reg = IXGB_READ_REG(hw, TORL);
+ temp_reg = IXGB_READ_REG(hw, TORH);
+ temp_reg = IXGB_READ_REG(hw, RNBC);
+ temp_reg = IXGB_READ_REG(hw, RUC);
+ temp_reg = IXGB_READ_REG(hw, ROC);
+ temp_reg = IXGB_READ_REG(hw, RLEC);
+ temp_reg = IXGB_READ_REG(hw, CRCERRS);
+ temp_reg = IXGB_READ_REG(hw, ICBC);
+ temp_reg = IXGB_READ_REG(hw, ECBC);
+ temp_reg = IXGB_READ_REG(hw, MPC);
+ temp_reg = IXGB_READ_REG(hw, TPTL);
+ temp_reg = IXGB_READ_REG(hw, TPTH);
+ temp_reg = IXGB_READ_REG(hw, GPTCL);
+ temp_reg = IXGB_READ_REG(hw, GPTCH);
+ temp_reg = IXGB_READ_REG(hw, BPTCL);
+ temp_reg = IXGB_READ_REG(hw, BPTCH);
+ temp_reg = IXGB_READ_REG(hw, MPTCL);
+ temp_reg = IXGB_READ_REG(hw, MPTCH);
+ temp_reg = IXGB_READ_REG(hw, UPTCL);
+ temp_reg = IXGB_READ_REG(hw, UPTCH);
+ temp_reg = IXGB_READ_REG(hw, VPTCL);
+ temp_reg = IXGB_READ_REG(hw, VPTCH);
+ temp_reg = IXGB_READ_REG(hw, JPTCL);
+ temp_reg = IXGB_READ_REG(hw, JPTCH);
+ temp_reg = IXGB_READ_REG(hw, GOTCL);
+ temp_reg = IXGB_READ_REG(hw, GOTCH);
+ temp_reg = IXGB_READ_REG(hw, TOTL);
+ temp_reg = IXGB_READ_REG(hw, TOTH);
+ temp_reg = IXGB_READ_REG(hw, DC);
+ temp_reg = IXGB_READ_REG(hw, PLT64C);
+ temp_reg = IXGB_READ_REG(hw, TSCTC);
+ temp_reg = IXGB_READ_REG(hw, TSCTFC);
+ temp_reg = IXGB_READ_REG(hw, IBIC);
+ temp_reg = IXGB_READ_REG(hw, RFC);
+ temp_reg = IXGB_READ_REG(hw, LFC);
+ temp_reg = IXGB_READ_REG(hw, PFRC);
+ temp_reg = IXGB_READ_REG(hw, PFTC);
+ temp_reg = IXGB_READ_REG(hw, MCFRC);
+ temp_reg = IXGB_READ_REG(hw, MCFTC);
+ temp_reg = IXGB_READ_REG(hw, XONRXC);
+ temp_reg = IXGB_READ_REG(hw, XONTXC);
+ temp_reg = IXGB_READ_REG(hw, XOFFRXC);
+ temp_reg = IXGB_READ_REG(hw, XOFFTXC);
+ temp_reg = IXGB_READ_REG(hw, RJC);
+ return;
+}
+
+/******************************************************************************
+ * Turns on the software controllable LED
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+void
+ixgb_led_on(struct ixgb_hw *hw)
+{
+ uint32_t ctrl0_reg = IXGB_READ_REG(hw, CTRL0);
+
+ /* To turn on the LED, clear software-definable pin 0 (SDP0). */
+ ctrl0_reg &= ~IXGB_CTRL0_SDP0;
+ IXGB_WRITE_REG(hw, CTRL0, ctrl0_reg);
+ return;
+}
+
+/******************************************************************************
+ * Turns off the software controllable LED
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+void
+ixgb_led_off(struct ixgb_hw *hw)
+{
+ uint32_t ctrl0_reg = IXGB_READ_REG(hw, CTRL0);
+
+ /* To turn off the LED, set software-definable pin 0 (SDP0). */
+ ctrl0_reg |= IXGB_CTRL0_SDP0;
+ IXGB_WRITE_REG(hw, CTRL0, ctrl0_reg);
+ return;
+}
+
+/******************************************************************************
+ * Gets the current PCI bus type, speed, and width of the hardware
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+static void
+ixgb_get_bus_info(struct ixgb_hw *hw)
+{
+ uint32_t status_reg;
+
+ status_reg = IXGB_READ_REG(hw, STATUS);
+
+ hw->bus.type = (status_reg & IXGB_STATUS_PCIX_MODE) ?
+ ixgb_bus_type_pcix : ixgb_bus_type_pci;
+
+ if (hw->bus.type == ixgb_bus_type_pci) {
+ hw->bus.speed = (status_reg & IXGB_STATUS_PCI_SPD) ?
+ ixgb_bus_speed_66 : ixgb_bus_speed_33;
+ } else {
+ switch (status_reg & IXGB_STATUS_PCIX_SPD_MASK) {
+ case IXGB_STATUS_PCIX_SPD_66:
+ hw->bus.speed = ixgb_bus_speed_66;
+ break;
+ case IXGB_STATUS_PCIX_SPD_100:
+ hw->bus.speed = ixgb_bus_speed_100;
+ break;
+ case IXGB_STATUS_PCIX_SPD_133:
+ hw->bus.speed = ixgb_bus_speed_133;
+ break;
+ default:
+ hw->bus.speed = ixgb_bus_speed_reserved;
+ break;
+ }
+ }
+
+ hw->bus.width = (status_reg & IXGB_STATUS_BUS64) ?
+ ixgb_bus_width_64 : ixgb_bus_width_32;
+
+ return;
+}
+
+/******************************************************************************
+ * Tests a MAC address to ensure it is a valid Individual Address
+ *
+ * mac_addr - pointer to MAC address.
+ *
+ *****************************************************************************/
+boolean_t
+mac_addr_valid(uint8_t *mac_addr)
+{
+ boolean_t is_valid = TRUE;
+ DEBUGFUNC("mac_addr_valid");
+
+ /* Make sure it is not a multicast address */
+ if (IS_MULTICAST(mac_addr)) {
+ DEBUGOUT("MAC address is multicast\n");
+ is_valid = FALSE;
+ }
+ /* Not a broadcast address */
+ else if (IS_BROADCAST(mac_addr)) {
+ DEBUGOUT("MAC address is broadcast\n");
+ is_valid = FALSE;
+ }
+ /* Reject the zero address */
+ else if (mac_addr[0] == 0 &&
+ mac_addr[1] == 0 &&
+ mac_addr[2] == 0 &&
+ mac_addr[3] == 0 &&
+ mac_addr[4] == 0 &&
+ mac_addr[5] == 0) {
+ DEBUGOUT("MAC address is all zeros\n");
+ is_valid = FALSE;
+ }
+ return (is_valid);
+}
+
+/******************************************************************************
+ * Resets the 10GbE link. Waits the settle time and returns the state of
+ * the link.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+boolean_t
+ixgb_link_reset(struct ixgb_hw *hw)
+{
+ boolean_t link_status = FALSE;
+ uint8_t wait_retries = MAX_RESET_ITERATIONS;
+ uint8_t lrst_retries = MAX_RESET_ITERATIONS;
+
+ do {
+ /* Reset the link */
+ IXGB_WRITE_REG(hw, CTRL0,
+ IXGB_READ_REG(hw, CTRL0) | IXGB_CTRL0_LRST);
+
+ /* Wait for link-up and lane re-alignment */
+ do {
+ udelay(IXGB_DELAY_USECS_AFTER_LINK_RESET);
+ link_status =
+ ((IXGB_READ_REG(hw, STATUS) & IXGB_STATUS_LU)
+ && (IXGB_READ_REG(hw, XPCSS) &
+ IXGB_XPCSS_ALIGN_STATUS)) ? TRUE : FALSE;
+ } while (!link_status && --wait_retries);
+
+ } while (!link_status && --lrst_retries);
+
+ return link_status;
+}
+
+/******************************************************************************
+ * Resets the 10GbE optics module.
+ *
+ * hw - Struct containing variables accessed by shared code
+ *****************************************************************************/
+void
+ixgb_optics_reset(struct ixgb_hw *hw)
+{
+ if (hw->phy_type == ixgb_phy_type_txn17401) {
+ uint16_t mdio_reg;
+
+ ixgb_write_phy_reg(hw,
+ MDIO_PMA_PMD_CR1,
+ IXGB_PHY_ADDRESS,
+ MDIO_PMA_PMD_DID,
+ MDIO_PMA_PMD_CR1_RESET);
+
+ mdio_reg = ixgb_read_phy_reg( hw,
+ MDIO_PMA_PMD_CR1,
+ IXGB_PHY_ADDRESS,
+ MDIO_PMA_PMD_DID);
+ }
+
+ return;
+}