From c0a34ebd43959ca83df86ba2ffc63ab0ddb2ea75 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Thu, 9 Apr 2015 21:02:02 -0700 Subject: [PATCH 1/9] igb: doc don't refer to ifconfig ifconfig command is obsolete, best to remove all references so that new users learn ip. Signed-off-by: Stephen Hemminger Signed-off-by: Jeff Kirsher --- Documentation/networking/igb.txt | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Documentation/networking/igb.txt b/Documentation/networking/igb.txt index 43d3549366a09..15534fdd09a88 100644 --- a/Documentation/networking/igb.txt +++ b/Documentation/networking/igb.txt @@ -42,10 +42,10 @@ Additional Configurations Jumbo Frames ------------ Jumbo Frames support is enabled by changing the MTU to a value larger than - the default of 1500. Use the ifconfig command to increase the MTU size. + the default of 1500. Use the ip command to increase the MTU size. For example: - ifconfig eth mtu 9000 up + ip link set dev eth mtu 9000 This setting is not saved across reboots. From d7018be0ae2aee988331d2908234d1a1ee3196ac Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Thu, 9 Apr 2015 22:03:21 -0700 Subject: [PATCH 2/9] ixgbe: fix documentation The MTU values in the documentation do not match the source. The source has frame limit of IXGBE_MAX_JUMBO_FRAME_SIZE (9728) which is MTU of 9710 because of the accounting for Ethernet header and CRC. Also, don't refer to the obsolete ifconfig command. Signed-off-by: Stephen Hemminger Signed-off-by: Jeff Kirsher --- Documentation/networking/ixgbe.txt | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/Documentation/networking/ixgbe.txt b/Documentation/networking/ixgbe.txt index 0ace6e776ac86..6f0cb57b59c6b 100644 --- a/Documentation/networking/ixgbe.txt +++ b/Documentation/networking/ixgbe.txt @@ -70,10 +70,10 @@ Avago 1000BASE-T SFP ABCU-5710RZ 82599-based adapters support all passive and active limiting direct attach cables that comply with SFF-8431 v4.1 and SFF-8472 v10.4 specifications. -Laser turns off for SFP+ when ifconfig down +Laser turns off for SFP+ when device is down ------------------------------------------- -"ifconfig down" turns off the laser for 82599-based SFP+ fiber adapters. -"ifconfig up" turns on the laser. +"ip link set down" turns off the laser for 82599-based SFP+ fiber adapters. +"ip link set up" turns on the laser. 82598-BASED ADAPTERS @@ -213,13 +213,13 @@ Additional Configurations ------------ The driver supports Jumbo Frames for all adapters. Jumbo Frames support is enabled by changing the MTU to a value larger than the default of 1500. - The maximum value for the MTU is 16110. Use the ifconfig command to + The maximum value for the MTU is 16110. Use the ip command to increase the MTU size. For example: - ifconfig ethx mtu 9000 up + ip link set dev ethx mtu 9000 - The maximum MTU setting for Jumbo Frames is 16110. This value coincides - with the maximum Jumbo Frames size of 16128. + The maximum MTU setting for Jumbo Frames is 9710. This value coincides + with the maximum Jumbo Frames size of 9728. Generic Receive Offload, aka GRO -------------------------------- From 322d3ed1679047e4ed0c9997d84acf3cdd4f1d29 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Thu, 9 Apr 2015 22:03:21 -0700 Subject: [PATCH 3/9] ixgb: remove references to ifconfig Move documentation into this century, even if this device hasn't been available for some time. Signed-off-by: Stephen Hemminger Signed-off-by: Jeff Kirsher --- Documentation/networking/ixgb.txt | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/Documentation/networking/ixgb.txt b/Documentation/networking/ixgb.txt index 1e0c045e89f7d..9b4a10a1cf50b 100644 --- a/Documentation/networking/ixgb.txt +++ b/Documentation/networking/ixgb.txt @@ -39,7 +39,7 @@ Channel Bonding documentation can be found in the Linux kernel source: The driver information previously displayed in the /proc filesystem is not supported in this release. Alternatively, you can use ethtool (version 1.6 -or later), lspci, and ifconfig to obtain the same information. +or later), lspci, and iproute2 to obtain the same information. Instructions on updating ethtool can be found in the section "Additional Configurations" later in this document. @@ -90,7 +90,7 @@ select m for "Intel(R) PRO/10GbE support" located at: 3. Assign an IP address to the interface by entering the following, where x is the interface number: - ifconfig ethx + ip addr add ethx 4. Verify that the interface works. Enter the following, where is the IP address for another machine on the same subnet as the interface @@ -177,7 +177,7 @@ NOTE: These changes are only suggestions, and serve as a starting point for tuning your network performance. The changes are made in three major ways, listed in order of greatest effect: -- Use ifconfig to modify the mtu (maximum transmission unit) and the txqueuelen +- Use ip link to modify the mtu (maximum transmission unit) and the txqueuelen parameter. - Use sysctl to modify /proc parameters (essentially kernel tuning) - Use setpci to modify the MMRBC field in PCI-X configuration space to increase @@ -202,7 +202,7 @@ setpci -d 8086:1a48 e6.b=2e # to change as well. # set the txqueuelen # your ixgb adapter should be loaded as eth1 for this to work, change if needed -ifconfig eth1 mtu 9000 txqueuelen 1000 up +ip li set dev eth1 mtu 9000 txqueuelen 1000 up # call the sysctl utility to modify /proc/sys entries sysctl -p ./sysctl_ixgb.conf - END ixgb_perf.sh @@ -297,10 +297,10 @@ Additional Configurations ------------ The driver supports Jumbo Frames for all adapters. Jumbo Frames support is enabled by changing the MTU to a value larger than the default of 1500. - The maximum value for the MTU is 16114. Use the ifconfig command to + The maximum value for the MTU is 16114. Use the ip command to increase the MTU size. For example: - ifconfig ethx mtu 9000 up + ip li set dev ethx mtu 9000 The maximum MTU setting for Jumbo Frames is 16114. This value coincides with the maximum Jumbo Frames size of 16128. From aa2bacb60963870919ec5bc59ac69f0c31572d26 Mon Sep 17 00:00:00 2001 From: Don Skidmore Date: Thu, 9 Apr 2015 22:03:22 -0700 Subject: [PATCH 4/9] ixgbe: Move bridge mode from flag to variable We are currently storing our BRIDGE_MODE as a bit in our adapter flags. This patch will store the actual mode instead which minimizes obfuscation and makes following patches for X550 simpler. Signed-off-by: Don Skidmore Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe.h | 3 +- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 33 +++++++++---------- .../net/ethernet/intel/ixgbe/ixgbe_sriov.c | 3 +- 3 files changed, 20 insertions(+), 19 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h index 7068e9c3691dd..ae5b8b22e7e00 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h @@ -642,7 +642,6 @@ struct ixgbe_adapter { #define IXGBE_FLAG2_RSS_FIELD_IPV4_UDP (u32)(1 << 8) #define IXGBE_FLAG2_RSS_FIELD_IPV6_UDP (u32)(1 << 9) #define IXGBE_FLAG2_PTP_PPS_ENABLED (u32)(1 << 10) -#define IXGBE_FLAG2_BRIDGE_MODE_VEB (u32)(1 << 11) /* Tx fast path data */ int num_tx_queues; @@ -722,6 +721,8 @@ struct ixgbe_adapter { u8 __iomem *io_addr; /* Mainly for iounmap use */ u32 wol; + u16 bridge_mode; + u16 eeprom_verh; u16 eeprom_verl; u16 eeprom_cap; diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index 395dc6bb5d821..aceb13c7d42c0 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -3553,7 +3553,7 @@ static void ixgbe_configure_virtualization(struct ixgbe_adapter *adapter) IXGBE_WRITE_REG(hw, IXGBE_VFRE(reg_offset ^ 1), reg_offset - 1); IXGBE_WRITE_REG(hw, IXGBE_VFTE(reg_offset), (~0) << vf_shift); IXGBE_WRITE_REG(hw, IXGBE_VFTE(reg_offset ^ 1), reg_offset - 1); - if (adapter->flags2 & IXGBE_FLAG2_BRIDGE_MODE_VEB) + if (adapter->bridge_mode == BRIDGE_MODE_VEB) IXGBE_WRITE_REG(hw, IXGBE_PFDTXGSWC, IXGBE_PFDTXGSWC_VT_LBEN); /* Map PF MAC address in RAR Entry 0 to first pool following VFs */ @@ -7886,7 +7886,6 @@ static int ixgbe_ndo_bridge_setlink(struct net_device *dev, nla_for_each_nested(attr, br_spec, rem) { __u16 mode; - u32 reg = 0; if (nla_type(attr) != IFLA_BRIDGE_MODE) continue; @@ -7895,19 +7894,24 @@ static int ixgbe_ndo_bridge_setlink(struct net_device *dev, return -EINVAL; mode = nla_get_u16(attr); - if (mode == BRIDGE_MODE_VEPA) { - reg = 0; - adapter->flags2 &= ~IXGBE_FLAG2_BRIDGE_MODE_VEB; - } else if (mode == BRIDGE_MODE_VEB) { - reg = IXGBE_PFDTXGSWC_VT_LBEN; - adapter->flags2 |= IXGBE_FLAG2_BRIDGE_MODE_VEB; - } else + switch (mode) { + case BRIDGE_MODE_VEPA: + IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, 0); + break; + case BRIDGE_MODE_VEB: + IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, + IXGBE_PFDTXGSWC_VT_LBEN); + break; + default: return -EINVAL; + } - IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, reg); + adapter->bridge_mode = mode; e_info(drv, "enabling bridge mode: %s\n", mode == BRIDGE_MODE_VEPA ? "VEPA" : "VEB"); + + break; } return 0; @@ -7918,17 +7922,12 @@ static int ixgbe_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, u32 filter_mask) { struct ixgbe_adapter *adapter = netdev_priv(dev); - u16 mode; if (!(adapter->flags & IXGBE_FLAG_SRIOV_ENABLED)) return 0; - if (adapter->flags2 & IXGBE_FLAG2_BRIDGE_MODE_VEB) - mode = BRIDGE_MODE_VEB; - else - mode = BRIDGE_MODE_VEPA; - - return ndo_dflt_bridge_getlink(skb, pid, seq, dev, mode, 0, 0); + return ndo_dflt_bridge_getlink(skb, pid, seq, dev, + adapter->bridge_mode, 0, 0); } static void *ixgbe_fwd_add(struct net_device *pdev, struct net_device *vdev) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c index 09a291bb7c343..2d98ecdbd3d67 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c @@ -36,6 +36,7 @@ #include #include #include +#include #ifdef NETIF_F_HW_VLAN_CTAG_TX #include #endif @@ -79,7 +80,7 @@ static int __ixgbe_enable_sriov(struct ixgbe_adapter *adapter) /* Initialize default switching mode VEB */ IXGBE_WRITE_REG(hw, IXGBE_PFDTXGSWC, IXGBE_PFDTXGSWC_VT_LBEN); - adapter->flags2 |= IXGBE_FLAG2_BRIDGE_MODE_VEB; + adapter->bridge_mode = BRIDGE_MODE_VEB; /* If call to enable VFs succeeded then allocate memory * for per VF control structures. From 219efe97bb8a6989dfc93ae6bf97cd72e5b72de3 Mon Sep 17 00:00:00 2001 From: Don Skidmore Date: Thu, 9 Apr 2015 22:03:22 -0700 Subject: [PATCH 5/9] ixgbe: add new bridge mode support function. This patch gathers together all the logic needed to configure bridge modes. Currently that it is rather simple but this is really laying the ground work for future X550 feature enhancement. Signed-off-by: Don Skidmore Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 50 +++++++++++++------ 1 file changed, 34 insertions(+), 16 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index aceb13c7d42c0..cb9c43f28e5ef 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -7870,6 +7870,36 @@ static int ixgbe_ndo_fdb_add(struct ndmsg *ndm, struct nlattr *tb[], return ndo_dflt_fdb_add(ndm, tb, dev, addr, vid, flags); } +/** + * ixgbe_configure_bridge_mode - set various bridge modes + * @adapter - the private structure + * @mode - requested bridge mode + * + * Configure some settings require for various bridge modes. + **/ +static int ixgbe_configure_bridge_mode(struct ixgbe_adapter *adapter, + __u16 mode) +{ + switch (mode) { + case BRIDGE_MODE_VEPA: + IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, 0); + break; + case BRIDGE_MODE_VEB: + IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, + IXGBE_PFDTXGSWC_VT_LBEN); + break; + default: + return -EINVAL; + } + + adapter->bridge_mode = mode; + + e_info(drv, "enabling bridge mode: %s\n", + mode == BRIDGE_MODE_VEPA ? "VEPA" : "VEB"); + + return 0; +} + static int ixgbe_ndo_bridge_setlink(struct net_device *dev, struct nlmsghdr *nlh, u16 flags) { @@ -7885,6 +7915,7 @@ static int ixgbe_ndo_bridge_setlink(struct net_device *dev, return -EINVAL; nla_for_each_nested(attr, br_spec, rem) { + u32 status; __u16 mode; if (nla_type(attr) != IFLA_BRIDGE_MODE) @@ -7894,22 +7925,9 @@ static int ixgbe_ndo_bridge_setlink(struct net_device *dev, return -EINVAL; mode = nla_get_u16(attr); - switch (mode) { - case BRIDGE_MODE_VEPA: - IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, 0); - break; - case BRIDGE_MODE_VEB: - IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, - IXGBE_PFDTXGSWC_VT_LBEN); - break; - default: - return -EINVAL; - } - - adapter->bridge_mode = mode; - - e_info(drv, "enabling bridge mode: %s\n", - mode == BRIDGE_MODE_VEPA ? "VEPA" : "VEB"); + status = ixgbe_configure_bridge_mode(adapter, mode); + if (status) + return status; break; } From 6d4c96ad4afd7265d2d214e02d28e66cfdaf9bb4 Mon Sep 17 00:00:00 2001 From: Don Skidmore Date: Thu, 9 Apr 2015 22:03:23 -0700 Subject: [PATCH 6/9] ixgbe: add support for X550 source_address_prunning This patch will enable X550 Source Address Prunning for VEPA bridge mode. This requires that we also have replication enabled as well, while in this mode. Signed-off-by: Don Skidmore Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 44 +++++++++++++++++++ drivers/net/ethernet/intel/ixgbe/ixgbe_type.h | 4 ++ drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c | 29 ++++++++++++ 3 files changed, 77 insertions(+) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index cb9c43f28e5ef..c696ad721153a 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -7880,13 +7880,57 @@ static int ixgbe_ndo_fdb_add(struct ndmsg *ndm, struct nlattr *tb[], static int ixgbe_configure_bridge_mode(struct ixgbe_adapter *adapter, __u16 mode) { + struct ixgbe_hw *hw = &adapter->hw; + unsigned int p, num_pools; + u32 vmdctl; + switch (mode) { case BRIDGE_MODE_VEPA: + /* disable Tx loopback, rely on switch hairpin mode */ IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, 0); + + /* must enable Rx switching replication to allow multicast + * packet reception on all VFs, and to enable source address + * pruning. + */ + vmdctl = IXGBE_READ_REG(hw, IXGBE_VMD_CTL); + vmdctl |= IXGBE_VT_CTL_REPLEN; + IXGBE_WRITE_REG(hw, IXGBE_VMD_CTL, vmdctl); + + /* enable Rx source address pruning. Note, this requires + * replication to be enabled or else it does nothing. + */ + num_pools = adapter->num_vfs + adapter->num_rx_pools; + for (p = 0; p < num_pools; p++) { + if (hw->mac.ops.set_source_address_pruning) + hw->mac.ops.set_source_address_pruning(hw, + true, + p); + } break; case BRIDGE_MODE_VEB: + /* enable Tx loopback for internal VF/PF communication */ IXGBE_WRITE_REG(&adapter->hw, IXGBE_PFDTXGSWC, IXGBE_PFDTXGSWC_VT_LBEN); + + /* disable Rx switching replication unless we have SR-IOV + * virtual functions + */ + vmdctl = IXGBE_READ_REG(hw, IXGBE_VMD_CTL); + if (!adapter->num_vfs) + vmdctl &= ~IXGBE_VT_CTL_REPLEN; + IXGBE_WRITE_REG(hw, IXGBE_VMD_CTL, vmdctl); + + /* disable Rx source address pruning, since we don't expect to + * be receiving external loopback of our transmitted frames. + */ + num_pools = adapter->num_vfs + adapter->num_rx_pools; + for (p = 0; p < num_pools; p++) { + if (hw->mac.ops.set_source_address_pruning) + hw->mac.ops.set_source_address_pruning(hw, + false, + p); + } break; default: return -EINVAL; diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h index c3ddc944f1e95..8e393098638a1 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h @@ -285,6 +285,8 @@ struct ixgbe_thermal_sensor_data { #define IXGBE_VLVF(_i) (0x0F100 + ((_i) * 4)) /* 64 of these (0-63) */ #define IXGBE_VLVFB(_i) (0x0F200 + ((_i) * 4)) /* 128 of these (0-127) */ #define IXGBE_VMVIR(_i) (0x08000 + ((_i) * 4)) /* 64 of these (0-63) */ +#define IXGBE_PFFLPL 0x050B0 +#define IXGBE_PFFLPH 0x050B4 #define IXGBE_VT_CTL 0x051B0 #define IXGBE_PFMAILBOX(_i) (0x04B00 + (4 * (_i))) /* 64 total */ #define IXGBE_PFMBMEM(_i) (0x13000 + (64 * (_i))) /* 64 Mailboxes, 16 DW each */ @@ -3069,6 +3071,8 @@ struct ixgbe_mac_operations { s32 (*init_thermal_sensor_thresh)(struct ixgbe_hw *hw); void (*disable_rx)(struct ixgbe_hw *hw); void (*enable_rx)(struct ixgbe_hw *hw); + void (*set_source_address_pruning)(struct ixgbe_hw *, bool, + unsigned int); void (*set_ethertype_anti_spoofing)(struct ixgbe_hw *, bool, int); /* DMA Coalescing */ diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c index 58a3155af7cd0..cf5cf819a6b89 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c @@ -1363,6 +1363,33 @@ static void ixgbe_set_ethertype_anti_spoofing_X550(struct ixgbe_hw *hw, IXGBE_WRITE_REG(hw, IXGBE_PFVFSPOOF(vf_target_reg), pfvfspoof); } +/** ixgbe_set_source_address_pruning_X550 - Enable/Disbale src address pruning + * @hw: pointer to hardware structure + * @enable: enable or disable source address pruning + * @pool: Rx pool to set source address pruning for + **/ +static void ixgbe_set_source_address_pruning_X550(struct ixgbe_hw *hw, + bool enable, + unsigned int pool) +{ + u64 pfflp; + + /* max rx pool is 63 */ + if (pool > 63) + return; + + pfflp = (u64)IXGBE_READ_REG(hw, IXGBE_PFFLPL); + pfflp |= (u64)IXGBE_READ_REG(hw, IXGBE_PFFLPH) << 32; + + if (enable) + pfflp |= (1ULL << pool); + else + pfflp &= ~(1ULL << pool); + + IXGBE_WRITE_REG(hw, IXGBE_PFFLPL, (u32)pfflp); + IXGBE_WRITE_REG(hw, IXGBE_PFFLPH, (u32)(pfflp >> 32)); +} + #define X550_COMMON_MAC \ .init_hw = &ixgbe_init_hw_generic, \ .start_hw = &ixgbe_start_hw_X540, \ @@ -1397,6 +1424,8 @@ static void ixgbe_set_ethertype_anti_spoofing_X550(struct ixgbe_hw *hw, .init_uta_tables = &ixgbe_init_uta_tables_generic, \ .set_mac_anti_spoofing = &ixgbe_set_mac_anti_spoofing, \ .set_vlan_anti_spoofing = &ixgbe_set_vlan_anti_spoofing, \ + .set_source_address_pruning = \ + &ixgbe_set_source_address_pruning_X550, \ .set_ethertype_anti_spoofing = \ &ixgbe_set_ethertype_anti_spoofing_X550, \ .acquire_swfw_sync = &ixgbe_acquire_swfw_sync_X540, \ From ea412015a2e72faf066e90765bdd1dc4e4ecea41 Mon Sep 17 00:00:00 2001 From: Vasu Dev Date: Thu, 9 Apr 2015 22:03:23 -0700 Subject: [PATCH 7/9] ixgbe: adds x550 specific FCoE offloads Adds x550 specific FCoE offloads for DDP context programming and increased DDP exchanges. Signed-off-by: Vasu Dev Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.c | 122 ++++++++++++++---- drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.h | 3 +- drivers/net/ethernet/intel/ixgbe/ixgbe_type.h | 9 ++ 3 files changed, 107 insertions(+), 27 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.c index 2ad91cb04dab9..631c603fc9664 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.c @@ -71,6 +71,7 @@ int ixgbe_fcoe_ddp_put(struct net_device *netdev, u16 xid) struct ixgbe_fcoe *fcoe; struct ixgbe_adapter *adapter; struct ixgbe_fcoe_ddp *ddp; + struct ixgbe_hw *hw; u32 fcbuff; if (!netdev) @@ -85,25 +86,51 @@ int ixgbe_fcoe_ddp_put(struct net_device *netdev, u16 xid) if (!ddp->udl) return 0; + hw = &adapter->hw; len = ddp->len; - /* if there an error, force to invalidate ddp context */ - if (ddp->err) { + /* if no error then skip ddp context invalidation */ + if (!ddp->err) + goto skip_ddpinv; + + if (hw->mac.type == ixgbe_mac_X550) { + /* X550 does not require DDP FCoE lock */ + + IXGBE_WRITE_REG(hw, IXGBE_FCDFC(0, xid), 0); + IXGBE_WRITE_REG(hw, IXGBE_FCDFC(3, xid), + (xid | IXGBE_FCFLTRW_WE)); + + /* program FCBUFF */ + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(2, xid), 0); + + /* program FCDMARW */ + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(3, xid), + (xid | IXGBE_FCDMARW_WE)); + + /* read FCBUFF to check context invalidated */ + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(3, xid), + (xid | IXGBE_FCDMARW_RE)); + fcbuff = IXGBE_READ_REG(hw, IXGBE_FCDDC(2, xid)); + } else { + /* other hardware requires DDP FCoE lock */ spin_lock_bh(&fcoe->lock); - IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCFLT, 0); - IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCFLTRW, + IXGBE_WRITE_REG(hw, IXGBE_FCFLT, 0); + IXGBE_WRITE_REG(hw, IXGBE_FCFLTRW, (xid | IXGBE_FCFLTRW_WE)); - IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCBUFF, 0); - IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCDMARW, + IXGBE_WRITE_REG(hw, IXGBE_FCBUFF, 0); + IXGBE_WRITE_REG(hw, IXGBE_FCDMARW, (xid | IXGBE_FCDMARW_WE)); /* guaranteed to be invalidated after 100us */ - IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCDMARW, + IXGBE_WRITE_REG(hw, IXGBE_FCDMARW, (xid | IXGBE_FCDMARW_RE)); - fcbuff = IXGBE_READ_REG(&adapter->hw, IXGBE_FCBUFF); + fcbuff = IXGBE_READ_REG(hw, IXGBE_FCBUFF); spin_unlock_bh(&fcoe->lock); - if (fcbuff & IXGBE_FCBUFF_VALID) - udelay(100); - } + } + + if (fcbuff & IXGBE_FCBUFF_VALID) + usleep_range(100, 150); + +skip_ddpinv: if (ddp->sgl) dma_unmap_sg(&adapter->pdev->dev, ddp->sgl, ddp->sgc, DMA_FROM_DEVICE); @@ -272,7 +299,6 @@ static int ixgbe_fcoe_ddp_setup(struct net_device *netdev, u16 xid, /* program DMA context */ hw = &adapter->hw; - spin_lock_bh(&fcoe->lock); /* turn on last frame indication for target mode as FCP_RSPtarget is * supposed to send FCP_RSP when it is done. */ @@ -283,16 +309,33 @@ static int ixgbe_fcoe_ddp_setup(struct net_device *netdev, u16 xid, IXGBE_WRITE_REG(hw, IXGBE_FCRXCTRL, fcrxctl); } - IXGBE_WRITE_REG(hw, IXGBE_FCPTRL, ddp->udp & DMA_BIT_MASK(32)); - IXGBE_WRITE_REG(hw, IXGBE_FCPTRH, (u64)ddp->udp >> 32); - IXGBE_WRITE_REG(hw, IXGBE_FCBUFF, fcbuff); - IXGBE_WRITE_REG(hw, IXGBE_FCDMARW, fcdmarw); - /* program filter context */ - IXGBE_WRITE_REG(hw, IXGBE_FCPARAM, 0); - IXGBE_WRITE_REG(hw, IXGBE_FCFLT, IXGBE_FCFLT_VALID); - IXGBE_WRITE_REG(hw, IXGBE_FCFLTRW, fcfltrw); + if (hw->mac.type == ixgbe_mac_X550) { + /* X550 does not require DDP lock */ + + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(0, xid), + ddp->udp & DMA_BIT_MASK(32)); + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(1, xid), (u64)ddp->udp >> 32); + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(2, xid), fcbuff); + IXGBE_WRITE_REG(hw, IXGBE_FCDDC(3, xid), fcdmarw); + /* program filter context */ + IXGBE_WRITE_REG(hw, IXGBE_FCDFC(0, xid), IXGBE_FCFLT_VALID); + IXGBE_WRITE_REG(hw, IXGBE_FCDFC(1, xid), 0); + IXGBE_WRITE_REG(hw, IXGBE_FCDFC(3, xid), fcfltrw); + } else { + /* DDP lock for indirect DDP context access */ + spin_lock_bh(&fcoe->lock); + + IXGBE_WRITE_REG(hw, IXGBE_FCPTRL, ddp->udp & DMA_BIT_MASK(32)); + IXGBE_WRITE_REG(hw, IXGBE_FCPTRH, (u64)ddp->udp >> 32); + IXGBE_WRITE_REG(hw, IXGBE_FCBUFF, fcbuff); + IXGBE_WRITE_REG(hw, IXGBE_FCDMARW, fcdmarw); + /* program filter context */ + IXGBE_WRITE_REG(hw, IXGBE_FCPARAM, 0); + IXGBE_WRITE_REG(hw, IXGBE_FCFLT, IXGBE_FCFLT_VALID); + IXGBE_WRITE_REG(hw, IXGBE_FCFLTRW, fcfltrw); - spin_unlock_bh(&fcoe->lock); + spin_unlock_bh(&fcoe->lock); + } return 1; @@ -371,6 +414,7 @@ int ixgbe_fcoe_ddp(struct ixgbe_adapter *adapter, struct fcoe_crc_eof *crc; __le32 fcerr = ixgbe_test_staterr(rx_desc, IXGBE_RXDADV_ERR_FCERR); __le32 ddp_err; + int ddp_max; u32 fctl; u16 xid; @@ -392,7 +436,11 @@ int ixgbe_fcoe_ddp(struct ixgbe_adapter *adapter, else xid = be16_to_cpu(fh->fh_rx_id); - if (xid >= IXGBE_FCOE_DDP_MAX) + ddp_max = IXGBE_FCOE_DDP_MAX; + /* X550 has different DDP Max limit */ + if (adapter->hw.mac.type == ixgbe_mac_X550) + ddp_max = IXGBE_FCOE_DDP_MAX_X550; + if (xid >= ddp_max) return -EINVAL; fcoe = &adapter->fcoe; @@ -612,7 +660,8 @@ void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter) { struct ixgbe_ring_feature *fcoe = &adapter->ring_feature[RING_F_FCOE]; struct ixgbe_hw *hw = &adapter->hw; - int i, fcoe_q, fcoe_i; + int i, fcoe_q, fcoe_i, fcoe_q_h = 0; + int fcreta_size; u32 etqf; /* Minimal functionality for FCoE requires at least CRC offloads */ @@ -633,10 +682,23 @@ void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter) return; /* Use one or more Rx queues for FCoE by redirection table */ - for (i = 0; i < IXGBE_FCRETA_SIZE; i++) { + fcreta_size = IXGBE_FCRETA_SIZE; + if (adapter->hw.mac.type == ixgbe_mac_X550) + fcreta_size = IXGBE_FCRETA_SIZE_X550; + + for (i = 0; i < fcreta_size; i++) { + if (adapter->hw.mac.type == ixgbe_mac_X550) { + int fcoe_i_h = fcoe->offset + ((i + fcreta_size) % + fcoe->indices); + fcoe_q_h = adapter->rx_ring[fcoe_i_h]->reg_idx; + fcoe_q_h = (fcoe_q_h << IXGBE_FCRETA_ENTRY_HIGH_SHIFT) & + IXGBE_FCRETA_ENTRY_HIGH_MASK; + } + fcoe_i = fcoe->offset + (i % fcoe->indices); fcoe_i &= IXGBE_FCRETA_ENTRY_MASK; fcoe_q = adapter->rx_ring[fcoe_i]->reg_idx; + fcoe_q |= fcoe_q_h; IXGBE_WRITE_REG(hw, IXGBE_FCRETA(i), fcoe_q); } IXGBE_WRITE_REG(hw, IXGBE_FCRECTL, IXGBE_FCRECTL_ENA); @@ -672,13 +734,18 @@ void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter) void ixgbe_free_fcoe_ddp_resources(struct ixgbe_adapter *adapter) { struct ixgbe_fcoe *fcoe = &adapter->fcoe; - int cpu, i; + int cpu, i, ddp_max; /* do nothing if no DDP pools were allocated */ if (!fcoe->ddp_pool) return; - for (i = 0; i < IXGBE_FCOE_DDP_MAX; i++) + ddp_max = IXGBE_FCOE_DDP_MAX; + /* X550 has different DDP Max limit */ + if (adapter->hw.mac.type == ixgbe_mac_X550) + ddp_max = IXGBE_FCOE_DDP_MAX_X550; + + for (i = 0; i < ddp_max; i++) ixgbe_fcoe_ddp_put(adapter->netdev, i); for_each_possible_cpu(cpu) @@ -758,6 +825,9 @@ static int ixgbe_fcoe_ddp_enable(struct ixgbe_adapter *adapter) } adapter->netdev->fcoe_ddp_xid = IXGBE_FCOE_DDP_MAX - 1; + /* X550 has different DDP Max limit */ + if (adapter->hw.mac.type == ixgbe_mac_X550) + adapter->netdev->fcoe_ddp_xid = IXGBE_FCOE_DDP_MAX_X550 - 1; return 0; } diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.h index 0772b7730fce9..38385876effb0 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.h +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.h @@ -46,6 +46,7 @@ #define IXGBE_FCBUFF_MAX 65536 /* 64KB max */ #define IXGBE_FCBUFF_MIN 4096 /* 4KB min */ #define IXGBE_FCOE_DDP_MAX 512 /* 9 bits xid */ +#define IXGBE_FCOE_DDP_MAX_X550 2048 /* 11 bits xid */ /* Default traffic class to use for FCoE */ #define IXGBE_FCOE_DEFTC 3 @@ -77,7 +78,7 @@ struct ixgbe_fcoe { struct ixgbe_fcoe_ddp_pool __percpu *ddp_pool; atomic_t refcnt; spinlock_t lock; - struct ixgbe_fcoe_ddp ddp[IXGBE_FCOE_DDP_MAX]; + struct ixgbe_fcoe_ddp ddp[IXGBE_FCOE_DDP_MAX_X550]; void *extra_ddp_buffer; dma_addr_t extra_ddp_buffer_dma; unsigned long mode; diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h index 8e393098638a1..dd6ba5916dfe0 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h @@ -610,6 +610,8 @@ struct ixgbe_thermal_sensor_data { #define IXGBE_RTTBCNRM 0x04980 #define IXGBE_RTTQCNRM 0x04980 +/* FCoE Direct DMA Context */ +#define IXGBE_FCDDC(_i, _j) (0x20000 + ((_i) * 0x4) + ((_j) * 0x10)) /* FCoE DMA Context Registers */ #define IXGBE_FCPTRL 0x02410 /* FC User Desc. PTR Low */ #define IXGBE_FCPTRH 0x02414 /* FC USer Desc. PTR High */ @@ -636,6 +638,9 @@ struct ixgbe_thermal_sensor_data { #define IXGBE_TSOFF 0x04A98 /* Tx FC SOF */ #define IXGBE_REOFF 0x05158 /* Rx FC EOF */ #define IXGBE_RSOFF 0x051F8 /* Rx FC SOF */ +/* FCoE Direct Filter Context */ +#define IXGBE_FCDFC(_i, _j) (0x28000 + ((_i) * 0x4) + ((_j) * 0x10)) +#define IXGBE_FCDFCD(_i) (0x30000 + ((_i) * 0x4)) /* FCoE Filter Context Registers */ #define IXGBE_FCFLT 0x05108 /* FC FLT Context */ #define IXGBE_FCFLTRW 0x05110 /* FC Filter RW Control */ @@ -666,6 +671,10 @@ struct ixgbe_thermal_sensor_data { #define IXGBE_FCRECTL_ENA 0x1 /* FCoE Redir Table Enable */ #define IXGBE_FCRETA_SIZE 8 /* Max entries in FCRETA */ #define IXGBE_FCRETA_ENTRY_MASK 0x0000007f /* 7 bits for the queue index */ +#define IXGBE_FCRETA_SIZE_X550 32 /* Max entries in FCRETA */ +/* Higher 7 bits for the queue index */ +#define IXGBE_FCRETA_ENTRY_HIGH_MASK 0x007F0000 +#define IXGBE_FCRETA_ENTRY_HIGH_SHIFT 16 /* Stats registers */ #define IXGBE_CRCERRS 0x04000 From 87031c0df521d80af3b168bb2425a8faa999b629 Mon Sep 17 00:00:00 2001 From: Alexander Duyck Date: Thu, 9 Apr 2015 22:03:24 -0700 Subject: [PATCH 8/9] ixgbe: Remove NETIF_F_HW_VLAN_CTAG_FILTER from hw_features This change makes it so that the HW_VLAN_CTAG_FILTER bit is not falsely advertised as being a feature that can be toggled on ixgbe parts. The driver was setting the bit in features and letting it be inherited by hw_features, however the driver was actually ignoring the value of the bit and just assuming it was always set. As a result VLAN filtering was always enabled which is a requirement for SR-IOV, VMDq, DCB, FCoE, and possibly other features within the adapters. Signed-off-by: Alexander Duyck Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index c696ad721153a..0b38be8a1a528 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -8455,7 +8455,6 @@ static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent) NETIF_F_IPV6_CSUM | NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX | - NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_RXHASH | @@ -8477,6 +8476,7 @@ static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent) } netdev->hw_features |= NETIF_F_RXALL; + netdev->features |= NETIF_F_HW_VLAN_CTAG_FILTER; netdev->vlan_features |= NETIF_F_TSO; netdev->vlan_features |= NETIF_F_TSO6; From 020f1d9d9ae3f476d887f2b043223d8a89652174 Mon Sep 17 00:00:00 2001 From: Alexander Duyck Date: Thu, 9 Apr 2015 22:03:24 -0700 Subject: [PATCH 9/9] ixgbe: Drop unnecessary call to rcu_barrier The ixgbe driver hasn't used call_rcu to free the rings for some time now. Since that is the case the call to rcu_barrier can be dropped since calls to kfree_rcu don't require it. Signed-off-by: Alexander Duyck Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index 0b38be8a1a528..a7acb2dee0099 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -9038,8 +9038,6 @@ static void __exit ixgbe_exit_module(void) pci_unregister_driver(&ixgbe_driver); ixgbe_dbg_exit(); - - rcu_barrier(); /* Wait for completion of call_rcu()'s */ } #ifdef CONFIG_IXGBE_DCA