From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id CF46FA0C41; Thu, 16 Sep 2021 11:50:18 +0200 (CEST) Received: from [217.70.189.124] (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 8CBC6410F8; Thu, 16 Sep 2021 11:50:08 +0200 (CEST) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by mails.dpdk.org (Postfix) with ESMTP id 2BD01410EF for ; Thu, 16 Sep 2021 11:50:07 +0200 (CEST) X-IronPort-AV: E=McAfee;i="6200,9189,10108"; a="222185854" X-IronPort-AV: E=Sophos;i="5.85,298,1624345200"; d="scan'208";a="222185854" Received: from fmsmga008.fm.intel.com ([10.253.24.58]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 16 Sep 2021 02:50:06 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.85,298,1624345200"; d="scan'208";a="509246900" Received: from dpdk51.sh.intel.com ([10.67.111.142]) by fmsmga008.fm.intel.com with ESMTP; 16 Sep 2021 02:50:04 -0700 From: Qi Zhang To: qiming.yang@intel.com Cc: junfeng.guo@intel.com, dev@dpdk.org, Qi Zhang , Jacob Keller Date: Thu, 16 Sep 2021 17:52:55 +0800 Message-Id: <20210916095304.3058210-4-qi.z.zhang@intel.com> X-Mailer: git-send-email 2.26.2 In-Reply-To: <20210916095304.3058210-1-qi.z.zhang@intel.com> References: <20210916095304.3058210-1-qi.z.zhang@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Subject: [dpdk-dev] [PATCH 03/12] net/ice/base: use macro instead of open-coded division X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" For some operating systems, 64-bit division requires using specific implementations. Use the DIV_64BIT macro to replace open-coded division so that the driver may convert this to the appropriate operating-system specific implementation when necessary. Signed-off-by: Jacob Keller Signed-off-by: Qi Zhang --- drivers/net/ice/base/ice_ptp_hw.c | 53 +++++++++++++++++++------------ 1 file changed, 32 insertions(+), 21 deletions(-) diff --git a/drivers/net/ice/base/ice_ptp_hw.c b/drivers/net/ice/base/ice_ptp_hw.c index 8ea75538fa..70eb87abf9 100644 --- a/drivers/net/ice/base/ice_ptp_hw.c +++ b/drivers/net/ice/base/ice_ptp_hw.c @@ -1634,7 +1634,7 @@ static enum ice_status ice_phy_cfg_uix_e822(struct ice_hw *hw, u8 port) #define LINE_UI_25G_100G 256 /* 6600 UIs is 256 nanoseconds at 25Gb/100Gb */ /* Program the 10Gb/40Gb conversion ratio */ - uix = (tu_per_sec * LINE_UI_10G_40G) / 390625000; + uix = DIV_64BIT(tu_per_sec * LINE_UI_10G_40G, 390625000); status = ice_write_64b_phy_reg_e822(hw, port, P_REG_UIX66_10G_40G_L, uix); @@ -1645,7 +1645,7 @@ static enum ice_status ice_phy_cfg_uix_e822(struct ice_hw *hw, u8 port) } /* Program the 25Gb/100Gb conversion ratio */ - uix = (tu_per_sec * LINE_UI_25G_100G) / 390625000; + uix = DIV_64BIT(tu_per_sec * LINE_UI_25G_100G, 390625000); status = ice_write_64b_phy_reg_e822(hw, port, P_REG_UIX66_25G_100G_L, uix); @@ -1727,7 +1727,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_PAR_TX_TUS */ if (e822_vernier[link_spd].tx_par_clk) - phy_tus = tu_per_sec / e822_vernier[link_spd].tx_par_clk; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].tx_par_clk); else phy_tus = 0; @@ -1738,7 +1739,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_PAR_RX_TUS */ if (e822_vernier[link_spd].rx_par_clk) - phy_tus = tu_per_sec / e822_vernier[link_spd].rx_par_clk; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].rx_par_clk); else phy_tus = 0; @@ -1749,7 +1751,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_PCS_TX_TUS */ if (e822_vernier[link_spd].tx_pcs_clk) - phy_tus = tu_per_sec / e822_vernier[link_spd].tx_pcs_clk; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].tx_pcs_clk); else phy_tus = 0; @@ -1760,7 +1763,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_PCS_RX_TUS */ if (e822_vernier[link_spd].rx_pcs_clk) - phy_tus = tu_per_sec / e822_vernier[link_spd].rx_pcs_clk; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].rx_pcs_clk); else phy_tus = 0; @@ -1771,7 +1775,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_DESK_PAR_TX_TUS */ if (e822_vernier[link_spd].tx_desk_rsgb_par) - phy_tus = tu_per_sec / e822_vernier[link_spd].tx_desk_rsgb_par; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].tx_desk_rsgb_par); else phy_tus = 0; @@ -1782,7 +1787,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_DESK_PAR_RX_TUS */ if (e822_vernier[link_spd].rx_desk_rsgb_par) - phy_tus = tu_per_sec / e822_vernier[link_spd].rx_desk_rsgb_par; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].rx_desk_rsgb_par); else phy_tus = 0; @@ -1793,7 +1799,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_DESK_PCS_TX_TUS */ if (e822_vernier[link_spd].tx_desk_rsgb_pcs) - phy_tus = tu_per_sec / e822_vernier[link_spd].tx_desk_rsgb_pcs; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].tx_desk_rsgb_pcs); else phy_tus = 0; @@ -1804,7 +1811,8 @@ static enum ice_status ice_phy_cfg_parpcs_e822(struct ice_hw *hw, u8 port) /* P_REG_DESK_PCS_RX_TUS */ if (e822_vernier[link_spd].rx_desk_rsgb_pcs) - phy_tus = tu_per_sec / e822_vernier[link_spd].rx_desk_rsgb_pcs; + phy_tus = DIV_64BIT(tu_per_sec, + e822_vernier[link_spd].rx_desk_rsgb_pcs); else phy_tus = 0; @@ -1836,9 +1844,9 @@ ice_calc_fixed_tx_offset_e822(struct ice_hw *hw, enum ice_ptp_link_spd link_spd) * overflows 64 bit integer arithmetic, so break it up into two * divisions by 1e4 first then by 1e7. */ - fixed_offset = tu_per_sec / 10000; + fixed_offset = DIV_64BIT(tu_per_sec, 10000); fixed_offset *= e822_vernier[link_spd].tx_fixed_delay; - fixed_offset /= 10000000; + fixed_offset = DIV_64BIT(fixed_offset, 10000000); return fixed_offset; } @@ -1982,9 +1990,9 @@ ice_phy_calc_pmd_adj_e822(struct ice_hw *hw, u8 port, enum ice_ptp_fec_mode fec_mode, u64 *pmd_adj) { u64 cur_freq, clk_incval, tu_per_sec, mult, adj; + u32 pmd_adj_divisor, val; enum ice_status status; u8 pmd_align; - u32 val; status = ice_read_phy_reg_e822(hw, port, P_REG_PMD_ALIGNMENT, &val); if (status) { @@ -2001,6 +2009,9 @@ ice_phy_calc_pmd_adj_e822(struct ice_hw *hw, u8 port, /* Calculate TUs per second */ tu_per_sec = cur_freq * clk_incval; + /* Get the link speed dependent PMD adjustment divisor */ + pmd_adj_divisor = e822_vernier[link_spd].pmd_adj_divisor; + /* The PMD alignment adjustment measurement depends on the link speed, * and whether FEC is enabled. For each link speed, the alignment * adjustment is calculated by dividing a value by the length of @@ -2063,9 +2074,9 @@ ice_phy_calc_pmd_adj_e822(struct ice_hw *hw, u8 port, * divide by 125, and then handle remaining divisor based on the link * speed pmd_adj_divisor value. */ - adj = tu_per_sec / 125; + adj = DIV_64BIT(tu_per_sec, 125); adj *= mult; - adj /= e822_vernier[link_spd].pmd_adj_divisor; + adj = DIV_64BIT(adj, pmd_adj_divisor); /* Finally, for 25G-RS and 50G-RS, a further adjustment for the Rx * cycle count is necessary. @@ -2086,9 +2097,9 @@ ice_phy_calc_pmd_adj_e822(struct ice_hw *hw, u8 port, if (rx_cycle) { mult = (4 - rx_cycle) * 40; - cycle_adj = tu_per_sec / 125; + cycle_adj = DIV_64BIT(tu_per_sec, 125); cycle_adj *= mult; - cycle_adj /= e822_vernier[link_spd].pmd_adj_divisor; + cycle_adj = DIV_64BIT(cycle_adj, pmd_adj_divisor); adj += cycle_adj; } @@ -2108,9 +2119,9 @@ ice_phy_calc_pmd_adj_e822(struct ice_hw *hw, u8 port, if (rx_cycle) { mult = rx_cycle * 40; - cycle_adj = tu_per_sec / 125; + cycle_adj = DIV_64BIT(tu_per_sec, 125); cycle_adj *= mult; - cycle_adj /= e822_vernier[link_spd].pmd_adj_divisor; + cycle_adj = DIV_64BIT(cycle_adj, pmd_adj_divisor); adj += cycle_adj; } @@ -2146,9 +2157,9 @@ ice_calc_fixed_rx_offset_e822(struct ice_hw *hw, enum ice_ptp_link_spd link_spd) * overflows 64 bit integer arithmetic, so break it up into two * divisions by 1e4 first then by 1e7. */ - fixed_offset = tu_per_sec / 10000; + fixed_offset = DIV_64BIT(tu_per_sec, 10000); fixed_offset *= e822_vernier[link_spd].rx_fixed_delay; - fixed_offset /= 10000000; + fixed_offset = DIV_64BIT(fixed_offset, 10000000); return fixed_offset; } -- 2.26.2