From 23c619190318376769ad7b61504c2ea0703fb783 Mon Sep 17 00:00:00 2001 From: Dave Ertman Date: Tue, 9 Aug 2022 10:24:23 -0700 Subject: ice: Don't double unplug aux on peer initiated reset In the IDC callback that is accessed when the aux drivers request a reset, the function to unplug the aux devices is called. This function is also called in the ice_prepare_for_reset function. This double call is causing a "scheduling while atomic" BUG. [ 662.676430] ice 0000:4c:00.0 rocep76s0: cqp opcode = 0x1 maj_err_code = 0xffff min_err_code = 0x8003 [ 662.676609] ice 0000:4c:00.0 rocep76s0: [Modify QP Cmd Error][op_code=8] status=-29 waiting=1 completion_err=1 maj=0xffff min=0x8003 [ 662.815006] ice 0000:4c:00.0 rocep76s0: ICE OICR event notification: oicr = 0x10000003 [ 662.815014] ice 0000:4c:00.0 rocep76s0: critical PE Error, GLPE_CRITERR=0x00011424 [ 662.815017] ice 0000:4c:00.0 rocep76s0: Requesting a reset [ 662.815475] BUG: scheduling while atomic: swapper/37/0/0x00010002 [ 662.815475] BUG: scheduling while atomic: swapper/37/0/0x00010002 [ 662.815477] Modules linked in: rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache netfs rfkill 8021q garp mrp stp llc vfat fat rpcrdma intel_rapl_msr intel_rapl_common sunrpc i10nm_edac rdma_ucm nfit ib_srpt libnvdimm ib_isert iscsi_target_mod x86_pkg_temp_thermal intel_powerclamp coretemp target_core_mod snd_hda_intel ib_iser snd_intel_dspcfg libiscsi snd_intel_sdw_acpi scsi_transport_iscsi kvm_intel iTCO_wdt rdma_cm snd_hda_codec kvm iw_cm ipmi_ssif iTCO_vendor_support snd_hda_core irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel snd_hwdep snd_seq snd_seq_device rapl snd_pcm snd_timer isst_if_mbox_pci pcspkr isst_if_mmio irdma intel_uncore idxd acpi_ipmi joydev isst_if_common snd mei_me idxd_bus ipmi_si soundcore i2c_i801 mei ipmi_devintf i2c_smbus i2c_ismt ipmi_msghandler acpi_power_meter acpi_pad rv(OE) ib_uverbs ib_cm ib_core xfs libcrc32c ast i2c_algo_bit drm_vram_helper drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm_ttm_helpe r ttm [ 662.815546] nvme nvme_core ice drm crc32c_intel i40e t10_pi wmi pinctrl_emmitsburg dm_mirror dm_region_hash dm_log dm_mod fuse [ 662.815557] Preemption disabled at: [ 662.815558] [<0000000000000000>] 0x0 [ 662.815563] CPU: 37 PID: 0 Comm: swapper/37 Kdump: loaded Tainted: G S OE 5.17.1 #2 [ 662.815566] Hardware name: Intel Corporation D50DNP/D50DNP, BIOS SE5C6301.86B.6624.D18.2111021741 11/02/2021 [ 662.815568] Call Trace: [ 662.815572] [ 662.815574] dump_stack_lvl+0x33/0x42 [ 662.815581] __schedule_bug.cold.147+0x7d/0x8a [ 662.815588] __schedule+0x798/0x990 [ 662.815595] schedule+0x44/0xc0 [ 662.815597] schedule_preempt_disabled+0x14/0x20 [ 662.815600] __mutex_lock.isra.11+0x46c/0x490 [ 662.815603] ? __ibdev_printk+0x76/0xc0 [ib_core] [ 662.815633] device_del+0x37/0x3d0 [ 662.815639] ice_unplug_aux_dev+0x1a/0x40 [ice] [ 662.815674] ice_schedule_reset+0x3c/0xd0 [ice] [ 662.815693] irdma_iidc_event_handler.cold.7+0xb6/0xd3 [irdma] [ 662.815712] ? bitmap_find_next_zero_area_off+0x45/0xa0 [ 662.815719] ice_send_event_to_aux+0x54/0x70 [ice] [ 662.815741] ice_misc_intr+0x21d/0x2d0 [ice] [ 662.815756] __handle_irq_event_percpu+0x4c/0x180 [ 662.815762] handle_irq_event_percpu+0xf/0x40 [ 662.815764] handle_irq_event+0x34/0x60 [ 662.815766] handle_edge_irq+0x9a/0x1c0 [ 662.815770] __common_interrupt+0x62/0x100 [ 662.815774] common_interrupt+0xb4/0xd0 [ 662.815779] [ 662.815780] [ 662.815780] asm_common_interrupt+0x1e/0x40 [ 662.815785] RIP: 0010:cpuidle_enter_state+0xd6/0x380 [ 662.815789] Code: 49 89 c4 0f 1f 44 00 00 31 ff e8 65 d7 95 ff 45 84 ff 74 12 9c 58 f6 c4 02 0f 85 64 02 00 00 31 ff e8 ae c5 9c ff fb 45 85 f6 <0f> 88 12 01 00 00 49 63 d6 4c 2b 24 24 48 8d 04 52 48 8d 04 82 49 [ 662.815791] RSP: 0018:ff2c2c4f18edbe80 EFLAGS: 00000202 [ 662.815793] RAX: ff280805df140000 RBX: 0000000000000002 RCX: 000000000000001f [ 662.815795] RDX: 0000009a52da2d08 RSI: ffffffff93f8240b RDI: ffffffff93f53ee7 [ 662.815796] RBP: ff5e2bd11ff41928 R08: 0000000000000000 R09: 000000000002f8c0 [ 662.815797] R10: 0000010c3f18e2cf R11: 000000000000000f R12: 0000009a52da2d08 [ 662.815798] R13: ffffffff94ad7e20 R14: 0000000000000002 R15: 0000000000000000 [ 662.815801] cpuidle_enter+0x29/0x40 [ 662.815803] do_idle+0x261/0x2b0 [ 662.815807] cpu_startup_entry+0x19/0x20 [ 662.815809] start_secondary+0x114/0x150 [ 662.815813] secondary_startup_64_no_verify+0xd5/0xdb [ 662.815818] [ 662.815846] bad: scheduling from the idle thread! [ 662.815849] CPU: 37 PID: 0 Comm: swapper/37 Kdump: loaded Tainted: G S W OE 5.17.1 #2 [ 662.815852] Hardware name: Intel Corporation D50DNP/D50DNP, BIOS SE5C6301.86B.6624.D18.2111021741 11/02/2021 [ 662.815853] Call Trace: [ 662.815855] [ 662.815856] dump_stack_lvl+0x33/0x42 [ 662.815860] dequeue_task_idle+0x20/0x30 [ 662.815863] __schedule+0x1c3/0x990 [ 662.815868] schedule+0x44/0xc0 [ 662.815871] schedule_preempt_disabled+0x14/0x20 [ 662.815873] __mutex_lock.isra.11+0x3a8/0x490 [ 662.815876] ? __ibdev_printk+0x76/0xc0 [ib_core] [ 662.815904] device_del+0x37/0x3d0 [ 662.815909] ice_unplug_aux_dev+0x1a/0x40 [ice] [ 662.815937] ice_schedule_reset+0x3c/0xd0 [ice] [ 662.815961] irdma_iidc_event_handler.cold.7+0xb6/0xd3 [irdma] [ 662.815979] ? bitmap_find_next_zero_area_off+0x45/0xa0 [ 662.815985] ice_send_event_to_aux+0x54/0x70 [ice] [ 662.816011] ice_misc_intr+0x21d/0x2d0 [ice] [ 662.816033] __handle_irq_event_percpu+0x4c/0x180 [ 662.816037] handle_irq_event_percpu+0xf/0x40 [ 662.816039] handle_irq_event+0x34/0x60 [ 662.816042] handle_edge_irq+0x9a/0x1c0 [ 662.816045] __common_interrupt+0x62/0x100 [ 662.816048] common_interrupt+0xb4/0xd0 [ 662.816052] [ 662.816053] [ 662.816054] asm_common_interrupt+0x1e/0x40 [ 662.816057] RIP: 0010:cpuidle_enter_state+0xd6/0x380 [ 662.816060] Code: 49 89 c4 0f 1f 44 00 00 31 ff e8 65 d7 95 ff 45 84 ff 74 12 9c 58 f6 c4 02 0f 85 64 02 00 00 31 ff e8 ae c5 9c ff fb 45 85 f6 <0f> 88 12 01 00 00 49 63 d6 4c 2b 24 24 48 8d 04 52 48 8d 04 82 49 [ 662.816063] RSP: 0018:ff2c2c4f18edbe80 EFLAGS: 00000202 [ 662.816065] RAX: ff280805df140000 RBX: 0000000000000002 RCX: 000000000000001f [ 662.816067] RDX: 0000009a52da2d08 RSI: ffffffff93f8240b RDI: ffffffff93f53ee7 [ 662.816068] RBP: ff5e2bd11ff41928 R08: 0000000000000000 R09: 000000000002f8c0 [ 662.816070] R10: 0000010c3f18e2cf R11: 000000000000000f R12: 0000009a52da2d08 [ 662.816071] R13: ffffffff94ad7e20 R14: 0000000000000002 R15: 0000000000000000 [ 662.816075] cpuidle_enter+0x29/0x40 [ 662.816077] do_idle+0x261/0x2b0 [ 662.816080] cpu_startup_entry+0x19/0x20 [ 662.816083] start_secondary+0x114/0x150 [ 662.816087] secondary_startup_64_no_verify+0xd5/0xdb [ 662.816091] [ 662.816169] bad: scheduling from the idle thread! The correct place to unplug the aux devices for a reset is in the prepare_for_reset function, as this is a common place for all reset flows. It also has built in protection from being called twice in a single reset instance before the aux devices are replugged. Fixes: f9f5301e7e2d4 ("ice: Register auxiliary device to provide RDMA") Signed-off-by: Dave Ertman Tested-by: Helena Anna Dubel Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_main.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 8c30eea61b6d..04836bbaf7d5 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -2399,8 +2399,6 @@ int ice_schedule_reset(struct ice_pf *pf, enum ice_reset_req reset) return -EBUSY; } - ice_unplug_aux_dev(pf); - switch (reset) { case ICE_RESET_PFR: set_bit(ICE_PFR_REQ, pf->state); -- cgit v1.2.3 From a509702cac95a8b450228a037c8542f57e538e5b Mon Sep 17 00:00:00 2001 From: Ding Hui Date: Wed, 17 Aug 2022 18:53:18 +0800 Subject: ice: Fix crash by keep old cfg when update TCs more than queues There are problems if allocated queues less than Traffic Classes. Commit a632b2a4c920 ("ice: ethtool: Prohibit improper channel config for DCB") already disallow setting less queues than TCs. Another case is if we first set less queues, and later update more TCs config due to LLDP, ice_vsi_cfg_tc() will failed but left dirty num_txq/rxq and tc_cfg in vsi, that will cause invalid pointer access. [ 95.968089] ice 0000:3b:00.1: More TCs defined than queues/rings allocated. [ 95.968092] ice 0000:3b:00.1: Trying to use more Rx queues (8), than were allocated (1)! [ 95.968093] ice 0000:3b:00.1: Failed to config TC for VSI index: 0 [ 95.969621] general protection fault: 0000 [#1] SMP NOPTI [ 95.969705] CPU: 1 PID: 58405 Comm: lldpad Kdump: loaded Tainted: G U W O --------- -t - 4.18.0 #1 [ 95.969867] Hardware name: O.E.M/BC11SPSCB10, BIOS 8.23 12/30/2021 [ 95.969992] RIP: 0010:devm_kmalloc+0xa/0x60 [ 95.970052] Code: 5c ff ff ff 31 c0 5b 5d 41 5c c3 b8 f4 ff ff ff eb f4 0f 1f 40 00 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 89 f8 89 d1 <8b> 97 60 02 00 00 48 8d 7e 18 48 39 f7 72 3f 55 89 ce 53 48 8b 4c [ 95.970344] RSP: 0018:ffffc9003f553888 EFLAGS: 00010206 [ 95.970425] RAX: dead000000000200 RBX: ffffea003c425b00 RCX: 00000000006080c0 [ 95.970536] RDX: 00000000006080c0 RSI: 0000000000000200 RDI: dead000000000200 [ 95.970648] RBP: dead000000000200 R08: 00000000000463c0 R09: ffff888ffa900000 [ 95.970760] R10: 0000000000000000 R11: 0000000000000002 R12: ffff888ff6b40100 [ 95.970870] R13: ffff888ff6a55018 R14: 0000000000000000 R15: ffff888ff6a55460 [ 95.970981] FS: 00007f51b7d24700(0000) GS:ffff88903ee80000(0000) knlGS:0000000000000000 [ 95.971108] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 95.971197] CR2: 00007fac5410d710 CR3: 0000000f2c1de002 CR4: 00000000007606e0 [ 95.971309] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 95.971419] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 95.971530] PKRU: 55555554 [ 95.971573] Call Trace: [ 95.971622] ice_setup_rx_ring+0x39/0x110 [ice] [ 95.971695] ice_vsi_setup_rx_rings+0x54/0x90 [ice] [ 95.971774] ice_vsi_open+0x25/0x120 [ice] [ 95.971843] ice_open_internal+0xb8/0x1f0 [ice] [ 95.971919] ice_ena_vsi+0x4f/0xd0 [ice] [ 95.971987] ice_dcb_ena_dis_vsi.constprop.5+0x29/0x90 [ice] [ 95.972082] ice_pf_dcb_cfg+0x29a/0x380 [ice] [ 95.972154] ice_dcbnl_setets+0x174/0x1b0 [ice] [ 95.972220] dcbnl_ieee_set+0x89/0x230 [ 95.972279] ? dcbnl_ieee_del+0x150/0x150 [ 95.972341] dcb_doit+0x124/0x1b0 [ 95.972392] rtnetlink_rcv_msg+0x243/0x2f0 [ 95.972457] ? dcb_doit+0x14d/0x1b0 [ 95.972510] ? __kmalloc_node_track_caller+0x1d3/0x280 [ 95.972591] ? rtnl_calcit.isra.31+0x100/0x100 [ 95.972661] netlink_rcv_skb+0xcf/0xf0 [ 95.972720] netlink_unicast+0x16d/0x220 [ 95.972781] netlink_sendmsg+0x2ba/0x3a0 [ 95.975891] sock_sendmsg+0x4c/0x50 [ 95.979032] ___sys_sendmsg+0x2e4/0x300 [ 95.982147] ? kmem_cache_alloc+0x13e/0x190 [ 95.985242] ? __wake_up_common_lock+0x79/0x90 [ 95.988338] ? __check_object_size+0xac/0x1b0 [ 95.991440] ? _copy_to_user+0x22/0x30 [ 95.994539] ? move_addr_to_user+0xbb/0xd0 [ 95.997619] ? __sys_sendmsg+0x53/0x80 [ 96.000664] __sys_sendmsg+0x53/0x80 [ 96.003747] do_syscall_64+0x5b/0x1d0 [ 96.006862] entry_SYSCALL_64_after_hwframe+0x65/0xca Only update num_txq/rxq when passed check, and restore tc_cfg if setup queue map failed. Fixes: a632b2a4c920 ("ice: ethtool: Prohibit improper channel config for DCB") Signed-off-by: Ding Hui Reviewed-by: Anatolii Gerasymenko Tested-by: Arpana Arland (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_lib.c | 42 ++++++++++++++++++++------------ 1 file changed, 26 insertions(+), 16 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c index 0c4ec9264071..58d483e2f539 100644 --- a/drivers/net/ethernet/intel/ice/ice_lib.c +++ b/drivers/net/ethernet/intel/ice/ice_lib.c @@ -914,7 +914,7 @@ static void ice_set_dflt_vsi_ctx(struct ice_hw *hw, struct ice_vsi_ctx *ctxt) */ static int ice_vsi_setup_q_map(struct ice_vsi *vsi, struct ice_vsi_ctx *ctxt) { - u16 offset = 0, qmap = 0, tx_count = 0, pow = 0; + u16 offset = 0, qmap = 0, tx_count = 0, rx_count = 0, pow = 0; u16 num_txq_per_tc, num_rxq_per_tc; u16 qcount_tx = vsi->alloc_txq; u16 qcount_rx = vsi->alloc_rxq; @@ -981,23 +981,25 @@ static int ice_vsi_setup_q_map(struct ice_vsi *vsi, struct ice_vsi_ctx *ctxt) * at least 1) */ if (offset) - vsi->num_rxq = offset; + rx_count = offset; else - vsi->num_rxq = num_rxq_per_tc; + rx_count = num_rxq_per_tc; - if (vsi->num_rxq > vsi->alloc_rxq) { + if (rx_count > vsi->alloc_rxq) { dev_err(ice_pf_to_dev(vsi->back), "Trying to use more Rx queues (%u), than were allocated (%u)!\n", - vsi->num_rxq, vsi->alloc_rxq); + rx_count, vsi->alloc_rxq); return -EINVAL; } - vsi->num_txq = tx_count; - if (vsi->num_txq > vsi->alloc_txq) { + if (tx_count > vsi->alloc_txq) { dev_err(ice_pf_to_dev(vsi->back), "Trying to use more Tx queues (%u), than were allocated (%u)!\n", - vsi->num_txq, vsi->alloc_txq); + tx_count, vsi->alloc_txq); return -EINVAL; } + vsi->num_txq = tx_count; + vsi->num_rxq = rx_count; + if (vsi->type == ICE_VSI_VF && vsi->num_txq != vsi->num_rxq) { dev_dbg(ice_pf_to_dev(vsi->back), "VF VSI should have same number of Tx and Rx queues. Hence making them equal\n"); /* since there is a chance that num_rxq could have been changed @@ -3490,6 +3492,7 @@ ice_vsi_setup_q_map_mqprio(struct ice_vsi *vsi, struct ice_vsi_ctx *ctxt, u16 pow, offset = 0, qcount_tx = 0, qcount_rx = 0, qmap; u16 tc0_offset = vsi->mqprio_qopt.qopt.offset[0]; int tc0_qcount = vsi->mqprio_qopt.qopt.count[0]; + u16 new_txq, new_rxq; u8 netdev_tc = 0; int i; @@ -3530,21 +3533,24 @@ ice_vsi_setup_q_map_mqprio(struct ice_vsi *vsi, struct ice_vsi_ctx *ctxt, } } - /* Set actual Tx/Rx queue pairs */ - vsi->num_txq = offset + qcount_tx; - if (vsi->num_txq > vsi->alloc_txq) { + new_txq = offset + qcount_tx; + if (new_txq > vsi->alloc_txq) { dev_err(ice_pf_to_dev(vsi->back), "Trying to use more Tx queues (%u), than were allocated (%u)!\n", - vsi->num_txq, vsi->alloc_txq); + new_txq, vsi->alloc_txq); return -EINVAL; } - vsi->num_rxq = offset + qcount_rx; - if (vsi->num_rxq > vsi->alloc_rxq) { + new_rxq = offset + qcount_rx; + if (new_rxq > vsi->alloc_rxq) { dev_err(ice_pf_to_dev(vsi->back), "Trying to use more Rx queues (%u), than were allocated (%u)!\n", - vsi->num_rxq, vsi->alloc_rxq); + new_rxq, vsi->alloc_rxq); return -EINVAL; } + /* Set actual Tx/Rx queue pairs */ + vsi->num_txq = new_txq; + vsi->num_rxq = new_rxq; + /* Setup queue TC[0].qmap for given VSI context */ ctxt->info.tc_mapping[0] = cpu_to_le16(qmap); ctxt->info.q_mapping[0] = cpu_to_le16(vsi->rxq_map[0]); @@ -3576,6 +3582,7 @@ int ice_vsi_cfg_tc(struct ice_vsi *vsi, u8 ena_tc) { u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 }; struct ice_pf *pf = vsi->back; + struct ice_tc_cfg old_tc_cfg; struct ice_vsi_ctx *ctx; struct device *dev; int i, ret = 0; @@ -3600,6 +3607,7 @@ int ice_vsi_cfg_tc(struct ice_vsi *vsi, u8 ena_tc) max_txqs[i] = vsi->num_txq; } + memcpy(&old_tc_cfg, &vsi->tc_cfg, sizeof(old_tc_cfg)); vsi->tc_cfg.ena_tc = ena_tc; vsi->tc_cfg.numtc = num_tc; @@ -3616,8 +3624,10 @@ int ice_vsi_cfg_tc(struct ice_vsi *vsi, u8 ena_tc) else ret = ice_vsi_setup_q_map(vsi, ctx); - if (ret) + if (ret) { + memcpy(&vsi->tc_cfg, &old_tc_cfg, sizeof(vsi->tc_cfg)); goto out; + } /* must to indicate which section of VSI context are being modified */ ctx->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_RXQ_MAP_VALID); -- cgit v1.2.3 From f66b98c868f2c1ffcb1139ea19d6f20983f1f9dc Mon Sep 17 00:00:00 2001 From: Sylwester Dziedziuch Date: Thu, 1 Sep 2022 16:32:06 +0200 Subject: iavf: Fix change VF's mac address Previously changing mac address gives false negative because ip link set address return with RTNLINK: Permission denied. In iavf_set_mac was check if PF handled our mac set request, even before filter was added to list. Because this check returns always true and it never waits for PF's response. Move iavf_is_mac_handled to wait_event_interruptible_timeout instead of false. Now it will wait for PF's response and then check if address was added or rejected. Fixes: 35a2443d0910 ("iavf: Add waiting for response from PF in set mac") Signed-off-by: Sylwester Dziedziuch Co-developed-by: Norbert Zulinski Signed-off-by: Norbert Zulinski Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_main.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 10aa99dfdcdb..0c89f16bf1e2 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -1077,7 +1077,6 @@ static int iavf_set_mac(struct net_device *netdev, void *p) { struct iavf_adapter *adapter = netdev_priv(netdev); struct sockaddr *addr = p; - bool handle_mac = iavf_is_mac_set_handled(netdev, addr->sa_data); int ret; if (!is_valid_ether_addr(addr->sa_data)) @@ -1094,10 +1093,9 @@ static int iavf_set_mac(struct net_device *netdev, void *p) return 0; } - if (handle_mac) - goto done; - - ret = wait_event_interruptible_timeout(adapter->vc_waitqueue, false, msecs_to_jiffies(2500)); + ret = wait_event_interruptible_timeout(adapter->vc_waitqueue, + iavf_is_mac_set_handled(netdev, addr->sa_data), + msecs_to_jiffies(2500)); /* If ret < 0 then it means wait was interrupted. * If ret == 0 then it means we got a timeout. @@ -1111,7 +1109,6 @@ static int iavf_set_mac(struct net_device *netdev, void *p) if (!ret) return -EAGAIN; -done: if (!ether_addr_equal(netdev->dev_addr, addr->sa_data)) return -EACCES; -- cgit v1.2.3 From 809f23c0423a43266e47a7dc67e95b5cb4d1cbfc Mon Sep 17 00:00:00 2001 From: Brett Creeley Date: Thu, 1 Sep 2022 16:34:40 +0200 Subject: iavf: Fix cached head and tail value for iavf_get_tx_pending The underlying hardware may or may not allow reading of the head or tail registers and it really makes no difference if we use the software cached values. So, always used the software cached values. Fixes: 9c6c12595b73 ("i40e: Detection and recovery of TX queue hung logic moved to service_task from tx_timeout") Signed-off-by: Brett Creeley Co-developed-by: Norbert Zulinski Signed-off-by: Norbert Zulinski Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_txrx.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/iavf/iavf_txrx.c b/drivers/net/ethernet/intel/iavf/iavf_txrx.c index 06d18797d25a..4c3f3f419110 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_txrx.c +++ b/drivers/net/ethernet/intel/iavf/iavf_txrx.c @@ -114,8 +114,11 @@ u32 iavf_get_tx_pending(struct iavf_ring *ring, bool in_sw) { u32 head, tail; + /* underlying hardware might not allow access and/or always return + * 0 for the head/tail registers so just use the cached values + */ head = ring->next_to_clean; - tail = readl(ring->tail); + tail = ring->next_to_use; if (head != tail) return (head < tail) ? -- cgit v1.2.3 From 66039eb9015eee4f7ff0c99b83c65c7ecb3c8190 Mon Sep 17 00:00:00 2001 From: Norbert Zulinski Date: Wed, 14 Sep 2022 15:39:13 +0200 Subject: iavf: Fix bad page state Fix bad page state, free inappropriate page in handling dummy descriptor. iavf_build_skb now has to check not only if rx_buffer is NULL but also if size is zero, same thing in iavf_clean_rx_irq. Without this patch driver would free page that will be used by napi_build_skb. Fixes: a9f49e006030 ("iavf: Fix handling of dummy receive descriptors") Signed-off-by: Norbert Zulinski Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_txrx.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/iavf/iavf_txrx.c b/drivers/net/ethernet/intel/iavf/iavf_txrx.c index 4c3f3f419110..18b6a702a1d6 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_txrx.c +++ b/drivers/net/ethernet/intel/iavf/iavf_txrx.c @@ -1393,7 +1393,7 @@ static struct sk_buff *iavf_build_skb(struct iavf_ring *rx_ring, #endif struct sk_buff *skb; - if (!rx_buffer) + if (!rx_buffer || !size) return NULL; /* prefetch first cache line of first page */ va = page_address(rx_buffer->page) + rx_buffer->page_offset; @@ -1551,7 +1551,7 @@ static int iavf_clean_rx_irq(struct iavf_ring *rx_ring, int budget) /* exit if we failed to retrieve a buffer */ if (!skb) { rx_ring->rx_stats.alloc_buff_failed++; - if (rx_buffer) + if (rx_buffer && size) rx_buffer->pagecnt_bias++; break; } -- cgit v1.2.3 From 399c98c4dc50b7eb7e9f24da7ffdda6f025676ef Mon Sep 17 00:00:00 2001 From: Michal Jaron Date: Tue, 13 Sep 2022 15:38:35 +0200 Subject: iavf: Fix set max MTU size with port VLAN and jumbo frames After setting port VLAN and MTU to 9000 on VF with ice driver there was an iavf error "PF returned error -5 (IAVF_ERR_PARAM) to our request 6". During queue configuration, VF's max packet size was set to IAVF_MAX_RXBUFFER but on ice max frame size was smaller by VLAN_HLEN due to making some space for port VLAN as VF is not aware whether it's in a port VLAN. This mismatch in sizes caused ice to reject queue configuration with ERR_PARAM error. Proper max_mtu is sent from ice PF to VF with GET_VF_RESOURCES msg but VF does not look at this. In iavf change max_frame from IAVF_MAX_RXBUFFER to max_mtu received from pf with GET_VF_RESOURCES msg to make vf's max_frame_size dependent from pf. Add check if received max_mtu is not in eligible range then set it to IAVF_MAX_RXBUFFER. Fixes: dab86afdbbd1 ("i40e/i40evf: Change the way we limit the maximum frame size for Rx") Signed-off-by: Michal Jaron Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c index 15ee85dc33bd..5a9e6563923e 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c @@ -269,11 +269,14 @@ int iavf_get_vf_vlan_v2_caps(struct iavf_adapter *adapter) void iavf_configure_queues(struct iavf_adapter *adapter) { struct virtchnl_vsi_queue_config_info *vqci; - struct virtchnl_queue_pair_info *vqpi; + int i, max_frame = adapter->vf_res->max_mtu; int pairs = adapter->num_active_queues; - int i, max_frame = IAVF_MAX_RXBUFFER; + struct virtchnl_queue_pair_info *vqpi; size_t len; + if (max_frame > IAVF_MAX_RXBUFFER || !max_frame) + max_frame = IAVF_MAX_RXBUFFER; + if (adapter->current_op != VIRTCHNL_OP_UNKNOWN) { /* bail because we already have a command pending */ dev_err(&adapter->pdev->dev, "Cannot configure queues, command %d pending\n", -- cgit v1.2.3 From 372539def2824c43b6afe2403045b140f65c5acc Mon Sep 17 00:00:00 2001 From: Michal Jaron Date: Tue, 13 Sep 2022 15:38:36 +0200 Subject: i40e: Fix VF set max MTU size Max MTU sent to VF is set to 0 during memory allocation. It cause that max MTU on VF is changed to IAVF_MAX_RXBUFFER and does not depend on data from HW. Set max_mtu field in virtchnl_vf_resource struct to inform VF in GET_VF_RESOURCES msg what size should be max frame. Fixes: dab86afdbbd1 ("i40e/i40evf: Change the way we limit the maximum frame size for Rx") Signed-off-by: Michal Jaron Signed-off-by: Mateusz Palczewski Tested-by: Konrad Jankowski Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c index 4f184c50f6e8..7e9f6a69eb10 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c @@ -2038,6 +2038,25 @@ static void i40e_del_qch(struct i40e_vf *vf) } } +/** + * i40e_vc_get_max_frame_size + * @vf: pointer to the VF + * + * Max frame size is determined based on the current port's max frame size and + * whether a port VLAN is configured on this VF. The VF is not aware whether + * it's in a port VLAN so the PF needs to account for this in max frame size + * checks and sending the max frame size to the VF. + **/ +static u16 i40e_vc_get_max_frame_size(struct i40e_vf *vf) +{ + u16 max_frame_size = vf->pf->hw.phy.link_info.max_frame_size; + + if (vf->port_vlan_id) + max_frame_size -= VLAN_HLEN; + + return max_frame_size; +} + /** * i40e_vc_get_vf_resources_msg * @vf: pointer to the VF info @@ -2139,6 +2158,7 @@ static int i40e_vc_get_vf_resources_msg(struct i40e_vf *vf, u8 *msg) vfres->max_vectors = pf->hw.func_caps.num_msix_vectors_vf; vfres->rss_key_size = I40E_HKEY_ARRAY_SIZE; vfres->rss_lut_size = I40E_VF_HLUT_ARRAY_SIZE; + vfres->max_mtu = i40e_vc_get_max_frame_size(vf); if (vf->lan_vsi_idx) { vfres->vsi_res[0].vsi_id = vf->lan_vsi_id; -- cgit v1.2.3 From 198eb7e1b81d8ba676d0f4f120c092032ae69a8e Mon Sep 17 00:00:00 2001 From: Michal Jaron Date: Thu, 1 Sep 2022 09:49:33 +0200 Subject: i40e: Fix set max_tx_rate when it is lower than 1 Mbps While converting max_tx_rate from bytes to Mbps, this value was set to 0, if the original value was lower than 125000 bytes (1 Mbps). This would cause no transmission rate limiting to occur. This happened due to lack of check of max_tx_rate against the 1 Mbps value for max_tx_rate and the following division by 125000. Fix this issue by adding a helper i40e_bw_bytes_to_mbits() which sets max_tx_rate to minimum usable value of 50 Mbps, if its value is less than 1 Mbps, otherwise do the required conversion by dividing by 125000. Fixes: 5ecae4120a6b ("i40e: Refactor VF BW rate limiting") Signed-off-by: Michal Jaron Signed-off-by: Andrii Staikov Tested-by: Bharathi Sreenivas Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/i40e/i40e_main.c | 32 +++++++++++++++++++++++------ 1 file changed, 26 insertions(+), 6 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index 10c1e1ea83a1..e3d9804aeb25 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -5908,6 +5908,26 @@ static int i40e_get_link_speed(struct i40e_vsi *vsi) } } +/** + * i40e_bw_bytes_to_mbits - Convert max_tx_rate from bytes to mbits + * @vsi: Pointer to vsi structure + * @max_tx_rate: max TX rate in bytes to be converted into Mbits + * + * Helper function to convert units before send to set BW limit + **/ +static u64 i40e_bw_bytes_to_mbits(struct i40e_vsi *vsi, u64 max_tx_rate) +{ + if (max_tx_rate < I40E_BW_MBPS_DIVISOR) { + dev_warn(&vsi->back->pdev->dev, + "Setting max tx rate to minimum usable value of 50Mbps.\n"); + max_tx_rate = I40E_BW_CREDIT_DIVISOR; + } else { + do_div(max_tx_rate, I40E_BW_MBPS_DIVISOR); + } + + return max_tx_rate; +} + /** * i40e_set_bw_limit - setup BW limit for Tx traffic based on max_tx_rate * @vsi: VSI to be configured @@ -5930,10 +5950,10 @@ int i40e_set_bw_limit(struct i40e_vsi *vsi, u16 seid, u64 max_tx_rate) max_tx_rate, seid); return -EINVAL; } - if (max_tx_rate && max_tx_rate < 50) { + if (max_tx_rate && max_tx_rate < I40E_BW_CREDIT_DIVISOR) { dev_warn(&pf->pdev->dev, "Setting max tx rate to minimum usable value of 50Mbps.\n"); - max_tx_rate = 50; + max_tx_rate = I40E_BW_CREDIT_DIVISOR; } /* Tx rate credits are in values of 50Mbps, 0 is disabled */ @@ -8224,9 +8244,9 @@ config_tc: if (i40e_is_tc_mqprio_enabled(pf)) { if (vsi->mqprio_qopt.max_rate[0]) { - u64 max_tx_rate = vsi->mqprio_qopt.max_rate[0]; + u64 max_tx_rate = i40e_bw_bytes_to_mbits(vsi, + vsi->mqprio_qopt.max_rate[0]); - do_div(max_tx_rate, I40E_BW_MBPS_DIVISOR); ret = i40e_set_bw_limit(vsi, vsi->seid, max_tx_rate); if (!ret) { u64 credits = max_tx_rate; @@ -10971,10 +10991,10 @@ static void i40e_rebuild(struct i40e_pf *pf, bool reinit, bool lock_acquired) } if (vsi->mqprio_qopt.max_rate[0]) { - u64 max_tx_rate = vsi->mqprio_qopt.max_rate[0]; + u64 max_tx_rate = i40e_bw_bytes_to_mbits(vsi, + vsi->mqprio_qopt.max_rate[0]); u64 credits = 0; - do_div(max_tx_rate, I40E_BW_MBPS_DIVISOR); ret = i40e_set_bw_limit(vsi, vsi->seid, max_tx_rate); if (ret) goto end_unlock; -- cgit v1.2.3 From 122045ca770492660005c22379995506f13efea8 Mon Sep 17 00:00:00 2001 From: Michal Swiatkowski Date: Mon, 8 Aug 2022 11:58:54 +0200 Subject: ice: config netdev tc before setting queues number After lowering number of tx queues the warning appears: "Number of in use tx queues changed invalidating tc mappings. Priority traffic classification disabled!" Example command to reproduce: ethtool -L enp24s0f0 tx 36 rx 36 Fix this by setting correct tc mapping before setting real number of queues on netdev. Fixes: 0754d65bd4be5 ("ice: Add infrastructure for mqprio support via ndo_setup_tc") Signed-off-by: Michal Swiatkowski Tested-by: Gurucharan (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_main.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 04836bbaf7d5..3042a76b848d 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -6858,6 +6858,8 @@ int ice_vsi_open(struct ice_vsi *vsi) if (err) goto err_setup_rx; + ice_vsi_cfg_netdev_tc(vsi, vsi->tc_cfg.ena_tc); + if (vsi->type == ICE_VSI_PF) { /* Notify the stack of the actual queue counts. */ err = netif_set_real_num_tx_queues(vsi->netdev, vsi->num_txq); -- cgit v1.2.3 From 8ac7132704f3fbd2095abb9459e5303ce8c9e559 Mon Sep 17 00:00:00 2001 From: Mateusz Palczewski Date: Fri, 26 Aug 2022 10:31:23 +0200 Subject: ice: Fix interface being down after reset with link-down-on-close flag on When performing a reset on ice driver with link-down-on-close flag on interface would always stay down. Fix this by moving a check of this flag to ice_stop() that is called only when user wants to bring interface down. Fixes: ab4ab73fc1ec ("ice: Add ethtool private flag to make forcing link down optional") Signed-off-by: Mateusz Palczewski Tested-by: Petr Oros Tested-by: Gurucharan (A Contingent worker at Intel) Signed-off-by: Tony Nguyen --- drivers/net/ethernet/intel/ice/ice_main.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 3042a76b848d..e109cb93886b 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -6649,7 +6649,7 @@ static void ice_napi_disable_all(struct ice_vsi *vsi) */ int ice_down(struct ice_vsi *vsi) { - int i, tx_err, rx_err, link_err = 0, vlan_err = 0; + int i, tx_err, rx_err, vlan_err = 0; WARN_ON(!test_bit(ICE_VSI_DOWN, vsi->state)); @@ -6683,20 +6683,13 @@ int ice_down(struct ice_vsi *vsi) ice_napi_disable_all(vsi); - if (test_bit(ICE_FLAG_LINK_DOWN_ON_CLOSE_ENA, vsi->back->flags)) { - link_err = ice_force_phys_link_state(vsi, false); - if (link_err) - netdev_err(vsi->netdev, "Failed to set physical link down, VSI %d error %d\n", - vsi->vsi_num, link_err); - } - ice_for_each_txq(vsi, i) ice_clean_tx_ring(vsi->tx_rings[i]); ice_for_each_rxq(vsi, i) ice_clean_rx_ring(vsi->rx_rings[i]); - if (tx_err || rx_err || link_err || vlan_err) { + if (tx_err || rx_err || vlan_err) { netdev_err(vsi->netdev, "Failed to close VSI 0x%04X on switch 0x%04X\n", vsi->vsi_num, vsi->vsw->sw_id); return -EIO; @@ -8892,6 +8885,16 @@ int ice_stop(struct net_device *netdev) return -EBUSY; } + if (test_bit(ICE_FLAG_LINK_DOWN_ON_CLOSE_ENA, vsi->back->flags)) { + int link_err = ice_force_phys_link_state(vsi, false); + + if (link_err) { + netdev_err(vsi->netdev, "Failed to set physical link down, VSI %d error %d\n", + vsi->vsi_num, link_err); + return -EIO; + } + } + ice_vsi_close(vsi); return 0; -- cgit v1.2.3 From 114f398d48c571bb628187a7b2dd42695156781f Mon Sep 17 00:00:00 2001 From: Larysa Zaremba Date: Mon, 19 Sep 2022 15:43:46 +0200 Subject: ice: Fix ice_xdp_xmit() when XDP TX queue number is not sufficient The original patch added the static branch to handle the situation, when assigning an XDP TX queue to every CPU is not possible, so they have to be shared. However, in the XDP transmit handler ice_xdp_xmit(), an error was returned in such cases even before static condition was checked, thus making queue sharing still impossible. Fixes: 22bf877e528f ("ice: introduce XDP_TX fallback path") Signed-off-by: Larysa Zaremba Reviewed-by: Alexander Lobakin Link: https://lore.kernel.org/r/20220919134346.25030-1-larysa.zaremba@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ice/ice_txrx.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/intel') diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c index 836dce840712..97453d1dfafe 100644 --- a/drivers/net/ethernet/intel/ice/ice_txrx.c +++ b/drivers/net/ethernet/intel/ice/ice_txrx.c @@ -610,7 +610,7 @@ ice_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, if (test_bit(ICE_VSI_DOWN, vsi->state)) return -ENETDOWN; - if (!ice_is_xdp_ena_vsi(vsi) || queue_index >= vsi->num_xdp_txq) + if (!ice_is_xdp_ena_vsi(vsi)) return -ENXIO; if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK)) @@ -621,6 +621,9 @@ ice_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, xdp_ring = vsi->xdp_rings[queue_index]; spin_lock(&xdp_ring->tx_lock); } else { + /* Generally, should not happen */ + if (unlikely(queue_index >= vsi->num_xdp_txq)) + return -ENXIO; xdp_ring = vsi->xdp_rings[queue_index]; } -- cgit v1.2.3