From patchwork Fri Jul 26 10:02:52 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742555 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgsg2.qq.com (smtpbgsg2.qq.com [54.254.200.128]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D4554225AF for ; Fri, 26 Jul 2024 10:03:30 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.254.200.128 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988215; cv=none; b=tqg0MucmfHCc5IxUs/w8v+ahdiiZ7yxzElruKjoKo5wqO1nKabSd2a1FL5OcjkQbY5GaGWKhz08nCNkQ1bdtdkZlyUKybOfmjXI5+M5jppLemyMPtfttfp2CDV8kkjiv9xSUmTM/cLCALHO0W8EFIz6okYjvutzyabrnmDPCWCg= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988215; c=relaxed/simple; bh=XGQFYd4TTTsO0ZdoGpaj9PokQ4NvXa5G6NXFVoL92e8=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=SfFa0JQnNrijxbDNsHJmUhWbya5EoL13AGqRkf6i0+0Kii1f4PXQqQ0hlfvv/LrBo5qp8kJDgF74sRAM44bdhGNM9EwTBFPQ1zabZNNO5gv9rMhCfmUOdPymbpMFlX5wDKLmXVCEvcmdR+b8ALF+PsDIZijNwflisEzBTGPRKyM= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.254.200.128 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988201ts08xi7z X-QQ-Originating-IP: DROWQN4n7DmZklOG7AkkAuTOlsM5KV1Rrb9LYnN7smc= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:19 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 5499599651212820689 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 01/10] net: libwx: Add malibox api for wangxun pf drivers Date: Fri, 26 Jul 2024 18:02:52 +0800 Message-ID: <0A63D5305A7F0B97+20240726100301.21416-2-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Implements the mailbox interfaces for wangxun pf drivers ngbe and txgbe. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/Makefile | 2 +- drivers/net/ethernet/wangxun/libwx/wx_mbx.c | 175 +++++++++++++++++++ drivers/net/ethernet/wangxun/libwx/wx_mbx.h | 32 ++++ drivers/net/ethernet/wangxun/libwx/wx_type.h | 8 + 4 files changed, 216 insertions(+), 1 deletion(-) create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_mbx.c create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_mbx.h diff --git a/drivers/net/ethernet/wangxun/libwx/Makefile b/drivers/net/ethernet/wangxun/libwx/Makefile index 42ccd6e4052e..913a978c9032 100644 --- a/drivers/net/ethernet/wangxun/libwx/Makefile +++ b/drivers/net/ethernet/wangxun/libwx/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_LIBWX) += libwx.o -libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o +libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o diff --git a/drivers/net/ethernet/wangxun/libwx/wx_mbx.c b/drivers/net/ethernet/wangxun/libwx/wx_mbx.c new file mode 100644 index 000000000000..5062ddb2ce39 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_mbx.c @@ -0,0 +1,175 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2015 - 2024 Beijing WangXun Technology Co., Ltd. */ +#include +#include "wx_type.h" +#include "wx_mbx.h" + +/** + * wx_obtain_mbx_lock_pf - obtain mailbox lock + * @wx: pointer to the HW structure + * @vf: the VF index + * + * return: return 0 on success and -EBUSY on failure + **/ +static int wx_obtain_mbx_lock_pf(struct wx *wx, u16 vf) +{ + int count = 5; + u32 mailbox; + + while (count--) { + /* Take ownership of the buffer */ + wr32(wx, WX_PXMAILBOX(vf), WX_PXMAILBOX_PFU); + + /* reserve mailbox for vf use */ + mailbox = rd32(wx, WX_PXMAILBOX(vf)); + if (mailbox & WX_PXMAILBOX_PFU) + return 0; + else if (count != 1) + udelay(10); + } + wx_err(wx, "Failed to obtain mailbox lock for PF%d", vf); + + return -EBUSY; +} + +static int wx_check_for_bit_pf(struct wx *wx, u32 mask, int index) +{ + u32 mbvficr = rd32(wx, WX_MBVFICR(index)); + + if ((mbvficr & mask) == 0) + return -EBUSY; + wr32(wx, WX_MBVFICR(index), mask); + + return 0; +} + +/** + * wx_check_for_ack_pf - checks to see if the VF has acked + * @wx: pointer to the HW structure + * @vf: the VF index + * + * return: return 0 if the VF has set the status bit or else -EBUSY + **/ +int wx_check_for_ack_pf(struct wx *wx, u16 vf) +{ + u32 index = vf / 16, vf_bit = vf % 16; + + return wx_check_for_bit_pf(wx, + FIELD_PREP(WX_MBVFICR_VFACK_MASK, + BIT(vf_bit)), + index); +} + +/** + * wx_check_for_msg_pf - checks to see if the VF has sent mail + * @wx: pointer to the HW structure + * @vf: the VF index + * + * return: return 0 if the VF has got req bit or else -EBUSY + **/ +int wx_check_for_msg_pf(struct wx *wx, u16 vf) +{ + u32 index = vf / 16, vf_bit = vf % 16; + + return wx_check_for_bit_pf(wx, + FIELD_PREP(WX_MBVFICR_VFREQ_MASK, + BIT(vf_bit)), + index); +} + +/** + * wx_write_mbx_pf - Places a message in the mailbox + * @wx: pointer to the HW structure + * @msg: The message buffer + * @size: Length of buffer + * @vf: the VF index + * + * return: return 0 on success and -EINVAL/-EBUSY on failure + **/ +int wx_write_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf) +{ + struct wx_mbx_info *mbx = &wx->mbx; + int ret, i; + + /* mbx->size is up to 15 */ + if (size > mbx->size) { + wx_err(wx, "Invalid mailbox message size %d", size); + return -EINVAL; + } + + /* lock the mailbox to prevent pf/vf race condition */ + ret = wx_obtain_mbx_lock_pf(wx, vf); + if (ret) + return ret; + + /* flush msg and acks as we are overwriting the message buffer */ + wx_check_for_msg_pf(wx, vf); + wx_check_for_ack_pf(wx, vf); + + /* copy the caller specified message to the mailbox memory buffer */ + for (i = 0; i < size; i++) + wr32a(wx, WX_PXMBMEM(vf), i, msg[i]); + + /* Interrupt VF to tell it a message has been sent and release buffer*/ + /* set mirrored mailbox flags */ + wr32a(wx, WX_PXMBMEM(vf), WX_VXMAILBOX_SIZE, WX_PXMAILBOX_STS); + wr32(wx, WX_PXMAILBOX(vf), WX_PXMAILBOX_STS); + + return 0; +} + +/** + * wx_read_mbx_pf - Read a message from the mailbox + * @wx: pointer to the HW structure + * @msg: The message buffer + * @size: Length of buffer + * @vf: the VF index + * + * return: return 0 on success and -EBUSY on failure + **/ +int wx_read_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf) +{ + struct wx_mbx_info *mbx = &wx->mbx; + int ret; + u16 i; + + /* limit read to size of mailbox and mbx->size is up to 15 */ + if (size > mbx->size) + size = mbx->size; + + /* lock the mailbox to prevent pf/vf race condition */ + ret = wx_obtain_mbx_lock_pf(wx, vf); + if (ret) + return ret; + + for (i = 0; i < size; i++) + msg[i] = rd32a(wx, WX_PXMBMEM(vf), i); + + /* Acknowledge the message and release buffer */ + /* set mirrored mailbox flags */ + wr32a(wx, WX_PXMBMEM(vf), WX_VXMAILBOX_SIZE, WX_PXMAILBOX_ACK); + wr32(wx, WX_PXMAILBOX(vf), WX_PXMAILBOX_ACK); + + return 0; +} + +/** + * wx_check_for_rst_pf - checks to see if the VF has reset + * @wx: pointer to the HW structure + * @vf: the VF index + * + * return: return 0 on success and -EBUSY on failure + **/ +int wx_check_for_rst_pf(struct wx *wx, u16 vf) +{ + u32 reg_offset = WX_VF_REG_OFFSET(vf); + u32 vf_shift = WX_VF_IND_SHIFT(vf); + u32 vflre = 0; + + vflre = rd32(wx, WX_VFLRE(reg_offset)); + if (!(vflre & BIT(vf_shift))) + return -EBUSY; + wr32(wx, WX_VFLREC(reg_offset), BIT(vf_shift)); + + return 0; +} diff --git a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h new file mode 100644 index 000000000000..1579096fb6ad --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h @@ -0,0 +1,32 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2015 - 2024 Beijing WangXun Technology Co., Ltd. */ +#ifndef _WX_MBX_H_ +#define _WX_MBX_H_ + +#define WX_VXMAILBOX_SIZE 15 + +/* PF Registers */ +#define WX_PXMAILBOX(i) (0x600 + (4 * (i))) /* i=[0,63] */ +#define WX_PXMAILBOX_STS BIT(0) /* Initiate message send to VF */ +#define WX_PXMAILBOX_ACK BIT(1) /* Ack message recv'd from VF */ +#define WX_PXMAILBOX_PFU BIT(3) /* PF owns the mailbox buffer */ + +#define WX_PXMBMEM(i) (0x5000 + (64 * (i))) /* i=[0,63] */ + +#define WX_VFLRE(i) (0x4A0 + (4 * (i))) /* i=[0,1] */ +#define WX_VFLREC(i) (0x4A8 + (4 * (i))) /* i=[0,1] */ + +/* SR-IOV specific macros */ +#define WX_MBVFICR(i) (0x480 + (4 * (i))) /* i=[0,3] */ +#define WX_MBVFICR_VFREQ_MASK GENMASK(15, 0) +#define WX_MBVFICR_VFACK_MASK GENMASK(31, 16) + +#define WX_VT_MSGINFO_MASK GENMASK(23, 16) + +int wx_write_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf); +int wx_read_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf); +int wx_check_for_rst_pf(struct wx *wx, u16 mbx_id); +int wx_check_for_msg_pf(struct wx *wx, u16 mbx_id); +int wx_check_for_ack_pf(struct wx *wx, u16 mbx_id); + +#endif /* _WX_MBX_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index 1d57b047817b..5b9d5fd7033d 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -20,6 +20,9 @@ #define WX_PCI_LINK_STATUS 0xB2 /**************** Global Registers ****************************/ +#define WX_VF_REG_OFFSET(_v) ((_v) >> 5) +#define WX_VF_IND_SHIFT(_v) ((_v) & 0x1f) + /* chip control Registers */ #define WX_MIS_PWR 0x10000 #define WX_MIS_RST 0x1000C @@ -707,6 +710,10 @@ struct wx_bus_info { u16 device; }; +struct wx_mbx_info { + u16 size; +}; + struct wx_thermal_sensor_data { s16 temp; s16 alarm_thresh; @@ -1046,6 +1053,7 @@ struct wx { struct pci_dev *pdev; struct net_device *netdev; struct wx_bus_info bus; + struct wx_mbx_info mbx; struct wx_mac_info mac; enum em_mac_type mac_type; enum sp_media_type media_type; From patchwork Fri Jul 26 10:02:53 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742558 X-Patchwork-Delegate: kuba@kernel.org Received: from bg1.exmail.qq.com (bg1.exmail.qq.com [114.132.77.159]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id E617F173332 for ; Fri, 26 Jul 2024 10:03:37 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=114.132.77.159 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988221; cv=none; b=DcShOmBw5GUBW6KzrOmHusCMEvjvJXA9FvhO7MjDJ645aE8QOlQVuAqtj9jJ1VknZfgyYw89DI8FgxUBk32qdEY636TO7HO8FcmXZXZ1v5y71SVT+qMMf+weVPTm7UiMHIXEZa+u1qE+WWAFpRpllPyyiSATyzX0bI2wJKF76+A= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988221; c=relaxed/simple; bh=EqgwY+V6Jh/SfKS1iygzNr7b8Oh8X6n1B8iTdUmX/ok=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=bRvXpZTKVi7ElGxaLWAfVsoZmhbPCoRApUCL0m07pyzwZ/AFGDHx7HA3Try10W3wzGMWmTSSvaNNf75dDknWj7KH7at+OaS8c+pcIPXlF019fFNuosDH6akGtS5VnufG2rxfY1KYuZsKsbR/m8JbDHi7l1pmWVAktAfFpnh31vU= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=114.132.77.159 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988203tvmmrcd6 X-QQ-Originating-IP: hIXZJ/3sx5RjUFIeKeGOWYqCEXCPwz3ranJnhM2neqc= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:22 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 14795770019563736396 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 02/10] net: libwx: Add sriov api for wangxun nics Date: Fri, 26 Jul 2024 18:02:53 +0800 Message-ID: <4437D4BD16BDFAEF+20240726100301.21416-3-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Implement sriov_configure interface for wangxun nics in libwx. Enable VT mode and initialize vf control structure, when sriov is enabled. Do not be allowed to disable sriov when vfs are assigned. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/Makefile | 2 +- drivers/net/ethernet/wangxun/libwx/wx_mbx.h | 4 + drivers/net/ethernet/wangxun/libwx/wx_sriov.c | 210 ++++++++++++++++++ drivers/net/ethernet/wangxun/libwx/wx_sriov.h | 10 + drivers/net/ethernet/wangxun/libwx/wx_type.h | 31 +++ 5 files changed, 256 insertions(+), 1 deletion(-) create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_sriov.c create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_sriov.h diff --git a/drivers/net/ethernet/wangxun/libwx/Makefile b/drivers/net/ethernet/wangxun/libwx/Makefile index 913a978c9032..5b996d973d29 100644 --- a/drivers/net/ethernet/wangxun/libwx/Makefile +++ b/drivers/net/ethernet/wangxun/libwx/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_LIBWX) += libwx.o -libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o +libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o wx_sriov.o diff --git a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h index 1579096fb6ad..3c70654a8b14 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h @@ -23,6 +23,10 @@ #define WX_VT_MSGINFO_MASK GENMASK(23, 16) +enum wxvf_xcast_modes { + WXVF_XCAST_MODE_NONE = 0, +}; + int wx_write_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf); int wx_read_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf); int wx_check_for_rst_pf(struct wx *wx, u16 mbx_id); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c new file mode 100644 index 000000000000..6b0a150abd76 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c @@ -0,0 +1,210 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2015 - 2023 Beijing WangXun Technology Co., Ltd. */ + +#include +#include + +#include "wx_type.h" +#include "wx_mbx.h" +#include "wx_sriov.h" + +static void wx_vf_configuration(struct pci_dev *pdev, int event_mask) +{ + unsigned int vfn = (event_mask & GENMASK(5, 0)); + struct wx *wx = pci_get_drvdata(pdev); + + bool enable = ((event_mask & BIT(31)) != 0); + + if (enable) + eth_zero_addr(wx->vfinfo[vfn].vf_mac_addr); +} + +static void wx_alloc_vf_macvlans(struct wx *wx, u8 num_vfs) +{ + struct vf_macvlans *mv_list; + int num_vf_macvlans, i; + + /* Initialize list of VF macvlans */ + INIT_LIST_HEAD(&wx->vf_mvs.mvlist); + + num_vf_macvlans = wx->mac.num_rar_entries - + (WX_MAX_PF_MACVLANS + 1 + num_vfs); + if (!num_vf_macvlans) + return; + + mv_list = kcalloc(num_vf_macvlans, sizeof(struct vf_macvlans), + GFP_KERNEL); + if (mv_list) { + for (i = 0; i < num_vf_macvlans; i++) { + mv_list[i].vf = -1; + mv_list[i].free = true; + list_add(&mv_list[i].mvlist, &wx->vf_mvs.mvlist); + } + wx->mv_list = mv_list; + } +} + +static void wx_sriov_clear_data(struct wx *wx) +{ + /* set num VFs to 0 to prevent access to vfinfo */ + wx->num_vfs = 0; + + /* free VF control structures */ + kfree(wx->vfinfo); + wx->vfinfo = NULL; + + /* free macvlan list */ + kfree(wx->mv_list); + wx->mv_list = NULL; + + /* set default pool back to 0 */ + wr32m(wx, WX_PSR_VM_CTL, WX_PSR_VM_CTL_POOL_MASK, 0); + wx->ring_feature[RING_F_VMDQ].offset = 0; + + clear_bit(WX_FLAG_SRIOV_ENABLED, wx->flags); + /* Disable VMDq flag so device will be set in NM mode */ + if (wx->ring_feature[RING_F_VMDQ].limit == 1) + clear_bit(WX_FLAG_VMDQ_ENABLED, wx->flags); +} + +static int __wx_enable_sriov(struct wx *wx, u8 num_vfs) +{ + u32 value = 0; + int i; + + set_bit(WX_FLAG_SRIOV_ENABLED, wx->flags); + wx_err(wx, "SR-IOV enabled with %d VFs\n", num_vfs); + + /* Enable VMDq flag so device will be set in VM mode */ + set_bit(WX_FLAG_VMDQ_ENABLED, wx->flags); + if (!wx->ring_feature[RING_F_VMDQ].limit) + wx->ring_feature[RING_F_VMDQ].limit = 1; + wx->ring_feature[RING_F_VMDQ].offset = num_vfs; + + wx->vfinfo = kcalloc(num_vfs, sizeof(struct vf_data_storage), + GFP_KERNEL); + if (!wx->vfinfo) + return -ENOMEM; + + wx_alloc_vf_macvlans(wx, num_vfs); + /* Initialize default switching mode VEB */ + wr32m(wx, WX_PSR_CTL, WX_PSR_CTL_SW_EN, WX_PSR_CTL_SW_EN); + + for (i = 0; i < num_vfs; i++) { + /* enable spoof checking for all VFs */ + wx->vfinfo[i].spoofchk_enabled = true; + wx->vfinfo[i].link_enable = true; + /* untrust all VFs */ + wx->vfinfo[i].trusted = false; + /* set the default xcast mode */ + wx->vfinfo[i].xcast_mode = WXVF_XCAST_MODE_NONE; + } + + if (wx->mac.type == wx_mac_sp) { + if (num_vfs < 32) + value = WX_CFG_PORT_CTL_NUM_VT_32; + else + value = WX_CFG_PORT_CTL_NUM_VT_64; + } else { + value = WX_CFG_PORT_CTL_NUM_VT_8; + } + wr32m(wx, WX_CFG_PORT_CTL, + WX_CFG_PORT_CTL_NUM_VT_MASK, + value); + + return 0; +} + +static void wx_sriov_reinit(struct wx *wx) +{ + rtnl_lock(); + wx->setup_tc(wx->netdev, netdev_get_num_tc(wx->netdev)); + rtnl_unlock(); +} + +void wx_disable_sriov(struct wx *wx) +{ + if (!pci_vfs_assigned(wx->pdev)) + pci_disable_sriov(wx->pdev); + else + wx_err(wx, "Unloading driver while VFs are assigned.\n"); + + /* clear flags and free allloced data */ + wx_sriov_clear_data(wx); +} +EXPORT_SYMBOL(wx_disable_sriov); + +static int wx_pci_sriov_enable(struct pci_dev *dev, + int num_vfs) +{ + struct wx *wx = pci_get_drvdata(dev); + int err = 0, i; + + err = __wx_enable_sriov(wx, num_vfs); + if (err) + return err; + + wx->num_vfs = num_vfs; + for (i = 0; i < wx->num_vfs; i++) + wx_vf_configuration(dev, (i | BIT(31))); + + /* reset before enabling SRIOV to avoid mailbox issues */ + wx_sriov_reinit(wx); + + err = pci_enable_sriov(dev, num_vfs); + if (err) { + wx_err(wx, "Failed to enable PCI sriov: %d\n", err); + goto err_out; + } + + return num_vfs; +err_out: + wx_sriov_clear_data(wx); + return err; +} + +static void wx_pci_sriov_disable(struct pci_dev *dev) +{ + struct wx *wx = pci_get_drvdata(dev); + + wx_disable_sriov(wx); + wx_sriov_reinit(wx); +} + +static int wx_check_sriov_allowed(struct wx *wx, int num_vfs) +{ + u16 max_vfs; + + max_vfs = (wx->mac.type == wx_mac_sp) ? 63 : 7; + if (num_vfs > max_vfs) + return -EPERM; + + return 0; +} + +int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs) +{ + struct wx *wx = pci_get_drvdata(pdev); + int err; + + err = wx_check_sriov_allowed(wx, num_vfs); + if (err) + return err; + + if (!num_vfs) { + if (!pci_vfs_assigned(pdev)) { + wx_pci_sriov_disable(pdev); + return 0; + } + + wx_err(wx, "can't free VFs because some are assigned to VMs.\n"); + return -EBUSY; + } + + err = wx_pci_sriov_enable(pdev, num_vfs); + if (err) + return err; + + return num_vfs; +} +EXPORT_SYMBOL(wx_pci_sriov_configure); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h new file mode 100644 index 000000000000..1ac03e9c0d55 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2015 - 2024 Beijing WangXun Technology Co., Ltd. */ + +#ifndef _WX_SRIOV_H_ +#define _WX_SRIOV_H_ + +void wx_disable_sriov(struct wx *wx); +int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs); + +#endif /* _WX_SRIOV_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index 5b9d5fd7033d..c4f73c5a366f 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -18,6 +18,7 @@ /* MSI-X capability fields masks */ #define WX_PCIE_MSIX_TBL_SZ_MASK 0x7FF #define WX_PCI_LINK_STATUS 0xB2 +#define WX_MAX_PF_MACVLANS 15 /**************** Global Registers ****************************/ #define WX_VF_REG_OFFSET(_v) ((_v) >> 5) @@ -91,6 +92,9 @@ #define WX_CFG_TAG_TPID(_i) (0x14430 + ((_i) * 4)) #define WX_CFG_PORT_CTL_NUM_VT_MASK GENMASK(13, 12) /* number of TVs */ +#define WX_CFG_PORT_CTL_NUM_VT_8 FIELD_PREP(GENMASK(13, 12), 1) +#define WX_CFG_PORT_CTL_NUM_VT_32 FIELD_PREP(GENMASK(13, 12), 2) +#define WX_CFG_PORT_CTL_NUM_VT_64 FIELD_PREP(GENMASK(13, 12), 3) /* GPIO Registers */ #define WX_GPIO_DR 0x14800 @@ -166,6 +170,7 @@ /******************************* PSR Registers *******************************/ /* psr control */ #define WX_PSR_CTL 0x15000 +#define WX_PSR_VM_CTL 0x151B0 /* Header split receive */ #define WX_PSR_CTL_SW_EN BIT(18) #define WX_PSR_CTL_RSC_ACK BIT(17) @@ -186,6 +191,7 @@ /* mcasst/ucast overflow tbl */ #define WX_PSR_MC_TBL(_i) (0x15200 + ((_i) * 4)) #define WX_PSR_UC_TBL(_i) (0x15400 + ((_i) * 4)) +#define WX_PSR_VM_CTL_POOL_MASK GENMASK(12, 7) /* VM L2 contorl */ #define WX_PSR_VM_L2CTL(_i) (0x15600 + ((_i) * 4)) @@ -984,6 +990,7 @@ struct wx_ring_feature { enum wx_ring_f_enum { RING_F_NONE = 0, + RING_F_VMDQ, RING_F_RSS, RING_F_FDIR, RING_F_ARRAY_SIZE /* must be last in enum set */ @@ -1036,7 +1043,26 @@ enum wx_state { WX_STATE_NBITS, /* must be last */ }; +struct vf_data_storage { + struct pci_dev *vfdev; + unsigned char vf_mac_addr[ETH_ALEN]; + bool spoofchk_enabled; + bool link_enable; + bool trusted; + int xcast_mode; +}; + +struct vf_macvlans { + struct list_head mvlist; + int vf; + bool free; + bool is_macvlan; + u8 vf_macvlan[ETH_ALEN]; +}; + enum wx_pf_flags { + WX_FLAG_VMDQ_ENABLED, + WX_FLAG_SRIOV_ENABLED, WX_FLAG_FDIR_CAPABLE, WX_FLAG_FDIR_HASH, WX_FLAG_FDIR_PERFECT, @@ -1136,10 +1162,15 @@ struct wx { u64 hw_csum_rx_good; u64 hw_csum_rx_error; u64 alloc_rx_buff_failed; + unsigned int num_vfs; + struct vf_data_storage *vfinfo; + struct vf_macvlans vf_mvs; + struct vf_macvlans *mv_list; u32 atr_sample_rate; void (*atr)(struct wx_ring *ring, struct wx_tx_buffer *first, u8 ptype); void (*configure_fdir)(struct wx *wx); + int (*setup_tc)(struct net_device *netdev, u8 tc); void (*do_reset)(struct net_device *netdev); }; From patchwork Fri Jul 26 10:02:54 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742556 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgbr2.qq.com (smtpbgbr2.qq.com [54.207.22.56]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 7466A178388 for ; Fri, 26 Jul 2024 10:03:32 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.207.22.56 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988218; cv=none; b=ZJd33MXC+4ip0XYNj5VuKrQLW2a2QimkVZMSZyNMYl19NtJH5RL3RhlB84+Bpq6sOdU5EDCrV2GzVoqCx2lim/O28JnANE+hD6in5UnNwz4Ye3631A8qCDgGElaBNY4SKWhNEd5z4C3mvJoDBuoAYtmwurhY5THAau/IXJ3QTJo= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988218; c=relaxed/simple; bh=bBTsn16Hoxg5OXuu868xJZi9Ae5+3rcBTr/+jy6e2w8=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=N+FLVvff9fMbyE7aJ6A5MQXjTwpl87QHlnsYKiQxlyxsQLcmOmPcUFwgKFfSwgFvRnzm0BmLM6o80ii+oVEOHhzUVQ41jsNYZMJV3Htcn/dD3NdYxTjXRvH36dvvEidqg6gwqdczh+oF7u9l3LhjOOowI23j1hYFvmaID6CEk8w= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.207.22.56 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988206tjm8lpr4 X-QQ-Originating-IP: eGD4GCOtjEvlvTeDPxE7mAxApHAUYF0pFO/uilTJH/Y= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:24 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 504848177141283072 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 03/10] net: libwx: Redesign flow when sriov is enabled Date: Fri, 26 Jul 2024 18:02:54 +0800 Message-ID: <66F650CBAFD187CE+20240726100301.21416-4-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Reallocate queue and int resources when sriov is enabled. Redefine macro VMDQ to make it work in VT mode. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/wx_hw.c | 295 ++++++++++++++++++- drivers/net/ethernet/wangxun/libwx/wx_lib.c | 128 +++++++- drivers/net/ethernet/wangxun/libwx/wx_type.h | 37 ++- 3 files changed, 443 insertions(+), 17 deletions(-) diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c index 1bf9c38e4125..d4a0c36eabb0 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c @@ -10,6 +10,7 @@ #include "wx_type.h" #include "wx_lib.h" +#include "wx_sriov.h" #include "wx_hw.h" static int wx_phy_read_reg_mdi(struct mii_bus *bus, int phy_addr, int devnum, int regnum) @@ -804,11 +805,28 @@ static void wx_sync_mac_table(struct wx *wx) } } +static void wx_full_sync_mac_table(struct wx *wx) +{ + int i; + + for (i = 0; i < wx->mac.num_rar_entries; i++) { + if (wx->mac_table[i].state & WX_MAC_STATE_IN_USE) { + wx_set_rar(wx, i, + wx->mac_table[i].addr, + wx->mac_table[i].pools, + WX_PSR_MAC_SWC_AD_H_AV); + } else { + wx_clear_rar(wx, i); + } + wx->mac_table[i].state &= ~(WX_MAC_STATE_MODIFIED); + } +} + /* this function destroys the first RAR entry */ void wx_mac_set_default_filter(struct wx *wx, u8 *addr) { memcpy(&wx->mac_table[0].addr, addr, ETH_ALEN); - wx->mac_table[0].pools = 1ULL; + wx->mac_table[0].pools = BIT(VMDQ_P(0)); wx->mac_table[0].state = (WX_MAC_STATE_DEFAULT | WX_MAC_STATE_IN_USE); wx_set_rar(wx, 0, wx->mac_table[0].addr, wx->mac_table[0].pools, @@ -1046,6 +1064,36 @@ static void wx_update_mc_addr_list(struct wx *wx, struct net_device *netdev) wx_dbg(wx, "Update mc addr list Complete\n"); } +static void wx_restore_vf_multicasts(struct wx *wx) +{ + u32 i, j, vector_bit, vector_reg; + struct vf_data_storage *vfinfo; + + for (i = 0; i < wx->num_vfs; i++) { + u32 vmolr = rd32(wx, WX_PSR_VM_L2CTL(i)); + + vfinfo = &wx->vfinfo[i]; + for (j = 0; j < vfinfo->num_vf_mc_hashes; j++) { + wx->addr_ctrl.mta_in_use++; + vector_reg = (vfinfo->vf_mc_hashes[j] >> 5) & + GENMASK(6, 0); + vector_bit = vfinfo->vf_mc_hashes[j] & GENMASK(4, 0); + wr32m(wx, WX_PSR_MC_TBL(vector_reg), + BIT(vector_bit), BIT(vector_bit)); + /* errata 5: maintain a copy of the reg table conf */ + wx->mac.mta_shadow[vector_reg] |= BIT(vector_bit); + } + if (vfinfo->num_vf_mc_hashes) + vmolr |= WX_PSR_VM_L2CTL_ROMPE; + else + vmolr &= ~WX_PSR_VM_L2CTL_ROMPE; + wr32(wx, WX_PSR_VM_L2CTL(i), vmolr); + } + + /* Restore any VF macvlans */ + wx_full_sync_mac_table(wx); +} + /** * wx_write_mc_addr_list - write multicast addresses to MTA * @netdev: network interface device structure @@ -1063,6 +1111,9 @@ static int wx_write_mc_addr_list(struct net_device *netdev) wx_update_mc_addr_list(wx, netdev); + if (test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags)) + wx_restore_vf_multicasts(wx); + return netdev_mc_count(netdev); } @@ -1083,7 +1134,7 @@ int wx_set_mac(struct net_device *netdev, void *p) if (retval) return retval; - wx_del_mac_filter(wx, wx->mac.addr, 0); + wx_del_mac_filter(wx, wx->mac.addr, VMDQ_P(0)); eth_hw_addr_set(netdev, addr->sa_data); memcpy(wx->mac.addr, addr->sa_data, netdev->addr_len); @@ -1185,6 +1236,10 @@ static int wx_hpbthresh(struct wx *wx) /* Calculate delay value for device */ dv_id = WX_DV(link, tc); + /* Loopback switch introduces additional latency */ + if (test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags)) + dv_id += WX_B2BT(tc); + /* Delay value is calculated in bit times convert to KB */ kb = WX_BT2KB(dv_id); rx_pba = rd32(wx, WX_RDB_PB_SZ(0)) >> WX_RDB_PB_SZ_SHIFT; @@ -1240,12 +1295,107 @@ static void wx_pbthresh_setup(struct wx *wx) wx->fc.low_water = 0; } +static void wx_set_ethertype_anti_spoofing(struct wx *wx, bool enable, int vf) +{ + u32 pfvfspoof, reg_offset, vf_shift; + + vf_shift = WX_VF_IND_SHIFT(vf); + reg_offset = WX_VF_REG_OFFSET(vf); + + pfvfspoof = rd32(wx, WX_TDM_ETYPE_AS(reg_offset)); + if (enable) + pfvfspoof |= BIT(vf_shift); + else + pfvfspoof &= ~BIT(vf_shift); + wr32(wx, WX_TDM_ETYPE_AS(reg_offset), pfvfspoof); +} + +static int wx_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting) +{ + u32 index = WX_VF_REG_OFFSET(vf), vf_bit = WX_VF_IND_SHIFT(vf); + struct wx *wx = netdev_priv(netdev); + u32 regval; + + if (vf >= wx->num_vfs) + return -EINVAL; + + wx->vfinfo[vf].spoofchk_enabled = setting; + + regval = (setting << vf_bit); + wr32m(wx, WX_TDM_MAC_AS(index), regval | BIT(vf_bit), regval); + + if (wx->vfinfo[vf].vlan_count) + wr32m(wx, WX_TDM_VLAN_AS(index), regval | BIT(vf_bit), regval); + + return 0; +} + +static void wx_configure_virtualization(struct wx *wx) +{ + u16 pool = wx->num_rx_pools; + u32 reg_offset, vf_shift; + u32 i; + + if (!test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags)) + return; + + wr32m(wx, WX_PSR_VM_CTL, + WX_PSR_VM_CTL_POOL_MASK | WX_PSR_VM_CTL_REPLEN, + FIELD_PREP(WX_PSR_VM_CTL_POOL_MASK, VMDQ_P(0)) | + WX_PSR_VM_CTL_REPLEN); + while (pool--) + wr32m(wx, WX_PSR_VM_L2CTL(pool), + WX_PSR_VM_L2CTL_AUPE, WX_PSR_VM_L2CTL_AUPE); + + if (wx->mac.type == wx_mac_sp) { + vf_shift = WX_VF_IND_SHIFT(VMDQ_P(0)); + reg_offset = WX_VF_REG_OFFSET(VMDQ_P(0)); + + /* Enable only the PF pools for Tx/Rx */ + wr32(wx, WX_RDM_VF_RE(reg_offset), GENMASK(31, vf_shift)); + wr32(wx, WX_RDM_VF_RE(reg_offset ^ 1), reg_offset - 1); + wr32(wx, WX_TDM_VF_TE(reg_offset), GENMASK(31, vf_shift)); + wr32(wx, WX_TDM_VF_TE(reg_offset ^ 1), reg_offset - 1); + } else { + vf_shift = BIT(VMDQ_P(0)); + /* Enable only the PF pools for Tx/Rx */ + wr32(wx, WX_RDM_VF_RE(0), vf_shift); + wr32(wx, WX_TDM_VF_TE(0), vf_shift); + } + + /* clear VLAN promisc flag so VFTA will be updated if necessary */ + clear_bit(WX_FLAG_VLAN_PROMISC, wx->flags); + + for (i = 0; i < wx->num_vfs; i++) { + if (!wx->vfinfo[i].spoofchk_enabled) + wx_set_vf_spoofchk(wx->netdev, i, false); + /* enable ethertype anti spoofing if hw supports it */ + wx_set_ethertype_anti_spoofing(wx, true, i); + } +} + static void wx_configure_port(struct wx *wx) { u32 value, i; - value = WX_CFG_PORT_CTL_D_VLAN | WX_CFG_PORT_CTL_QINQ; + if (wx->mac.type == wx_mac_em) { + value = (wx->num_vfs == 0) ? + WX_CFG_PORT_CTL_NUM_VT_NONE : + WX_CFG_PORT_CTL_NUM_VT_8; + } else { + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) { + if (wx->ring_feature[RING_F_RSS].indices == 4) + value = WX_CFG_PORT_CTL_NUM_VT_32; + else + value = WX_CFG_PORT_CTL_NUM_VT_64; + } else { + value = 0; + } + } + + value |= WX_CFG_PORT_CTL_D_VLAN | WX_CFG_PORT_CTL_QINQ; wr32m(wx, WX_CFG_PORT_CTL, + WX_CFG_PORT_CTL_NUM_VT_MASK | WX_CFG_PORT_CTL_D_VLAN | WX_CFG_PORT_CTL_QINQ, value); @@ -1306,6 +1456,83 @@ static void wx_vlan_strip_control(struct wx *wx, bool enable) } } +static void wx_vlan_promisc_enable(struct wx *wx) +{ + u32 vlnctrl, i, vind, bits, reg_idx; + + vlnctrl = rd32(wx, WX_PSR_VLAN_CTL); + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) { + /* we need to keep the VLAN filter on in SRIOV */ + vlnctrl |= WX_PSR_VLAN_CTL_VFE; + wr32(wx, WX_PSR_VLAN_CTL, vlnctrl); + } else { + vlnctrl &= ~WX_PSR_VLAN_CTL_VFE; + wr32(wx, WX_PSR_VLAN_CTL, vlnctrl); + return; + } + /* We are already in VLAN promisc, nothing to do */ + if (test_bit(WX_FLAG_VLAN_PROMISC, wx->flags)) + return; + /* Set flag so we don't redo unnecessary work */ + set_bit(WX_FLAG_VLAN_PROMISC, wx->flags); + /* Add PF to all active pools */ + for (i = WX_PSR_VLAN_SWC_ENTRIES; --i;) { + wr32(wx, WX_PSR_VLAN_SWC_IDX, i); + vind = WX_VF_IND_SHIFT(VMDQ_P(0)); + reg_idx = WX_VF_REG_OFFSET(VMDQ_P(0)); + bits = rd32(wx, WX_PSR_VLAN_SWC_VM(reg_idx)); + bits |= BIT(vind); + wr32(wx, WX_PSR_VLAN_SWC_VM(reg_idx), bits); + } + /* Set all bits in the VLAN filter table array */ + for (i = 0; i < wx->mac.vft_size; i++) + wr32(wx, WX_PSR_VLAN_TBL(i), U32_MAX); +} + +static void wx_scrub_vfta(struct wx *wx) +{ + u32 i, vid, bits, vfta, vind, vlvf, reg_idx; + + for (i = WX_PSR_VLAN_SWC_ENTRIES; --i;) { + wr32(wx, WX_PSR_VLAN_SWC_IDX, i); + vlvf = rd32(wx, WX_PSR_VLAN_SWC_IDX); + /* pull VLAN ID from VLVF */ + vid = vlvf & ~WX_PSR_VLAN_SWC_VIEN; + if (vlvf & WX_PSR_VLAN_SWC_VIEN) { + /* if PF is part of this then continue */ + if (test_bit(vid, wx->active_vlans)) + continue; + } + /* remove PF from the pool */ + vind = WX_VF_IND_SHIFT(VMDQ_P(0)); + reg_idx = WX_VF_REG_OFFSET(VMDQ_P(0)); + bits = rd32(wx, WX_PSR_VLAN_SWC_VM(reg_idx)); + bits &= ~BIT(vind); + wr32(wx, WX_PSR_VLAN_SWC_VM(reg_idx), bits); + } + /* extract values from vft_shadow and write back to VFTA */ + for (i = 0; i < wx->mac.vft_size; i++) { + vfta = wx->mac.vft_shadow[i]; + wr32(wx, WX_PSR_VLAN_TBL(i), vfta); + } +} + +static void wx_vlan_promisc_disable(struct wx *wx) +{ + u32 vlnctrl; + + /* configure vlan filtering */ + vlnctrl = rd32(wx, WX_PSR_VLAN_CTL); + vlnctrl |= WX_PSR_VLAN_CTL_VFE; + wr32(wx, WX_PSR_VLAN_CTL, vlnctrl); + /* We are not in VLAN promisc, nothing to do */ + if (!test_bit(WX_FLAG_VLAN_PROMISC, wx->flags)) + return; + /* Set flag so we don't redo unnecessary work */ + clear_bit(WX_FLAG_VLAN_PROMISC, wx->flags); + wx_scrub_vfta(wx); +} + void wx_set_rx_mode(struct net_device *netdev) { struct wx *wx = netdev_priv(netdev); @@ -1318,7 +1545,7 @@ void wx_set_rx_mode(struct net_device *netdev) /* Check for Promiscuous and All Multicast modes */ fctrl = rd32(wx, WX_PSR_CTL); fctrl &= ~(WX_PSR_CTL_UPE | WX_PSR_CTL_MPE); - vmolr = rd32(wx, WX_PSR_VM_L2CTL(0)); + vmolr = rd32(wx, WX_PSR_VM_L2CTL(VMDQ_P(0))); vmolr &= ~(WX_PSR_VM_L2CTL_UPE | WX_PSR_VM_L2CTL_MPE | WX_PSR_VM_L2CTL_ROPE | @@ -1339,7 +1566,10 @@ void wx_set_rx_mode(struct net_device *netdev) fctrl |= WX_PSR_CTL_UPE | WX_PSR_CTL_MPE; /* pf don't want packets routing to vf, so clear UPE */ vmolr |= WX_PSR_VM_L2CTL_MPE; - vlnctrl &= ~WX_PSR_VLAN_CTL_VFE; + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags) && + test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags)) + vlnctrl |= WX_PSR_VLAN_CTL_VFE; + features &= ~NETIF_F_HW_VLAN_CTAG_FILTER; } if (netdev->flags & IFF_ALLMULTI) { @@ -1362,7 +1592,7 @@ void wx_set_rx_mode(struct net_device *netdev) * sufficient space to store all the addresses then enable * unicast promiscuous mode */ - count = wx_write_uc_addr_list(netdev, 0); + count = wx_write_uc_addr_list(netdev, VMDQ_P(0)); if (count < 0) { vmolr &= ~WX_PSR_VM_L2CTL_ROPE; vmolr |= WX_PSR_VM_L2CTL_UPE; @@ -1380,7 +1610,7 @@ void wx_set_rx_mode(struct net_device *netdev) wr32(wx, WX_PSR_VLAN_CTL, vlnctrl); wr32(wx, WX_PSR_CTL, fctrl); - wr32(wx, WX_PSR_VM_L2CTL(0), vmolr); + wr32(wx, WX_PSR_VM_L2CTL(VMDQ_P(0)), vmolr); if ((features & NETIF_F_HW_VLAN_CTAG_RX) && (features & NETIF_F_HW_VLAN_STAG_RX)) @@ -1388,6 +1618,10 @@ void wx_set_rx_mode(struct net_device *netdev) else wx_vlan_strip_control(wx, false); + if (features & NETIF_F_HW_VLAN_CTAG_FILTER) + wx_vlan_promisc_disable(wx); + else + wx_vlan_promisc_enable(wx); } EXPORT_SYMBOL(wx_set_rx_mode); @@ -1637,6 +1871,13 @@ static void wx_setup_reta(struct wx *wx) u32 random_key_size = WX_RSS_KEY_SIZE / 4; u32 i, j; + if (test_bit(WX_FLAG_SRIOV_ENABLED, wx->flags)) { + if (wx->mac.type == wx_mac_sp) + rss_i = rss_i < 4 ? 4 : rss_i; + else if (wx->mac.type == wx_mac_em) + rss_i = 1; + } + /* Fill out hash function seeds */ for (i = 0; i < random_key_size; i++) wr32(wx, WX_RDB_RSSRK(i), wx->rss_key[i]); @@ -1654,10 +1895,40 @@ static void wx_setup_reta(struct wx *wx) wx_store_reta(wx); } +static void wx_setup_psrtype(struct wx *wx) +{ + int rss_i = wx->ring_feature[RING_F_RSS].indices; + u32 psrtype; + int pool; + + psrtype = WX_RDB_PL_CFG_L4HDR | + WX_RDB_PL_CFG_L3HDR | + WX_RDB_PL_CFG_L2HDR | + WX_RDB_PL_CFG_TUN_OUTL2HDR | + WX_RDB_PL_CFG_TUN_TUNHDR; + + if (wx->mac.type == wx_mac_sp) { + if (rss_i > 3) + psrtype |= FIELD_PREP(GENMASK(31, 29), 2); + else if (rss_i > 1) + psrtype |= FIELD_PREP(GENMASK(31, 29), 1); + + for_each_set_bit(pool, &wx->fwd_bitmask, 32) + wr32(wx, WX_RDB_PL_CFG(VMDQ_P(pool)), psrtype); + } else { + for_each_set_bit(pool, &wx->fwd_bitmask, 8) + wr32(wx, WX_RDB_PL_CFG(VMDQ_P(pool)), psrtype); + } +} + static void wx_setup_mrqc(struct wx *wx) { u32 rss_field = 0; + /* VT, and RSS do not coexist at the same time */ + if (test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) + return; + /* Disable indicating checksum in descriptor, enables RSS hash */ wr32m(wx, WX_PSR_CTL, WX_PSR_CTL_PCSD, WX_PSR_CTL_PCSD); @@ -1687,16 +1958,11 @@ static void wx_setup_mrqc(struct wx *wx) **/ void wx_configure_rx(struct wx *wx) { - u32 psrtype, i; int ret; + u32 i; wx_disable_rx(wx); - - psrtype = WX_RDB_PL_CFG_L4HDR | - WX_RDB_PL_CFG_L3HDR | - WX_RDB_PL_CFG_L2HDR | - WX_RDB_PL_CFG_TUN_TUNHDR; - wr32(wx, WX_RDB_PL_CFG(0), psrtype); + wx_setup_psrtype(wx); /* enable hw crc stripping */ wr32m(wx, WX_RSC_CTL, WX_RSC_CTL_CRC_STRIP, WX_RSC_CTL_CRC_STRIP); @@ -1744,6 +2010,7 @@ void wx_configure(struct wx *wx) { wx_set_rxpba(wx); wx_pbthresh_setup(wx); + wx_configure_virtualization(wx); wx_configure_port(wx); wx_set_rx_mode(wx->netdev); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_lib.c index 1eecba984f3b..8171b603a5ca 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_lib.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_lib.c @@ -1564,6 +1564,65 @@ void wx_napi_disable_all(struct wx *wx) } EXPORT_SYMBOL(wx_napi_disable_all); +static bool wx_set_vmdq_queues(struct wx *wx) +{ + u16 vmdq_i = wx->ring_feature[RING_F_VMDQ].limit; + u16 rss_i = wx->ring_feature[RING_F_RSS].limit; + u16 rss_m = WX_RSS_DISABLED_MASK; + u16 vmdq_m = 0; + + /* only proceed if VMDq is enabled */ + if (!test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) + return false; + /* Add starting offset to total pool count */ + vmdq_i += wx->ring_feature[RING_F_VMDQ].offset; + + if (wx->mac.type == wx_mac_sp) { + /* double check we are limited to maximum pools */ + vmdq_i = min_t(u16, 64, vmdq_i); + + /* 64 pool mode with 2 queues per pool, or + * 16/32/64 pool mode with 1 queue per pool + */ + if (vmdq_i > 32 || rss_i < 4) { + vmdq_m = WX_VMDQ_2Q_MASK; + rss_m = WX_RSS_2Q_MASK; + rss_i = min_t(u16, rss_i, 2); + /* 32 pool mode with 4 queues per pool */ + } else { + vmdq_m = WX_VMDQ_4Q_MASK; + rss_m = WX_RSS_4Q_MASK; + rss_i = 4; + } + } else { + /* double check we are limited to maximum pools */ + vmdq_i = min_t(u16, 8, vmdq_i); + + /* when VMDQ on, disable RSS */ + rss_i = 1; + } + + /* remove the starting offset from the pool count */ + vmdq_i -= wx->ring_feature[RING_F_VMDQ].offset; + + /* save features for later use */ + wx->ring_feature[RING_F_VMDQ].indices = vmdq_i; + wx->ring_feature[RING_F_VMDQ].mask = vmdq_m; + + /* limit RSS based on user input and save for later use */ + wx->ring_feature[RING_F_RSS].indices = rss_i; + wx->ring_feature[RING_F_RSS].mask = rss_m; + + wx->queues_per_pool = rss_i;/*maybe same to num_rx_queues_per_pool*/ + wx->num_rx_pools = vmdq_i; + wx->num_rx_queues_per_pool = rss_i; + + wx->num_rx_queues = vmdq_i * rss_i; + wx->num_tx_queues = vmdq_i * rss_i; + + return true; +} + /** * wx_set_rss_queues: Allocate queues for RSS * @wx: board private structure to initialize @@ -1578,6 +1637,10 @@ static void wx_set_rss_queues(struct wx *wx) /* set mask for 16 queue limit of RSS */ f = &wx->ring_feature[RING_F_RSS]; + if (wx->mac.type == wx_mac_sp) + f->mask = WX_RSS_64Q_MASK; + else + f->mask = WX_RSS_8Q_MASK; f->indices = f->limit; if (!(test_bit(WX_FLAG_FDIR_CAPABLE, wx->flags))) @@ -1610,6 +1673,9 @@ static void wx_set_num_queues(struct wx *wx) wx->num_tx_queues = 1; wx->queues_per_pool = 1; + if (wx_set_vmdq_queues(wx)) + return; + wx_set_rss_queues(wx); } @@ -1690,6 +1756,10 @@ static int wx_set_interrupt_capability(struct wx *wx) if (ret == 0 || (ret == -ENOMEM)) return ret; + /* Disable VMDq support */ + dev_warn(&wx->pdev->dev, "Disabling VMQQ support\n"); + clear_bit(WX_FLAG_VMDQ_ENABLED, wx->flags); + /* Disable RSS */ dev_warn(&wx->pdev->dev, "Disabling RSS support\n"); wx->ring_feature[RING_F_RSS].limit = 1; @@ -1716,6 +1786,49 @@ static int wx_set_interrupt_capability(struct wx *wx) return 0; } +static bool wx_cache_ring_vmdq(struct wx *wx) +{ + struct wx_ring_feature *vmdq = &wx->ring_feature[RING_F_VMDQ]; + struct wx_ring_feature *rss = &wx->ring_feature[RING_F_RSS]; + u16 reg_idx; + int i; + + /* only proceed if VMDq is enabled */ + if (!test_bit(WX_FLAG_VMDQ_ENABLED, wx->flags)) + return false; + + if (wx->mac.type == wx_mac_sp) { + /* start at VMDq register offset for SR-IOV enabled setups */ + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask); + for (i = 0; i < wx->num_rx_queues; i++, reg_idx++) { + /* If we are greater than indices move to next pool */ + if ((reg_idx & ~vmdq->mask) >= rss->indices) + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask); + wx->rx_ring[i]->reg_idx = reg_idx; + } + reg_idx = vmdq->offset * __ALIGN_MASK(1, ~vmdq->mask); + for (i = 0; i < wx->num_tx_queues; i++, reg_idx++) { + /* If we are greater than indices move to next pool */ + if ((reg_idx & rss->mask) >= rss->indices) + reg_idx = __ALIGN_MASK(reg_idx, ~vmdq->mask); + wx->tx_ring[i]->reg_idx = reg_idx; + } + } else { + /* start at VMDq register offset for SR-IOV enabled setups */ + reg_idx = vmdq->offset; + for (i = 0; i < wx->num_rx_queues; i++) + /* If we are greater than indices move to next pool */ + wx->rx_ring[i]->reg_idx = reg_idx + i; + + reg_idx = vmdq->offset; + for (i = 0; i < wx->num_tx_queues; i++) + /* If we are greater than indices move to next pool */ + wx->tx_ring[i]->reg_idx = reg_idx + i; + } + + return true; +} + /** * wx_cache_ring_rss - Descriptor ring to register mapping for RSS * @wx: board private structure to initialize @@ -1727,6 +1840,9 @@ static void wx_cache_ring_rss(struct wx *wx) { u16 i; + if (wx_cache_ring_vmdq(wx)) + return; + for (i = 0; i < wx->num_rx_queues; i++) wx->rx_ring[i]->reg_idx = i; @@ -2119,7 +2235,8 @@ static void wx_set_ivar(struct wx *wx, s8 direction, wr32(wx, WX_PX_MISC_IVAR, ivar); } else { /* tx or rx causes */ - msix_vector += 1; /* offset for queue vectors */ + if (!(wx->mac.type == wx_mac_em && wx->num_vfs == 7)) + msix_vector += 1; /* offset for queue vectors */ msix_vector |= WX_PX_IVAR_ALLOC_VAL; index = ((16 * (queue & 1)) + (8 * direction)); ivar = rd32(wx, WX_PX_IVAR(queue >> 1)); @@ -2164,10 +2281,17 @@ void wx_configure_vectors(struct wx *wx) { struct pci_dev *pdev = wx->pdev; u32 eitrsel = 0; - u16 v_idx; + u16 v_idx, i; if (pdev->msix_enabled) { /* Populate MSIX to EITR Select */ + if (wx->mac.type == wx_mac_sp) { + if (wx->num_vfs >= 32) + eitrsel = BIT(wx->num_vfs % 32) - 1; + } else if (wx->mac.type == wx_mac_em) { + for (i = 0; i < wx->num_vfs; i++) + eitrsel |= BIT(i); + } wr32(wx, WX_PX_ITRSEL, eitrsel); /* use EIAM to auto-mask when MSI-X interrupt is asserted * this saves a register write for every interrupt diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index c4f73c5a366f..ecfe7357d952 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -19,6 +19,7 @@ #define WX_PCIE_MSIX_TBL_SZ_MASK 0x7FF #define WX_PCI_LINK_STATUS 0xB2 #define WX_MAX_PF_MACVLANS 15 +#define WX_MAX_VF_MC_ENTRIES 30 /**************** Global Registers ****************************/ #define WX_VF_REG_OFFSET(_v) ((_v) >> 5) @@ -78,6 +79,7 @@ #define WX_MAC_LXONOFFRXC 0x11E0C /*********************** Receive DMA registers **************************/ +#define WX_RDM_VF_RE(_i) (0x12004 + ((_i) * 4)) #define WX_RDM_DRP_PKT 0x12500 #define WX_RDM_PKT_CNT 0x12504 #define WX_RDM_BYTE_CNT_LSB 0x12508 @@ -92,6 +94,7 @@ #define WX_CFG_TAG_TPID(_i) (0x14430 + ((_i) * 4)) #define WX_CFG_PORT_CTL_NUM_VT_MASK GENMASK(13, 12) /* number of TVs */ +#define WX_CFG_PORT_CTL_NUM_VT_NONE 0 #define WX_CFG_PORT_CTL_NUM_VT_8 FIELD_PREP(GENMASK(13, 12), 1) #define WX_CFG_PORT_CTL_NUM_VT_32 FIELD_PREP(GENMASK(13, 12), 2) #define WX_CFG_PORT_CTL_NUM_VT_64 FIELD_PREP(GENMASK(13, 12), 3) @@ -117,6 +120,10 @@ /*********************** Transmit DMA registers **************************/ /* transmit global control */ #define WX_TDM_CTL 0x18000 +#define WX_TDM_VF_TE(_i) (0x18004 + ((_i) * 4)) +#define WX_TDM_MAC_AS(_i) (0x18060 + ((_i) * 4)) +#define WX_TDM_VLAN_AS(_i) (0x18070 + ((_i) * 4)) + /* TDM CTL BIT */ #define WX_TDM_CTL_TE BIT(0) /* Transmit Enable */ #define WX_TDM_PB_THRE(_i) (0x18020 + ((_i) * 4)) @@ -191,6 +198,7 @@ /* mcasst/ucast overflow tbl */ #define WX_PSR_MC_TBL(_i) (0x15200 + ((_i) * 4)) #define WX_PSR_UC_TBL(_i) (0x15400 + ((_i) * 4)) +#define WX_PSR_VM_CTL_REPLEN BIT(30) /* replication enabled */ #define WX_PSR_VM_CTL_POOL_MASK GENMASK(12, 7) /* VM L2 contorl */ @@ -235,6 +243,7 @@ #define WX_PSR_VLAN_SWC 0x16220 #define WX_PSR_VLAN_SWC_VM_L 0x16224 #define WX_PSR_VLAN_SWC_VM_H 0x16228 +#define WX_PSR_VLAN_SWC_VM(_i) (0x16224 + ((_i) * 4)) #define WX_PSR_VLAN_SWC_IDX 0x16230 /* 64 vlan entries */ /* VLAN pool filtering masks */ #define WX_PSR_VLAN_SWC_VIEN BIT(31) /* filter is valid */ @@ -249,6 +258,10 @@ #define WX_RSC_ST 0x17004 #define WX_RSC_ST_RSEC_RDY BIT(0) +/*********************** Transmit DMA registers **************************/ +/* transmit global control */ +#define WX_TDM_ETYPE_AS(_i) (0x18058 + ((_i) * 4)) + /****************************** TDB ******************************************/ #define WX_TDB_PB_SZ(_i) (0x1CC00 + ((_i) * 4)) #define WX_TXPKT_SIZE_MAX 0xA /* Max Tx Packet size */ @@ -376,6 +389,15 @@ enum WX_MSCA_CMD_value { /* Number of 80 microseconds we wait for PCI Express master disable */ #define WX_PCI_MASTER_DISABLE_TIMEOUT 80000 +#define WX_RSS_64Q_MASK 0x3F +#define WX_RSS_8Q_MASK 0x7 +#define WX_RSS_4Q_MASK 0x3 +#define WX_RSS_2Q_MASK 0x1 +#define WX_RSS_DISABLED_MASK 0x0 + +#define WX_VMDQ_4Q_MASK 0x7C +#define WX_VMDQ_2Q_MASK 0x7E + /****************** Manageablility Host Interface defines ********************/ #define WX_HI_MAX_BLOCK_BYTE_LENGTH 256 /* Num of bytes in range */ #define WX_HI_COMMAND_TIMEOUT 1000 /* Process HI command limit */ @@ -440,7 +462,12 @@ enum WX_MSCA_CMD_value { #define WX_REQ_TX_DESCRIPTOR_MULTIPLE 8 #define WX_MAX_JUMBO_FRAME_SIZE 9432 /* max payload 9414 */ -#define VMDQ_P(p) p +/* must account for pools assigned to VFs. */ +#ifdef CONFIG_PCI_IOV +#define VMDQ_P(p) ((p) + wx->ring_feature[RING_F_VMDQ].offset) +#else +#define VMDQ_P(p) (p) +#endif /* Supported Rx Buffer Sizes */ #define WX_RXBUFFER_256 256 /* Used for skb receive header */ @@ -1050,6 +1077,10 @@ struct vf_data_storage { bool link_enable; bool trusted; int xcast_mode; + + u16 vf_mc_hashes[WX_MAX_VF_MC_ENTRIES]; + u16 num_vf_mc_hashes; + u16 vlan_count; }; struct vf_macvlans { @@ -1062,6 +1093,7 @@ struct vf_macvlans { enum wx_pf_flags { WX_FLAG_VMDQ_ENABLED, + WX_FLAG_VLAN_PROMISC, WX_FLAG_SRIOV_ENABLED, WX_FLAG_FDIR_CAPABLE, WX_FLAG_FDIR_HASH, @@ -1133,6 +1165,8 @@ struct wx { struct wx_ring *tx_ring[64] ____cacheline_aligned_in_smp; struct wx_ring *rx_ring[64]; struct wx_q_vector *q_vector[64]; + int num_rx_pools; + int num_rx_queues_per_pool; unsigned int queues_per_pool; struct msix_entry *msix_q_entries; @@ -1166,6 +1200,7 @@ struct wx { struct vf_data_storage *vfinfo; struct vf_macvlans vf_mvs; struct vf_macvlans *mv_list; + unsigned long fwd_bitmask; u32 atr_sample_rate; void (*atr)(struct wx_ring *ring, struct wx_tx_buffer *first, u8 ptype); From patchwork Fri Jul 26 10:02:55 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742561 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbguseast1.qq.com (smtpbguseast1.qq.com [54.204.34.129]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 618E5178CF2 for ; Fri, 26 Jul 2024 10:03:40 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.204.34.129 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988225; cv=none; b=Ih++NM5doIl85mzEqS5rs/ZKV+Ie5DAG0RMTYxjcWufjSqMyB552MCSUhz9/aDaCTEr45WFa2oRigOpeBJJkkU4Z1qoTiiGfi/Ws4lq/BoexTsrWzB6iAGXXRYDvmuPyKChXbd/4ur4lTtpWvvXmFsOYX+GeoMTyMrJGbV1Cjwg= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988225; c=relaxed/simple; bh=/HOkI3AASxaYSqrGl1Ms+xJvkmPVqsmBYZlxBVz8/8g=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=Oyubk+yKi4xP0D8pkmLBioZMEhVymQ8bJ5rgjCYPRrR6ID7EblABSczn3Fn3BXGWE99CR7GttR72oKsBM7jeC5U10VN7gOtvJTaQBF1JEpz+d83Wudlfp6xYsvL37OIdStaF7mq53EkJsjvAj4TJnfesidwrgkb9szTNSjA6PCs= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.204.34.129 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988208tryrlav2 X-QQ-Originating-IP: ogijMBj0HoStckR4CvzrfdLN6tUb3gjtfYoWiEAW3S4= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:27 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 1 X-BIZMAIL-ID: 886751299633612328 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 04/10] net: libwx: Add msg task func Date: Fri, 26 Jul 2024 18:02:55 +0800 Message-ID: <135920C1A6A3E085+20240726100301.21416-5-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Implement wx_msg_task which is used to process mailbox messages sent by vf. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/wx_hw.c | 8 +- drivers/net/ethernet/wangxun/libwx/wx_hw.h | 4 + drivers/net/ethernet/wangxun/libwx/wx_mbx.h | 50 ++ drivers/net/ethernet/wangxun/libwx/wx_sriov.c | 672 ++++++++++++++++++ drivers/net/ethernet/wangxun/libwx/wx_sriov.h | 1 + drivers/net/ethernet/wangxun/libwx/wx_type.h | 17 + 6 files changed, 748 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c index d4a0c36eabb0..21eafd1a9904 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c @@ -851,7 +851,7 @@ void wx_flush_sw_mac_table(struct wx *wx) } EXPORT_SYMBOL(wx_flush_sw_mac_table); -static int wx_add_mac_filter(struct wx *wx, u8 *addr, u16 pool) +int wx_add_mac_filter(struct wx *wx, u8 *addr, u16 pool) { u32 i; @@ -882,7 +882,7 @@ static int wx_add_mac_filter(struct wx *wx, u8 *addr, u16 pool) return -ENOMEM; } -static int wx_del_mac_filter(struct wx *wx, u8 *addr, u16 pool) +int wx_del_mac_filter(struct wx *wx, u8 *addr, u16 pool) { u32 i; @@ -1310,7 +1310,7 @@ static void wx_set_ethertype_anti_spoofing(struct wx *wx, bool enable, int vf) wr32(wx, WX_TDM_ETYPE_AS(reg_offset), pfvfspoof); } -static int wx_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting) +int wx_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting) { u32 index = WX_VF_REG_OFFSET(vf), vf_bit = WX_VF_IND_SHIFT(vf); struct wx *wx = netdev_priv(netdev); @@ -2369,7 +2369,7 @@ static int wx_set_vlvf(struct wx *wx, u32 vlan, u32 vind, bool vlan_on, * * Turn on/off specified VLAN in the VLAN filter table. **/ -static int wx_set_vfta(struct wx *wx, u32 vlan, u32 vind, bool vlan_on) +int wx_set_vfta(struct wx *wx, u32 vlan, u32 vind, bool vlan_on) { u32 bitindex, vfta, targetbit; bool vfta_changed = false; diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.h b/drivers/net/ethernet/wangxun/libwx/wx_hw.h index 11fb33349482..224f89218bfb 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_hw.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.h @@ -25,9 +25,12 @@ void wx_init_eeprom_params(struct wx *wx); void wx_get_mac_addr(struct wx *wx, u8 *mac_addr); void wx_init_rx_addrs(struct wx *wx); void wx_mac_set_default_filter(struct wx *wx, u8 *addr); +int wx_add_mac_filter(struct wx *wx, u8 *addr, u16 pool); +int wx_del_mac_filter(struct wx *wx, u8 *addr, u16 pool); void wx_flush_sw_mac_table(struct wx *wx); int wx_set_mac(struct net_device *netdev, void *p); void wx_disable_rx(struct wx *wx); +int wx_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting); int wx_disable_sec_rx_path(struct wx *wx); void wx_enable_sec_rx_path(struct wx *wx); void wx_set_rx_mode(struct net_device *netdev); @@ -41,6 +44,7 @@ int wx_stop_adapter(struct wx *wx); void wx_reset_misc(struct wx *wx); int wx_get_pcie_msix_counts(struct wx *wx, u16 *msix_count, u16 max_msix_count); int wx_sw_init(struct wx *wx); +int wx_set_vfta(struct wx *wx, u32 vlan, u32 vind, bool vlan_on); int wx_vlan_rx_add_vid(struct net_device *netdev, __be16 proto, u16 vid); int wx_vlan_rx_kill_vid(struct net_device *netdev, __be16 proto, u16 vid); int wx_fc_enable(struct wx *wx, bool tx_pause, bool rx_pause); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h index 3c70654a8b14..725f29c9cb11 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h @@ -21,10 +21,60 @@ #define WX_MBVFICR_VFREQ_MASK GENMASK(15, 0) #define WX_MBVFICR_VFACK_MASK GENMASK(31, 16) +#define WX_VT_MSGTYPE_ACK BIT(31) +#define WX_VT_MSGTYPE_NACK BIT(30) +#define WX_VT_MSGTYPE_CTS BIT(29) +#define WX_VT_MSGINFO_SHIFT 16 #define WX_VT_MSGINFO_MASK GENMASK(23, 16) +enum wx_pfvf_api_rev { + wx_mbox_api_null, + wx_mbox_api_10, /* API version 1.0, linux/freebsd VF driver */ + wx_mbox_api_11, /* API version 1.1, linux/freebsd VF driver */ + wx_mbox_api_12, /* API version 1.2, linux/freebsd VF driver */ + wx_mbox_api_13, /* API version 1.3, linux/freebsd VF driver */ + wx_mbox_api_20, /* API version 2.0, solaris Phase1 VF driver */ + wx_mbox_api_unknown, /* indicates that API version is not known */ +}; + +/* mailbox API, legacy requests */ +#define WX_VF_RESET 0x01 /* VF requests reset */ +#define WX_VF_SET_MAC_ADDR 0x02 /* VF requests PF to set MAC addr */ +#define WX_VF_SET_MULTICAST 0x03 /* VF requests PF to set MC addr */ +#define WX_VF_SET_VLAN 0x04 /* VF requests PF to set VLAN */ + +/* mailbox API, version 1.0 VF requests */ +#define WX_VF_SET_LPE 0x05 /* VF requests PF to set VMOLR.LPE */ +#define WX_VF_SET_MACVLAN 0x06 /* VF requests PF unicast filter */ +#define WX_VF_API_NEGOTIATE 0x08 /* negotiate API version */ + +/* mailbox API, version 1.1 VF requests */ +#define WX_VF_GET_QUEUES 0x09 /* get queue configuration */ + +/* mailbox API, version 1.2 VF requests */ +#define WX_VF_GET_RETA 0x0a /* VF request for RETA */ +#define WX_VF_GET_RSS_KEY 0x0b /* get RSS key */ +#define WX_VF_UPDATE_XCAST_MODE 0x0c +#define WX_VF_GET_LINK_STATE 0x10 /* get vf link state */ +#define WX_VF_GET_FW_VERSION 0x11 /* get fw version */ +#define WX_VF_BACKUP 0x8001 /* VF requests backup */ + +#define WX_PF_CONTROL_MSG BIT(8) /* PF control message */ +#define WX_PF_NOFITY_VF_LINK_STATUS 0x1 +#define WX_PF_NOFITY_VF_NET_NOT_RUNNING BIT(31) + +#define WX_VF_TX_QUEUES 1 /* number of Tx queues supported */ +#define WX_VF_RX_QUEUES 2 /* number of Rx queues supported */ +#define WX_VF_TRANS_VLAN 3 /* Indication of port vlan */ +#define WX_VF_DEF_QUEUE 4 /* Default queue offset */ + +#define WX_VF_PERMADDR_MSG_LEN 4 + enum wxvf_xcast_modes { WXVF_XCAST_MODE_NONE = 0, + WXVF_XCAST_MODE_MULTI, + WXVF_XCAST_MODE_ALLMULTI, + WXVF_XCAST_MODE_PROMISC, }; int wx_write_mbx_pf(struct wx *wx, u32 *msg, u16 size, u16 vf); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c index 6b0a150abd76..f97743a292b5 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c @@ -5,6 +5,7 @@ #include #include "wx_type.h" +#include "wx_hw.h" #include "wx_mbx.h" #include "wx_sriov.h" @@ -208,3 +209,674 @@ int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs) return num_vfs; } EXPORT_SYMBOL(wx_pci_sriov_configure); + +static int wx_set_vf_mac(struct wx *wx, u16 vf, const u8 *mac_addr) +{ + u8 hw_addr[ETH_ALEN]; + int ret = 0; + + ether_addr_copy(hw_addr, mac_addr); + wx_del_mac_filter(wx, wx->vfinfo[vf].vf_mac_addr, vf); + ret = wx_add_mac_filter(wx, hw_addr, vf); + if (ret >= 0) + ether_addr_copy(wx->vfinfo[vf].vf_mac_addr, mac_addr); + else + eth_zero_addr(wx->vfinfo[vf].vf_mac_addr); + + return ret; +} + +static void wx_set_vmolr(struct wx *wx, u16 vf, bool aupe) +{ + u32 vmolr = rd32(wx, WX_PSR_VM_L2CTL(vf)); + + vmolr |= WX_PSR_VM_L2CTL_BAM; + if (aupe) + vmolr |= WX_PSR_VM_L2CTL_AUPE; + else + vmolr &= ~WX_PSR_VM_L2CTL_AUPE; + wr32(wx, WX_PSR_VM_L2CTL(vf), vmolr); +} + +static void wx_set_vmvir(struct wx *wx, u16 vid, u16 qos, u16 vf) +{ + u32 vmvir = vid | (qos << VLAN_PRIO_SHIFT) | + WX_TDM_VLAN_INS_VLANA_DEFAULT; + + wr32(wx, WX_TDM_VLAN_INS(vf), vmvir); +} + +static int wx_set_vf_vlan(struct wx *wx, int add, int vid, u16 vf) +{ + if (!vid && !add) + return 0; + + return wx_set_vfta(wx, vid, vf, (bool)add); +} + +static void wx_set_vlan_anti_spoofing(struct wx *wx, bool enable, int vf) +{ + u32 index = WX_VF_REG_OFFSET(vf), vf_bit = WX_VF_IND_SHIFT(vf); + u32 pfvfspoof; + + pfvfspoof = rd32(wx, WX_TDM_VLAN_AS(index)); + if (enable) + pfvfspoof |= BIT(vf_bit); + else + pfvfspoof &= ~BIT(vf_bit); + wr32(wx, WX_TDM_VLAN_AS(index), pfvfspoof); +} + +static void wx_write_qde(struct wx *wx, u32 vf, u32 qde) +{ + struct wx_ring_feature *vmdq = &wx->ring_feature[RING_F_VMDQ]; + u32 q_per_pool = __ALIGN_MASK(1, ~vmdq->mask); + u32 reg = 0, n = vf * q_per_pool / 32; + u32 i = vf * q_per_pool; + + reg = rd32(wx, WX_RDM_PF_QDE(n)); + for (i = (vf * q_per_pool - n * 32); + i < ((vf + 1) * q_per_pool - n * 32); + i++) { + if (qde == 1) + reg |= qde << i; + else + reg &= qde << i; + } + + wr32(wx, WX_RDM_PF_QDE(n), reg); +} + +static void wx_clear_vmvir(struct wx *wx, u32 vf) +{ + wr32(wx, WX_TDM_VLAN_INS(vf), 0); +} + +static void wx_set_vf_rx_tx(struct wx *wx, int vf) +{ + u32 index = WX_VF_REG_OFFSET(vf), vf_bit = WX_VF_IND_SHIFT(vf); + u32 reg_cur_tx, reg_cur_rx, reg_req_tx, reg_req_rx; + + reg_cur_tx = rd32(wx, WX_TDM_VF_TE(index)); + reg_cur_rx = rd32(wx, WX_RDM_VF_RE(index)); + + if (wx->vfinfo[vf].link_enable) { + reg_req_tx = reg_cur_tx | BIT(vf_bit); + reg_req_rx = reg_cur_rx | BIT(vf_bit); + /* Enable particular VF */ + if (reg_cur_tx != reg_req_tx) + wr32(wx, WX_TDM_VF_TE(index), reg_req_tx); + if (reg_cur_rx != reg_req_rx) + wr32(wx, WX_RDM_VF_RE(index), reg_req_rx); + } else { + reg_req_tx = BIT(vf_bit); + reg_req_rx = BIT(vf_bit); + /* Disable particular VF */ + if (reg_cur_tx & reg_req_tx) + wr32(wx, WX_TDM_VFTE_CLR(index), reg_req_tx); + if (reg_cur_rx & reg_req_rx) + wr32(wx, WX_RDM_VFRE_CLR(index), reg_req_rx); + } +} + +static int wx_get_vf_queues(struct wx *wx, u32 *msgbuf, u32 vf) +{ + struct wx_ring_feature *vmdq = &wx->ring_feature[RING_F_VMDQ]; + unsigned int default_tc = 0; + + switch (wx->vfinfo[vf].vf_api) { + case wx_mbox_api_11 ... wx_mbox_api_20: + break; + default: + return -EINVAL; + } + + msgbuf[WX_VF_TX_QUEUES] = __ALIGN_MASK(1, ~vmdq->mask); + msgbuf[WX_VF_RX_QUEUES] = __ALIGN_MASK(1, ~vmdq->mask); + + if (wx->vfinfo[vf].pf_vlan || wx->vfinfo[vf].pf_qos) + msgbuf[WX_VF_TRANS_VLAN] = 1; + else + msgbuf[WX_VF_TRANS_VLAN] = 0; + + /* notify VF of default queue */ + msgbuf[WX_VF_DEF_QUEUE] = default_tc; + + return 0; +} + +static void wx_vf_reset_event(struct wx *wx, u16 vf) +{ + struct vf_data_storage *vfinfo = &wx->vfinfo[vf]; + u8 num_tcs = netdev_get_num_tc(wx->netdev); + + /* add PF assigned VLAN */ + wx_set_vf_vlan(wx, true, vfinfo->pf_vlan, vf); + + /* reset offloads to defaults */ + wx_set_vmolr(wx, vf, !vfinfo->pf_vlan); + + /* set outgoing tags for VFs */ + if (!vfinfo->pf_vlan && !vfinfo->pf_qos && !num_tcs) { + wx_clear_vmvir(wx, vf); + } else { + if (vfinfo->pf_qos || !num_tcs) + wx_set_vmvir(wx, vfinfo->pf_vlan, + vfinfo->pf_qos, vf); + else + wx_set_vmvir(wx, vfinfo->pf_vlan, + wx->default_up, vf); + } + + /* reset multicast table array for vf */ + wx->vfinfo[vf].num_vf_mc_hashes = 0; + + /* Flush and reset the mta with the new values */ + wx_set_rx_mode(wx->netdev); + + wx_del_mac_filter(wx, wx->vfinfo[vf].vf_mac_addr, vf); + + wx->vfinfo[vf].vf_api = wx_mbox_api_10; +} + +static void wx_vf_reset_msg(struct wx *wx, u16 vf) +{ + const u8 *vf_mac = wx->vfinfo[vf].vf_mac_addr; + struct net_device *dev = wx->netdev; + u32 msgbuf[5] = {0, 0, 0, 0, 0}; + u8 *addr = (u8 *)(&msgbuf[1]); + u32 reg = 0, index, vf_bit; + int pf_max_frame; + + /* reset the filters for the device */ + wx_vf_reset_event(wx, vf); + + /* set vf mac address */ + if (!is_zero_ether_addr(vf_mac)) + wx_set_vf_mac(wx, vf, vf_mac); + + index = WX_VF_REG_OFFSET(vf); + vf_bit = WX_VF_IND_SHIFT(vf); + + /* force drop enable for all VF Rx queues */ + wx_write_qde(wx, vf, 1); + + /* set transmit and receive for vf */ + wx_set_vf_rx_tx(wx, vf); + + pf_max_frame = dev->mtu + ETH_HLEN; + + if (pf_max_frame > ETH_FRAME_LEN) + reg = BIT(vf_bit); + wr32(wx, WX_RDM_VFRE_CLR(index), reg); + + /* enable VF mailbox for further messages */ + wx->vfinfo[vf].clear_to_send = true; + + /* reply to reset with ack and vf mac address */ + msgbuf[0] = WX_VF_RESET; + if (!is_zero_ether_addr(vf_mac)) { + msgbuf[0] |= WX_VT_MSGTYPE_ACK; + memcpy(addr, vf_mac, ETH_ALEN); + } else { + msgbuf[0] |= WX_VT_MSGTYPE_NACK; + wx_err(wx, "VF %d has no MAC address assigned", vf); + } + + msgbuf[3] = wx->mac.mc_filter_type; + wx_write_mbx_pf(wx, msgbuf, WX_VF_PERMADDR_MSG_LEN, vf); +} + +static int wx_set_vf_mac_addr(struct wx *wx, u32 *msgbuf, u16 vf) +{ + const u8 *new_mac = ((u8 *)(&msgbuf[1])); + int ret; + + if (!is_valid_ether_addr(new_mac)) { + wx_err(wx, "VF %d attempted to set invalid mac\n", vf); + return -EINVAL; + } + + if (wx->vfinfo[vf].pf_set_mac && + memcmp(wx->vfinfo[vf].vf_mac_addr, new_mac, ETH_ALEN)) { + wx_err(wx, + "VF %d attempt to set a MAC but it already had a MAC.", + vf); + return -EBUSY; + } + + ret = wx_set_vf_mac(wx, vf, new_mac); + if (ret < 0) + return ret; + + return 0; +} + +static void wx_set_vf_multicasts(struct wx *wx, u32 *msgbuf, u32 vf) +{ + struct vf_data_storage *vfinfo = &wx->vfinfo[vf]; + u16 entries = (msgbuf[0] & WX_VT_MSGINFO_MASK) + >> WX_VT_MSGINFO_SHIFT; + u32 vmolr = rd32(wx, WX_PSR_VM_L2CTL(vf)); + u32 vector_bit, vector_reg, mta_reg, i; + u16 *hash_list = (u16 *)&msgbuf[1]; + + /* only so many hash values supported */ + entries = min_t(u16, entries, WX_MAX_VF_MC_ENTRIES); + vfinfo->num_vf_mc_hashes = entries; + + for (i = 0; i < entries; i++) + vfinfo->vf_mc_hashes[i] = hash_list[i]; + + for (i = 0; i < vfinfo->num_vf_mc_hashes; i++) { + vector_reg = (vfinfo->vf_mc_hashes[i] >> 5) & 0x7F; + vector_bit = vfinfo->vf_mc_hashes[i] & 0x1F; + mta_reg = wx->mac.mta_shadow[vector_reg]; + mta_reg |= (1 << vector_bit); + wx->mac.mta_shadow[vector_reg] = mta_reg; + wr32(wx, WX_PSR_MC_TBL(vector_reg), mta_reg); + } + vmolr |= WX_PSR_VM_L2CTL_ROMPE; + wr32(wx, WX_PSR_VM_L2CTL(vf), vmolr); +} + +static void wx_set_vf_lpe(struct wx *wx, u32 max_frame, u32 vf) +{ + u32 index, vf_bit, vfre; + u32 max_frs, reg_val; + + /* determine VF receive enable location */ + index = WX_VF_REG_OFFSET(vf); + vf_bit = WX_VF_IND_SHIFT(vf); + + vfre = rd32(wx, WX_RDM_VF_RE(index)); + vfre |= BIT(vf_bit); + wr32(wx, WX_RDM_VF_RE(index), vfre); + + /* pull current max frame size from hardware */ + max_frs = DIV_ROUND_UP(max_frame, 1024); + reg_val = rd32(wx, WX_MAC_WDG_TIMEOUT) & WX_MAC_WDG_TIMEOUT_WTO_MASK; + if (max_frs > (reg_val + WX_MAC_WDG_TIMEOUT_WTO_DELTA)) + wr32(wx, WX_MAC_WDG_TIMEOUT, + max_frs - WX_MAC_WDG_TIMEOUT_WTO_DELTA); +} + +static int wx_find_vlvf_entry(struct wx *wx, u32 vlan) +{ + int regindex; + u32 vlvf; + + /* short cut the special case */ + if (vlan == 0) + return 0; + + /* Search for the vlan id in the VLVF entries */ + for (regindex = 1; regindex < WX_PSR_VLAN_SWC_ENTRIES; regindex++) { + wr32(wx, WX_PSR_VLAN_SWC_IDX, regindex); + vlvf = rd32(wx, WX_PSR_VLAN_SWC); + if ((vlvf & VLAN_VID_MASK) == vlan) + break; + } + + /* Return a negative value if not found */ + if (regindex >= WX_PSR_VLAN_SWC_ENTRIES) + regindex = -EINVAL; + + return regindex; +} + +static int wx_set_vf_macvlan(struct wx *wx, + u16 vf, int index, unsigned char *mac_addr) +{ + struct vf_macvlans *entry; + struct list_head *pos; + int retval = 0; + + if (index <= 1) { + list_for_each(pos, &wx->vf_mvs.mvlist) { + entry = list_entry(pos, struct vf_macvlans, mvlist); + if (entry->vf == vf) { + entry->vf = -1; + entry->free = true; + entry->is_macvlan = false; + wx_del_mac_filter(wx, entry->vf_macvlan, vf); + } + } + } + + if (!index) + return 0; + + entry = NULL; + list_for_each(pos, &wx->vf_mvs.mvlist) { + entry = list_entry(pos, struct vf_macvlans, mvlist); + if (entry->free) + break; + } + + if (!entry || !entry->free) + return -ENOSPC; + + retval = wx_add_mac_filter(wx, mac_addr, vf); + if (retval >= 0) { + entry->free = false; + entry->is_macvlan = true; + entry->vf = vf; + memcpy(entry->vf_macvlan, mac_addr, ETH_ALEN); + } + + return retval; +} + +static int wx_set_vf_vlan_msg(struct wx *wx, u32 *msgbuf, u16 vf) +{ + int add = (msgbuf[0] & WX_VT_MSGINFO_MASK) >> WX_VT_MSGINFO_SHIFT; + int vid = (msgbuf[1] & WX_PSR_VLAN_SWC_VLANID_MASK); + int ret; + + if (add) + wx->vfinfo[vf].vlan_count++; + else if (wx->vfinfo[vf].vlan_count) + wx->vfinfo[vf].vlan_count--; + + /* in case of promiscuous mode any VLAN filter set for a VF must + * also have the PF pool added to it. + */ + if (add && wx->netdev->flags & IFF_PROMISC) + wx_set_vf_vlan(wx, add, vid, VMDQ_P(0)); + + ret = wx_set_vf_vlan(wx, add, vid, vf); + if (!ret && wx->vfinfo[vf].spoofchk_enabled) + wx_set_vlan_anti_spoofing(wx, true, vf); + + /* Go through all the checks to see if the VLAN filter should + * be wiped completely. + */ + if (!add && wx->netdev->flags & IFF_PROMISC) { + u32 bits = 0, vlvf; + int reg_ndx; + + reg_ndx = wx_find_vlvf_entry(wx, vid); + if (reg_ndx < 0) + return -ENOSPC; + wr32(wx, WX_PSR_VLAN_SWC_IDX, reg_ndx); + vlvf = rd32(wx, WX_PSR_VLAN_SWC); + /* See if any other pools are set for this VLAN filter + * entry other than the PF. + */ + if (VMDQ_P(0) < 32) { + bits = rd32(wx, WX_PSR_VLAN_SWC_VM_L); + bits &= ~BIT(VMDQ_P(0)); + if (wx->mac.type == wx_mac_sp) + bits |= rd32(wx, WX_PSR_VLAN_SWC_VM_H); + } else { + if (wx->mac.type == wx_mac_sp) + bits = rd32(wx, WX_PSR_VLAN_SWC_VM_H); + bits &= ~BIT(VMDQ_P(0) % 32); + bits |= rd32(wx, WX_PSR_VLAN_SWC_VM_L); + } + /* If the filter was removed then ensure PF pool bit + * is cleared if the PF only added itself to the pool + * because the PF is in promiscuous mode. + */ + if ((vlvf & VLAN_VID_MASK) == vid && !bits) + wx_set_vf_vlan(wx, add, vid, VMDQ_P(0)); + } + + return 0; +} + +static int wx_set_vf_macvlan_msg(struct wx *wx, u32 *msgbuf, u16 vf) +{ + int index = (msgbuf[0] & WX_VT_MSGINFO_MASK) >> + WX_VT_MSGINFO_SHIFT; + u8 *new_mac = ((u8 *)(&msgbuf[1])); + int err; + + if (wx->vfinfo[vf].pf_set_mac && index > 0) { + wx_err(wx, "VF %d request MACVLAN filter but is denied\n", vf); + return -EINVAL; + } + + /* An non-zero index indicates the VF is setting a filter */ + if (index) { + if (!is_valid_ether_addr(new_mac)) { + wx_err(wx, "VF %d attempted to set invalid mac\n", vf); + return -EINVAL; + } + /* If the VF is allowed to set MAC filters then turn off + * anti-spoofing to avoid false positives. + */ + if (wx->vfinfo[vf].spoofchk_enabled) + wx_set_vf_spoofchk(wx->netdev, vf, false); + } + + err = wx_set_vf_macvlan(wx, vf, index, new_mac); + if (err == -ENOSPC) + wx_err(wx, + "VF %d request MACVLAN filter but there is no space\n", + vf); + if (err < 0) + return err; + + return 0; +} + +static int wx_negotiate_vf_api(struct wx *wx, u32 *msgbuf, u32 vf) +{ + int api = msgbuf[1]; + + switch (api) { + case wx_mbox_api_10 ... wx_mbox_api_13: + wx->vfinfo[vf].vf_api = api; + return 0; + default: + wx_err(wx, "VF %d requested invalid api version %u\n", vf, api); + return -EINVAL; + } +} + +static int wx_get_vf_link_state(struct wx *wx, u32 *msgbuf, u32 vf) +{ + /* verify the PF is supporting the correct API */ + switch (wx->vfinfo[vf].vf_api) { + case wx_mbox_api_12 ... wx_mbox_api_13: + break; + default: + return -EOPNOTSUPP; + } + + msgbuf[1] = wx->vfinfo[vf].link_enable; + + return 0; +} + +static int wx_get_fw_version(struct wx *wx, u32 *msgbuf, u32 vf) +{ + unsigned long fw_version = 0ULL; + int ret = 0; + + /* verify the PF is supporting the correct API */ + switch (wx->vfinfo[vf].vf_api) { + case wx_mbox_api_12 ... wx_mbox_api_13: + break; + default: + return -EOPNOTSUPP; + } + + ret = kstrtoul(wx->eeprom_id, 16, &fw_version); + if (ret) + return -EOPNOTSUPP; + msgbuf[1] = fw_version; + + return 0; +} + +static int wx_update_vf_xcast_mode(struct wx *wx, u32 *msgbuf, u32 vf) +{ + int xcast_mode = msgbuf[1]; + u32 vmolr, disable, enable; + + /* verify the PF is supporting the correct APIs */ + switch (wx->vfinfo[vf].vf_api) { + case wx_mbox_api_12: + /* promisc introduced in 1.3 version */ + if (xcast_mode == WXVF_XCAST_MODE_PROMISC) + return -EOPNOTSUPP; + fallthrough; + case wx_mbox_api_13: + break; + default: + return -EOPNOTSUPP; + } + + if (wx->vfinfo[vf].xcast_mode == xcast_mode) + return 0; + + switch (xcast_mode) { + case WXVF_XCAST_MODE_NONE: + disable = WX_PSR_VM_L2CTL_BAM | WX_PSR_VM_L2CTL_ROMPE | + WX_PSR_VM_L2CTL_MPE | WX_PSR_VM_L2CTL_UPE | + WX_PSR_VM_L2CTL_VPE; + enable = 0; + break; + case WXVF_XCAST_MODE_MULTI: + disable = WX_PSR_VM_L2CTL_MPE | WX_PSR_VM_L2CTL_UPE | + WX_PSR_VM_L2CTL_VPE; + enable = WX_PSR_VM_L2CTL_BAM | WX_PSR_VM_L2CTL_ROMPE; + break; + case WXVF_XCAST_MODE_ALLMULTI: + disable = WX_PSR_VM_L2CTL_UPE | WX_PSR_VM_L2CTL_VPE; + enable = WX_PSR_VM_L2CTL_BAM | WX_PSR_VM_L2CTL_ROMPE | + WX_PSR_VM_L2CTL_MPE; + break; + case WXVF_XCAST_MODE_PROMISC: + disable = 0; + enable = WX_PSR_VM_L2CTL_BAM | WX_PSR_VM_L2CTL_ROMPE | + WX_PSR_VM_L2CTL_MPE | WX_PSR_VM_L2CTL_UPE | + WX_PSR_VM_L2CTL_VPE; + break; + default: + return -EOPNOTSUPP; + } + + vmolr = rd32(wx, WX_PSR_VM_L2CTL(vf)); + vmolr &= ~disable; + vmolr |= enable; + wr32(wx, WX_PSR_VM_L2CTL(vf), vmolr); + + wx->vfinfo[vf].xcast_mode = xcast_mode; + msgbuf[1] = xcast_mode; + + return 0; +} + +static void wx_rcv_msg_from_vf(struct wx *wx, u16 vf) +{ + u16 mbx_size = WX_VXMAILBOX_SIZE; + u32 msgbuf[WX_VXMAILBOX_SIZE]; + int retval; + + retval = wx_read_mbx_pf(wx, msgbuf, mbx_size, vf); + if (retval) { + wx_err(wx, "Error receiving message from VF\n"); + return; + } + + /* this is a message we already processed, do nothing */ + if (msgbuf[0] & (WX_VT_MSGTYPE_ACK | WX_VT_MSGTYPE_NACK)) + return; + + if (msgbuf[0] == WX_VF_RESET) { + wx_vf_reset_msg(wx, vf); + return; + } + + /* until the vf completes a virtual function reset it should not be + * allowed to start any configuration. + */ + if (!wx->vfinfo[vf].clear_to_send) { + msgbuf[0] |= WX_VT_MSGTYPE_NACK; + wx_write_mbx_pf(wx, msgbuf, 1, vf); + return; + } + + switch ((msgbuf[0] & U16_MAX)) { + case WX_VF_SET_MAC_ADDR: + retval = wx_set_vf_mac_addr(wx, msgbuf, vf); + break; + case WX_VF_SET_MULTICAST: + wx_set_vf_multicasts(wx, msgbuf, vf); + retval = 0; + break; + case WX_VF_SET_VLAN: + retval = wx_set_vf_vlan_msg(wx, msgbuf, vf); + break; + case WX_VF_SET_LPE: + wx_set_vf_lpe(wx, msgbuf[1], vf); + retval = 0; + break; + case WX_VF_SET_MACVLAN: + retval = wx_set_vf_macvlan_msg(wx, msgbuf, vf); + break; + case WX_VF_API_NEGOTIATE: + retval = wx_negotiate_vf_api(wx, msgbuf, vf); + break; + case WX_VF_GET_QUEUES: + retval = wx_get_vf_queues(wx, msgbuf, vf); + break; + case WX_VF_GET_LINK_STATE: + retval = wx_get_vf_link_state(wx, msgbuf, vf); + break; + case WX_VF_GET_FW_VERSION: + retval = wx_get_fw_version(wx, msgbuf, vf); + break; + case WX_VF_UPDATE_XCAST_MODE: + retval = wx_update_vf_xcast_mode(wx, msgbuf, vf); + break; + case WX_VF_BACKUP: + break; + default: + wx_err(wx, "Unhandled Msg %8.8x\n", msgbuf[0]); + break; + } + + /* notify the VF of the results of what it sent us */ + if (retval) + msgbuf[0] |= WX_VT_MSGTYPE_NACK; + else + msgbuf[0] |= WX_VT_MSGTYPE_ACK; + + msgbuf[0] |= WX_VT_MSGTYPE_CTS; + + wx_write_mbx_pf(wx, msgbuf, mbx_size, vf); +} + +static void wx_rcv_ack_from_vf(struct wx *wx, u16 vf) +{ + u32 msg = WX_VT_MSGTYPE_NACK; + + /* if device isn't clear to send it shouldn't be reading either */ + if (!wx->vfinfo[vf].clear_to_send) + wx_write_mbx_pf(wx, &msg, 1, vf); +} + +void wx_msg_task(struct wx *wx) +{ + u16 vf; + + for (vf = 0; vf < wx->num_vfs; vf++) { + /* process any reset requests */ + if (!wx_check_for_rst_pf(wx, vf)) + wx_vf_reset_event(wx, vf); + + /* process any messages pending */ + if (!wx_check_for_msg_pf(wx, vf)) + wx_rcv_msg_from_vf(wx, vf); + + /* process any acks */ + if (!wx_check_for_ack_pf(wx, vf)) + wx_rcv_ack_from_vf(wx, vf); + } +} +EXPORT_SYMBOL(wx_msg_task); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h index 1ac03e9c0d55..400c37040e7c 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h @@ -6,5 +6,6 @@ void wx_disable_sriov(struct wx *wx); int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs); +void wx_msg_task(struct wx *wx); #endif /* _WX_SRIOV_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index ecfe7357d952..43c68c9314cf 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -80,6 +80,8 @@ /*********************** Receive DMA registers **************************/ #define WX_RDM_VF_RE(_i) (0x12004 + ((_i) * 4)) +#define WX_RDM_PF_QDE(_i) (0x12080 + ((_i) * 4)) +#define WX_RDM_VFRE_CLR(_i) (0x120A0 + ((_i) * 4)) #define WX_RDM_DRP_PKT 0x12500 #define WX_RDM_PKT_CNT 0x12504 #define WX_RDM_BYTE_CNT_LSB 0x12508 @@ -123,6 +125,7 @@ #define WX_TDM_VF_TE(_i) (0x18004 + ((_i) * 4)) #define WX_TDM_MAC_AS(_i) (0x18060 + ((_i) * 4)) #define WX_TDM_VLAN_AS(_i) (0x18070 + ((_i) * 4)) +#define WX_TDM_VFTE_CLR(_i) (0x180A0 + ((_i) * 4)) /* TDM CTL BIT */ #define WX_TDM_CTL_TE BIT(0) /* Transmit Enable */ @@ -205,6 +208,7 @@ #define WX_PSR_VM_L2CTL(_i) (0x15600 + ((_i) * 4)) #define WX_PSR_VM_L2CTL_UPE BIT(4) /* unicast promiscuous */ #define WX_PSR_VM_L2CTL_VACC BIT(6) /* accept nomatched vlan */ +#define WX_PSR_VM_L2CTL_VPE BIT(7) /* vlan promiscuous mode */ #define WX_PSR_VM_L2CTL_AUPE BIT(8) /* accept untagged packets */ #define WX_PSR_VM_L2CTL_ROMPE BIT(9) /* accept packets in MTA tbl */ #define WX_PSR_VM_L2CTL_ROPE BIT(10) /* accept packets in UC tbl */ @@ -248,6 +252,7 @@ /* VLAN pool filtering masks */ #define WX_PSR_VLAN_SWC_VIEN BIT(31) /* filter is valid */ #define WX_PSR_VLAN_SWC_ENTRIES 64 +#define WX_PSR_VLAN_SWC_VLANID_MASK GENMASK(11, 0) /********************************* RSEC **************************************/ /* general rsec */ @@ -261,6 +266,9 @@ /*********************** Transmit DMA registers **************************/ /* transmit global control */ #define WX_TDM_ETYPE_AS(_i) (0x18058 + ((_i) * 4)) +#define WX_TDM_VLAN_INS(_i) (0x18100 + ((_i) * 4)) +/* Per VF Port VLAN insertion rules */ +#define WX_TDM_VLAN_INS_VLANA_DEFAULT BIT(30) /* Always use default VLAN*/ /****************************** TDB ******************************************/ #define WX_TDB_PB_SZ(_i) (0x1CC00 + ((_i) * 4)) @@ -301,6 +309,9 @@ #define WX_MAC_WDG_TIMEOUT 0x1100C #define WX_MAC_RX_FLOW_CTRL 0x11090 #define WX_MAC_RX_FLOW_CTRL_RFE BIT(0) /* receive fc enable */ + +#define WX_MAC_WDG_TIMEOUT_WTO_MASK GENMASK(3, 0) +#define WX_MAC_WDG_TIMEOUT_WTO_DELTA 2 /* MDIO Registers */ #define WX_MSCA 0x11200 #define WX_MSCA_RA(v) FIELD_PREP(U16_MAX, v) @@ -1077,6 +1088,11 @@ struct vf_data_storage { bool link_enable; bool trusted; int xcast_mode; + unsigned int vf_api; + bool clear_to_send; + u16 pf_vlan; /* When set, guest VLAN config not allowed. */ + u16 pf_qos; + bool pf_set_mac; u16 vf_mc_hashes[WX_MAX_VF_MC_ENTRIES]; u16 num_vf_mc_hashes; @@ -1188,6 +1204,7 @@ struct wx { u32 wol; u16 bd_number; + bool default_up; struct wx_hw_stats stats; u64 tx_busy; From patchwork Fri Jul 26 10:02:56 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742559 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgsg1.qq.com (smtpbgsg1.qq.com [54.254.200.92]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 9876E17799F for ; Fri, 26 Jul 2024 10:03:40 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.254.200.92 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988223; cv=none; b=JLM6orCTx5aFeexoO/NbvLRcobvOLWv9Z1Y5b/1lQYLNKSsun0BvPe5v/dUhNjbZhx2YcFtNHvwotlRz2Gknb6iSaSyyhrUadpSXjxff1Rgsj8CSfvWmfL8PQuIkCGskqkYYpI31thW9bvHEL1HB2LFejUlkdQfYp8SHu7BPtKI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988223; c=relaxed/simple; bh=ueaN+BOo2MCmVn7HXmgFsLnCAF+85gFHSzim+Jv2DbI=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=aqgJZNdfUZiBI5Q3IrgHuVs2HwzYqEDjkfyK3A5LX+EQyTegcu7i9vaU0Co1IRETSOv4Pqw48Lm6MFoiiI2AwJa0+7HtWcc+cfnTCnBZXKloe86KpFQETKhMBnTp7DoyoiY+/gpXPoQ9uCgew3PimwANjDcaV4xFC5JYxZWyDP0= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.254.200.92 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988211tg2a3bdb X-QQ-Originating-IP: Zk2fo2Q3mskLIcyFWOMq/zMHhMg3o5lL8n9ddOSa41Q= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:29 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 8662454427076051992 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 05/10] net: ngbe: add sriov function support Date: Fri, 26 Jul 2024 18:02:56 +0800 Message-ID: <92E9E8B68E27AE46+20240726100301.21416-6-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Add sriov_configure for driver ops. Add mailbox handler wx_msg_task for ngbe in the interrupt handler. Add the notification flow when the vfs exist. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/wx_sriov.c | 31 ++++++++++ drivers/net/ethernet/wangxun/libwx/wx_sriov.h | 2 + drivers/net/ethernet/wangxun/libwx/wx_type.h | 2 + drivers/net/ethernet/wangxun/ngbe/ngbe_main.c | 58 +++++++++++++++++-- drivers/net/ethernet/wangxun/ngbe/ngbe_mdio.c | 10 ++++ drivers/net/ethernet/wangxun/ngbe/ngbe_type.h | 2 + 6 files changed, 101 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c index f97743a292b5..26f9cce454b2 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c @@ -880,3 +880,34 @@ void wx_msg_task(struct wx *wx) } } EXPORT_SYMBOL(wx_msg_task); + +void wx_disable_vf_rx_tx(struct wx *wx) +{ + wr32(wx, WX_TDM_VFTE_CLR(0), 0); + wr32(wx, WX_RDM_VFRE_CLR(0), 0); + if (wx->mac.type == wx_mac_sp) { + wr32(wx, WX_TDM_VFTE_CLR(1), 0); + wr32(wx, WX_RDM_VFRE_CLR(1), 0); + } +} +EXPORT_SYMBOL(wx_disable_vf_rx_tx); + +void wx_ping_all_vfs_with_link_status(struct wx *wx, bool link_up) +{ + u32 msgbuf[2] = {0, 0}; + u16 i; + + if (!wx->num_vfs) + return; + msgbuf[0] = WX_PF_NOFITY_VF_LINK_STATUS | WX_PF_CONTROL_MSG; + if (link_up) + msgbuf[1] = (wx->speed << 1) | link_up; + if (wx->notify_down) + msgbuf[1] |= WX_PF_NOFITY_VF_NET_NOT_RUNNING; + for (i = 0 ; i < wx->num_vfs; i++) { + if (wx->vfinfo[i].clear_to_send) + msgbuf[0] |= WX_VT_MSGTYPE_CTS; + wx_write_mbx_pf(wx, msgbuf, 2, i); + } +} +EXPORT_SYMBOL(wx_ping_all_vfs_with_link_status); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h index 400c37040e7c..fa1b2b0ecf0a 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h @@ -7,5 +7,7 @@ void wx_disable_sriov(struct wx *wx); int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs); void wx_msg_task(struct wx *wx); +void wx_disable_vf_rx_tx(struct wx *wx); +void wx_ping_all_vfs_with_link_status(struct wx *wx, bool link_up); #endif /* _WX_SRIOV_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index 43c68c9314cf..a7d2a27316e4 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -90,6 +90,7 @@ /************************* Port Registers ************************************/ /* port cfg Registers */ #define WX_CFG_PORT_CTL 0x14400 +#define WX_CFG_PORT_CTL_PFRSTD BIT(14) #define WX_CFG_PORT_CTL_DRV_LOAD BIT(3) #define WX_CFG_PORT_CTL_QINQ BIT(2) #define WX_CFG_PORT_CTL_D_VLAN BIT(0) /* double vlan*/ @@ -1150,6 +1151,7 @@ struct wx { enum wx_reset_type reset_type; /* PHY stuff */ + bool notify_down; unsigned int link; int speed; int duplex; diff --git a/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c b/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c index 53aeae2f884b..a03a4b5f2766 100644 --- a/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c +++ b/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c @@ -14,6 +14,8 @@ #include "../libwx/wx_type.h" #include "../libwx/wx_hw.h" #include "../libwx/wx_lib.h" +#include "../libwx/wx_mbx.h" +#include "../libwx/wx_sriov.h" #include "ngbe_type.h" #include "ngbe_mdio.h" #include "ngbe_hw.h" @@ -128,6 +130,10 @@ static int ngbe_sw_init(struct wx *wx) wx->tx_work_limit = NGBE_DEFAULT_TX_WORK; wx->rx_work_limit = NGBE_DEFAULT_RX_WORK; + wx->mbx.size = WX_VXMAILBOX_SIZE; + wx->setup_tc = ngbe_setup_tc; + set_bit(0, &wx->fwd_bitmask); + return 0; } @@ -197,11 +203,25 @@ static irqreturn_t ngbe_intr(int __always_unused irq, void *data) static irqreturn_t ngbe_msix_other(int __always_unused irq, void *data) { - struct wx *wx = data; + struct wx_q_vector *q_vector; + struct wx *wx = data; + u32 eicr; - /* re-enable the original interrupt state, no lsc, no queues */ - if (netif_running(wx->netdev)) - ngbe_irq_enable(wx, false); + q_vector = wx->q_vector[0]; + + eicr = wx_misc_isb(wx, WX_ISB_MISC); + + if (eicr & NGBE_PX_MISC_IC_VF_MBOX) + wx_msg_task(wx); + + if (wx->num_vfs == 7) { + napi_schedule_irqoff(&q_vector->napi); + ngbe_irq_enable(wx, true); + } else { + /* re-enable the original interrupt state, no lsc, no queues */ + if (netif_running(wx->netdev)) + ngbe_irq_enable(wx, false); + } return IRQ_HANDLED; } @@ -291,6 +311,22 @@ static void ngbe_disable_device(struct wx *wx) struct net_device *netdev = wx->netdev; u32 i; + if (wx->num_vfs) { + /* Clear EITR Select mapping */ + wr32(wx, WX_PX_ITRSEL, 0); + + /* Mark all the VFs as inactive */ + for (i = 0 ; i < wx->num_vfs; i++) + wx->vfinfo[i].clear_to_send = 0; + wx->notify_down = true; + /* ping all the active vfs to let them know we are going down */ + wx_ping_all_vfs_with_link_status(wx, false); + wx->notify_down = false; + + /* Disable all VFTE/VFRE TX/RX */ + wx_disable_vf_rx_tx(wx); + } + /* disable all enabled rx queues */ for (i = 0; i < wx->num_rx_queues; i++) /* this call also flushes the previous write */ @@ -313,10 +349,17 @@ static void ngbe_disable_device(struct wx *wx) wx_update_stats(wx); } +static void ngbe_reset(struct wx *wx) +{ + wx_flush_sw_mac_table(wx); + wx_mac_set_default_filter(wx, wx->mac.addr); +} + void ngbe_down(struct wx *wx) { phylink_stop(wx->phylink); ngbe_disable_device(wx); + ngbe_reset(wx); wx_clean_all_tx_rings(wx); wx_clean_all_rx_rings(wx); } @@ -339,6 +382,11 @@ void ngbe_up(struct wx *wx) ngbe_sfp_modules_txrx_powerctl(wx, true); phylink_start(wx->phylink); + /* Set PF Reset Done bit so PF/VF Mail Ops can work */ + wr32m(wx, WX_CFG_PORT_CTL, + WX_CFG_PORT_CTL_PFRSTD, WX_CFG_PORT_CTL_PFRSTD); + if (wx->num_vfs) + wx_ping_all_vfs_with_link_status(wx, false); } /** @@ -725,6 +773,7 @@ static void ngbe_remove(struct pci_dev *pdev) struct net_device *netdev; netdev = wx->netdev; + wx_disable_sriov(wx); unregister_netdev(netdev); phylink_destroy(wx->phylink); pci_release_selected_regions(pdev, @@ -784,6 +833,7 @@ static struct pci_driver ngbe_driver = { .suspend = ngbe_suspend, .resume = ngbe_resume, .shutdown = ngbe_shutdown, + .sriov_configure = wx_pci_sriov_configure, }; module_pci_driver(ngbe_driver); diff --git a/drivers/net/ethernet/wangxun/ngbe/ngbe_mdio.c b/drivers/net/ethernet/wangxun/ngbe/ngbe_mdio.c index ec54b18c5fe7..dd01aec87b02 100644 --- a/drivers/net/ethernet/wangxun/ngbe/ngbe_mdio.c +++ b/drivers/net/ethernet/wangxun/ngbe/ngbe_mdio.c @@ -8,6 +8,7 @@ #include "../libwx/wx_type.h" #include "../libwx/wx_hw.h" +#include "../libwx/wx_sriov.h" #include "ngbe_type.h" #include "ngbe_mdio.h" @@ -64,6 +65,11 @@ static void ngbe_mac_config(struct phylink_config *config, unsigned int mode, static void ngbe_mac_link_down(struct phylink_config *config, unsigned int mode, phy_interface_t interface) { + struct wx *wx = phylink_to_wx(config); + + wx->speed = 0; + /* ping all the active vfs to let them know we are going down */ + wx_ping_all_vfs_with_link_status(wx, false); } static void ngbe_mac_link_up(struct phylink_config *config, @@ -103,6 +109,10 @@ static void ngbe_mac_link_up(struct phylink_config *config, wr32(wx, WX_MAC_PKT_FLT, WX_MAC_PKT_FLT_PR); reg = rd32(wx, WX_MAC_WDG_TIMEOUT); wr32(wx, WX_MAC_WDG_TIMEOUT, reg); + + wx->speed = speed; + /* ping all the active vfs to let them know we are going up */ + wx_ping_all_vfs_with_link_status(wx, true); } static const struct phylink_mac_ops ngbe_mac_ops = { diff --git a/drivers/net/ethernet/wangxun/ngbe/ngbe_type.h b/drivers/net/ethernet/wangxun/ngbe/ngbe_type.h index f48ed7fc1805..bb70af035c39 100644 --- a/drivers/net/ethernet/wangxun/ngbe/ngbe_type.h +++ b/drivers/net/ethernet/wangxun/ngbe/ngbe_type.h @@ -72,11 +72,13 @@ #define NGBE_PX_MISC_IEN_DEV_RST BIT(10) #define NGBE_PX_MISC_IEN_ETH_LK BIT(18) #define NGBE_PX_MISC_IEN_INT_ERR BIT(20) +#define NGBE_PX_MISC_IC_VF_MBOX BIT(23) #define NGBE_PX_MISC_IEN_GPIO BIT(26) #define NGBE_PX_MISC_IEN_MASK ( \ NGBE_PX_MISC_IEN_DEV_RST | \ NGBE_PX_MISC_IEN_ETH_LK | \ NGBE_PX_MISC_IEN_INT_ERR | \ + NGBE_PX_MISC_IC_VF_MBOX | \ NGBE_PX_MISC_IEN_GPIO) #define NGBE_INTR_ALL 0x1FF From patchwork Fri Jul 26 10:02:57 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742557 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgjp3.qq.com (smtpbgjp3.qq.com [54.92.39.34]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6D72B225AF for ; Fri, 26 Jul 2024 10:03:37 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.92.39.34 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988220; cv=none; b=iMOqgfKdSstdNmZKWqUZ54TbdyMUT0cRS+lNjXKSOZVmuMCRuxLxixeeMQih0hbqKSiXaes5U7tA0FzII59j/0fLhgbwU6s/xkaK1WHn8nZ5c37jRyv7nQLwyPVoOFTmsf/Ki4Hj5g+gdB5FaYrSzN46hOOngALU4KVqk3Ij3DA= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988220; c=relaxed/simple; bh=d6cQ+mTQsuZYw2H9h1Bs2fbW77E7QxvvzILJRUuzdAU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=Xgni0/TrUztcdzOw5COVN6ubZlpRBOnpCKBm0MeqlH3M6EacUwMDrSJYuHc3hQmzVbqCeXuKqtNZQqgCok3HftfBiVbgC6k6kEku7xwBPZqZtLjNu/I5h7H4hDT4zGF8El6LTCbZsFFZewGPQ2HO7BYxJPMa1dtpi8doMHDp6co= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.92.39.34 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988213t24orqkm X-QQ-Originating-IP: FkgNe2C+YZ7ftQeeSOQ7LPVJQzD2whkIsPGaccy1U7E= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:32 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 10597209048635700018 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 06/10] net: txgbe: add sriov function support Date: Fri, 26 Jul 2024 18:02:57 +0800 Message-ID: <3FCB7037C8FE4DE3+20240726100301.21416-7-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Add sriov_configure for driver ops. Add mailbox handler wx_msg_task for txgbe. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/wx_sriov.c | 42 +++++++++++++++++++ drivers/net/ethernet/wangxun/libwx/wx_sriov.h | 1 + drivers/net/ethernet/wangxun/libwx/wx_type.h | 1 + .../net/ethernet/wangxun/txgbe/txgbe_irq.c | 21 ++++++++-- .../net/ethernet/wangxun/txgbe/txgbe_main.c | 23 ++++++++++ .../net/ethernet/wangxun/txgbe/txgbe_phy.c | 8 ++++ .../net/ethernet/wangxun/txgbe/txgbe_type.h | 4 +- 7 files changed, 96 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c index 26f9cce454b2..08738738d02e 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c @@ -292,6 +292,15 @@ static void wx_clear_vmvir(struct wx *wx, u32 vf) wr32(wx, WX_TDM_VLAN_INS(vf), 0); } +static void wx_ping_vf(struct wx *wx, int vf) +{ + u32 ping = WX_PF_CONTROL_MSG; + + if (wx->vfinfo[vf].clear_to_send) + ping |= WX_VT_MSGTYPE_CTS; + wx_write_mbx_pf(wx, &ping, 1, vf); +} + static void wx_set_vf_rx_tx(struct wx *wx, int vf) { u32 index = WX_VF_REG_OFFSET(vf), vf_bit = WX_VF_IND_SHIFT(vf); @@ -911,3 +920,36 @@ void wx_ping_all_vfs_with_link_status(struct wx *wx, bool link_up) } } EXPORT_SYMBOL(wx_ping_all_vfs_with_link_status); + +static void wx_set_vf_link_state(struct wx *wx, int vf, int state) +{ + wx->vfinfo[vf].link_state = state; + switch (state) { + case IFLA_VF_LINK_STATE_AUTO: + if (netif_running(wx->netdev)) + wx->vfinfo[vf].link_enable = true; + else + wx->vfinfo[vf].link_enable = false; + break; + case IFLA_VF_LINK_STATE_ENABLE: + wx->vfinfo[vf].link_enable = true; + break; + case IFLA_VF_LINK_STATE_DISABLE: + wx->vfinfo[vf].link_enable = false; + break; + } + /* restart the VF */ + wx->vfinfo[vf].clear_to_send = false; + wx_ping_vf(wx, vf); + + wx_set_vf_rx_tx(wx, vf); +} + +void wx_set_all_vfs(struct wx *wx) +{ + int i; + + for (i = 0 ; i < wx->num_vfs; i++) + wx_set_vf_link_state(wx, i, wx->vfinfo[i].link_state); +} +EXPORT_SYMBOL(wx_set_all_vfs); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h index fa1b2b0ecf0a..2aa68947c0d3 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h @@ -9,5 +9,6 @@ int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs); void wx_msg_task(struct wx *wx); void wx_disable_vf_rx_tx(struct wx *wx); void wx_ping_all_vfs_with_link_status(struct wx *wx, bool link_up); +void wx_set_all_vfs(struct wx *wx); #endif /* _WX_SRIOV_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index a7d2a27316e4..e50a75e7c58b 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -1098,6 +1098,7 @@ struct vf_data_storage { u16 vf_mc_hashes[WX_MAX_VF_MC_ENTRIES]; u16 num_vf_mc_hashes; u16 vlan_count; + int link_state; }; struct vf_macvlans { diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_irq.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_irq.c index a4cf682dca65..0d414e92ba0b 100644 --- a/drivers/net/ethernet/wangxun/txgbe/txgbe_irq.c +++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_irq.c @@ -7,6 +7,7 @@ #include "../libwx/wx_type.h" #include "../libwx/wx_lib.h" #include "../libwx/wx_hw.h" +#include "../libwx/wx_sriov.h" #include "txgbe_type.h" #include "txgbe_phy.h" #include "txgbe_irq.h" @@ -117,8 +118,17 @@ static irqreturn_t txgbe_misc_irq_handle(int irq, void *data) struct wx *wx = txgbe->wx; u32 eicr; - if (wx->pdev->msix_enabled) + if (wx->pdev->msix_enabled) { + eicr = wx_misc_isb(wx, WX_ISB_MISC) & TXGBE_PX_MISC_IEN_MASK; + if (!eicr) + return IRQ_NONE; + txgbe->eicr = eicr; + if (eicr & TXGBE_PX_MISC_IC_VF_MBOX) { + wx_msg_task(txgbe->wx); + wx_intr_enable(wx, TXGBE_INTR_MISC); + } return IRQ_WAKE_THREAD; + } eicr = wx_misc_isb(wx, WX_ISB_VEC0); if (!eicr) { @@ -137,6 +147,11 @@ static irqreturn_t txgbe_misc_irq_handle(int irq, void *data) q_vector = wx->q_vector[0]; napi_schedule_irqoff(&q_vector->napi); + eicr = wx_misc_isb(wx, WX_ISB_MISC) & TXGBE_PX_MISC_IEN_MASK; + if (!eicr) + return IRQ_NONE; + txgbe->eicr = eicr; + return IRQ_WAKE_THREAD; } @@ -148,7 +163,7 @@ static irqreturn_t txgbe_misc_irq_thread_fn(int irq, void *data) unsigned int sub_irq; u32 eicr; - eicr = wx_misc_isb(wx, WX_ISB_MISC); + eicr = txgbe->eicr; if (eicr & TXGBE_PX_MISC_GPIO) { sub_irq = irq_find_mapping(txgbe->misc.domain, TXGBE_IRQ_GPIO); handle_nested_irq(sub_irq); @@ -191,7 +206,7 @@ int txgbe_setup_misc_irq(struct txgbe *txgbe) struct wx *wx = txgbe->wx; int hwirq, err; - txgbe->misc.nirqs = 2; + txgbe->misc.nirqs = TXGBE_IRQ_MAX; txgbe->misc.domain = irq_domain_add_simple(NULL, txgbe->misc.nirqs, 0, &txgbe_misc_irq_domain_ops, txgbe); if (!txgbe->misc.domain) diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c index 93180225a6f1..c95d0611cf25 100644 --- a/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c +++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c @@ -14,6 +14,8 @@ #include "../libwx/wx_type.h" #include "../libwx/wx_lib.h" #include "../libwx/wx_hw.h" +#include "../libwx/wx_mbx.h" +#include "../libwx/wx_sriov.h" #include "txgbe_type.h" #include "txgbe_hw.h" #include "txgbe_phy.h" @@ -100,6 +102,12 @@ static void txgbe_up_complete(struct wx *wx) /* enable transmits */ netif_tx_start_all_queues(netdev); + + /* Set PF Reset Done bit so PF/VF Mail Ops can work */ + wr32m(wx, WX_CFG_PORT_CTL, WX_CFG_PORT_CTL_PFRSTD, + WX_CFG_PORT_CTL_PFRSTD); + /* update setting rx tx for all active vfs */ + wx_set_all_vfs(wx); } static void txgbe_reset(struct wx *wx) @@ -145,6 +153,16 @@ static void txgbe_disable_device(struct wx *wx) wx_err(wx, "%s: invalid bus lan id %d\n", __func__, wx->bus.func); + if (wx->num_vfs) { + /* Clear EITR Select mapping */ + wr32(wx, WX_PX_ITRSEL, 0); + /* Mark all the VFs as inactive */ + for (i = 0 ; i < wx->num_vfs; i++) + wx->vfinfo[i].clear_to_send = 0; + /* update setting rx tx for all active vfs */ + wx_set_all_vfs(wx); + } + if (!(((wx->subsystem_device_id & WX_NCSI_MASK) == WX_NCSI_SUP) || ((wx->subsystem_device_id & WX_WOL_MASK) == WX_WOL_SUP))) { /* disable mac transmiter */ @@ -277,8 +295,11 @@ static int txgbe_sw_init(struct wx *wx) /* set default work limits */ wx->tx_work_limit = TXGBE_DEFAULT_TX_WORK; wx->rx_work_limit = TXGBE_DEFAULT_RX_WORK; + wx->mbx.size = WX_VXMAILBOX_SIZE; + wx->setup_tc = txgbe_setup_tc; wx->do_reset = txgbe_do_reset; + set_bit(0, &wx->fwd_bitmask); return 0; } @@ -743,6 +764,7 @@ static void txgbe_remove(struct pci_dev *pdev) struct net_device *netdev; netdev = wx->netdev; + wx_disable_sriov(wx); unregister_netdev(netdev); txgbe_remove_phy(txgbe); @@ -765,6 +787,7 @@ static struct pci_driver txgbe_driver = { .probe = txgbe_probe, .remove = txgbe_remove, .shutdown = txgbe_shutdown, + .sriov_configure = wx_pci_sriov_configure, }; module_pci_driver(txgbe_driver); diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c index 5f502265f0a6..76635d4366e4 100644 --- a/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c +++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c @@ -16,6 +16,7 @@ #include "../libwx/wx_type.h" #include "../libwx/wx_lib.h" #include "../libwx/wx_hw.h" +#include "../libwx/wx_sriov.h" #include "txgbe_type.h" #include "txgbe_phy.h" #include "txgbe_hw.h" @@ -179,6 +180,9 @@ static void txgbe_mac_link_down(struct phylink_config *config, struct wx *wx = phylink_to_wx(config); wr32m(wx, WX_MAC_TX_CFG, WX_MAC_TX_CFG_TE, 0); + wx->speed = 0; + /* ping all the active vfs to let them know we are going down */ + wx_ping_all_vfs_with_link_status(wx, false); } static void txgbe_mac_link_up(struct phylink_config *config, @@ -215,6 +219,10 @@ static void txgbe_mac_link_up(struct phylink_config *config, wr32(wx, WX_MAC_PKT_FLT, WX_MAC_PKT_FLT_PR); wdg = rd32(wx, WX_MAC_WDG_TIMEOUT); wr32(wx, WX_MAC_WDG_TIMEOUT, wdg); + + wx->speed = speed; + /* ping all the active vfs to let them know we are going up */ + wx_ping_all_vfs_with_link_status(wx, true); } static int txgbe_mac_prepare(struct phylink_config *config, unsigned int mode, diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_type.h b/drivers/net/ethernet/wangxun/txgbe/txgbe_type.h index 959102c4c379..6a95423bb253 100644 --- a/drivers/net/ethernet/wangxun/txgbe/txgbe_type.h +++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_type.h @@ -71,12 +71,13 @@ #define TXGBE_PX_MISC_ETH_LK BIT(18) #define TXGBE_PX_MISC_ETH_AN BIT(19) #define TXGBE_PX_MISC_INT_ERR BIT(20) +#define TXGBE_PX_MISC_IC_VF_MBOX BIT(23) #define TXGBE_PX_MISC_GPIO BIT(26) #define TXGBE_PX_MISC_IEN_MASK \ (TXGBE_PX_MISC_ETH_LKDN | TXGBE_PX_MISC_DEV_RST | \ TXGBE_PX_MISC_ETH_EVENT | TXGBE_PX_MISC_ETH_LK | \ TXGBE_PX_MISC_ETH_AN | TXGBE_PX_MISC_INT_ERR | \ - TXGBE_PX_MISC_GPIO) + TXGBE_PX_MISC_IC_VF_MBOX | TXGBE_PX_MISC_GPIO) /* Port cfg registers */ #define TXGBE_CFG_PORT_ST 0x14404 @@ -337,6 +338,7 @@ struct txgbe { struct gpio_chip *gpio; unsigned int gpio_irq; unsigned int link_irq; + u32 eicr; /* flow director */ struct hlist_head fdir_filter_list; From patchwork Fri Jul 26 10:02:58 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742560 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgjp3.qq.com (smtpbgjp3.qq.com [54.92.39.34]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 0A196178CE2 for ; Fri, 26 Jul 2024 10:03:40 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.92.39.34 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988223; cv=none; b=c+8mBnyhgcWOu6JBH7xTlEMQXhVF9Ru2Eils9HFqYkLTonlk8ZYOZWzHvdjz2D8xDZMDUD+bCbuEXbxSIhdshlWaqDqsxYfTIlZyNxVFDJ7U3DJ9k7H9EakPfHE5i8QEtfUJOL++JUCWy3fMf8oo41vI7PF3VtqcPTIRuCcE5Hg= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988223; c=relaxed/simple; bh=Be4jhs6/NJHXOfMwC6gQ+zenPHrnAjjAQ1Zhq/k1sp0=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=trTvxi2LKHUn7WWgzn+8812rOvH2sFlvO+L+nl71frdrWd4k9HbOfPbfJWRyoydDqIycrImUk2g0uirOc68KvriIx/WXhtGRjfGhKURZJ+3eWjDlCTIuipNnLwTAb1snSK93LyNlGYAET5MTufldIy/ZGOj7UkEZpNUx/GJFotA= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.92.39.34 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988216th58f9z7 X-QQ-Originating-IP: xMvyEGjQzsu6ac4mNcBBcQxA4A9abQ8R9Irr2RKGg3E= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:34 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 3932851911391721343 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 07/10] net: libwx: allocate devlink for devlink port Date: Fri, 26 Jul 2024 18:02:58 +0800 Message-ID: <576BCDCE2AE928D0+20240726100301.21416-8-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Make devlink allocation function generic to use it for PF and for VF. Add function for PF/VF devlink port creation. It will be used in wangxun NICs. Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/Makefile | 2 +- .../net/ethernet/wangxun/libwx/wx_devlink.c | 208 ++++++++++++++++++ .../net/ethernet/wangxun/libwx/wx_devlink.h | 13 ++ drivers/net/ethernet/wangxun/libwx/wx_sriov.c | 13 +- drivers/net/ethernet/wangxun/libwx/wx_sriov.h | 1 + drivers/net/ethernet/wangxun/libwx/wx_type.h | 14 ++ 6 files changed, 249 insertions(+), 2 deletions(-) create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_devlink.c create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_devlink.h diff --git a/drivers/net/ethernet/wangxun/libwx/Makefile b/drivers/net/ethernet/wangxun/libwx/Makefile index 5b996d973d29..643a5e947ba9 100644 --- a/drivers/net/ethernet/wangxun/libwx/Makefile +++ b/drivers/net/ethernet/wangxun/libwx/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_LIBWX) += libwx.o -libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o wx_sriov.o +libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o wx_sriov.o wx_devlink.o diff --git a/drivers/net/ethernet/wangxun/libwx/wx_devlink.c b/drivers/net/ethernet/wangxun/libwx/wx_devlink.c new file mode 100644 index 000000000000..b39da37c0842 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_devlink.c @@ -0,0 +1,208 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2015 - 2024 Beijing WangXun Technology Co., Ltd. */ + +#include +#include + +#include "wx_type.h" +#include "wx_sriov.h" +#include "wx_devlink.h" + +static const struct devlink_ops wx_pf_devlink_ops = { +}; + +static void wx_devlink_free(void *devlink_ptr) +{ + devlink_unregister((struct devlink *)devlink_ptr); + devlink_free((struct devlink *)devlink_ptr); +} + +struct wx_dl_priv *wx_create_devlink(struct device *dev) +{ + struct devlink *devlink; + + devlink = devlink_alloc(&wx_pf_devlink_ops, sizeof(devlink), dev); + if (!devlink) + return NULL; + + /* Add an action to teardown the devlink when unwinding the driver */ + if (devm_add_action_or_reset(dev, wx_devlink_free, devlink)) + return NULL; + + devlink_register(devlink); + + return devlink_priv(devlink); +} +EXPORT_SYMBOL(wx_create_devlink); + +/** + * wx_devlink_set_switch_id - Set unique switch id based on pci dsn + * @wx: the PF to create a devlink port for + * @ppid: struct with switch id information + */ +static void wx_devlink_set_switch_id(struct wx *wx, + struct netdev_phys_item_id *ppid) +{ + struct pci_dev *pdev = wx->pdev; + u64 id; + + id = pci_get_dsn(pdev); + + ppid->id_len = sizeof(id); + put_unaligned_be64(id, &ppid->id); +} + +/** + * wx_devlink_create_pf_port - Create a devlink port for this PF + * @wx: the PF to create a devlink port for + * + * Create and register a devlink_port for this PF. + * This function has to be called under devl_lock. + * + * Return: zero on success or an error code on failure. + */ +int wx_devlink_create_pf_port(struct wx *wx) +{ + struct devlink *devlink = priv_to_devlink(wx->dl_priv); + struct devlink_port_attrs attrs = {}; + int err; + + attrs.flavour = DEVLINK_PORT_FLAVOUR_PHYSICAL; + attrs.phys.port_number = wx->bus.func; + wx_devlink_set_switch_id(wx, &attrs.switch_id); + devlink_port_attrs_set(&wx->devlink_port, &attrs); + err = devlink_port_register(devlink, &wx->devlink_port, 0); + if (err) { + wx_err(wx, "Failed to create devlink port for PF%d, error %d\n", + wx->bus.func, err); + return err; + } + + return 0; +} +EXPORT_SYMBOL(wx_devlink_create_pf_port); + +/** + * wx_devlink_destroy_pf_port - Destroy the devlink_port for this PF + * @wx: the PF to cleanup + * + * Unregisters the devlink_port structure associated with this PF. + * This function has to be called under devl_lock. + */ +void wx_devlink_destroy_pf_port(struct wx *wx) +{ + devl_port_unregister(&wx->devlink_port); +} +EXPORT_SYMBOL(wx_devlink_destroy_pf_port); + +/** + * wx_devlink_port_get_vf_fn_mac - .port_fn_hw_addr_get devlink handler + * @port: devlink port structure + * @hw_addr: MAC address of the port + * @hw_addr_len: length of MAC address + * @extack: extended netdev ack structure + * + * Callback for the devlink .port_fn_hw_addr_get operation + * Return: zero on success or an error code on failure. + */ +static int wx_devlink_port_get_vf_fn_mac(struct devlink_port *port, + u8 *hw_addr, int *hw_addr_len, + struct netlink_ext_ack *extack) +{ + struct vf_data_storage *vfinfo = container_of(port, + struct vf_data_storage, + devlink_port); + struct devlink_port_attrs *attrs = &port->attrs; + struct devlink_port_pci_vf_attrs *pci_vf; + struct wx *wx = vfinfo->vf_priv_wx; + u16 vf_id; + + pci_vf = &attrs->pci_vf; + vf_id = pci_vf->vf; + + ether_addr_copy(hw_addr, wx->vfinfo[vf_id].vf_mac_addr); + *hw_addr_len = ETH_ALEN; + + return 0; +} + +/** + * wx_devlink_port_set_vf_fn_mac - .port_fn_hw_addr_set devlink handler + * @port: devlink port structure + * @hw_addr: MAC address of the port + * @hw_addr_len: length of MAC address + * @extack: extended netdev ack structure + * + * Callback for the devlink .port_fn_hw_addr_set operation + * Return: zero on success or an error code on failure. + */ +static int wx_devlink_port_set_vf_fn_mac(struct devlink_port *port, + const u8 *hw_addr, + int hw_addr_len, + struct netlink_ext_ack *extack) + +{ + struct vf_data_storage *vfinfo = container_of(port, + struct vf_data_storage, + devlink_port); + struct devlink_port_attrs *attrs = &port->attrs; + struct devlink_port_pci_vf_attrs *pci_vf; + struct wx *wx = vfinfo->vf_priv_wx; + int ret = 0; + u16 vf_id; + + pci_vf = &attrs->pci_vf; + vf_id = pci_vf->vf; + + if (!is_valid_ether_addr(hw_addr) || vf_id >= wx->num_vfs) + return -EINVAL; + + ret = wx_set_vf_mac(wx, vf_id, hw_addr); + if (ret >= 0) { + wx->vfinfo[vf_id].pf_set_mac = true; + if (!netif_running(wx->netdev)) + wx_err(wx, "Bring the PF device up before use vfs\n"); + } else { + wx_err(wx, "The VF MAC address was NOT set due to invalid\n"); + } + + return 0; +} + +static const struct devlink_port_ops wx_devlink_vf_port_ops = { + .port_fn_hw_addr_get = wx_devlink_port_get_vf_fn_mac, + .port_fn_hw_addr_set = wx_devlink_port_set_vf_fn_mac, +}; + +int wx_devlink_create_vf_port(struct wx *wx, int vf_id) +{ + struct devlink *devlink = priv_to_devlink(wx->dl_priv); + struct devlink_port_attrs attrs = {}; + struct devlink_port *devlink_port; + int err; + + devlink_port = &wx->vfinfo[vf_id].devlink_port; + attrs.flavour = DEVLINK_PORT_FLAVOUR_PCI_VF; + attrs.pci_vf.pf = wx->bus.func; + attrs.pci_vf.vf = vf_id; + + wx_devlink_set_switch_id(wx, &attrs.switch_id); + devlink_port_attrs_set(devlink_port, &attrs); + err = devl_port_register_with_ops(devlink, devlink_port, vf_id + 1, + &wx_devlink_vf_port_ops); + if (err) { + wx_err(wx, "Failed to create devlink port for VF %d, error %d\n", + vf_id, err); + return err; + } + + return 0; +} + +void wx_devlink_destroy_vf_port(struct wx *wx) +{ + int i; + + for (i = 0; i < wx->num_vfs; i++) + devl_port_unregister(&wx->vfinfo[i].devlink_port); +} diff --git a/drivers/net/ethernet/wangxun/libwx/wx_devlink.h b/drivers/net/ethernet/wangxun/libwx/wx_devlink.h new file mode 100644 index 000000000000..0754579ed304 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_devlink.h @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2015 - 2024 Beijing WangXun Technology Co., Ltd. */ + +#ifndef _WX_DEVLINK_H_ +#define _WX_DEVLINK_H_ + +struct wx_dl_priv *wx_create_devlink(struct device *dev); +int wx_devlink_create_pf_port(struct wx *wx); +void wx_devlink_destroy_pf_port(struct wx *wx); +int wx_devlink_create_vf_port(struct wx *wx, int vf_id); +void wx_devlink_destroy_vf_port(struct wx *wx); + +#endif /* _WX_DEVLINK_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c index 08738738d02e..88ef82f1a1a3 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.c @@ -7,6 +7,7 @@ #include "wx_type.h" #include "wx_hw.h" #include "wx_mbx.h" +#include "wx_devlink.h" #include "wx_sriov.h" static void wx_vf_configuration(struct pci_dev *pdev, int event_mask) @@ -92,6 +93,7 @@ static int __wx_enable_sriov(struct wx *wx, u8 num_vfs) wr32m(wx, WX_PSR_CTL, WX_PSR_CTL_SW_EN, WX_PSR_CTL_SW_EN); for (i = 0; i < num_vfs; i++) { + wx->vfinfo[i].vf_priv_wx = wx; /* enable spoof checking for all VFs */ wx->vfinfo[i].spoofchk_enabled = true; wx->vfinfo[i].link_enable = true; @@ -130,6 +132,7 @@ void wx_disable_sriov(struct wx *wx) else wx_err(wx, "Unloading driver while VFs are assigned.\n"); + wx_devlink_destroy_vf_port(wx); /* clear flags and free allloced data */ wx_sriov_clear_data(wx); } @@ -158,7 +161,15 @@ static int wx_pci_sriov_enable(struct pci_dev *dev, goto err_out; } + for (i = 0; i < wx->num_vfs; i++) { + err = wx_devlink_create_vf_port(wx, i); + if (err) + goto err_dis_sriov; + } + return num_vfs; +err_dis_sriov: + pci_disable_sriov(dev); err_out: wx_sriov_clear_data(wx); return err; @@ -210,7 +221,7 @@ int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs) } EXPORT_SYMBOL(wx_pci_sriov_configure); -static int wx_set_vf_mac(struct wx *wx, u16 vf, const u8 *mac_addr) +int wx_set_vf_mac(struct wx *wx, u16 vf, const u8 *mac_addr) { u8 hw_addr[ETH_ALEN]; int ret = 0; diff --git a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h index 2aa68947c0d3..e876cb03c2c9 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_sriov.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_sriov.h @@ -6,6 +6,7 @@ void wx_disable_sriov(struct wx *wx); int wx_pci_sriov_configure(struct pci_dev *pdev, int num_vfs); +int wx_set_vf_mac(struct wx *wx, u16 vf, const u8 *mac_addr); void wx_msg_task(struct wx *wx); void wx_disable_vf_rx_tx(struct wx *wx); void wx_ping_all_vfs_with_link_status(struct wx *wx, bool link_up); diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index e50a75e7c58b..a8722f69cebb 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -8,6 +8,7 @@ #include #include #include +#include #include #define WX_NCSI_SUP 0x8000 @@ -1084,6 +1085,10 @@ enum wx_state { struct vf_data_storage { struct pci_dev *vfdev; + struct wx *vf_priv_wx; + /* vf devlink port data */ + struct devlink_port devlink_port; + unsigned char vf_mac_addr[ETH_ALEN]; bool spoofchk_enabled; bool link_enable; @@ -1119,6 +1124,10 @@ enum wx_pf_flags { WX_PF_FLAGS_NBITS /* must be last */ }; +struct wx_dl_priv { + struct wx *priv_wx; +}; + struct wx { unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)]; DECLARE_BITMAP(state, WX_STATE_NBITS); @@ -1128,6 +1137,11 @@ struct wx { u8 __iomem *hw_addr; struct pci_dev *pdev; struct net_device *netdev; + + /* devlink port data */ + struct devlink_port devlink_port; + struct wx_dl_priv *dl_priv; + struct wx_bus_info bus; struct wx_mbx_info mbx; struct wx_mac_info mac; From patchwork Fri Jul 26 10:02:59 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742563 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgau1.qq.com (smtpbgau1.qq.com [54.206.16.166]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4C73B17623E for ; Fri, 26 Jul 2024 10:03:49 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.206.16.166 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988233; cv=none; b=kMNdP1QGzg86wkRSIdnQpDmPmENq6WkmSg/4lEVq9FC3SMkAnBSbmHF2FKPuWgNvuBl86iEpyJBWhBUtGjBE0BAzppJNjJHEQAPDheLEr2XK4So/Lunb7V2dqBdbZrMzcfh1hKr0C29NXbbEi1OJqxMhRfNFQcLQFnQR9gCzdYM= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988233; c=relaxed/simple; bh=NrSi7W58Zl3GtwT4ouV3P5T+IOF7mIF+e/rGL7gFjk0=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=AV7VyicwjcMDcC26HVLecJg5gU3r9rHgSvocFB5LgQDyEiGI/wKlcCubix8/7Zs0tVBQBW5697kcERy3WkeNEJVAfkeTWjrK0eHEhfEMYuozD8KWpmtJUBcQAPKSKNbk1kIzEaLnfyPvg5P0GaQDOuusVpqcicQEea9nUZk4Lg0= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.206.16.166 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988218thokokn2 X-QQ-Originating-IP: mYI4iGImCpBYP5/7w5me1KdWim+uuTh2doBppatINrU= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:37 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 1 X-BIZMAIL-ID: 7546676071251062894 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 08/10] net: libwx: add eswitch switch api for devlink ops Date: Fri, 26 Jul 2024 18:02:59 +0800 Message-ID: <0F90835A90AD58A2+20240726100301.21416-9-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/libwx/Makefile | 3 +- .../net/ethernet/wangxun/libwx/wx_devlink.c | 3 ++ .../net/ethernet/wangxun/libwx/wx_eswitch.c | 53 +++++++++++++++++++ .../net/ethernet/wangxun/libwx/wx_eswitch.h | 13 +++++ drivers/net/ethernet/wangxun/libwx/wx_type.h | 1 + 5 files changed, 72 insertions(+), 1 deletion(-) create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_eswitch.c create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_eswitch.h diff --git a/drivers/net/ethernet/wangxun/libwx/Makefile b/drivers/net/ethernet/wangxun/libwx/Makefile index 643a5e947ba9..a7065ad924d1 100644 --- a/drivers/net/ethernet/wangxun/libwx/Makefile +++ b/drivers/net/ethernet/wangxun/libwx/Makefile @@ -4,4 +4,5 @@ obj-$(CONFIG_LIBWX) += libwx.o -libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o wx_sriov.o wx_devlink.o +libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_mbx.o wx_sriov.o wx_devlink.o \ + wx_eswitch.o diff --git a/drivers/net/ethernet/wangxun/libwx/wx_devlink.c b/drivers/net/ethernet/wangxun/libwx/wx_devlink.c index b39da37c0842..f37362af1449 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_devlink.c +++ b/drivers/net/ethernet/wangxun/libwx/wx_devlink.c @@ -6,9 +6,12 @@ #include "wx_type.h" #include "wx_sriov.h" +#include "wx_eswitch.h" #include "wx_devlink.h" static const struct devlink_ops wx_pf_devlink_ops = { + .eswitch_mode_get = wx_eswitch_mode_get, + .eswitch_mode_set = wx_eswitch_mode_set, }; static void wx_devlink_free(void *devlink_ptr) diff --git a/drivers/net/ethernet/wangxun/libwx/wx_eswitch.c b/drivers/net/ethernet/wangxun/libwx/wx_eswitch.c new file mode 100644 index 000000000000..a426a352bf96 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_eswitch.c @@ -0,0 +1,53 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (C) 2019-2021, Intel Corporation. */ + +#include + +#include "wx_type.h" +#include "wx_eswitch.h" +#include "wx_devlink.h" + +int wx_eswitch_mode_set(struct devlink *devlink, u16 mode, + struct netlink_ext_ack *extack) +{ + struct wx_dl_priv *dl_priv = devlink_priv(devlink); + struct wx *wx = dl_priv->priv_wx; + + if (wx->eswitch_mode == mode) + return 0; + + if (wx->num_vfs) { + dev_info(&(wx)->pdev->dev, + "Change eswitch mode is allowed if there is no VFs."); + return -EOPNOTSUPP; + } + + switch (mode) { + case DEVLINK_ESWITCH_MODE_LEGACY: + dev_info(&(wx)->pdev->dev, + "PF%d changed eswitch mode to legacy", + wx->bus.func); + NL_SET_ERR_MSG_MOD(extack, "Changed eswitch mode to legacy"); + break; + case DEVLINK_ESWITCH_MODE_SWITCHDEV: + dev_info(&(wx)->pdev->dev, + "Do not support switchdev in eswitch mode."); + NL_SET_ERR_MSG_MOD(extack, "Do not support switchdev mode."); + return -EINVAL; + default: + NL_SET_ERR_MSG_MOD(extack, "Unknown eswitch mode"); + return -EINVAL; + } + + wx->eswitch_mode = mode; + return 0; +} + +int wx_eswitch_mode_get(struct devlink *devlink, u16 *mode) +{ + struct wx_dl_priv *dl_priv = devlink_priv(devlink); + struct wx *wx = dl_priv->priv_wx; + + *mode = wx->eswitch_mode; + return 0; +} diff --git a/drivers/net/ethernet/wangxun/libwx/wx_eswitch.h b/drivers/net/ethernet/wangxun/libwx/wx_eswitch.h new file mode 100644 index 000000000000..0323931e7df1 --- /dev/null +++ b/drivers/net/ethernet/wangxun/libwx/wx_eswitch.h @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (C) 2019-2021, Intel Corporation. */ + +#ifndef _WX_ESWITCH_H_ +#define _WX_ESWITCH_H_ + +#include + +int wx_eswitch_mode_get(struct devlink *devlink, u16 *mode); +int wx_eswitch_mode_set(struct devlink *devlink, u16 mode, + struct netlink_ext_ack *extack); + +#endif /* _WX_ESWITCH_H_ */ diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h index a8722f69cebb..a3e103e0c365 100644 --- a/drivers/net/ethernet/wangxun/libwx/wx_type.h +++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h @@ -1141,6 +1141,7 @@ struct wx { /* devlink port data */ struct devlink_port devlink_port; struct wx_dl_priv *dl_priv; + u16 eswitch_mode; /* current mode of eswitch */ struct wx_bus_info bus; struct wx_mbx_info mbx; From patchwork Fri Jul 26 10:03:00 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742564 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgbr1.qq.com (smtpbgbr1.qq.com [54.207.19.206]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CAB67176AD5 for ; Fri, 26 Jul 2024 10:03:50 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.207.19.206 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988233; cv=none; b=fALnb3isxjH+mCTQ0lJtwG8NfVykU+3cNXxKvPbXPn9KHM5RNSR2kJfFs8zonxNbPg7j27sb24XIfzSHOyY2Y32lkxXM5Hxjs23HH0nL38SH/DLlkjs0hLODjuKpUcZmG0Kc0Dv3jY21TKLnotL6c9EPXZwbFvISCXdUOVteCRM= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988233; c=relaxed/simple; bh=ixypbAcaozQ6hFgpS39bj8Z4DjGqSTk9XuFkBcaqq+I=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=mp9nR0dX3kuqFhurjELob0iYR/8MVoXKHfM6Puj1RpBuuK6w1hHsV+WCk8xxARE5tpDs0+pd2uXtIlgENhctT7JF/LvaahbACpSqqq+dViaVFY9BaB7qcXuPdCB46PQBvswQcCWTrAI1SL2ye4sCZ9yNZwlGI2sT9DqLsUEN47k= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.207.19.206 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988220t7lr9p41 X-QQ-Originating-IP: XEzhFo4bp/ABUD9TOfEVpr5oYpSjJ4f0BmkeLvZzAy0= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:39 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 1 X-BIZMAIL-ID: 686133209948113762 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 09/10] net: txgbe: add devlink and devlink port created Date: Fri, 26 Jul 2024 18:03:00 +0800 Message-ID: X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Signed-off-by: Mengyuan Lou --- .../net/ethernet/wangxun/txgbe/txgbe_main.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c index c95d0611cf25..bf9756814f76 100644 --- a/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c +++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_main.c @@ -16,6 +16,7 @@ #include "../libwx/wx_hw.h" #include "../libwx/wx_mbx.h" #include "../libwx/wx_sriov.h" +#include "../libwx/wx_devlink.h" #include "txgbe_type.h" #include "txgbe_hw.h" #include "txgbe_phy.h" @@ -564,6 +565,13 @@ static int txgbe_probe(struct pci_dev *pdev, wx->netdev = netdev; wx->pdev = pdev; + wx->dl_priv = wx_create_devlink(&pdev->dev); + if (!wx->dl_priv) { + err = -ENOMEM; + goto err_pci_release_regions; + } + + wx->dl_priv->priv_wx = wx; wx->msg_enable = (1 << DEFAULT_DEBUG_LEVEL_SHIFT) - 1; wx->hw_addr = devm_ioremap(&pdev->dev, @@ -707,9 +715,14 @@ static int txgbe_probe(struct pci_dev *pdev, if (err) goto err_free_misc_irq; - err = register_netdev(netdev); + err = wx_devlink_create_pf_port(wx); if (err) goto err_remove_phy; + SET_NETDEV_DEVLINK_PORT(netdev, &wx->devlink_port); + + err = register_netdev(netdev); + if (err) + goto err_devlink_create_pf_port; pci_set_drvdata(pdev, wx); @@ -731,6 +744,8 @@ static int txgbe_probe(struct pci_dev *pdev, return 0; +err_devlink_create_pf_port: + devl_port_unregister(&wx->devlink_port); err_remove_phy: txgbe_remove_phy(txgbe); err_free_misc_irq: @@ -767,6 +782,7 @@ static void txgbe_remove(struct pci_dev *pdev) wx_disable_sriov(wx); unregister_netdev(netdev); + devl_port_unregister(&wx->devlink_port); txgbe_remove_phy(txgbe); txgbe_free_misc_irq(txgbe); wx_free_isb_resources(wx); From patchwork Fri Jul 26 10:03:01 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mengyuan Lou X-Patchwork-Id: 13742562 X-Patchwork-Delegate: kuba@kernel.org Received: from smtpbgsg1.qq.com (smtpbgsg1.qq.com [54.254.200.92]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id E0383179663 for ; Fri, 26 Jul 2024 10:03:48 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=54.254.200.92 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988231; cv=none; b=YK7OMiAxOTOfO78xXtwyvawabeYYxIlNxQAMlMsx6GvyH7MYY+GQKfNaWhb8RqV9mXrV6dAlCxKezDvfmv47jRfDLPZ0OcaEUajHAiMFntHosjV/HFgzqkCW2+rEG+nR5S7R8PGo8co5mfcstKyeTaVwzsUvpbOsgfrt2o12pTA= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1721988231; c=relaxed/simple; bh=0/oYikfuKK20djbAz28xrKtDpeB717aq+cK6lhEUOdI=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=BpZYwRiPraR0+9DerNiBhwOhGg7Vgwjs9rDxoXfGvy3X9FG/tiki1i+tb//WVdVAF7+ZDoKbCk0XI/aEaAyIyDF8r14VrbMmMqF6jL/TzYKJrAxb7Nq/2BMIHggTiAaA+SRoMogFDI7NCZnwFssmmbS04ZUcRCUVqpuq5ToAEHE= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com; spf=pass smtp.mailfrom=net-swift.com; arc=none smtp.client-ip=54.254.200.92 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=net-swift.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=net-swift.com X-QQ-mid: bizesmtp81t1721988223t7jgz9xy X-QQ-Originating-IP: XnLlgt/j+JNTntotQ7vPTkUErVqGHwVleDb4wK8dcTo= Received: from localhost.localdomain ( [122.231.252.211]) by bizesmtp.qq.com (ESMTP) with id ; Fri, 26 Jul 2024 18:03:41 +0800 (CST) X-QQ-SSF: 0001000000000000000000000000000 X-QQ-GoodBg: 2 X-BIZMAIL-ID: 1084126950472790650 From: Mengyuan Lou To: netdev@vger.kernel.org Cc: jiawenwu@net-swift.com, Mengyuan Lou Subject: [RFC PATCH net-next v5 10/10] net: ngbe: add devlink and devlink port created Date: Fri, 26 Jul 2024 18:03:01 +0800 Message-ID: <57F6E5845914FDC4+20240726100301.21416-11-mengyuanlou@net-swift.com> X-Mailer: git-send-email 2.43.2 In-Reply-To: <20240726100301.21416-1-mengyuanlou@net-swift.com> References: <20240726100301.21416-1-mengyuanlou@net-swift.com> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-QQ-SENDSIZE: 520 Feedback-ID: bizesmtp:net-swift.com:qybglogicsvrsz:qybglogicsvrsz4a-0 X-Patchwork-Delegate: kuba@kernel.org X-Patchwork-State: RFC Signed-off-by: Mengyuan Lou --- drivers/net/ethernet/wangxun/ngbe/ngbe_main.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c b/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c index a03a4b5f2766..784819f8fcd5 100644 --- a/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c +++ b/drivers/net/ethernet/wangxun/ngbe/ngbe_main.c @@ -16,6 +16,7 @@ #include "../libwx/wx_lib.h" #include "../libwx/wx_mbx.h" #include "../libwx/wx_sriov.h" +#include "../libwx/wx_devlink.h" #include "ngbe_type.h" #include "ngbe_mdio.h" #include "ngbe_hw.h" @@ -616,6 +617,13 @@ static int ngbe_probe(struct pci_dev *pdev, wx = netdev_priv(netdev); wx->netdev = netdev; wx->pdev = pdev; + + wx->dl_priv = wx_create_devlink(&pdev->dev); + if (!wx->dl_priv) { + err = -ENOMEM; + goto err_pci_release_regions; + } + wx->dl_priv->priv_wx = wx; wx->msg_enable = BIT(3) - 1; wx->hw_addr = devm_ioremap(&pdev->dev, @@ -735,6 +743,10 @@ static int ngbe_probe(struct pci_dev *pdev, if (err) goto err_clear_interrupt_scheme; + err = wx_devlink_create_pf_port(wx); + if (err) + goto err_devlink_create_pf_port; + err = register_netdev(netdev); if (err) goto err_register; @@ -744,6 +756,8 @@ static int ngbe_probe(struct pci_dev *pdev, return 0; err_register: + devl_port_unregister(&wx->devlink_port); +err_devlink_create_pf_port: phylink_destroy(wx->phylink); wx_control_hw(wx, false); err_clear_interrupt_scheme: @@ -775,6 +789,7 @@ static void ngbe_remove(struct pci_dev *pdev) netdev = wx->netdev; wx_disable_sriov(wx); unregister_netdev(netdev); + devl_port_unregister(&wx->devlink_port); phylink_destroy(wx->phylink); pci_release_selected_regions(pdev, pci_select_bars(pdev, IORESOURCE_MEM));