From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 895B1C433FE for ; Tue, 14 Dec 2021 19:00:19 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237277AbhLNTAS (ORCPT ); Tue, 14 Dec 2021 14:00:18 -0500 Received: from mga01.intel.com ([192.55.52.88]:50835 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237279AbhLNTAO (ORCPT ); Tue, 14 Dec 2021 14:00:14 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1639508414; x=1671044414; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=XynYfaHPcwnAoEIBTfMczTgLVzMSgBjVMQiRatUW/Qs=; b=gfkdDpKJAa0dIUya2Nn4FcsHENUdNfFzoVvOwt0aEmJZN6RTgcDknJZT 71cGnUDMQNVQF2+8S1QdMEu6IZf9Hm9ONSfqJxWfQEa99PEUd5mbxOMGC iZIV6JF01SF07Y2uffcMMQYoFqp3U4QOWOBfOaDYocxSfFTH8lxrpx19u PZSUMCy3xxgPl+FRYWJ6w628bIGHudte0RqfYEP0H5FN+ssdGW7Tqi/XV rdVgAUkYr0NR72+HCJBJ/OqAuD8Jn9KdAU6vmxR3IVarWkFW4LrARDOwK ykKRw+qaarCMAReJksYa5pWy5HXtCk5ZhMsqDLn4NfKOivIXMht6K/khh w==; X-IronPort-AV: E=McAfee;i="6200,9189,10197"; a="263200035" X-IronPort-AV: E=Sophos;i="5.88,205,1635231600"; d="scan'208";a="263200035" Received: from fmsmga004.fm.intel.com ([10.253.24.48]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 14 Dec 2021 10:30:11 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.88,205,1635231600"; d="scan'208";a="583712776" Received: from anguy11-desk2.jf.intel.com ([10.166.244.147]) by fmsmga004.fm.intel.com with ESMTP; 14 Dec 2021 10:30:10 -0800 From: Tony Nguyen To: davem@davemloft.net, kuba@kernel.org Cc: Tony Nguyen , netdev@vger.kernel.org, Gurucharan G Subject: [PATCH net-next 08/12] ice: Cleanup after ice_status removal Date: Tue, 14 Dec 2021 10:29:04 -0800 Message-Id: <20211214182908.1513343-9-anthony.l.nguyen@intel.com> X-Mailer: git-send-email 2.31.1 In-Reply-To: <20211214182908.1513343-1-anthony.l.nguyen@intel.com> References: <20211214182908.1513343-1-anthony.l.nguyen@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org Clean up code after changing ice_status to int. Rearrange to fix reverse Christmas tree and pull lines up where applicable. Signed-off-by: Tony Nguyen Tested-by: Gurucharan G --- drivers/net/ethernet/intel/ice/ice_common.c | 45 ++++++----- drivers/net/ethernet/intel/ice/ice_controlq.c | 18 ++--- drivers/net/ethernet/intel/ice/ice_dcb.c | 19 ++--- drivers/net/ethernet/intel/ice/ice_dcb.h | 3 +- drivers/net/ethernet/intel/ice/ice_devlink.c | 4 +- drivers/net/ethernet/intel/ice/ice_ethtool.c | 35 ++++----- .../net/ethernet/intel/ice/ice_ethtool_fdir.c | 4 +- drivers/net/ethernet/intel/ice/ice_fdir.c | 6 +- drivers/net/ethernet/intel/ice/ice_fdir.h | 6 +- .../net/ethernet/intel/ice/ice_flex_pipe.c | 59 +++++++-------- .../net/ethernet/intel/ice/ice_flex_pipe.h | 3 +- drivers/net/ethernet/intel/ice/ice_flow.c | 17 ++--- drivers/net/ethernet/intel/ice/ice_flow.h | 9 +-- drivers/net/ethernet/intel/ice/ice_fltr.c | 25 +++---- drivers/net/ethernet/intel/ice/ice_fltr.h | 6 +- .../net/ethernet/intel/ice/ice_fw_update.c | 22 +++--- drivers/net/ethernet/intel/ice/ice_lib.c | 29 ++++--- drivers/net/ethernet/intel/ice/ice_lib.h | 3 +- drivers/net/ethernet/intel/ice/ice_main.c | 60 +++++++-------- drivers/net/ethernet/intel/ice/ice_nvm.c | 41 +++++----- drivers/net/ethernet/intel/ice/ice_nvm.h | 12 +-- drivers/net/ethernet/intel/ice/ice_sched.c | 54 +++++++------ drivers/net/ethernet/intel/ice/ice_sched.h | 3 +- drivers/net/ethernet/intel/ice/ice_switch.c | 75 +++++++++---------- drivers/net/ethernet/intel/ice/ice_switch.h | 20 ++--- drivers/net/ethernet/intel/ice/ice_tc_lib.c | 2 +- .../ethernet/intel/ice/ice_virtchnl_fdir.c | 4 +- .../net/ethernet/intel/ice/ice_virtchnl_pf.c | 31 ++++---- 28 files changed, 265 insertions(+), 350 deletions(-) diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c index 2f74ab437129..2728cea1a490 100644 --- a/drivers/net/ethernet/intel/ice/ice_common.c +++ b/drivers/net/ethernet/intel/ice/ice_common.c @@ -183,8 +183,8 @@ ice_aq_get_phy_caps(struct ice_port_info *pi, bool qual_mods, u8 report_mode, struct ice_aqc_get_phy_caps *cmd; u16 pcaps_size = sizeof(*pcaps); struct ice_aq_desc desc; - int status; struct ice_hw *hw; + int status; cmd = &desc.params.get_phy; @@ -428,9 +428,9 @@ ice_aq_get_link_info(struct ice_port_info *pi, bool ena_lse, struct ice_fc_info *hw_fc_info; bool tx_pause, rx_pause; struct ice_aq_desc desc; - int status; struct ice_hw *hw; u16 cmd_flags; + int status; if (!pi) return -EINVAL; @@ -668,8 +668,8 @@ static void ice_cleanup_fltr_mgmt_struct(struct ice_hw *hw) static int ice_get_fw_log_cfg(struct ice_hw *hw) { struct ice_aq_desc desc; - int status; __le16 *config; + int status; u16 size; size = sizeof(*config) * ICE_AQC_FW_LOG_ID_MAX; @@ -740,12 +740,12 @@ static int ice_get_fw_log_cfg(struct ice_hw *hw) static int ice_cfg_fw_log(struct ice_hw *hw, bool enable) { struct ice_aqc_fw_logging *cmd; - int status = 0; u16 i, chgs = 0, len = 0; struct ice_aq_desc desc; __le16 *data = NULL; u8 actv_evnts = 0; void *buf = NULL; + int status = 0; if (!hw->fw_log.cq_en && !hw->fw_log.uart_en) return 0; @@ -906,9 +906,9 @@ static void ice_get_itr_intrl_gran(struct ice_hw *hw) int ice_init_hw(struct ice_hw *hw) { struct ice_aqc_get_phy_caps_data *pcaps; - int status; u16 mac_buf_len; void *mac_buf; + int status; /* Set MAC type based on DeviceID */ status = ice_set_mac_type(hw); @@ -1457,11 +1457,11 @@ ice_sq_send_cmd_retry(struct ice_hw *hw, struct ice_ctl_q_info *cq, struct ice_sq_cd *cd) { struct ice_aq_desc desc_cpy; - int status; bool is_cmd_for_retry; u8 *buf_cpy = NULL; u8 idx = 0; u16 opcode; + int status; opcode = le16_to_cpu(desc->opcode); is_cmd_for_retry = ice_should_retry_sq_send_cmd(opcode); @@ -1820,16 +1820,15 @@ ice_acquire_res(struct ice_hw *hw, enum ice_aq_res_ids res, */ void ice_release_res(struct ice_hw *hw, enum ice_aq_res_ids res) { - int status; u32 total_delay = 0; + int status; status = ice_aq_release_res(hw, res, 0, NULL); /* there are some rare cases when trying to release the resource * results in an admin queue timeout, so handle them correctly */ - while ((status == -EIO) && - (total_delay < hw->adminq.sq_cmd_timeout)) { + while ((status == -EIO) && (total_delay < hw->adminq.sq_cmd_timeout)) { mdelay(1); status = ice_aq_release_res(hw, res, 0, NULL); total_delay++; @@ -1884,8 +1883,8 @@ int ice_alloc_hw_res(struct ice_hw *hw, u16 type, u16 num, bool btm, u16 *res) { struct ice_aqc_alloc_free_res_elem *buf; - int status; u16 buf_len; + int status; buf_len = struct_size(buf, elem, num); buf = kzalloc(buf_len, GFP_KERNEL); @@ -1921,8 +1920,8 @@ ice_alloc_hw_res(struct ice_hw *hw, u16 type, u16 num, bool btm, u16 *res) int ice_free_hw_res(struct ice_hw *hw, u16 type, u16 num, u16 *res) { struct ice_aqc_alloc_free_res_elem *buf; - int status; u16 buf_len; + int status; buf_len = struct_size(buf, elem, num); buf = kzalloc(buf_len, GFP_KERNEL); @@ -2518,9 +2517,9 @@ ice_aq_list_caps(struct ice_hw *hw, void *buf, u16 buf_size, u32 *cap_count, int ice_discover_dev_caps(struct ice_hw *hw, struct ice_hw_dev_caps *dev_caps) { - int status; u32 cap_count = 0; void *cbuf; + int status; cbuf = kzalloc(ICE_AQ_MAX_BUF_LEN, GFP_KERNEL); if (!cbuf) @@ -2552,9 +2551,9 @@ ice_discover_dev_caps(struct ice_hw *hw, struct ice_hw_dev_caps *dev_caps) static int ice_discover_func_caps(struct ice_hw *hw, struct ice_hw_func_caps *func_caps) { - int status; u32 cap_count = 0; void *cbuf; + int status; cbuf = kzalloc(ICE_AQ_MAX_BUF_LEN, GFP_KERNEL); if (!cbuf) @@ -3120,8 +3119,8 @@ ice_set_fc(struct ice_port_info *pi, u8 *aq_failures, bool ena_auto_link_update) { struct ice_aqc_set_phy_cfg_data cfg = { 0 }; struct ice_aqc_get_phy_caps_data *pcaps; - int status; struct ice_hw *hw; + int status; if (!pi || !aq_failures) return -EINVAL; @@ -3261,8 +3260,8 @@ ice_cfg_phy_fec(struct ice_port_info *pi, struct ice_aqc_set_phy_cfg_data *cfg, enum ice_fec_mode fec) { struct ice_aqc_get_phy_caps_data *pcaps; - int status; struct ice_hw *hw; + int status; if (!pi || !cfg) return -EINVAL; @@ -3652,9 +3651,9 @@ ice_aq_set_rss_lut(struct ice_hw *hw, struct ice_aq_get_set_rss_lut_params *set_ * * get (0x0B04) or set (0x0B02) the RSS key per VSI */ -static int __ice_aq_get_set_rss_key(struct ice_hw *hw, u16 vsi_id, - struct ice_aqc_get_set_rss_keys *key, - bool set) +static int +__ice_aq_get_set_rss_key(struct ice_hw *hw, u16 vsi_id, + struct ice_aqc_get_set_rss_keys *key, bool set) { struct ice_aqc_get_set_rss_key *cmd_resp; u16 key_size = sizeof(*key); @@ -3793,8 +3792,8 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps, struct ice_aqc_dis_txq_item *item; struct ice_aqc_dis_txqs *cmd; struct ice_aq_desc desc; - int status; u16 i, sz = 0; + int status; cmd = &desc.params.dis_txqs; ice_fill_dflt_direct_cmd_desc(&desc, ice_aqc_opc_dis_txqs); @@ -4189,8 +4188,8 @@ ice_ena_vsi_txq(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u16 q_handle, struct ice_aqc_txsched_elem_data node = { 0 }; struct ice_sched_node *parent; struct ice_q_ctx *q_ctx; - int status; struct ice_hw *hw; + int status; if (!pi || pi->port_state != ICE_SCHED_PORT_STATE_READY) return -EIO; @@ -4292,9 +4291,9 @@ ice_dis_vsi_txq(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u8 num_queues, enum ice_disq_rst_src rst_src, u16 vmvf_num, struct ice_sq_cd *cd) { - int status = -ENOENT; struct ice_aqc_dis_txq_item *qg_list; struct ice_q_ctx *q_ctx; + int status = -ENOENT; struct ice_hw *hw; u16 i, buf_size; @@ -4446,9 +4445,9 @@ ice_ena_vsi_rdma_qset(struct ice_port_info *pi, u16 vsi_handle, u8 tc, struct ice_aqc_txsched_elem_data node = { 0 }; struct ice_aqc_add_rdma_qset_data *buf; struct ice_sched_node *parent; - int status; struct ice_hw *hw; u16 i, buf_size; + int status; int ret; if (!pi || pi->port_state != ICE_SCHED_PORT_STATE_READY) @@ -4523,8 +4522,8 @@ ice_dis_vsi_rdma_qset(struct ice_port_info *pi, u16 count, u32 *qset_teid, u16 *q_id) { struct ice_aqc_dis_txq_item *qg_list; - int status = 0; struct ice_hw *hw; + int status = 0; u16 qg_size; int i; diff --git a/drivers/net/ethernet/intel/ice/ice_controlq.c b/drivers/net/ethernet/intel/ice/ice_controlq.c index 5b0c1456698e..6bcfee295991 100644 --- a/drivers/net/ethernet/intel/ice/ice_controlq.c +++ b/drivers/net/ethernet/intel/ice/ice_controlq.c @@ -295,8 +295,7 @@ ice_cfg_cq_regs(struct ice_hw *hw, struct ice_ctl_q_ring *ring, u16 num_entries) * * Configure base address and length registers for the transmit queue */ -static int -ice_cfg_sq_regs(struct ice_hw *hw, struct ice_ctl_q_info *cq) +static int ice_cfg_sq_regs(struct ice_hw *hw, struct ice_ctl_q_info *cq) { return ice_cfg_cq_regs(hw, &cq->sq, cq->num_sq_entries); } @@ -308,8 +307,7 @@ ice_cfg_sq_regs(struct ice_hw *hw, struct ice_ctl_q_info *cq) * * Configure base address and length registers for the receive (event queue) */ -static int -ice_cfg_rq_regs(struct ice_hw *hw, struct ice_ctl_q_info *cq) +static int ice_cfg_rq_regs(struct ice_hw *hw, struct ice_ctl_q_info *cq) { int status; @@ -474,8 +472,7 @@ static int ice_init_rq(struct ice_hw *hw, struct ice_ctl_q_info *cq) * * The main shutdown routine for the Control Transmit Queue */ -static int -ice_shutdown_sq(struct ice_hw *hw, struct ice_ctl_q_info *cq) +static int ice_shutdown_sq(struct ice_hw *hw, struct ice_ctl_q_info *cq) { int ret_code = 0; @@ -541,8 +538,7 @@ static bool ice_aq_ver_check(struct ice_hw *hw) * * The main shutdown routine for the Control Receive Queue */ -static int -ice_shutdown_rq(struct ice_hw *hw, struct ice_ctl_q_info *cq) +static int ice_shutdown_rq(struct ice_hw *hw, struct ice_ctl_q_info *cq) { int ret_code = 0; @@ -753,8 +749,8 @@ void ice_shutdown_all_ctrlq(struct ice_hw *hw) */ int ice_init_all_ctrlq(struct ice_hw *hw) { - int status; u32 retry = 0; + int status; /* Init FW admin queue */ do { @@ -970,9 +966,9 @@ ice_sq_send_cmd(struct ice_hw *hw, struct ice_ctl_q_info *cq, struct ice_dma_mem *dma_buf = NULL; struct ice_aq_desc *desc_on_ring; bool cmd_completed = false; - int status = 0; struct ice_sq_cd *details; u32 total_delay = 0; + int status = 0; u16 retval = 0; u32 val = 0; @@ -1160,9 +1156,9 @@ ice_clean_rq_elem(struct ice_hw *hw, struct ice_ctl_q_info *cq, { u16 ntc = cq->rq.next_to_clean; enum ice_aq_err rq_last_status; - int ret_code = 0; struct ice_aq_desc *desc; struct ice_dma_mem *bi; + int ret_code = 0; u16 desc_idx; u16 datalen; u16 flags; diff --git a/drivers/net/ethernet/intel/ice/ice_dcb.c b/drivers/net/ethernet/intel/ice/ice_dcb.c index 14cb301210fc..0b146a0d4205 100644 --- a/drivers/net/ethernet/intel/ice/ice_dcb.c +++ b/drivers/net/ethernet/intel/ice/ice_dcb.c @@ -116,8 +116,7 @@ ice_aq_stop_lldp(struct ice_hw *hw, bool shutdown_lldp_agent, bool persist, * * Start the embedded LLDP Agent on all ports. (0x0A06) */ -int -ice_aq_start_lldp(struct ice_hw *hw, bool persist, struct ice_sq_cd *cd) +int ice_aq_start_lldp(struct ice_hw *hw, bool persist, struct ice_sq_cd *cd) { struct ice_aqc_lldp_start *cmd; struct ice_aq_desc desc; @@ -597,12 +596,11 @@ ice_parse_org_tlv(struct ice_lldp_org_tlv *tlv, struct ice_dcbx_cfg *dcbcfg) * * Parse DCB configuration from the LLDPDU */ -static int -ice_lldp_to_dcb_cfg(u8 *lldpmib, struct ice_dcbx_cfg *dcbcfg) +static int ice_lldp_to_dcb_cfg(u8 *lldpmib, struct ice_dcbx_cfg *dcbcfg) { struct ice_lldp_org_tlv *tlv; - int ret = 0; u16 offset = 0; + int ret = 0; u16 typelen; u16 type; u16 len; @@ -652,8 +650,8 @@ int ice_aq_get_dcb_cfg(struct ice_hw *hw, u8 mib_type, u8 bridgetype, struct ice_dcbx_cfg *dcbcfg) { - int ret; u8 *lldpmib; + int ret; /* Allocate the LLDPDU */ lldpmib = devm_kzalloc(ice_hw_to_dev(hw), ICE_LLDPDU_SIZE, GFP_KERNEL); @@ -691,9 +689,9 @@ ice_aq_start_stop_dcbx(struct ice_hw *hw, bool start_dcbx_agent, bool *dcbx_agent_status, struct ice_sq_cd *cd) { struct ice_aqc_lldp_stop_start_specific_agent *cmd; - int status; struct ice_aq_desc desc; u16 opcode; + int status; cmd = &desc.params.lldp_agent_ctrl; @@ -902,8 +900,7 @@ ice_cee_to_dcb_cfg(struct ice_aqc_get_cee_dcb_cfg_resp *cee_cfg, * * Get IEEE or CEE mode DCB configuration from the Firmware */ -static int -ice_get_ieee_or_cee_dcb_cfg(struct ice_port_info *pi, u8 dcbx_mode) +static int ice_get_ieee_or_cee_dcb_cfg(struct ice_port_info *pi, u8 dcbx_mode) { struct ice_dcbx_cfg *dcbx_cfg = NULL; int ret; @@ -1472,9 +1469,9 @@ int ice_set_dcb_cfg(struct ice_port_info *pi) { u8 mib_type, *lldpmib = NULL; struct ice_dcbx_cfg *dcbcfg; - int ret; struct ice_hw *hw; u16 miblen; + int ret; if (!pi) return -EINVAL; @@ -1542,8 +1539,8 @@ ice_update_port_tc_tree_cfg(struct ice_port_info *pi, { struct ice_sched_node *node, *tc_node; struct ice_aqc_txsched_elem_data elem; - int status = 0; u32 teid1, teid2; + int status = 0; u8 i, j; if (!pi) diff --git a/drivers/net/ethernet/intel/ice/ice_dcb.h b/drivers/net/ethernet/intel/ice/ice_dcb.h index 4eb6383c363e..d73348f279f7 100644 --- a/drivers/net/ethernet/intel/ice/ice_dcb.h +++ b/drivers/net/ethernet/intel/ice/ice_dcb.h @@ -152,8 +152,7 @@ ice_query_port_ets(struct ice_port_info *pi, int ice_aq_stop_lldp(struct ice_hw *hw, bool shutdown_lldp_agent, bool persist, struct ice_sq_cd *cd); -int -ice_aq_start_lldp(struct ice_hw *hw, bool persist, struct ice_sq_cd *cd); +int ice_aq_start_lldp(struct ice_hw *hw, bool persist, struct ice_sq_cd *cd); int ice_aq_start_stop_dcbx(struct ice_hw *hw, bool start_dcbx_agent, bool *dcbx_agent_status, struct ice_sq_cd *cd); diff --git a/drivers/net/ethernet/intel/ice/ice_devlink.c b/drivers/net/ethernet/intel/ice/ice_devlink.c index e838494bf060..dde112e74d41 100644 --- a/drivers/net/ethernet/intel/ice/ice_devlink.c +++ b/drivers/net/ethernet/intel/ice/ice_devlink.c @@ -762,9 +762,9 @@ static int ice_devlink_nvm_snapshot(struct devlink *devlink, struct ice_pf *pf = devlink_priv(devlink); struct device *dev = ice_pf_to_dev(pf); struct ice_hw *hw = &pf->hw; - int status; void *nvm_data; u32 nvm_size; + int status; nvm_size = hw->flash.flash_size; nvm_data = vzalloc(nvm_size); @@ -819,8 +819,8 @@ ice_devlink_devcaps_snapshot(struct devlink *devlink, struct ice_pf *pf = devlink_priv(devlink); struct device *dev = ice_pf_to_dev(pf); struct ice_hw *hw = &pf->hw; - int status; void *devcaps; + int status; devcaps = vzalloc(ICE_AQ_MAX_BUF_LEN); if (!devcaps) diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c index ca7e0ccc81bc..900572181f69 100644 --- a/drivers/net/ethernet/intel/ice/ice_ethtool.c +++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c @@ -270,9 +270,9 @@ ice_get_eeprom(struct net_device *netdev, struct ethtool_eeprom *eeprom, struct ice_vsi *vsi = np->vsi; struct ice_pf *pf = vsi->back; struct ice_hw *hw = &pf->hw; - int status; struct device *dev; int ret = 0; + int status; u8 *buf; dev = ice_pf_to_dev(pf); @@ -288,8 +288,7 @@ ice_get_eeprom(struct net_device *netdev, struct ethtool_eeprom *eeprom, status = ice_acquire_nvm(hw, ICE_RES_READ); if (status) { dev_err(dev, "ice_acquire_nvm failed, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); ret = -EIO; goto out; } @@ -298,8 +297,7 @@ ice_get_eeprom(struct net_device *netdev, struct ethtool_eeprom *eeprom, false); if (status) { dev_err(dev, "ice_read_flat_nvm failed, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); ret = -EIO; goto release; } @@ -342,8 +340,8 @@ static bool ice_active_vfs(struct ice_pf *pf) static u64 ice_link_test(struct net_device *netdev) { struct ice_netdev_priv *np = netdev_priv(netdev); - int status; bool link_up = false; + int status; netdev_info(netdev, "link test\n"); status = ice_get_link_status(np->vsi->port_info, &link_up); @@ -1052,8 +1050,8 @@ ice_get_fecparam(struct net_device *netdev, struct ethtool_fecparam *fecparam) struct ice_link_status *link_info; struct ice_vsi *vsi = np->vsi; struct ice_port_info *pi; - int status; int err = 0; + int status; pi = vsi->port_info; @@ -1232,8 +1230,8 @@ static int ice_set_priv_flags(struct net_device *netdev, u32 flags) pf->dcbx_cap &= ~DCB_CAP_DCBX_LLD_MANAGED; pf->dcbx_cap |= DCB_CAP_DCBX_HOST; } else { - int status; bool dcbx_agent_status; + int status; if (ice_get_pfc_mode(pf) == ICE_QOS_MODE_DSCP) { clear_bit(ICE_FLAG_FW_LLDP_AGENT, pf->flags); @@ -1938,8 +1936,8 @@ ice_get_link_ksettings(struct net_device *netdev, struct ice_aqc_get_phy_caps_data *caps; struct ice_link_status *hw_link_info; struct ice_vsi *vsi = np->vsi; - int status; int err = 0; + int status; ethtool_link_ksettings_zero_link_mode(ks, supported); ethtool_link_ksettings_zero_link_mode(ks, advertising); @@ -2210,11 +2208,11 @@ ice_set_link_ksettings(struct net_device *netdev, struct ice_pf *pf = np->vsi->back; struct ice_port_info *pi; u8 autoneg_changed = 0; - int status; u64 phy_type_high = 0; u64 phy_type_low = 0; int err = 0; bool linkup; + int status; pi = np->vsi->port_info; @@ -2522,9 +2520,9 @@ static int ice_set_rss_hash_opt(struct ice_vsi *vsi, struct ethtool_rxnfc *nfc) { struct ice_pf *pf = vsi->back; - int status; struct device *dev; u64 hashed_flds; + int status; u32 hdrs; dev = ice_pf_to_dev(pf); @@ -3003,10 +3001,10 @@ ice_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause) struct ice_vsi *vsi = np->vsi; struct ice_hw *hw = &pf->hw; struct ice_port_info *pi; - int status; u8 aq_failures; bool link_up; int err = 0; + int status; u32 is_an; pi = vsi->port_info; @@ -3077,18 +3075,15 @@ ice_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause) if (aq_failures & ICE_SET_FC_AQ_FAIL_GET) { netdev_info(netdev, "Set fc failed on the get_phy_capabilities call with err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); err = -EAGAIN; } else if (aq_failures & ICE_SET_FC_AQ_FAIL_SET) { netdev_info(netdev, "Set fc failed on the set_phy_config call with err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); err = -EAGAIN; } else if (aq_failures & ICE_SET_FC_AQ_FAIL_UPDATE) { netdev_info(netdev, "Set fc failed on the get_link_info call with err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); err = -EAGAIN; } @@ -3928,11 +3923,11 @@ ice_get_module_info(struct net_device *netdev, struct ice_vsi *vsi = np->vsi; struct ice_pf *pf = vsi->back; struct ice_hw *hw = &pf->hw; - int status; u8 sff8472_comp = 0; u8 sff8472_swap = 0; u8 sff8636_rev = 0; u8 value = 0; + int status; status = ice_aq_sff_eeprom(hw, 0, ICE_I2C_EEPROM_DEV_ADDR, 0x00, 0x00, 0, &value, 1, 0, NULL); @@ -4005,11 +4000,11 @@ ice_get_module_eeprom(struct net_device *netdev, struct ice_vsi *vsi = np->vsi; struct ice_pf *pf = vsi->back; struct ice_hw *hw = &pf->hw; - int status; bool is_sfp = false; unsigned int i, j; u16 offset = 0; u8 page = 0; + int status; if (!ee || !ee->len || !data) return -EINVAL; diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool_fdir.c b/drivers/net/ethernet/intel/ice/ice_ethtool_fdir.c index 226ec42abbc7..deb4e3ad38f6 100644 --- a/drivers/net/ethernet/intel/ice/ice_ethtool_fdir.c +++ b/drivers/net/ethernet/intel/ice/ice_ethtool_fdir.c @@ -530,10 +530,10 @@ ice_fdir_set_hw_fltr_rule(struct ice_pf *pf, struct ice_flow_seg_info *seg, struct ice_flow_prof *prof = NULL; struct ice_fd_hw_prof *hw_prof; struct ice_hw *hw = &pf->hw; - int status; u64 entry1_h = 0; u64 entry2_h = 0; u64 prof_id; + int status; int err; main_vsi = ice_get_main_vsi(pf); @@ -1190,9 +1190,9 @@ ice_fdir_write_fltr(struct ice_pf *pf, struct ice_fdir_fltr *input, bool add, struct ice_hw *hw = &pf->hw; struct ice_fltr_desc desc; struct ice_vsi *ctrl_vsi; - int status; u8 *pkt, *frag_pkt; bool has_frag; + int status; int err; ctrl_vsi = ice_get_ctrl_vsi(pf); diff --git a/drivers/net/ethernet/intel/ice/ice_fdir.c b/drivers/net/ethernet/intel/ice/ice_fdir.c index bbef6f4366de..ae089d32ee9d 100644 --- a/drivers/net/ethernet/intel/ice/ice_fdir.c +++ b/drivers/net/ethernet/intel/ice/ice_fdir.c @@ -735,8 +735,7 @@ int ice_free_fd_res_cntr(struct ice_hw *hw, u16 cntr_id) * @cntr_id: returns counter index * @num_fltr: number of filter entries to be allocated */ -int -ice_alloc_fd_guar_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr) +int ice_alloc_fd_guar_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr) { return ice_alloc_res_cntr(hw, ICE_AQC_RES_TYPE_FDIR_GUARANTEED_ENTRIES, ICE_AQC_RES_TYPE_FLAG_DEDICATED, num_fltr, @@ -749,8 +748,7 @@ ice_alloc_fd_guar_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr) * @cntr_id: returns counter index * @num_fltr: number of filter entries to be allocated */ -int -ice_alloc_fd_shrd_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr) +int ice_alloc_fd_shrd_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr) { return ice_alloc_res_cntr(hw, ICE_AQC_RES_TYPE_FDIR_SHARED_ENTRIES, ICE_AQC_RES_TYPE_FLAG_DEDICATED, num_fltr, diff --git a/drivers/net/ethernet/intel/ice/ice_fdir.h b/drivers/net/ethernet/intel/ice/ice_fdir.h index 49a2453731d2..b6c7c6903f35 100644 --- a/drivers/net/ethernet/intel/ice/ice_fdir.h +++ b/drivers/net/ethernet/intel/ice/ice_fdir.h @@ -203,10 +203,8 @@ struct ice_fdir_base_pkt { int ice_alloc_fd_res_cntr(struct ice_hw *hw, u16 *cntr_id); int ice_free_fd_res_cntr(struct ice_hw *hw, u16 cntr_id); -int -ice_alloc_fd_guar_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr); -int -ice_alloc_fd_shrd_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr); +int ice_alloc_fd_guar_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr); +int ice_alloc_fd_shrd_item(struct ice_hw *hw, u16 *cntr_id, u16 num_fltr); void ice_fdir_get_prgm_desc(struct ice_hw *hw, struct ice_fdir_fltr *input, struct ice_fltr_desc *fdesc, bool add); diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c index b52ea684ee9c..d29197ab3d02 100644 --- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.c +++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.c @@ -964,11 +964,10 @@ ice_find_seg_in_pkg(struct ice_hw *hw, u32 seg_type, * * Obtains change lock and updates package. */ -static int -ice_update_pkg(struct ice_hw *hw, struct ice_buf *bufs, u32 count) +static int ice_update_pkg(struct ice_hw *hw, struct ice_buf *bufs, u32 count) { - int status; u32 offset, info, i; + int status; status = ice_acquire_change_lock(hw, ICE_RES_WRITE); if (status) @@ -1023,10 +1022,10 @@ static enum ice_ddp_state ice_dwnld_cfg_bufs(struct ice_hw *hw, struct ice_buf *bufs, u32 count) { enum ice_ddp_state state = ICE_DDP_PKG_SUCCESS; - int status; struct ice_buf_hdr *bh; enum ice_aq_err err; u32 offset, info, i; + int status; if (!bufs || !count) return ICE_DDP_PKG_ERR; @@ -2101,8 +2100,8 @@ ice_create_tunnel(struct ice_hw *hw, u16 index, enum ice_tunnel_type type, u16 port) { struct ice_boost_tcam_section *sect_rx, *sect_tx; - int status = -ENOSPC; struct ice_buf_build *bld; + int status = -ENOSPC; mutex_lock(&hw->tnl_lock); @@ -2172,8 +2171,8 @@ ice_destroy_tunnel(struct ice_hw *hw, u16 index, enum ice_tunnel_type type, u16 port) { struct ice_boost_tcam_section *sect_rx, *sect_tx; - int status = -ENOSPC; struct ice_buf_build *bld; + int status = -ENOSPC; mutex_lock(&hw->tnl_lock); @@ -2399,8 +2398,8 @@ ice_ptg_remove_ptype(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 ptg) static int ice_ptg_add_mv_ptype(struct ice_hw *hw, enum ice_block blk, u16 ptype, u8 ptg) { - int status; u8 original_ptg; + int status; if (ptype > ICE_XLT1_CNT - 1) return -EINVAL; @@ -2636,8 +2635,7 @@ ice_find_dup_props_vsig(struct ice_hw *hw, enum ice_block blk, * The function will remove all VSIs associated with the input VSIG and move * them to the DEFAULT_VSIG and mark the VSIG available. */ -static int -ice_vsig_free(struct ice_hw *hw, enum ice_block blk, u16 vsig) +static int ice_vsig_free(struct ice_hw *hw, enum ice_block blk, u16 vsig) { struct ice_vsig_prof *dtmp, *del; struct ice_vsig_vsi *vsi_cur; @@ -2759,8 +2757,8 @@ static int ice_vsig_add_mv_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig) { struct ice_vsig_vsi *tmp; - int status; u16 orig_vsig, idx; + int status; idx = vsig & ICE_VSIG_IDX_M; @@ -2998,12 +2996,11 @@ ice_free_tcam_ent(struct ice_hw *hw, enum ice_block blk, u16 tcam_idx) * This function allocates a new profile ID, which also corresponds to a Field * Vector (Extraction Sequence) entry. */ -static int -ice_alloc_prof_id(struct ice_hw *hw, enum ice_block blk, u8 *prof_id) +static int ice_alloc_prof_id(struct ice_hw *hw, enum ice_block blk, u8 *prof_id) { - int status; u16 res_type; u16 get_prof; + int status; if (!ice_prof_id_rsrc_type(blk, &res_type)) return -EINVAL; @@ -3023,8 +3020,7 @@ ice_alloc_prof_id(struct ice_hw *hw, enum ice_block blk, u8 *prof_id) * * This function frees a profile ID, which also corresponds to a Field Vector. */ -static int -ice_free_prof_id(struct ice_hw *hw, enum ice_block blk, u8 prof_id) +static int ice_free_prof_id(struct ice_hw *hw, enum ice_block blk, u8 prof_id) { u16 tmp_prof_id = (u16)prof_id; u16 res_type; @@ -3041,8 +3037,7 @@ ice_free_prof_id(struct ice_hw *hw, enum ice_block blk, u8 prof_id) * @blk: the block from which to free the profile ID * @prof_id: the profile ID for which to increment the reference count */ -static int -ice_prof_inc_ref(struct ice_hw *hw, enum ice_block blk, u8 prof_id) +static int ice_prof_inc_ref(struct ice_hw *hw, enum ice_block blk, u8 prof_id) { if (prof_id > hw->blk[blk].es.count) return -EINVAL; @@ -3167,8 +3162,8 @@ ice_alloc_prof_mask(struct ice_hw *hw, enum ice_block blk, u16 idx, u16 mask, u16 *mask_idx) { bool found_unused = false, found_copy = false; - int status = -ENOSPC; u16 unused_idx = 0, copy_idx = 0; + int status = -ENOSPC; u16 i; if (blk != ICE_BLK_RSS && blk != ICE_BLK_FD) @@ -4267,12 +4262,12 @@ ice_upd_prof_hw(struct ice_hw *hw, enum ice_block blk, { struct ice_buf_build *b; struct ice_chs_chg *tmp; - int status; u16 pkg_sects; u16 xlt1 = 0; u16 xlt2 = 0; u16 tcam = 0; u16 es = 0; + int status; u16 sects; /* count number of sections we need */ @@ -4638,9 +4633,9 @@ ice_add_prof(struct ice_hw *hw, enum ice_block blk, u64 id, u8 ptypes[], u32 bytes = DIV_ROUND_UP(ICE_FLOW_PTYPE_MAX, BITS_PER_BYTE); DECLARE_BITMAP(ptgs_used, ICE_XLT1_CNT); struct ice_prof_map *prof; - int status; u8 byte = 0; u8 prof_id; + int status; bitmap_zero(ptgs_used, ICE_XLT1_CNT); @@ -4798,8 +4793,7 @@ ice_vsig_prof_id_count(struct ice_hw *hw, enum ice_block blk, u16 vsig) * @blk: hardware block * @idx: the index to release */ -static int -ice_rel_tcam_idx(struct ice_hw *hw, enum ice_block blk, u16 idx) +static int ice_rel_tcam_idx(struct ice_hw *hw, enum ice_block blk, u16 idx) { /* Masks to invoke a never match entry */ u8 vl_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF }; @@ -4941,12 +4935,11 @@ ice_rem_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, * @blk: hardware block * @id: profile tracking ID */ -static int -ice_rem_flow_all(struct ice_hw *hw, enum ice_block blk, u64 id) +static int ice_rem_flow_all(struct ice_hw *hw, enum ice_block blk, u64 id) { struct ice_chs_chg *del, *tmp; - int status; struct list_head chg; + int status; u16 i; INIT_LIST_HEAD(&chg); @@ -5022,9 +5015,9 @@ static int ice_get_prof(struct ice_hw *hw, enum ice_block blk, u64 hdl, struct list_head *chg) { - int status = 0; struct ice_prof_map *map; struct ice_chs_chg *p; + int status = 0; u16 i; mutex_lock(&hw->blk[blk].es.prof_map_lock); @@ -5115,9 +5108,9 @@ static int ice_add_prof_to_lst(struct ice_hw *hw, enum ice_block blk, struct list_head *lst, u64 hdl) { - int status = 0; struct ice_prof_map *map; struct ice_vsig_prof *p; + int status = 0; u16 i; mutex_lock(&hw->blk[blk].es.prof_map_lock); @@ -5162,9 +5155,9 @@ static int ice_move_vsi(struct ice_hw *hw, enum ice_block blk, u16 vsi, u16 vsig, struct list_head *chg) { - int status; struct ice_chs_chg *p; u16 orig_vsig; + int status; p = devm_kzalloc(ice_hw_to_dev(hw), sizeof(*p), GFP_KERNEL); if (!p) @@ -5223,8 +5216,8 @@ ice_prof_tcam_ena_dis(struct ice_hw *hw, enum ice_block blk, bool enable, u16 vsig, struct ice_tcam_inf *tcam, struct list_head *chg) { - int status; struct ice_chs_chg *p; + int status; u8 vl_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF }; u8 dc_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0x00, 0x00, 0x00 }; @@ -5370,11 +5363,11 @@ ice_add_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsig, u64 hdl, u8 vl_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF }; u8 dc_msk[ICE_TCAM_KEY_VAL_SZ] = { 0xFF, 0xFF, 0x00, 0x00, 0x00 }; u8 nm_msk[ICE_TCAM_KEY_VAL_SZ] = { 0x00, 0x00, 0x00, 0x00, 0x00 }; - int status = 0; struct ice_prof_map *map; struct ice_vsig_prof *t; struct ice_chs_chg *p; u16 vsig_idx, i; + int status = 0; /* Error, if this VSIG already has this profile */ if (ice_has_prof_vsig(hw, blk, vsig, hdl)) @@ -5478,9 +5471,9 @@ static int ice_create_prof_id_vsig(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl, struct list_head *chg) { - int status; struct ice_chs_chg *p; u16 new_vsig; + int status; p = devm_kzalloc(ice_hw_to_dev(hw), sizeof(*p), GFP_KERNEL); if (!p) @@ -5565,8 +5558,8 @@ static bool ice_find_prof_vsig(struct ice_hw *hw, enum ice_block blk, u64 hdl, u16 *vsig) { struct ice_vsig_prof *t; - int status; struct list_head lst; + int status; INIT_LIST_HEAD(&lst); @@ -5602,8 +5595,8 @@ ice_add_prof_id_flow(struct ice_hw *hw, enum ice_block blk, u16 vsi, u64 hdl) struct ice_vsig_prof *tmp1, *del1; struct ice_chs_chg *tmp, *del; struct list_head union_lst; - int status; struct list_head chg; + int status; u16 vsig; INIT_LIST_HEAD(&union_lst); diff --git a/drivers/net/ethernet/intel/ice/ice_flex_pipe.h b/drivers/net/ethernet/intel/ice/ice_flex_pipe.h index dd602285c78e..6cbc29bcb02f 100644 --- a/drivers/net/ethernet/intel/ice/ice_flex_pipe.h +++ b/drivers/net/ethernet/intel/ice/ice_flex_pipe.h @@ -118,6 +118,5 @@ void ice_free_seg(struct ice_hw *hw); void ice_fill_blk_tbls(struct ice_hw *hw); void ice_clear_hw_tbls(struct ice_hw *hw); void ice_free_hw_tbls(struct ice_hw *hw); -int -ice_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 id); +int ice_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 id); #endif /* _ICE_FLEX_PIPE_H_ */ diff --git a/drivers/net/ethernet/intel/ice/ice_flow.c b/drivers/net/ethernet/intel/ice/ice_flow.c index ee6a8fab4c4c..b17b984196a2 100644 --- a/drivers/net/ethernet/intel/ice/ice_flow.c +++ b/drivers/net/ethernet/intel/ice/ice_flow.c @@ -625,8 +625,7 @@ struct ice_flow_prof_params { * @segs: array of one or more packet segments that describe the flow * @segs_cnt: number of packet segments provided */ -static int -ice_flow_val_hdrs(struct ice_flow_seg_info *segs, u8 segs_cnt) +static int ice_flow_val_hdrs(struct ice_flow_seg_info *segs, u8 segs_cnt) { u8 i; @@ -700,8 +699,7 @@ static u16 ice_flow_calc_seg_sz(struct ice_flow_prof_params *params, u8 seg) * This function identifies the packet types associated with the protocol * headers being present in packet segments of the specified flow profile. */ -static int -ice_flow_proc_seg_hdrs(struct ice_flow_prof_params *params) +static int ice_flow_proc_seg_hdrs(struct ice_flow_prof_params *params) { struct ice_flow_prof *prof; u8 i; @@ -1574,8 +1572,7 @@ ice_flow_add_prof(struct ice_hw *hw, enum ice_block blk, enum ice_flow_dir dir, * @blk: the block for which the flow profile is to be removed * @prof_id: unique ID of the flow profile to be removed */ -int -ice_flow_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 prof_id) +int ice_flow_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 prof_id) { struct ice_flow_prof *prof; int status; @@ -1680,8 +1677,7 @@ ice_flow_add_entry(struct ice_hw *hw, enum ice_block blk, u64 prof_id, * @blk: classification stage * @entry_h: handle to the flow entry to be removed */ -int ice_flow_rem_entry(struct ice_hw *hw, enum ice_block blk, - u64 entry_h) +int ice_flow_rem_entry(struct ice_hw *hw, enum ice_block blk, u64 entry_h) { struct ice_flow_entry *entry; struct ice_flow_prof *prof; @@ -2279,8 +2275,7 @@ ice_rem_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds, * message, convert it to ICE-compatible values, and configure RSS flow * profiles. */ -int -ice_add_avf_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 avf_hash) +int ice_add_avf_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 avf_hash) { int status = 0; u64 hash_flds; @@ -2370,8 +2365,8 @@ ice_add_avf_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 avf_hash) */ int ice_replay_rss_cfg(struct ice_hw *hw, u16 vsi_handle) { - int status = 0; struct ice_rss_cfg *r; + int status = 0; if (!ice_is_vsi_valid(hw, vsi_handle)) return -EINVAL; diff --git a/drivers/net/ethernet/intel/ice/ice_flow.h b/drivers/net/ethernet/intel/ice/ice_flow.h index 818986437001..d8782b28323e 100644 --- a/drivers/net/ethernet/intel/ice/ice_flow.h +++ b/drivers/net/ethernet/intel/ice/ice_flow.h @@ -387,14 +387,12 @@ int ice_flow_add_prof(struct ice_hw *hw, enum ice_block blk, enum ice_flow_dir dir, u64 prof_id, struct ice_flow_seg_info *segs, u8 segs_cnt, struct ice_flow_prof **prof); -int -ice_flow_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 prof_id); +int ice_flow_rem_prof(struct ice_hw *hw, enum ice_block blk, u64 prof_id); int ice_flow_add_entry(struct ice_hw *hw, enum ice_block blk, u64 prof_id, u64 entry_id, u16 vsi, enum ice_flow_priority prio, void *data, u64 *entry_h); -int -ice_flow_rem_entry(struct ice_hw *hw, enum ice_block blk, u64 entry_h); +int ice_flow_rem_entry(struct ice_hw *hw, enum ice_block blk, u64 entry_h); void ice_flow_set_fld(struct ice_flow_seg_info *seg, enum ice_flow_field fld, u16 val_loc, u16 mask_loc, u16 last_loc, bool range); @@ -403,8 +401,7 @@ ice_flow_add_fld_raw(struct ice_flow_seg_info *seg, u16 off, u8 len, u16 val_loc, u16 mask_loc); void ice_rem_vsi_rss_list(struct ice_hw *hw, u16 vsi_handle); int ice_replay_rss_cfg(struct ice_hw *hw, u16 vsi_handle); -int -ice_add_avf_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds); +int ice_add_avf_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds); int ice_rem_vsi_rss_cfg(struct ice_hw *hw, u16 vsi_handle); int ice_add_rss_cfg(struct ice_hw *hw, u16 vsi_handle, u64 hashed_flds, diff --git a/drivers/net/ethernet/intel/ice/ice_fltr.c b/drivers/net/ethernet/intel/ice/ice_fltr.c index 52d9d78039e2..94903b450345 100644 --- a/drivers/net/ethernet/intel/ice/ice_fltr.c +++ b/drivers/net/ethernet/intel/ice/ice_fltr.c @@ -109,8 +109,7 @@ ice_fltr_set_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask, * @vsi: pointer to VSI struct * @list: list of filters */ -int -ice_fltr_add_mac_list(struct ice_vsi *vsi, struct list_head *list) +int ice_fltr_add_mac_list(struct ice_vsi *vsi, struct list_head *list) { return ice_add_mac(&vsi->back->hw, list); } @@ -120,8 +119,7 @@ ice_fltr_add_mac_list(struct ice_vsi *vsi, struct list_head *list) * @vsi: pointer to VSI struct * @list: list of filters */ -int -ice_fltr_remove_mac_list(struct ice_vsi *vsi, struct list_head *list) +int ice_fltr_remove_mac_list(struct ice_vsi *vsi, struct list_head *list) { return ice_remove_mac(&vsi->back->hw, list); } @@ -131,8 +129,7 @@ ice_fltr_remove_mac_list(struct ice_vsi *vsi, struct list_head *list) * @vsi: pointer to VSI struct * @list: list of filters */ -static int -ice_fltr_add_vlan_list(struct ice_vsi *vsi, struct list_head *list) +static int ice_fltr_add_vlan_list(struct ice_vsi *vsi, struct list_head *list) { return ice_add_vlan(&vsi->back->hw, list); } @@ -153,8 +150,7 @@ ice_fltr_remove_vlan_list(struct ice_vsi *vsi, struct list_head *list) * @vsi: pointer to VSI struct * @list: list of filters */ -static int -ice_fltr_add_eth_list(struct ice_vsi *vsi, struct list_head *list) +static int ice_fltr_add_eth_list(struct ice_vsi *vsi, struct list_head *list) { return ice_add_eth_mac(&vsi->back->hw, list); } @@ -164,8 +160,7 @@ ice_fltr_add_eth_list(struct ice_vsi *vsi, struct list_head *list) * @vsi: pointer to VSI struct * @list: list of filters */ -static int -ice_fltr_remove_eth_list(struct ice_vsi *vsi, struct list_head *list) +static int ice_fltr_remove_eth_list(struct ice_vsi *vsi, struct list_head *list) { return ice_remove_eth_mac(&vsi->back->hw, list); } @@ -270,8 +265,8 @@ ice_fltr_prepare_mac(struct ice_vsi *vsi, const u8 *mac, enum ice_sw_fwd_act_type action, int (*mac_action)(struct ice_vsi *, struct list_head *)) { - int result; LIST_HEAD(tmp_list); + int result; if (ice_fltr_add_mac_to_list(vsi, &tmp_list, mac, action)) { ice_fltr_free_list(ice_pf_to_dev(vsi->back), &tmp_list); @@ -297,8 +292,8 @@ ice_fltr_prepare_mac_and_broadcast(struct ice_vsi *vsi, const u8 *mac, (struct ice_vsi *, struct list_head *)) { u8 broadcast[ETH_ALEN]; - int result; LIST_HEAD(tmp_list); + int result; eth_broadcast_addr(broadcast); if (ice_fltr_add_mac_to_list(vsi, &tmp_list, mac, action) || @@ -324,8 +319,8 @@ ice_fltr_prepare_vlan(struct ice_vsi *vsi, u16 vlan_id, enum ice_sw_fwd_act_type action, int (*vlan_action)(struct ice_vsi *, struct list_head *)) { - int result; LIST_HEAD(tmp_list); + int result; if (ice_fltr_add_vlan_to_list(vsi, &tmp_list, vlan_id, action)) return -ENOMEM; @@ -348,8 +343,8 @@ ice_fltr_prepare_eth(struct ice_vsi *vsi, u16 ethertype, u16 flag, enum ice_sw_fwd_act_type action, int (*eth_action)(struct ice_vsi *, struct list_head *)) { - int result; LIST_HEAD(tmp_list); + int result; if (ice_fltr_add_eth_to_list(vsi, &tmp_list, ethertype, flag, action)) return -ENOMEM; @@ -466,8 +461,8 @@ ice_fltr_update_rule_flags(struct ice_hw *hw, u16 rule_id, u16 recipe_id, u32 act, u16 type, u16 src, u32 new_flags) { struct ice_aqc_sw_rules_elem *s_rule; - int err; u32 flags_mask; + int err; s_rule = kzalloc(ICE_SW_RULE_RX_TX_NO_HDR_SIZE, GFP_KERNEL); if (!s_rule) diff --git a/drivers/net/ethernet/intel/ice/ice_fltr.h b/drivers/net/ethernet/intel/ice/ice_fltr.h index bb04f87376da..0e65fd021ad2 100644 --- a/drivers/net/ethernet/intel/ice/ice_fltr.h +++ b/drivers/net/ethernet/intel/ice/ice_fltr.h @@ -26,13 +26,11 @@ ice_fltr_add_mac(struct ice_vsi *vsi, const u8 *mac, int ice_fltr_add_mac_and_broadcast(struct ice_vsi *vsi, const u8 *mac, enum ice_sw_fwd_act_type action); -int -ice_fltr_add_mac_list(struct ice_vsi *vsi, struct list_head *list); +int ice_fltr_add_mac_list(struct ice_vsi *vsi, struct list_head *list); int ice_fltr_remove_mac(struct ice_vsi *vsi, const u8 *mac, enum ice_sw_fwd_act_type action); -int -ice_fltr_remove_mac_list(struct ice_vsi *vsi, struct list_head *list); +int ice_fltr_remove_mac_list(struct ice_vsi *vsi, struct list_head *list); int ice_fltr_add_vlan(struct ice_vsi *vsi, u16 vid, diff --git a/drivers/net/ethernet/intel/ice/ice_fw_update.c b/drivers/net/ethernet/intel/ice/ice_fw_update.c index c6ed12cf36eb..d12f292a576a 100644 --- a/drivers/net/ethernet/intel/ice/ice_fw_update.c +++ b/drivers/net/ethernet/intel/ice/ice_fw_update.c @@ -40,8 +40,8 @@ ice_send_package_data(struct pldmfw *context, const u8 *data, u16 length) struct device *dev = context->dev; struct ice_pf *pf = priv->pf; struct ice_hw *hw = &pf->hw; - int status; u8 *package_data; + int status; dev_dbg(dev, "Sending PLDM record package data to firmware\n"); @@ -55,8 +55,7 @@ ice_send_package_data(struct pldmfw *context, const u8 *data, u16 length) if (status) { dev_err(dev, "Failed to send record package data to firmware, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); NL_SET_ERR_MSG_MOD(extack, "Failed to record package data to firmware"); return -EIO; } @@ -203,8 +202,8 @@ ice_send_component_table(struct pldmfw *context, struct pldmfw_component *compon struct device *dev = context->dev; struct ice_pf *pf = priv->pf; struct ice_hw *hw = &pf->hw; - int status; size_t length; + int status; switch (component->identifier) { case NVM_COMP_ID_OROM: @@ -241,8 +240,7 @@ ice_send_component_table(struct pldmfw *context, struct pldmfw_component *compon if (status) { dev_err(dev, "Failed to transfer component table to firmware, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); NL_SET_ERR_MSG_MOD(extack, "Failed to transfer component table to firmware"); return -EIO; } @@ -277,8 +275,8 @@ ice_write_one_nvm_block(struct ice_pf *pf, u16 module, u32 offset, struct device *dev = ice_pf_to_dev(pf); struct ice_rq_event_info event; struct ice_hw *hw = &pf->hw; - int status; u32 completion_offset; + int status; int err; memset(&event, 0, sizeof(event)); @@ -524,8 +522,8 @@ static int ice_switch_flash_banks(struct ice_pf *pf, u8 activate_flags, struct device *dev = ice_pf_to_dev(pf); struct ice_rq_event_info event; struct ice_hw *hw = &pf->hw; - int status; u16 completion_retval; + int status; int err; memset(&event, 0, sizeof(event)); @@ -533,8 +531,7 @@ static int ice_switch_flash_banks(struct ice_pf *pf, u8 activate_flags, status = ice_nvm_write_activate(hw, activate_flags); if (status) { dev_err(dev, "Failed to switch active flash banks, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); NL_SET_ERR_MSG_MOD(extack, "Failed to switch active flash banks"); return -EIO; } @@ -695,8 +692,7 @@ int ice_flash_pldm_image(struct ice_pf *pf, const struct firmware *fw, status = ice_acquire_nvm(hw, ICE_RES_WRITE); if (status) { dev_err(dev, "Failed to acquire device flash lock, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); NL_SET_ERR_MSG_MOD(extack, "Failed to acquire device flash lock"); return -EIO; } @@ -736,8 +732,8 @@ int ice_check_for_pending_update(struct ice_pf *pf, const char *component, struct device *dev = ice_pf_to_dev(pf); struct ice_hw_dev_caps *dev_caps; struct ice_hw *hw = &pf->hw; - int status; u8 pending = 0; + int status; int err; dev_caps = kzalloc(sizeof(*dev_caps), GFP_KERNEL); diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c index 758af9bc225a..1902aa250eac 100644 --- a/drivers/net/ethernet/intel/ice/ice_lib.c +++ b/drivers/net/ethernet/intel/ice/ice_lib.c @@ -1545,8 +1545,8 @@ int ice_vsi_cfg_rss_lut_key(struct ice_vsi *vsi) static void ice_vsi_set_vf_rss_flow_fld(struct ice_vsi *vsi) { struct ice_pf *pf = vsi->back; - int status; struct device *dev; + int status; dev = ice_pf_to_dev(pf); if (ice_is_safe_mode(pf)) { @@ -1577,8 +1577,8 @@ static void ice_vsi_set_rss_flow_fld(struct ice_vsi *vsi) u16 vsi_handle = vsi->idx, vsi_num = vsi->vsi_num; struct ice_pf *pf = vsi->back; struct ice_hw *hw = &pf->hw; - int status; struct device *dev; + int status; dev = ice_pf_to_dev(pf); if (ice_is_safe_mode(pf)) { @@ -1749,9 +1749,9 @@ ice_vsi_add_vlan(struct ice_vsi *vsi, u16 vid, enum ice_sw_fwd_act_type action) int ice_vsi_kill_vlan(struct ice_vsi *vsi, u16 vid) { struct ice_pf *pf = vsi->back; - int status; struct device *dev; int err = 0; + int status; dev = ice_pf_to_dev(pf); @@ -2105,8 +2105,8 @@ int ice_vsi_manage_vlan_insertion(struct ice_vsi *vsi) { struct ice_hw *hw = &vsi->back->hw; struct ice_vsi_ctx *ctxt; - int status; int ret = 0; + int status; ctxt = kzalloc(sizeof(*ctxt), GFP_KERNEL); if (!ctxt) @@ -2326,8 +2326,7 @@ int ice_cfg_vlan_pruning(struct ice_vsi *vsi, bool ena) if (status) { netdev_err(vsi->netdev, "%sabling VLAN pruning on VSI handle: %d, VSI HW ID: %d failed, err = %d, aq_err = %s\n", ena ? "En" : "Dis", vsi->idx, vsi->vsi_num, - status, - ice_aq_str(pf->hw.adminq.sq_last_status)); + status, ice_aq_str(pf->hw.adminq.sq_last_status)); goto err_out; } @@ -2408,8 +2407,8 @@ void ice_cfg_sw_lldp(struct ice_vsi *vsi, bool tx, bool create) int (*eth_fltr)(struct ice_vsi *v, u16 type, u16 flag, enum ice_sw_fwd_act_type act); struct ice_pf *pf = vsi->back; - int status; struct device *dev; + int status; dev = ice_pf_to_dev(pf); eth_fltr = create ? ice_fltr_add_eth : ice_fltr_remove_eth; @@ -2576,9 +2575,9 @@ ice_vsi_setup(struct ice_pf *pf, struct ice_port_info *pi, { u16 max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 }; struct device *dev = ice_pf_to_dev(pf); - int status; struct ice_vsi *vsi; int ret, i; + int status; if (vsi_type == ICE_VSI_CHNL) vsi = ice_vsi_alloc(pf, vsi_type, ch, ICE_INVAL_VFID); @@ -3036,8 +3035,8 @@ void ice_napi_del(struct ice_vsi *vsi) */ int ice_vsi_release(struct ice_vsi *vsi) { - int err; struct ice_pf *pf; + int err; if (!vsi->back) return -ENODEV; @@ -3273,9 +3272,9 @@ int ice_vsi_rebuild(struct ice_vsi *vsi, bool init_vsi) int prev_num_q_vectors = 0; struct ice_vf *vf = NULL; enum ice_vsi_type vtype; - int status; struct ice_pf *pf; int ret, i; + int status; if (!vsi) return -EINVAL; @@ -3663,10 +3662,10 @@ 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_vsi_ctx *ctx; - int status; struct device *dev; int i, ret = 0; u8 num_tc = 0; + int status; dev = ice_pf_to_dev(pf); if (vsi->tc_cfg.ena_tc == ena_tc && @@ -3816,8 +3815,8 @@ bool ice_is_vsi_dflt_vsi(struct ice_sw *sw, struct ice_vsi *vsi) */ int ice_set_dflt_vsi(struct ice_sw *sw, struct ice_vsi *vsi) { - int status; struct device *dev; + int status; if (!sw || !vsi) return -EINVAL; @@ -3862,8 +3861,8 @@ int ice_set_dflt_vsi(struct ice_sw *sw, struct ice_vsi *vsi) int ice_clear_dflt_vsi(struct ice_sw *sw) { struct ice_vsi *dflt_vsi; - int status; struct device *dev; + int status; if (!sw) return -EINVAL; @@ -3954,8 +3953,8 @@ int ice_get_link_speed_kbps(struct ice_vsi *vsi) int ice_set_min_bw_limit(struct ice_vsi *vsi, u64 min_tx_rate) { struct ice_pf *pf = vsi->back; - int status; struct device *dev; + int status; int speed; dev = ice_pf_to_dev(pf); @@ -4015,8 +4014,8 @@ int ice_set_min_bw_limit(struct ice_vsi *vsi, u64 min_tx_rate) int ice_set_max_bw_limit(struct ice_vsi *vsi, u64 max_tx_rate) { struct ice_pf *pf = vsi->back; - int status; struct device *dev; + int status; int speed; dev = ice_pf_to_dev(pf); diff --git a/drivers/net/ethernet/intel/ice/ice_lib.h b/drivers/net/ethernet/intel/ice/ice_lib.h index 7865a743efee..b2ed189527d6 100644 --- a/drivers/net/ethernet/intel/ice/ice_lib.h +++ b/drivers/net/ethernet/intel/ice/ice_lib.h @@ -107,8 +107,7 @@ void ice_write_intrl(struct ice_q_vector *q_vector, u8 intrl); void ice_write_itr(struct ice_ring_container *rc, u16 itr); void ice_set_q_vector_intrl(struct ice_q_vector *q_vector); -int -ice_vsi_cfg_mac_fltr(struct ice_vsi *vsi, const u8 *macaddr, bool set); +int ice_vsi_cfg_mac_fltr(struct ice_vsi *vsi, const u8 *macaddr, bool set); bool ice_is_safe_mode(struct ice_pf *pf); bool ice_is_aux_ena(struct ice_pf *pf); diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 154a90989897..c4676f289ad8 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -155,9 +155,9 @@ static void ice_check_for_hang_subtask(struct ice_pf *pf) */ static int ice_init_mac_fltr(struct ice_pf *pf) { - int status; struct ice_vsi *vsi; u8 *perm_addr; + int status; vsi = ice_get_main_vsi(pf); if (!vsi) @@ -295,8 +295,8 @@ static int ice_vsi_sync_fltr(struct ice_vsi *vsi) bool promisc_forced_on = false; struct ice_pf *pf = vsi->back; struct ice_hw *hw = &pf->hw; - int status = 0; u32 changed_flags = 0; + int status = 0; u8 promisc_m; int err = 0; @@ -708,12 +708,12 @@ void ice_print_link_msg(struct ice_vsi *vsi, bool isup) { struct ice_aqc_get_phy_caps_data *caps; const char *an_advertised; - int status; const char *fec_req; const char *speed; const char *fec; const char *fc; const char *an; + int status; if (!vsi) return; @@ -1033,10 +1033,10 @@ ice_link_event(struct ice_pf *pf, struct ice_port_info *pi, bool link_up, { struct device *dev = ice_pf_to_dev(pf); struct ice_phy_info *phy_info; - int status; struct ice_vsi *vsi; u16 old_link_speed; bool old_link; + int status; phy_info = &pi->phy; phy_info->link_info_old = phy_info->link_info; @@ -1420,8 +1420,8 @@ static int __ice_clean_ctrlq(struct ice_pf *pf, enum ice_ctl_q q_type) return 0; do { - int ret; u16 opcode; + int ret; ret = ice_clean_rq_elem(hw, cq, &event, &pending); if (ret == -EALREADY) @@ -1879,8 +1879,8 @@ static int ice_init_nvm_phy_type(struct ice_port_info *pi) { struct ice_aqc_get_phy_caps_data *pcaps; struct ice_pf *pf = pi->hw->back; - int status; int err = 0; + int status; pcaps = kzalloc(sizeof(*pcaps), GFP_KERNEL); if (!pcaps) @@ -1990,8 +1990,8 @@ static int ice_init_phy_user_cfg(struct ice_port_info *pi) struct ice_aqc_get_phy_caps_data *pcaps; struct ice_phy_info *phy = &pi->phy; struct ice_pf *pf = pi->hw->back; - int status; int err = 0; + int status; if (!(phy->link_info.link_info & ICE_AQ_MEDIA_AVAILABLE)) return -EIO; @@ -2062,8 +2062,8 @@ static int ice_configure_phy(struct ice_vsi *vsi) struct ice_aqc_set_phy_cfg_data *cfg; struct ice_phy_info *phy = &pi->phy; struct ice_pf *pf = vsi->back; - int status; int err = 0; + int status; /* Ensure we have media as we cannot configure a medialess port */ if (!(phy->link_info.link_info & ICE_AQ_MEDIA_AVAILABLE)) @@ -2562,9 +2562,9 @@ int ice_prepare_xdp_rings(struct ice_vsi *vsi, struct bpf_prog *prog) .vsi_map_offset = vsi->alloc_txq, .mapping_mode = ICE_VSI_MAP_CONTIG }; - int status; struct device *dev; int i, v_idx; + int status; dev = ice_pf_to_dev(pf); vsi->xdp_rings = devm_kcalloc(dev, vsi->num_xdp_txq, @@ -4036,8 +4036,8 @@ static void ice_set_safe_mode_vlan_cfg(struct ice_pf *pf) { struct ice_vsi *vsi = ice_get_main_vsi(pf); struct ice_vsi_ctx *ctxt; - int status; struct ice_hw *hw; + int status; if (!vsi) return; @@ -4068,8 +4068,7 @@ static void ice_set_safe_mode_vlan_cfg(struct ice_pf *pf) status = ice_update_vsi(hw, vsi->idx, ctxt, NULL); if (status) { dev_err(ice_pf_to_dev(vsi->back), "Failed to update VSI for safe mode VLANs, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); } else { vsi->info.sec_flags = ctxt->info.sec_flags; vsi->info.sw_flags2 = ctxt->info.sw_flags2; @@ -4792,9 +4791,9 @@ static void ice_setup_mc_magic_wake(struct ice_pf *pf) { struct device *dev = ice_pf_to_dev(pf); struct ice_hw *hw = &pf->hw; - int status; u8 mac_addr[ETH_ALEN]; struct ice_vsi *vsi; + int status; u8 flags; if (!pf->wol_ena) @@ -4817,8 +4816,7 @@ static void ice_setup_mc_magic_wake(struct ice_pf *pf) status = ice_aq_manage_mac_write(hw, mac_addr, flags, NULL); if (status) dev_err(dev, "Failed to enable Multicast Magic Packet wake, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); } /** @@ -5357,10 +5355,10 @@ static int ice_set_mac_address(struct net_device *netdev, void *pi) struct ice_pf *pf = vsi->back; struct ice_hw *hw = &pf->hw; struct sockaddr *addr = pi; - int status; u8 old_mac[ETH_ALEN]; u8 flags = 0; int err = 0; + int status; u8 *mac; mac = (u8 *)addr->sa_data; @@ -5472,8 +5470,8 @@ ice_set_tx_maxrate(struct net_device *netdev, int queue_index, u32 maxrate) { struct ice_netdev_priv *np = netdev_priv(netdev); struct ice_vsi *vsi = np->vsi; - int status; u16 q_handle; + int status; u8 tc; /* Validate maxrate requested is within permitted range */ @@ -6554,8 +6552,7 @@ static int ice_vsi_rebuild_by_type(struct ice_pf *pf, enum ice_vsi_type type) status = ice_replay_vsi(&pf->hw, vsi->idx); if (status) { dev_err(dev, "replay VSI failed, error %d, VSI index %d, type %s\n", - status, vsi->idx, - ice_vsi_type_str(type)); + status, vsi->idx, ice_vsi_type_str(type)); return -EIO; } @@ -6629,8 +6626,7 @@ static void ice_rebuild(struct ice_pf *pf, enum ice_reset_req reset_type) ret = ice_init_all_ctrlq(hw); if (ret) { - dev_err(dev, "control queues init failed %d\n", - ret); + dev_err(dev, "control queues init failed %d\n", ret); goto err_init_ctrlq; } @@ -6646,8 +6642,7 @@ static void ice_rebuild(struct ice_pf *pf, enum ice_reset_req reset_type) ret = ice_clear_pf_cfg(hw); if (ret) { - dev_err(dev, "clear PF configuration failed %d\n", - ret); + dev_err(dev, "clear PF configuration failed %d\n", ret); goto err_init_ctrlq; } @@ -6964,8 +6959,7 @@ int ice_set_rss_lut(struct ice_vsi *vsi, u8 *lut, u16 lut_size) status = ice_aq_set_rss_lut(hw, ¶ms); if (status) { dev_err(ice_pf_to_dev(vsi->back), "Cannot set RSS lut, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); return -EIO; } @@ -6990,8 +6984,7 @@ int ice_set_rss_key(struct ice_vsi *vsi, u8 *seed) status = ice_aq_set_rss_key(hw, vsi->idx, (struct ice_aqc_get_set_rss_keys *)seed); if (status) { dev_err(ice_pf_to_dev(vsi->back), "Cannot set RSS key, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); return -EIO; } @@ -7023,8 +7016,7 @@ int ice_get_rss_lut(struct ice_vsi *vsi, u8 *lut, u16 lut_size) status = ice_aq_get_rss_lut(hw, ¶ms); if (status) { dev_err(ice_pf_to_dev(vsi->back), "Cannot get RSS lut, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); return -EIO; } @@ -7049,8 +7041,7 @@ int ice_get_rss_key(struct ice_vsi *vsi, u8 *seed) status = ice_aq_get_rss_key(hw, vsi->idx, (struct ice_aqc_get_set_rss_keys *)seed); if (status) { dev_err(ice_pf_to_dev(vsi->back), "Cannot get RSS key, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); return -EIO; } @@ -7095,8 +7086,8 @@ static int ice_vsi_update_bridge_mode(struct ice_vsi *vsi, u16 bmode) struct ice_aqc_vsi_props *vsi_props; struct ice_hw *hw = &vsi->back->hw; struct ice_vsi_ctx *ctxt; - int status; int ret = 0; + int status; vsi_props = &vsi->info; @@ -7117,8 +7108,7 @@ static int ice_vsi_update_bridge_mode(struct ice_vsi *vsi, u16 bmode) status = ice_update_vsi(hw, vsi->idx, ctxt, NULL); if (status) { dev_err(ice_pf_to_dev(vsi->back), "update VSI for bridge mode failed, bmode = %d err %d aq_err %s\n", - bmode, status, - ice_aq_str(hw->adminq.sq_last_status)); + bmode, status, ice_aq_str(hw->adminq.sq_last_status)); ret = -EIO; goto out; } @@ -7151,9 +7141,9 @@ ice_bridge_setlink(struct net_device *dev, struct nlmsghdr *nlh, struct ice_pf *pf = np->vsi->back; struct nlattr *attr, *br_spec; struct ice_hw *hw = &pf->hw; - int status; struct ice_sw *pf_sw; int rem, v, err = 0; + int status; pf_sw = pf->first_sw; /* find the attribute in the netlink message */ diff --git a/drivers/net/ethernet/intel/ice/ice_nvm.c b/drivers/net/ethernet/intel/ice/ice_nvm.c index 40c9bcb7889a..2d95fe533d1e 100644 --- a/drivers/net/ethernet/intel/ice/ice_nvm.c +++ b/drivers/net/ethernet/intel/ice/ice_nvm.c @@ -64,10 +64,10 @@ int ice_read_flat_nvm(struct ice_hw *hw, u32 offset, u32 *length, u8 *data, bool read_shadow_ram) { - int status; u32 inlen = *length; u32 bytes_read = 0; bool last_cmd; + int status; *length = 0; @@ -158,8 +158,7 @@ ice_aq_update_nvm(struct ice_hw *hw, u16 module_typeid, u32 offset, * * Erase the NVM sector using the admin queue commands (0x0702) */ -int -ice_aq_erase_nvm(struct ice_hw *hw, u16 module_typeid, struct ice_sq_cd *cd) +int ice_aq_erase_nvm(struct ice_hw *hw, u16 module_typeid, struct ice_sq_cd *cd) { struct ice_aq_desc desc; struct ice_aqc_nvm *cmd; @@ -184,12 +183,11 @@ ice_aq_erase_nvm(struct ice_hw *hw, u16 module_typeid, struct ice_sq_cd *cd) * * Reads one 16 bit word from the Shadow RAM using ice_read_flat_nvm. */ -static int -ice_read_sr_word_aq(struct ice_hw *hw, u16 offset, u16 *data) +static int ice_read_sr_word_aq(struct ice_hw *hw, u16 offset, u16 *data) { u32 bytes = sizeof(u16); - int status; __le16 data_local; + int status; /* Note that ice_read_flat_nvm takes into account the 4Kb AdminQ and * Shadow RAM sector restrictions necessary when reading from the NVM. @@ -210,8 +208,7 @@ ice_read_sr_word_aq(struct ice_hw *hw, u16 offset, u16 *data) * * This function will request NVM ownership. */ -int -ice_acquire_nvm(struct ice_hw *hw, enum ice_aq_res_access_type access) +int ice_acquire_nvm(struct ice_hw *hw, enum ice_aq_res_access_type access) { if (hw->flash.blank_nvm_mode) return 0; @@ -356,8 +353,8 @@ ice_read_flash_module(struct ice_hw *hw, enum ice_bank_select bank, u16 module, static int ice_read_nvm_module(struct ice_hw *hw, enum ice_bank_select bank, u32 offset, u16 *data) { - int status; __le16 data_local; + int status; status = ice_read_flash_module(hw, bank, ICE_SR_1ST_NVM_BANK_PTR, offset * sizeof(u16), (__force u8 *)&data_local, sizeof(u16)); @@ -395,8 +392,8 @@ ice_read_nvm_sr_copy(struct ice_hw *hw, enum ice_bank_select bank, u32 offset, u static int ice_read_netlist_module(struct ice_hw *hw, enum ice_bank_select bank, u32 offset, u16 *data) { - int status; __le16 data_local; + int status; status = ice_read_flash_module(hw, bank, ICE_SR_NETLIST_BANK_PTR, offset * sizeof(u16), (__force u8 *)&data_local, sizeof(u16)); @@ -442,9 +439,9 @@ int ice_get_pfa_module_tlv(struct ice_hw *hw, u16 *module_tlv, u16 *module_tlv_len, u16 module_type) { - int status; u16 pfa_len, pfa_ptr; u16 next_tlv; + int status; status = ice_read_sr_word(hw, ICE_SR_PFA_PTR, &pfa_ptr); if (status) { @@ -501,12 +498,11 @@ ice_get_pfa_module_tlv(struct ice_hw *hw, u16 *module_tlv, u16 *module_tlv_len, * * Reads the part number string from the NVM. */ -int -ice_read_pba_string(struct ice_hw *hw, u8 *pba_num, u32 pba_num_size) +int ice_read_pba_string(struct ice_hw *hw, u8 *pba_num, u32 pba_num_size) { u16 pba_tlv, pba_tlv_len; - int status; u16 pba_word, pba_size; + int status; u16 i; status = ice_get_pfa_module_tlv(hw, &pba_tlv, &pba_tlv_len, @@ -673,8 +669,8 @@ static int ice_get_orom_ver_info(struct ice_hw *hw, enum ice_bank_select bank, struct ice_orom_info *orom) { struct ice_orom_civd_info civd; - int status; u32 combo_ver; + int status; status = ice_get_orom_civd_data(hw, bank, &civd); if (status) { @@ -720,8 +716,8 @@ ice_get_netlist_info(struct ice_hw *hw, enum ice_bank_select bank, struct ice_netlist_info *netlist) { u16 module_id, length, node_count, i; - int status; u16 *id_blk; + int status; status = ice_read_netlist_module(hw, bank, ICE_NETLIST_TYPE_OFFSET, &module_id); if (status) @@ -859,8 +855,7 @@ static int ice_discover_flash_size(struct ice_hw *hw) * sector size by using the highest bit. The reported pointer value will be in * bytes, intended for flat NVM reads. */ -static int -ice_read_sr_pointer(struct ice_hw *hw, u16 offset, u32 *pointer) +static int ice_read_sr_pointer(struct ice_hw *hw, u16 offset, u32 *pointer) { int status; u16 value; @@ -891,8 +886,7 @@ ice_read_sr_pointer(struct ice_hw *hw, u16 offset, u32 *pointer) * Each area size word is specified in 4KB sector units. This function reports * the size in bytes, intended for flat NVM reads. */ -static int -ice_read_sr_area_size(struct ice_hw *hw, u16 offset, u32 *size) +static int ice_read_sr_area_size(struct ice_hw *hw, u16 offset, u32 *size) { int status; u16 value; @@ -917,12 +911,11 @@ ice_read_sr_area_size(struct ice_hw *hw, u16 offset, u32 *size) * structure for later use in order to calculate the correct offset to read * from the active module. */ -static int -ice_determine_active_flash_banks(struct ice_hw *hw) +static int ice_determine_active_flash_banks(struct ice_hw *hw) { struct ice_bank_info *banks = &hw->flash.banks; - int status; u16 ctrl_word; + int status; status = ice_read_sr_word(hw, ICE_SR_NVM_CTRL_WORD, &ctrl_word); if (status) { @@ -1000,9 +993,9 @@ ice_determine_active_flash_banks(struct ice_hw *hw) int ice_init_nvm(struct ice_hw *hw) { struct ice_flash_info *flash = &hw->flash; - int status; u32 fla, gens_stat; u8 sr_size; + int status; /* The SR size is stored regardless of the NVM programming mode * as the blank mode may be used in the factory line. diff --git a/drivers/net/ethernet/intel/ice/ice_nvm.h b/drivers/net/ethernet/intel/ice/ice_nvm.h index ed5165d24cb1..f1d754ffac16 100644 --- a/drivers/net/ethernet/intel/ice/ice_nvm.h +++ b/drivers/net/ethernet/intel/ice/ice_nvm.h @@ -12,8 +12,7 @@ struct ice_orom_civd_info { __le16 combo_name[32]; /* Unicode string representing the Combo Image version */ } __packed; -int -ice_acquire_nvm(struct ice_hw *hw, enum ice_aq_res_access_type access); +int ice_acquire_nvm(struct ice_hw *hw, enum ice_aq_res_access_type access); void ice_release_nvm(struct ice_hw *hw); int ice_read_flat_nvm(struct ice_hw *hw, u32 offset, u32 *length, u8 *data, @@ -21,14 +20,11 @@ ice_read_flat_nvm(struct ice_hw *hw, u32 offset, u32 *length, u8 *data, int ice_get_pfa_module_tlv(struct ice_hw *hw, u16 *module_tlv, u16 *module_tlv_len, u16 module_type); -int -ice_get_inactive_orom_ver(struct ice_hw *hw, struct ice_orom_info *orom); -int -ice_get_inactive_nvm_ver(struct ice_hw *hw, struct ice_nvm_info *nvm); +int ice_get_inactive_orom_ver(struct ice_hw *hw, struct ice_orom_info *orom); +int ice_get_inactive_nvm_ver(struct ice_hw *hw, struct ice_nvm_info *nvm); int ice_get_inactive_netlist_ver(struct ice_hw *hw, struct ice_netlist_info *netlist); -int -ice_read_pba_string(struct ice_hw *hw, u8 *pba_num, u32 pba_num_size); +int ice_read_pba_string(struct ice_hw *hw, u8 *pba_num, u32 pba_num_size); int ice_init_nvm(struct ice_hw *hw); int ice_read_sr_word(struct ice_hw *hw, u16 offset, u16 *data); int diff --git a/drivers/net/ethernet/intel/ice/ice_sched.c b/drivers/net/ethernet/intel/ice/ice_sched.c index e80f35cf7eab..b4bf18f9d4e3 100644 --- a/drivers/net/ethernet/intel/ice/ice_sched.c +++ b/drivers/net/ethernet/intel/ice/ice_sched.c @@ -152,8 +152,8 @@ ice_sched_add_node(struct ice_port_info *pi, u8 layer, struct ice_aqc_txsched_elem_data elem; struct ice_sched_node *parent; struct ice_sched_node *node; - int status; struct ice_hw *hw; + int status; if (!pi) return -EINVAL; @@ -234,8 +234,8 @@ ice_sched_remove_elems(struct ice_hw *hw, struct ice_sched_node *parent, { struct ice_aqc_delete_elem *buf; u16 i, num_groups_removed = 0; - int status; u16 buf_size; + int status; buf_size = struct_size(buf, teid, num_nodes); buf = devm_kzalloc(ice_hw_to_dev(hw), buf_size, GFP_KERNEL); @@ -525,8 +525,8 @@ ice_sched_suspend_resume_elems(struct ice_hw *hw, u8 num_nodes, u32 *node_teids, bool suspend) { u16 i, buf_size, num_elem_ret = 0; - int status; __le32 *buf; + int status; buf_size = sizeof(*buf) * num_nodes; buf = devm_kzalloc(ice_hw_to_dev(hw), buf_size, GFP_KERNEL); @@ -727,8 +727,8 @@ ice_sched_del_rl_profile(struct ice_hw *hw, { struct ice_aqc_rl_profile_elem *buf; u16 num_profiles_removed; - int status; u16 num_profiles = 1; + int status; if (rl_info->prof_id_ref != 0) return -EBUSY; @@ -883,9 +883,9 @@ ice_sched_add_elems(struct ice_port_info *pi, struct ice_sched_node *tc_node, struct ice_sched_node *prev, *new_node; struct ice_aqc_add_elem *buf; u16 i, num_groups_added = 0; - int status = 0; struct ice_hw *hw = pi->hw; size_t buf_size; + int status = 0; u32 teid; buf_size = struct_size(buf, generic, num_nodes); @@ -1201,10 +1201,10 @@ static void ice_sched_rm_dflt_nodes(struct ice_port_info *pi) int ice_sched_init_port(struct ice_port_info *pi) { struct ice_aqc_get_topo_elem *buf; - int status; struct ice_hw *hw; u8 num_branches; u16 num_elems; + int status; u8 i, j; if (!pi) @@ -1303,8 +1303,8 @@ int ice_sched_init_port(struct ice_port_info *pi) int ice_sched_query_res_alloc(struct ice_hw *hw) { struct ice_aqc_query_txsched_res_resp *buf; - int status = 0; __le16 max_sibl; + int status = 0; u16 i; if (hw->layer_info) @@ -1621,10 +1621,10 @@ ice_sched_add_vsi_child_nodes(struct ice_port_info *pi, u16 vsi_handle, { struct ice_sched_node *parent, *node; struct ice_hw *hw = pi->hw; - int status; u32 first_node_teid; u16 num_added = 0; u8 i, qgl, vsil; + int status; qgl = ice_sched_get_qgrp_layer(hw); vsil = ice_sched_get_vsi_layer(hw); @@ -1722,10 +1722,10 @@ ice_sched_add_vsi_support_nodes(struct ice_port_info *pi, u16 vsi_handle, struct ice_sched_node *tc_node, u16 *num_nodes) { struct ice_sched_node *parent = tc_node; - int status; u32 first_node_teid; u16 num_added = 0; u8 i, vsil; + int status; if (!pi) return -EINVAL; @@ -1802,9 +1802,9 @@ ice_sched_update_vsi_child_nodes(struct ice_port_info *pi, u16 vsi_handle, struct ice_sched_node *vsi_node; struct ice_sched_node *tc_node; struct ice_vsi_ctx *vsi_ctx; - int status = 0; struct ice_hw *hw = pi->hw; u16 prev_numqs; + int status = 0; tc_node = ice_sched_get_tc_node(pi, tc); if (!tc_node) @@ -1875,8 +1875,8 @@ ice_sched_cfg_vsi(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u16 maxqs, { struct ice_sched_node *vsi_node, *tc_node; struct ice_vsi_ctx *vsi_ctx; - int status = 0; struct ice_hw *hw = pi->hw; + int status = 0; ice_debug(pi->hw, ICE_DBG_SCHED, "add/config VSI %d\n", vsi_handle); tc_node = ice_sched_get_tc_node(pi, tc); @@ -1996,8 +1996,8 @@ static bool ice_sched_is_leaf_node_present(struct ice_sched_node *node) static int ice_sched_rm_vsi_cfg(struct ice_port_info *pi, u16 vsi_handle, u8 owner) { - int status = -EINVAL; struct ice_vsi_ctx *vsi_ctx; + int status = -EINVAL; u8 i; ice_debug(pi->hw, ICE_DBG_SCHED, "removing VSI %d\n", vsi_handle); @@ -2194,9 +2194,9 @@ ice_sched_move_nodes(struct ice_port_info *pi, struct ice_sched_node *parent, { struct ice_aqc_move_elem *buf; struct ice_sched_node *node; - int status = 0; u16 i, grps_movd = 0; struct ice_hw *hw; + int status = 0; u16 buf_len; hw = pi->hw; @@ -2258,9 +2258,9 @@ ice_sched_move_vsi_to_agg(struct ice_port_info *pi, u16 vsi_handle, u32 agg_id, struct ice_sched_node *vsi_node, *agg_node, *tc_node, *parent; u16 num_nodes[ICE_AQC_TOPO_MAX_LEVEL_NUM] = { 0 }; u32 first_node_teid, vsi_teid; - int status; u16 num_nodes_added; u8 aggl, vsil, i; + int status; tc_node = ice_sched_get_tc_node(pi, tc); if (!tc_node) @@ -2506,10 +2506,10 @@ ice_sched_add_agg_cfg(struct ice_port_info *pi, u32 agg_id, u8 tc) { struct ice_sched_node *parent, *agg_node, *tc_node; u16 num_nodes[ICE_AQC_TOPO_MAX_LEVEL_NUM] = { 0 }; - int status = 0; struct ice_hw *hw = pi->hw; u32 first_node_teid; u16 num_nodes_added; + int status = 0; u8 i, aggl; tc_node = ice_sched_get_tc_node(pi, tc); @@ -2596,8 +2596,8 @@ ice_sched_cfg_agg(struct ice_port_info *pi, u32 agg_id, enum ice_agg_type agg_type, unsigned long *tc_bitmap) { struct ice_sched_agg_info *agg_info; - int status = 0; struct ice_hw *hw = pi->hw; + int status = 0; u8 tc; agg_info = ice_get_agg_info(hw, agg_id); @@ -2760,8 +2760,8 @@ ice_sched_assoc_vsi_to_agg(struct ice_port_info *pi, u32 agg_id, { struct ice_sched_agg_vsi_info *agg_vsi_info, *old_agg_vsi_info = NULL; struct ice_sched_agg_info *agg_info, *old_agg_info; - int status = 0; struct ice_hw *hw = pi->hw; + int status = 0; u8 tc; if (!ice_is_vsi_valid(pi->hw, vsi_handle)) @@ -2856,9 +2856,9 @@ ice_sched_update_elem(struct ice_hw *hw, struct ice_sched_node *node, struct ice_aqc_txsched_elem_data *info) { struct ice_aqc_txsched_elem_data buf; - int status; u16 elem_cfgd = 0; u16 num_elems = 1; + int status; buf = *info; /* Parent TEID is reserved field in this aq call */ @@ -3102,8 +3102,8 @@ static int ice_sched_bw_to_rl_profile(struct ice_hw *hw, u32 bw, struct ice_aqc_rl_profile_elem *profile) { - int status = -EINVAL; s64 bytes_per_sec, ts_rate, mv_tmp; + int status = -EINVAL; bool found = false; s32 encode = 0; s64 mv = 0; @@ -3176,9 +3176,9 @@ ice_sched_add_rl_profile(struct ice_port_info *pi, struct ice_aqc_rl_profile_info *rl_prof_elem; u16 profiles_added = 0, num_profiles = 1; struct ice_aqc_rl_profile_elem *buf; - int status; struct ice_hw *hw; u8 profile_type; + int status; if (layer_num >= ICE_AQC_TOPO_MAX_LEVEL_NUM) return NULL; @@ -3466,11 +3466,11 @@ ice_sched_set_node_bw_dflt(struct ice_port_info *pi, struct ice_sched_node *node, enum ice_rl_type rl_type, u8 layer_num) { - int status; struct ice_hw *hw; u8 profile_type; u16 rl_prof_id; u16 old_id; + int status; hw = pi->hw; switch (rl_type) { @@ -3567,9 +3567,9 @@ ice_sched_set_node_bw(struct ice_port_info *pi, struct ice_sched_node *node, enum ice_rl_type rl_type, u32 bw, u8 layer_num) { struct ice_aqc_rl_profile_info *rl_prof_info; - int status = -EINVAL; struct ice_hw *hw = pi->hw; u16 old_id, rl_prof_id; + int status = -EINVAL; rl_prof_info = ice_sched_add_rl_profile(pi, rl_type, bw, layer_num); if (!rl_prof_info) @@ -3735,9 +3735,9 @@ static int ice_sched_set_q_bw_lmt(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u16 q_handle, enum ice_rl_type rl_type, u32 bw) { - int status = -EINVAL; struct ice_sched_node *node; struct ice_q_ctx *q_ctx; + int status = -EINVAL; if (!ice_is_vsi_valid(pi->hw, vsi_handle)) return -EINVAL; @@ -3885,8 +3885,8 @@ ice_sched_set_node_bw_lmt_per_tc(struct ice_port_info *pi, u32 id, enum ice_agg_type agg_type, u8 tc, enum ice_rl_type rl_type, u32 bw) { - int status = -EINVAL; struct ice_sched_node *node; + int status = -EINVAL; if (!pi) return status; @@ -4191,8 +4191,7 @@ void ice_sched_replay_agg_vsi_preinit(struct ice_hw *hw) * their node bandwidth information. This function needs to be called with * scheduler lock held. */ -static int -ice_sched_replay_vsi_agg(struct ice_hw *hw, u16 vsi_handle) +static int ice_sched_replay_vsi_agg(struct ice_hw *hw, u16 vsi_handle) { DECLARE_BITMAP(replay_bitmap, ICE_MAX_TRAFFIC_CLASS); struct ice_sched_agg_vsi_info *agg_vsi_info; @@ -4252,8 +4251,7 @@ int ice_replay_vsi_agg(struct ice_hw *hw, u16 vsi_handle) * This function replays queue type node bandwidth. This function needs to be * called with scheduler lock held. */ -int -ice_sched_replay_q_bw(struct ice_port_info *pi, struct ice_q_ctx *q_ctx) +int ice_sched_replay_q_bw(struct ice_port_info *pi, struct ice_q_ctx *q_ctx) { struct ice_sched_node *q_node; diff --git a/drivers/net/ethernet/intel/ice/ice_sched.h b/drivers/net/ethernet/intel/ice/ice_sched.h index 6f21355ddd44..4f91577fed56 100644 --- a/drivers/net/ethernet/intel/ice/ice_sched.h +++ b/drivers/net/ethernet/intel/ice/ice_sched.h @@ -120,6 +120,5 @@ int ice_cfg_rl_burst_size(struct ice_hw *hw, u32 bytes); void ice_sched_replay_agg_vsi_preinit(struct ice_hw *hw); void ice_sched_replay_agg(struct ice_hw *hw); int ice_replay_vsi_agg(struct ice_hw *hw, u16 vsi_handle); -int -ice_sched_replay_q_bw(struct ice_port_info *pi, struct ice_q_ctx *q_ctx); +int ice_sched_replay_q_bw(struct ice_port_info *pi, struct ice_q_ctx *q_ctx); #endif /* _ICE_SCHED_H_ */ diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c index 12e76717cecd..e477c2b1d5bb 100644 --- a/drivers/net/ethernet/intel/ice/ice_switch.c +++ b/drivers/net/ethernet/intel/ice/ice_switch.c @@ -946,8 +946,8 @@ ice_aq_alloc_free_vsi_list(struct ice_hw *hw, u16 *vsi_list_id, { struct ice_aqc_alloc_free_res_elem *sw_buf; struct ice_aqc_res_elem *vsi_ele; - int status; u16 buf_len; + int status; buf_len = struct_size(sw_buf, elem, 1); sw_buf = devm_kzalloc(ice_hw_to_dev(hw), buf_len, GFP_KERNEL); @@ -1076,8 +1076,8 @@ ice_aq_get_recipe(struct ice_hw *hw, { struct ice_aqc_add_get_recipe *cmd; struct ice_aq_desc desc; - int status; u16 buf_size; + int status; if (*num_recipes != ICE_MAX_NUM_RECIPES) return -EINVAL; @@ -1157,8 +1157,8 @@ ice_aq_get_recipe_to_profile(struct ice_hw *hw, u32 profile_id, u8 *r_bitmap, static int ice_alloc_recipe(struct ice_hw *hw, u16 *rid) { struct ice_aqc_alloc_free_res_elem *sw_buf; - int status; u16 buf_len; + int status; buf_len = struct_size(sw_buf, elem, 1); sw_buf = kzalloc(buf_len, GFP_KERNEL); @@ -1238,9 +1238,9 @@ ice_get_recp_frm_fw(struct ice_hw *hw, struct ice_sw_recipe *recps, u8 rid, struct ice_aqc_recipe_data_elem *tmp; u16 num_recps = ICE_MAX_NUM_RECIPES; struct ice_prot_lkup_ext *lkup_exts; - int status; u8 fv_word_idx = 0; u16 sub_recps; + int status; bitmap_zero(result_bm, ICE_MAX_FV_WORDS); @@ -1410,9 +1410,9 @@ ice_init_port_info(struct ice_port_info *pi, u16 vsi_port_num, u8 type, int ice_get_initial_sw_cfg(struct ice_hw *hw) { struct ice_aqc_get_sw_cfg_resp_elem *rbuf; - int status; u16 req_desc = 0; u16 num_elems; + int status; u16 i; rbuf = devm_kzalloc(ice_hw_to_dev(hw), ICE_SW_CFG_MAX_BUF_LEN, @@ -1681,9 +1681,9 @@ ice_add_marker_act(struct ice_hw *hw, struct ice_fltr_mgmt_list_entry *m_ent, * 3. GENERIC VALUE action to hold the software marker ID */ const u16 num_lg_acts = 3; - int status; u16 lg_act_size; u16 rules_size; + int status; u32 act; u16 id; @@ -1814,9 +1814,9 @@ ice_update_vsi_list_rule(struct ice_hw *hw, u16 *vsi_handle_arr, u16 num_vsi, enum ice_sw_lkup_type lkup_type) { struct ice_aqc_sw_rules_elem *s_rule; - int status; u16 s_rule_size; u16 rule_type; + int status; int i; if (!num_vsi) @@ -1992,9 +1992,9 @@ int ice_update_sw_rule_bridge_mode(struct ice_hw *hw) { struct ice_switch_info *sw = hw->switch_info; struct ice_fltr_mgmt_list_entry *fm_entry; - int status = 0; struct list_head *rule_head; struct mutex *rule_lock; /* Lock to protect filter rule list */ + int status = 0; rule_lock = &sw->recp_list[ICE_SW_LKUP_MAC].filt_rule_lock; rule_head = &sw->recp_list[ICE_SW_LKUP_MAC].filt_rules; @@ -2050,8 +2050,8 @@ ice_add_update_vsi_list(struct ice_hw *hw, struct ice_fltr_info *cur_fltr, struct ice_fltr_info *new_fltr) { - int status = 0; u16 vsi_list_id = 0; + int status = 0; if ((cur_fltr->fltr_act == ICE_FWD_TO_Q || cur_fltr->fltr_act == ICE_FWD_TO_QGRP)) @@ -2260,8 +2260,8 @@ ice_remove_vsi_list_rule(struct ice_hw *hw, u16 vsi_list_id, enum ice_sw_lkup_type lkup_type) { struct ice_aqc_sw_rules_elem *s_rule; - int status; u16 s_rule_size; + int status; s_rule_size = (u16)ICE_SW_RULE_VSI_LIST_SIZE(0); s_rule = devm_kzalloc(ice_hw_to_dev(hw), s_rule_size, GFP_KERNEL); @@ -2293,8 +2293,8 @@ ice_rem_update_vsi_list(struct ice_hw *hw, u16 vsi_handle, struct ice_fltr_mgmt_list_entry *fm_list) { enum ice_sw_lkup_type lkup_type; - int status = 0; u16 vsi_list_id; + int status = 0; if (fm_list->fltr_info.fltr_act != ICE_FWD_TO_VSI_LIST || fm_list->vsi_count == 0) @@ -2382,9 +2382,9 @@ ice_remove_rule_internal(struct ice_hw *hw, u8 recp_id, struct ice_switch_info *sw = hw->switch_info; struct ice_fltr_mgmt_list_entry *list_elem; struct mutex *rule_lock; /* Lock to protect filter rule list */ - int status = 0; bool remove_rule = false; u16 vsi_handle; + int status = 0; if (!ice_is_vsi_valid(hw, f_entry->fltr_info.vsi_handle)) return -EINVAL; @@ -2598,8 +2598,8 @@ int ice_add_mac(struct ice_hw *hw, struct list_head *m_list) u16 total_elem_left, s_rule_size; struct ice_switch_info *sw; struct mutex *rule_lock; /* Lock to protect filter rule list */ - int status = 0; u16 num_unicast = 0; + int status = 0; u8 elem_sent; if (!m_list || !hw) @@ -2917,8 +2917,7 @@ int ice_add_vlan(struct ice_hw *hw, struct list_head *v_list) * the filter list with the necessary fields (including flags to * indicate Tx or Rx rules). */ -int -ice_add_eth_mac(struct ice_hw *hw, struct list_head *em_list) +int ice_add_eth_mac(struct ice_hw *hw, struct list_head *em_list) { struct ice_fltr_list_entry *em_list_itr; @@ -2946,8 +2945,7 @@ ice_add_eth_mac(struct ice_hw *hw, struct list_head *em_list) * @hw: pointer to the hardware structure * @em_list: list of ethertype or ethertype MAC entries */ -int -ice_remove_eth_mac(struct ice_hw *hw, struct list_head *em_list) +int ice_remove_eth_mac(struct ice_hw *hw, struct list_head *em_list) { struct ice_fltr_list_entry *em_list_itr, *tmp; @@ -3020,15 +3018,14 @@ ice_rem_adv_rule_info(struct ice_hw *hw, struct list_head *rule_head) * add filter rule to set/unset given VSI as default VSI for the switch * (represented by swid) */ -int -ice_cfg_dflt_vsi(struct ice_hw *hw, u16 vsi_handle, bool set, u8 direction) +int ice_cfg_dflt_vsi(struct ice_hw *hw, u16 vsi_handle, bool set, u8 direction) { struct ice_aqc_sw_rules_elem *s_rule; struct ice_fltr_info f_info; enum ice_adminq_opc opcode; - int status; u16 s_rule_size; u16 hw_vsi_id; + int status; if (!ice_is_vsi_valid(hw, vsi_handle)) return -EINVAL; @@ -3137,10 +3134,10 @@ ice_find_ucast_rule_entry(struct ice_hw *hw, u8 recp_id, * This function removes either a MAC filter rule or a specific VSI from a * VSI list for a multicast MAC address. * - * Returns -ENOENT if a given entry was not added by - * ice_add_mac. Caller should be aware that this call will only work if all - * the entries passed into m_list were added previously. It will not attempt to - * do a partial remove of entries that were found. + * Returns -ENOENT if a given entry was not added by ice_add_mac. Caller should + * be aware that this call will only work if all the entries passed into m_list + * were added previously. It will not attempt to do a partial remove of entries + * that were found. */ int ice_remove_mac(struct ice_hw *hw, struct list_head *m_list) { @@ -3192,8 +3189,7 @@ int ice_remove_mac(struct ice_hw *hw, struct list_head *m_list) * @hw: pointer to the hardware structure * @v_list: list of VLAN entries and forwarding information */ -int -ice_remove_vlan(struct ice_hw *hw, struct list_head *v_list) +int ice_remove_vlan(struct ice_hw *hw, struct list_head *v_list) { struct ice_fltr_list_entry *v_list_itr, *tmp; @@ -3350,8 +3346,7 @@ static u8 ice_determine_promisc_mask(struct ice_fltr_info *fi) * @v_list: list of promisc entries */ static int -ice_remove_promisc(struct ice_hw *hw, u8 recp_id, - struct list_head *v_list) +ice_remove_promisc(struct ice_hw *hw, u8 recp_id, struct list_head *v_list) { struct ice_fltr_list_entry *v_list_itr, *tmp; @@ -3450,8 +3445,8 @@ ice_set_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask, u16 vid) enum { UCAST_FLTR = 1, MCAST_FLTR, BCAST_FLTR }; struct ice_fltr_list_entry f_list_entry; struct ice_fltr_info new_fltr; - int status = 0; bool is_tx_fltr; + int status = 0; u16 hw_vsi_id; int pkt_type; u8 recipe_id; @@ -3567,8 +3562,8 @@ ice_set_vlan_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask, struct list_head vsi_list_head; struct list_head *vlan_head; struct mutex *vlan_lock; /* Lock to protect filter rule list */ - int status; u16 vlan_id; + int status; INIT_LIST_HEAD(&vsi_list_head); vlan_lock = &sw->recp_list[ICE_SW_LKUP_VLAN].filt_rule_lock; @@ -3686,8 +3681,8 @@ ice_alloc_res_cntr(struct ice_hw *hw, u8 type, u8 alloc_shared, u16 num_items, u16 *counter_id) { struct ice_aqc_alloc_free_res_elem *buf; - int status; u16 buf_len; + int status; /* Allocate resource */ buf_len = struct_size(buf, elem, 1); @@ -3724,8 +3719,8 @@ ice_free_res_cntr(struct ice_hw *hw, u8 type, u8 alloc_shared, u16 num_items, u16 counter_id) { struct ice_aqc_alloc_free_res_elem *buf; - int status; u16 buf_len; + int status; /* Free resource */ buf_len = struct_size(buf, elem, 1); @@ -4123,11 +4118,11 @@ ice_add_sw_recipe(struct ice_hw *hw, struct ice_sw_recipe *rm, struct ice_aqc_recipe_data_elem *tmp; struct ice_aqc_recipe_data_elem *buf; struct ice_recp_grp_entry *entry; - int status; u16 free_res_idx; u16 recipe_count; u8 chain_idx; u8 recps = 0; + int status; /* When more than one recipe are required, another recipe is needed to * chain them together. Matching a tunnel metadata ID takes up one of @@ -4411,8 +4406,8 @@ static int ice_create_recipe_group(struct ice_hw *hw, struct ice_sw_recipe *rm, struct ice_prot_lkup_ext *lkup_exts) { - int status; u8 recp_count = 0; + int status; rm->n_grp_count = 0; @@ -4446,8 +4441,8 @@ static int ice_get_fv(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, u16 lkups_cnt, unsigned long *bm, struct list_head *fv_list) { - int status; u8 *prot_ids; + int status; u16 i; prot_ids = kcalloc(lkups_cnt, sizeof(*prot_ids), GFP_KERNEL); @@ -4572,8 +4567,8 @@ ice_add_adv_recipe(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, struct ice_sw_fv_list_entry *fvit; struct ice_recp_grp_entry *r_tmp; struct ice_sw_fv_list_entry *tmp; - int status = 0; struct ice_sw_recipe *rm; + int status = 0; u8 i; if (!lkups_cnt) @@ -5060,8 +5055,8 @@ ice_adv_add_update_vsi_list(struct ice_hw *hw, struct ice_adv_rule_info *cur_fltr, struct ice_adv_rule_info *new_fltr) { - int status; u16 vsi_list_id = 0; + int status; if (cur_fltr->sw_act.fltr_act == ICE_FWD_TO_Q || cur_fltr->sw_act.fltr_act == ICE_FWD_TO_QGRP || @@ -5171,10 +5166,10 @@ ice_add_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, struct ice_aqc_sw_rules_elem *s_rule = NULL; struct list_head *rule_head; struct ice_switch_info *sw; - int status; const u8 *pkt = NULL; u16 word_cnt; u32 act = 0; + int status; u8 q_rgn; /* Initialize profile to result index bitmap */ @@ -5439,8 +5434,8 @@ ice_adv_rem_update_vsi_list(struct ice_hw *hw, u16 vsi_handle, { struct ice_vsi_list_map_info *vsi_list_info; enum ice_sw_lkup_type lkup_type; - int status; u16 vsi_list_id; + int status; if (fm_list->rule_info.sw_act.fltr_act != ICE_FWD_TO_VSI_LIST || fm_list->vsi_count == 0) @@ -5538,10 +5533,10 @@ ice_rem_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, { struct ice_adv_fltr_mgmt_list_entry *list_elem; struct ice_prot_lkup_ext lkup_exts; - int status = 0; bool remove_rule = false; struct mutex *rule_lock; /* Lock to protect filter rule list */ u16 i, rid, vsi_handle; + int status = 0; memset(&lkup_exts, 0, sizeof(lkup_exts)); for (i = 0; i < lkups_cnt; i++) { diff --git a/drivers/net/ethernet/intel/ice/ice_switch.h b/drivers/net/ethernet/intel/ice/ice_switch.h index 8238dae0babd..d8334beaaa8a 100644 --- a/drivers/net/ethernet/intel/ice/ice_switch.h +++ b/drivers/net/ethernet/intel/ice/ice_switch.h @@ -330,24 +330,19 @@ ice_add_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups, u16 lkups_cnt, struct ice_adv_rule_info *rinfo, struct ice_rule_query_data *added_entry); int ice_update_sw_rule_bridge_mode(struct ice_hw *hw); +int ice_add_vlan(struct ice_hw *hw, struct list_head *m_list); +int ice_remove_vlan(struct ice_hw *hw, struct list_head *v_list); int ice_add_mac(struct ice_hw *hw, struct list_head *m_lst); int ice_remove_mac(struct ice_hw *hw, struct list_head *m_lst); -int -ice_add_eth_mac(struct ice_hw *hw, struct list_head *em_list); -int -ice_remove_eth_mac(struct ice_hw *hw, struct list_head *em_list); -int -ice_cfg_rdma_fltr(struct ice_hw *hw, u16 vsi_handle, bool enable); bool ice_mac_fltr_exist(struct ice_hw *hw, u8 *mac, u16 vsi_handle); bool ice_vlan_fltr_exist(struct ice_hw *hw, u16 vlan_id, u16 vsi_handle); +int ice_add_eth_mac(struct ice_hw *hw, struct list_head *em_list); +int ice_remove_eth_mac(struct ice_hw *hw, struct list_head *em_list); +int ice_cfg_rdma_fltr(struct ice_hw *hw, u16 vsi_handle, bool enable); void ice_remove_vsi_fltr(struct ice_hw *hw, u16 vsi_handle); -int -ice_add_vlan(struct ice_hw *hw, struct list_head *m_list); -int ice_remove_vlan(struct ice_hw *hw, struct list_head *v_list); /* Promisc/defport setup for VSIs */ -int -ice_cfg_dflt_vsi(struct ice_hw *hw, u16 vsi_handle, bool set, u8 direction); +int ice_cfg_dflt_vsi(struct ice_hw *hw, u16 vsi_handle, bool set, u8 direction); int ice_set_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask, u16 vid); @@ -358,8 +353,7 @@ int ice_set_vlan_vsi_promisc(struct ice_hw *hw, u16 vsi_handle, u8 promisc_mask, bool rm_vlan_promisc); -int -ice_rem_adv_rule_for_vsi(struct ice_hw *hw, u16 vsi_handle); +int ice_rem_adv_rule_for_vsi(struct ice_hw *hw, u16 vsi_handle); int ice_rem_adv_rule_by_id(struct ice_hw *hw, struct ice_rule_query_data *remove_entry); diff --git a/drivers/net/ethernet/intel/ice/ice_tc_lib.c b/drivers/net/ethernet/intel/ice/ice_tc_lib.c index 21c4bc0db33b..476fcdf9c794 100644 --- a/drivers/net/ethernet/intel/ice/ice_tc_lib.c +++ b/drivers/net/ethernet/intel/ice/ice_tc_lib.c @@ -398,9 +398,9 @@ ice_eswitch_add_tc_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr) struct ice_hw *hw = &vsi->back->hw; struct ice_adv_lkup_elem *list; u32 flags = fltr->flags; - int status; int lkups_cnt; int ret = 0; + int status; int i; if (!flags || (flags & ICE_TC_FLWR_FIELD_ENC_SRC_L4_PORT)) { diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_fdir.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_fdir.c index 5db87f7999d3..4ce30ff96a79 100644 --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_fdir.c +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_fdir.c @@ -560,13 +560,13 @@ ice_vc_fdir_write_flow_prof(struct ice_vf *vf, enum ice_fltr_ptype flow, struct ice_flow_seg_info *old_seg; struct ice_flow_prof *prof = NULL; struct ice_fd_hw_prof *vf_prof; - int status; struct device *dev; struct ice_pf *pf; struct ice_hw *hw; u64 entry1_h = 0; u64 entry2_h = 0; u64 prof_id; + int status; int ret; pf = vf->pf; @@ -1199,10 +1199,10 @@ static int ice_vc_fdir_write_fltr(struct ice_vf *vf, struct ice_fdir_fltr *input = &conf->input; struct ice_vsi *vsi, *ctrl_vsi; struct ice_fltr_desc desc; - int status; struct device *dev; struct ice_pf *pf; struct ice_hw *hw; + int status; int ret; u8 *pkt; diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c index ece5aa897ab8..92ddc3042d0d 100644 --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c @@ -653,8 +653,8 @@ static int ice_vsi_manage_pvid(struct ice_vsi *vsi, u16 pvid_info, bool enable) struct ice_hw *hw = &vsi->back->hw; struct ice_aqc_vsi_props *info; struct ice_vsi_ctx *ctxt; - int status; int ret = 0; + int status; ctxt = kzalloc(sizeof(*ctxt), GFP_KERNEL); if (!ctxt) @@ -680,8 +680,7 @@ static int ice_vsi_manage_pvid(struct ice_vsi *vsi, u16 pvid_info, bool enable) status = ice_update_vsi(hw, vsi->idx, ctxt, NULL); if (status) { dev_info(ice_hw_to_dev(hw), "update VSI for port VLAN failed, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); ret = -EIO; goto out; } @@ -851,8 +850,8 @@ static int ice_vf_rebuild_host_mac_cfg(struct ice_vf *vf) { struct device *dev = ice_pf_to_dev(vf->pf); struct ice_vsi *vsi = ice_get_vf_vsi(vf); - int status; u8 broadcast[ETH_ALEN]; + int status; if (ice_is_eswitch_mode_switchdev(vf->pf)) return 0; @@ -1303,8 +1302,8 @@ static void ice_vf_pre_vsi_rebuild(struct ice_vf *vf) static void ice_vf_rebuild_aggregator_node_cfg(struct ice_vsi *vsi) { struct ice_pf *pf = vsi->back; - int status; struct device *dev; + int status; if (!vsi->agg_node) return; @@ -1734,9 +1733,9 @@ static int ice_init_vf_vsi_res(struct ice_vf *vf) { struct ice_pf *pf = vf->pf; u8 broadcast[ETH_ALEN]; - int status; struct ice_vsi *vsi; struct device *dev; + int status; int err; vf->first_vector_idx = ice_calc_vf_first_vector_idx(pf, vf); @@ -2160,9 +2159,9 @@ int ice_vc_send_msg_to_vf(struct ice_vf *vf, u32 v_opcode, enum virtchnl_status_code v_retval, u8 *msg, u16 msglen) { - int aq_ret; struct device *dev; struct ice_pf *pf; + int aq_ret; if (!vf) return -EINVAL; @@ -2672,8 +2671,8 @@ static int ice_vc_handle_rss_cfg(struct ice_vf *vf, u8 *msg, bool add) if (rss_cfg->rss_algorithm == VIRTCHNL_RSS_ALG_R_ASYMMETRIC) { struct ice_vsi_ctx *ctx; - int status; u8 lut_type, hash_type; + int status; lut_type = ICE_AQ_VSI_Q_OPT_RSS_LUT_VSI; hash_type = add ? ICE_AQ_VSI_Q_OPT_RSS_XOR : @@ -2703,8 +2702,7 @@ static int ice_vc_handle_rss_cfg(struct ice_vf *vf, u8 *msg, bool add) status = ice_update_vsi(hw, vsi->idx, ctx, NULL); if (status) { dev_err(dev, "update VSI for RSS failed, err %d aq_err %s\n", - status, - ice_aq_str(hw->adminq.sq_last_status)); + status, ice_aq_str(hw->adminq.sq_last_status)); v_ret = VIRTCHNL_STATUS_ERR_PARAM; } else { vsi->info.q_opt_rss = ctx->info.q_opt_rss; @@ -2733,10 +2731,9 @@ static int ice_vc_handle_rss_cfg(struct ice_vf *vf, u8 *msg, bool add) status = ice_rem_rss_cfg(hw, vsi->idx, hash_flds, addl_hdrs); - /* We just ignore -ENOENT, because - * if two configurations share the same profile remove - * one of them actually removes both, since the - * profile is deleted. + /* We just ignore -ENOENT, because if two configurations + * share the same profile remove one of them actually + * removes both, since the profile is deleted. */ if (status && status != -ENOENT) { v_ret = VIRTCHNL_STATUS_ERR_PARAM; @@ -2899,9 +2896,9 @@ int ice_set_vf_spoofchk(struct net_device *netdev, int vf_id, bool ena) struct ice_pf *pf = np->vsi->back; struct ice_vsi_ctx *ctx; struct ice_vsi *vf_vsi; - int status; struct device *dev; struct ice_vf *vf; + int status; int ret; dev = ice_pf_to_dev(pf); @@ -3789,8 +3786,8 @@ ice_vc_add_mac_addr(struct ice_vf *vf, struct ice_vsi *vsi, { struct device *dev = ice_pf_to_dev(vf->pf); u8 *mac_addr = vc_ether_addr->addr; - int status; int ret = 0; + int status; /* device MAC already added */ if (ether_addr_equal(mac_addr, vf->dev_lan_addr.addr)) @@ -5265,9 +5262,9 @@ ice_is_malicious_vf(struct ice_pf *pf, struct ice_rq_event_info *event, s16 vf_id = le16_to_cpu(event->desc.retval); struct device *dev = ice_pf_to_dev(pf); struct ice_mbx_data mbxdata; - int status; bool malvf = false; struct ice_vf *vf; + int status; if (ice_validate_vf_id(pf, vf_id)) return false; -- 2.31.1