From patchwork Fri Jul 23 06:16:06 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "huangguangbin \(A\)" X-Patchwork-Id: 485628 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, URIBL_BLOCKED, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 4F62BC4320E for ; Fri, 23 Jul 2021 06:19:59 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 3773660EB6 for ; Fri, 23 Jul 2021 06:19:59 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S234014AbhGWFjQ (ORCPT ); Fri, 23 Jul 2021 01:39:16 -0400 Received: from szxga03-in.huawei.com ([45.249.212.189]:12287 "EHLO szxga03-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233835AbhGWFjL (ORCPT ); Fri, 23 Jul 2021 01:39:11 -0400 Received: from dggemv704-chm.china.huawei.com (unknown [172.30.72.56]) by szxga03-in.huawei.com (SkyGuard) with ESMTP id 4GWJvX3Gfhz7wpK; Fri, 23 Jul 2021 14:15:04 +0800 (CST) Received: from dggemi759-chm.china.huawei.com (10.1.198.145) by dggemv704-chm.china.huawei.com (10.3.19.47) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:42 +0800 Received: from localhost.localdomain (10.67.165.24) by dggemi759-chm.china.huawei.com (10.1.198.145) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:41 +0800 From: Guangbin Huang To: , , , , , , , CC: , , , , Subject: [RFC PATCH net-next 1/4] arm64: barrier: add DGH macros to control memory accesses merging Date: Fri, 23 Jul 2021 14:16:06 +0800 Message-ID: <1627020969-32945-2-git-send-email-huangguangbin2@huawei.com> X-Mailer: git-send-email 2.8.1 In-Reply-To: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> References: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.67.165.24] X-ClientProxiedBy: dggems706-chm.china.huawei.com (10.3.19.183) To dggemi759-chm.china.huawei.com (10.1.198.145) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: Xiongfeng Wang DGH prohibits merging memory accesses with Normal-NC or Device-GRE attributes before the hint instruction with any memory accesses appearing after the hint instruction. Provide macros to expose it to the arch code. Signed-off-by: Xiongfeng Wang Signed-off-by: Cheng Jian Signed-off-by: Yufeng Mo --- arch/arm64/include/asm/assembler.h | 7 +++++++ arch/arm64/include/asm/barrier.h | 1 + 2 files changed, 8 insertions(+) diff --git a/arch/arm64/include/asm/assembler.h b/arch/arm64/include/asm/assembler.h index 89faca0e740d..5a3348b5e9f3 100644 --- a/arch/arm64/include/asm/assembler.h +++ b/arch/arm64/include/asm/assembler.h @@ -90,6 +90,13 @@ .endm /* + * Data gathering hint + */ + .macro dgh + hint #6 + .endm + +/* * RAS Error Synchronization barrier */ .macro esb diff --git a/arch/arm64/include/asm/barrier.h b/arch/arm64/include/asm/barrier.h index 451e11e5fd23..02e1735706d2 100644 --- a/arch/arm64/include/asm/barrier.h +++ b/arch/arm64/include/asm/barrier.h @@ -22,6 +22,7 @@ #define dmb(opt) asm volatile("dmb " #opt : : : "memory") #define dsb(opt) asm volatile("dsb " #opt : : : "memory") +#define dgh() asm volatile("hint #6" : : : "memory") #define psb_csync() asm volatile("hint #17" : : : "memory") #define tsb_csync() asm volatile("hint #18" : : : "memory") #define csdb() asm volatile("hint #20" : : : "memory") From patchwork Fri Jul 23 06:16:07 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "huangguangbin \(A\)" X-Patchwork-Id: 485065 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, URIBL_BLOCKED, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 77E4DC4338F for ; Fri, 23 Jul 2021 06:19:48 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 571A960EB2 for ; Fri, 23 Jul 2021 06:19:48 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233940AbhGWFjM (ORCPT ); Fri, 23 Jul 2021 01:39:12 -0400 Received: from szxga02-in.huawei.com ([45.249.212.188]:7417 "EHLO szxga02-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233733AbhGWFjL (ORCPT ); Fri, 23 Jul 2021 01:39:11 -0400 Received: from dggemv703-chm.china.huawei.com (unknown [172.30.72.53]) by szxga02-in.huawei.com (SkyGuard) with ESMTP id 4GWJwd6Jh0z7wqY; Fri, 23 Jul 2021 14:16:01 +0800 (CST) Received: from dggemi759-chm.china.huawei.com (10.1.198.145) by dggemv703-chm.china.huawei.com (10.3.19.46) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:42 +0800 Received: from localhost.localdomain (10.67.165.24) by dggemi759-chm.china.huawei.com (10.1.198.145) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:42 +0800 From: Guangbin Huang To: , , , , , , , CC: , , , , Subject: [RFC PATCH net-next 2/4] io: add function to flush the write combine buffer to device immediately Date: Fri, 23 Jul 2021 14:16:07 +0800 Message-ID: <1627020969-32945-3-git-send-email-huangguangbin2@huawei.com> X-Mailer: git-send-email 2.8.1 In-Reply-To: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> References: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.67.165.24] X-ClientProxiedBy: dggems706-chm.china.huawei.com (10.3.19.183) To dggemi759-chm.china.huawei.com (10.1.198.145) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: Xiongfeng Wang Device registers can be mapped as write-combine type. In this case, data are not written into the device immediately. They are temporarily stored in the write combine buffer and written into the device when the buffer is full. But in some situation, we need to flush the write combine buffer to device immediately for better performance. So we add a general function called 'flush_wc_write()'. We use DGH instruction to implement this function for ARM64. Signed-off-by: Xiongfeng Wang --- arch/arm64/include/asm/io.h | 2 ++ include/linux/io.h | 6 ++++++ 2 files changed, 8 insertions(+) diff --git a/arch/arm64/include/asm/io.h b/arch/arm64/include/asm/io.h index 7fd836bea7eb..5315d023b2dd 100644 --- a/arch/arm64/include/asm/io.h +++ b/arch/arm64/include/asm/io.h @@ -112,6 +112,8 @@ static inline u64 __raw_readq(const volatile void __iomem *addr) #define __iowmb() dma_wmb() #define __iomb() dma_mb() +#define flush_wc_write() dgh() + /* * Relaxed I/O memory access primitives. These follow the Device memory * ordering rules but do not guarantee any ordering relative to Normal memory diff --git a/include/linux/io.h b/include/linux/io.h index 9595151d800d..469d53444218 100644 --- a/include/linux/io.h +++ b/include/linux/io.h @@ -166,4 +166,10 @@ static inline void arch_io_free_memtype_wc(resource_size_t base, } #endif +/* IO barriers */ + +#ifndef flush_wc_write +#define flush_wc_write() do { } while (0) +#endif + #endif /* _LINUX_IO_H */ From patchwork Fri Jul 23 06:16:08 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "huangguangbin \(A\)" X-Patchwork-Id: 485629 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, URIBL_BLOCKED, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 615D5C4338F for ; Fri, 23 Jul 2021 06:19:53 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 4223D60EB6 for ; Fri, 23 Jul 2021 06:19:53 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233835AbhGWFjS (ORCPT ); Fri, 23 Jul 2021 01:39:18 -0400 Received: from szxga08-in.huawei.com ([45.249.212.255]:12241 "EHLO szxga08-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233875AbhGWFjL (ORCPT ); Fri, 23 Jul 2021 01:39:11 -0400 Received: from dggemv711-chm.china.huawei.com (unknown [172.30.72.55]) by szxga08-in.huawei.com (SkyGuard) with ESMTP id 4GWJt95XrLz1CMw8; Fri, 23 Jul 2021 14:13:53 +0800 (CST) Received: from dggemi759-chm.china.huawei.com (10.1.198.145) by dggemv711-chm.china.huawei.com (10.1.198.66) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:42 +0800 Received: from localhost.localdomain (10.67.165.24) by dggemi759-chm.china.huawei.com (10.1.198.145) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:42 +0800 From: Guangbin Huang To: , , , , , , , CC: , , , , Subject: [RFC PATCH net-next 3/4] net: hns3: add support for TX push mode Date: Fri, 23 Jul 2021 14:16:08 +0800 Message-ID: <1627020969-32945-4-git-send-email-huangguangbin2@huawei.com> X-Mailer: git-send-email 2.8.1 In-Reply-To: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> References: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.67.165.24] X-ClientProxiedBy: dggems706-chm.china.huawei.com (10.3.19.183) To dggemi759-chm.china.huawei.com (10.1.198.145) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: Huazhong Tan For the device that supports the TX push capability, the BD can be directly copied to the device memory. However, due to hardware restrictions, the push mode can be used only when there are no more than two BDs, otherwise, the doorbell mode based on device memory is used. Signed-off-by: Huazhong Tan Signed-off-by: Yufeng Mo --- drivers/net/ethernet/hisilicon/hns3/hnae3.h | 1 + drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 79 ++++++++++++++++++++-- drivers/net/ethernet/hisilicon/hns3/hns3_enet.h | 6 ++ drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c | 2 + .../net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c | 2 + .../ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 11 ++- .../ethernet/hisilicon/hns3/hns3pf/hclge_main.h | 8 +++ .../ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c | 2 + .../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 11 ++- .../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 8 +++ 10 files changed, 122 insertions(+), 8 deletions(-) diff --git a/drivers/net/ethernet/hisilicon/hns3/hnae3.h b/drivers/net/ethernet/hisilicon/hns3/hnae3.h index e0b7c3c44e7b..f19336bbd88a 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hnae3.h +++ b/drivers/net/ethernet/hisilicon/hns3/hnae3.h @@ -163,6 +163,7 @@ struct hnae3_handle; struct hnae3_queue { void __iomem *io_base; + void __iomem *mem_base; struct hnae3_ae_algo *ae_algo; struct hnae3_handle *handle; int tqp_index; /* index in a handle */ diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c index cdb5f14fb6bc..3fe4d9a165fa 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c @@ -2002,9 +2002,73 @@ static int hns3_fill_skb_to_desc(struct hns3_enet_ring *ring, return bd_num; } +static void hns3_tx_push_bd(struct hns3_enet_ring *ring, int num) +{ +#define HNS3_BYTES_PER_64BIT 8 + + struct hns3_desc desc[HNS3_MAX_PUSH_BD_NUM] = {}; + int offset = 0; + + /* make sure everything is visible to device before + * excuting tx push or updating doorbell + */ + dma_wmb(); + + do { + int idx = (ring->next_to_use - num + ring->desc_num) % + ring->desc_num; + + u64_stats_update_begin(&ring->syncp); + ring->stats.tx_push++; + u64_stats_update_end(&ring->syncp); + memcpy(&desc[offset], &ring->desc[idx], + sizeof(struct hns3_desc)); + offset++; + } while (--num); + + __iowrite64_copy(ring->tqp->mem_base, desc, + (sizeof(struct hns3_desc) * HNS3_MAX_PUSH_BD_NUM) / + HNS3_BYTES_PER_64BIT); + + flush_wc_write(); +} + +static void hns3_tx_mem_doorbell(struct hns3_enet_ring *ring) +{ +#define HNS3_MEM_DOORBELL_OFFSET 64 + + __le64 bd_num = cpu_to_le64((u64)ring->pending_buf); + + /* make sure everything is visible to device before + * excuting tx push or updating doorbell + */ + dma_wmb(); + + __iowrite64_copy(ring->tqp->mem_base + HNS3_MEM_DOORBELL_OFFSET, + &bd_num, 1); + u64_stats_update_begin(&ring->syncp); + ring->stats.tx_mem_doorbell += ring->pending_buf; + u64_stats_update_end(&ring->syncp); + + flush_wc_write(); +} + static void hns3_tx_doorbell(struct hns3_enet_ring *ring, int num, bool doorbell) { + struct net_device *netdev = ring_to_netdev(ring); + struct hns3_nic_priv *priv = netdev_priv(netdev); + + /* when tx push is enabled, the packet whose number of BD below + * HNS3_MAX_PUSH_BD_NUM can be pushed directly. + */ + if (test_bit(HNS3_NIC_STATE_TX_PUSH_ENABLE, &priv->state) && num && + !ring->pending_buf && num <= HNS3_MAX_PUSH_BD_NUM && doorbell) { + hns3_tx_push_bd(ring, num); + WRITE_ONCE(ring->last_to_use, ring->next_to_use); + return; + } + ring->pending_buf += num; if (!doorbell) { @@ -2014,11 +2078,12 @@ static void hns3_tx_doorbell(struct hns3_enet_ring *ring, int num, return; } - if (!ring->pending_buf) - return; + if (ring->tqp->mem_base) + hns3_tx_mem_doorbell(ring); + else + writel(ring->pending_buf, + ring->tqp->io_base + HNS3_RING_TX_RING_TAIL_REG); - writel(ring->pending_buf, - ring->tqp->io_base + HNS3_RING_TX_RING_TAIL_REG); ring->pending_buf = 0; WRITE_ONCE(ring->last_to_use, ring->next_to_use); } @@ -2713,6 +2778,9 @@ static bool hns3_get_tx_timeo_queue_info(struct net_device *ndev) tx_ring->stats.seg_pkt_cnt, tx_ring->stats.tx_more, tx_ring->stats.restart_queue, tx_ring->stats.tx_busy); + netdev_info(ndev, "tx_push: %llu, tx_mem_doorbell: %llu\n", + tx_ring->stats.tx_push, tx_ring->stats.tx_mem_doorbell); + /* When mac received many pause frames continuous, it's unable to send * packets, which may cause tx timeout */ @@ -5060,6 +5128,9 @@ static int hns3_client_init(struct hnae3_handle *handle) if (hnae3_ae_dev_rxd_adv_layout_supported(ae_dev)) set_bit(HNS3_NIC_STATE_RXD_ADV_LAYOUT_ENABLE, &priv->state); + if (test_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps)) + set_bit(HNS3_NIC_STATE_TX_PUSH_ENABLE, &priv->state); + set_bit(HNS3_NIC_STATE_INITED, &priv->state); if (ae_dev->dev_version >= HNAE3_DEVICE_VERSION_V3) diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h index 15af3d93857b..277c4e1bdfa1 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h @@ -6,6 +6,7 @@ #include #include +#include #include "hnae3.h" @@ -21,9 +22,12 @@ enum hns3_nic_state { HNS3_NIC_STATE2_RESET_REQUESTED, HNS3_NIC_STATE_HW_TX_CSUM_ENABLE, HNS3_NIC_STATE_RXD_ADV_LAYOUT_ENABLE, + HNS3_NIC_STATE_TX_PUSH_ENABLE, HNS3_NIC_STATE_MAX }; +#define HNS3_MAX_PUSH_BD_NUM 2 + #define HNS3_RING_RX_RING_BASEADDR_L_REG 0x00000 #define HNS3_RING_RX_RING_BASEADDR_H_REG 0x00004 #define HNS3_RING_RX_RING_BD_NUM_REG 0x00008 @@ -399,6 +403,8 @@ struct ring_stats { u64 tx_pkts; u64 tx_bytes; u64 tx_more; + u64 tx_push; + u64 tx_mem_doorbell; u64 restart_queue; u64 tx_busy; u64 tx_copy; diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c index 82061ab6930f..155a58e11089 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c @@ -37,6 +37,8 @@ static const struct hns3_stats hns3_txq_stats[] = { HNS3_TQP_STAT("packets", tx_pkts), HNS3_TQP_STAT("bytes", tx_bytes), HNS3_TQP_STAT("more", tx_more), + HNS3_TQP_STAT("push", tx_push), + HNS3_TQP_STAT("mem_doorbell", tx_mem_doorbell), HNS3_TQP_STAT("wake", restart_queue), HNS3_TQP_STAT("busy", tx_busy), HNS3_TQP_STAT("copy", tx_copy), diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c index 887297e37cf3..fe985fd65870 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c @@ -395,6 +395,8 @@ static void hclge_parse_capability(struct hclge_dev *hdev, set_bit(HNAE3_DEV_SUPPORT_PORT_VLAN_BYPASS_B, ae_dev->caps); set_bit(HNAE3_DEV_SUPPORT_VLAN_FLTR_MDF_B, ae_dev->caps); } + if (hnae3_get_bit(caps, HCLGE_CAP_TX_PUSH_B)) + set_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps); } static __le32 hclge_build_api_caps(void) diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c index f15d76ec0068..d97ebc9ee12a 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c @@ -1643,6 +1643,7 @@ static int hclge_config_gro(struct hclge_dev *hdev, bool en) static int hclge_alloc_tqps(struct hclge_dev *hdev) { + struct hnae3_ae_dev *ae_dev = pci_get_drvdata(hdev->pdev); struct hclge_tqp *tqp; int i; @@ -1676,6 +1677,14 @@ static int hclge_alloc_tqps(struct hclge_dev *hdev) (i - HCLGE_TQP_MAX_SIZE_DEV_V2) * HCLGE_TQP_REG_SIZE; + /* when device supports tx push and has device memory, + * the queue can execute push mode or doorbell mode on + * device memory. + */ + if (test_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps)) + tqp->q.mem_base = hdev->hw.mem_base + + HCLGE_TQP_MEM_OFFSET(hdev, i); + tqp++; } @@ -11260,8 +11269,6 @@ static void hclge_uninit_client_instance(struct hnae3_client *client, static int hclge_dev_mem_map(struct hclge_dev *hdev) { -#define HCLGE_MEM_BAR 4 - struct pci_dev *pdev = hdev->pdev; struct hclge_hw *hw = &hdev->hw; diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h index cc31b12904ad..50cc45e06697 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h @@ -195,6 +195,14 @@ enum HLCGE_PORT_TYPE { #define HCLGE_VECTOR0_IMP_RD_POISON_B 5U #define HCLGE_VECTOR0_ALL_MSIX_ERR_B 6U +#define HCLGE_TQP_MEM_SIZE 0x10000 +#define HCLGE_MEM_BAR 4 +/* in the bar4, the first half is for roce, and the second half is for nic */ +#define HCLGE_NIC_MEM_OFFSET(hdev) \ + (pci_resource_len((hdev)->pdev, HCLGE_MEM_BAR) >> 1) +#define HCLGE_TQP_MEM_OFFSET(hdev, i) \ + (HCLGE_NIC_MEM_OFFSET(hdev) + HCLGE_TQP_MEM_SIZE * (i)) + #define HCLGE_MAC_DEFAULT_FRAME \ (ETH_HLEN + ETH_FCS_LEN + 2 * VLAN_HLEN + ETH_DATA_LEN) #define HCLGE_MAC_MIN_FRAME 64 diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c index bd19a2d89f6c..55c56c28bb81 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c @@ -361,6 +361,8 @@ static void hclgevf_parse_capability(struct hclgevf_dev *hdev, set_bit(HNAE3_DEV_SUPPORT_UDP_TUNNEL_CSUM_B, ae_dev->caps); if (hnae3_get_bit(caps, HCLGEVF_CAP_RXD_ADV_LAYOUT_B)) set_bit(HNAE3_DEV_SUPPORT_RXD_ADV_LAYOUT_B, ae_dev->caps); + if (hnae3_get_bit(caps, HCLGEVF_CAP_TX_PUSH_B)) + set_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps); } static __le32 hclgevf_build_api_caps(void) diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c index 3a19f08bfff3..586be7d3de00 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c @@ -397,6 +397,7 @@ static int hclgevf_get_pf_media_type(struct hclgevf_dev *hdev) static int hclgevf_alloc_tqps(struct hclgevf_dev *hdev) { + struct hnae3_ae_dev *ae_dev = pci_get_drvdata(hdev->pdev); struct hclgevf_tqp *tqp; int i; @@ -430,6 +431,14 @@ static int hclgevf_alloc_tqps(struct hclgevf_dev *hdev) (i - HCLGEVF_TQP_MAX_SIZE_DEV_V2) * HCLGEVF_TQP_REG_SIZE; + /* when device supports tx push and has device memory, + * the queue can execute push mode or doorbell mode on + * device memory. + */ + if (test_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps)) + tqp->q.mem_base = hdev->hw.mem_base + + HCLGEVF_TQP_MEM_OFFSET(hdev, i); + tqp++; } @@ -3012,8 +3021,6 @@ static void hclgevf_uninit_client_instance(struct hnae3_client *client, static int hclgevf_dev_mem_map(struct hclgevf_dev *hdev) { -#define HCLGEVF_MEM_BAR 4 - struct pci_dev *pdev = hdev->pdev; struct hclgevf_hw *hw = &hdev->hw; diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h index 6f222a3a0bf2..6f0358bcba7e 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h +++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h @@ -126,6 +126,14 @@ #define HCLGEVF_RSS_INPUT_TUPLE_SCTP_NO_PORT \ (HCLGEVF_D_IP_BIT | HCLGEVF_S_IP_BIT | HCLGEVF_V_TAG_BIT) +#define HCLGEVF_TQP_MEM_SIZE 0x10000 +#define HCLGEVF_MEM_BAR 4 +/* in the bar4, the first half is for roce, and the second half is for nic */ +#define HCLGEVF_NIC_MEM_OFFSET(hdev) \ + (pci_resource_len((hdev)->pdev, HCLGEVF_MEM_BAR) >> 1) +#define HCLGEVF_TQP_MEM_OFFSET(hdev, i) \ + (HCLGEVF_NIC_MEM_OFFSET(hdev) + HCLGEVF_TQP_MEM_SIZE * (i)) + #define HCLGEVF_MAC_MAX_FRAME 9728 #define HCLGEVF_STATS_TIMER_INTERVAL 36U From patchwork Fri Jul 23 06:16:09 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "huangguangbin \(A\)" X-Patchwork-Id: 485064 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, URIBL_BLOCKED, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 6FBC2C4338F for ; Fri, 23 Jul 2021 06:19:56 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 56CEF60EB2 for ; Fri, 23 Jul 2021 06:19:56 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S234066AbhGWFjU (ORCPT ); Fri, 23 Jul 2021 01:39:20 -0400 Received: from szxga03-in.huawei.com ([45.249.212.189]:12288 "EHLO szxga03-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233846AbhGWFjL (ORCPT ); Fri, 23 Jul 2021 01:39:11 -0400 Received: from dggemv704-chm.china.huawei.com (unknown [172.30.72.56]) by szxga03-in.huawei.com (SkyGuard) with ESMTP id 4GWJvX6tWNz7xBb; Fri, 23 Jul 2021 14:15:04 +0800 (CST) Received: from dggemi759-chm.china.huawei.com (10.1.198.145) by dggemv704-chm.china.huawei.com (10.3.19.47) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:43 +0800 Received: from localhost.localdomain (10.67.165.24) by dggemi759-chm.china.huawei.com (10.1.198.145) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.2176.2; Fri, 23 Jul 2021 14:19:42 +0800 From: Guangbin Huang To: , , , , , , , CC: , , , , Subject: [RFC PATCH net-next 4/4] net: hns3: add ethtool priv-flag for TX push Date: Fri, 23 Jul 2021 14:16:09 +0800 Message-ID: <1627020969-32945-5-git-send-email-huangguangbin2@huawei.com> X-Mailer: git-send-email 2.8.1 In-Reply-To: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> References: <1627020969-32945-1-git-send-email-huangguangbin2@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.67.165.24] X-ClientProxiedBy: dggems706-chm.china.huawei.com (10.3.19.183) To dggemi759-chm.china.huawei.com (10.1.198.145) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: Huazhong Tan Add a control private flag in ethtool for enable/disable TX push feature. Signed-off-by: Huazhong Tan Signed-off-by: Yufeng Mo --- drivers/net/ethernet/hisilicon/hns3/hnae3.h | 1 + drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 5 ++++- drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c | 19 ++++++++++++++++++- 3 files changed, 23 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/hisilicon/hns3/hnae3.h b/drivers/net/ethernet/hisilicon/hns3/hnae3.h index f19336bbd88a..48d1f369f00e 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hnae3.h +++ b/drivers/net/ethernet/hisilicon/hns3/hnae3.h @@ -807,6 +807,7 @@ struct hnae3_roce_private_info { enum hnae3_pflag { HNAE3_PFLAG_LIMIT_PROMISC, + HNAE3_PFLAG_PUSH_ENABLE, HNAE3_PFLAG_MAX }; diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c index 3fe4d9a165fa..a6434416006c 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c @@ -5128,8 +5128,11 @@ static int hns3_client_init(struct hnae3_handle *handle) if (hnae3_ae_dev_rxd_adv_layout_supported(ae_dev)) set_bit(HNS3_NIC_STATE_RXD_ADV_LAYOUT_ENABLE, &priv->state); - if (test_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps)) + if (test_bit(HNAE3_DEV_SUPPORT_TX_PUSH_B, ae_dev->caps)) { set_bit(HNS3_NIC_STATE_TX_PUSH_ENABLE, &priv->state); + handle->priv_flags |= BIT(HNAE3_PFLAG_PUSH_ENABLE); + set_bit(HNAE3_PFLAG_PUSH_ENABLE, &handle->supported_pflags); + } set_bit(HNS3_NIC_STATE_INITED, &priv->state); diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c index 155a58e11089..0b2557d4441d 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c @@ -423,8 +423,25 @@ static void hns3_update_limit_promisc_mode(struct net_device *netdev, hns3_request_update_promisc_mode(handle); } +static void hns3_update_state(struct net_device *netdev, + enum hns3_nic_state state, bool enable) +{ + struct hns3_nic_priv *priv = netdev_priv(netdev); + + if (enable) + set_bit(state, &priv->state); + else + clear_bit(state, &priv->state); +} + +static void hns3_update_push_state(struct net_device *netdev, bool enable) +{ + hns3_update_state(netdev, HNS3_NIC_STATE_TX_PUSH_ENABLE, enable); +} + static const struct hns3_pflag_desc hns3_priv_flags[HNAE3_PFLAG_MAX] = { - { "limit_promisc", hns3_update_limit_promisc_mode } + { "limit_promisc", hns3_update_limit_promisc_mode }, + { "tx_push_enable", hns3_update_push_state } }; static int hns3_get_sset_count(struct net_device *netdev, int stringset)