From patchwork Fri Jan 5 20:38:17 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Shinas Rasheed X-Patchwork-Id: 185535 Return-Path: Delivered-To: ouuuleilei@gmail.com Received: by 2002:a05:7301:6f82:b0:100:9c79:88ff with SMTP id tb2csp6458004dyb; Fri, 5 Jan 2024 12:39:52 -0800 (PST) X-Google-Smtp-Source: AGHT+IHomiRclLsDbIOxF0xkQ37eIIZs+0cshy04Zwzr1mGipSCjgf4jKkDyWhkZ544obgFauY4A X-Received: by 2002:ac2:4572:0:b0:50e:7c9c:3e1 with SMTP id k18-20020ac24572000000b0050e7c9c03e1mr1195lfm.71.1704487191830; Fri, 05 Jan 2024 12:39:51 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1704487191; cv=none; d=google.com; s=arc-20160816; b=uF17ecBujhTg8M55FRM4+6VWU6jRKLqVWb7yZKlcXYx3RWVDiG0KjKkUEjmSVpDn0I SUoO4kpJhjRaDusvApaS/ato4RmxMGMRPM628jJ2d/mIo/WwVyHnfaRtUThKdMU1NEm9 j8qqnMa7xx+RhXkiTN1J+oWkmY+Jfzo+ZgVNVWJjzS/CBwd1B0kgf6hAbo1PlZs3t7GT GsZzxlxtYhQdMFJtm9M6CjgB7NryzRhBUJaFw3/gleTVIbGvaGueeVxX3zyejrAkBWWL smMXZdfVNK6AtctjxSgU2vMWiSC2ud3atq7pcfbhYiVc487mOCi0pHU9jvJEdnd557OT DSsg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=content-transfer-encoding:mime-version:list-unsubscribe :list-subscribe:list-id:precedence:references:in-reply-to:message-id :date:subject:cc:to:from:dkim-signature; bh=PM3RM2FvW+uUPY40fNW96LhUIWf4Mu83NE0t01Zda2s=; fh=Y/Zcr0YXYeEsWchgSSogMl8sJJcf/EqHGWMcPJ+i19E=; b=CUBJphonn6hmZRb1nlstpsH9Ulssm576soU0VVHhFm0BQT25mFFJoI+3TJC4bRoE0O TLtFSrk5YQxOchs3m88vpcGCITxK9F81pgMTWz/WH4fi/BdtSHWE9qy8qIuuCv0oWoQG HTG5mmpSXP2+FnoVGAJbx+DOUIwpv+3FH0AHI+Tii2ALTMyGQLVi0DXQ5MUKphZuQjMW H/C6XYeF2xfWF/+wvmTgtT9Kkn6dMy6br+1LCtk35RRkLsGDB8O2NqzpM+tGMHqUGBqj zDgUct4n8fvTPCtHFXPI03KJp91L1fCEEh1/jcKhcCIHMSzwWMThejuWWKmizA/PVHx7 Dahg== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@marvell.com header.s=pfpt0220 header.b=hsmIGOJA; spf=pass (google.com: domain of linux-kernel+bounces-18334-ouuuleilei=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) smtp.mailfrom="linux-kernel+bounces-18334-ouuuleilei=gmail.com@vger.kernel.org"; dmarc=pass (p=NONE sp=REJECT dis=NONE) header.from=marvell.com Received: from am.mirrors.kernel.org (am.mirrors.kernel.org. [147.75.80.249]) by mx.google.com with ESMTPS id j9-20020a508a89000000b005522a1302a4si950716edj.677.2024.01.05.12.39.51 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 05 Jan 2024 12:39:51 -0800 (PST) Received-SPF: pass (google.com: domain of linux-kernel+bounces-18334-ouuuleilei=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249; Authentication-Results: mx.google.com; dkim=pass header.i=@marvell.com header.s=pfpt0220 header.b=hsmIGOJA; spf=pass (google.com: domain of linux-kernel+bounces-18334-ouuuleilei=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) smtp.mailfrom="linux-kernel+bounces-18334-ouuuleilei=gmail.com@vger.kernel.org"; dmarc=pass (p=NONE sp=REJECT dis=NONE) header.from=marvell.com Received: from smtp.subspace.kernel.org (wormhole.subspace.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by am.mirrors.kernel.org (Postfix) with ESMTPS id 20A761F245E7 for ; Fri, 5 Jan 2024 20:39:51 +0000 (UTC) Received: from localhost.localdomain (localhost.localdomain [127.0.0.1]) by smtp.subspace.kernel.org (Postfix) with ESMTP id C8CF52D059; Fri, 5 Jan 2024 20:39:14 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=marvell.com header.i=@marvell.com header.b="hsmIGOJA" X-Original-To: linux-kernel@vger.kernel.org Received: from mx0b-0016f401.pphosted.com (mx0a-0016f401.pphosted.com [67.231.148.174]) (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 C557A22EF7; Fri, 5 Jan 2024 20:39:09 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=marvell.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=marvell.com Received: from pps.filterd (m0045849.ppops.net [127.0.0.1]) by mx0a-0016f401.pphosted.com (8.17.1.24/8.17.1.24) with ESMTP id 4058MNIu030833; Fri, 5 Jan 2024 12:38:49 -0800 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com; h= from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding:content-type; s= pfpt0220; bh=PM3RM2FvW+uUPY40fNW96LhUIWf4Mu83NE0t01Zda2s=; b=hsm IGOJAOT+PbTczJlAF70fSwPeCFdYB43xiblzBovupzoaLoLmXwLGRrUfxyhfsOEZ xZdDuUC3w2gUXjgx93AeKq5YmxPT+Uw3YyphF/rJ2djrRxHOIxHlugVIu0sj8TeQ 69zMRU7kfdHaqMiQDoCHuunO96zX5e/E5B3OgODA/cAzwjbeGnd4L7Z7kKcJozyT aAlGWkmXDEgm4eyemk1rqr0l/zEXYsIqdGo+J0JKRLB/Zr7B4vQRRsvpl/ccxea3 dUOYjVBqMt2fQS7bQo/cEjY2DkGPjZMXfOCBU0WojajyEbv1Gn4PFEwJ2MxoTnlk SI38Ni4/JVbmWydpXyg== Received: from dc5-exch02.marvell.com ([199.233.59.182]) by mx0a-0016f401.pphosted.com (PPS) with ESMTPS id 3vee6xa44g-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT); Fri, 05 Jan 2024 12:38:49 -0800 (PST) Received: from DC5-EXCH01.marvell.com (10.69.176.38) by DC5-EXCH02.marvell.com (10.69.176.39) with Microsoft SMTP Server (TLS) id 15.0.1497.48; Fri, 5 Jan 2024 12:38:47 -0800 Received: from maili.marvell.com (10.69.176.80) by DC5-EXCH01.marvell.com (10.69.176.38) with Microsoft SMTP Server id 15.0.1497.48 via Frontend Transport; Fri, 5 Jan 2024 12:38:47 -0800 Received: from ubuntu-PowerEdge-T110-II.sclab.marvell.com (unknown [10.106.27.86]) by maili.marvell.com (Postfix) with ESMTP id 16F693F707C; Fri, 5 Jan 2024 12:38:47 -0800 (PST) From: Shinas Rasheed To: , CC: , , , , , , , , , , , , Veerasenareddy Burru , Satananda Burla , "David S. Miller" , Eric Dumazet Subject: [PATCH net-next v3 2/8] octeon_ep_vf: add hardware configuration APIs Date: Fri, 5 Jan 2024 12:38:17 -0800 Message-ID: <20240105203823.2953604-3-srasheed@marvell.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20240105203823.2953604-1-srasheed@marvell.com> References: <20240105203823.2953604-1-srasheed@marvell.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Proofpoint-GUID: D_Ur7PQ-yteShaQmZIl834ANeivJ20d3 X-Proofpoint-ORIG-GUID: D_Ur7PQ-yteShaQmZIl834ANeivJ20d3 X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.272,Aquarius:18.0.997,Hydra:6.0.619,FMLib:17.11.176.26 definitions=2023-12-09_02,2023-12-07_01,2023-05-22_02 X-getmail-retrieved-from-mailbox: INBOX X-GMAIL-THRID: 1787284361607383513 X-GMAIL-MSGID: 1787284361607383513 Implement hardware resource init and shutdown helper APIs, like hardware Tx/Rx queue init/enable/disable/reset. Signed-off-by: Shinas Rasheed --- V3: - Replaced masks and ULL declarations with GENMASK_ULL(), ULL() and other linux/bits.h macros - Corrected declarations to conform to xmas tree format V2: https://lore.kernel.org/all/20231223134000.2906144-3-srasheed@marvell.com/ - No changes V1: https://lore.kernel.org/all/20231221092844.2885872-3-srasheed@marvell.com/ .../marvell/octeon_ep_vf/octep_vf_cn9k.c | 336 ++++++++++++++++- .../marvell/octeon_ep_vf/octep_vf_cnxk.c | 346 +++++++++++++++++- 2 files changed, 680 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cn9k.c b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cn9k.c index c24ef2265205..bb8c58dd4d7c 100644 --- a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cn9k.c +++ b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cn9k.c @@ -13,9 +13,125 @@ #include "octep_vf_main.h" #include "octep_vf_regs_cn9k.h" +/* Dump useful hardware IQ/OQ CSRs for debug purpose */ +static void cn93_vf_dump_q_regs(struct octep_vf_device *oct, int qno) +{ + struct device *dev = &oct->pdev->dev; + + dev_info(dev, "IQ-%d register dump\n", qno); + dev_info(dev, "R[%d]_IN_INSTR_DBELL[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_INSTR_DBELL(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INSTR_DBELL(qno))); + dev_info(dev, "R[%d]_IN_CONTROL[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_CONTROL(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_CONTROL(qno))); + dev_info(dev, "R[%d]_IN_ENABLE[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_ENABLE(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_ENABLE(qno))); + dev_info(dev, "R[%d]_IN_INSTR_BADDR[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_INSTR_BADDR(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INSTR_BADDR(qno))); + dev_info(dev, "R[%d]_IN_INSTR_RSIZE[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_INSTR_RSIZE(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INSTR_RSIZE(qno))); + dev_info(dev, "R[%d]_IN_CNTS[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_CNTS(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_CNTS(qno))); + dev_info(dev, "R[%d]_IN_INT_LEVELS[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_INT_LEVELS(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(qno))); + dev_info(dev, "R[%d]_IN_PKT_CNT[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_PKT_CNT(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_PKT_CNT(qno))); + dev_info(dev, "R[%d]_IN_BYTE_CNT[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_IN_BYTE_CNT(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_BYTE_CNT(qno))); + + dev_info(dev, "OQ-%d register dump\n", qno); + dev_info(dev, "R[%d]_OUT_SLIST_DBELL[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_SLIST_DBELL(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_DBELL(qno))); + dev_info(dev, "R[%d]_OUT_CONTROL[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_CONTROL(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_CONTROL(qno))); + dev_info(dev, "R[%d]_OUT_ENABLE[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_ENABLE(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_ENABLE(qno))); + dev_info(dev, "R[%d]_OUT_SLIST_BADDR[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_SLIST_BADDR(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_BADDR(qno))); + dev_info(dev, "R[%d]_OUT_SLIST_RSIZE[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_SLIST_RSIZE(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_RSIZE(qno))); + dev_info(dev, "R[%d]_OUT_CNTS[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_CNTS(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_CNTS(qno))); + dev_info(dev, "R[%d]_OUT_INT_LEVELS[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_INT_LEVELS(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(qno))); + dev_info(dev, "R[%d]_OUT_PKT_CNT[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_PKT_CNT(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_PKT_CNT(qno))); + dev_info(dev, "R[%d]_OUT_BYTE_CNT[0x%llx]: 0x%016llx\n", + qno, CN93_VF_SDP_R_OUT_BYTE_CNT(qno), + octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_BYTE_CNT(qno))); +} + +/* Reset Hardware Tx queue */ +static int cn93_vf_reset_iq(struct octep_vf_device *oct, int q_no) +{ + u64 val = ULL(0); + + dev_dbg(&oct->pdev->dev, "Reset VF IQ-%d\n", q_no); + + /* Disable the Tx/Instruction Ring */ + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_ENABLE(q_no), val); + + /* clear the Instruction Ring packet/byte counts and doorbell CSRs */ + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(q_no), val); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_PKT_CNT(q_no), val); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_BYTE_CNT(q_no), val); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INSTR_BADDR(q_no), val); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INSTR_RSIZE(q_no), val); + + val = GENMASK_ULL(31, 0); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INSTR_DBELL(q_no), val); + + val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_CNTS(q_no)); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_CNTS(q_no), + val & GENMASK_ULL(31, 0)); + + return 0; +} + +/* Reset Hardware Rx queue */ +static void cn93_vf_reset_oq(struct octep_vf_device *oct, int q_no) +{ + u64 val = ULL(0); + + /* Disable Output (Rx) Ring */ + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_ENABLE(q_no), val); + + /* Clear count CSRs */ + val = octep_vf_read_csr(oct, CN93_VF_SDP_R_OUT_CNTS(q_no)); + octep_vf_write_csr(oct, CN93_VF_SDP_R_OUT_CNTS(q_no), val); + + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_PKT_CNT(q_no), GENMASK_ULL(35, 0)); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_DBELL(q_no), GENMASK_ULL(31, 0)); +} + /* Reset all hardware Tx/Rx queues */ static void octep_vf_reset_io_queues_cn93(struct octep_vf_device *oct) { + struct pci_dev *pdev = oct->pdev; + int q; + + dev_dbg(&pdev->dev, "Reset OCTEP_CN93 VF IO Queues\n"); + + for (q = 0; q < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); q++) { + cn93_vf_reset_iq(oct, q); + cn93_vf_reset_oq(oct, q); + } } /* Initialize configuration limits and initial active config */ @@ -46,78 +162,296 @@ static void octep_vf_init_config_cn93_vf(struct octep_vf_device *oct) /* Setup registers for a hardware Tx Queue */ static void octep_vf_setup_iq_regs_cn93(struct octep_vf_device *oct, int iq_no) { + struct octep_vf_iq *iq = oct->iq[iq_no]; + u32 reset_instr_cnt; + u64 reg_val; + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_CONTROL(iq_no)); + + /* wait for IDLE to set to 1 */ + if (!(reg_val & CN93_VF_R_IN_CTL_IDLE)) { + do { + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_CONTROL(iq_no)); + } while (!(reg_val & CN93_VF_R_IN_CTL_IDLE)); + } + reg_val |= CN93_VF_R_IN_CTL_RDSIZE; + reg_val |= CN93_VF_R_IN_CTL_IS_64B; + reg_val |= CN93_VF_R_IN_CTL_ESR; + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_CONTROL(iq_no), reg_val); + + /* Write the start of the input queue's ring and its size */ + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INSTR_BADDR(iq_no), iq->desc_ring_dma); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INSTR_RSIZE(iq_no), iq->max_count); + + /* Remember the doorbell & instruction count register addr for this queue */ + iq->doorbell_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_IN_INSTR_DBELL(iq_no); + iq->inst_cnt_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_IN_CNTS(iq_no); + iq->intr_lvl_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_IN_INT_LEVELS(iq_no); + + /* Store the current instruction counter (used in flush_iq calculation) */ + reset_instr_cnt = readl(iq->inst_cnt_reg); + writel(reset_instr_cnt, iq->inst_cnt_reg); + + /* INTR_THRESHOLD is set to max(FFFFFFFF) to disable the INTR */ + reg_val = CFG_GET_IQ_INTR_THRESHOLD(oct->conf) & GENMASK_ULL(31, 0); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(iq_no), reg_val); } /* Setup registers for a hardware Rx Queue */ static void octep_vf_setup_oq_regs_cn93(struct octep_vf_device *oct, int oq_no) { + struct octep_vf_oq *oq = oct->oq[oq_no]; + u32 time_threshold = 0; + u64 oq_ctl = ULL(0); + u64 reg_val; + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_CONTROL(oq_no)); + + /* wait for IDLE to set to 1 */ + if (!(reg_val & CN93_VF_R_OUT_CTL_IDLE)) { + do { + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_CONTROL(oq_no)); + } while (!(reg_val & CN93_VF_R_OUT_CTL_IDLE)); + } + + reg_val &= ~(CN93_VF_R_OUT_CTL_IMODE); + reg_val &= ~(CN93_VF_R_OUT_CTL_ROR_P); + reg_val &= ~(CN93_VF_R_OUT_CTL_NSR_P); + reg_val &= ~(CN93_VF_R_OUT_CTL_ROR_I); + reg_val &= ~(CN93_VF_R_OUT_CTL_NSR_I); + reg_val &= ~(CN93_VF_R_OUT_CTL_ES_I); + reg_val &= ~(CN93_VF_R_OUT_CTL_ROR_D); + reg_val &= ~(CN93_VF_R_OUT_CTL_NSR_D); + reg_val &= ~(CN93_VF_R_OUT_CTL_ES_D); + reg_val |= (CN93_VF_R_OUT_CTL_ES_P); + + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_CONTROL(oq_no), reg_val); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_BADDR(oq_no), oq->desc_ring_dma); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_RSIZE(oq_no), oq->max_count); + + oq_ctl = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_CONTROL(oq_no)); + oq_ctl &= ~GENMASK_ULL(22, 0); //clear the ISIZE and BSIZE (22-0) + oq_ctl |= (oq->buffer_size & GENMASK_ULL(15, 0)); //populate the BSIZE (15-0) + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_CONTROL(oq_no), oq_ctl); + + /* Get the mapped address of the pkt_sent and pkts_credit regs */ + oq->pkts_sent_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_OUT_CNTS(oq_no); + oq->pkts_credit_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_OUT_SLIST_DBELL(oq_no); + + time_threshold = CFG_GET_OQ_INTR_TIME(oct->conf); + reg_val = ((u64)time_threshold << 32) | CFG_GET_OQ_INTR_PKT(oct->conf); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(oq_no), reg_val); } /* Setup registers for a VF mailbox */ static void octep_vf_setup_mbox_regs_cn93(struct octep_vf_device *oct, int q_no) { + struct octep_vf_mbox *mbox = oct->mbox; + + /* PF to VF DATA reg. VF reads from this reg */ + mbox->mbox_read_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_MBOX_PF_VF_DATA(q_no); + + /* VF mbox interrupt reg */ + mbox->mbox_int_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_MBOX_PF_VF_INT(q_no); + + /* VF to PF DATA reg. VF writes into this reg */ + mbox->mbox_write_reg = oct->mmio.hw_addr + CN93_VF_SDP_R_MBOX_VF_PF_DATA(q_no); +} + +/* Mailbox Interrupt handler */ +static void cn93_handle_vf_mbox_intr(struct octep_vf_device *oct) +{ + if (oct->mbox) + schedule_work(&oct->mbox->wk.work); + else + dev_err(&oct->pdev->dev, "cannot schedule work on invalid mbox\n"); } /* Tx/Rx queue interrupt handler */ static irqreturn_t octep_vf_ioq_intr_handler_cn93(void *data) { + struct octep_vf_ioq_vector *vector = data; + struct octep_vf_device *oct; + struct octep_vf_oq *oq; + u64 reg_val = ULL(0); + + oct = vector->octep_vf_dev; + oq = vector->oq; + /* Mailbox interrupt arrives along with interrupt of tx/rx ring pair 0 */ + if (oq->q_no == 0) { + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_MBOX_PF_VF_INT(0)); + if (reg_val & CN93_VF_SDP_R_MBOX_PF_VF_INT_STATUS) { + cn93_handle_vf_mbox_intr(oct); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_MBOX_PF_VF_INT(0), reg_val); + } + } + napi_schedule_irqoff(oq->napi); return IRQ_HANDLED; } /* Re-initialize Octeon hardware registers */ static void octep_vf_reinit_regs_cn93(struct octep_vf_device *oct) { + u32 i; + + for (i = 0; i < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); i++) + oct->hw_ops.setup_iq_regs(oct, i); + + for (i = 0; i < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); i++) + oct->hw_ops.setup_oq_regs(oct, i); + + oct->hw_ops.enable_interrupts(oct); + oct->hw_ops.enable_io_queues(oct); + + for (i = 0; i < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); i++) + writel(oct->oq[i]->max_count, oct->oq[i]->pkts_credit_reg); } /* Enable all interrupts */ static void octep_vf_enable_interrupts_cn93(struct octep_vf_device *oct) { + int num_rings, q; + u64 reg_val; + + num_rings = CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); + for (q = 0; q < num_rings; q++) { + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(q)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(q), reg_val); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(q)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(q), reg_val); + } + /* Enable PF to VF mbox interrupt by setting 2nd bit*/ + octep_vf_write_csr64(oct, CN93_VF_SDP_R_MBOX_PF_VF_INT(0), + CN93_VF_SDP_R_MBOX_PF_VF_INT_ENAB); } /* Disable all interrupts */ static void octep_vf_disable_interrupts_cn93(struct octep_vf_device *oct) { + int num_rings, q; + u64 reg_val; + + /* Disable PF to VF mbox interrupt by setting 2nd bit*/ + if (oct->mbox) + octep_vf_write_csr64(oct, CN93_VF_SDP_R_MBOX_PF_VF_INT(0), 0x0); + + num_rings = CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); + for (q = 0; q < num_rings; q++) { + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(q)); + reg_val &= ~BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(q), reg_val); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(q)); + reg_val &= ~BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(q), reg_val); + } } /* Get new Octeon Read Index: index of descriptor that Octeon reads next. */ static u32 octep_vf_update_iq_read_index_cn93(struct octep_vf_iq *iq) { - return 0; + u32 pkt_in_done = readl(iq->inst_cnt_reg); + u32 last_done, new_idx; + + last_done = pkt_in_done - iq->pkt_in_done; + iq->pkt_in_done = pkt_in_done; + + new_idx = (iq->octep_vf_read_index + last_done) % iq->max_count; + + return new_idx; } /* Enable a hardware Tx Queue */ static void octep_vf_enable_iq_cn93(struct octep_vf_device *oct, int iq_no) { + u64 loop = HZ; + u64 reg_val; + + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INSTR_DBELL(iq_no), GENMASK_ULL(31, 0)); + + while (octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INSTR_DBELL(iq_no)) && + loop--) { + schedule_timeout_interruptible(1); + } + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(iq_no)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_INT_LEVELS(iq_no), reg_val); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_ENABLE(iq_no)); + reg_val |= ULL(1); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_ENABLE(iq_no), reg_val); } /* Enable a hardware Rx Queue */ static void octep_vf_enable_oq_cn93(struct octep_vf_device *oct, int oq_no) { + u64 reg_val = ULL(0); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(oq_no)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_INT_LEVELS(oq_no), reg_val); + + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_SLIST_DBELL(oq_no), GENMASK_ULL(31, 0)); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_ENABLE(oq_no)); + reg_val |= ULL(1); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_ENABLE(oq_no), reg_val); } /* Enable all hardware Tx/Rx Queues assigned to VF */ static void octep_vf_enable_io_queues_cn93(struct octep_vf_device *oct) { + u8 q; + + for (q = 0; q < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); q++) { + octep_vf_enable_iq_cn93(oct, q); + octep_vf_enable_oq_cn93(oct, q); + } } /* Disable a hardware Tx Queue assigned to VF */ static void octep_vf_disable_iq_cn93(struct octep_vf_device *oct, int iq_no) { + u64 reg_val = ULL(0); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_IN_ENABLE(iq_no)); + reg_val &= ~ULL(1); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_IN_ENABLE(iq_no), reg_val); } /* Disable a hardware Rx Queue assigned to VF */ static void octep_vf_disable_oq_cn93(struct octep_vf_device *oct, int oq_no) { + u64 reg_val = ULL(0); + + reg_val = octep_vf_read_csr64(oct, CN93_VF_SDP_R_OUT_ENABLE(oq_no)); + reg_val &= ~ULL(1); + octep_vf_write_csr64(oct, CN93_VF_SDP_R_OUT_ENABLE(oq_no), reg_val); } /* Disable all hardware Tx/Rx Queues assigned to VF */ static void octep_vf_disable_io_queues_cn93(struct octep_vf_device *oct) { + int q = 0; + + for (q = 0; q < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); q++) { + octep_vf_disable_iq_cn93(oct, q); + octep_vf_disable_oq_cn93(oct, q); + } } /* Dump hardware registers (including Tx/Rx queues) for debugging. */ static void octep_vf_dump_registers_cn93(struct octep_vf_device *oct) { + u8 num_rings, q; + + num_rings = CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); + for (q = 0; q < num_rings; q++) + cn93_vf_dump_q_regs(oct, q); } /** diff --git a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cnxk.c b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cnxk.c index af07a4a6edc5..3440d4345985 100644 --- a/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cnxk.c +++ b/drivers/net/ethernet/marvell/octeon_ep_vf/octep_vf_cnxk.c @@ -13,9 +13,127 @@ #include "octep_vf_main.h" #include "octep_vf_regs_cnxk.h" +/* Dump useful hardware IQ/OQ CSRs for debug purpose */ +static void cnxk_vf_dump_q_regs(struct octep_vf_device *oct, int qno) +{ + struct device *dev = &oct->pdev->dev; + + dev_info(dev, "IQ-%d register dump\n", qno); + dev_info(dev, "R[%d]_IN_INSTR_DBELL[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_INSTR_DBELL(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_DBELL(qno))); + dev_info(dev, "R[%d]_IN_CONTROL[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_CONTROL(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_CONTROL(qno))); + dev_info(dev, "R[%d]_IN_ENABLE[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_ENABLE(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_ENABLE(qno))); + dev_info(dev, "R[%d]_IN_INSTR_BADDR[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_INSTR_BADDR(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_BADDR(qno))); + dev_info(dev, "R[%d]_IN_INSTR_RSIZE[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_INSTR_RSIZE(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_RSIZE(qno))); + dev_info(dev, "R[%d]_IN_CNTS[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_CNTS(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_CNTS(qno))); + dev_info(dev, "R[%d]_IN_INT_LEVELS[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_INT_LEVELS(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(qno))); + dev_info(dev, "R[%d]_IN_PKT_CNT[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_PKT_CNT(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_PKT_CNT(qno))); + dev_info(dev, "R[%d]_IN_BYTE_CNT[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_IN_BYTE_CNT(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_BYTE_CNT(qno))); + + dev_info(dev, "OQ-%d register dump\n", qno); + dev_info(dev, "R[%d]_OUT_SLIST_DBELL[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_SLIST_DBELL(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_DBELL(qno))); + dev_info(dev, "R[%d]_OUT_CONTROL[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_CONTROL(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_CONTROL(qno))); + dev_info(dev, "R[%d]_OUT_ENABLE[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_ENABLE(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_ENABLE(qno))); + dev_info(dev, "R[%d]_OUT_SLIST_BADDR[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_SLIST_BADDR(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_BADDR(qno))); + dev_info(dev, "R[%d]_OUT_SLIST_RSIZE[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_SLIST_RSIZE(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_RSIZE(qno))); + dev_info(dev, "R[%d]_OUT_CNTS[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_CNTS(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_CNTS(qno))); + dev_info(dev, "R[%d]_OUT_INT_LEVELS[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_INT_LEVELS(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(qno))); + dev_info(dev, "R[%d]_OUT_PKT_CNT[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_PKT_CNT(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_PKT_CNT(qno))); + dev_info(dev, "R[%d]_OUT_BYTE_CNT[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_OUT_BYTE_CNT(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_BYTE_CNT(qno))); + dev_info(dev, "R[%d]_ERR_TYPE[0x%llx]: 0x%016llx\n", + qno, CNXK_VF_SDP_R_ERR_TYPE(qno), + octep_vf_read_csr64(oct, CNXK_VF_SDP_R_ERR_TYPE(qno))); +} + +/* Reset Hardware Tx queue */ +static int cnxk_vf_reset_iq(struct octep_vf_device *oct, int q_no) +{ + u64 val = ULL(0); + + dev_dbg(&oct->pdev->dev, "Reset VF IQ-%d\n", q_no); + + /* Disable the Tx/Instruction Ring */ + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_ENABLE(q_no), val); + + /* clear the Instruction Ring packet/byte counts and doorbell CSRs */ + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(q_no), val); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_PKT_CNT(q_no), val); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_BYTE_CNT(q_no), val); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_BADDR(q_no), val); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_RSIZE(q_no), val); + + val = GENMASK_ULL(31, 0); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_DBELL(q_no), val); + + val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_CNTS(q_no)); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_CNTS(q_no), val & GENMASK_ULL(31, 0)); + + return 0; +} + +/* Reset Hardware Rx queue */ +static void cnxk_vf_reset_oq(struct octep_vf_device *oct, int q_no) +{ + u64 val = ULL(0); + + /* Disable Output (Rx) Ring */ + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_ENABLE(q_no), val); + + /* Clear count CSRs */ + val = octep_vf_read_csr(oct, CNXK_VF_SDP_R_OUT_CNTS(q_no)); + octep_vf_write_csr(oct, CNXK_VF_SDP_R_OUT_CNTS(q_no), val); + + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_PKT_CNT(q_no), GENMASK_ULL(35, 0)); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_DBELL(q_no), GENMASK_ULL(31, 0)); +} + /* Reset all hardware Tx/Rx queues */ static void octep_vf_reset_io_queues_cnxk(struct octep_vf_device *oct) { + struct pci_dev *pdev = oct->pdev; + int q; + + dev_dbg(&pdev->dev, "Reset OCTEP_CNXK VF IO Queues\n"); + + for (q = 0; q < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); q++) { + cnxk_vf_reset_iq(oct, q); + cnxk_vf_reset_oq(oct, q); + } } /* Initialize configuration limits and initial active config */ @@ -47,78 +165,304 @@ static void octep_vf_init_config_cnxk_vf(struct octep_vf_device *oct) /* Setup registers for a hardware Tx Queue */ static void octep_vf_setup_iq_regs_cnxk(struct octep_vf_device *oct, int iq_no) { + struct octep_vf_iq *iq = oct->iq[iq_no]; + u32 reset_instr_cnt; + u64 reg_val; + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_CONTROL(iq_no)); + + /* wait for IDLE to set to 1 */ + if (!(reg_val & CNXK_VF_R_IN_CTL_IDLE)) { + do { + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_CONTROL(iq_no)); + } while (!(reg_val & CNXK_VF_R_IN_CTL_IDLE)); + } + reg_val |= CNXK_VF_R_IN_CTL_RDSIZE; + reg_val |= CNXK_VF_R_IN_CTL_IS_64B; + reg_val |= CNXK_VF_R_IN_CTL_ESR; + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_CONTROL(iq_no), reg_val); + + /* Write the start of the input queue's ring and its size */ + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_BADDR(iq_no), iq->desc_ring_dma); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_RSIZE(iq_no), iq->max_count); + + /* Remember the doorbell & instruction count register addr for this queue */ + iq->doorbell_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_IN_INSTR_DBELL(iq_no); + iq->inst_cnt_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_IN_CNTS(iq_no); + iq->intr_lvl_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_IN_INT_LEVELS(iq_no); + + /* Store the current instruction counter (used in flush_iq calculation) */ + reset_instr_cnt = readl(iq->inst_cnt_reg); + writel(reset_instr_cnt, iq->inst_cnt_reg); + + /* INTR_THRESHOLD is set to max(FFFFFFFF) to disable the INTR */ + reg_val = CFG_GET_IQ_INTR_THRESHOLD(oct->conf) & GENMASK_ULL(31, 0); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(iq_no), reg_val); } /* Setup registers for a hardware Rx Queue */ static void octep_vf_setup_oq_regs_cnxk(struct octep_vf_device *oct, int oq_no) { + struct octep_vf_oq *oq = oct->oq[oq_no]; + u32 time_threshold = 0; + u64 oq_ctl = ULL(0); + u64 reg_val; + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_CONTROL(oq_no)); + + /* wait for IDLE to set to 1 */ + if (!(reg_val & CNXK_VF_R_OUT_CTL_IDLE)) { + do { + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_CONTROL(oq_no)); + } while (!(reg_val & CNXK_VF_R_OUT_CTL_IDLE)); + } + + reg_val &= ~(CNXK_VF_R_OUT_CTL_IMODE); + reg_val &= ~(CNXK_VF_R_OUT_CTL_ROR_P); + reg_val &= ~(CNXK_VF_R_OUT_CTL_NSR_P); + reg_val &= ~(CNXK_VF_R_OUT_CTL_ROR_I); + reg_val &= ~(CNXK_VF_R_OUT_CTL_NSR_I); + reg_val &= ~(CNXK_VF_R_OUT_CTL_ES_I); + reg_val &= ~(CNXK_VF_R_OUT_CTL_ROR_D); + reg_val &= ~(CNXK_VF_R_OUT_CTL_NSR_D); + reg_val &= ~(CNXK_VF_R_OUT_CTL_ES_D); + reg_val |= (CNXK_VF_R_OUT_CTL_ES_P); + + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_CONTROL(oq_no), reg_val); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_BADDR(oq_no), oq->desc_ring_dma); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_RSIZE(oq_no), oq->max_count); + + oq_ctl = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_CONTROL(oq_no)); + /* Clear the ISIZE and BSIZE (22-0) */ + oq_ctl &= ~GENMASK_ULL(22, 0); + /* Populate the BSIZE (15-0) */ + oq_ctl |= (oq->buffer_size & GENMASK_ULL(15, 0)); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_CONTROL(oq_no), oq_ctl); + + /* Get the mapped address of the pkt_sent and pkts_credit regs */ + oq->pkts_sent_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_OUT_CNTS(oq_no); + oq->pkts_credit_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_OUT_SLIST_DBELL(oq_no); + + time_threshold = CFG_GET_OQ_INTR_TIME(oct->conf); + reg_val = ((u64)time_threshold << 32) | CFG_GET_OQ_INTR_PKT(oct->conf); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(oq_no), reg_val); + + /* set watermark for backpressure */ + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_WMARK(oq_no)); + reg_val &= ~GENMASK_ULL(31, 0); + reg_val |= CFG_GET_OQ_WMARK(oct->conf); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_WMARK(oq_no), reg_val); } /* Setup registers for a VF mailbox */ static void octep_vf_setup_mbox_regs_cnxk(struct octep_vf_device *oct, int q_no) { + struct octep_vf_mbox *mbox = oct->mbox; + + /* PF to VF DATA reg. VF reads from this reg */ + mbox->mbox_read_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_MBOX_PF_VF_DATA(q_no); + + /* VF mbox interrupt reg */ + mbox->mbox_int_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_MBOX_PF_VF_INT(q_no); + + /* VF to PF DATA reg. VF writes into this reg */ + mbox->mbox_write_reg = oct->mmio.hw_addr + CNXK_VF_SDP_R_MBOX_VF_PF_DATA(q_no); +} + +/* Mailbox Interrupt handler */ +static void cnxk_handle_vf_mbox_intr(struct octep_vf_device *oct) +{ + if (oct->mbox) + schedule_work(&oct->mbox->wk.work); + else + dev_err(&oct->pdev->dev, "cannot schedule work on invalid mbox\n"); } /* Tx/Rx queue interrupt handler */ static irqreturn_t octep_vf_ioq_intr_handler_cnxk(void *data) { + struct octep_vf_ioq_vector *vector = data; + struct octep_vf_device *oct; + struct octep_vf_oq *oq; + u64 reg_val = ULL(0); + + oct = vector->octep_vf_dev; + oq = vector->oq; + /* Mailbox interrupt arrives along with interrupt of tx/rx ring pair 0 */ + if (oq->q_no == 0) { + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_MBOX_PF_VF_INT(0)); + if (reg_val & CNXK_VF_SDP_R_MBOX_PF_VF_INT_STATUS) { + cnxk_handle_vf_mbox_intr(oct); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_MBOX_PF_VF_INT(0), reg_val); + } + } + napi_schedule_irqoff(oq->napi); return IRQ_HANDLED; } /* Re-initialize Octeon hardware registers */ static void octep_vf_reinit_regs_cnxk(struct octep_vf_device *oct) { + u32 i; + + for (i = 0; i < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); i++) + oct->hw_ops.setup_iq_regs(oct, i); + + for (i = 0; i < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); i++) + oct->hw_ops.setup_oq_regs(oct, i); + + oct->hw_ops.enable_interrupts(oct); + oct->hw_ops.enable_io_queues(oct); + + for (i = 0; i < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); i++) + writel(oct->oq[i]->max_count, oct->oq[i]->pkts_credit_reg); } /* Enable all interrupts */ static void octep_vf_enable_interrupts_cnxk(struct octep_vf_device *oct) { + int num_rings, q; + u64 reg_val; + + num_rings = CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); + for (q = 0; q < num_rings; q++) { + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(q)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(q), reg_val); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(q)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(q), reg_val); + } + /* Enable PF to VF mbox interrupt by setting 2nd bit*/ + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_MBOX_PF_VF_INT(0), + CNXK_VF_SDP_R_MBOX_PF_VF_INT_ENAB); } /* Disable all interrupts */ static void octep_vf_disable_interrupts_cnxk(struct octep_vf_device *oct) { + int num_rings, q; + u64 reg_val; + + /* Disable PF to VF mbox interrupt by setting 2nd bit*/ + if (oct->mbox) + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_MBOX_PF_VF_INT(0), 0x0); + + num_rings = CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); + for (q = 0; q < num_rings; q++) { + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(q)); + reg_val &= ~BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(q), reg_val); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(q)); + reg_val &= ~BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(q), reg_val); + } } /* Get new Octeon Read Index: index of descriptor that Octeon reads next. */ static u32 octep_vf_update_iq_read_index_cnxk(struct octep_vf_iq *iq) { - return 0; + u32 pkt_in_done = readl(iq->inst_cnt_reg); + u32 last_done, new_idx; + + last_done = pkt_in_done - iq->pkt_in_done; + iq->pkt_in_done = pkt_in_done; + + new_idx = (iq->octep_vf_read_index + last_done) % iq->max_count; + + return new_idx; } /* Enable a hardware Tx Queue */ static void octep_vf_enable_iq_cnxk(struct octep_vf_device *oct, int iq_no) { + u64 loop = HZ; + u64 reg_val; + + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_DBELL(iq_no), GENMASK_ULL(31, 0)); + + while (octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INSTR_DBELL(iq_no)) && + loop--) { + schedule_timeout_interruptible(1); + } + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(iq_no)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_INT_LEVELS(iq_no), reg_val); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_ENABLE(iq_no)); + reg_val |= ULL(1); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_ENABLE(iq_no), reg_val); } /* Enable a hardware Rx Queue */ static void octep_vf_enable_oq_cnxk(struct octep_vf_device *oct, int oq_no) { + u64 reg_val = ULL(0); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(oq_no)); + reg_val |= BIT_ULL_MASK(62); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_INT_LEVELS(oq_no), reg_val); + + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_SLIST_DBELL(oq_no), GENMASK_ULL(31, 0)); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_ENABLE(oq_no)); + reg_val |= ULL(1); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_ENABLE(oq_no), reg_val); } /* Enable all hardware Tx/Rx Queues assigned to VF */ static void octep_vf_enable_io_queues_cnxk(struct octep_vf_device *oct) { + u8 q; + + for (q = 0; q < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); q++) { + octep_vf_enable_iq_cnxk(oct, q); + octep_vf_enable_oq_cnxk(oct, q); + } } /* Disable a hardware Tx Queue assigned to VF */ static void octep_vf_disable_iq_cnxk(struct octep_vf_device *oct, int iq_no) { + u64 reg_val = ULL(0); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_IN_ENABLE(iq_no)); + reg_val &= ~ULL(1); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_IN_ENABLE(iq_no), reg_val); } /* Disable a hardware Rx Queue assigned to VF */ static void octep_vf_disable_oq_cnxk(struct octep_vf_device *oct, int oq_no) { + u64 reg_val = ULL(0); + + reg_val = octep_vf_read_csr64(oct, CNXK_VF_SDP_R_OUT_ENABLE(oq_no)); + reg_val &= ~ULL(1); + octep_vf_write_csr64(oct, CNXK_VF_SDP_R_OUT_ENABLE(oq_no), reg_val); } /* Disable all hardware Tx/Rx Queues assigned to VF */ static void octep_vf_disable_io_queues_cnxk(struct octep_vf_device *oct) { + int q = 0; + + for (q = 0; q < CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); q++) { + octep_vf_disable_iq_cnxk(oct, q); + octep_vf_disable_oq_cnxk(oct, q); + } } /* Dump hardware registers (including Tx/Rx queues) for debugging. */ static void octep_vf_dump_registers_cnxk(struct octep_vf_device *oct) { + u8 num_rings, q; + + num_rings = CFG_GET_PORTS_ACTIVE_IO_RINGS(oct->conf); + for (q = 0; q < num_rings; q++) + cnxk_vf_dump_q_regs(oct, q); } /**