From b352c3465bb808ab700d03f5bac2f7a6f37c5350 Mon Sep 17 00:00:00 2001 From: Xiaoke Wang Date: Fri, 18 Feb 2022 10:19:39 +0800 Subject: net: ll_temac: check the return value of devm_kmalloc() devm_kmalloc() returns a pointer to allocated memory on success, NULL on failure. While lp->indirect_lock is allocated by devm_kmalloc() without proper check. It is better to check the value of it to prevent potential wrong memory access. Fixes: f14f5c11f051 ("net: ll_temac: Support indirect_mutex share within TEMAC IP") Signed-off-by: Xiaoke Wang Signed-off-by: David S. Miller --- drivers/net/ethernet/xilinx/ll_temac_main.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/xilinx/ll_temac_main.c b/drivers/net/ethernet/xilinx/ll_temac_main.c index b900ab5aef2a..64c7e26c3b75 100644 --- a/drivers/net/ethernet/xilinx/ll_temac_main.c +++ b/drivers/net/ethernet/xilinx/ll_temac_main.c @@ -1433,6 +1433,8 @@ static int temac_probe(struct platform_device *pdev) lp->indirect_lock = devm_kmalloc(&pdev->dev, sizeof(*lp->indirect_lock), GFP_KERNEL); + if (!lp->indirect_lock) + return -ENOMEM; spin_lock_init(lp->indirect_lock); } -- cgit v1.2.3-70-g09d2 From b70bc066d77b460a63a8c3fb2ea0d811ce862a83 Mon Sep 17 00:00:00 2001 From: Wojciech Drewek Date: Fri, 17 Dec 2021 12:36:25 +0100 Subject: ice: Match on all profiles in slow-path In switchdev mode, slow-path rules need to match all protocols, in order to correctly redirect unfiltered or missed packets to the uplink. To set this up for the virtual function to uplink flow, the rule that redirects packets to the control VSI must have the tunnel type set to ICE_SW_TUN_AND_NON_TUN. As a result of that new tunnel type being set, ice_get_compat_fv_bitmap will select ICE_PROF_ALL. At that point all profiles would be selected for this rule, resulting in the desired behavior. Without this change slow-path would not work with tunnel protocols. Fixes: 8b032a55c1bd ("ice: low level support for tunnels") Signed-off-by: Wojciech Drewek Tested-by: Sandeep Penigalapati Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_eswitch.c | 1 + drivers/net/ethernet/intel/ice/ice_protocol_type.h | 1 + drivers/net/ethernet/intel/ice/ice_switch.c | 4 +++- 3 files changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_eswitch.c b/drivers/net/ethernet/intel/ice/ice_eswitch.c index 864692b157b6..73edc24d81d5 100644 --- a/drivers/net/ethernet/intel/ice/ice_eswitch.c +++ b/drivers/net/ethernet/intel/ice/ice_eswitch.c @@ -44,6 +44,7 @@ ice_eswitch_add_vf_mac_rule(struct ice_pf *pf, struct ice_vf *vf, const u8 *mac) ctrl_vsi->rxq_map[vf->vf_id]; rule_info.flags_info.act |= ICE_SINGLE_ACT_LB_ENABLE; rule_info.flags_info.act_valid = true; + rule_info.tun_type = ICE_SW_TUN_AND_NON_TUN; err = ice_add_adv_rule(hw, list, lkups_cnt, &rule_info, vf->repr->mac_rule); diff --git a/drivers/net/ethernet/intel/ice/ice_protocol_type.h b/drivers/net/ethernet/intel/ice/ice_protocol_type.h index dc1b0e9e6df5..695b6dd61dc2 100644 --- a/drivers/net/ethernet/intel/ice/ice_protocol_type.h +++ b/drivers/net/ethernet/intel/ice/ice_protocol_type.h @@ -47,6 +47,7 @@ enum ice_protocol_type { enum ice_sw_tunnel_type { ICE_NON_TUN = 0, + ICE_SW_TUN_AND_NON_TUN, ICE_SW_TUN_VXLAN, ICE_SW_TUN_GENEVE, ICE_SW_TUN_NVGRE, diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c index 11ae0bee3590..475ec2afa210 100644 --- a/drivers/net/ethernet/intel/ice/ice_switch.c +++ b/drivers/net/ethernet/intel/ice/ice_switch.c @@ -4537,6 +4537,7 @@ ice_get_compat_fv_bitmap(struct ice_hw *hw, struct ice_adv_rule_info *rinfo, case ICE_SW_TUN_NVGRE: prof_type = ICE_PROF_TUN_GRE; break; + case ICE_SW_TUN_AND_NON_TUN: default: prof_type = ICE_PROF_ALL; break; @@ -5305,7 +5306,8 @@ ice_add_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, if (status) goto err_ice_add_adv_rule; - if (rinfo->tun_type != ICE_NON_TUN) { + if (rinfo->tun_type != ICE_NON_TUN && + rinfo->tun_type != ICE_SW_TUN_AND_NON_TUN) { status = ice_fill_adv_packet_tun(hw, rinfo->tun_type, s_rule->pdata.lkup_tx_rx.hdr, pkt_offsets); -- cgit v1.2.3-70-g09d2 From 932645c298c41aad64ef13016ff4c2034eef5aed Mon Sep 17 00:00:00 2001 From: Michal Swiatkowski Date: Mon, 3 Jan 2022 07:41:21 +0100 Subject: ice: fix setting l4 port flag when adding filter Accidentally filter flag for none encapsulated l4 port field is always set. Even if user wants to add encapsulated l4 port field. Remove this unnecessary flag setting. Fixes: 9e300987d4a81 ("ice: VXLAN and Geneve TC support") Signed-off-by: Michal Swiatkowski Tested-by: Sandeep Penigalapati Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_tc_lib.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_tc_lib.c b/drivers/net/ethernet/intel/ice/ice_tc_lib.c index e8aab664270a..65cf32eb4046 100644 --- a/drivers/net/ethernet/intel/ice/ice_tc_lib.c +++ b/drivers/net/ethernet/intel/ice/ice_tc_lib.c @@ -709,7 +709,7 @@ ice_tc_set_port(struct flow_match_ports match, fltr->flags |= ICE_TC_FLWR_FIELD_ENC_DEST_L4_PORT; else fltr->flags |= ICE_TC_FLWR_FIELD_DEST_L4_PORT; - fltr->flags |= ICE_TC_FLWR_FIELD_DEST_L4_PORT; + headers->l4_key.dst_port = match.key->dst; headers->l4_mask.dst_port = match.mask->dst; } @@ -718,7 +718,7 @@ ice_tc_set_port(struct flow_match_ports match, fltr->flags |= ICE_TC_FLWR_FIELD_ENC_SRC_L4_PORT; else fltr->flags |= ICE_TC_FLWR_FIELD_SRC_L4_PORT; - fltr->flags |= ICE_TC_FLWR_FIELD_SRC_L4_PORT; + headers->l4_key.src_port = match.key->src; headers->l4_mask.src_port = match.mask->src; } -- cgit v1.2.3-70-g09d2 From fadead80fe4c033b5e514fcbadd20b55c4494112 Mon Sep 17 00:00:00 2001 From: Jacob Keller Date: Mon, 7 Feb 2022 10:23:29 -0800 Subject: ice: fix concurrent reset and removal of VFs Commit c503e63200c6 ("ice: Stop processing VF messages during teardown") introduced a driver state flag, ICE_VF_DEINIT_IN_PROGRESS, which is intended to prevent some issues with concurrently handling messages from VFs while tearing down the VFs. This change was motivated by crashes caused while tearing down and bringing up VFs in rapid succession. It turns out that the fix actually introduces issues with the VF driver caused because the PF no longer responds to any messages sent by the VF during its .remove routine. This results in the VF potentially removing its DMA memory before the PF has shut down the device queues. Additionally, the fix doesn't actually resolve concurrency issues within the ice driver. It is possible for a VF to initiate a reset just prior to the ice driver removing VFs. This can result in the remove task concurrently operating while the VF is being reset. This results in similar memory corruption and panics purportedly fixed by that commit. Fix this concurrency at its root by protecting both the reset and removal flows using the existing VF cfg_lock. This ensures that we cannot remove the VF while any outstanding critical tasks such as a virtchnl message or a reset are occurring. This locking change also fixes the root cause originally fixed by commit c503e63200c6 ("ice: Stop processing VF messages during teardown"), so we can simply revert it. Note that I kept these two changes together because simply reverting the original commit alone would leave the driver vulnerable to worse race conditions. Fixes: c503e63200c6 ("ice: Stop processing VF messages during teardown") Signed-off-by: Jacob Keller Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice.h | 1 - drivers/net/ethernet/intel/ice/ice_main.c | 2 ++ drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 42 ++++++++++++++---------- 3 files changed, 27 insertions(+), 18 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h index a9fa701aaa95..473b1f6be9de 100644 --- a/drivers/net/ethernet/intel/ice/ice.h +++ b/drivers/net/ethernet/intel/ice/ice.h @@ -280,7 +280,6 @@ enum ice_pf_state { ICE_VFLR_EVENT_PENDING, ICE_FLTR_OVERFLOW_PROMISC, ICE_VF_DIS, - ICE_VF_DEINIT_IN_PROGRESS, ICE_CFG_BUSY, ICE_SERVICE_SCHED, ICE_SERVICE_DIS, diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 17a9bb461dc3..f3c346e13b7a 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -1799,7 +1799,9 @@ static void ice_handle_mdd_event(struct ice_pf *pf) * reset, so print the event prior to reset. */ ice_print_vf_rx_mdd_event(vf); + mutex_lock(&pf->vf[i].cfg_lock); ice_reset_vf(&pf->vf[i], false); + mutex_unlock(&pf->vf[i].cfg_lock); } } } diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c index 39b80124d282..408f78e3eb13 100644 --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c @@ -500,8 +500,6 @@ void ice_free_vfs(struct ice_pf *pf) struct ice_hw *hw = &pf->hw; unsigned int tmp, i; - set_bit(ICE_VF_DEINIT_IN_PROGRESS, pf->state); - if (!pf->vf) return; @@ -519,22 +517,26 @@ void ice_free_vfs(struct ice_pf *pf) else dev_warn(dev, "VFs are assigned - not disabling SR-IOV\n"); - /* Avoid wait time by stopping all VFs at the same time */ - ice_for_each_vf(pf, i) - ice_dis_vf_qs(&pf->vf[i]); - tmp = pf->num_alloc_vfs; pf->num_qps_per_vf = 0; pf->num_alloc_vfs = 0; for (i = 0; i < tmp; i++) { - if (test_bit(ICE_VF_STATE_INIT, pf->vf[i].vf_states)) { + struct ice_vf *vf = &pf->vf[i]; + + mutex_lock(&vf->cfg_lock); + + ice_dis_vf_qs(vf); + + if (test_bit(ICE_VF_STATE_INIT, vf->vf_states)) { /* disable VF qp mappings and set VF disable state */ - ice_dis_vf_mappings(&pf->vf[i]); - set_bit(ICE_VF_STATE_DIS, pf->vf[i].vf_states); - ice_free_vf_res(&pf->vf[i]); + ice_dis_vf_mappings(vf); + set_bit(ICE_VF_STATE_DIS, vf->vf_states); + ice_free_vf_res(vf); } - mutex_destroy(&pf->vf[i].cfg_lock); + mutex_unlock(&vf->cfg_lock); + + mutex_destroy(&vf->cfg_lock); } if (ice_sriov_free_msix_res(pf)) @@ -570,7 +572,6 @@ void ice_free_vfs(struct ice_pf *pf) i); clear_bit(ICE_VF_DIS, pf->state); - clear_bit(ICE_VF_DEINIT_IN_PROGRESS, pf->state); clear_bit(ICE_FLAG_SRIOV_ENA, pf->flags); } @@ -1498,6 +1499,8 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr) ice_for_each_vf(pf, v) { vf = &pf->vf[v]; + mutex_lock(&vf->cfg_lock); + vf->driver_caps = 0; ice_vc_set_default_allowlist(vf); @@ -1512,6 +1515,8 @@ bool ice_reset_all_vfs(struct ice_pf *pf, bool is_vflr) ice_vf_pre_vsi_rebuild(vf); ice_vf_rebuild_vsi(vf); ice_vf_post_vsi_rebuild(vf); + + mutex_unlock(&vf->cfg_lock); } if (ice_is_eswitch_mode_switchdev(pf)) @@ -1562,6 +1567,8 @@ bool ice_reset_vf(struct ice_vf *vf, bool is_vflr) u32 reg; int i; + lockdep_assert_held(&vf->cfg_lock); + dev = ice_pf_to_dev(pf); if (test_bit(ICE_VF_RESETS_DISABLED, pf->state)) { @@ -2061,9 +2068,12 @@ void ice_process_vflr_event(struct ice_pf *pf) bit_idx = (hw->func_caps.vf_base_id + vf_id) % 32; /* read GLGEN_VFLRSTAT register to find out the flr VFs */ reg = rd32(hw, GLGEN_VFLRSTAT(reg_idx)); - if (reg & BIT(bit_idx)) + if (reg & BIT(bit_idx)) { /* GLGEN_VFLRSTAT bit will be cleared in ice_reset_vf */ + mutex_lock(&vf->cfg_lock); ice_reset_vf(vf, true); + mutex_unlock(&vf->cfg_lock); + } } } @@ -2140,7 +2150,9 @@ ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event) if (!vf) return; + mutex_lock(&vf->cfg_lock); ice_vc_reset_vf(vf); + mutex_unlock(&vf->cfg_lock); } /** @@ -4625,10 +4637,6 @@ void ice_vc_process_vf_msg(struct ice_pf *pf, struct ice_rq_event_info *event) struct device *dev; int err = 0; - /* if de-init is underway, don't process messages from VF */ - if (test_bit(ICE_VF_DEINIT_IN_PROGRESS, pf->state)) - return; - dev = ice_pf_to_dev(pf); if (ice_validate_vf_id(pf, vf_id)) { err = -EINVAL; -- cgit v1.2.3-70-g09d2 From ed22d9c8d128293fc7b0b086c7d3654bcb99a8dd Mon Sep 17 00:00:00 2001 From: Tom Rix Date: Mon, 14 Feb 2022 06:33:27 -0800 Subject: ice: check the return of ice_ptp_gettimex64 Clang static analysis reports this issue time64.h:69:50: warning: The left operand of '+' is a garbage value set_normalized_timespec64(&ts_delta, lhs.tv_sec + rhs.tv_sec, ~~~~~~~~~~ ^ In ice_ptp_adjtime_nonatomic(), the timespec64 variable 'now' is set by ice_ptp_gettimex64(). This function can fail with -EBUSY, so 'now' can have a gargbage value. So check the return. Fixes: 06c16d89d2cb ("ice: register 1588 PTP clock device object for E810 devices") Signed-off-by: Tom Rix Tested-by: Gurucharan G (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_ptp.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.c b/drivers/net/ethernet/intel/ice/ice_ptp.c index ae291d442539..000c39d163a2 100644 --- a/drivers/net/ethernet/intel/ice/ice_ptp.c +++ b/drivers/net/ethernet/intel/ice/ice_ptp.c @@ -1533,9 +1533,12 @@ exit: static int ice_ptp_adjtime_nonatomic(struct ptp_clock_info *info, s64 delta) { struct timespec64 now, then; + int ret; then = ns_to_timespec64(delta); - ice_ptp_gettimex64(info, &now, NULL); + ret = ice_ptp_gettimex64(info, &now, NULL); + if (ret) + return ret; now = timespec64_add(now, then); return ice_ptp_settime64(info, (const struct timespec64 *)&now); -- cgit v1.2.3-70-g09d2 From 5950bdc88dd1d158f2845fdff8fb1de86476806c Mon Sep 17 00:00:00 2001 From: Tom Rix Date: Mon, 14 Feb 2022 07:40:43 -0800 Subject: ice: initialize local variable 'tlv' Clang static analysis reports this issues ice_common.c:5008:21: warning: The left expression of the compound assignment is an uninitialized value. The computed value will also be garbage ldo->phy_type_low |= ((u64)buf << (i * 16)); ~~~~~~~~~~~~~~~~~ ^ When called from ice_cfg_phy_fec() ldo is the uninitialized local variable tlv. So initialize. Fixes: ea78ce4dab05 ("ice: add link lenient and default override support") Signed-off-by: Tom Rix Tested-by: Gurucharan G (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_common.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c index a6d7d3eff186..e2af99a763ed 100644 --- a/drivers/net/ethernet/intel/ice/ice_common.c +++ b/drivers/net/ethernet/intel/ice/ice_common.c @@ -3340,7 +3340,7 @@ ice_cfg_phy_fec(struct ice_port_info *pi, struct ice_aqc_set_phy_cfg_data *cfg, if (fec == ICE_FEC_AUTO && ice_fw_supports_link_override(hw) && !ice_fw_supports_report_dflt_cfg(hw)) { - struct ice_link_default_override_tlv tlv; + struct ice_link_default_override_tlv tlv = { 0 }; status = ice_get_link_default_override(&tlv, pi); if (status) -- cgit v1.2.3-70-g09d2 From 5a2aba71cd2610d3ed08867a1b1bf617cd8f89b8 Mon Sep 17 00:00:00 2001 From: Jeremy Linton Date: Mon, 14 Feb 2022 17:18:52 -0600 Subject: net: mvpp2: always set port pcs ops Booting a MACCHIATObin with 5.17, the system OOPs with a null pointer deref when the network is started. This is caused by the pcs->ops structure being null in mcpp2_acpi_start() when it tries to call pcs_config(). Hoisting the code which sets pcs_gmac.ops and pcs_xlg.ops, assuring they are always set, fixes the problem. The OOPs looks like: [ 18.687760] Unable to handle kernel access to user memory outside uaccess routines at virtual address 0000000000000010 [ 18.698561] Mem abort info: [ 18.698564] ESR = 0x96000004 [ 18.698567] EC = 0x25: DABT (current EL), IL = 32 bits [ 18.709821] SET = 0, FnV = 0 [ 18.714292] EA = 0, S1PTW = 0 [ 18.718833] FSC = 0x04: level 0 translation fault [ 18.725126] Data abort info: [ 18.729408] ISV = 0, ISS = 0x00000004 [ 18.734655] CM = 0, WnR = 0 [ 18.738933] user pgtable: 4k pages, 48-bit VAs, pgdp=0000000111bbf000 [ 18.745409] [0000000000000010] pgd=0000000000000000, p4d=0000000000000000 [ 18.752235] Internal error: Oops: 96000004 [#1] SMP [ 18.757134] Modules linked in: rfkill ip_set nf_tables nfnetlink qrtr sunrpc vfat fat omap_rng fuse zram xfs crct10dif_ce mvpp2 ghash_ce sbsa_gwdt phylink xhci_plat_hcd ahci_plam [ 18.773481] CPU: 0 PID: 681 Comm: NetworkManager Not tainted 5.17.0-0.rc3.89.fc36.aarch64 #1 [ 18.781954] Hardware name: Marvell Armada 7k/8k Family Board /Armada 7k/8k Family Board , BIOS EDK II Jun 4 2019 [ 18.795222] pstate: 60400005 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) [ 18.802213] pc : mvpp2_start_dev+0x2b0/0x300 [mvpp2] [ 18.807208] lr : mvpp2_start_dev+0x298/0x300 [mvpp2] [ 18.812197] sp : ffff80000b4732c0 [ 18.815522] x29: ffff80000b4732c0 x28: 0000000000000000 x27: ffffccab38ae57f8 [ 18.822689] x26: ffff6eeb03065a10 x25: ffff80000b473a30 x24: ffff80000b4735b8 [ 18.829855] x23: 0000000000000000 x22: 00000000000001e0 x21: ffff6eeb07b6ab68 [ 18.837021] x20: ffff6eeb07b6ab30 x19: ffff6eeb07b6a9c0 x18: 0000000000000014 [ 18.844187] x17: 00000000f6232bfe x16: ffffccab899b1dc0 x15: 000000006a30f9fa [ 18.851353] x14: 000000003b77bd50 x13: 000006dc896f0e8e x12: 001bbbfccfd0d3a2 [ 18.858519] x11: 0000000000001528 x10: 0000000000001548 x9 : ffffccab38ad0fb0 [ 18.865685] x8 : ffff80000b473330 x7 : 0000000000000000 x6 : 0000000000000000 [ 18.872851] x5 : 0000000000000000 x4 : 0000000000000000 x3 : ffff80000b4732f8 [ 18.880017] x2 : 000000000000001a x1 : 0000000000000002 x0 : ffff6eeb07b6ab68 [ 18.887183] Call trace: [ 18.889637] mvpp2_start_dev+0x2b0/0x300 [mvpp2] [ 18.894279] mvpp2_open+0x134/0x2b4 [mvpp2] [ 18.898483] __dev_open+0x128/0x1e4 [ 18.901988] __dev_change_flags+0x17c/0x1d0 [ 18.906187] dev_change_flags+0x30/0x70 [ 18.910038] do_setlink+0x278/0xa7c [ 18.913540] __rtnl_newlink+0x44c/0x7d0 [ 18.917391] rtnl_newlink+0x5c/0x8c [ 18.920892] rtnetlink_rcv_msg+0x254/0x314 [ 18.925006] netlink_rcv_skb+0x48/0x10c [ 18.928858] rtnetlink_rcv+0x24/0x30 [ 18.932449] netlink_unicast+0x290/0x2f4 [ 18.936386] netlink_sendmsg+0x1d0/0x41c [ 18.940323] sock_sendmsg+0x60/0x70 [ 18.943825] ____sys_sendmsg+0x248/0x260 [ 18.947762] ___sys_sendmsg+0x74/0xa0 [ 18.951438] __sys_sendmsg+0x64/0xcc [ 18.955027] __arm64_sys_sendmsg+0x30/0x40 [ 18.959140] invoke_syscall+0x50/0x120 [ 18.962906] el0_svc_common.constprop.0+0x4c/0xf4 [ 18.967629] do_el0_svc+0x30/0x9c [ 18.970958] el0_svc+0x28/0xb0 [ 18.974025] el0t_64_sync_handler+0x10c/0x140 [ 18.978400] el0t_64_sync+0x1a4/0x1a8 [ 18.982078] Code: 52800004 b9416262 aa1503e0 52800041 (f94008a5) [ 18.988196] ---[ end trace 0000000000000000 ]--- Fixes: cff056322372 ("net: mvpp2: use .mac_select_pcs() interface") Suggested-by: Russell King (Oracle) Signed-off-by: Jeremy Linton Reviewed-by: Marcin Wojtas Link: https://lore.kernel.org/r/20220214231852.3331430-1-jeremy.linton@arm.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c index 7cdbf8b8bbf6..1a835b48791b 100644 --- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c @@ -6870,6 +6870,9 @@ static int mvpp2_port_probe(struct platform_device *pdev, dev->max_mtu = MVPP2_BM_JUMBO_PKT_SIZE; dev->dev.of_node = port_node; + port->pcs_gmac.ops = &mvpp2_phylink_gmac_pcs_ops; + port->pcs_xlg.ops = &mvpp2_phylink_xlg_pcs_ops; + if (!mvpp2_use_acpi_compat_mode(port_fwnode)) { port->phylink_config.dev = &dev->dev; port->phylink_config.type = PHYLINK_NETDEV; @@ -6940,9 +6943,6 @@ static int mvpp2_port_probe(struct platform_device *pdev, port->phylink_config.supported_interfaces); } - port->pcs_gmac.ops = &mvpp2_phylink_gmac_pcs_ops; - port->pcs_xlg.ops = &mvpp2_phylink_xlg_pcs_ops; - phylink = phylink_create(&port->phylink_config, port_fwnode, phy_mode, &mvpp2_phylink_ops); if (IS_ERR(phylink)) { -- cgit v1.2.3-70-g09d2 From 3a14d0888eb4b0045884126acc69abfb7b87814d Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Fri, 18 Feb 2022 14:15:35 +0100 Subject: nfp: flower: Fix a potential leak in nfp_tunnel_add_shared_mac() ida_simple_get() returns an id between min (0) and max (NFP_MAX_MAC_INDEX) inclusive. So NFP_MAX_MAC_INDEX (0xff) is a valid id. In order for the error handling path to work correctly, the 'invalid' value for 'ida_idx' should not be in the 0..NFP_MAX_MAC_INDEX range, inclusive. So set it to -1. Fixes: 20cce8865098 ("nfp: flower: enable MAC address sharing for offloadable devs") Signed-off-by: Christophe JAILLET Signed-off-by: Simon Horman Link: https://lore.kernel.org/r/20220218131535.100258-1-simon.horman@corigine.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c index 0a326e04e692..cb43651ea9ba 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c +++ b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c @@ -922,8 +922,8 @@ nfp_tunnel_add_shared_mac(struct nfp_app *app, struct net_device *netdev, int port, bool mod) { struct nfp_flower_priv *priv = app->priv; - int ida_idx = NFP_MAX_MAC_INDEX, err; struct nfp_tun_offloaded_mac *entry; + int ida_idx = -1, err; u16 nfp_mac_idx = 0; entry = nfp_tunnel_lookup_offloaded_macs(app, netdev->dev_addr); @@ -997,7 +997,7 @@ err_remove_hash: err_free_entry: kfree(entry); err_free_ida: - if (ida_idx != NFP_MAX_MAC_INDEX) + if (ida_idx != -1) ida_simple_remove(&priv->tun.mac_off_ids, ida_idx); return err; -- cgit v1.2.3-70-g09d2 From 84d3c83e6ea7d46cf3de3a54578af73eb24a64f2 Mon Sep 17 00:00:00 2001 From: Somnath Kotur Date: Sun, 20 Feb 2022 04:05:47 -0500 Subject: bnxt_en: Fix active FEC reporting to ethtool ethtool --show-fec does not show anything when the Active FEC setting in the chip is set to None. Fix it to properly return ETHTOOL_FEC_OFF in that case. Fixes: 8b2775890ad8 ("bnxt_en: Report FEC settings to ethtool.") Signed-off-by: Somnath Kotur Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c index 003330e8cd58..e195f4a669d8 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c @@ -1969,6 +1969,9 @@ static int bnxt_get_fecparam(struct net_device *dev, case PORT_PHY_QCFG_RESP_ACTIVE_FEC_FEC_RS272_IEEE_ACTIVE: fec->active_fec |= ETHTOOL_FEC_LLRS; break; + case PORT_PHY_QCFG_RESP_ACTIVE_FEC_FEC_NONE_ACTIVE: + fec->active_fec |= ETHTOOL_FEC_OFF; + break; } return 0; } -- cgit v1.2.3-70-g09d2 From 6758f937669dba14c6aac7ca004edda42ec1b18d Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Sun, 20 Feb 2022 04:05:48 -0500 Subject: bnxt_en: Fix offline ethtool selftest with RDMA enabled For offline (destructive) self tests, we need to stop the RDMA driver first. Otherwise, the RDMA driver will run into unrecoverable errors when destructive firmware tests are being performed. The irq_re_init parameter used in the half close and half open sequence when preparing the NIC for offline tests should be set to true because the RDMA driver will free all IRQs before the offline tests begin. Fixes: 55fd0cf320c3 ("bnxt_en: Add external loopback test to ethtool selftest.") Reviewed-by: Edwin Peer Reviewed-by: Ben Li Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 10 +++++----- drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 12 +++++++++--- 2 files changed, 14 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 4f94136a011a..23bbb1c5812d 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -10330,12 +10330,12 @@ int bnxt_half_open_nic(struct bnxt *bp) goto half_open_err; } - rc = bnxt_alloc_mem(bp, false); + rc = bnxt_alloc_mem(bp, true); if (rc) { netdev_err(bp->dev, "bnxt_alloc_mem err: %x\n", rc); goto half_open_err; } - rc = bnxt_init_nic(bp, false); + rc = bnxt_init_nic(bp, true); if (rc) { netdev_err(bp->dev, "bnxt_init_nic err: %x\n", rc); goto half_open_err; @@ -10344,7 +10344,7 @@ int bnxt_half_open_nic(struct bnxt *bp) half_open_err: bnxt_free_skbs(bp); - bnxt_free_mem(bp, false); + bnxt_free_mem(bp, true); dev_close(bp->dev); return rc; } @@ -10354,9 +10354,9 @@ half_open_err: */ void bnxt_half_close_nic(struct bnxt *bp) { - bnxt_hwrm_resource_free(bp, false, false); + bnxt_hwrm_resource_free(bp, false, true); bnxt_free_skbs(bp); - bnxt_free_mem(bp, false); + bnxt_free_mem(bp, true); } void bnxt_reenable_sriov(struct bnxt *bp) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c index e195f4a669d8..a85b18858b32 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c @@ -25,6 +25,7 @@ #include "bnxt_hsi.h" #include "bnxt.h" #include "bnxt_hwrm.h" +#include "bnxt_ulp.h" #include "bnxt_xdp.h" #include "bnxt_ptp.h" #include "bnxt_ethtool.h" @@ -3551,9 +3552,12 @@ static void bnxt_self_test(struct net_device *dev, struct ethtool_test *etest, if (!offline) { bnxt_run_fw_tests(bp, test_mask, &test_results); } else { - rc = bnxt_close_nic(bp, false, false); - if (rc) + bnxt_ulp_stop(bp); + rc = bnxt_close_nic(bp, true, false); + if (rc) { + bnxt_ulp_start(bp, rc); return; + } bnxt_run_fw_tests(bp, test_mask, &test_results); buf[BNXT_MACLPBK_TEST_IDX] = 1; @@ -3563,6 +3567,7 @@ static void bnxt_self_test(struct net_device *dev, struct ethtool_test *etest, if (rc) { bnxt_hwrm_mac_loopback(bp, false); etest->flags |= ETH_TEST_FL_FAILED; + bnxt_ulp_start(bp, rc); return; } if (bnxt_run_loopback(bp)) @@ -3588,7 +3593,8 @@ static void bnxt_self_test(struct net_device *dev, struct ethtool_test *etest, } bnxt_hwrm_phy_loopback(bp, false, false); bnxt_half_close_nic(bp); - rc = bnxt_open_nic(bp, false, true); + rc = bnxt_open_nic(bp, true, true); + bnxt_ulp_start(bp, rc); } if (rc || bnxt_test_irq(bp)) { buf[BNXT_IRQ_TEST_IDX] = 1; -- cgit v1.2.3-70-g09d2 From cfcab3b3b61584a02bb523ffa99564eafa761dfe Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Sun, 20 Feb 2022 04:05:49 -0500 Subject: bnxt_en: Fix occasional ethtool -t loopback test failures In the current code, we setup the port to PHY or MAC loopback mode and then transmit a test broadcast packet for the loopback test. This scheme fails sometime if the port is shared with management firmware that can also send packets. The driver may receive the management firmware's packet and the test will fail when the contents don't match the test packet. Change the test packet to use it's own MAC address as the destination and setup the port to only receive it's own MAC address. This should filter out other packets sent by management firmware. Fixes: 91725d89b97a ("bnxt_en: Add PHY loopback to ethtool self-test.") Reviewed-by: Pavan Chebbi Reviewed-by: Edwin Peer Reviewed-by: Andy Gospodarek Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 7 +++++++ drivers/net/ethernet/broadcom/bnxt/bnxt.h | 1 + drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 2 +- 3 files changed, 9 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 23bbb1c5812d..785436f6dd24 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -8639,6 +8639,9 @@ static int bnxt_init_chip(struct bnxt *bp, bool irq_re_init) vnic->uc_filter_count = 1; vnic->rx_mask = 0; + if (test_bit(BNXT_STATE_HALF_OPEN, &bp->state)) + goto skip_rx_mask; + if (bp->dev->flags & IFF_BROADCAST) vnic->rx_mask |= CFA_L2_SET_RX_MASK_REQ_MASK_BCAST; @@ -8659,6 +8662,7 @@ static int bnxt_init_chip(struct bnxt *bp, bool irq_re_init) if (rc) goto err_out; +skip_rx_mask: rc = bnxt_hwrm_set_coal(bp); if (rc) netdev_warn(bp->dev, "HWRM set coalescing failure rc: %x\n", @@ -10335,8 +10339,10 @@ int bnxt_half_open_nic(struct bnxt *bp) netdev_err(bp->dev, "bnxt_alloc_mem err: %x\n", rc); goto half_open_err; } + set_bit(BNXT_STATE_HALF_OPEN, &bp->state); rc = bnxt_init_nic(bp, true); if (rc) { + clear_bit(BNXT_STATE_HALF_OPEN, &bp->state); netdev_err(bp->dev, "bnxt_init_nic err: %x\n", rc); goto half_open_err; } @@ -10357,6 +10363,7 @@ void bnxt_half_close_nic(struct bnxt *bp) bnxt_hwrm_resource_free(bp, false, true); bnxt_free_skbs(bp); bnxt_free_mem(bp, true); + clear_bit(BNXT_STATE_HALF_OPEN, &bp->state); } void bnxt_reenable_sriov(struct bnxt *bp) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h index 440dfeb4948b..666fc1e7a7d2 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h @@ -1921,6 +1921,7 @@ struct bnxt { #define BNXT_STATE_RECOVER 12 #define BNXT_STATE_FW_NON_FATAL_COND 13 #define BNXT_STATE_FW_ACTIVATE_RESET 14 +#define BNXT_STATE_HALF_OPEN 15 /* For offline ethtool tests */ #define BNXT_NO_FW_ACCESS(bp) \ (test_bit(BNXT_STATE_FW_FATAL_COND, &(bp)->state) || \ diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c index a85b18858b32..8aaa2335f848 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c @@ -3458,7 +3458,7 @@ static int bnxt_run_loopback(struct bnxt *bp) if (!skb) return -ENOMEM; data = skb_put(skb, pkt_size); - eth_broadcast_addr(data); + ether_addr_copy(&data[i], bp->dev->dev_addr); i += ETH_ALEN; ether_addr_copy(&data[i], bp->dev->dev_addr); i += ETH_ALEN; -- cgit v1.2.3-70-g09d2 From 8cdb15924252e27af16c4a8fe0fc606ce5fd04dc Mon Sep 17 00:00:00 2001 From: Pavan Chebbi Date: Sun, 20 Feb 2022 04:05:50 -0500 Subject: bnxt_en: Fix incorrect multicast rx mask setting when not requested We should setup multicast only when net_device flags explicitly has IFF_MULTICAST set. Otherwise we will incorrectly turn it on even when not asked. Fix it by only passing the multicast table to the firmware if IFF_MULTICAST is set. Fixes: 7d2837dd7a32 ("bnxt_en: Setup multicast properly after resetting device.") Signed-off-by: Pavan Chebbi Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 785436f6dd24..fc5b1d816bdb 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -4747,8 +4747,10 @@ static int bnxt_hwrm_cfa_l2_set_rx_mask(struct bnxt *bp, u16 vnic_id) return rc; req->vnic_id = cpu_to_le32(vnic->fw_vnic_id); - req->num_mc_entries = cpu_to_le32(vnic->mc_list_count); - req->mc_tbl_addr = cpu_to_le64(vnic->mc_list_mapping); + if (vnic->rx_mask & CFA_L2_SET_RX_MASK_REQ_MASK_MCAST) { + req->num_mc_entries = cpu_to_le32(vnic->mc_list_count); + req->mc_tbl_addr = cpu_to_le64(vnic->mc_list_mapping); + } req->mask = cpu_to_le32(vnic->rx_mask); return hwrm_req_send_silent(bp, req); } @@ -8651,7 +8653,7 @@ static int bnxt_init_chip(struct bnxt *bp, bool irq_re_init) if (bp->dev->flags & IFF_ALLMULTI) { vnic->rx_mask |= CFA_L2_SET_RX_MASK_REQ_MASK_ALL_MCAST; vnic->mc_list_count = 0; - } else { + } else if (bp->dev->flags & IFF_MULTICAST) { u32 mask = 0; bnxt_mc_list_updated(bp, &mask); @@ -10779,7 +10781,7 @@ static void bnxt_set_rx_mode(struct net_device *dev) if (dev->flags & IFF_ALLMULTI) { mask |= CFA_L2_SET_RX_MASK_REQ_MASK_ALL_MCAST; vnic->mc_list_count = 0; - } else { + } else if (dev->flags & IFF_MULTICAST) { mc_update = bnxt_mc_list_updated(bp, &mask); } @@ -10856,9 +10858,10 @@ skip_uc: !bnxt_promisc_ok(bp)) vnic->rx_mask &= ~CFA_L2_SET_RX_MASK_REQ_MASK_PROMISCUOUS; rc = bnxt_hwrm_cfa_l2_set_rx_mask(bp, 0); - if (rc && vnic->mc_list_count) { + if (rc && (vnic->rx_mask & CFA_L2_SET_RX_MASK_REQ_MASK_MCAST)) { netdev_info(bp->dev, "Failed setting MC filters rc: %d, turning on ALL_MCAST mode\n", rc); + vnic->rx_mask &= ~CFA_L2_SET_RX_MASK_REQ_MASK_MCAST; vnic->rx_mask |= CFA_L2_SET_RX_MASK_REQ_MASK_ALL_MCAST; vnic->mc_list_count = 0; rc = bnxt_hwrm_cfa_l2_set_rx_mask(bp, 0); -- cgit v1.2.3-70-g09d2 From 0e0e3c5358470cbad10bd7ca29f84a44d179d286 Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Sun, 20 Feb 2022 04:05:51 -0500 Subject: bnxt_en: Restore the resets_reliable flag in bnxt_open() During ifdown, we call bnxt_inv_fw_health_reg() which will clear both the status_reliable and resets_reliable flags if these registers are mapped. This is correct because a FW reset during ifdown will clear these register mappings. If we detect that FW has gone through reset during the next ifup, we will remap these registers. But during normal ifup with no FW reset, we need to restore the resets_reliable flag otherwise we will not show the reset counter during devlink diagnose. Fixes: 8cc95ceb7087 ("bnxt_en: improve fw diagnose devlink health messages") Reviewed-by: Vikas Gupta Reviewed-by: Pavan Chebbi Reviewed-by: Somnath Kotur Signed-off-by: Kalesh AP Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index fc5b1d816bdb..b1c98d1408b8 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -7789,6 +7789,19 @@ static int bnxt_map_fw_health_regs(struct bnxt *bp) return 0; } +static void bnxt_remap_fw_health_regs(struct bnxt *bp) +{ + if (!bp->fw_health) + return; + + if (bp->fw_cap & BNXT_FW_CAP_ERROR_RECOVERY) { + bp->fw_health->status_reliable = true; + bp->fw_health->resets_reliable = true; + } else { + bnxt_try_map_fw_health_reg(bp); + } +} + static int bnxt_hwrm_error_recovery_qcfg(struct bnxt *bp) { struct bnxt_fw_health *fw_health = bp->fw_health; @@ -9856,8 +9869,8 @@ static int bnxt_hwrm_if_change(struct bnxt *bp, bool up) resc_reinit = true; if (flags & FUNC_DRV_IF_CHANGE_RESP_FLAGS_HOT_FW_RESET_DONE) fw_reset = true; - else if (bp->fw_health && !bp->fw_health->status_reliable) - bnxt_try_map_fw_health_reg(bp); + else + bnxt_remap_fw_health_regs(bp); if (test_bit(BNXT_STATE_IN_FW_RESET, &bp->state) && !fw_reset) { netdev_err(bp->dev, "RESET_DONE not set during FW reset.\n"); -- cgit v1.2.3-70-g09d2 From b891106da52b2c12dbaf73400f6d225b06a38d80 Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Sun, 20 Feb 2022 04:05:52 -0500 Subject: bnxt_en: Increase firmware message response DMA wait time When polling for the firmware message response, we first poll for the response message header. Once the valid length is detected in the header, we poll for the valid bit at the end of the message which signals DMA completion. Normally, this poll time for DMA completion is extremely short (0 to a few usec). But on some devices under some rare conditions, it can be up to about 20 msec. Increase this delay to 50 msec and use udelay() for the first 10 usec for the common case, and usleep_range() beyond that. Also, change the error message to include the above delay time when printing the timeout value. Fixes: 3c8c20db769c ("bnxt_en: move HWRM API implementation into separate file") Reviewed-by: Vladimir Olovyannikov Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.c | 12 +++++++++--- drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.h | 2 +- 2 files changed, 10 insertions(+), 4 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.c index 566c9487ef55..b01d42928a53 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.c @@ -644,17 +644,23 @@ static int __hwrm_send(struct bnxt *bp, struct bnxt_hwrm_ctx *ctx) /* Last byte of resp contains valid bit */ valid = ((u8 *)ctx->resp) + len - 1; - for (j = 0; j < HWRM_VALID_BIT_DELAY_USEC; j++) { + for (j = 0; j < HWRM_VALID_BIT_DELAY_USEC; ) { /* make sure we read from updated DMA memory */ dma_rmb(); if (*valid) break; - usleep_range(1, 5); + if (j < 10) { + udelay(1); + j++; + } else { + usleep_range(20, 30); + j += 20; + } } if (j >= HWRM_VALID_BIT_DELAY_USEC) { hwrm_err(bp, ctx, "Error (timeout: %u) msg {0x%x 0x%x} len:%d v:%d\n", - hwrm_total_timeout(i), req_type, + hwrm_total_timeout(i) + j, req_type, le16_to_cpu(ctx->req->seq_id), len, *valid); goto exit; } diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.h index d52bd2d63aec..c98032e38188 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.h +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_hwrm.h @@ -90,7 +90,7 @@ static inline unsigned int hwrm_total_timeout(unsigned int n) } -#define HWRM_VALID_BIT_DELAY_USEC 150 +#define HWRM_VALID_BIT_DELAY_USEC 50000 static inline bool bnxt_cfa_hwrm_message(u16 req_type) { -- cgit v1.2.3-70-g09d2 From 1278d17a1fb860e7eab4bc3ff4b026a87cbf5105 Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Sun, 20 Feb 2022 04:05:53 -0500 Subject: bnxt_en: Fix devlink fw_activate To install a livepatch, first flash the package to NVM, and then activate the patch through the "HWRM_FW_LIVEPATCH" fw command. To uninstall a patch from NVM, flash the removal package and then activate it through the "HWRM_FW_LIVEPATCH" fw command. The "HWRM_FW_LIVEPATCH" fw command has to consider following scenarios: 1. no patch in NVM and no patch active. Do nothing. 2. patch in NVM, but not active. Activate the patch currently in NVM. 3. patch is not in NVM, but active. Deactivate the patch. 4. patch in NVM and the patch active. Do nothing. Fix the code to handle these scenarios during devlink "fw_activate". To install and activate a live patch: devlink dev flash pci/0000:c1:00.0 file thor_patch.pkg devlink -f dev reload pci/0000:c1:00.0 action fw_activate limit no_reset To remove and deactivate a live patch: devlink dev flash pci/0000:c1:00.0 file thor_patch_rem.pkg devlink -f dev reload pci/0000:c1:00.0 action fw_activate limit no_reset Fixes: 3c4153394e2c ("bnxt_en: implement firmware live patching") Reviewed-by: Vikas Gupta Reviewed-by: Somnath Kotur Signed-off-by: Kalesh AP Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c | 39 ++++++++++++++++++----- 1 file changed, 31 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c index 4da31b1b84f9..f6e21fac0e69 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c @@ -367,6 +367,16 @@ bnxt_dl_livepatch_report_err(struct bnxt *bp, struct netlink_ext_ack *extack, } } +/* Live patch status in NVM */ +#define BNXT_LIVEPATCH_NOT_INSTALLED 0 +#define BNXT_LIVEPATCH_INSTALLED FW_LIVEPATCH_QUERY_RESP_STATUS_FLAGS_INSTALL +#define BNXT_LIVEPATCH_REMOVED FW_LIVEPATCH_QUERY_RESP_STATUS_FLAGS_ACTIVE +#define BNXT_LIVEPATCH_MASK (FW_LIVEPATCH_QUERY_RESP_STATUS_FLAGS_INSTALL | \ + FW_LIVEPATCH_QUERY_RESP_STATUS_FLAGS_ACTIVE) +#define BNXT_LIVEPATCH_ACTIVATED BNXT_LIVEPATCH_MASK + +#define BNXT_LIVEPATCH_STATE(flags) ((flags) & BNXT_LIVEPATCH_MASK) + static int bnxt_dl_livepatch_activate(struct bnxt *bp, struct netlink_ext_ack *extack) { @@ -374,8 +384,9 @@ bnxt_dl_livepatch_activate(struct bnxt *bp, struct netlink_ext_ack *extack) struct hwrm_fw_livepatch_query_input *query_req; struct hwrm_fw_livepatch_output *patch_resp; struct hwrm_fw_livepatch_input *patch_req; + u16 flags, live_patch_state; + bool activated = false; u32 installed = 0; - u16 flags; u8 target; int rc; @@ -394,7 +405,6 @@ bnxt_dl_livepatch_activate(struct bnxt *bp, struct netlink_ext_ack *extack) hwrm_req_drop(bp, query_req); return rc; } - patch_req->opcode = FW_LIVEPATCH_REQ_OPCODE_ACTIVATE; patch_req->loadtype = FW_LIVEPATCH_REQ_LOADTYPE_NVM_INSTALL; patch_resp = hwrm_req_hold(bp, patch_req); @@ -407,12 +417,20 @@ bnxt_dl_livepatch_activate(struct bnxt *bp, struct netlink_ext_ack *extack) } flags = le16_to_cpu(query_resp->status_flags); - if (~flags & FW_LIVEPATCH_QUERY_RESP_STATUS_FLAGS_INSTALL) + live_patch_state = BNXT_LIVEPATCH_STATE(flags); + + if (live_patch_state == BNXT_LIVEPATCH_NOT_INSTALLED) continue; - if ((flags & FW_LIVEPATCH_QUERY_RESP_STATUS_FLAGS_ACTIVE) && - !strncmp(query_resp->active_ver, query_resp->install_ver, - sizeof(query_resp->active_ver))) + + if (live_patch_state == BNXT_LIVEPATCH_ACTIVATED) { + activated = true; continue; + } + + if (live_patch_state == BNXT_LIVEPATCH_INSTALLED) + patch_req->opcode = FW_LIVEPATCH_REQ_OPCODE_ACTIVATE; + else if (live_patch_state == BNXT_LIVEPATCH_REMOVED) + patch_req->opcode = FW_LIVEPATCH_REQ_OPCODE_DEACTIVATE; patch_req->fw_target = target; rc = hwrm_req_send(bp, patch_req); @@ -424,8 +442,13 @@ bnxt_dl_livepatch_activate(struct bnxt *bp, struct netlink_ext_ack *extack) } if (!rc && !installed) { - NL_SET_ERR_MSG_MOD(extack, "No live patches found"); - rc = -ENOENT; + if (activated) { + NL_SET_ERR_MSG_MOD(extack, "Live patch already activated"); + rc = -EEXIST; + } else { + NL_SET_ERR_MSG_MOD(extack, "No live patches found"); + rc = -ENOENT; + } } hwrm_req_drop(bp, query_req); hwrm_req_drop(bp, patch_req); -- cgit v1.2.3-70-g09d2 From 277f2bb14361790a70e4b3c649e794b75a91a597 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Mon, 21 Feb 2022 15:05:45 -0600 Subject: ibmvnic: schedule failover only if vioctl fails If client is unable to initiate a failover reset via H_VIOCTL hcall, then it should schedule a failover reset as a last resort. Otherwise, there is no need to do a last resort. Fixes: 334c42414729 ("ibmvnic: improve failover sysfs entry") Reported-by: Cris Forno Signed-off-by: Sukadev Bhattiprolu Signed-off-by: Dany Madden Link: https://lore.kernel.org/r/20220221210545.115283-1-drt@linux.ibm.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/ibm/ibmvnic.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 29617a86b299..dee05a353dbd 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -5917,10 +5917,14 @@ static ssize_t failover_store(struct device *dev, struct device_attribute *attr, be64_to_cpu(session_token)); rc = plpar_hcall_norets(H_VIOCTL, adapter->vdev->unit_address, H_SESSION_ERR_DETECTED, session_token, 0, 0); - if (rc) + if (rc) { netdev_err(netdev, "H_VIOCTL initiated failover failed, rc %ld\n", rc); + goto last_resort; + } + + return count; last_resort: netdev_dbg(netdev, "Trying to send CRQ_CMD, the last resort\n"); -- cgit v1.2.3-70-g09d2 From 4f1e72850d452e5c3302faa82a01f179ff5f9482 Mon Sep 17 00:00:00 2001 From: Heyi Guo Date: Wed, 23 Feb 2022 11:14:34 +0800 Subject: drivers/net/ftgmac100: refactor ftgmac100_reset_task to enable direct function call This is to prepare for ftgmac100_adjust_link() to call reset function directly, instead of task schedule. Signed-off-by: Heyi Guo Signed-off-by: David S. Miller --- drivers/net/ethernet/faraday/ftgmac100.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/faraday/ftgmac100.c b/drivers/net/ethernet/faraday/ftgmac100.c index 691605c15265..1f3eb4431475 100644 --- a/drivers/net/ethernet/faraday/ftgmac100.c +++ b/drivers/net/ethernet/faraday/ftgmac100.c @@ -1410,10 +1410,8 @@ static int ftgmac100_init_all(struct ftgmac100 *priv, bool ignore_alloc_err) return err; } -static void ftgmac100_reset_task(struct work_struct *work) +static void ftgmac100_reset(struct ftgmac100 *priv) { - struct ftgmac100 *priv = container_of(work, struct ftgmac100, - reset_task); struct net_device *netdev = priv->netdev; int err; @@ -1459,6 +1457,14 @@ static void ftgmac100_reset_task(struct work_struct *work) rtnl_unlock(); } +static void ftgmac100_reset_task(struct work_struct *work) +{ + struct ftgmac100 *priv = container_of(work, struct ftgmac100, + reset_task); + + ftgmac100_reset(priv); +} + static int ftgmac100_open(struct net_device *netdev) { struct ftgmac100 *priv = netdev_priv(netdev); -- cgit v1.2.3-70-g09d2 From 3c773dba8182cdfea7b32caafe9290240ab8de5f Mon Sep 17 00:00:00 2001 From: Heyi Guo Date: Wed, 23 Feb 2022 11:14:35 +0800 Subject: drivers/net/ftgmac100: adjust code place for function call dependency This is to prepare for ftgmac100_adjust_link() to call ftgmac100_reset() directly. Only code places are changed. Signed-off-by: Heyi Guo Signed-off-by: David S. Miller --- drivers/net/ethernet/faraday/ftgmac100.c | 222 +++++++++++++++---------------- 1 file changed, 111 insertions(+), 111 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/faraday/ftgmac100.c b/drivers/net/ethernet/faraday/ftgmac100.c index 1f3eb4431475..c1deb6e5d26c 100644 --- a/drivers/net/ethernet/faraday/ftgmac100.c +++ b/drivers/net/ethernet/faraday/ftgmac100.c @@ -989,117 +989,6 @@ static int ftgmac100_alloc_rx_buffers(struct ftgmac100 *priv) return 0; } -static void ftgmac100_adjust_link(struct net_device *netdev) -{ - struct ftgmac100 *priv = netdev_priv(netdev); - struct phy_device *phydev = netdev->phydev; - bool tx_pause, rx_pause; - int new_speed; - - /* We store "no link" as speed 0 */ - if (!phydev->link) - new_speed = 0; - else - new_speed = phydev->speed; - - /* Grab pause settings from PHY if configured to do so */ - if (priv->aneg_pause) { - rx_pause = tx_pause = phydev->pause; - if (phydev->asym_pause) - tx_pause = !rx_pause; - } else { - rx_pause = priv->rx_pause; - tx_pause = priv->tx_pause; - } - - /* Link hasn't changed, do nothing */ - if (phydev->speed == priv->cur_speed && - phydev->duplex == priv->cur_duplex && - rx_pause == priv->rx_pause && - tx_pause == priv->tx_pause) - return; - - /* Print status if we have a link or we had one and just lost it, - * don't print otherwise. - */ - if (new_speed || priv->cur_speed) - phy_print_status(phydev); - - priv->cur_speed = new_speed; - priv->cur_duplex = phydev->duplex; - priv->rx_pause = rx_pause; - priv->tx_pause = tx_pause; - - /* Link is down, do nothing else */ - if (!new_speed) - return; - - /* Disable all interrupts */ - iowrite32(0, priv->base + FTGMAC100_OFFSET_IER); - - /* Reset the adapter asynchronously */ - schedule_work(&priv->reset_task); -} - -static int ftgmac100_mii_probe(struct net_device *netdev) -{ - struct ftgmac100 *priv = netdev_priv(netdev); - struct platform_device *pdev = to_platform_device(priv->dev); - struct device_node *np = pdev->dev.of_node; - struct phy_device *phydev; - phy_interface_t phy_intf; - int err; - - /* Default to RGMII. It's a gigabit part after all */ - err = of_get_phy_mode(np, &phy_intf); - if (err) - phy_intf = PHY_INTERFACE_MODE_RGMII; - - /* Aspeed only supports these. I don't know about other IP - * block vendors so I'm going to just let them through for - * now. Note that this is only a warning if for some obscure - * reason the DT really means to lie about it or it's a newer - * part we don't know about. - * - * On the Aspeed SoC there are additionally straps and SCU - * control bits that could tell us what the interface is - * (or allow us to configure it while the IP block is held - * in reset). For now I chose to keep this driver away from - * those SoC specific bits and assume the device-tree is - * right and the SCU has been configured properly by pinmux - * or the firmware. - */ - if (priv->is_aspeed && !(phy_interface_mode_is_rgmii(phy_intf))) { - netdev_warn(netdev, - "Unsupported PHY mode %s !\n", - phy_modes(phy_intf)); - } - - phydev = phy_find_first(priv->mii_bus); - if (!phydev) { - netdev_info(netdev, "%s: no PHY found\n", netdev->name); - return -ENODEV; - } - - phydev = phy_connect(netdev, phydev_name(phydev), - &ftgmac100_adjust_link, phy_intf); - - if (IS_ERR(phydev)) { - netdev_err(netdev, "%s: Could not attach to PHY\n", netdev->name); - return PTR_ERR(phydev); - } - - /* Indicate that we support PAUSE frames (see comment in - * Documentation/networking/phy.rst) - */ - phy_support_asym_pause(phydev); - - /* Display what we found */ - phy_attached_info(phydev); - - return 0; -} - static int ftgmac100_mdiobus_read(struct mii_bus *bus, int phy_addr, int regnum) { struct net_device *netdev = bus->priv; @@ -1465,6 +1354,117 @@ static void ftgmac100_reset_task(struct work_struct *work) ftgmac100_reset(priv); } +static void ftgmac100_adjust_link(struct net_device *netdev) +{ + struct ftgmac100 *priv = netdev_priv(netdev); + struct phy_device *phydev = netdev->phydev; + bool tx_pause, rx_pause; + int new_speed; + + /* We store "no link" as speed 0 */ + if (!phydev->link) + new_speed = 0; + else + new_speed = phydev->speed; + + /* Grab pause settings from PHY if configured to do so */ + if (priv->aneg_pause) { + rx_pause = tx_pause = phydev->pause; + if (phydev->asym_pause) + tx_pause = !rx_pause; + } else { + rx_pause = priv->rx_pause; + tx_pause = priv->tx_pause; + } + + /* Link hasn't changed, do nothing */ + if (phydev->speed == priv->cur_speed && + phydev->duplex == priv->cur_duplex && + rx_pause == priv->rx_pause && + tx_pause == priv->tx_pause) + return; + + /* Print status if we have a link or we had one and just lost it, + * don't print otherwise. + */ + if (new_speed || priv->cur_speed) + phy_print_status(phydev); + + priv->cur_speed = new_speed; + priv->cur_duplex = phydev->duplex; + priv->rx_pause = rx_pause; + priv->tx_pause = tx_pause; + + /* Link is down, do nothing else */ + if (!new_speed) + return; + + /* Disable all interrupts */ + iowrite32(0, priv->base + FTGMAC100_OFFSET_IER); + + /* Reset the adapter asynchronously */ + schedule_work(&priv->reset_task); +} + +static int ftgmac100_mii_probe(struct net_device *netdev) +{ + struct ftgmac100 *priv = netdev_priv(netdev); + struct platform_device *pdev = to_platform_device(priv->dev); + struct device_node *np = pdev->dev.of_node; + struct phy_device *phydev; + phy_interface_t phy_intf; + int err; + + /* Default to RGMII. It's a gigabit part after all */ + err = of_get_phy_mode(np, &phy_intf); + if (err) + phy_intf = PHY_INTERFACE_MODE_RGMII; + + /* Aspeed only supports these. I don't know about other IP + * block vendors so I'm going to just let them through for + * now. Note that this is only a warning if for some obscure + * reason the DT really means to lie about it or it's a newer + * part we don't know about. + * + * On the Aspeed SoC there are additionally straps and SCU + * control bits that could tell us what the interface is + * (or allow us to configure it while the IP block is held + * in reset). For now I chose to keep this driver away from + * those SoC specific bits and assume the device-tree is + * right and the SCU has been configured properly by pinmux + * or the firmware. + */ + if (priv->is_aspeed && !(phy_interface_mode_is_rgmii(phy_intf))) { + netdev_warn(netdev, + "Unsupported PHY mode %s !\n", + phy_modes(phy_intf)); + } + + phydev = phy_find_first(priv->mii_bus); + if (!phydev) { + netdev_info(netdev, "%s: no PHY found\n", netdev->name); + return -ENODEV; + } + + phydev = phy_connect(netdev, phydev_name(phydev), + &ftgmac100_adjust_link, phy_intf); + + if (IS_ERR(phydev)) { + netdev_err(netdev, "%s: Could not attach to PHY\n", netdev->name); + return PTR_ERR(phydev); + } + + /* Indicate that we support PAUSE frames (see comment in + * Documentation/networking/phy.rst) + */ + phy_support_asym_pause(phydev); + + /* Display what we found */ + phy_attached_info(phydev); + + return 0; +} + static int ftgmac100_open(struct net_device *netdev) { struct ftgmac100 *priv = netdev_priv(netdev); -- cgit v1.2.3-70-g09d2 From 1baf2e50e48f10f0ea07d53e13381fd0da1546d2 Mon Sep 17 00:00:00 2001 From: Heyi Guo Date: Wed, 23 Feb 2022 11:14:36 +0800 Subject: drivers/net/ftgmac100: fix DHCP potential failure with systemd DHCP failures were observed with systemd 247.6. The issue could be reproduced by rebooting Aspeed 2600 and then running ifconfig ethX down/up. It is caused by below procedures in the driver: 1. ftgmac100_open() enables net interface and call phy_start() 2. When PHY is link up, it calls netif_carrier_on() and then adjust_link callback 3. ftgmac100_adjust_link() will schedule the reset task 4. ftgmac100_reset_task() will then reset the MAC in another schedule After step 2, systemd will be notified to send DHCP discover packet, while the packet might be corrupted by MAC reset operation in step 4. Call ftgmac100_reset() directly instead of scheduling task to fix the issue. Signed-off-by: Heyi Guo Signed-off-by: David S. Miller --- drivers/net/ethernet/faraday/ftgmac100.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/faraday/ftgmac100.c b/drivers/net/ethernet/faraday/ftgmac100.c index c1deb6e5d26c..d5356db7539a 100644 --- a/drivers/net/ethernet/faraday/ftgmac100.c +++ b/drivers/net/ethernet/faraday/ftgmac100.c @@ -1402,8 +1402,17 @@ static void ftgmac100_adjust_link(struct net_device *netdev) /* Disable all interrupts */ iowrite32(0, priv->base + FTGMAC100_OFFSET_IER); - /* Reset the adapter asynchronously */ - schedule_work(&priv->reset_task); + /* Release phy lock to allow ftgmac100_reset to aquire it, keeping lock + * order consistent to prevent dead lock. + */ + if (netdev->phydev) + mutex_unlock(&netdev->phydev->lock); + + ftgmac100_reset(priv); + + if (netdev->phydev) + mutex_lock(&netdev->phydev->lock); + } static int ftgmac100_mii_probe(struct net_device *netdev) -- cgit v1.2.3-70-g09d2 From f908a35b22180c4da64cf2647e4f5f0cd3054da7 Mon Sep 17 00:00:00 2001 From: Meir Lichtinger Date: Mon, 10 Jan 2022 10:14:41 +0200 Subject: net/mlx5: Update the list of the PCI supported devices Add the upcoming BlueField-4 and ConnectX-8 device IDs. Fixes: 2e9d3e83ab82 ("net/mlx5: Update the list of the PCI supported devices") Signed-off-by: Meir Lichtinger Reviewed-by: Gal Pressman Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/main.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 2c774f367199..13f913c13a2d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -1840,10 +1840,12 @@ static const struct pci_device_id mlx5_core_pci_table[] = { { PCI_VDEVICE(MELLANOX, 0x101e), MLX5_PCI_DEV_IS_VF}, /* ConnectX Family mlx5Gen Virtual Function */ { PCI_VDEVICE(MELLANOX, 0x101f) }, /* ConnectX-6 LX */ { PCI_VDEVICE(MELLANOX, 0x1021) }, /* ConnectX-7 */ + { PCI_VDEVICE(MELLANOX, 0x1023) }, /* ConnectX-8 */ { PCI_VDEVICE(MELLANOX, 0xa2d2) }, /* BlueField integrated ConnectX-5 network controller */ { PCI_VDEVICE(MELLANOX, 0xa2d3), MLX5_PCI_DEV_IS_VF}, /* BlueField integrated ConnectX-5 network controller VF */ { PCI_VDEVICE(MELLANOX, 0xa2d6) }, /* BlueField-2 integrated ConnectX-6 Dx network controller */ { PCI_VDEVICE(MELLANOX, 0xa2dc) }, /* BlueField-3 integrated ConnectX-7 network controller */ + { PCI_VDEVICE(MELLANOX, 0xa2df) }, /* BlueField-4 integrated ConnectX-8 network controller */ { 0, } }; -- cgit v1.2.3-70-g09d2 From e5b2bc30c21139ae10f0e56989389d0bc7b7b1d6 Mon Sep 17 00:00:00 2001 From: Yevgeny Kliteynik Date: Fri, 24 Dec 2021 01:07:30 +0200 Subject: net/mlx5: DR, Cache STE shadow memory During rule insertion on each ICM memory chunk we also allocate shadow memory used for management. This includes the hw_ste, dr_ste and miss list per entry. Since the scale of these allocations is large we noticed a performance hiccup that happens once malloc and free are stressed. In extreme usecases when ~1M chunks are freed at once, it might take up to 40 seconds to complete this, up to the point the kernel sees this as self-detected stall on CPU: rcu: INFO: rcu_sched self-detected stall on CPU To resolve this we will increase the reuse of shadow memory. Doing this we see that a time in the aforementioned usecase dropped from ~40 seconds to ~8-10 seconds. Fixes: 29cf8febd185 ("net/mlx5: DR, ICM pool memory allocator") Signed-off-by: Alex Vesker Signed-off-by: Yevgeny Kliteynik Signed-off-by: Saeed Mahameed --- .../mellanox/mlx5/core/steering/dr_icm_pool.c | 109 ++++++++++++++------- .../ethernet/mellanox/mlx5/core/steering/mlx5dr.h | 5 + 2 files changed, 79 insertions(+), 35 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c index 7f6fd9c5e371..f496b7e9401b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c @@ -136,37 +136,35 @@ static void dr_icm_pool_mr_destroy(struct mlx5dr_icm_mr *icm_mr) kvfree(icm_mr); } -static int dr_icm_chunk_ste_init(struct mlx5dr_icm_chunk *chunk) +static int dr_icm_buddy_get_ste_size(struct mlx5dr_icm_buddy_mem *buddy) { - chunk->ste_arr = kvzalloc(chunk->num_of_entries * - sizeof(chunk->ste_arr[0]), GFP_KERNEL); - if (!chunk->ste_arr) - return -ENOMEM; - - chunk->hw_ste_arr = kvzalloc(chunk->num_of_entries * - DR_STE_SIZE_REDUCED, GFP_KERNEL); - if (!chunk->hw_ste_arr) - goto out_free_ste_arr; - - chunk->miss_list = kvmalloc(chunk->num_of_entries * - sizeof(chunk->miss_list[0]), GFP_KERNEL); - if (!chunk->miss_list) - goto out_free_hw_ste_arr; + /* We support only one type of STE size, both for ConnectX-5 and later + * devices. Once the support for match STE which has a larger tag is + * added (32B instead of 16B), the STE size for devices later than + * ConnectX-5 needs to account for that. + */ + return DR_STE_SIZE_REDUCED; +} - return 0; +static void dr_icm_chunk_ste_init(struct mlx5dr_icm_chunk *chunk, int offset) +{ + struct mlx5dr_icm_buddy_mem *buddy = chunk->buddy_mem; + int index = offset / DR_STE_SIZE; -out_free_hw_ste_arr: - kvfree(chunk->hw_ste_arr); -out_free_ste_arr: - kvfree(chunk->ste_arr); - return -ENOMEM; + chunk->ste_arr = &buddy->ste_arr[index]; + chunk->miss_list = &buddy->miss_list[index]; + chunk->hw_ste_arr = buddy->hw_ste_arr + + index * dr_icm_buddy_get_ste_size(buddy); } static void dr_icm_chunk_ste_cleanup(struct mlx5dr_icm_chunk *chunk) { - kvfree(chunk->miss_list); - kvfree(chunk->hw_ste_arr); - kvfree(chunk->ste_arr); + struct mlx5dr_icm_buddy_mem *buddy = chunk->buddy_mem; + + memset(chunk->hw_ste_arr, 0, + chunk->num_of_entries * dr_icm_buddy_get_ste_size(buddy)); + memset(chunk->ste_arr, 0, + chunk->num_of_entries * sizeof(chunk->ste_arr[0])); } static enum mlx5dr_icm_type @@ -189,6 +187,44 @@ static void dr_icm_chunk_destroy(struct mlx5dr_icm_chunk *chunk, kvfree(chunk); } +static int dr_icm_buddy_init_ste_cache(struct mlx5dr_icm_buddy_mem *buddy) +{ + int num_of_entries = + mlx5dr_icm_pool_chunk_size_to_entries(buddy->pool->max_log_chunk_sz); + + buddy->ste_arr = kvcalloc(num_of_entries, + sizeof(struct mlx5dr_ste), GFP_KERNEL); + if (!buddy->ste_arr) + return -ENOMEM; + + /* Preallocate full STE size on non-ConnectX-5 devices since + * we need to support both full and reduced with the same cache. + */ + buddy->hw_ste_arr = kvcalloc(num_of_entries, + dr_icm_buddy_get_ste_size(buddy), GFP_KERNEL); + if (!buddy->hw_ste_arr) + goto free_ste_arr; + + buddy->miss_list = kvmalloc(num_of_entries * sizeof(struct list_head), GFP_KERNEL); + if (!buddy->miss_list) + goto free_hw_ste_arr; + + return 0; + +free_hw_ste_arr: + kvfree(buddy->hw_ste_arr); +free_ste_arr: + kvfree(buddy->ste_arr); + return -ENOMEM; +} + +static void dr_icm_buddy_cleanup_ste_cache(struct mlx5dr_icm_buddy_mem *buddy) +{ + kvfree(buddy->ste_arr); + kvfree(buddy->hw_ste_arr); + kvfree(buddy->miss_list); +} + static int dr_icm_buddy_create(struct mlx5dr_icm_pool *pool) { struct mlx5dr_icm_buddy_mem *buddy; @@ -208,11 +244,19 @@ static int dr_icm_buddy_create(struct mlx5dr_icm_pool *pool) buddy->icm_mr = icm_mr; buddy->pool = pool; + if (pool->icm_type == DR_ICM_TYPE_STE) { + /* Reduce allocations by preallocating and reusing the STE structures */ + if (dr_icm_buddy_init_ste_cache(buddy)) + goto err_cleanup_buddy; + } + /* add it to the -start- of the list in order to search in it first */ list_add(&buddy->list_node, &pool->buddy_mem_list); return 0; +err_cleanup_buddy: + mlx5dr_buddy_cleanup(buddy); err_free_buddy: kvfree(buddy); free_mr: @@ -234,6 +278,9 @@ static void dr_icm_buddy_destroy(struct mlx5dr_icm_buddy_mem *buddy) mlx5dr_buddy_cleanup(buddy); + if (buddy->pool->icm_type == DR_ICM_TYPE_STE) + dr_icm_buddy_cleanup_ste_cache(buddy); + kvfree(buddy); } @@ -261,26 +308,18 @@ dr_icm_chunk_create(struct mlx5dr_icm_pool *pool, chunk->byte_size = mlx5dr_icm_pool_chunk_size_to_byte(chunk_size, pool->icm_type); chunk->seg = seg; + chunk->buddy_mem = buddy_mem_pool; - if (pool->icm_type == DR_ICM_TYPE_STE && dr_icm_chunk_ste_init(chunk)) { - mlx5dr_err(pool->dmn, - "Failed to init ste arrays (order: %d)\n", - chunk_size); - goto out_free_chunk; - } + if (pool->icm_type == DR_ICM_TYPE_STE) + dr_icm_chunk_ste_init(chunk, offset); buddy_mem_pool->used_memory += chunk->byte_size; - chunk->buddy_mem = buddy_mem_pool; INIT_LIST_HEAD(&chunk->chunk_list); /* chunk now is part of the used_list */ list_add_tail(&chunk->chunk_list, &buddy_mem_pool->used_list); return chunk; - -out_free_chunk: - kvfree(chunk); - return NULL; } static bool dr_icm_pool_is_sync_required(struct mlx5dr_icm_pool *pool) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h index c7c93131b762..dfa223415fe2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h @@ -160,6 +160,11 @@ struct mlx5dr_icm_buddy_mem { * sync_ste command sets them free. */ struct list_head hot_list; + + /* Memory optimisation */ + struct mlx5dr_ste *ste_arr; + struct list_head *miss_list; + u8 *hw_ste_arr; }; int mlx5dr_buddy_init(struct mlx5dr_icm_buddy_mem *buddy, -- cgit v1.2.3-70-g09d2 From 0aec12d97b2036af0946e3d582144739860ac07b Mon Sep 17 00:00:00 2001 From: Yevgeny Kliteynik Date: Tue, 11 Jan 2022 03:00:03 +0200 Subject: net/mlx5: DR, Fix slab-out-of-bounds in mlx5_cmd_dr_create_fte When adding a rule with 32 destinations, we hit the following out-of-band access issue: BUG: KASAN: slab-out-of-bounds in mlx5_cmd_dr_create_fte+0x18ee/0x1e70 This patch fixes the issue by both increasing the allocated buffers to accommodate for the needed actions and by checking the number of actions to prevent this issue when a rule with too many actions is provided. Fixes: 1ffd498901c1 ("net/mlx5: DR, Increase supported num of actions to 32") Signed-off-by: Yevgeny Kliteynik Reviewed-by: Alex Vesker Signed-off-by: Saeed Mahameed --- .../ethernet/mellanox/mlx5/core/steering/fs_dr.c | 33 +++++++++++++++++----- 1 file changed, 26 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c index a476da2424f8..3f311462bedf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c @@ -233,7 +233,11 @@ static bool contain_vport_reformat_action(struct mlx5_flow_rule *dst) dst->dest_attr.vport.flags & MLX5_FLOW_DEST_VPORT_REFORMAT_ID; } -#define MLX5_FLOW_CONTEXT_ACTION_MAX 32 +/* We want to support a rule with 32 destinations, which means we need to + * account for 32 destinations plus usually a counter plus one more action + * for a multi-destination flow table. + */ +#define MLX5_FLOW_CONTEXT_ACTION_MAX 34 static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, struct mlx5_flow_table *ft, struct mlx5_flow_group *group, @@ -403,9 +407,9 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, enum mlx5_flow_destination_type type = dst->dest_attr.type; u32 id; - if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX || - num_term_actions >= MLX5_FLOW_CONTEXT_ACTION_MAX) { - err = -ENOSPC; + if (fs_dr_num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX || + num_term_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { + err = -EOPNOTSUPP; goto free_actions; } @@ -478,8 +482,9 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, MLX5_FLOW_DESTINATION_TYPE_COUNTER) continue; - if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { - err = -ENOSPC; + if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX || + fs_dr_num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { + err = -EOPNOTSUPP; goto free_actions; } @@ -499,14 +504,28 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, params.match_sz = match_sz; params.match_buf = (u64 *)fte->val; if (num_term_actions == 1) { - if (term_actions->reformat) + if (term_actions->reformat) { + if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { + err = -EOPNOTSUPP; + goto free_actions; + } actions[num_actions++] = term_actions->reformat; + } + if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { + err = -EOPNOTSUPP; + goto free_actions; + } actions[num_actions++] = term_actions->dest; } else if (num_term_actions > 1) { bool ignore_flow_level = !!(fte->action.flags & FLOW_ACT_IGNORE_FLOW_LEVEL); + if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX || + fs_dr_num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { + err = -EOPNOTSUPP; + goto free_actions; + } tmp_action = mlx5dr_action_create_mult_dest_tbl(domain, term_actions, num_term_actions, -- cgit v1.2.3-70-g09d2 From ffb0753b954763d94f52c901adfe58ed0d4005e6 Mon Sep 17 00:00:00 2001 From: Yevgeny Kliteynik Date: Thu, 13 Jan 2022 14:52:48 +0200 Subject: net/mlx5: DR, Don't allow match on IP w/o matching on full ethertype/ip_version Currently SMFS allows adding rule with matching on src/dst IP w/o matching on full ethertype or ip_version, which is not supported by HW. This patch fixes this issue and adds the check as it is done in DMFS. Fixes: 26d688e33f88 ("net/mlx5: DR, Add Steering entry (STE) utilities") Signed-off-by: Yevgeny Kliteynik Reviewed-by: Alex Vesker Signed-off-by: Saeed Mahameed --- .../mellanox/mlx5/core/steering/dr_matcher.c | 20 +++----------- .../ethernet/mellanox/mlx5/core/steering/dr_ste.c | 32 +++++++++++++++++++++- .../mellanox/mlx5/core/steering/dr_types.h | 10 +++++++ 3 files changed, 45 insertions(+), 17 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c index e87cf498c77b..38971fe1dfe1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c @@ -13,18 +13,6 @@ static bool dr_mask_is_dmac_set(struct mlx5dr_match_spec *spec) return (spec->dmac_47_16 || spec->dmac_15_0); } -static bool dr_mask_is_src_addr_set(struct mlx5dr_match_spec *spec) -{ - return (spec->src_ip_127_96 || spec->src_ip_95_64 || - spec->src_ip_63_32 || spec->src_ip_31_0); -} - -static bool dr_mask_is_dst_addr_set(struct mlx5dr_match_spec *spec) -{ - return (spec->dst_ip_127_96 || spec->dst_ip_95_64 || - spec->dst_ip_63_32 || spec->dst_ip_31_0); -} - static bool dr_mask_is_l3_base_set(struct mlx5dr_match_spec *spec) { return (spec->ip_protocol || spec->frag || spec->tcp_flags || @@ -503,11 +491,11 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher, &mask, inner, rx); if (outer_ipv == DR_RULE_IPV6) { - if (dr_mask_is_dst_addr_set(&mask.outer)) + if (DR_MASK_IS_DST_IP_SET(&mask.outer)) mlx5dr_ste_build_eth_l3_ipv6_dst(ste_ctx, &sb[idx++], &mask, inner, rx); - if (dr_mask_is_src_addr_set(&mask.outer)) + if (DR_MASK_IS_SRC_IP_SET(&mask.outer)) mlx5dr_ste_build_eth_l3_ipv6_src(ste_ctx, &sb[idx++], &mask, inner, rx); @@ -610,11 +598,11 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher, &mask, inner, rx); if (inner_ipv == DR_RULE_IPV6) { - if (dr_mask_is_dst_addr_set(&mask.inner)) + if (DR_MASK_IS_DST_IP_SET(&mask.inner)) mlx5dr_ste_build_eth_l3_ipv6_dst(ste_ctx, &sb[idx++], &mask, inner, rx); - if (dr_mask_is_src_addr_set(&mask.inner)) + if (DR_MASK_IS_SRC_IP_SET(&mask.inner)) mlx5dr_ste_build_eth_l3_ipv6_src(ste_ctx, &sb[idx++], &mask, inner, rx); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c index 7e61742e58a0..187e29b409b6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c @@ -602,12 +602,34 @@ int mlx5dr_ste_set_action_decap_l3_list(struct mlx5dr_ste_ctx *ste_ctx, used_hw_action_num); } +static int dr_ste_build_pre_check_spec(struct mlx5dr_domain *dmn, + struct mlx5dr_match_spec *spec) +{ + if (spec->ip_version) { + if (spec->ip_version != 0xf) { + mlx5dr_err(dmn, + "Partial ip_version mask with src/dst IP is not supported\n"); + return -EINVAL; + } + } else if (spec->ethertype != 0xffff && + (DR_MASK_IS_SRC_IP_SET(spec) || DR_MASK_IS_DST_IP_SET(spec))) { + mlx5dr_err(dmn, + "Partial/no ethertype mask with src/dst IP is not supported\n"); + return -EINVAL; + } + + return 0; +} + int mlx5dr_ste_build_pre_check(struct mlx5dr_domain *dmn, u8 match_criteria, struct mlx5dr_match_param *mask, struct mlx5dr_match_param *value) { - if (!value && (match_criteria & DR_MATCHER_CRITERIA_MISC)) { + if (value) + return 0; + + if (match_criteria & DR_MATCHER_CRITERIA_MISC) { if (mask->misc.source_port && mask->misc.source_port != 0xffff) { mlx5dr_err(dmn, "Partial mask source_port is not supported\n"); @@ -621,6 +643,14 @@ int mlx5dr_ste_build_pre_check(struct mlx5dr_domain *dmn, } } + if ((match_criteria & DR_MATCHER_CRITERIA_OUTER) && + dr_ste_build_pre_check_spec(dmn, &mask->outer)) + return -EINVAL; + + if ((match_criteria & DR_MATCHER_CRITERIA_INNER) && + dr_ste_build_pre_check_spec(dmn, &mask->inner)) + return -EINVAL; + return 0; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h index 1b3d484b99be..55fcb751e24a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h @@ -798,6 +798,16 @@ struct mlx5dr_match_param { (_misc3)->icmpv4_code || \ (_misc3)->icmpv4_header_data) +#define DR_MASK_IS_SRC_IP_SET(_spec) ((_spec)->src_ip_127_96 || \ + (_spec)->src_ip_95_64 || \ + (_spec)->src_ip_63_32 || \ + (_spec)->src_ip_31_0) + +#define DR_MASK_IS_DST_IP_SET(_spec) ((_spec)->dst_ip_127_96 || \ + (_spec)->dst_ip_95_64 || \ + (_spec)->dst_ip_63_32 || \ + (_spec)->dst_ip_31_0) + struct mlx5dr_esw_caps { u64 drop_icm_address_rx; u64 drop_icm_address_tx; -- cgit v1.2.3-70-g09d2 From ecd9c5cd46e013659e2fad433057bad1ba66888e Mon Sep 17 00:00:00 2001 From: Yevgeny Kliteynik Date: Wed, 29 Dec 2021 22:22:05 +0200 Subject: net/mlx5: DR, Fix the threshold that defines when pool sync is initiated When deciding whether to start syncing and actually free all the "hot" ICM chunks, we need to consider the type of the ICM chunks that we're dealing with. For instance, the amount of available ICM for MODIFY_ACTION is significantly lower than the usual STE ICM, so the threshold should account for that - otherwise we can deplete MODIFY_ACTION memory just by creating and deleting the same modify header action in a continuous loop. This patch replaces the hard-coded threshold with a dynamic value. Fixes: 1c58651412bb ("net/mlx5: DR, ICM memory pools sync optimization") Signed-off-by: Yevgeny Kliteynik Reviewed-by: Alex Vesker Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c index f496b7e9401b..e289cfdbce07 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c @@ -4,7 +4,6 @@ #include "dr_types.h" #define DR_ICM_MODIFY_HDR_ALIGN_BASE 64 -#define DR_ICM_SYNC_THRESHOLD_POOL (64 * 1024 * 1024) struct mlx5dr_icm_pool { enum mlx5dr_icm_type icm_type; @@ -324,10 +323,14 @@ dr_icm_chunk_create(struct mlx5dr_icm_pool *pool, static bool dr_icm_pool_is_sync_required(struct mlx5dr_icm_pool *pool) { - if (pool->hot_memory_size > DR_ICM_SYNC_THRESHOLD_POOL) - return true; + int allow_hot_size; - return false; + /* sync when hot memory reaches half of the pool size */ + allow_hot_size = + mlx5dr_icm_pool_chunk_size_to_byte(pool->max_log_chunk_sz, + pool->icm_type) / 2; + + return pool->hot_memory_size > allow_hot_size; } static int dr_icm_pool_sync_all_buddy_pools(struct mlx5dr_icm_pool *pool) -- cgit v1.2.3-70-g09d2 From 7f839965b2d77e1926ad08b23c51d60988f10a99 Mon Sep 17 00:00:00 2001 From: Maher Sanalla Date: Wed, 16 Feb 2022 11:01:04 +0200 Subject: net/mlx5: Update log_max_qp value to be 17 at most Currently, log_max_qp value is dependent on what FW reports as its max capability. In reality, due to a bug, some FWs report a value greater than 17, even though they don't support log_max_qp > 17. This FW issue led the driver to exhaust memory on startup. Thus, log_max_qp value is set to be no more than 17 regardless of what FW reports, as it was before the cited commit. Fixes: f79a609ea6bf ("net/mlx5: Update log_max_qp value to FW max capability") Signed-off-by: Maher Sanalla Reviewed-by: Avihai Horon Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 13f913c13a2d..bba72b220cc3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -526,7 +526,7 @@ static int handle_hca_cap(struct mlx5_core_dev *dev, void *set_ctx) /* Check log_max_qp from HCA caps to set in current profile */ if (prof->log_max_qp == LOG_MAX_SUPPORTED_QPS) { - prof->log_max_qp = MLX5_CAP_GEN_MAX(dev, log_max_qp); + prof->log_max_qp = min_t(u8, 17, MLX5_CAP_GEN_MAX(dev, log_max_qp)); } else if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < prof->log_max_qp) { mlx5_core_warn(dev, "log_max_qp value in current profile is %d, changing it to HCA capability limit (%d)\n", prof->log_max_qp, -- cgit v1.2.3-70-g09d2 From 07666c75ad17d7389b18ac0235c8cf41e1504ea8 Mon Sep 17 00:00:00 2001 From: Ariel Levkovich Date: Sat, 29 Jan 2022 01:39:24 +0200 Subject: net/mlx5: Fix wrong limitation of metadata match on ecpf Match metadata support check returns false for ecpf device. However, this support does exist for ecpf and therefore this limitation should be removed to allow feature such as stacked devices and internal port offloaded to be supported. Fixes: 92ab1eb392c6 ("net/mlx5: E-Switch, Enable vport metadata matching if firmware supports it") Signed-off-by: Ariel Levkovich Reviewed-by: Maor Dickman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 9a7b25692505..cfcd72bad9af 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -2838,10 +2838,6 @@ bool mlx5_esw_vport_match_metadata_supported(const struct mlx5_eswitch *esw) if (!MLX5_CAP_ESW_FLOWTABLE(esw->dev, flow_source)) return false; - if (mlx5_core_is_ecpf_esw_manager(esw->dev) || - mlx5_ecpf_vport_exists(esw->dev)) - return false; - return true; } -- cgit v1.2.3-70-g09d2 From be7f4b0ab149afd19514929fad824b2117d238c9 Mon Sep 17 00:00:00 2001 From: Chris Mi Date: Tue, 14 Dec 2021 03:52:53 +0200 Subject: net/mlx5: Fix tc max supported prio for nic mode Only prio 1 is supported if firmware doesn't support ignore flow level for nic mode. The offending commit removed the check wrongly. Add it back. Fixes: 9a99c8f1253a ("net/mlx5e: E-Switch, Offload all chain 0 priorities when modify header and forward action is not supported") Signed-off-by: Chris Mi Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c index df58cba37930..1e8ec4f236b2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c @@ -121,6 +121,9 @@ u32 mlx5_chains_get_nf_ft_chain(struct mlx5_fs_chains *chains) u32 mlx5_chains_get_prio_range(struct mlx5_fs_chains *chains) { + if (!mlx5_chains_prios_supported(chains)) + return 1; + if (mlx5_chains_ignore_flow_level_supported(chains)) return UINT_MAX; -- cgit v1.2.3-70-g09d2 From b645e57debca846f51b3209907546ea857ddd3f5 Mon Sep 17 00:00:00 2001 From: Maor Gottlieb Date: Mon, 24 Jan 2022 21:25:04 +0200 Subject: net/mlx5: Fix possible deadlock on rule deletion Add missing call to up_write_ref_node() which releases the semaphore in case the FTE doesn't have destinations, such in drop rule case. Fixes: 465e7baab6d9 ("net/mlx5: Fix deletion of duplicate rules") Signed-off-by: Maor Gottlieb Reviewed-by: Mark Bloch Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index b628917e38e4..537c82b9aa53 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -2074,6 +2074,8 @@ void mlx5_del_flow_rules(struct mlx5_flow_handle *handle) fte->node.del_hw_func = NULL; up_write_ref_node(&fte->node, false); tree_put_node(&fte->node, false); + } else { + up_write_ref_node(&fte->node, false); } kfree(handle); } -- cgit v1.2.3-70-g09d2 From 0b89429722353d112f8b8b29ca397e95fa994d27 Mon Sep 17 00:00:00 2001 From: Gal Pressman Date: Wed, 2 Feb 2022 16:07:21 +0200 Subject: net/mlx5e: Fix wrong return value on ioctl EEPROM query failure The ioctl EEPROM query wrongly returns success on read failures, fix that by returning the appropriate error code. Fixes: bb64143eee8c ("net/mlx5e: Add ethtool support for dump module EEPROM") Signed-off-by: Gal Pressman Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index 57d755db1cf5..6e80585d731f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -1792,7 +1792,7 @@ static int mlx5e_get_module_eeprom(struct net_device *netdev, if (size_read < 0) { netdev_err(priv->netdev, "%s: mlx5_query_eeprom failed:0x%x\n", __func__, size_read); - return 0; + return size_read; } i += size_read; -- cgit v1.2.3-70-g09d2 From 7eaf1f37b8817c608c4e959d69986ef459d345cd Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Mon, 31 Jan 2022 10:26:19 +0200 Subject: net/mlx5e: kTLS, Use CHECKSUM_UNNECESSARY for device-offloaded packets For RX TLS device-offloaded packets, the HW spec guarantees checksum validation for the offloaded packets, but does not define whether the CQE.checksum field matches the original packet (ciphertext) or the decrypted one (plaintext). This latitude allows architetctural improvements between generations of chips, resulting in different decisions regarding the value type of CQE.checksum. Hence, for these packets, the device driver should not make use of this CQE field. Here we block CHECKSUM_COMPLETE usage for RX TLS device-offloaded packets, and use CHECKSUM_UNNECESSARY instead. Value of the packet's tcp_hdr.csum is not modified by the HW, and it always matches the original ciphertext. Fixes: 1182f3659357 ("net/mlx5e: kTLS, Add kTLS RX HW offload support") Signed-off-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index ee0a8f5206e3..6530d7bd5045 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -1349,7 +1349,8 @@ static inline void mlx5e_handle_csum(struct net_device *netdev, } /* True when explicitly set via priv flag, or XDP prog is loaded */ - if (test_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &rq->state)) + if (test_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &rq->state) || + get_cqe_tls_offload(cqe)) goto csum_unnecessary; /* CQE csum doesn't cover padding octets in short ethernet -- cgit v1.2.3-70-g09d2 From 23216d387c40b090b221ad457c95912fb47eb11e Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Tue, 4 Jan 2022 10:38:02 +0200 Subject: net/mlx5e: TC, Reject rules with drop and modify hdr action This kind of action is not supported by firmware and generates a syndrome. kernel: mlx5_core 0000:08:00.0: mlx5_cmd_check:777:(pid 102063): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x8708c3) Fixes: d7e75a325cb2 ("net/mlx5e: Add offloading of E-Switch TC pedit (header re-write) actions") Signed-off-by: Roi Dayan Reviewed-by: Maor Dickman Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 2022fa4a9598..34700cf1285e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -3204,6 +3204,12 @@ actions_match_supported(struct mlx5e_priv *priv, return false; } + if (actions & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR && + actions & MLX5_FLOW_CONTEXT_ACTION_DROP) { + NL_SET_ERR_MSG_MOD(extack, "Drop with modify header action is not supported"); + return false; + } + if (actions & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR && !modify_header_match_supported(priv, &parse_attr->spec, flow_action, actions, ct_flow, ct_clear, extack)) -- cgit v1.2.3-70-g09d2 From 3d65492a86d4e6675734646929759138a023d914 Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Mon, 17 Jan 2022 15:00:30 +0200 Subject: net/mlx5e: TC, Reject rules with forward and drop actions Such rules are redundant but allowed and passed to the driver. The driver does not support offloading such rules so return an error. Fixes: 03a9d11e6eeb ("net/mlx5e: Add TC drop and mirred/redirect action parsing for SRIOV offloads") Signed-off-by: Roi Dayan Reviewed-by: Oz Shlomo Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 34700cf1285e..b27532a9301e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -3204,6 +3204,12 @@ actions_match_supported(struct mlx5e_priv *priv, return false; } + if (!(~actions & + (MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_DROP))) { + NL_SET_ERR_MSG_MOD(extack, "Rule cannot support forward+drop action"); + return false; + } + if (actions & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR && actions & MLX5_FLOW_CONTEXT_ACTION_DROP) { NL_SET_ERR_MSG_MOD(extack, "Drop with modify header action is not supported"); -- cgit v1.2.3-70-g09d2 From fb7e76ea3f3b6238dda2f19a4212052d2caf00aa Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Thu, 3 Feb 2022 09:42:19 +0200 Subject: net/mlx5e: TC, Skip redundant ct clear actions Offload of ct clear action is just resetting the reg_c register. It's done by allocating modify hdr resources which is limited. Doing it multiple times is redundant and wasting modify hdr resources and if resources depleted the driver will fail offloading the rule. Ignore redundant ct clear actions after the first one. Fixes: 806401c20a0f ("net/mlx5e: CT, Fix multiple allocations and memleak of mod acts") Signed-off-by: Roi Dayan Reviewed-by: Ariel Levkovich Reviewed-by: Maor Dickman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c | 7 +++++++ 2 files changed, 8 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h index 26efa33de56f..10a40487d536 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h @@ -16,6 +16,7 @@ struct mlx5e_tc_act_parse_state { unsigned int num_actions; struct mlx5e_tc_flow *flow; struct netlink_ext_ack *extack; + bool ct_clear; bool encap; bool decap; bool mpls_push; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c index 06ec30cdb269..58cc33f1363d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c @@ -27,8 +27,13 @@ tc_act_parse_ct(struct mlx5e_tc_act_parse_state *parse_state, struct mlx5e_priv *priv, struct mlx5_flow_attr *attr) { + bool clear_action = act->ct.action & TCA_CT_ACT_CLEAR; int err; + /* It's redundant to do ct clear more than once. */ + if (clear_action && parse_state->ct_clear) + return 0; + err = mlx5_tc_ct_parse_action(parse_state->ct_priv, attr, &attr->parse_attr->mod_hdr_acts, act, parse_state->extack); @@ -40,6 +45,8 @@ tc_act_parse_ct(struct mlx5e_tc_act_parse_state *parse_state, if (mlx5e_is_eswitch_flow(parse_state->flow)) attr->esw_attr->split_count = attr->esw_attr->out_count; + parse_state->ct_clear = clear_action; + return 0; } -- cgit v1.2.3-70-g09d2 From 7fac0529038021919ef56a9c3218d8012f187cbb Mon Sep 17 00:00:00 2001 From: Lama Kayal Date: Tue, 1 Feb 2022 11:24:41 +0200 Subject: net/mlx5e: Add feature check for set fec counters Fec counters support is checked via the PCAM feature_cap_mask, bit 0: PPCNT_counter_group_Phy_statistical_counter_group. Add feature check to avoid faulty behavior. Fixes: 0a1498ebfa55 ("net/mlx5e: Expose FEC counters via ethtool") Signed-off-by: Lama Kayal Reviewed-by: Gal Pressman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 26e326fe503c..00f1d16db456 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -1254,9 +1254,6 @@ static void fec_set_corrected_bits_total(struct mlx5e_priv *priv, u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {}; int sz = MLX5_ST_SZ_BYTES(ppcnt_reg); - if (!MLX5_CAP_PCAM_FEATURE(mdev, ppcnt_statistical_group)) - return; - MLX5_SET(ppcnt_reg, in, local_port, 1); MLX5_SET(ppcnt_reg, in, grp, MLX5_PHYSICAL_LAYER_STATISTICAL_GROUP); if (mlx5_core_access_reg(mdev, in, sz, ppcnt_phy_statistical, @@ -1272,6 +1269,9 @@ static void fec_set_corrected_bits_total(struct mlx5e_priv *priv, void mlx5e_stats_fec_get(struct mlx5e_priv *priv, struct ethtool_fec_stats *fec_stats) { + if (!MLX5_CAP_PCAM_FEATURE(priv->mdev, ppcnt_statistical_group)) + return; + fec_set_corrected_bits_total(priv, fec_stats); fec_set_block_stats(priv, fec_stats); } -- cgit v1.2.3-70-g09d2 From c63741b426e11062631b013c3396f5452bbc0034 Mon Sep 17 00:00:00 2001 From: Maor Dickman Date: Thu, 6 Jan 2022 14:10:18 +0200 Subject: net/mlx5e: Fix MPLSoUDP encap to use MPLS action information Currently the MPLSoUDP encap builds the MPLS header using encap action information (tunnel id, ttl and tos) instead of the MPLS action information (label, ttl, tc and bos) which is wrong. Fix by storing the MPLS action information during the flow action parse and later using it to create the encap MPLS header. Fixes: f828ca6a2fb6 ("net/mlx5e: Add support for hw encapsulation of MPLS over UDP") Signed-off-by: Maor Dickman Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mirred.c | 6 ++++++ drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mpls.c | 11 +++++++++++ drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c | 3 +++ drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c | 5 ++--- drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 8 ++++++++ 7 files changed, 32 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h index 10a40487d536..9cc844bd00f5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h @@ -22,6 +22,7 @@ struct mlx5e_tc_act_parse_state { bool mpls_push; bool ptype_host; const struct ip_tunnel_info *tun_info; + struct mlx5e_mpls_info mpls_info; struct pedit_headers_action hdrs[__PEDIT_CMD_MAX]; int ifindexes[MLX5_MAX_FLOW_FWD_VPORTS]; int if_count; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mirred.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mirred.c index c614fc7fdc9c..2e615e0ba972 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mirred.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mirred.c @@ -177,6 +177,12 @@ parse_mirred_encap(struct mlx5e_tc_act_parse_state *parse_state, return -ENOMEM; parse_state->encap = false; + + if (parse_state->mpls_push) { + memcpy(&parse_attr->mpls_info[esw_attr->out_count], + &parse_state->mpls_info, sizeof(parse_state->mpls_info)); + parse_state->mpls_push = false; + } esw_attr->dests[esw_attr->out_count].flags |= MLX5_ESW_DEST_ENCAP; esw_attr->out_count++; /* attr->dests[].rep is resolved when we handle encap */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mpls.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mpls.c index 784fc4f68b1e..89ca88c78840 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mpls.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mpls.c @@ -22,6 +22,16 @@ tc_act_can_offload_mpls_push(struct mlx5e_tc_act_parse_state *parse_state, return true; } +static void +copy_mpls_info(struct mlx5e_mpls_info *mpls_info, + const struct flow_action_entry *act) +{ + mpls_info->label = act->mpls_push.label; + mpls_info->tc = act->mpls_push.tc; + mpls_info->bos = act->mpls_push.bos; + mpls_info->ttl = act->mpls_push.ttl; +} + static int tc_act_parse_mpls_push(struct mlx5e_tc_act_parse_state *parse_state, const struct flow_action_entry *act, @@ -29,6 +39,7 @@ tc_act_parse_mpls_push(struct mlx5e_tc_act_parse_state *parse_state, struct mlx5_flow_attr *attr) { parse_state->mpls_push = true; + copy_mpls_info(&parse_state->mpls_info, act); return 0; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h index f832c26ff2c3..70b40ae384e4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h @@ -35,6 +35,7 @@ enum { struct mlx5e_tc_flow_parse_attr { const struct ip_tunnel_info *tun_info[MLX5_MAX_FLOW_FWD_VPORTS]; + struct mlx5e_mpls_info mpls_info[MLX5_MAX_FLOW_FWD_VPORTS]; struct net_device *filter_dev; struct mlx5_flow_spec spec; struct mlx5e_tc_mod_hdr_acts mod_hdr_acts; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c index 9918ed8c059b..d39d0dae22fc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c @@ -750,6 +750,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv, struct mlx5e_tc_flow_parse_attr *parse_attr; struct mlx5_flow_attr *attr = flow->attr; const struct ip_tunnel_info *tun_info; + const struct mlx5e_mpls_info *mpls_info; unsigned long tbl_time_before = 0; struct mlx5e_encap_entry *e; struct mlx5e_encap_key key; @@ -760,6 +761,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv, parse_attr = attr->parse_attr; tun_info = parse_attr->tun_info[out_index]; + mpls_info = &parse_attr->mpls_info[out_index]; family = ip_tunnel_info_af(tun_info); key.ip_tun_key = &tun_info->key; key.tc_tunnel = mlx5e_get_tc_tun(mirred_dev); @@ -810,6 +812,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv, goto out_err_init; } e->tun_info = tun_info; + memcpy(&e->mpls_info, mpls_info, sizeof(*mpls_info)); err = mlx5e_tc_tun_init_encap_attr(mirred_dev, priv, e, extack); if (err) goto out_err_init; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c index 60952b33b568..f40dbfcb6437 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c @@ -30,16 +30,15 @@ static int generate_ip_tun_hdr(char buf[], struct mlx5e_encap_entry *r) { const struct ip_tunnel_key *tun_key = &r->tun_info->key; + const struct mlx5e_mpls_info *mpls_info = &r->mpls_info; struct udphdr *udp = (struct udphdr *)(buf); struct mpls_shim_hdr *mpls; - u32 tun_id; - tun_id = be32_to_cpu(tunnel_id_to_key32(tun_key->tun_id)); mpls = (struct mpls_shim_hdr *)(udp + 1); *ip_proto = IPPROTO_UDP; udp->dest = tun_key->tp_dst; - *mpls = mpls_entry_encode(tun_id, tun_key->ttl, tun_key->tos, true); + *mpls = mpls_entry_encode(mpls_info->label, mpls_info->ttl, mpls_info->tc, mpls_info->bos); return 0; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h index b01dacb6f527..b3f7520dfd08 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h @@ -183,6 +183,13 @@ struct mlx5e_decap_entry { struct rcu_head rcu; }; +struct mlx5e_mpls_info { + u32 label; + u8 tc; + u8 bos; + u8 ttl; +}; + struct mlx5e_encap_entry { /* attached neigh hash entry */ struct mlx5e_neigh_hash_entry *nhe; @@ -196,6 +203,7 @@ struct mlx5e_encap_entry { struct list_head route_list; struct mlx5_pkt_reformat *pkt_reformat; const struct ip_tunnel_info *tun_info; + struct mlx5e_mpls_info mpls_info; unsigned char h_dest[ETH_ALEN]; /* destination eth addr */ struct net_device *out_dev; -- cgit v1.2.3-70-g09d2 From fdc18e4e4bded2a08638cdcd22dc087a64b9ddad Mon Sep 17 00:00:00 2001 From: Maor Dickman Date: Thu, 6 Jan 2022 14:46:24 +0200 Subject: net/mlx5e: MPLSoUDP decap, fix check for unsupported matches Currently offload of rule on bareudp device require tunnel key in order to match on mpls fields and without it the mpls fields are ignored, this is incorrect due to the fact udp tunnel doesn't have key to match on. Fix by returning error in case flow is matching on tunnel key. Fixes: 72046a91d134 ("net/mlx5e: Allow to match on mpls parameters") Signed-off-by: Maor Dickman Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../mellanox/mlx5/core/en/tc_tun_mplsoudp.c | 28 +++++++++------------- 1 file changed, 11 insertions(+), 17 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c index f40dbfcb6437..c5b1617d556f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_mplsoudp.c @@ -59,37 +59,31 @@ static int parse_tunnel(struct mlx5e_priv *priv, void *headers_v) { struct flow_rule *rule = flow_cls_offload_flow_rule(f); - struct flow_match_enc_keyid enc_keyid; struct flow_match_mpls match; void *misc2_c; void *misc2_v; - misc2_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, - misc_parameters_2); - misc2_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, - misc_parameters_2); - - if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_MPLS)) - return 0; - - if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) - return 0; - - flow_rule_match_enc_keyid(rule, &enc_keyid); - - if (!enc_keyid.mask->keyid) - return 0; - if (!MLX5_CAP_ETH(priv->mdev, tunnel_stateless_mpls_over_udp) && !(MLX5_CAP_GEN(priv->mdev, flex_parser_protocols) & MLX5_FLEX_PROTO_CW_MPLS_UDP)) return -EOPNOTSUPP; + if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) + return -EOPNOTSUPP; + + if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_MPLS)) + return 0; + flow_rule_match_mpls(rule, &match); /* Only support matching the first LSE */ if (match.mask->used_lses != 1) return -EOPNOTSUPP; + misc2_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, + misc_parameters_2); + misc2_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, + misc_parameters_2); + MLX5_SET(fte_match_set_misc2, misc2_c, outer_first_mpls_over_udp.mpls_label, match.mask->ls[0].mpls_label); -- cgit v1.2.3-70-g09d2 From 5ee02b7a800654ff9549807bcf0b4c9fd5cf25f9 Mon Sep 17 00:00:00 2001 From: Lama Kayal Date: Mon, 21 Feb 2022 12:26:11 +0200 Subject: net/mlx5e: Add missing increment of count Add mistakenly missing increment of count variable when looping over output buffer in mlx5e_self_test(). This resolves the issue of garbage values output when querying with self test via ethtool. before: $ ethtool -t eth2 The test result is PASS The test extra info: Link Test 0 Speed Test 1768697188 Health Test 758528120 Loopback Test 3288687 after: $ ethtool -t eth2 The test result is PASS The test extra info: Link Test 0 Speed Test 0 Health Test 0 Loopback Test 0 Fixes: 7990b1b5e8bd ("net/mlx5e: loopback test is not supported in switchdev mode") Signed-off-by: Lama Kayal Reviewed-by: Gal Pressman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c b/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c index 8c9163d2c646..08a75654f5f1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c @@ -334,6 +334,7 @@ void mlx5e_self_test(struct net_device *ndev, struct ethtool_test *etest, netdev_info(ndev, "\t[%d] %s start..\n", i, st.name); buf[count] = st.st_func(priv); netdev_info(ndev, "\t[%d] %s end: result(%lld)\n", i, st.name, buf[count]); + count++; } mutex_unlock(&priv->state_lock); -- cgit v1.2.3-70-g09d2 From ca49df96f9f5efd4f0f1e64f7c4c0c63a3329cb9 Mon Sep 17 00:00:00 2001 From: Gal Pressman Date: Mon, 21 Feb 2022 17:54:34 +0200 Subject: net/mlx5e: Fix VF min/max rate parameters interchange mistake The VF min and max rate were passed incorrectly and resulted in wrongly interchanging them. Fix the order of parameters in mlx5_esw_qos_set_vport_rate(). Fixes: d7df09f5e7b4 ("net/mlx5: E-switch, Enable vport QoS on demand") Signed-off-by: Gal Pressman Reviewed-by: Aya Levin Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c index 11bbcd5f5b8b..694c54066955 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c @@ -697,7 +697,7 @@ void mlx5_esw_qos_vport_disable(struct mlx5_eswitch *esw, struct mlx5_vport *vpo } int mlx5_esw_qos_set_vport_rate(struct mlx5_eswitch *esw, struct mlx5_vport *vport, - u32 min_rate, u32 max_rate) + u32 max_rate, u32 min_rate) { int err; -- cgit v1.2.3-70-g09d2 From e13ad1443684f7afaff24cf207e85e97885256bd Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Wed, 23 Feb 2022 00:57:20 -0800 Subject: bnx2x: fix driver load from initrd Commit b7a49f73059f ("bnx2x: Utilize firmware 7.13.21.0") added new firmware support in the driver with maintaining older firmware compatibility. However, older firmware was not added in MODULE_FIRMWARE() which caused missing firmware files in initrd image leading to driver load failure from initrd. This patch adds MODULE_FIRMWARE() for older firmware version to have firmware files included in initrd. Fixes: b7a49f73059f ("bnx2x: Utilize firmware 7.13.21.0") Link: https://bugzilla.kernel.org/show_bug.cgi?id=215627 Signed-off-by: Manish Chopra Signed-off-by: Alok Prasad Signed-off-by: Ariel Elior Link: https://lore.kernel.org/r/20220223085720.12021-1-manishc@marvell.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c index 774c1f1a57c3..eedb48d945ed 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c @@ -100,6 +100,9 @@ MODULE_LICENSE("GPL"); MODULE_FIRMWARE(FW_FILE_NAME_E1); MODULE_FIRMWARE(FW_FILE_NAME_E1H); MODULE_FIRMWARE(FW_FILE_NAME_E2); +MODULE_FIRMWARE(FW_FILE_NAME_E1_V15); +MODULE_FIRMWARE(FW_FILE_NAME_E1H_V15); +MODULE_FIRMWARE(FW_FILE_NAME_E2_V15); int bnx2x_num_queues; module_param_named(num_queues, bnx2x_num_queues, int, 0444); -- cgit v1.2.3-70-g09d2 From fe20371578ef640069e6ae9fa8038f60e7908565 Mon Sep 17 00:00:00 2001 From: Mateusz Palczewski Date: Wed, 23 Feb 2022 09:53:47 -0800 Subject: Revert "i40e: Fix reset bw limit when DCB enabled with 1 TC" Revert of a patch that instead of fixing a AQ error when trying to reset BW limit introduced several regressions related to creation and managing TC. Currently there are errors when creating a TC on both PF and VF. Error log: [17428.783095] i40e 0000:3b:00.1: AQ command Config VSI BW allocation per TC failed = 14 [17428.783107] i40e 0000:3b:00.1: Failed configuring TC map 0 for VSI 391 [17428.783254] i40e 0000:3b:00.1: AQ command Config VSI BW allocation per TC failed = 14 [17428.783259] i40e 0000:3b:00.1: Unable to configure TC map 0 for VSI 391 This reverts commit 3d2504663c41104b4359a15f35670cfa82de1bbf. Fixes: 3d2504663c41 (i40e: Fix reset bw limit when DCB enabled with 1 TC) Signed-off-by: Mateusz Palczewski Signed-off-by: Tony Nguyen Link: https://lore.kernel.org/r/20220223175347.1690692-1-anthony.l.nguyen@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/i40e/i40e_main.c | 12 +----------- 1 file changed, 1 insertion(+), 11 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index 0c4b7dfb3b35..31b03fe78d3b 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -5372,15 +5372,7 @@ static int i40e_vsi_configure_bw_alloc(struct i40e_vsi *vsi, u8 enabled_tc, /* There is no need to reset BW when mqprio mode is on. */ if (pf->flags & I40E_FLAG_TC_MQPRIO) return 0; - - if (!vsi->mqprio_qopt.qopt.hw) { - if (pf->flags & I40E_FLAG_DCB_ENABLED) - goto skip_reset; - - if (IS_ENABLED(CONFIG_I40E_DCB) && - i40e_dcb_hw_get_num_tc(&pf->hw) == 1) - goto skip_reset; - + if (!vsi->mqprio_qopt.qopt.hw && !(pf->flags & I40E_FLAG_DCB_ENABLED)) { ret = i40e_set_bw_limit(vsi, vsi->seid, 0); if (ret) dev_info(&pf->pdev->dev, @@ -5388,8 +5380,6 @@ static int i40e_vsi_configure_bw_alloc(struct i40e_vsi *vsi, u8 enabled_tc, vsi->seid); return ret; } - -skip_reset: memset(&bw_data, 0, sizeof(bw_data)); bw_data.tc_valid_bits = enabled_tc; for (i = 0; i < I40E_MAX_TRAFFIC_CLASS; i++) -- cgit v1.2.3-70-g09d2 From 42404d8f1c01861b22ccfa1d70f950242720ae57 Mon Sep 17 00:00:00 2001 From: Mauri Sandberg Date: Wed, 23 Feb 2022 16:23:37 +0200 Subject: net: mv643xx_eth: process retval from of_get_mac_address Obtaining a MAC address may be deferred in cases when the MAC is stored in an NVMEM block, for example, and it may not be ready upon the first retrieval attempt and return EPROBE_DEFER. It is also possible that a port that does not rely on NVMEM has been already created when getting the defer request. Thus, also the resources allocated previously must be freed when doing a roll-back. Fixes: 76723bca2802 ("net: mv643xx_eth: add DT parsing support") Signed-off-by: Mauri Sandberg Reviewed-by: Andrew Lunn Link: https://lore.kernel.org/r/20220223142337.41757-1-maukka@ext.kapsi.fi Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/marvell/mv643xx_eth.c | 24 ++++++++++++++---------- 1 file changed, 14 insertions(+), 10 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c index 105247582684..143ca8be5eb5 100644 --- a/drivers/net/ethernet/marvell/mv643xx_eth.c +++ b/drivers/net/ethernet/marvell/mv643xx_eth.c @@ -2704,6 +2704,16 @@ MODULE_DEVICE_TABLE(of, mv643xx_eth_shared_ids); static struct platform_device *port_platdev[3]; +static void mv643xx_eth_shared_of_remove(void) +{ + int n; + + for (n = 0; n < 3; n++) { + platform_device_del(port_platdev[n]); + port_platdev[n] = NULL; + } +} + static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev, struct device_node *pnp) { @@ -2740,7 +2750,9 @@ static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev, return -EINVAL; } - of_get_mac_address(pnp, ppd.mac_addr); + ret = of_get_mac_address(pnp, ppd.mac_addr); + if (ret) + return ret; mv643xx_eth_property(pnp, "tx-queue-size", ppd.tx_queue_size); mv643xx_eth_property(pnp, "tx-sram-addr", ppd.tx_sram_addr); @@ -2804,21 +2816,13 @@ static int mv643xx_eth_shared_of_probe(struct platform_device *pdev) ret = mv643xx_eth_shared_of_add_port(pdev, pnp); if (ret) { of_node_put(pnp); + mv643xx_eth_shared_of_remove(); return ret; } } return 0; } -static void mv643xx_eth_shared_of_remove(void) -{ - int n; - - for (n = 0; n < 3; n++) { - platform_device_del(port_platdev[n]); - port_platdev[n] = NULL; - } -} #else static inline int mv643xx_eth_shared_of_probe(struct platform_device *pdev) { -- cgit v1.2.3-70-g09d2 From 087a7b944c5db409f7c1a68bf4896c56ba54eaff Mon Sep 17 00:00:00 2001 From: Vincent Whitchurch Date: Thu, 24 Feb 2022 12:38:29 +0100 Subject: net: stmmac: only enable DMA interrupts when ready In this driver's ->ndo_open() callback, it enables DMA interrupts, starts the DMA channels, then requests interrupts with request_irq(), and then finally enables napi. If RX DMA interrupts are received before napi is enabled, no processing is done because napi_schedule_prep() will return false. If the network has a lot of broadcast/multicast traffic, then the RX ring could fill up completely before napi is enabled. When this happens, no further RX interrupts will be delivered, and the driver will fail to receive any packets. Fix this by only enabling DMA interrupts after all other initialization is complete. Fixes: 523f11b5d4fd72efb ("net: stmmac: move hardware setup for stmmac_open to new function") Reported-by: Lars Persson Signed-off-by: Vincent Whitchurch Signed-off-by: David S. Miller --- drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 28 +++++++++++++++++++++-- 1 file changed, 26 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index bde76ea2deec..cb9b6e08780c 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -2262,6 +2262,23 @@ static void stmmac_stop_tx_dma(struct stmmac_priv *priv, u32 chan) stmmac_stop_tx(priv, priv->ioaddr, chan); } +static void stmmac_enable_all_dma_irq(struct stmmac_priv *priv) +{ + u32 rx_channels_count = priv->plat->rx_queues_to_use; + u32 tx_channels_count = priv->plat->tx_queues_to_use; + u32 dma_csr_ch = max(rx_channels_count, tx_channels_count); + u32 chan; + + for (chan = 0; chan < dma_csr_ch; chan++) { + struct stmmac_channel *ch = &priv->channel[chan]; + unsigned long flags; + + spin_lock_irqsave(&ch->lock, flags); + stmmac_enable_dma_irq(priv, priv->ioaddr, chan, 1, 1); + spin_unlock_irqrestore(&ch->lock, flags); + } +} + /** * stmmac_start_all_dma - start all RX and TX DMA channels * @priv: driver private structure @@ -2904,8 +2921,10 @@ static int stmmac_init_dma_engine(struct stmmac_priv *priv) stmmac_axi(priv, priv->ioaddr, priv->plat->axi); /* DMA CSR Channel configuration */ - for (chan = 0; chan < dma_csr_ch; chan++) + for (chan = 0; chan < dma_csr_ch; chan++) { stmmac_init_chan(priv, priv->ioaddr, priv->plat->dma_cfg, chan); + stmmac_disable_dma_irq(priv, priv->ioaddr, chan, 1, 1); + } /* DMA RX Channel Configuration */ for (chan = 0; chan < rx_channels_count; chan++) { @@ -3761,6 +3780,7 @@ static int stmmac_open(struct net_device *dev) stmmac_enable_all_queues(priv); netif_tx_start_all_queues(priv->dev); + stmmac_enable_all_dma_irq(priv); return 0; @@ -6510,8 +6530,10 @@ int stmmac_xdp_open(struct net_device *dev) } /* DMA CSR Channel configuration */ - for (chan = 0; chan < dma_csr_ch; chan++) + for (chan = 0; chan < dma_csr_ch; chan++) { stmmac_init_chan(priv, priv->ioaddr, priv->plat->dma_cfg, chan); + stmmac_disable_dma_irq(priv, priv->ioaddr, chan, 1, 1); + } /* Adjust Split header */ sph_en = (priv->hw->rx_csum > 0) && priv->sph; @@ -6572,6 +6594,7 @@ int stmmac_xdp_open(struct net_device *dev) stmmac_enable_all_queues(priv); netif_carrier_on(dev); netif_tx_start_all_queues(dev); + stmmac_enable_all_dma_irq(priv); return 0; @@ -7451,6 +7474,7 @@ int stmmac_resume(struct device *dev) stmmac_restore_hw_vlan_rx_fltr(priv, ndev, priv->hw); stmmac_enable_all_queues(priv); + stmmac_enable_all_dma_irq(priv); mutex_unlock(&priv->lock); rtnl_unlock(); -- cgit v1.2.3-70-g09d2 From 8d0657f39f487d904fca713e0bc39c2707382553 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:51 -0800 Subject: ibmvnic: free reset-work-item when flushing Fix a tiny memory leak when flushing the reset work queue. Fixes: 2770a7984db5 ("ibmvnic: Introduce hard reset recovery") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index dee05a353dbd..86d522171892 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -2784,8 +2784,10 @@ static int ibmvnic_reset(struct ibmvnic_adapter *adapter, * flush reset queue and process this reset */ if (adapter->force_reset_recovery && !list_empty(&adapter->rwi_list)) { - list_for_each_safe(entry, tmp_entry, &adapter->rwi_list) + list_for_each_safe(entry, tmp_entry, &adapter->rwi_list) { list_del(entry); + kfree(list_entry(entry, struct ibmvnic_rwi, list)); + } } rwi->reset_reason = reason; list_add_tail(&rwi->list, &adapter->rwi_list); -- cgit v1.2.3-70-g09d2 From 765559b10ce514eb1576595834f23cdc92125fee Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:52 -0800 Subject: ibmvnic: initialize rc before completing wait We should initialize ->init_done_rc before calling complete(). Otherwise the waiting thread may see ->init_done_rc as 0 before we have updated it and may assume that the CRQ was successful. Fixes: 6b278c0cb378 ("ibmvnic delay complete()") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 86d522171892..c73c699600a8 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -5323,9 +5323,9 @@ static void ibmvnic_handle_crq(union ibmvnic_crq *crq, } if (!completion_done(&adapter->init_done)) { - complete(&adapter->init_done); if (!adapter->init_done_rc) adapter->init_done_rc = -EAGAIN; + complete(&adapter->init_done); } break; -- cgit v1.2.3-70-g09d2 From 83da53f7e4bd86dca4b2edc1e2bb324fb3c033a1 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:53 -0800 Subject: ibmvnic: define flush_reset_queue helper Define and use a helper to flush the reset queue. Fixes: 2770a7984db5 ("ibmvnic: Introduce hard reset recovery") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index c73c699600a8..42c3ac9ebb75 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -2735,12 +2735,23 @@ static void __ibmvnic_delayed_reset(struct work_struct *work) __ibmvnic_reset(&adapter->ibmvnic_reset); } +static void flush_reset_queue(struct ibmvnic_adapter *adapter) +{ + struct list_head *entry, *tmp_entry; + + if (!list_empty(&adapter->rwi_list)) { + list_for_each_safe(entry, tmp_entry, &adapter->rwi_list) { + list_del(entry); + kfree(list_entry(entry, struct ibmvnic_rwi, list)); + } + } +} + static int ibmvnic_reset(struct ibmvnic_adapter *adapter, enum ibmvnic_reset_reason reason) { - struct list_head *entry, *tmp_entry; - struct ibmvnic_rwi *rwi, *tmp; struct net_device *netdev = adapter->netdev; + struct ibmvnic_rwi *rwi, *tmp; unsigned long flags; int ret; @@ -2783,12 +2794,9 @@ static int ibmvnic_reset(struct ibmvnic_adapter *adapter, /* if we just received a transport event, * flush reset queue and process this reset */ - if (adapter->force_reset_recovery && !list_empty(&adapter->rwi_list)) { - list_for_each_safe(entry, tmp_entry, &adapter->rwi_list) { - list_del(entry); - kfree(list_entry(entry, struct ibmvnic_rwi, list)); - } - } + if (adapter->force_reset_recovery) + flush_reset_queue(adapter); + rwi->reset_reason = reason; list_add_tail(&rwi->list, &adapter->rwi_list); netdev_dbg(adapter->netdev, "Scheduling reset (reason %s)\n", -- cgit v1.2.3-70-g09d2 From 36491f2df9ad2501e5a4ec25d3d95d72bafd2781 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:54 -0800 Subject: ibmvnic: complete init_done on transport events If we get a transport event, set the error and mark the init as complete so the attempt to send crq-init or login fail sooner rather than wait for the timeout. Fixes: bbd669a868bb ("ibmvnic: Fix completion structure initialization") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 42c3ac9ebb75..5913d372bc27 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -5356,6 +5356,13 @@ static void ibmvnic_handle_crq(union ibmvnic_crq *crq, adapter->fw_done_rc = -EIO; complete(&adapter->fw_done); } + + /* if we got here during crq-init, retry crq-init */ + if (!completion_done(&adapter->init_done)) { + adapter->init_done_rc = -EAGAIN; + complete(&adapter->init_done); + } + if (!completion_done(&adapter->stats_done)) complete(&adapter->stats_done); if (test_bit(0, &adapter->resetting)) -- cgit v1.2.3-70-g09d2 From 570425f8c7c18b14fa8a2a58a0adb431968ad118 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:55 -0800 Subject: ibmvnic: register netdev after init of adapter Finish initializing the adapter before registering netdev so state is consistent. Fixes: c26eba03e407 ("ibmvnic: Update reset infrastructure to support tunable parameters") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 5913d372bc27..a7b03ca109d8 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -5826,12 +5826,6 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) goto ibmvnic_dev_file_err; netif_carrier_off(netdev); - rc = register_netdev(netdev); - if (rc) { - dev_err(&dev->dev, "failed to register netdev rc=%d\n", rc); - goto ibmvnic_register_fail; - } - dev_info(&dev->dev, "ibmvnic registered\n"); if (init_success) { adapter->state = VNIC_PROBED; @@ -5844,6 +5838,14 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) adapter->wait_for_reset = false; adapter->last_reset_time = jiffies; + + rc = register_netdev(netdev); + if (rc) { + dev_err(&dev->dev, "failed to register netdev rc=%d\n", rc); + goto ibmvnic_register_fail; + } + dev_info(&dev->dev, "ibmvnic registered\n"); + return 0; ibmvnic_register_fail: -- cgit v1.2.3-70-g09d2 From ae16bf15374d8b055e040ac6f3f1147ab1c9bb7d Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:56 -0800 Subject: ibmvnic: init init_done_rc earlier We currently initialize the ->init_done completion/return code fields before issuing a CRQ_INIT command. But if we get a transport event soon after registering the CRQ the taskslet may already have recorded the completion and error code. If we initialize here, we might overwrite/ lose that and end up issuing the CRQ_INIT only to timeout later. If that timeout happens during probe, we will leave the adapter in the DOWN state rather than retrying to register/init the CRQ. Initialize the completion before registering the CRQ so we don't lose the notification. Fixes: 032c5e82847a ("Driver for IBM System i/p VNIC protocol") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index a7b03ca109d8..16a8f92adbda 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -2212,6 +2212,19 @@ static const char *reset_reason_to_string(enum ibmvnic_reset_reason reason) return "UNKNOWN"; } +/* + * Initialize the init_done completion and return code values. We + * can get a transport event just after registering the CRQ and the + * tasklet will use this to communicate the transport event. To ensure + * we don't miss the notification/error, initialize these _before_ + * regisering the CRQ. + */ +static inline void reinit_init_done(struct ibmvnic_adapter *adapter) +{ + reinit_completion(&adapter->init_done); + adapter->init_done_rc = 0; +} + /* * do_reset returns zero if we are able to keep processing reset events, or * non-zero if we hit a fatal error and must halt. @@ -2318,6 +2331,8 @@ static int do_reset(struct ibmvnic_adapter *adapter, */ adapter->state = VNIC_PROBED; + reinit_init_done(adapter); + if (adapter->reset_reason == VNIC_RESET_CHANGE_PARAM) { rc = init_crq_queue(adapter); } else if (adapter->reset_reason == VNIC_RESET_MOBILITY) { @@ -2461,7 +2476,8 @@ static int do_hard_reset(struct ibmvnic_adapter *adapter, */ adapter->state = VNIC_PROBED; - reinit_completion(&adapter->init_done); + reinit_init_done(adapter); + rc = init_crq_queue(adapter); if (rc) { netdev_err(adapter->netdev, @@ -5679,10 +5695,6 @@ static int ibmvnic_reset_init(struct ibmvnic_adapter *adapter, bool reset) adapter->from_passive_init = false; - if (reset) - reinit_completion(&adapter->init_done); - - adapter->init_done_rc = 0; rc = ibmvnic_send_crq_init(adapter); if (rc) { dev_err(dev, "Send crq init failed with error %d\n", rc); @@ -5696,12 +5708,14 @@ static int ibmvnic_reset_init(struct ibmvnic_adapter *adapter, bool reset) if (adapter->init_done_rc) { release_crq_queue(adapter); + dev_err(dev, "CRQ-init failed, %d\n", adapter->init_done_rc); return adapter->init_done_rc; } if (adapter->from_passive_init) { adapter->state = VNIC_OPEN; adapter->from_passive_init = false; + dev_err(dev, "CRQ-init failed, passive-init\n"); return -EINVAL; } @@ -5795,6 +5809,8 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) init_success = false; do { + reinit_init_done(adapter); + rc = init_crq_queue(adapter); if (rc) { dev_err(&dev->dev, "Couldn't initialize crq. rc=%d\n", -- cgit v1.2.3-70-g09d2 From f628ad531b4f34fdba0984255b4a2850dd369513 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:57 -0800 Subject: ibmvnic: clear fop when retrying probe Clear ->failover_pending flag that may have been set in the previous pass of registering CRQ. If we don't clear, a subsequent ibmvnic_open() call would be misled into thinking a failover is pending and assuming that the reset worker thread would open the adapter. If this pass of registering the CRQ succeeds (i.e there is no transport event), there wouldn't be a reset worker thread. This would leave the adapter unconfigured and require manual intervention to bring it up during boot. Fixes: 5a18e1e0c193 ("ibmvnic: Fix failover case for non-redundant configuration") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 16a8f92adbda..93580c68600a 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -5811,6 +5811,11 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) do { reinit_init_done(adapter); + /* clear any failovers we got in the previous pass + * since we are reinitializing the CRQ + */ + adapter->failover_pending = false; + rc = init_crq_queue(adapter); if (rc) { dev_err(&dev->dev, "Couldn't initialize crq. rc=%d\n", -- cgit v1.2.3-70-g09d2 From fd98693cb0721317f27341951593712c580c36a1 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Thu, 24 Feb 2022 22:23:58 -0800 Subject: ibmvnic: Allow queueing resets during probe We currently don't allow queuing resets when adapter is in VNIC_PROBING state - instead we throw away the reset and return EBUSY. The reasoning is probably that during ibmvnic_probe() the ibmvnic_adapter itself is being initialized so performing a reset during this time can lead us to accessing fields in the ibmvnic_adapter that are not fully initialized. A review of the code shows that all the adapter state neede to process a reset is initialized before registering the CRQ so that should no longer be a concern. Further the expectation is that if we do get a reset (transport event) during probe, the do..while() loop in ibmvnic_probe() will handle this by reinitializing the CRQ. While that is true to some extent, it is possible that the reset might occur _after_ the CRQ is registered and CRQ_INIT message was exchanged but _before_ the adapter state is set to VNIC_PROBED. As mentioned above, such a reset will be thrown away. While the client assumes that the adapter is functional, the vnic server will wait for the client to reinit the adapter. This disconnect between the two leaves the adapter down needing manual intervention. Because ibmvnic_probe() has other work to do after initializing the CRQ (such as registering the netdev at a minimum) and because the reset event can occur at any instant after the CRQ is initialized, there will always be a window between initializing the CRQ and considering the adapter ready for resets (ie state == PROBED). So rather than discarding resets during this window, allow queueing them - but only process them after the adapter is fully initialized. To do this, introduce a new completion state ->probe_done and have the reset worker thread wait on this before processing resets. This change brings up two new situations in or just after ibmvnic_probe(). First after one or more resets were queued, we encounter an error and decide to retry the initialization. At that point the queued resets are no longer relevant since we could be talking to a new vnic server. So we must purge/flush the queued resets before restarting the initialization. As a side note, since we are still in the probing stage and we have not registered the netdev, it will not be CHANGE_PARAM reset. Second this change opens up a potential race between the worker thread in __ibmvnic_reset(), the tasklet and the ibmvnic_open() due to the following sequence of events: 1. Register CRQ 2. Get transport event before CRQ_INIT completes. 3. Tasklet schedules reset: a) add rwi to list b) schedule_work() to start worker thread which runs and waits for ->probe_done. 4. ibmvnic_probe() decides to retry, purges rwi_list 5. Re-register crq and this time rest of probe succeeds - register netdev and complete(->probe_done). 6. Worker thread resumes in __ibmvnic_reset() from 3b. 7. Worker thread sets ->resetting bit 8. ibmvnic_open() comes in, notices ->resetting bit, sets state to IBMVNIC_OPEN and returns early expecting worker thread to finish the open. 9. Worker thread finds rwi_list empty and returns without opening the interface. If this happens, the ->ndo_open() call is effectively lost and the interface remains down. To address this, ensure that ->rwi_list is not empty before setting the ->resetting bit. See also comments in __ibmvnic_reset(). Fixes: 6a2fb0e99f9c ("ibmvnic: driver initialization for kdump/kexec") Signed-off-by: Sukadev Bhattiprolu Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmvnic.c | 107 +++++++++++++++++++++++++++++++++---- drivers/net/ethernet/ibm/ibmvnic.h | 1 + 2 files changed, 98 insertions(+), 10 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c index 93580c68600a..b423e94956f1 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.c +++ b/drivers/net/ethernet/ibm/ibmvnic.c @@ -2618,23 +2618,82 @@ out: static void __ibmvnic_reset(struct work_struct *work) { struct ibmvnic_adapter *adapter; - bool saved_state = false; + unsigned int timeout = 5000; struct ibmvnic_rwi *tmprwi; + bool saved_state = false; struct ibmvnic_rwi *rwi; unsigned long flags; - u32 reset_state; + struct device *dev; + bool need_reset; int num_fails = 0; + u32 reset_state; int rc = 0; adapter = container_of(work, struct ibmvnic_adapter, ibmvnic_reset); + dev = &adapter->vdev->dev; - if (test_and_set_bit_lock(0, &adapter->resetting)) { + /* Wait for ibmvnic_probe() to complete. If probe is taking too long + * or if another reset is in progress, defer work for now. If probe + * eventually fails it will flush and terminate our work. + * + * Three possibilities here: + * 1. Adpater being removed - just return + * 2. Timed out on probe or another reset in progress - delay the work + * 3. Completed probe - perform any resets in queue + */ + if (adapter->state == VNIC_PROBING && + !wait_for_completion_timeout(&adapter->probe_done, timeout)) { + dev_err(dev, "Reset thread timed out on probe"); queue_delayed_work(system_long_wq, &adapter->ibmvnic_delayed_reset, IBMVNIC_RESET_DELAY); return; } + /* adapter is done with probe (i.e state is never VNIC_PROBING now) */ + if (adapter->state == VNIC_REMOVING) + return; + + /* ->rwi_list is stable now (no one else is removing entries) */ + + /* ibmvnic_probe() may have purged the reset queue after we were + * scheduled to process a reset so there maybe no resets to process. + * Before setting the ->resetting bit though, we have to make sure + * that there is infact a reset to process. Otherwise we may race + * with ibmvnic_open() and end up leaving the vnic down: + * + * __ibmvnic_reset() ibmvnic_open() + * ----------------- -------------- + * + * set ->resetting bit + * find ->resetting bit is set + * set ->state to IBMVNIC_OPEN (i.e + * assume reset will open device) + * return + * find reset queue empty + * return + * + * Neither performed vnic login/open and vnic stays down + * + * If we hold the lock and conditionally set the bit, either we + * or ibmvnic_open() will complete the open. + */ + need_reset = false; + spin_lock(&adapter->rwi_lock); + if (!list_empty(&adapter->rwi_list)) { + if (test_and_set_bit_lock(0, &adapter->resetting)) { + queue_delayed_work(system_long_wq, + &adapter->ibmvnic_delayed_reset, + IBMVNIC_RESET_DELAY); + } else { + need_reset = true; + } + } + spin_unlock(&adapter->rwi_lock); + + if (!need_reset) + return; + rwi = get_next_rwi(adapter); while (rwi) { spin_lock_irqsave(&adapter->state_lock, flags); @@ -2786,13 +2845,6 @@ static int ibmvnic_reset(struct ibmvnic_adapter *adapter, goto err; } - if (adapter->state == VNIC_PROBING) { - netdev_warn(netdev, "Adapter reset during probe\n"); - adapter->init_done_rc = -EAGAIN; - ret = EAGAIN; - goto err; - } - list_for_each_entry(tmp, &adapter->rwi_list, list) { if (tmp->reset_reason == reason) { netdev_dbg(netdev, "Skipping matching reset, reason=%s\n", @@ -5755,6 +5807,7 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) struct ibmvnic_adapter *adapter; struct net_device *netdev; unsigned char *mac_addr_p; + unsigned long flags; bool init_success; int rc; @@ -5799,6 +5852,7 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) spin_lock_init(&adapter->rwi_lock); spin_lock_init(&adapter->state_lock); mutex_init(&adapter->fw_lock); + init_completion(&adapter->probe_done); init_completion(&adapter->init_done); init_completion(&adapter->fw_done); init_completion(&adapter->reset_done); @@ -5816,6 +5870,26 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) */ adapter->failover_pending = false; + /* If we had already initialized CRQ, we may have one or + * more resets queued already. Discard those and release + * the CRQ before initializing the CRQ again. + */ + release_crq_queue(adapter); + + /* Since we are still in PROBING state, __ibmvnic_reset() + * will not access the ->rwi_list and since we released CRQ, + * we won't get _new_ transport events. But there maybe an + * ongoing ibmvnic_reset() call. So serialize access to + * rwi_list. If we win the race, ibvmnic_reset() could add + * a reset after we purged but thats ok - we just may end + * up with an extra reset (i.e similar to having two or more + * resets in the queue at once). + * CHECK. + */ + spin_lock_irqsave(&adapter->rwi_lock, flags); + flush_reset_queue(adapter); + spin_unlock_irqrestore(&adapter->rwi_lock, flags); + rc = init_crq_queue(adapter); if (rc) { dev_err(&dev->dev, "Couldn't initialize crq. rc=%d\n", @@ -5867,6 +5941,8 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id) } dev_info(&dev->dev, "ibmvnic registered\n"); + complete(&adapter->probe_done); + return 0; ibmvnic_register_fail: @@ -5881,6 +5957,17 @@ ibmvnic_stats_fail: ibmvnic_init_fail: release_sub_crqs(adapter, 1); release_crq_queue(adapter); + + /* cleanup worker thread after releasing CRQ so we don't get + * transport events (i.e new work items for the worker thread). + */ + adapter->state = VNIC_REMOVING; + complete(&adapter->probe_done); + flush_work(&adapter->ibmvnic_reset); + flush_delayed_work(&adapter->ibmvnic_delayed_reset); + + flush_reset_queue(adapter); + mutex_destroy(&adapter->fw_lock); free_netdev(netdev); diff --git a/drivers/net/ethernet/ibm/ibmvnic.h b/drivers/net/ethernet/ibm/ibmvnic.h index 4a7a56ff74ce..fa2d607a7b1b 100644 --- a/drivers/net/ethernet/ibm/ibmvnic.h +++ b/drivers/net/ethernet/ibm/ibmvnic.h @@ -930,6 +930,7 @@ struct ibmvnic_adapter { struct ibmvnic_tx_pool *tx_pool; struct ibmvnic_tx_pool *tso_pool; + struct completion probe_done; struct completion init_done; int init_done_rc; -- cgit v1.2.3-70-g09d2 From 767b9825ed1765894e569a3d698749d40d83762a Mon Sep 17 00:00:00 2001 From: Jia-Ju Bai Date: Fri, 25 Feb 2022 04:37:27 -0800 Subject: net: chelsio: cxgb3: check the return value of pci_find_capability() The function pci_find_capability() in t3_prep_adapter() can fail, so its return value should be checked. Fixes: 4d22de3e6cc4 ("Add support for the latest 1G/10G Chelsio adapter, T3") Reported-by: TOTE Robot Signed-off-by: Jia-Ju Bai Signed-off-by: David S. Miller --- drivers/net/ethernet/chelsio/cxgb3/t3_hw.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/chelsio/cxgb3/t3_hw.c b/drivers/net/ethernet/chelsio/cxgb3/t3_hw.c index da41eee2f25c..a06003bfa04b 100644 --- a/drivers/net/ethernet/chelsio/cxgb3/t3_hw.c +++ b/drivers/net/ethernet/chelsio/cxgb3/t3_hw.c @@ -3613,6 +3613,8 @@ int t3_prep_adapter(struct adapter *adapter, const struct adapter_info *ai, MAC_STATS_ACCUM_SECS : (MAC_STATS_ACCUM_SECS * 10); adapter->params.pci.vpd_cap_addr = pci_find_capability(adapter->pdev, PCI_CAP_ID_VPD); + if (!adapter->params.pci.vpd_cap_addr) + return -ENODEV; ret = get_vpd_params(adapter, &adapter->params.vpd); if (ret < 0) return ret; -- cgit v1.2.3-70-g09d2 From b3a34dc362c03215031b268fcc0b988e69490231 Mon Sep 17 00:00:00 2001 From: Casper Andersson Date: Fri, 25 Feb 2022 11:15:16 +0100 Subject: net: sparx5: Fix add vlan when invalid operation Check if operation is valid before changing any settings in hardware. Otherwise it results in changes being made despite it not being a valid operation. Fixes: 78eab33bb68b ("net: sparx5: add vlan support") Signed-off-by: Casper Andersson Signed-off-by: David S. Miller --- drivers/net/ethernet/microchip/sparx5/sparx5_vlan.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/microchip/sparx5/sparx5_vlan.c b/drivers/net/ethernet/microchip/sparx5/sparx5_vlan.c index 4ce490a25f33..8e56ffa1c4f7 100644 --- a/drivers/net/ethernet/microchip/sparx5/sparx5_vlan.c +++ b/drivers/net/ethernet/microchip/sparx5/sparx5_vlan.c @@ -58,16 +58,6 @@ int sparx5_vlan_vid_add(struct sparx5_port *port, u16 vid, bool pvid, struct sparx5 *sparx5 = port->sparx5; int ret; - /* Make the port a member of the VLAN */ - set_bit(port->portno, sparx5->vlan_mask[vid]); - ret = sparx5_vlant_set_mask(sparx5, vid); - if (ret) - return ret; - - /* Default ingress vlan classification */ - if (pvid) - port->pvid = vid; - /* Untagged egress vlan classification */ if (untagged && port->vid != vid) { if (port->vid) { @@ -79,6 +69,16 @@ int sparx5_vlan_vid_add(struct sparx5_port *port, u16 vid, bool pvid, port->vid = vid; } + /* Make the port a member of the VLAN */ + set_bit(port->portno, sparx5->vlan_mask[vid]); + ret = sparx5_vlant_set_mask(sparx5, vid); + if (ret) + return ret; + + /* Default ingress vlan classification */ + if (pvid) + port->pvid = vid; + sparx5_vlan_port_apply(sparx5, port); return 0; -- cgit v1.2.3-70-g09d2 From 50e06ddceeea263f57fe92baa677c638ecd65bb6 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 23 Feb 2022 19:35:28 -0800 Subject: net: sxgbe: fix return value of __setup handler __setup() handlers should return 1 on success, i.e., the parameter has been handled. A return of 0 causes the "option=value" string to be added to init's environment strings, polluting it. Fixes: acc18c147b22 ("net: sxgbe: add EEE(Energy Efficient Ethernet) for Samsung sxgbe") Fixes: 1edb9ca69e8a ("net: sxgbe: add basic framework for Samsung 10Gb ethernet driver") Signed-off-by: Randy Dunlap Reported-by: Igor Zhbanov Link: lore.kernel.org/r/64644a2f-4a20-bab3-1e15-3b2cdd0defe3@omprussia.ru Cc: Siva Reddy Cc: Girish K S Cc: Byungho An Link: https://lore.kernel.org/r/20220224033528.24640-1-rdunlap@infradead.org Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c b/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c index 32161a56726c..2881f5b2b5f4 100644 --- a/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c +++ b/drivers/net/ethernet/samsung/sxgbe/sxgbe_main.c @@ -2285,18 +2285,18 @@ static int __init sxgbe_cmdline_opt(char *str) char *opt; if (!str || !*str) - return -EINVAL; + return 1; while ((opt = strsep(&str, ",")) != NULL) { if (!strncmp(opt, "eee_timer:", 10)) { if (kstrtoint(opt + 10, 0, &eee_timer)) goto err; } } - return 0; + return 1; err: pr_err("%s: ERROR broken module parameter conversion\n", __func__); - return -EINVAL; + return 1; } __setup("sxgbeeth=", sxgbe_cmdline_opt); -- cgit v1.2.3-70-g09d2 From e01b042e580f1fbf4fd8da467442451da00c7a90 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 23 Feb 2022 19:35:36 -0800 Subject: net: stmmac: fix return value of __setup handler __setup() handlers should return 1 on success, i.e., the parameter has been handled. A return of 0 causes the "option=value" string to be added to init's environment strings, polluting it. Fixes: 47dd7a540b8a ("net: add support for STMicroelectronics Ethernet controllers.") Fixes: f3240e2811f0 ("stmmac: remove warning when compile as built-in (V2)") Signed-off-by: Randy Dunlap Reported-by: Igor Zhbanov Link: lore.kernel.org/r/64644a2f-4a20-bab3-1e15-3b2cdd0defe3@omprussia.ru Cc: Giuseppe Cavallaro Cc: Alexandre Torgue Cc: Jose Abreu Link: https://lore.kernel.org/r/20220224033536.25056-1-rdunlap@infradead.org Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index cb9b6e08780c..422e3225f476 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -7491,7 +7491,7 @@ static int __init stmmac_cmdline_opt(char *str) char *opt; if (!str || !*str) - return -EINVAL; + return 1; while ((opt = strsep(&str, ",")) != NULL) { if (!strncmp(opt, "debug:", 6)) { if (kstrtoint(opt + 6, 0, &debug)) @@ -7522,11 +7522,11 @@ static int __init stmmac_cmdline_opt(char *str) goto err; } } - return 0; + return 1; err: pr_err("%s: ERROR broken module parameter conversion", __func__); - return -EINVAL; + return 1; } __setup("stmmaceth=", stmmac_cmdline_opt); -- cgit v1.2.3-70-g09d2 From fc2e6b3b132a907378f6af08356b105a4139c4fb Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:35:49 +0100 Subject: iavf: Rework mutexes for better synchronisation The driver used to crash in multiple spots when put to stress testing of the init, reset and remove paths. The user would experience call traces or hangs when creating, resetting, removing VFs. Depending on the machines, the call traces are happening in random spots, like reset restoring resources racing with driver remove. Make adapter->crit_lock mutex a mandatory lock for guarding the operations performed on all workqueues and functions dealing with resource allocation and disposal. Make __IAVF_REMOVE a final state of the driver respected by workqueues that shall not requeue, when they fail to obtain the crit_lock. Make the IRQ handler not to queue the new work for adminq_task when the __IAVF_REMOVE state is set. Fixes: 5ac49f3c2702 ("iavf: use mutexes for locking of critical sections") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf.h | 1 - drivers/net/ethernet/intel/iavf/iavf_main.c | 66 ++++++++++++++++------------- 2 files changed, 36 insertions(+), 31 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf.h b/drivers/net/ethernet/intel/iavf/iavf.h index 59806d1f7e79..44f83e06486d 100644 --- a/drivers/net/ethernet/intel/iavf/iavf.h +++ b/drivers/net/ethernet/intel/iavf/iavf.h @@ -246,7 +246,6 @@ struct iavf_adapter { struct list_head mac_filter_list; struct mutex crit_lock; struct mutex client_lock; - struct mutex remove_lock; /* Lock to protect accesses to MAC and VLAN lists */ spinlock_t mac_vlan_list_lock; char misc_vector_name[IFNAMSIZ + 9]; diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 8125b9120615..84ae96e912d7 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -302,8 +302,9 @@ static irqreturn_t iavf_msix_aq(int irq, void *data) rd32(hw, IAVF_VFINT_ICR01); rd32(hw, IAVF_VFINT_ICR0_ENA1); - /* schedule work on the private workqueue */ - queue_work(iavf_wq, &adapter->adminq_task); + if (adapter->state != __IAVF_REMOVE) + /* schedule work on the private workqueue */ + queue_work(iavf_wq, &adapter->adminq_task); return IRQ_HANDLED; } @@ -2374,8 +2375,12 @@ static void iavf_watchdog_task(struct work_struct *work) struct iavf_hw *hw = &adapter->hw; u32 reg_val; - if (!mutex_trylock(&adapter->crit_lock)) + if (!mutex_trylock(&adapter->crit_lock)) { + if (adapter->state == __IAVF_REMOVE) + return; + goto restart_watchdog; + } if (adapter->flags & IAVF_FLAG_PF_COMMS_FAILED) iavf_change_state(adapter, __IAVF_COMM_FAILED); @@ -2601,13 +2606,13 @@ static void iavf_reset_task(struct work_struct *work) /* When device is being removed it doesn't make sense to run the reset * task, just return in such a case. */ - if (mutex_is_locked(&adapter->remove_lock)) - return; + if (!mutex_trylock(&adapter->crit_lock)) { + if (adapter->state != __IAVF_REMOVE) + queue_work(iavf_wq, &adapter->reset_task); - if (iavf_lock_timeout(&adapter->crit_lock, 200)) { - schedule_work(&adapter->reset_task); return; } + while (!mutex_trylock(&adapter->client_lock)) usleep_range(500, 1000); if (CLIENT_ENABLED(adapter)) { @@ -2826,13 +2831,19 @@ static void iavf_adminq_task(struct work_struct *work) if (adapter->flags & IAVF_FLAG_PF_COMMS_FAILED) goto out; + if (!mutex_trylock(&adapter->crit_lock)) { + if (adapter->state == __IAVF_REMOVE) + return; + + queue_work(iavf_wq, &adapter->adminq_task); + goto out; + } + event.buf_len = IAVF_MAX_AQ_BUF_SIZE; event.msg_buf = kzalloc(event.buf_len, GFP_KERNEL); if (!event.msg_buf) goto out; - if (iavf_lock_timeout(&adapter->crit_lock, 200)) - goto freedom; do { ret = iavf_clean_arq_element(hw, &event, &pending); v_op = (enum virtchnl_ops)le32_to_cpu(event.desc.cookie_high); @@ -3800,11 +3811,12 @@ static int iavf_close(struct net_device *netdev) struct iavf_adapter *adapter = netdev_priv(netdev); int status; - if (adapter->state <= __IAVF_DOWN_PENDING) - return 0; + mutex_lock(&adapter->crit_lock); - while (!mutex_trylock(&adapter->crit_lock)) - usleep_range(500, 1000); + if (adapter->state <= __IAVF_DOWN_PENDING) { + mutex_unlock(&adapter->crit_lock); + return 0; + } set_bit(__IAVF_VSI_DOWN, adapter->vsi.state); if (CLIENT_ENABLED(adapter)) @@ -4431,7 +4443,6 @@ static int iavf_probe(struct pci_dev *pdev, const struct pci_device_id *ent) */ mutex_init(&adapter->crit_lock); mutex_init(&adapter->client_lock); - mutex_init(&adapter->remove_lock); mutex_init(&hw->aq.asq_mutex); mutex_init(&hw->aq.arq_mutex); @@ -4556,11 +4567,7 @@ static void iavf_remove(struct pci_dev *pdev) struct iavf_cloud_filter *cf, *cftmp; struct iavf_hw *hw = &adapter->hw; int err; - /* Indicate we are in remove and not to run reset_task */ - mutex_lock(&adapter->remove_lock); - cancel_work_sync(&adapter->reset_task); - cancel_delayed_work_sync(&adapter->watchdog_task); - cancel_delayed_work_sync(&adapter->client_task); + if (adapter->netdev_registered) { unregister_netdev(netdev); adapter->netdev_registered = false; @@ -4572,6 +4579,10 @@ static void iavf_remove(struct pci_dev *pdev) err); } + mutex_lock(&adapter->crit_lock); + dev_info(&adapter->pdev->dev, "Remove device\n"); + iavf_change_state(adapter, __IAVF_REMOVE); + iavf_request_reset(adapter); msleep(50); /* If the FW isn't responding, kick it once, but only once. */ @@ -4579,18 +4590,19 @@ static void iavf_remove(struct pci_dev *pdev) iavf_request_reset(adapter); msleep(50); } - if (iavf_lock_timeout(&adapter->crit_lock, 5000)) - dev_warn(&adapter->pdev->dev, "failed to acquire crit_lock in %s\n", __FUNCTION__); - dev_info(&adapter->pdev->dev, "Removing device\n"); + iavf_misc_irq_disable(adapter); /* Shut down all the garbage mashers on the detention level */ - iavf_change_state(adapter, __IAVF_REMOVE); + cancel_work_sync(&adapter->reset_task); + cancel_delayed_work_sync(&adapter->watchdog_task); + cancel_work_sync(&adapter->adminq_task); + cancel_delayed_work_sync(&adapter->client_task); + adapter->aq_required = 0; adapter->flags &= ~IAVF_FLAG_REINIT_ITR_NEEDED; iavf_free_all_tx_resources(adapter); iavf_free_all_rx_resources(adapter); - iavf_misc_irq_disable(adapter); iavf_free_misc_irq(adapter); /* In case we enter iavf_remove from erroneous state, free traffic irqs @@ -4606,10 +4618,6 @@ static void iavf_remove(struct pci_dev *pdev) iavf_reset_interrupt_capability(adapter); iavf_free_q_vectors(adapter); - cancel_delayed_work_sync(&adapter->watchdog_task); - - cancel_work_sync(&adapter->adminq_task); - iavf_free_rss(adapter); if (hw->aq.asq.count) @@ -4621,8 +4629,6 @@ static void iavf_remove(struct pci_dev *pdev) mutex_destroy(&adapter->client_lock); mutex_unlock(&adapter->crit_lock); mutex_destroy(&adapter->crit_lock); - mutex_unlock(&adapter->remove_lock); - mutex_destroy(&adapter->remove_lock); iounmap(hw->hw_addr); pci_release_regions(pdev); -- cgit v1.2.3-70-g09d2 From 974578017fc1fdd06cea8afb9dfa32602e8529ed Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:36:56 +0100 Subject: iavf: Add waiting so the port is initialized in remove There exist races when port is being configured and remove is triggered. unregister_netdev is not and can't be called under crit_lock mutex since it is calling ndo_stop -> iavf_close which requires this lock. Depending on init state the netdev could be still unregistered so unregister_netdev never cleans up, when shortly after that the device could become registered. Make iavf_remove wait until port finishes initialization. All critical state changes are atomic (under crit_lock). Crashes that come from iavf_reset_interrupt_capability and iavf_free_traffic_irqs should now be solved in a graceful manner. Fixes: 605ca7c5c6707 ("iavf: Fix kernel BUG in free_msi_irqs") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_main.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 84ae96e912d7..5e71b38e9154 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -4558,7 +4558,6 @@ static int __maybe_unused iavf_resume(struct device *dev_d) static void iavf_remove(struct pci_dev *pdev) { struct iavf_adapter *adapter = iavf_pdev_to_adapter(pdev); - enum iavf_state_t prev_state = adapter->last_state; struct net_device *netdev = adapter->netdev; struct iavf_fdir_fltr *fdir, *fdirtmp; struct iavf_vlan_filter *vlf, *vlftmp; @@ -4568,6 +4567,22 @@ static void iavf_remove(struct pci_dev *pdev) struct iavf_hw *hw = &adapter->hw; int err; + /* Wait until port initialization is complete. + * There are flows where register/unregister netdev may race. + */ + while (1) { + mutex_lock(&adapter->crit_lock); + if (adapter->state == __IAVF_RUNNING || + adapter->state == __IAVF_DOWN) { + mutex_unlock(&adapter->crit_lock); + break; + } + + mutex_unlock(&adapter->crit_lock); + usleep_range(500, 1000); + } + cancel_delayed_work_sync(&adapter->watchdog_task); + if (adapter->netdev_registered) { unregister_netdev(netdev); adapter->netdev_registered = false; @@ -4605,16 +4620,6 @@ static void iavf_remove(struct pci_dev *pdev) iavf_free_all_rx_resources(adapter); iavf_free_misc_irq(adapter); - /* In case we enter iavf_remove from erroneous state, free traffic irqs - * here, so as to not cause a kernel crash, when calling - * iavf_reset_interrupt_capability. - */ - if ((adapter->last_state == __IAVF_RESETTING && - prev_state != __IAVF_DOWN) || - (adapter->last_state == __IAVF_RUNNING && - !(netdev->flags & IFF_UP))) - iavf_free_traffic_irqs(adapter); - iavf_reset_interrupt_capability(adapter); iavf_free_q_vectors(adapter); -- cgit v1.2.3-70-g09d2 From 3ccd54ef44ebfa0792c5441b6d9c86618f3378d1 Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:37:10 +0100 Subject: iavf: Fix init state closure on remove When init states of the adapter work, the errors like lack of communication with the PF might hop in. If such events occur the driver restores previous states in order to retry initialization in a proper way. When remove task kicks in, this situation could lead to races with unregistering the netdevice as well as resources cleanup. With the commit introducing the waiting in remove for init to complete, this problem turns into an endless waiting if init never recovers from errors. Introduce __IAVF_IN_REMOVE_TASK bit to indicate that the remove thread has started. Make __IAVF_COMM_FAILED adapter state respect the __IAVF_IN_REMOVE_TASK bit and set the __IAVF_INIT_FAILED state and return without any action instead of trying to recover. Make __IAVF_INIT_FAILED adapter state respect the __IAVF_IN_REMOVE_TASK bit and return without any further actions. Make the loop in the remove handler break when adapter has __IAVF_INIT_FAILED state set. Fixes: 898ef1cb1cb2 ("iavf: Combine init and watchdog state machines") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf.h | 4 ++++ drivers/net/ethernet/intel/iavf/iavf_main.c | 24 +++++++++++++++++++++++- 2 files changed, 27 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf.h b/drivers/net/ethernet/intel/iavf/iavf.h index 44f83e06486d..f259fd517b2c 100644 --- a/drivers/net/ethernet/intel/iavf/iavf.h +++ b/drivers/net/ethernet/intel/iavf/iavf.h @@ -201,6 +201,10 @@ enum iavf_state_t { __IAVF_RUNNING, /* opened, working */ }; +enum iavf_critical_section_t { + __IAVF_IN_REMOVE_TASK, /* device being removed */ +}; + #define IAVF_CLOUD_FIELD_OMAC 0x01 #define IAVF_CLOUD_FIELD_IMAC 0x02 #define IAVF_CLOUD_FIELD_IVLAN 0x04 diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 5e71b38e9154..be51da978e7c 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -2424,6 +2424,15 @@ static void iavf_watchdog_task(struct work_struct *work) msecs_to_jiffies(1)); return; case __IAVF_INIT_FAILED: + if (test_bit(__IAVF_IN_REMOVE_TASK, + &adapter->crit_section)) { + /* Do not update the state and do not reschedule + * watchdog task, iavf_remove should handle this state + * as it can loop forever + */ + mutex_unlock(&adapter->crit_lock); + return; + } if (++adapter->aq_wait_count > IAVF_AQ_MAX_ERR) { dev_err(&adapter->pdev->dev, "Failed to communicate with PF; waiting before retry\n"); @@ -2440,6 +2449,17 @@ static void iavf_watchdog_task(struct work_struct *work) queue_delayed_work(iavf_wq, &adapter->watchdog_task, HZ); return; case __IAVF_COMM_FAILED: + if (test_bit(__IAVF_IN_REMOVE_TASK, + &adapter->crit_section)) { + /* Set state to __IAVF_INIT_FAILED and perform remove + * steps. Remove IAVF_FLAG_PF_COMMS_FAILED so the task + * doesn't bring the state back to __IAVF_COMM_FAILED. + */ + iavf_change_state(adapter, __IAVF_INIT_FAILED); + adapter->flags &= ~IAVF_FLAG_PF_COMMS_FAILED; + mutex_unlock(&adapter->crit_lock); + return; + } reg_val = rd32(hw, IAVF_VFGEN_RSTAT) & IAVF_VFGEN_RSTAT_VFR_STATE_MASK; if (reg_val == VIRTCHNL_VFR_VFACTIVE || @@ -4567,13 +4587,15 @@ static void iavf_remove(struct pci_dev *pdev) struct iavf_hw *hw = &adapter->hw; int err; + set_bit(__IAVF_IN_REMOVE_TASK, &adapter->crit_section); /* Wait until port initialization is complete. * There are flows where register/unregister netdev may race. */ while (1) { mutex_lock(&adapter->crit_lock); if (adapter->state == __IAVF_RUNNING || - adapter->state == __IAVF_DOWN) { + adapter->state == __IAVF_DOWN || + adapter->state == __IAVF_INIT_FAILED) { mutex_unlock(&adapter->crit_lock); break; } -- cgit v1.2.3-70-g09d2 From 0579fafd37fb7efe091f0e6c8ccf968864f40f3e Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:37:50 +0100 Subject: iavf: Fix locking for VIRTCHNL_OP_GET_OFFLOAD_VLAN_V2_CAPS iavf_virtchnl_completion is called under crit_lock but when the code for VIRTCHNL_OP_GET_OFFLOAD_VLAN_V2_CAPS is called, this lock is released in order to obtain rtnl_lock to avoid ABBA deadlock with unregister_netdev. Along with the new way iavf_remove behaves, there exist many risks related to the lock release and attmepts to regrab it. The driver faces crashes related to races between unregister_netdev and netdev_update_features. Yet another risk is that the driver could already obtain the crit_lock in order to destroy it and iavf_virtchnl_completion could crash or block forever. Make iavf_virtchnl_completion never relock crit_lock in it's call paths. Extract rtnl_lock locking logic to the driver for unregister_netdev in order to set the netdev_registered flag inside the lock. Introduce a new flag that will inform adminq_task to perform the code from VIRTCHNL_OP_GET_OFFLOAD_VLAN_V2_CAPS right after it finishes processing messages. Guard this code with remove flags so it's never called when the driver is in remove state. Fixes: 5951a2b9812d ("iavf: Fix VLAN feature flags after VFR") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf.h | 1 + drivers/net/ethernet/intel/iavf/iavf_main.c | 22 +++++++++++++++++++++- drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 24 +----------------------- 3 files changed, 23 insertions(+), 24 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf.h b/drivers/net/ethernet/intel/iavf/iavf.h index f259fd517b2c..89423947ee65 100644 --- a/drivers/net/ethernet/intel/iavf/iavf.h +++ b/drivers/net/ethernet/intel/iavf/iavf.h @@ -287,6 +287,7 @@ struct iavf_adapter { #define IAVF_FLAG_LEGACY_RX BIT(15) #define IAVF_FLAG_REINIT_ITR_NEEDED BIT(16) #define IAVF_FLAG_QUEUES_DISABLED BIT(17) +#define IAVF_FLAG_SETUP_NETDEV_FEATURES BIT(18) /* duplicates for common code */ #define IAVF_FLAG_DCB_ENABLED 0 /* flags for admin queue service task */ diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index be51da978e7c..67349d24dc90 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -2879,6 +2879,24 @@ static void iavf_adminq_task(struct work_struct *work) } while (pending); mutex_unlock(&adapter->crit_lock); + if ((adapter->flags & IAVF_FLAG_SETUP_NETDEV_FEATURES)) { + if (adapter->netdev_registered || + !test_bit(__IAVF_IN_REMOVE_TASK, &adapter->crit_section)) { + struct net_device *netdev = adapter->netdev; + + rtnl_lock(); + netdev_update_features(netdev); + rtnl_unlock(); + /* Request VLAN offload settings */ + if (VLAN_V2_ALLOWED(adapter)) + iavf_set_vlan_offload_features + (adapter, 0, netdev->features); + + iavf_set_queue_vlan_tag_loc(adapter); + } + + adapter->flags &= ~IAVF_FLAG_SETUP_NETDEV_FEATURES; + } if ((adapter->flags & (IAVF_FLAG_RESET_PENDING | IAVF_FLAG_RESET_NEEDED)) || adapter->state == __IAVF_RESETTING) @@ -4606,8 +4624,10 @@ static void iavf_remove(struct pci_dev *pdev) cancel_delayed_work_sync(&adapter->watchdog_task); if (adapter->netdev_registered) { - unregister_netdev(netdev); + rtnl_lock(); + unregister_netdevice(netdev); adapter->netdev_registered = false; + rtnl_unlock(); } if (CLIENT_ALLOWED(adapter)) { err = iavf_lan_del_device(adapter); diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c index 5ee1d118fd30..88844d68e150 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c @@ -2146,29 +2146,7 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter, sizeof(adapter->vlan_v2_caps))); iavf_process_config(adapter); - - /* unlock crit_lock before acquiring rtnl_lock as other - * processes holding rtnl_lock could be waiting for the same - * crit_lock - */ - mutex_unlock(&adapter->crit_lock); - /* VLAN capabilities can change during VFR, so make sure to - * update the netdev features with the new capabilities - */ - rtnl_lock(); - netdev_update_features(netdev); - rtnl_unlock(); - if (iavf_lock_timeout(&adapter->crit_lock, 10000)) - dev_warn(&adapter->pdev->dev, "failed to acquire crit_lock in %s\n", - __FUNCTION__); - - /* Request VLAN offload settings */ - if (VLAN_V2_ALLOWED(adapter)) - iavf_set_vlan_offload_features(adapter, 0, - netdev->features); - - iavf_set_queue_vlan_tag_loc(adapter); - + adapter->flags |= IAVF_FLAG_SETUP_NETDEV_FEATURES; } break; case VIRTCHNL_OP_ENABLE_QUEUES: -- cgit v1.2.3-70-g09d2 From a472eb5cbaebb5774672c565e024336c039e9128 Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:38:01 +0100 Subject: iavf: Fix race in init state When iavf_init_version_check sends VIRTCHNL_OP_GET_VF_RESOURCES message, the driver will wait for the response after requeueing the watchdog task in iavf_init_get_resources call stack. The logic is implemented this way that iavf_init_get_resources has to be called in order to allocate adapter->vf_res. It is polling for the AQ response in iavf_get_vf_config function. Expect a call trace from kernel when adminq_task worker handles this message first. adapter->vf_res will be NULL in iavf_virtchnl_completion. Make the watchdog task not queue the adminq_task if the init process is not finished yet. Fixes: 898ef1cb1cb2 ("iavf: Combine init and watchdog state machines") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 67349d24dc90..36433d6504b7 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -2532,7 +2532,8 @@ static void iavf_watchdog_task(struct work_struct *work) schedule_delayed_work(&adapter->client_task, msecs_to_jiffies(5)); mutex_unlock(&adapter->crit_lock); restart_watchdog: - queue_work(iavf_wq, &adapter->adminq_task); + if (adapter->state >= __IAVF_DOWN) + queue_work(iavf_wq, &adapter->adminq_task); if (adapter->aq_required) queue_delayed_work(iavf_wq, &adapter->watchdog_task, msecs_to_jiffies(20)); -- cgit v1.2.3-70-g09d2 From e85ff9c631e1bf109ce8428848dfc8e8b0041f48 Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:38:31 +0100 Subject: iavf: Fix deadlock in iavf_reset_task There exists a missing mutex_unlock call on crit_lock in iavf_reset_task call path. Unlock the crit_lock before returning from reset task. Fixes: 5ac49f3c2702 ("iavf: use mutexes for locking of critical sections") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_main.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 36433d6504b7..da50ea3e44c2 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -2688,6 +2688,7 @@ static void iavf_reset_task(struct work_struct *work) reg_val); iavf_disable_vf(adapter); mutex_unlock(&adapter->client_lock); + mutex_unlock(&adapter->crit_lock); return; /* Do not attempt to reinit. It's dead, Jim. */ } -- cgit v1.2.3-70-g09d2 From d2c0f45fcceb0995f208c441d9c9a453623f9ccf Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:38:43 +0100 Subject: iavf: Fix missing check for running netdev The driver was queueing reset_task regardless of the netdev state. Do not queue the reset task in iavf_change_mtu if netdev is not running. Fixes: fdd4044ffdc8 ("iavf: Remove timer for work triggering, use delaying work instead") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_main.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index da50ea3e44c2..be97ac251802 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -3905,8 +3905,11 @@ static int iavf_change_mtu(struct net_device *netdev, int new_mtu) iavf_notify_client_l2_params(&adapter->vsi); adapter->flags |= IAVF_FLAG_SERVICE_CLIENT_REQUESTED; } - adapter->flags |= IAVF_FLAG_RESET_NEEDED; - queue_work(iavf_wq, &adapter->reset_task); + + if (netif_running(netdev)) { + adapter->flags |= IAVF_FLAG_RESET_NEEDED; + queue_work(iavf_wq, &adapter->reset_task); + } return 0; } -- cgit v1.2.3-70-g09d2 From 14756b2ae265d526b8356e86729090b01778fdf6 Mon Sep 17 00:00:00 2001 From: Slawomir Laba Date: Wed, 23 Feb 2022 13:38:55 +0100 Subject: iavf: Fix __IAVF_RESETTING state usage The setup of __IAVF_RESETTING state in watchdog task had no effect and could lead to slow resets in the driver as the task for __IAVF_RESETTING state only requeues watchdog. Till now the __IAVF_RESETTING was interpreted by reset task as running state which could lead to errors with allocating and resources disposal. Make watchdog_task queue the reset task when it's necessary. Do not update the state to __IAVF_RESETTING so the reset task knows exactly what is the current state of the adapter. Fixes: 898ef1cb1cb2 ("iavf: Combine init and watchdog state machines") Signed-off-by: Slawomir Laba Signed-off-by: Phani Burra Signed-off-by: Jacob Keller Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_main.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index be97ac251802..dcf24264c7ea 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -1137,8 +1137,7 @@ void iavf_down(struct iavf_adapter *adapter) rss->state = IAVF_ADV_RSS_DEL_REQUEST; spin_unlock_bh(&adapter->adv_rss_lock); - if (!(adapter->flags & IAVF_FLAG_PF_COMMS_FAILED) && - adapter->state != __IAVF_RESETTING) { + if (!(adapter->flags & IAVF_FLAG_PF_COMMS_FAILED)) { /* cancel any current operation */ adapter->current_op = VIRTCHNL_OP_UNKNOWN; /* Schedule operations to close down the HW. Don't wait @@ -2385,11 +2384,12 @@ static void iavf_watchdog_task(struct work_struct *work) if (adapter->flags & IAVF_FLAG_PF_COMMS_FAILED) iavf_change_state(adapter, __IAVF_COMM_FAILED); - if (adapter->flags & IAVF_FLAG_RESET_NEEDED && - adapter->state != __IAVF_RESETTING) { - iavf_change_state(adapter, __IAVF_RESETTING); + if (adapter->flags & IAVF_FLAG_RESET_NEEDED) { adapter->aq_required = 0; adapter->current_op = VIRTCHNL_OP_UNKNOWN; + mutex_unlock(&adapter->crit_lock); + queue_work(iavf_wq, &adapter->reset_task); + return; } switch (adapter->state) { @@ -2697,8 +2697,7 @@ continue_reset: * ndo_open() returning, so we can't assume it means all our open * tasks have finished, since we're not holding the rtnl_lock here. */ - running = ((adapter->state == __IAVF_RUNNING) || - (adapter->state == __IAVF_RESETTING)); + running = adapter->state == __IAVF_RUNNING; if (running) { netdev->flags &= ~IFF_UP; -- cgit v1.2.3-70-g09d2 From 90d4025285748448809701a44cf466a3f5443eaa Mon Sep 17 00:00:00 2001 From: Casper Andersson Date: Fri, 25 Feb 2022 13:43:27 +0100 Subject: net: sparx5: Add #include to remove warning main.h uses NUM_TARGETS from main_regs.h, but the missing include never causes any errors because everywhere main.h is (currently) included, main_regs.h is included before. But since it is dependent on main_regs.h it should always be included. Signed-off-by: Casper Andersson Reviewed-by: Joacim Zetterling Signed-off-by: David S. Miller --- drivers/net/ethernet/microchip/sparx5/sparx5_main.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/microchip/sparx5/sparx5_main.h b/drivers/net/ethernet/microchip/sparx5/sparx5_main.h index a1acc9b461f2..d40e18ce3293 100644 --- a/drivers/net/ethernet/microchip/sparx5/sparx5_main.h +++ b/drivers/net/ethernet/microchip/sparx5/sparx5_main.h @@ -16,6 +16,8 @@ #include #include +#include "sparx5_main_regs.h" + /* Target chip type */ enum spx5_target_chiptype { SPX5_TARGET_CT_7546 = 0x7546, /* SparX-5-64 Enterprise */ -- cgit v1.2.3-70-g09d2 From fda2635466cd26ad237e1bc5d3f6a60f97ad09b6 Mon Sep 17 00:00:00 2001 From: Corinna Vinschen Date: Wed, 16 Feb 2022 14:31:35 +0100 Subject: igc: igc_read_phy_reg_gpy: drop premature return igc_read_phy_reg_gpy checks the return value from igc_read_phy_reg_mdic and if it's not 0, returns immediately. By doing this, it leaves the HW semaphore in the acquired state. Drop this premature return statement, the function returns after releasing the semaphore immediately anyway. Fixes: 5586838fe9ce ("igc: Add code for PHY support") Signed-off-by: Corinna Vinschen Acked-by: Sasha Neftin Tested-by: Naama Meir Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/igc/igc_phy.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/igc/igc_phy.c b/drivers/net/ethernet/intel/igc/igc_phy.c index 5cad31c3c7b0..df91d07ce82a 100644 --- a/drivers/net/ethernet/intel/igc/igc_phy.c +++ b/drivers/net/ethernet/intel/igc/igc_phy.c @@ -779,8 +779,6 @@ s32 igc_read_phy_reg_gpy(struct igc_hw *hw, u32 offset, u16 *data) if (ret_val) return ret_val; ret_val = igc_read_phy_reg_mdic(hw, offset, data); - if (ret_val) - return ret_val; hw->phy.ops.release(hw); } else { ret_val = igc_read_xmdio_reg(hw, (u16)offset, dev_addr, -- cgit v1.2.3-70-g09d2 From c4208653a327a09da1e9e7b10299709b6d9b17bf Mon Sep 17 00:00:00 2001 From: Sasha Neftin Date: Sun, 20 Feb 2022 09:29:15 +0200 Subject: igc: igc_write_phy_reg_gpy: drop premature return Similar to "igc_read_phy_reg_gpy: drop premature return" patch. igc_write_phy_reg_gpy checks the return value from igc_write_phy_reg_mdic and if it's not 0, returns immediately. By doing this, it leaves the HW semaphore in the acquired state. Drop this premature return statement, the function returns after releasing the semaphore immediately anyway. Fixes: 5586838fe9ce ("igc: Add code for PHY support") Suggested-by: Dima Ruinskiy Reported-by: Corinna Vinschen Signed-off-by: Sasha Neftin Tested-by: Naama Meir Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/igc/igc_phy.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/igc/igc_phy.c b/drivers/net/ethernet/intel/igc/igc_phy.c index df91d07ce82a..40dbf4b43234 100644 --- a/drivers/net/ethernet/intel/igc/igc_phy.c +++ b/drivers/net/ethernet/intel/igc/igc_phy.c @@ -746,8 +746,6 @@ s32 igc_write_phy_reg_gpy(struct igc_hw *hw, u32 offset, u16 data) if (ret_val) return ret_val; ret_val = igc_write_phy_reg_mdic(hw, offset, data); - if (ret_val) - return ret_val; hw->phy.ops.release(hw); } else { ret_val = igc_write_xmdio_reg(hw, (u16)offset, dev_addr, -- cgit v1.2.3-70-g09d2 From 1866aa0d0d6492bc2f8d22d0df49abaccf50cddd Mon Sep 17 00:00:00 2001 From: Sasha Neftin Date: Tue, 25 Jan 2022 19:31:23 +0200 Subject: e1000e: Fix possible HW unit hang after an s0ix exit Disable the OEM bit/Gig Disable/restart AN impact and disable the PHY LAN connected device (LCD) reset during power management flows. This fixes possible HW unit hangs on the s0ix exit on some corporate ADL platforms. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=214821 Fixes: 3e55d231716e ("e1000e: Add handshake with the CSME to support S0ix") Suggested-by: Dima Ruinskiy Suggested-by: Nir Efrati Signed-off-by: Sasha Neftin Tested-by: Kai-Heng Feng Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/e1000e/hw.h | 1 + drivers/net/ethernet/intel/e1000e/ich8lan.c | 4 ++++ drivers/net/ethernet/intel/e1000e/ich8lan.h | 1 + drivers/net/ethernet/intel/e1000e/netdev.c | 26 ++++++++++++++++++++++++++ 4 files changed, 32 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/e1000e/hw.h b/drivers/net/ethernet/intel/e1000e/hw.h index bcf680e83811..13382df2f2ef 100644 --- a/drivers/net/ethernet/intel/e1000e/hw.h +++ b/drivers/net/ethernet/intel/e1000e/hw.h @@ -630,6 +630,7 @@ struct e1000_phy_info { bool disable_polarity_correction; bool is_mdix; bool polarity_correction; + bool reset_disable; bool speed_downgraded; bool autoneg_wait_to_complete; }; diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c index c908c84b86d2..e298da712758 100644 --- a/drivers/net/ethernet/intel/e1000e/ich8lan.c +++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c @@ -2050,6 +2050,10 @@ static s32 e1000_check_reset_block_ich8lan(struct e1000_hw *hw) bool blocked = false; int i = 0; + /* Check the PHY (LCD) reset flag */ + if (hw->phy.reset_disable) + return true; + while ((blocked = !(er32(FWSM) & E1000_ICH_FWSM_RSPCIPHY)) && (i++ < 30)) usleep_range(10000, 11000); diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.h b/drivers/net/ethernet/intel/e1000e/ich8lan.h index 2504b11c3169..638a3ddd7ada 100644 --- a/drivers/net/ethernet/intel/e1000e/ich8lan.h +++ b/drivers/net/ethernet/intel/e1000e/ich8lan.h @@ -271,6 +271,7 @@ #define I217_CGFREG_ENABLE_MTA_RESET 0x0002 #define I217_MEMPWR PHY_REG(772, 26) #define I217_MEMPWR_DISABLE_SMB_RELEASE 0x0010 +#define I217_MEMPWR_MOEM 0x1000 /* Receive Address Initial CRC Calculation */ #define E1000_PCH_RAICC(_n) (0x05F50 + ((_n) * 4)) diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c index a42aeb555f34..c5bdef3ffe26 100644 --- a/drivers/net/ethernet/intel/e1000e/netdev.c +++ b/drivers/net/ethernet/intel/e1000e/netdev.c @@ -6987,8 +6987,21 @@ static __maybe_unused 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; + u16 phy_data; int rc; + if (er32(FWSM) & E1000_ICH_FWSM_FW_VALID && + hw->mac.type >= e1000_pch_adp) { + /* Mask OEM Bits / Gig Disable / Restart AN (772_26[12] = 1) */ + e1e_rphy(hw, I217_MEMPWR, &phy_data); + phy_data |= I217_MEMPWR_MOEM; + e1e_wphy(hw, I217_MEMPWR, phy_data); + + /* Disable LCD reset */ + hw->phy.reset_disable = true; + } + e1000e_flush_lpic(pdev); e1000e_pm_freeze(dev); @@ -7010,6 +7023,8 @@ static __maybe_unused 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; + u16 phy_data; int rc; /* Introduce S0ix implementation */ @@ -7020,6 +7035,17 @@ static __maybe_unused int e1000e_pm_resume(struct device *dev) if (rc) return rc; + if (er32(FWSM) & E1000_ICH_FWSM_FW_VALID && + hw->mac.type >= e1000_pch_adp) { + /* Unmask OEM Bits / Gig Disable / Restart AN 772_26[12] = 0 */ + e1e_rphy(hw, I217_MEMPWR, &phy_data); + phy_data &= ~I217_MEMPWR_MOEM; + e1e_wphy(hw, I217_MEMPWR, phy_data); + + /* Enable LCD reset */ + hw->phy.reset_disable = false; + } + return e1000e_pm_thaw(dev); } -- cgit v1.2.3-70-g09d2 From ffd24fa2fcc76ecb2e61e7a4ef8588177bcb42a6 Mon Sep 17 00:00:00 2001 From: Sasha Neftin Date: Thu, 3 Feb 2022 14:21:49 +0200 Subject: e1000e: Correct NVM checksum verification flow Update MAC type check e1000_pch_tgp because for e1000_pch_cnp, NVM checksum update is still possible. Emit a more detailed warning message. Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=1191663 Fixes: 4051f68318ca ("e1000e: Do not take care about recovery NVM checksum") Reported-by: Thomas Bogendoerfer Signed-off-by: Sasha Neftin Tested-by: Naama Meir Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/e1000e/ich8lan.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c index e298da712758..d60e2016d03c 100644 --- a/drivers/net/ethernet/intel/e1000e/ich8lan.c +++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c @@ -4140,9 +4140,9 @@ static s32 e1000_validate_nvm_checksum_ich8lan(struct e1000_hw *hw) return ret_val; if (!(data & valid_csum_mask)) { - e_dbg("NVM Checksum Invalid\n"); + e_dbg("NVM Checksum valid bit not set\n"); - if (hw->mac.type < e1000_pch_cnp) { + if (hw->mac.type < e1000_pch_tgp) { data |= valid_csum_mask; ret_val = e1000_write_nvm(hw, word, 1, &data); if (ret_val) -- cgit v1.2.3-70-g09d2 From f1fb205efb0ccca55626fd4ef38570dd16b44719 Mon Sep 17 00:00:00 2001 From: Niels Dossche Date: Tue, 1 Mar 2022 23:28:22 +0100 Subject: sfc: extend the locking on mcdi->seqno seqno could be read as a stale value outside of the lock. The lock is already acquired to protect the modification of seqno against a possible race condition. Place the reading of this value also inside this locking to protect it against a possible race condition. Signed-off-by: Niels Dossche Acked-by: Martin Habets Signed-off-by: David S. Miller --- drivers/net/ethernet/sfc/mcdi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c index be6bfd6b7ec7..50baf62b2cbc 100644 --- a/drivers/net/ethernet/sfc/mcdi.c +++ b/drivers/net/ethernet/sfc/mcdi.c @@ -163,9 +163,9 @@ static void efx_mcdi_send_request(struct efx_nic *efx, unsigned cmd, /* Serialise with efx_mcdi_ev_cpl() and efx_mcdi_ev_death() */ spin_lock_bh(&mcdi->iface_lock); ++mcdi->seqno; + seqno = mcdi->seqno & SEQ_MASK; spin_unlock_bh(&mcdi->iface_lock); - seqno = mcdi->seqno & SEQ_MASK; xflags = 0; if (mcdi->mode == MCDI_MODE_EVENTS) xflags |= MCDI_HEADER_XFLAGS_EVREQ; -- cgit v1.2.3-70-g09d2 From 8ccffe9ac3239e549beaa0a9d5e1a1eac94e866c Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Wed, 2 Mar 2022 21:21:15 +0100 Subject: bnx2: Fix an error message Fix an error message and report the correct failing function. Signed-off-by: Christophe JAILLET Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnx2.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/bnx2.c b/drivers/net/ethernet/broadcom/bnx2.c index e20aafeb4ca9..b97ed9b5f685 100644 --- a/drivers/net/ethernet/broadcom/bnx2.c +++ b/drivers/net/ethernet/broadcom/bnx2.c @@ -8216,7 +8216,7 @@ bnx2_init_board(struct pci_dev *pdev, struct net_device *dev) rc = dma_set_coherent_mask(&pdev->dev, persist_dma_mask); if (rc) { dev_err(&pdev->dev, - "pci_set_consistent_dma_mask failed, aborting\n"); + "dma_set_coherent_mask failed, aborting\n"); goto err_out_unmap; } } else if ((rc = dma_set_mask(&pdev->dev, DMA_BIT_MASK(32))) != 0) { -- cgit v1.2.3-70-g09d2 From 6c7273a266759d9d36f7c862149f248bcdeddc0f Mon Sep 17 00:00:00 2001 From: Maciej Fijalkowski Date: Wed, 2 Mar 2022 09:59:27 -0800 Subject: ixgbe: xsk: change !netif_carrier_ok() handling in ixgbe_xmit_zc() Commit c685c69fba71 ("ixgbe: don't do any AF_XDP zero-copy transmit if netif is not OK") addressed the ring transient state when MEM_TYPE_XSK_BUFF_POOL was being configured which in turn caused the interface to through down/up. Maurice reported that when carrier is not ok and xsk_pool is present on ring pair, ksoftirqd will consume 100% CPU cycles due to the constant NAPI rescheduling as ixgbe_poll() states that there is still some work to be done. To fix this, do not set work_done to false for a !netif_carrier_ok(). Fixes: c685c69fba71 ("ixgbe: don't do any AF_XDP zero-copy transmit if netif is not OK") Reported-by: Maurice Baijens Tested-by: Maurice Baijens Signed-off-by: Maciej Fijalkowski Tested-by: Sandeep Penigalapati Signed-off-by: Tony Nguyen Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c index b3fd8e5cd85b..6a5e9cf6b5da 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_xsk.c @@ -390,12 +390,14 @@ static bool ixgbe_xmit_zc(struct ixgbe_ring *xdp_ring, unsigned int budget) u32 cmd_type; while (budget-- > 0) { - if (unlikely(!ixgbe_desc_unused(xdp_ring)) || - !netif_carrier_ok(xdp_ring->netdev)) { + if (unlikely(!ixgbe_desc_unused(xdp_ring))) { work_done = false; break; } + if (!netif_carrier_ok(xdp_ring->netdev)) + break; + if (!xsk_tx_peek_desc(pool, &desc)) break; -- cgit v1.2.3-70-g09d2 From 0bf476fc3624e3a72af4ba7340d430a91c18cd67 Mon Sep 17 00:00:00 2001 From: Robert Hancock Date: Thu, 3 Mar 2022 12:10:27 -0600 Subject: net: macb: Fix lost RX packet wakeup race in NAPI receive There is an oddity in the way the RSR register flags propagate to the ISR register (and the actual interrupt output) on this hardware: it appears that RSR register bits only result in ISR being asserted if the interrupt was actually enabled at the time, so enabling interrupts with RSR bits already set doesn't trigger an interrupt to be raised. There was already a partial fix for this race in the macb_poll function where it checked for RSR bits being set and re-triggered NAPI receive. However, there was a still a race window between checking RSR and actually enabling interrupts, where a lost wakeup could happen. It's necessary to check again after enabling interrupts to see if RSR was set just prior to the interrupt being enabled, and re-trigger receive in that case. This issue was noticed in a point-to-point UDP request-response protocol which periodically saw timeouts or abnormally high response times due to received packets not being processed in a timely fashion. In many applications, more packets arriving, including TCP retransmissions, would cause the original packet to be processed, thus masking the issue. Fixes: 02f7a34f34e3 ("net: macb: Re-enable RX interrupt only when RX is done") Cc: stable@vger.kernel.org Co-developed-by: Scott McNutt Signed-off-by: Scott McNutt Signed-off-by: Robert Hancock Tested-by: Claudiu Beznea Signed-off-by: David S. Miller --- drivers/net/ethernet/cadence/macb_main.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/cadence/macb_main.c b/drivers/net/ethernet/cadence/macb_main.c index 98498a76ae16..d13f06cf0308 100644 --- a/drivers/net/ethernet/cadence/macb_main.c +++ b/drivers/net/ethernet/cadence/macb_main.c @@ -1573,7 +1573,14 @@ static int macb_poll(struct napi_struct *napi, int budget) if (work_done < budget) { napi_complete_done(napi, work_done); - /* Packets received while interrupts were disabled */ + /* RSR bits only seem to propagate to raise interrupts when + * interrupts are enabled at the time, so if bits are already + * set due to packets received while interrupts were disabled, + * they will not cause another interrupt to be generated when + * interrupts are re-enabled. + * Check for this case here. This has been seen to happen + * around 30% of the time under heavy network load. + */ status = macb_readl(bp, RSR); if (status) { if (bp->caps & MACB_CAPS_ISR_CLEAR_ON_WRITE) @@ -1581,6 +1588,22 @@ static int macb_poll(struct napi_struct *napi, int budget) napi_reschedule(napi); } else { queue_writel(queue, IER, bp->rx_intr_mask); + + /* In rare cases, packets could have been received in + * the window between the check above and re-enabling + * interrupts. Therefore, a double-check is required + * to avoid losing a wakeup. This can potentially race + * with the interrupt handler doing the same actions + * if an interrupt is raised just after enabling them, + * but this should be harmless. + */ + status = macb_readl(bp, RSR); + if (unlikely(status)) { + queue_writel(queue, IDR, bp->rx_intr_mask); + if (bp->caps & MACB_CAPS_ISR_CLEAR_ON_WRITE) + queue_writel(queue, ISR, MACB_BIT(RCOMP)); + napi_schedule(napi); + } } } -- cgit v1.2.3-70-g09d2 From e0058f0fa80f6e09c4d363779c241c45a3c56b94 Mon Sep 17 00:00:00 2001 From: Jia-Ju Bai Date: Sat, 5 Mar 2022 01:14:11 -0800 Subject: net: qlogic: check the return value of dma_alloc_coherent() in qed_vf_hw_prepare() The function dma_alloc_coherent() in qed_vf_hw_prepare() can fail, so its return value should be checked. Fixes: 1408cc1fa48c ("qed: Introduce VFs") Reported-by: TOTE Robot Signed-off-by: Jia-Ju Bai Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_vf.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/qlogic/qed/qed_vf.c b/drivers/net/ethernet/qlogic/qed/qed_vf.c index 597cd9cd57b5..7b0e390c0b07 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_vf.c +++ b/drivers/net/ethernet/qlogic/qed/qed_vf.c @@ -513,6 +513,9 @@ int qed_vf_hw_prepare(struct qed_hwfn *p_hwfn) p_iov->bulletin.size, &p_iov->bulletin.phys, GFP_KERNEL); + if (!p_iov->bulletin.p_virt) + goto free_pf2vf_reply; + DP_VERBOSE(p_hwfn, QED_MSG_IOV, "VF's bulletin Board [%p virt 0x%llx phys 0x%08x bytes]\n", p_iov->bulletin.p_virt, @@ -552,6 +555,10 @@ int qed_vf_hw_prepare(struct qed_hwfn *p_hwfn) return rc; +free_pf2vf_reply: + dma_free_coherent(&p_hwfn->cdev->pdev->dev, + sizeof(union pfvf_tlvs), + p_iov->pf2vf_reply, p_iov->pf2vf_reply_phys); free_vf2pf_request: dma_free_coherent(&p_hwfn->cdev->pdev->dev, sizeof(union vfpf_tlvs), -- cgit v1.2.3-70-g09d2 From dd830aed23c6e07cd8e2a163742bf3d63c9add08 Mon Sep 17 00:00:00 2001 From: Aleksander Jan Bajkowski Date: Sat, 5 Mar 2022 12:20:39 +0100 Subject: net: lantiq_xrx200: fix use after free bug The skb->len field is read after the packet is sent to the network stack. In the meantime, skb can be freed. This patch fixes this bug. Fixes: c3e6b2c35b34 ("net: lantiq_xrx200: add ingress SG DMA support") Reported-by: Eric Dumazet Signed-off-by: Aleksander Jan Bajkowski Acked-by: Hauke Mehrtens Signed-off-by: David S. Miller --- drivers/net/ethernet/lantiq_xrx200.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/lantiq_xrx200.c b/drivers/net/ethernet/lantiq_xrx200.c index 41d11137cde0..5712c3e94be8 100644 --- a/drivers/net/ethernet/lantiq_xrx200.c +++ b/drivers/net/ethernet/lantiq_xrx200.c @@ -260,9 +260,9 @@ static int xrx200_hw_receive(struct xrx200_chan *ch) if (ctl & LTQ_DMA_EOP) { ch->skb_head->protocol = eth_type_trans(ch->skb_head, net_dev); - netif_receive_skb(ch->skb_head); net_dev->stats.rx_packets++; net_dev->stats.rx_bytes += ch->skb_head->len; + netif_receive_skb(ch->skb_head); ch->skb_head = NULL; ch->skb_tail = NULL; ret = XRX200_DMA_PACKET_COMPLETE; -- cgit v1.2.3-70-g09d2 From bb77bd31c281f70ec77c9c4f584950a779e05cf8 Mon Sep 17 00:00:00 2001 From: Zheyu Ma Date: Sat, 5 Mar 2022 14:55:04 +0000 Subject: ethernet: sun: Free the coherent when failing in probing When the driver fails to register net device, it should free the DMA region first, and then do other cleanup. Signed-off-by: Zheyu Ma Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller --- drivers/net/ethernet/sun/sunhme.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/sun/sunhme.c b/drivers/net/ethernet/sun/sunhme.c index ad9029ae6848..77e5dffb558f 100644 --- a/drivers/net/ethernet/sun/sunhme.c +++ b/drivers/net/ethernet/sun/sunhme.c @@ -3146,7 +3146,7 @@ static int happy_meal_pci_probe(struct pci_dev *pdev, if (err) { printk(KERN_ERR "happymeal(PCI): Cannot register net device, " "aborting.\n"); - goto err_out_iounmap; + goto err_out_free_coherent; } pci_set_drvdata(pdev, hp); @@ -3179,6 +3179,10 @@ static int happy_meal_pci_probe(struct pci_dev *pdev, return 0; +err_out_free_coherent: + dma_free_coherent(hp->dma_dev, PAGE_SIZE, + hp->happy_block, hp->hblock_dvma); + err_out_iounmap: iounmap(hp->gregs); -- cgit v1.2.3-70-g09d2 From d9dc0c84ad2d4cc911ba252c973d1bf18d5eb9cf Mon Sep 17 00:00:00 2001 From: Tom Rix Date: Sat, 5 Mar 2022 07:06:42 -0800 Subject: qed: return status of qed_iov_get_link Clang static analysis reports this issue qed_sriov.c:4727:19: warning: Assigned value is garbage or undefined ivi->max_tx_rate = tx_rate ? tx_rate : link.speed; ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ link is only sometimes set by the call to qed_iov_get_link() qed_iov_get_link fails without setting link or returning status. So change the decl to return status. Fixes: 73390ac9d82b ("qed*: support ndo_get_vf_config") Signed-off-by: Tom Rix Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_sriov.c | 18 +++++++++++------- 1 file changed, 11 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/qlogic/qed/qed_sriov.c b/drivers/net/ethernet/qlogic/qed/qed_sriov.c index 8ac38828ba45..48cf4355bc47 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_sriov.c +++ b/drivers/net/ethernet/qlogic/qed/qed_sriov.c @@ -3806,11 +3806,11 @@ bool qed_iov_mark_vf_flr(struct qed_hwfn *p_hwfn, u32 *p_disabled_vfs) return found; } -static void qed_iov_get_link(struct qed_hwfn *p_hwfn, - u16 vfid, - struct qed_mcp_link_params *p_params, - struct qed_mcp_link_state *p_link, - struct qed_mcp_link_capabilities *p_caps) +static int qed_iov_get_link(struct qed_hwfn *p_hwfn, + u16 vfid, + struct qed_mcp_link_params *p_params, + struct qed_mcp_link_state *p_link, + struct qed_mcp_link_capabilities *p_caps) { struct qed_vf_info *p_vf = qed_iov_get_vf_info(p_hwfn, vfid, @@ -3818,7 +3818,7 @@ static void qed_iov_get_link(struct qed_hwfn *p_hwfn, struct qed_bulletin_content *p_bulletin; if (!p_vf) - return; + return -EINVAL; p_bulletin = p_vf->bulletin.p_virt; @@ -3828,6 +3828,7 @@ static void qed_iov_get_link(struct qed_hwfn *p_hwfn, __qed_vf_get_link_state(p_hwfn, p_link, p_bulletin); if (p_caps) __qed_vf_get_link_caps(p_hwfn, p_caps, p_bulletin); + return 0; } static int @@ -4686,6 +4687,7 @@ static int qed_get_vf_config(struct qed_dev *cdev, struct qed_public_vf_info *vf_info; struct qed_mcp_link_state link; u32 tx_rate; + int ret; /* Sanitize request */ if (IS_VF(cdev)) @@ -4699,7 +4701,9 @@ static int qed_get_vf_config(struct qed_dev *cdev, vf_info = qed_iov_get_public_vf_info(hwfn, vf_id, true); - qed_iov_get_link(hwfn, vf_id, NULL, &link, NULL); + ret = qed_iov_get_link(hwfn, vf_id, NULL, &link, NULL); + if (ret) + return ret; /* Fill information about VF */ ivi->vf = vf_id; -- cgit v1.2.3-70-g09d2 From 2cf29e55894886965722e6625f6a03630b4db31d Mon Sep 17 00:00:00 2001 From: Michal Maloszewski Date: Mon, 24 Jan 2022 13:35:43 +0000 Subject: iavf: Fix handling of vlan strip virtual channel messages Modify netdev->features for vlan stripping based on virtual channel messages received from the PF. Change is needed to synchronize vlan strip status between PF sysfs and iavf ethtool. Fixes: 5951a2b9812d ("iavf: Fix VLAN feature flags after VFR") Signed-off-by: Norbert Ciosek Signed-off-by: Michal Maloszewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 40 +++++++++++++++++++++++++ 1 file changed, 40 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c index 88844d68e150..5263cefe46f5 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c @@ -1834,6 +1834,22 @@ void iavf_request_reset(struct iavf_adapter *adapter) adapter->current_op = VIRTCHNL_OP_UNKNOWN; } +/** + * iavf_netdev_features_vlan_strip_set - update vlan strip status + * @netdev: ptr to netdev being adjusted + * @enable: enable or disable vlan strip + * + * Helper function to change vlan strip status in netdev->features. + */ +static void iavf_netdev_features_vlan_strip_set(struct net_device *netdev, + const bool enable) +{ + if (enable) + netdev->features |= NETIF_F_HW_VLAN_CTAG_RX; + else + netdev->features &= ~NETIF_F_HW_VLAN_CTAG_RX; +} + /** * iavf_virtchnl_completion * @adapter: adapter structure @@ -2057,8 +2073,18 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter, } break; case VIRTCHNL_OP_ENABLE_VLAN_STRIPPING: + dev_warn(&adapter->pdev->dev, "Changing VLAN Stripping is not allowed when Port VLAN is configured\n"); + /* Vlan stripping could not be enabled by ethtool. + * Disable it in netdev->features. + */ + iavf_netdev_features_vlan_strip_set(netdev, false); + break; case VIRTCHNL_OP_DISABLE_VLAN_STRIPPING: dev_warn(&adapter->pdev->dev, "Changing VLAN Stripping is not allowed when Port VLAN is configured\n"); + /* Vlan stripping could not be disabled by ethtool. + * Enable it in netdev->features. + */ + iavf_netdev_features_vlan_strip_set(netdev, true); break; default: dev_err(&adapter->pdev->dev, "PF returned error %d (%s) to our request %d\n", @@ -2312,6 +2338,20 @@ void iavf_virtchnl_completion(struct iavf_adapter *adapter, spin_unlock_bh(&adapter->adv_rss_lock); } break; + case VIRTCHNL_OP_ENABLE_VLAN_STRIPPING: + /* PF enabled vlan strip on this VF. + * Update netdev->features if needed to be in sync with ethtool. + */ + if (!v_retval) + iavf_netdev_features_vlan_strip_set(netdev, true); + break; + case VIRTCHNL_OP_DISABLE_VLAN_STRIPPING: + /* PF disabled vlan strip on this VF. + * Update netdev->features if needed to be in sync with ethtool. + */ + if (!v_retval) + iavf_netdev_features_vlan_strip_set(netdev, false); + break; default: if (adapter->current_op && (v_opcode != adapter->current_op)) dev_warn(&adapter->pdev->dev, "Expected response %d from PF, received %d\n", -- cgit v1.2.3-70-g09d2 From 57d03f5608c34079f6f15031f4e8b1e2ae95dcb0 Mon Sep 17 00:00:00 2001 From: Michal Maloszewski Date: Wed, 2 Feb 2022 12:44:54 +0000 Subject: iavf: Fix adopting new combined setting In some cases overloaded flag IAVF_FLAG_REINIT_ITR_NEEDED which should indicate that interrupts need to be completely reinitialized during reset leads to RTNL deadlocks using ethtool -C while a reset is in progress. To fix, it was added a new flag IAVF_FLAG_REINIT_MSIX_NEEDED used to trigger MSI-X reinit. New combined setting is fixed adopt after VF reset. This has been implemented by call reinit interrupt scheme during VF reset. Without this fix new combined setting has never been adopted. Fixes: 209f2f9c7181 ("iavf: Add support for VIRTCHNL_VF_OFFLOAD_VLAN_V2 negotiation") Signed-off-by: Grzegorz Szczurek Signed-off-by: Mitch Williams Signed-off-by: Michal Maloszewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf.h | 1 + drivers/net/ethernet/intel/iavf/iavf_main.c | 13 +++++++++---- 2 files changed, 10 insertions(+), 4 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/iavf/iavf.h b/drivers/net/ethernet/intel/iavf/iavf.h index 89423947ee65..4babe4705a55 100644 --- a/drivers/net/ethernet/intel/iavf/iavf.h +++ b/drivers/net/ethernet/intel/iavf/iavf.h @@ -288,6 +288,7 @@ struct iavf_adapter { #define IAVF_FLAG_REINIT_ITR_NEEDED BIT(16) #define IAVF_FLAG_QUEUES_DISABLED BIT(17) #define IAVF_FLAG_SETUP_NETDEV_FEATURES BIT(18) +#define IAVF_FLAG_REINIT_MSIX_NEEDED BIT(20) /* duplicates for common code */ #define IAVF_FLAG_DCB_ENABLED 0 /* flags for admin queue service task */ diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index dcf24264c7ea..8e644e9ed8da 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -2120,7 +2120,7 @@ int iavf_parse_vf_resource_msg(struct iavf_adapter *adapter) "Requested %d queues, but PF only gave us %d.\n", num_req_queues, adapter->vsi_res->num_queue_pairs); - adapter->flags |= IAVF_FLAG_REINIT_ITR_NEEDED; + adapter->flags |= IAVF_FLAG_REINIT_MSIX_NEEDED; adapter->num_req_queues = adapter->vsi_res->num_queue_pairs; iavf_schedule_reset(adapter); @@ -2727,7 +2727,8 @@ continue_reset: err); adapter->aq_required = 0; - if (adapter->flags & IAVF_FLAG_REINIT_ITR_NEEDED) { + if ((adapter->flags & IAVF_FLAG_REINIT_MSIX_NEEDED) || + (adapter->flags & IAVF_FLAG_REINIT_ITR_NEEDED)) { err = iavf_reinit_interrupt_scheme(adapter); if (err) goto reset_err; @@ -2799,12 +2800,13 @@ continue_reset: if (err) goto reset_err; - if (adapter->flags & IAVF_FLAG_REINIT_ITR_NEEDED) { + if ((adapter->flags & IAVF_FLAG_REINIT_MSIX_NEEDED) || + (adapter->flags & IAVF_FLAG_REINIT_ITR_NEEDED)) { err = iavf_request_traffic_irqs(adapter, netdev->name); if (err) goto reset_err; - adapter->flags &= ~IAVF_FLAG_REINIT_ITR_NEEDED; + adapter->flags &= ~IAVF_FLAG_REINIT_MSIX_NEEDED; } iavf_configure(adapter); @@ -2819,6 +2821,9 @@ continue_reset: iavf_change_state(adapter, __IAVF_DOWN); wake_up(&adapter->down_waitqueue); } + + adapter->flags &= ~IAVF_FLAG_REINIT_ITR_NEEDED; + mutex_unlock(&adapter->client_lock); mutex_unlock(&adapter->crit_lock); -- cgit v1.2.3-70-g09d2 From 5710ab79166504013f7c0ae6a57e7d2fd26e5c43 Mon Sep 17 00:00:00 2001 From: Jacob Keller Date: Wed, 16 Feb 2022 16:51:35 -0800 Subject: i40e: stop disabling VFs due to PF error responses The i40e_vc_send_msg_to_vf_ex (and its wrapper i40e_vc_send_msg_to_vf) function has logic to detect "failure" responses sent to the VF. If a VF is sent more than I40E_DEFAULT_NUM_INVALID_MSGS_ALLOWED, then the VF is marked as disabled. In either case, a dev_info message is printed stating that a VF opcode failed. This logic originates from the early implementation of VF support in commit 5c3c48ac6bf5 ("i40e: implement virtual device interface"). That commit did not go far enough. The "logic" for this behavior seems to be that error responses somehow indicate a malicious VF. This is not really true. The PF might be sending an error for any number of reasons such as lacking resources, an unsupported operation, etc. This does not indicate a malicious VF. We already have a separate robust malicious VF detection which relies on hardware logic to detect and prevent a variety of behaviors. There is no justification for this behavior in the original implementation. In fact, a later commit 18b7af57d9c1 ("i40e: Lower some message levels") reduced the opcode failure message from a dev_err to a dev_info. In addition, recent commit 01cbf50877e6 ("i40e: Fix to not show opcode msg on unsuccessful VF MAC change") changed the logic to allow quieting it for expected failures. That commit prevented this logic from kicking in for specific circumstances. This change did not go far enough. The behavior is not documented nor is it part of any requirement for our products. Other operating systems such as the FreeBSD implementation of our driver do not include this logic. It is clear this check does not make sense, and causes problems which led to ugly workarounds. Fix this by just removing the entire logic and the need for the i40e_vc_send_msg_to_vf_ex function. Fixes: 01cbf50877e6 ("i40e: Fix to not show opcode msg on unsuccessful VF MAC change") Fixes: 5c3c48ac6bf5 ("i40e: implement virtual device interface") Signed-off-by: Jacob Keller Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 6 +-- drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 57 +++------------------- drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h | 5 -- 3 files changed, 9 insertions(+), 59 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c index 1e57cc8c47d7..9db5001297c7 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c +++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c @@ -742,10 +742,8 @@ static void i40e_dbg_dump_vf(struct i40e_pf *pf, int vf_id) vsi = pf->vsi[vf->lan_vsi_idx]; dev_info(&pf->pdev->dev, "vf %2d: VSI id=%d, seid=%d, qps=%d\n", vf_id, vf->lan_vsi_id, vsi->seid, vf->num_queue_pairs); - dev_info(&pf->pdev->dev, " num MDD=%lld, invalid msg=%lld, valid msg=%lld\n", - vf->num_mdd_events, - vf->num_invalid_msgs, - vf->num_valid_msgs); + dev_info(&pf->pdev->dev, " num MDD=%lld\n", + vf->num_mdd_events); } else { dev_info(&pf->pdev->dev, "invalid VF id %d\n", vf_id); } diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c index dfdb6e786461..2606e8f0f19b 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c @@ -1917,19 +1917,17 @@ sriov_configure_out: /***********************virtual channel routines******************/ /** - * i40e_vc_send_msg_to_vf_ex + * i40e_vc_send_msg_to_vf * @vf: pointer to the VF info * @v_opcode: virtual channel opcode * @v_retval: virtual channel return value * @msg: pointer to the msg buffer * @msglen: msg length - * @is_quiet: true for not printing unsuccessful return values, false otherwise * * send msg to VF **/ -static int i40e_vc_send_msg_to_vf_ex(struct i40e_vf *vf, u32 v_opcode, - u32 v_retval, u8 *msg, u16 msglen, - bool is_quiet) +static int i40e_vc_send_msg_to_vf(struct i40e_vf *vf, u32 v_opcode, + u32 v_retval, u8 *msg, u16 msglen) { struct i40e_pf *pf; struct i40e_hw *hw; @@ -1944,25 +1942,6 @@ static int i40e_vc_send_msg_to_vf_ex(struct i40e_vf *vf, u32 v_opcode, hw = &pf->hw; abs_vf_id = vf->vf_id + hw->func_caps.vf_base_id; - /* single place to detect unsuccessful return values */ - if (v_retval && !is_quiet) { - vf->num_invalid_msgs++; - dev_info(&pf->pdev->dev, "VF %d failed opcode %d, retval: %d\n", - vf->vf_id, v_opcode, v_retval); - if (vf->num_invalid_msgs > - I40E_DEFAULT_NUM_INVALID_MSGS_ALLOWED) { - dev_err(&pf->pdev->dev, - "Number of invalid messages exceeded for VF %d\n", - vf->vf_id); - dev_err(&pf->pdev->dev, "Use PF Control I/F to enable the VF\n"); - set_bit(I40E_VF_STATE_DISABLED, &vf->vf_states); - } - } else { - vf->num_valid_msgs++; - /* reset the invalid counter, if a valid message is received. */ - vf->num_invalid_msgs = 0; - } - aq_ret = i40e_aq_send_msg_to_vf(hw, abs_vf_id, v_opcode, v_retval, msg, msglen, NULL); if (aq_ret) { @@ -1975,23 +1954,6 @@ static int i40e_vc_send_msg_to_vf_ex(struct i40e_vf *vf, u32 v_opcode, return 0; } -/** - * i40e_vc_send_msg_to_vf - * @vf: pointer to the VF info - * @v_opcode: virtual channel opcode - * @v_retval: virtual channel return value - * @msg: pointer to the msg buffer - * @msglen: msg length - * - * send msg to VF - **/ -static int i40e_vc_send_msg_to_vf(struct i40e_vf *vf, u32 v_opcode, - u32 v_retval, u8 *msg, u16 msglen) -{ - return i40e_vc_send_msg_to_vf_ex(vf, v_opcode, v_retval, - msg, msglen, false); -} - /** * i40e_vc_send_resp_to_vf * @vf: pointer to the VF info @@ -2822,7 +2784,6 @@ error_param: * i40e_check_vf_permission * @vf: pointer to the VF info * @al: MAC address list from virtchnl - * @is_quiet: set true for printing msg without opcode info, false otherwise * * Check that the given list of MAC addresses is allowed. Will return -EPERM * if any address in the list is not valid. Checks the following conditions: @@ -2837,8 +2798,7 @@ error_param: * addresses might not be accurate. **/ static inline int i40e_check_vf_permission(struct i40e_vf *vf, - struct virtchnl_ether_addr_list *al, - bool *is_quiet) + struct virtchnl_ether_addr_list *al) { struct i40e_pf *pf = vf->pf; struct i40e_vsi *vsi = pf->vsi[vf->lan_vsi_idx]; @@ -2846,7 +2806,6 @@ static inline int i40e_check_vf_permission(struct i40e_vf *vf, int mac2add_cnt = 0; int i; - *is_quiet = false; for (i = 0; i < al->num_elements; i++) { struct i40e_mac_filter *f; u8 *addr = al->list[i].addr; @@ -2870,7 +2829,6 @@ static inline int i40e_check_vf_permission(struct i40e_vf *vf, !ether_addr_equal(addr, vf->default_lan_addr.addr)) { dev_err(&pf->pdev->dev, "VF attempting to override administratively set MAC address, bring down and up the VF interface to resume normal operation\n"); - *is_quiet = true; return -EPERM; } @@ -2921,7 +2879,6 @@ static int i40e_vc_add_mac_addr_msg(struct i40e_vf *vf, u8 *msg) (struct virtchnl_ether_addr_list *)msg; struct i40e_pf *pf = vf->pf; struct i40e_vsi *vsi = NULL; - bool is_quiet = false; i40e_status ret = 0; int i; @@ -2938,7 +2895,7 @@ static int i40e_vc_add_mac_addr_msg(struct i40e_vf *vf, u8 *msg) */ spin_lock_bh(&vsi->mac_filter_hash_lock); - ret = i40e_check_vf_permission(vf, al, &is_quiet); + ret = i40e_check_vf_permission(vf, al); if (ret) { spin_unlock_bh(&vsi->mac_filter_hash_lock); goto error_param; @@ -2976,8 +2933,8 @@ static int i40e_vc_add_mac_addr_msg(struct i40e_vf *vf, u8 *msg) error_param: /* send the response to the VF */ - return i40e_vc_send_msg_to_vf_ex(vf, VIRTCHNL_OP_ADD_ETH_ADDR, - ret, NULL, 0, is_quiet); + return i40e_vc_send_msg_to_vf(vf, VIRTCHNL_OP_ADD_ETH_ADDR, + ret, NULL, 0); } /** diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h index 03c42fd0fea1..a554d0a0b09b 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h @@ -10,8 +10,6 @@ #define I40E_VIRTCHNL_SUPPORTED_QTYPES 2 -#define I40E_DEFAULT_NUM_INVALID_MSGS_ALLOWED 10 - #define I40E_VLAN_PRIORITY_SHIFT 13 #define I40E_VLAN_MASK 0xFFF #define I40E_PRIORITY_MASK 0xE000 @@ -92,9 +90,6 @@ struct i40e_vf { u8 num_queue_pairs; /* num of qps assigned to VF vsis */ u8 num_req_queues; /* num of requested qps */ u64 num_mdd_events; /* num of mdd events detected */ - /* num of continuous malformed or invalid msgs detected */ - u64 num_invalid_msgs; - u64 num_valid_msgs; /* num of valid msgs detected */ unsigned long vf_caps; /* vf's adv. capabilities */ unsigned long vf_states; /* vf's runtime states */ -- cgit v1.2.3-70-g09d2 From 79498d5af8e458102242d1667cf44df1f1564e63 Mon Sep 17 00:00:00 2001 From: Jacob Keller Date: Wed, 16 Feb 2022 16:51:36 -0800 Subject: ice: stop disabling VFs due to PF error responses The ice_vc_send_msg_to_vf function has logic to detect "failure" responses being sent to a VF. If a VF is sent more than ICE_DFLT_NUM_INVAL_MSGS_ALLOWED then the VF is marked as disabled. Almost identical logic also existed in the i40e driver. This logic was added to the ice driver in commit 1071a8358a28 ("ice: Implement virtchnl commands for AVF support") which itself copied from the i40e implementation in commit 5c3c48ac6bf5 ("i40e: implement virtual device interface"). Neither commit provides a proper explanation or justification of the check. In fact, later commits to i40e changed the logic to allow bypassing the check in some specific instances. The "logic" for this seems to be that error responses somehow indicate a malicious VF. This is not really true. The PF might be sending an error for any number of reasons such as lack of resources, etc. Additionally, this causes the PF to log an info message for every failed VF response which may confuse users, and can spam the kernel log. This behavior is not documented as part of any requirement for our products and other operating system drivers such as the FreeBSD implementation of our drivers do not include this type of check. In fact, the change from dev_err to dev_info in i40e commit 18b7af57d9c1 ("i40e: Lower some message levels") explains that these messages typically don't actually indicate a real issue. It is quite likely that a user who hits this in practice will be very confused as the VF will be disabled without an obvious way to recover. We already have robust malicious driver detection logic using actual hardware detection mechanisms that detect and prevent invalid device usage. Remove the logic since its not a documented requirement and the behavior is not intuitive. Fixes: 1071a8358a28 ("ice: Implement virtchnl commands for AVF support") Signed-off-by: Jacob Keller Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 18 ------------------ drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h | 3 --- 2 files changed, 21 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c index 408f78e3eb13..1be3cd4b2bef 100644 --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c @@ -2182,24 +2182,6 @@ ice_vc_send_msg_to_vf(struct ice_vf *vf, u32 v_opcode, dev = ice_pf_to_dev(pf); - /* single place to detect unsuccessful return values */ - if (v_retval) { - vf->num_inval_msgs++; - dev_info(dev, "VF %d failed opcode %d, retval: %d\n", vf->vf_id, - v_opcode, v_retval); - if (vf->num_inval_msgs > ICE_DFLT_NUM_INVAL_MSGS_ALLOWED) { - dev_err(dev, "Number of invalid messages exceeded for VF %d\n", - vf->vf_id); - dev_err(dev, "Use PF Control I/F to enable the VF\n"); - set_bit(ICE_VF_STATE_DIS, vf->vf_states); - return -EIO; - } - } else { - vf->num_valid_msgs++; - /* reset the invalid counter, if a valid message is received. */ - vf->num_inval_msgs = 0; - } - aq_ret = ice_aq_send_msg_to_vf(&pf->hw, vf->vf_id, v_opcode, v_retval, msg, msglen, NULL); if (aq_ret && pf->hw.mailboxq.sq_last_status != ICE_AQ_RC_ENOSYS) { diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h index 752487a1bdd6..8f27255cc0cc 100644 --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h @@ -14,7 +14,6 @@ #define ICE_MAX_MACADDR_PER_VF 18 /* Malicious Driver Detection */ -#define ICE_DFLT_NUM_INVAL_MSGS_ALLOWED 10 #define ICE_MDD_EVENTS_THRESHOLD 30 /* Static VF transaction/status register def */ @@ -134,8 +133,6 @@ struct ice_vf { unsigned int max_tx_rate; /* Maximum Tx bandwidth limit in Mbps */ DECLARE_BITMAP(vf_states, ICE_VF_STATES_NBITS); /* VF runtime states */ - u64 num_inval_msgs; /* number of continuous invalid msgs */ - u64 num_valid_msgs; /* number of valid msgs detected */ unsigned long vf_caps; /* VF's adv. capabilities */ u8 num_req_qs; /* num of queue pairs requested by VF */ u16 num_mac; -- cgit v1.2.3-70-g09d2 From 97b0129146b1544bbb0773585327896da3bb4e0a Mon Sep 17 00:00:00 2001 From: Dave Ertman Date: Fri, 18 Feb 2022 12:39:25 -0800 Subject: ice: Fix error with handling of bonding MTU When a bonded interface is destroyed, .ndo_change_mtu can be called during the tear-down process while the RTNL lock is held. This is a problem since the auxiliary driver linked to the LAN driver needs to be notified of the MTU change, and this requires grabbing a device_lock on the auxiliary_device's dev. Currently this is being attempted in the same execution context as the call to .ndo_change_mtu which is causing a dead-lock. Move the notification of the changed MTU to a separate execution context (watchdog service task) and eliminate the "before" notification. Fixes: 348048e724a0e ("ice: Implement iidc operations") Signed-off-by: Dave Ertman Tested-by: Jonathan Toppins Tested-by: Gurucharan G (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice.h | 1 + drivers/net/ethernet/intel/ice/ice_main.c | 29 ++++++++++++++--------------- 2 files changed, 15 insertions(+), 15 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h index 473b1f6be9de..3121f9b04f59 100644 --- a/drivers/net/ethernet/intel/ice/ice.h +++ b/drivers/net/ethernet/intel/ice/ice.h @@ -483,6 +483,7 @@ enum ice_pf_flags { ICE_FLAG_MDD_AUTO_RESET_VF, ICE_FLAG_LINK_LENIENT_MODE_ENA, ICE_FLAG_PLUG_AUX_DEV, + ICE_FLAG_MTU_CHANGED, ICE_PF_FLAGS_NBITS /* must be last */ }; diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index f3c346e13b7a..6fc6514eb007 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -2258,6 +2258,17 @@ static void ice_service_task(struct work_struct *work) if (test_and_clear_bit(ICE_FLAG_PLUG_AUX_DEV, pf->flags)) ice_plug_aux_dev(pf); + if (test_and_clear_bit(ICE_FLAG_MTU_CHANGED, pf->flags)) { + struct iidc_event *event; + + event = kzalloc(sizeof(*event), GFP_KERNEL); + if (event) { + set_bit(IIDC_EVENT_AFTER_MTU_CHANGE, event->type); + ice_send_event_to_aux(pf, event); + kfree(event); + } + } + ice_clean_adminq_subtask(pf); ice_check_media_subtask(pf); ice_check_for_hang_subtask(pf); @@ -6822,7 +6833,6 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu) struct ice_netdev_priv *np = netdev_priv(netdev); struct ice_vsi *vsi = np->vsi; struct ice_pf *pf = vsi->back; - struct iidc_event *event; u8 count = 0; int err = 0; @@ -6857,14 +6867,6 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu) return -EBUSY; } - event = kzalloc(sizeof(*event), GFP_KERNEL); - if (!event) - return -ENOMEM; - - set_bit(IIDC_EVENT_BEFORE_MTU_CHANGE, event->type); - ice_send_event_to_aux(pf, event); - clear_bit(IIDC_EVENT_BEFORE_MTU_CHANGE, event->type); - netdev->mtu = (unsigned int)new_mtu; /* if VSI is up, bring it down and then back up */ @@ -6872,21 +6874,18 @@ static int ice_change_mtu(struct net_device *netdev, int new_mtu) err = ice_down(vsi); if (err) { netdev_err(netdev, "change MTU if_down err %d\n", err); - goto event_after; + return err; } err = ice_up(vsi); if (err) { netdev_err(netdev, "change MTU if_up err %d\n", err); - goto event_after; + return err; } } netdev_dbg(netdev, "changed MTU to %d\n", new_mtu); -event_after: - set_bit(IIDC_EVENT_AFTER_MTU_CHANGE, event->type); - ice_send_event_to_aux(pf, event); - kfree(event); + set_bit(ICE_FLAG_MTU_CHANGED, pf->flags); return err; } -- cgit v1.2.3-70-g09d2 From 3d97f1afd8d831e0c0dc1157418f94b8faa97b54 Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Sun, 16 Jan 2022 19:46:20 +0100 Subject: ice: Don't use GFP_KERNEL in atomic context ice_misc_intr() is an irq handler. It should not sleep. Use GFP_ATOMIC instead of GFP_KERNEL when allocating some memory. Fixes: 348048e724a0 ("ice: Implement iidc operations") Signed-off-by: Christophe JAILLET Tested-by: Leszek Kaliszczuk Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 6fc6514eb007..83e3e8aae6cf 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -3034,7 +3034,7 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data) struct iidc_event *event; ena_mask &= ~ICE_AUX_CRIT_ERR; - event = kzalloc(sizeof(*event), GFP_KERNEL); + event = kzalloc(sizeof(*event), GFP_ATOMIC); if (event) { set_bit(IIDC_EVENT_CRIT_ERR, event->type); /* report the entire OICR value to AUX driver */ -- cgit v1.2.3-70-g09d2 From ad35ffa252af67d4cc7c744b9377a2b577748e3f Mon Sep 17 00:00:00 2001 From: Jedrzej Jagielski Date: Tue, 22 Feb 2022 11:43:04 +0000 Subject: ice: Fix curr_link_speed advertised speed Change curr_link_speed advertised speed, due to link_info.link_speed is not equal phy.curr_user_speed_req. Without this patch it is impossible to set advertised speed to same as link_speed. Testing Hints: Try to set advertised speed to 25G only with 25G default link (use ethtool -s 0x80000000) Fixes: 48cb27f2fd18 ("ice: Implement handlers for ethtool PHY/link operations") Signed-off-by: Grzegorz Siwik Signed-off-by: Jedrzej Jagielski Tested-by: Gurucharan (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_ethtool.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c index e2e3ef7fba7f..a5dc9824e255 100644 --- a/drivers/net/ethernet/intel/ice/ice_ethtool.c +++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c @@ -2298,7 +2298,7 @@ ice_set_link_ksettings(struct net_device *netdev, if (err) goto done; - curr_link_speed = pi->phy.link_info.link_speed; + curr_link_speed = pi->phy.curr_user_speed_req; adv_link_speed = ice_ksettings_find_adv_link_speed(ks); /* If speed didn't get set, set it to what it currently is. -- cgit v1.2.3-70-g09d2 From b19ab4b38b06aae12442b2de95ccf58b5dc53584 Mon Sep 17 00:00:00 2001 From: Miaoqian Lin Date: Tue, 8 Mar 2022 02:47:49 +0000 Subject: ethernet: Fix error handling in xemaclite_of_probe This node pointer is returned by of_parse_phandle() with refcount incremented in this function. Calling of_node_put() to avoid the refcount leak. As the remove function do. Fixes: 5cdaaa12866e ("net: emaclite: adding MDIO and phy lib support") Signed-off-by: Miaoqian Lin Reviewed-by: Andrew Lunn Link: https://lore.kernel.org/r/20220308024751.2320-1-linmq006@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/xilinx/xilinx_emaclite.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/xilinx/xilinx_emaclite.c b/drivers/net/ethernet/xilinx/xilinx_emaclite.c index 519599480b15..77fa2cb03aca 100644 --- a/drivers/net/ethernet/xilinx/xilinx_emaclite.c +++ b/drivers/net/ethernet/xilinx/xilinx_emaclite.c @@ -1183,7 +1183,7 @@ static int xemaclite_of_probe(struct platform_device *ofdev) if (rc) { dev_err(dev, "Cannot register network device, aborting\n"); - goto error; + goto put_node; } dev_info(dev, @@ -1191,6 +1191,8 @@ static int xemaclite_of_probe(struct platform_device *ofdev) (unsigned long __force)ndev->mem_start, lp->base_addr, ndev->irq); return 0; +put_node: + of_node_put(lp->phy_node); error: free_netdev(ndev); return rc; -- cgit v1.2.3-70-g09d2 From 6babfc6e6fab068018c36e8f6605184b8c0b349d Mon Sep 17 00:00:00 2001 From: Jiasheng Jiang Date: Tue, 8 Mar 2022 14:40:07 +0800 Subject: net: ethernet: ti: cpts: Handle error for clk_enable As the potential failure of the clk_enable(), it should be better to check it and return error if fails. Fixes: 8a2c9a5ab4b9 ("net: ethernet: ti: cpts: rework initialization/deinitialization") Signed-off-by: Jiasheng Jiang Signed-off-by: David S. Miller --- drivers/net/ethernet/ti/cpts.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/ti/cpts.c b/drivers/net/ethernet/ti/cpts.c index dc70a6bfaa6a..92ca739fac01 100644 --- a/drivers/net/ethernet/ti/cpts.c +++ b/drivers/net/ethernet/ti/cpts.c @@ -568,7 +568,9 @@ int cpts_register(struct cpts *cpts) for (i = 0; i < CPTS_MAX_EVENTS; i++) list_add(&cpts->pool_data[i].list, &cpts->pool); - clk_enable(cpts->refclk); + err = clk_enable(cpts->refclk); + if (err) + return err; cpts_write32(cpts, CPTS_EN, control); cpts_write32(cpts, TS_PEND_EN, int_enable); -- cgit v1.2.3-70-g09d2 From 2a760554dcba450d3ad61b32375b50ed6d59a87c Mon Sep 17 00:00:00 2001 From: "Minghao Chi (CGEL ZTE)" Date: Tue, 8 Mar 2022 06:43:09 +0000 Subject: net:mcf8390: Use platform_get_irq() to get the interrupt It is not recommened to use platform_get_resource(pdev, IORESOURCE_IRQ) for requesting IRQ's resources any more, as they can be not ready yet in case of DT-booting. platform_get_irq() instead is a recommended way for getting IRQ even if it was not retrieved earlier. It also makes code simpler because we're getting "int" value right away and no conversion from resource to int is required. Reported-by: Zeal Robot Signed-off-by: Minghao Chi (CGEL ZTE) Signed-off-by: David S. Miller --- drivers/net/ethernet/8390/mcf8390.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/8390/mcf8390.c b/drivers/net/ethernet/8390/mcf8390.c index e320cccba61a..90cd7bdf06f5 100644 --- a/drivers/net/ethernet/8390/mcf8390.c +++ b/drivers/net/ethernet/8390/mcf8390.c @@ -405,12 +405,12 @@ static int mcf8390_init(struct net_device *dev) static int mcf8390_probe(struct platform_device *pdev) { struct net_device *dev; - struct resource *mem, *irq; + struct resource *mem; resource_size_t msize; - int ret; + int ret, irq; - irq = platform_get_resource(pdev, IORESOURCE_IRQ, 0); - if (irq == NULL) { + irq = platform_get_irq(pdev, 0); + if (irq < 0) { dev_err(&pdev->dev, "no IRQ specified?\n"); return -ENXIO; } @@ -433,7 +433,7 @@ static int mcf8390_probe(struct platform_device *pdev) SET_NETDEV_DEV(dev, &pdev->dev); platform_set_drvdata(pdev, dev); - dev->irq = irq->start; + dev->irq = irq; dev->base_addr = mem->start; ret = mcf8390_init(dev); -- cgit v1.2.3-70-g09d2 From 2169b79258c8be803d2595d6456b1e77129fe154 Mon Sep 17 00:00:00 2001 From: Jiasheng Jiang Date: Tue, 8 Mar 2022 14:57:39 +0800 Subject: net: ethernet: lpc_eth: Handle error for clk_enable As the potential failure of the clk_enable(), it should be better to check it and return error if fails. Fixes: b7370112f519 ("lpc32xx: Added ethernet driver") Signed-off-by: Jiasheng Jiang Signed-off-by: David S. Miller --- drivers/net/ethernet/nxp/lpc_eth.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/nxp/lpc_eth.c b/drivers/net/ethernet/nxp/lpc_eth.c index bc39558fe82b..756f97dce85b 100644 --- a/drivers/net/ethernet/nxp/lpc_eth.c +++ b/drivers/net/ethernet/nxp/lpc_eth.c @@ -1471,6 +1471,7 @@ static int lpc_eth_drv_resume(struct platform_device *pdev) { struct net_device *ndev = platform_get_drvdata(pdev); struct netdata_local *pldat; + int ret; if (device_may_wakeup(&pdev->dev)) disable_irq_wake(ndev->irq); @@ -1480,7 +1481,9 @@ static int lpc_eth_drv_resume(struct platform_device *pdev) pldat = netdev_priv(ndev); /* Enable interface clock */ - clk_enable(pldat->clk); + ret = clk_enable(pldat->clk); + if (ret) + return ret; /* Reset and initialize */ __lpc_eth_reset(pldat); -- cgit v1.2.3-70-g09d2 From c9ffa3e2bc451816ce0295e40063514fabf2bd36 Mon Sep 17 00:00:00 2001 From: Miaoqian Lin Date: Tue, 8 Mar 2022 07:42:47 +0000 Subject: net: marvell: prestera: Add missing of_node_put() in prestera_switch_set_base_mac_addr This node pointer is returned by of_find_compatible_node() with refcount incremented. Calling of_node_put() to aovid the refcount leak. Fixes: 501ef3066c89 ("net: marvell: prestera: Add driver for Prestera family ASIC devices") Signed-off-by: Miaoqian Lin Signed-off-by: David S. Miller --- drivers/net/ethernet/marvell/prestera/prestera_main.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/marvell/prestera/prestera_main.c b/drivers/net/ethernet/marvell/prestera/prestera_main.c index cad93f747d0c..73cd0a4b7291 100644 --- a/drivers/net/ethernet/marvell/prestera/prestera_main.c +++ b/drivers/net/ethernet/marvell/prestera/prestera_main.c @@ -554,6 +554,7 @@ static int prestera_switch_set_base_mac_addr(struct prestera_switch *sw) dev_info(prestera_dev(sw), "using random base mac address\n"); } of_node_put(base_mac_np); + of_node_put(np); return prestera_hw_switch_mac_set(sw, sw->base_mac); } -- cgit v1.2.3-70-g09d2 From 063bd355595428750803d8736a9bb7c8db67d42d Mon Sep 17 00:00:00 2001 From: Moshe Shemesh Date: Fri, 4 Feb 2022 11:47:44 +0200 Subject: net/mlx5: Fix a race on command flush flow Fix a refcount use after free warning due to a race on command entry. Such race occurs when one of the commands releases its last refcount and frees its index and entry while another process running command flush flow takes refcount to this command entry. The process which handles commands flush may see this command as needed to be flushed if the other process released its refcount but didn't release the index yet. Fix it by adding the needed spin lock. It fixes the following warning trace: refcount_t: addition on 0; use-after-free. WARNING: CPU: 11 PID: 540311 at lib/refcount.c:25 refcount_warn_saturate+0x80/0xe0 ... RIP: 0010:refcount_warn_saturate+0x80/0xe0 ... Call Trace: mlx5_cmd_trigger_completions+0x293/0x340 [mlx5_core] mlx5_cmd_flush+0x3a/0xf0 [mlx5_core] enter_error_state+0x44/0x80 [mlx5_core] mlx5_fw_fatal_reporter_err_work+0x37/0xe0 [mlx5_core] process_one_work+0x1be/0x390 worker_thread+0x4d/0x3d0 ? rescuer_thread+0x350/0x350 kthread+0x141/0x160 ? set_kthread_struct+0x40/0x40 ret_from_fork+0x1f/0x30 Fixes: 50b2412b7e78 ("net/mlx5: Avoid possible free of command entry while timeout comp handler") Signed-off-by: Moshe Shemesh Reviewed-by: Eran Ben Elisha Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index 17fe05809653..3eacd8739929 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -131,11 +131,8 @@ static int cmd_alloc_index(struct mlx5_cmd *cmd) static void cmd_free_index(struct mlx5_cmd *cmd, int idx) { - unsigned long flags; - - spin_lock_irqsave(&cmd->alloc_lock, flags); + lockdep_assert_held(&cmd->alloc_lock); set_bit(idx, &cmd->bitmask); - spin_unlock_irqrestore(&cmd->alloc_lock, flags); } static void cmd_ent_get(struct mlx5_cmd_work_ent *ent) @@ -145,17 +142,21 @@ static void cmd_ent_get(struct mlx5_cmd_work_ent *ent) static void cmd_ent_put(struct mlx5_cmd_work_ent *ent) { + struct mlx5_cmd *cmd = ent->cmd; + unsigned long flags; + + spin_lock_irqsave(&cmd->alloc_lock, flags); if (!refcount_dec_and_test(&ent->refcnt)) - return; + goto out; if (ent->idx >= 0) { - struct mlx5_cmd *cmd = ent->cmd; - cmd_free_index(cmd, ent->idx); up(ent->page_queue ? &cmd->pages_sem : &cmd->sem); } cmd_free_ent(ent); +out: + spin_unlock_irqrestore(&cmd->alloc_lock, flags); } static struct mlx5_cmd_layout *get_inst(struct mlx5_cmd *cmd, int idx) -- cgit v1.2.3-70-g09d2 From 39bab83b119faac4bf7f07173a42ed35be95147e Mon Sep 17 00:00:00 2001 From: Dima Chumak Date: Mon, 17 Jan 2022 15:32:16 +0200 Subject: net/mlx5: Fix offloading with ESWITCH_IPV4_TTL_MODIFY_ENABLE Only prio 1 is supported for nic mode when there is no ignore flow level support in firmware. But for switchdev mode, which supports fixed number of statically pre-allocated prios, this restriction is not relevant so it can be relaxed. Fixes: d671e109bd85 ("net/mlx5: Fix tc max supported prio for nic mode") Signed-off-by: Dima Chumak Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c index 1e8ec4f236b2..df58cba37930 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c @@ -121,9 +121,6 @@ u32 mlx5_chains_get_nf_ft_chain(struct mlx5_fs_chains *chains) u32 mlx5_chains_get_prio_range(struct mlx5_fs_chains *chains) { - if (!mlx5_chains_prios_supported(chains)) - return 1; - if (mlx5_chains_ignore_flow_level_supported(chains)) return UINT_MAX; -- cgit v1.2.3-70-g09d2 From ad11c4f1d8fd1f03639460e425a36f7fd0ea83f5 Mon Sep 17 00:00:00 2001 From: Roi Dayan Date: Wed, 16 Feb 2022 13:56:57 +0200 Subject: net/mlx5e: Lag, Only handle events from highest priority multipath entry There could be multiple multipath entries but changing the port affinity for each one doesn't make much sense and there should be a default one. So only track the entry with lowest priority value. The commit doesn't affect existing users with a single entry. Fixes: 544fe7c2e654 ("net/mlx5e: Activate HW multipath and handle port affinity based on FIB events") Signed-off-by: Roi Dayan Reviewed-by: Maor Dickman Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c index 1ca01a5b6cdd..626aa60b6099 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c @@ -126,6 +126,10 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev, return; } + /* Handle multipath entry with lower priority value */ + if (mp->mfi && mp->mfi != fi && fi->fib_priority >= mp->mfi->fib_priority) + return; + /* Handle add/replace event */ nhs = fib_info_num_path(fi); if (nhs == 1) { @@ -135,12 +139,13 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev, int i = mlx5_lag_dev_get_netdev_idx(ldev, nh_dev); if (i < 0) - i = MLX5_LAG_NORMAL_AFFINITY; - else - ++i; + return; + i++; mlx5_lag_set_port_affinity(ldev, i); } + + mp->mfi = fi; return; } -- cgit v1.2.3-70-g09d2 From 99a2b9be077ae3a5d97fbf5f7782e0f2e9812978 Mon Sep 17 00:00:00 2001 From: Ben Ben-Ishay Date: Wed, 2 Mar 2022 17:07:08 +0200 Subject: net/mlx5e: SHAMPO, reduce TIR indication SHAMPO is an RQ / WQ feature, an indication was added to the TIR in the first place to enforce suitability between connected TIR and RQ, this enforcement does not exist in current the Firmware implementation and was redundant in the first place. Fixes: 83439f3c37aa ("net/mlx5e: Add HW-GRO offload") Signed-off-by: Ben Ben-Ishay Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en/tir.c | 3 --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 3 +-- include/linux/mlx5/mlx5_ifc.h | 1 - 3 files changed, 1 insertion(+), 6 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tir.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tir.c index da169b816665..d4239e3b3c88 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tir.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tir.c @@ -88,9 +88,6 @@ void mlx5e_tir_builder_build_packet_merge(struct mlx5e_tir_builder *builder, (MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ - rough_max_l2_l3_hdr_sz) >> 8); MLX5_SET(tirc, tirc, lro_timeout_period_usecs, pkt_merge_param->timeout); break; - case MLX5E_PACKET_MERGE_SHAMPO: - MLX5_SET(tirc, tirc, packet_merge_mask, MLX5_TIRC_PACKET_MERGE_MASK_SHAMPO); - break; default: break; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index bf80fb612449..3667f5ef5990 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -3616,8 +3616,7 @@ static int set_feature_hw_gro(struct net_device *netdev, bool enable) goto out; } - err = mlx5e_safe_switch_params(priv, &new_params, - mlx5e_modify_tirs_packet_merge_ctx, NULL, reset); + err = mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, reset); out: mutex_unlock(&priv->state_lock); return err; diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 5743f5b3414b..49a48d7709ac 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -3434,7 +3434,6 @@ enum { enum { MLX5_TIRC_PACKET_MERGE_MASK_IPV4_LRO = BIT(0), MLX5_TIRC_PACKET_MERGE_MASK_IPV6_LRO = BIT(1), - MLX5_TIRC_PACKET_MERGE_MASK_SHAMPO = BIT(2), }; enum { -- cgit v1.2.3-70-g09d2 From 2ac5b58e645c66932438bb021cb5b52097ce70b0 Mon Sep 17 00:00:00 2001 From: Miaoqian Lin Date: Thu, 10 Mar 2022 01:53:13 +0000 Subject: gianfar: ethtool: Fix refcount leak in gfar_get_ts_info The of_find_compatible_node() function returns a node pointer with refcount incremented, We should use of_node_put() on it when done Add the missing of_node_put() to release the refcount. Fixes: 7349a74ea75c ("net: ethernet: gianfar_ethtool: get phc index through drvdata") Signed-off-by: Miaoqian Lin Reviewed-by: Jesse Brandeburg Reviewed-by: Claudiu Manoil Link: https://lore.kernel.org/r/20220310015313.14938-1-linmq006@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/freescale/gianfar_ethtool.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/freescale/gianfar_ethtool.c b/drivers/net/ethernet/freescale/gianfar_ethtool.c index ff756265d58f..9a2c16d69e2c 100644 --- a/drivers/net/ethernet/freescale/gianfar_ethtool.c +++ b/drivers/net/ethernet/freescale/gianfar_ethtool.c @@ -1464,6 +1464,7 @@ static int gfar_get_ts_info(struct net_device *dev, ptp_node = of_find_compatible_node(NULL, NULL, "fsl,etsec-ptp"); if (ptp_node) { ptp_dev = of_find_device_by_node(ptp_node); + of_node_put(ptp_node); if (ptp_dev) ptp = platform_get_drvdata(ptp_dev); } -- cgit v1.2.3-70-g09d2 From bc0e610a6eb0d46e4123fafdbe5e6141d9fff3be Mon Sep 17 00:00:00 2001 From: Jianglei Nie Date: Wed, 9 Mar 2022 20:18:24 +0800 Subject: net: arc_emac: Fix use after free in arc_mdio_probe() If bus->state is equal to MDIOBUS_ALLOCATED, mdiobus_free(bus) will free the "bus". But bus->name is still used in the next line, which will lead to a use after free. We can fix it by putting the name in a local variable and make the bus->name point to the rodata section "name",then use the name in the error message without referring to bus to avoid the uaf. Fixes: 95b5fc03c189 ("net: arc_emac: Make use of the helper function dev_err_probe()") Signed-off-by: Jianglei Nie Link: https://lore.kernel.org/r/20220309121824.36529-1-niejianglei2021@163.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/arc/emac_mdio.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/arc/emac_mdio.c b/drivers/net/ethernet/arc/emac_mdio.c index 9acf589b1178..87f40c2ba904 100644 --- a/drivers/net/ethernet/arc/emac_mdio.c +++ b/drivers/net/ethernet/arc/emac_mdio.c @@ -132,6 +132,7 @@ int arc_mdio_probe(struct arc_emac_priv *priv) { struct arc_emac_mdio_bus_data *data = &priv->bus_data; struct device_node *np = priv->dev->of_node; + const char *name = "Synopsys MII Bus"; struct mii_bus *bus; int error; @@ -142,7 +143,7 @@ int arc_mdio_probe(struct arc_emac_priv *priv) priv->bus = bus; bus->priv = priv; bus->parent = priv->dev; - bus->name = "Synopsys MII Bus"; + bus->name = name; bus->read = &arc_mdio_read; bus->write = &arc_mdio_write; bus->reset = &arc_mdio_reset; @@ -167,7 +168,7 @@ int arc_mdio_probe(struct arc_emac_priv *priv) if (error) { mdiobus_free(bus); return dev_err_probe(priv->dev, error, - "cannot register MDIO bus %s\n", bus->name); + "cannot register MDIO bus %s\n", name); } return 0; -- cgit v1.2.3-70-g09d2 From 00b022f8f876a3a036b0df7f971001bef6398605 Mon Sep 17 00:00:00 2001 From: Jeremy Linton Date: Wed, 9 Mar 2022 22:55:35 -0600 Subject: net: bcmgenet: Don't claim WOL when its not available Some of the bcmgenet platforms don't correctly support WOL, yet ethtool returns: "Supports Wake-on: gsf" which is false. Ideally if there isn't a wol_irq, or there is something else that keeps the device from being able to wakeup it should display: "Supports Wake-on: d" This patch checks whether the device can wakup, before using the hard-coded supported flags. This corrects the ethtool reporting, as well as the WOL configuration because ethtool verifies that the mode is supported before attempting it. Fixes: c51de7f3976b ("net: bcmgenet: add Wake-on-LAN support code") Signed-off-by: Jeremy Linton Tested-by: Peter Robinson Acked-by: Florian Fainelli Link: https://lore.kernel.org/r/20220310045535.224450-1-jeremy.linton@arm.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/genet/bcmgenet_wol.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/broadcom/genet/bcmgenet_wol.c b/drivers/net/ethernet/broadcom/genet/bcmgenet_wol.c index e31a5a397f11..f55d9d9c01a8 100644 --- a/drivers/net/ethernet/broadcom/genet/bcmgenet_wol.c +++ b/drivers/net/ethernet/broadcom/genet/bcmgenet_wol.c @@ -40,6 +40,13 @@ void bcmgenet_get_wol(struct net_device *dev, struct ethtool_wolinfo *wol) { struct bcmgenet_priv *priv = netdev_priv(dev); + struct device *kdev = &priv->pdev->dev; + + if (!device_can_wakeup(kdev)) { + wol->supported = 0; + wol->wolopts = 0; + return; + } wol->supported = WAKE_MAGIC | WAKE_MAGICSECURE | WAKE_FILTER; wol->wolopts = priv->wolopts; -- cgit v1.2.3-70-g09d2 From 5cb1ebdbc4342b1c2ce89516e19808d64417bdbc Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Thu, 10 Mar 2022 18:16:41 +0100 Subject: ice: Fix race condition during interface enslave Commit 5dbbbd01cbba83 ("ice: Avoid RTNL lock when re-creating auxiliary device") changes a process of re-creation of aux device so ice_plug_aux_dev() is called from ice_service_task() context. This unfortunately opens a race window that can result in dead-lock when interface has left LAG and immediately enters LAG again. Reproducer: ``` #!/bin/sh ip link add lag0 type bond mode 1 miimon 100 ip link set lag0 for n in {1..10}; do echo Cycle: $n ip link set ens7f0 master lag0 sleep 1 ip link set ens7f0 nomaster done ``` This results in: [20976.208697] Workqueue: ice ice_service_task [ice] [20976.213422] Call Trace: [20976.215871] __schedule+0x2d1/0x830 [20976.219364] schedule+0x35/0xa0 [20976.222510] schedule_preempt_disabled+0xa/0x10 [20976.227043] __mutex_lock.isra.7+0x310/0x420 [20976.235071] enum_all_gids_of_dev_cb+0x1c/0x100 [ib_core] [20976.251215] ib_enum_roce_netdev+0xa4/0xe0 [ib_core] [20976.256192] ib_cache_setup_one+0x33/0xa0 [ib_core] [20976.261079] ib_register_device+0x40d/0x580 [ib_core] [20976.266139] irdma_ib_register_device+0x129/0x250 [irdma] [20976.281409] irdma_probe+0x2c1/0x360 [irdma] [20976.285691] auxiliary_bus_probe+0x45/0x70 [20976.289790] really_probe+0x1f2/0x480 [20976.298509] driver_probe_device+0x49/0xc0 [20976.302609] bus_for_each_drv+0x79/0xc0 [20976.306448] __device_attach+0xdc/0x160 [20976.310286] bus_probe_device+0x9d/0xb0 [20976.314128] device_add+0x43c/0x890 [20976.321287] __auxiliary_device_add+0x43/0x60 [20976.325644] ice_plug_aux_dev+0xb2/0x100 [ice] [20976.330109] ice_service_task+0xd0c/0xed0 [ice] [20976.342591] process_one_work+0x1a7/0x360 [20976.350536] worker_thread+0x30/0x390 [20976.358128] kthread+0x10a/0x120 [20976.365547] ret_from_fork+0x1f/0x40 ... [20976.438030] task:ip state:D stack: 0 pid:213658 ppid:213627 flags:0x00004084 [20976.446469] Call Trace: [20976.448921] __schedule+0x2d1/0x830 [20976.452414] schedule+0x35/0xa0 [20976.455559] schedule_preempt_disabled+0xa/0x10 [20976.460090] __mutex_lock.isra.7+0x310/0x420 [20976.464364] device_del+0x36/0x3c0 [20976.467772] ice_unplug_aux_dev+0x1a/0x40 [ice] [20976.472313] ice_lag_event_handler+0x2a2/0x520 [ice] [20976.477288] notifier_call_chain+0x47/0x70 [20976.481386] __netdev_upper_dev_link+0x18b/0x280 [20976.489845] bond_enslave+0xe05/0x1790 [bonding] [20976.494475] do_setlink+0x336/0xf50 [20976.502517] __rtnl_newlink+0x529/0x8b0 [20976.543441] rtnl_newlink+0x43/0x60 [20976.546934] rtnetlink_rcv_msg+0x2b1/0x360 [20976.559238] netlink_rcv_skb+0x4c/0x120 [20976.563079] netlink_unicast+0x196/0x230 [20976.567005] netlink_sendmsg+0x204/0x3d0 [20976.570930] sock_sendmsg+0x4c/0x50 [20976.574423] ____sys_sendmsg+0x1eb/0x250 [20976.586807] ___sys_sendmsg+0x7c/0xc0 [20976.606353] __sys_sendmsg+0x57/0xa0 [20976.609930] do_syscall_64+0x5b/0x1a0 [20976.613598] entry_SYSCALL_64_after_hwframe+0x65/0xca 1. Command 'ip link ... set nomaster' causes that ice_plug_aux_dev() is called from ice_service_task() context, aux device is created and associated device->lock is taken. 2. Command 'ip link ... set master...' calls ice's notifier under RTNL lock and that notifier calls ice_unplug_aux_dev(). That function tries to take aux device->lock but this is already taken by ice_plug_aux_dev() in step 1 3. Later ice_plug_aux_dev() tries to take RTNL lock but this is already taken in step 2 4. Dead-lock The patch fixes this issue by following changes: - Bit ICE_FLAG_PLUG_AUX_DEV is kept to be set during ice_plug_aux_dev() call in ice_service_task() - The bit is checked in ice_clear_rdma_cap() and only if it is not set then ice_unplug_aux_dev() is called. If it is set (in other words plugging of aux device was requested and ice_plug_aux_dev() is potentially running) then the function only clears the bit - Once ice_plug_aux_dev() call (in ice_service_task) is finished the bit ICE_FLAG_PLUG_AUX_DEV is cleared but it is also checked whether it was already cleared by ice_clear_rdma_cap(). If so then aux device is unplugged. Signed-off-by: Ivan Vecera Co-developed-by: Petr Oros Signed-off-by: Petr Oros Reviewed-by: Dave Ertman Link: https://lore.kernel.org/r/20220310171641.3863659-1-ivecera@redhat.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ice/ice.h | 11 ++++++++++- drivers/net/ethernet/intel/ice/ice_main.c | 12 +++++++++++- 2 files changed, 21 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet') diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h index 3121f9b04f59..bea1d1e39fa2 100644 --- a/drivers/net/ethernet/intel/ice/ice.h +++ b/drivers/net/ethernet/intel/ice/ice.h @@ -898,7 +898,16 @@ static inline void ice_set_rdma_cap(struct ice_pf *pf) */ static inline void ice_clear_rdma_cap(struct ice_pf *pf) { - ice_unplug_aux_dev(pf); + /* We can directly unplug aux device here only if the flag bit + * ICE_FLAG_PLUG_AUX_DEV is not set because ice_unplug_aux_dev() + * could race with ice_plug_aux_dev() called from + * ice_service_task(). In this case we only clear that bit now and + * aux device will be unplugged later once ice_plug_aux_device() + * called from ice_service_task() finishes (see ice_service_task()). + */ + if (!test_and_clear_bit(ICE_FLAG_PLUG_AUX_DEV, pf->flags)) + ice_unplug_aux_dev(pf); + clear_bit(ICE_FLAG_RDMA_ENA, pf->flags); clear_bit(ICE_FLAG_AUX_ENA, pf->flags); } diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 83e3e8aae6cf..493942e910be 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -2255,9 +2255,19 @@ static void ice_service_task(struct work_struct *work) return; } - if (test_and_clear_bit(ICE_FLAG_PLUG_AUX_DEV, pf->flags)) + if (test_bit(ICE_FLAG_PLUG_AUX_DEV, pf->flags)) { + /* Plug aux device per request */ ice_plug_aux_dev(pf); + /* Mark plugging as done but check whether unplug was + * requested during ice_plug_aux_dev() call + * (e.g. from ice_clear_rdma_cap()) and if so then + * plug aux device. + */ + if (!test_and_clear_bit(ICE_FLAG_PLUG_AUX_DEV, pf->flags)) + ice_unplug_aux_dev(pf); + } + if (test_and_clear_bit(ICE_FLAG_MTU_CHANGED, pf->flags)) { struct iidc_event *event; -- cgit v1.2.3-70-g09d2