From c9ebea04cb1b0a84d08de00f4800f726ca544ad2 Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Fri, 18 Jan 2019 15:57:55 +0000 Subject: [PATCH 1/8] mlxsw: pci: Ring CQ's doorbell before RDQ's When a packet should be trapped to the CPU the device consumes a WQE (work queue element) from an RDQ (receive descriptor queue) and copies the packet to the address specified in the WQE. The device then tries to post a CQE (completion queue element) that contains various metadata (e.g., ingress port) about the packet to a CQ (completion queue). In case the device managed to consume a WQE, but did not manage to post the corresponding CQE, it will get stuck. This unlikely situation can be triggered due to the scheme the driver is currently using to process CQEs. The driver will consume up to 512 CQEs at a time and after processing each corresponding WQE it will ring the RDQ's doorbell, letting the device know that a new WQE was posted for it to consume. Only after processing all the CQEs (up to 512), the driver will ring the CQ's doorbell, letting the device know that new ones can be posted. Fix this by having the driver ring the CQ's doorbell for every processed CQE, but before ringing the RDQ's doorbell. This guarantees that whenever we post a new WQE, there is a corresponding CQE available. Copy the currently processed CQE to prevent the device from overwriting it with a new CQE after ringing the doorbell. Note that the driver still arms the CQ only after processing all the pending CQEs, so that interrupts for this CQ will only be delivered after the driver finished its processing. Before commit 8404f6f2e8ed ("mlxsw: pci: Allow to use CQEs of version 1 and version 2") the issue was virtually impossible to trigger since the number of CQEs was twice the number of WQEs and the number of CQEs processed at a time was equal to the number of available WQEs. Fixes: 8404f6f2e8ed ("mlxsw: pci: Allow to use CQEs of version 1 and version 2") Signed-off-by: Ido Schimmel Reported-by: Semion Lisyansky Tested-by: Semion Lisyansky Acked-by: Jiri Pirko Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlxsw/pci.c | 12 +++++++----- drivers/net/ethernet/mellanox/mlxsw/pci_hw.h | 1 + 2 files changed, 8 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci.c b/drivers/net/ethernet/mellanox/mlxsw/pci.c index 66b8098c6fd2e2..91bf294f76776d 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/pci.c +++ b/drivers/net/ethernet/mellanox/mlxsw/pci.c @@ -604,29 +604,31 @@ static void mlxsw_pci_cq_tasklet(unsigned long data) u16 wqe_counter = mlxsw_pci_cqe_wqe_counter_get(cqe); u8 sendq = mlxsw_pci_cqe_sr_get(q->u.cq.v, cqe); u8 dqn = mlxsw_pci_cqe_dqn_get(q->u.cq.v, cqe); + char ncqe[MLXSW_PCI_CQE_SIZE_MAX]; + + memcpy(ncqe, cqe, q->elem_size); + mlxsw_pci_queue_doorbell_consumer_ring(mlxsw_pci, q); if (sendq) { struct mlxsw_pci_queue *sdq; sdq = mlxsw_pci_sdq_get(mlxsw_pci, dqn); mlxsw_pci_cqe_sdq_handle(mlxsw_pci, sdq, - wqe_counter, cqe); + wqe_counter, ncqe); q->u.cq.comp_sdq_count++; } else { struct mlxsw_pci_queue *rdq; rdq = mlxsw_pci_rdq_get(mlxsw_pci, dqn); mlxsw_pci_cqe_rdq_handle(mlxsw_pci, rdq, - wqe_counter, q->u.cq.v, cqe); + wqe_counter, q->u.cq.v, ncqe); q->u.cq.comp_rdq_count++; } if (++items == credits) break; } - if (items) { - mlxsw_pci_queue_doorbell_consumer_ring(mlxsw_pci, q); + if (items) mlxsw_pci_queue_doorbell_arm_consumer_ring(mlxsw_pci, q); - } } static u16 mlxsw_pci_cq_elem_count(const struct mlxsw_pci_queue *q) diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h b/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h index bb99f6d41fe0b9..4265dd394f2711 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h +++ b/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h @@ -53,6 +53,7 @@ #define MLXSW_PCI_WQE_SIZE 32 /* 32 bytes per element */ #define MLXSW_PCI_CQE01_SIZE 16 /* 16 bytes per element */ #define MLXSW_PCI_CQE2_SIZE 32 /* 32 bytes per element */ +#define MLXSW_PCI_CQE_SIZE_MAX MLXSW_PCI_CQE2_SIZE #define MLXSW_PCI_EQE_SIZE 16 /* 16 bytes per element */ #define MLXSW_PCI_WQE_COUNT (MLXSW_PCI_AQ_SIZE / MLXSW_PCI_WQE_SIZE) #define MLXSW_PCI_CQE01_COUNT (MLXSW_PCI_AQ_SIZE / MLXSW_PCI_CQE01_SIZE) From d2f372ba0914e5722ac28e15f2ed2db61bcf0e44 Mon Sep 17 00:00:00 2001 From: Nir Dotan Date: Fri, 18 Jan 2019 15:57:56 +0000 Subject: [PATCH 2/8] mlxsw: pci: Increase PCI SW reset timeout Spectrum-2 PHY layer introduces a calibration period which is a part of the Spectrum-2 firmware boot process. Hence increase the SW timeout waiting for the firmware to come out of boot. This does not increase system boot time in cases where the firmware PHY calibration process is done quickly. Fixes: c3ab435466d5 ("mlxsw: spectrum: Extend to support Spectrum-2 ASIC") Signed-off-by: Nir Dotan Acked-by: Jiri Pirko Signed-off-by: Ido Schimmel Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlxsw/pci_hw.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h b/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h index 4265dd394f2711..ffee38e36ce899 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h +++ b/drivers/net/ethernet/mellanox/mlxsw/pci_hw.h @@ -27,7 +27,7 @@ #define MLXSW_PCI_SW_RESET 0xF0010 #define MLXSW_PCI_SW_RESET_RST_BIT BIT(0) -#define MLXSW_PCI_SW_RESET_TIMEOUT_MSECS 5000 +#define MLXSW_PCI_SW_RESET_TIMEOUT_MSECS 13000 #define MLXSW_PCI_SW_RESET_WAIT_MSECS 100 #define MLXSW_PCI_FW_READY 0xA1844 #define MLXSW_PCI_FW_READY_MASK 0xFFFF From 67c14cc9b35055264fc0efed00159a7de1819f1b Mon Sep 17 00:00:00 2001 From: Nir Dotan Date: Fri, 18 Jan 2019 15:57:57 +0000 Subject: [PATCH 3/8] mlxsw: pci: Return error on PCI reset timeout Return an appropriate error in the case when the driver timeouts on waiting for firmware to go out of PCI reset. Fixes: 233fa44bd67a ("mlxsw: pci: Implement reset done check") Signed-off-by: Nir Dotan Acked-by: Jiri Pirko Signed-off-by: Ido Schimmel Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlxsw/pci.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci.c b/drivers/net/ethernet/mellanox/mlxsw/pci.c index 91bf294f76776d..a2321fe8d6a0f2 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/pci.c +++ b/drivers/net/ethernet/mellanox/mlxsw/pci.c @@ -1367,10 +1367,10 @@ static int mlxsw_pci_sw_reset(struct mlxsw_pci *mlxsw_pci, u32 val = mlxsw_pci_read32(mlxsw_pci, FW_READY); if ((val & MLXSW_PCI_FW_READY_MASK) == MLXSW_PCI_FW_READY_MAGIC) - break; + return 0; cond_resched(); } while (time_before(jiffies, end)); - return 0; + return -EBUSY; } static int mlxsw_pci_alloc_irq_vectors(struct mlxsw_pci *mlxsw_pci) From a11dcd6497915ba79d95ef4fe2541aaac27f6201 Mon Sep 17 00:00:00 2001 From: Nir Dotan Date: Fri, 18 Jan 2019 15:57:59 +0000 Subject: [PATCH 4/8] mlxsw: spectrum_fid: Update dummy FID index When using a tc flower action of egress mirred redirect, the driver adds an implicit FID setting action. This implicit action sets a dummy FID to the packet and is used as part of a design for trapping unmatched flows in OVS. While this implicit FID setting action is supposed to be a NOP when a redirect action is added, in Spectrum-2 the FID record is consulted as the dummy FID index is an 802.1D FID index and the packet is dropped instead of being redirected. Set the dummy FID index value to be within 802.1Q range. This satisfies both Spectrum-1 which ignores the FID and Spectrum-2 which identifies it as an 802.1Q FID and will then follow the redirect action. Fixes: c3ab435466d5 ("mlxsw: spectrum: Extend to support Spectrum-2 ASIC") Signed-off-by: Nir Dotan Signed-off-by: Ido Schimmel Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c index 055cc6943b348d..9d9aa28684af7a 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c +++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_fid.c @@ -997,8 +997,8 @@ static const struct mlxsw_sp_fid_ops mlxsw_sp_fid_dummy_ops = { static const struct mlxsw_sp_fid_family mlxsw_sp_fid_dummy_family = { .type = MLXSW_SP_FID_TYPE_DUMMY, .fid_size = sizeof(struct mlxsw_sp_fid), - .start_index = MLXSW_SP_RFID_BASE - 1, - .end_index = MLXSW_SP_RFID_BASE - 1, + .start_index = VLAN_N_VID - 1, + .end_index = VLAN_N_VID - 1, .ops = &mlxsw_sp_fid_dummy_ops, }; From 710ae72877378e7cde611efd30fe90502a6e5b30 Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Fri, 18 Jan 2019 15:58:00 +0000 Subject: [PATCH 5/8] net: bridge: Mark FDB entries that were added by user as such Externally learned entries can be added by a user or by a switch driver that is notifying the bridge driver about entries that were learned in hardware. In the first case, the entries are not marked with the 'added_by_user' flag, which causes switch drivers to ignore them and not offload them. The 'added_by_user' flag can be set on externally learned FDB entries based on the 'swdev_notify' parameter in br_fdb_external_learn_add(), which effectively means if the created / updated FDB entry was added by a user or not. Fixes: 816a3bed9549 ("switchdev: Add fdb.added_by_user to switchdev notifications") Signed-off-by: Ido Schimmel Reported-by: Alexander Petrovskiy Reviewed-by: Petr Machata Cc: Roopa Prabhu Cc: Nikolay Aleksandrov Cc: bridge@lists.linux-foundation.org Signed-off-by: David S. Miller --- net/bridge/br_fdb.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/net/bridge/br_fdb.c b/net/bridge/br_fdb.c index fe3c758791ca99..9e14767500ea89 100644 --- a/net/bridge/br_fdb.c +++ b/net/bridge/br_fdb.c @@ -1128,6 +1128,8 @@ int br_fdb_external_learn_add(struct net_bridge *br, struct net_bridge_port *p, err = -ENOMEM; goto err_unlock; } + if (swdev_notify) + fdb->added_by_user = 1; fdb->added_by_external_learn = 1; fdb_notify(br, fdb, RTM_NEWNEIGH, swdev_notify); } else { @@ -1147,6 +1149,9 @@ int br_fdb_external_learn_add(struct net_bridge *br, struct net_bridge_port *p, modified = true; } + if (swdev_notify) + fdb->added_by_user = 1; + if (modified) fdb_notify(br, fdb, RTM_NEWNEIGH, swdev_notify); } From 64254a2054611205798e6bde634639bc704573ac Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Fri, 18 Jan 2019 15:58:01 +0000 Subject: [PATCH 6/8] mlxsw: spectrum_switchdev: Do not treat static FDB entries as sticky The driver currently treats static FDB entries as both static and sticky. This is incorrect and prevents such entries from being roamed to a different port via learning. Fix this by configuring static entries with ageing disabled and roaming enabled. In net-next we can add proper support for the newly introduced 'sticky' flag. Fixes: 56ade8fe3fe1 ("mlxsw: spectrum: Add initial support for Spectrum ASIC") Signed-off-by: Ido Schimmel Reported-by: Alexander Petrovskiy Reviewed-by: Petr Machata Signed-off-by: David S. Miller --- .../net/ethernet/mellanox/mlxsw/spectrum_switchdev.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c index 0abbaa0fbf14c3..c772109b638d62 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c +++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c @@ -1233,7 +1233,7 @@ mlxsw_sp_bridge_port_fdb_flush(struct mlxsw_sp *mlxsw_sp, static enum mlxsw_reg_sfd_rec_policy mlxsw_sp_sfd_rec_policy(bool dynamic) { return dynamic ? MLXSW_REG_SFD_REC_POLICY_DYNAMIC_ENTRY_INGRESS : - MLXSW_REG_SFD_REC_POLICY_STATIC_ENTRY; + MLXSW_REG_SFD_REC_POLICY_DYNAMIC_ENTRY_MLAG; } static enum mlxsw_reg_sfd_op mlxsw_sp_sfd_op(bool adding) @@ -1290,7 +1290,7 @@ static int mlxsw_sp_port_fdb_tunnel_uc_op(struct mlxsw_sp *mlxsw_sp, static int __mlxsw_sp_port_fdb_uc_op(struct mlxsw_sp *mlxsw_sp, u8 local_port, const char *mac, u16 fid, bool adding, enum mlxsw_reg_sfd_rec_action action, - bool dynamic) + enum mlxsw_reg_sfd_rec_policy policy) { char *sfd_pl; u8 num_rec; @@ -1301,8 +1301,7 @@ static int __mlxsw_sp_port_fdb_uc_op(struct mlxsw_sp *mlxsw_sp, u8 local_port, return -ENOMEM; mlxsw_reg_sfd_pack(sfd_pl, mlxsw_sp_sfd_op(adding), 0); - mlxsw_reg_sfd_uc_pack(sfd_pl, 0, mlxsw_sp_sfd_rec_policy(dynamic), - mac, fid, action, local_port); + mlxsw_reg_sfd_uc_pack(sfd_pl, 0, policy, mac, fid, action, local_port); num_rec = mlxsw_reg_sfd_num_rec_get(sfd_pl); err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sfd), sfd_pl); if (err) @@ -1321,7 +1320,8 @@ static int mlxsw_sp_port_fdb_uc_op(struct mlxsw_sp *mlxsw_sp, u8 local_port, bool dynamic) { return __mlxsw_sp_port_fdb_uc_op(mlxsw_sp, local_port, mac, fid, adding, - MLXSW_REG_SFD_REC_ACTION_NOP, dynamic); + MLXSW_REG_SFD_REC_ACTION_NOP, + mlxsw_sp_sfd_rec_policy(dynamic)); } int mlxsw_sp_rif_fdb_op(struct mlxsw_sp *mlxsw_sp, const char *mac, u16 fid, @@ -1329,7 +1329,7 @@ int mlxsw_sp_rif_fdb_op(struct mlxsw_sp *mlxsw_sp, const char *mac, u16 fid, { return __mlxsw_sp_port_fdb_uc_op(mlxsw_sp, 0, mac, fid, adding, MLXSW_REG_SFD_REC_ACTION_FORWARD_IP_ROUTER, - false); + MLXSW_REG_SFD_REC_POLICY_STATIC_ENTRY); } static int mlxsw_sp_port_fdb_uc_lag_op(struct mlxsw_sp *mlxsw_sp, u16 lag_id, From 72159272fc08230a76791dd5d0a0019ede1b802c Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Fri, 18 Jan 2019 15:58:02 +0000 Subject: [PATCH 7/8] selftests: mlxsw: Test FDB offload indication Test that externally learned FDB entries added from user space are marked as offloaded. Signed-off-by: Ido Schimmel Reviewed-by: Petr Machata Signed-off-by: David S. Miller --- .../selftests/drivers/net/mlxsw/rtnetlink.sh | 20 +++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/tools/testing/selftests/drivers/net/mlxsw/rtnetlink.sh b/tools/testing/selftests/drivers/net/mlxsw/rtnetlink.sh index 94fdbf215c145e..c4cf6e6d800ebe 100755 --- a/tools/testing/selftests/drivers/net/mlxsw/rtnetlink.sh +++ b/tools/testing/selftests/drivers/net/mlxsw/rtnetlink.sh @@ -25,6 +25,7 @@ ALL_TESTS=" lag_unlink_slaves_test lag_dev_deletion_test vlan_interface_uppers_test + bridge_extern_learn_test devlink_reload_test " NUM_NETIFS=2 @@ -541,6 +542,25 @@ vlan_interface_uppers_test() ip link del dev br0 } +bridge_extern_learn_test() +{ + # Test that externally learned entries added from user space are + # marked as offloaded + RET=0 + + ip link add name br0 type bridge + ip link set dev $swp1 master br0 + + bridge fdb add de:ad:be:ef:13:37 dev $swp1 master extern_learn + + bridge fdb show brport $swp1 | grep de:ad:be:ef:13:37 | grep -q offload + check_err $? "fdb entry not marked as offloaded when should" + + log_test "externally learned fdb entry" + + ip link del dev br0 +} + devlink_reload_test() { # Test that after executing all the above configuration tests, a From 479a2b761d61c04e2ae97325aa391a8a8c99c23e Mon Sep 17 00:00:00 2001 From: Ido Schimmel Date: Fri, 18 Jan 2019 15:58:03 +0000 Subject: [PATCH 8/8] selftests: forwarding: Add a test case for externally learned FDB entries Test that externally learned FDB entries can roam, but not age out. Signed-off-by: Ido Schimmel Reviewed-by: Petr Machata Signed-off-by: David S. Miller --- .../net/forwarding/bridge_vlan_aware.sh | 34 ++++++++++++++++++- 1 file changed, 33 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/net/forwarding/bridge_vlan_aware.sh b/tools/testing/selftests/net/forwarding/bridge_vlan_aware.sh index 04c6431b2bd8c6..b90dff8d3a94bf 100755 --- a/tools/testing/selftests/net/forwarding/bridge_vlan_aware.sh +++ b/tools/testing/selftests/net/forwarding/bridge_vlan_aware.sh @@ -1,7 +1,7 @@ #!/bin/bash # SPDX-License-Identifier: GPL-2.0 -ALL_TESTS="ping_ipv4 ping_ipv6 learning flooding vlan_deletion" +ALL_TESTS="ping_ipv4 ping_ipv6 learning flooding vlan_deletion extern_learn" NUM_NETIFS=4 CHECK_TC="yes" source lib.sh @@ -109,6 +109,38 @@ vlan_deletion() ping_ipv6 } +extern_learn() +{ + local mac=de:ad:be:ef:13:37 + local ageing_time + + # Test that externally learned FDB entries can roam, but not age out + RET=0 + + bridge fdb add de:ad:be:ef:13:37 dev $swp1 master extern_learn vlan 1 + + bridge fdb show brport $swp1 | grep -q de:ad:be:ef:13:37 + check_err $? "Did not find FDB entry when should" + + # Wait for 10 seconds after the ageing time to make sure the FDB entry + # was not aged out + ageing_time=$(bridge_ageing_time_get br0) + sleep $((ageing_time + 10)) + + bridge fdb show brport $swp1 | grep -q de:ad:be:ef:13:37 + check_err $? "FDB entry was aged out when should not" + + $MZ $h2 -c 1 -p 64 -a $mac -t ip -q + + bridge fdb show brport $swp2 | grep -q de:ad:be:ef:13:37 + check_err $? "FDB entry did not roam when should" + + log_test "Externally learned FDB entry - ageing & roaming" + + bridge fdb del de:ad:be:ef:13:37 dev $swp2 master vlan 1 &> /dev/null + bridge fdb del de:ad:be:ef:13:37 dev $swp1 master vlan 1 &> /dev/null +} + trap cleanup EXIT setup_prepare