From patchwork Tue Apr 25 19:29:52 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Selvin Xavier X-Patchwork-Id: 13223700 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 9C261C77B61 for ; Tue, 25 Apr 2023 19:30:30 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235500AbjDYTa3 (ORCPT ); Tue, 25 Apr 2023 15:30:29 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:50674 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235496AbjDYTa2 (ORCPT ); Tue, 25 Apr 2023 15:30:28 -0400 Received: from mail-pl1-x636.google.com (mail-pl1-x636.google.com [IPv6:2607:f8b0:4864:20::636]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 900542D59 for ; Tue, 25 Apr 2023 12:30:25 -0700 (PDT) Received: by mail-pl1-x636.google.com with SMTP id d9443c01a7336-1a98ac97d47so12683915ad.3 for ; Tue, 25 Apr 2023 12:30:25 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=broadcom.com; s=google; t=1682451025; x=1685043025; h=references:in-reply-to:message-id:date:subject:cc:to:from:from:to :cc:subject:date:message-id:reply-to; bh=ttA1BRffw5dtAL9Q3IwCApq0n2JZ4bFvUS2bIAkfFqE=; b=Ld4i1mmP7HKv+WA7QbBDIXcfeypSVe9nAMHtlBi2s4d+Qguw1c7h5VMdLrtRz1jRO5 kFvgyhf7lsXFTPQcmkQaBXuXRGXqeCbd4uvHGKNTp+EVW4m/or/hY+YXaXVYb/R0xop1 oGycSGOwHo/rACqHxZej2Xi5SzdOh0bemLFXw= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1682451025; x=1685043025; h=references:in-reply-to:message-id:date:subject:cc:to:from :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=ttA1BRffw5dtAL9Q3IwCApq0n2JZ4bFvUS2bIAkfFqE=; b=iNEw3FEiLfJIa59t57DcBaZUeXc1o6tfznIXhTXYtm9o2nFUHkLLwDq+s12lNhIOXB OOf3kVFXs5sokgNFUM/6+GYj6AZclxWmV7Jvs9M8anT476fYYsmVDc+SPwkORm3iIP1h 09FJvvHbJMPTpV1kqM37bKGfOyCJwX2Hlh4UNOqZzwmpiliZ9xoa0BaNZa2Wrqt4Xnyk 8zgnpr3+py2uUBf5dhZ4IKIuAXIpYkGHkXeBHzk+rNCDBNGD8uEZNwxLRk9nxbvKROVS 4pcsnooocdRjlCr9ZwZbsyeHQMJGx8Qvm8+249uT4xLy4GaPzqvCepqFyQH98Bvn3u6U Gi5g== X-Gm-Message-State: AAQBX9eWMySB6JRL2M68w7PfdqRgH94ucDAwhBagP1eyP/m+wtNFAU+n 7icn4cuwTU8jzooOqU8tgaAA3g== X-Google-Smtp-Source: AKy350brnhYiAVrio8bkNxh65wzPKETnq7mEcyt1G2vj9thUxlZqEq4fj3/5Cy+OmNKafy7aQ/8Gng== X-Received: by 2002:a17:902:d486:b0:1a9:5aef:1aea with SMTP id c6-20020a170902d48600b001a95aef1aeamr15589655plg.66.1682451024685; Tue, 25 Apr 2023 12:30:24 -0700 (PDT) Received: from dhcp-10-192-206-197.iig.avagotech.net.net ([192.19.234.250]) by smtp.gmail.com with ESMTPSA id w3-20020a170902c78300b001a64ed4621csm8618217pla.182.2023.04.25.12.30.22 (version=TLS1_2 cipher=ECDHE-ECDSA-AES128-GCM-SHA256 bits=128/128); Tue, 25 Apr 2023 12:30:23 -0700 (PDT) From: Selvin Xavier To: jgg@ziepe.ca, leon@kernel.org Cc: linux-rdma@vger.kernel.org, andrew.gospodarek@broadcom.com, Selvin Xavier Subject: [PATCH v2 for-next 5/6] RDMA/bnxt_re: Reorg the bar mapping Date: Tue, 25 Apr 2023 12:29:52 -0700 Message-Id: <1682450993-17711-6-git-send-email-selvin.xavier@broadcom.com> X-Mailer: git-send-email 2.5.5 In-Reply-To: <1682450993-17711-1-git-send-email-selvin.xavier@broadcom.com> References: <1682450993-17711-1-git-send-email-selvin.xavier@broadcom.com> Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Reorganize the code for allocation and mapping of Doorbell pages. Implements new HW command to get the BAR length used by L2 driver. These changes are used by the future patch which maps the WC Doorbell pages. Also, introduced a new lock dpi_tbl_lock for synchronize the DB page allocation from users. Signed-off-by: Selvin Xavier --- v1 - v2: - Few generic cleanup based on internal review drivers/infiniband/hw/bnxt_re/ib_verbs.c | 7 +- drivers/infiniband/hw/bnxt_re/main.c | 71 +++++++++++- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 4 +- drivers/infiniband/hw/bnxt_re/qplib_rcfw.h | 1 + drivers/infiniband/hw/bnxt_re/qplib_res.c | 174 +++++++++++++++++++---------- drivers/infiniband/hw/bnxt_re/qplib_res.h | 30 +++-- drivers/infiniband/hw/bnxt_re/qplib_sp.c | 3 + drivers/infiniband/hw/bnxt_re/qplib_sp.h | 1 + 8 files changed, 213 insertions(+), 78 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index 2434174..bc85433 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -606,8 +606,8 @@ int bnxt_re_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata) * ibv_devinfo and family of application when DPIs * are depleted. */ - if (bnxt_qplib_alloc_dpi(&rdev->qplib_res.dpi_tbl, - &ucntx->dpi, ucntx)) { + if (bnxt_qplib_alloc_dpi(&rdev->qplib_res, + &ucntx->dpi, ucntx, BNXT_QPLIB_DPI_TYPE_UC)) { rc = -ENOMEM; goto dbfail; } @@ -4075,8 +4075,7 @@ void bnxt_re_dealloc_ucontext(struct ib_ucontext *ib_uctx) /* Free DPI only if this is the first PD allocated by the * application and mark the context dpi as NULL */ - bnxt_qplib_dealloc_dpi(&rdev->qplib_res, - &rdev->qplib_res.dpi_tbl, &uctx->dpi); + bnxt_qplib_dealloc_dpi(&rdev->qplib_res, &uctx->dpi); uctx->dpi.dbr = NULL; } } diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c index 873ee75..913cb80 100644 --- a/drivers/infiniband/hw/bnxt_re/main.c +++ b/drivers/infiniband/hw/bnxt_re/main.c @@ -85,6 +85,40 @@ static struct bnxt_re_dev *bnxt_re_from_netdev(struct net_device *netdev); static void bnxt_re_dev_uninit(struct bnxt_re_dev *rdev); static int bnxt_re_hwrm_qcaps(struct bnxt_re_dev *rdev); +static int bnxt_re_hwrm_qcfg(struct bnxt_re_dev *rdev, u32 *db_len, + u32 *offset); +static void bnxt_re_set_db_offset(struct bnxt_re_dev *rdev) +{ + struct bnxt_qplib_chip_ctx *cctx; + struct bnxt_en_dev *en_dev; + struct bnxt_qplib_res *res; + u32 l2db_len = 0; + u32 offset = 0; + u32 barlen; + int rc; + + res = &rdev->qplib_res; + en_dev = rdev->en_dev; + cctx = rdev->chip_ctx; + + /* Issue qcfg */ + rc = bnxt_re_hwrm_qcfg(rdev, &l2db_len, &offset); + if (rc) + dev_info(rdev_to_dev(rdev), + "Couldn't get DB bar size, Low latency framework is disabled\n"); + /* set register offsets for both UC and WC */ + res->dpi_tbl.ucreg.offset = res->is_vf ? BNXT_QPLIB_DBR_VF_DB_OFFSET : + BNXT_QPLIB_DBR_PF_DB_OFFSET; + res->dpi_tbl.wcreg.offset = res->dpi_tbl.ucreg.offset; + + /* If WC mapping is disabled by L2 driver then en_dev->l2_db_size + * is equal to the DB-Bar actual size. This indicates that L2 + * is mapping entire bar as UC-. RoCE driver can't enable WC mapping + * in such cases and DB-push will be disabled. + */ + barlen = pci_resource_len(res->pdev, RCFW_DBR_PCI_BAR_REGION); +} + static void bnxt_re_set_drv_mode(struct bnxt_re_dev *rdev, u8 mode) { struct bnxt_qplib_chip_ctx *cctx; @@ -116,6 +150,7 @@ static int bnxt_re_setup_chip_ctx(struct bnxt_re_dev *rdev, u8 wqe_mode) { struct bnxt_qplib_chip_ctx *chip_ctx; struct bnxt_en_dev *en_dev; + int rc; en_dev = rdev->en_dev; @@ -134,6 +169,12 @@ static int bnxt_re_setup_chip_ctx(struct bnxt_re_dev *rdev, u8 wqe_mode) rdev->qplib_res.is_vf = BNXT_EN_VF(en_dev); bnxt_re_set_drv_mode(rdev, wqe_mode); + + bnxt_re_set_db_offset(rdev); + rc = bnxt_qplib_map_db_bar(&rdev->qplib_res); + if (rc) + return rc; + if (bnxt_qplib_determine_atomics(en_dev->pdev)) ibdev_info(&rdev->ibdev, "platform doesn't support global atomics."); @@ -338,6 +379,30 @@ static void bnxt_re_fill_fw_msg(struct bnxt_fw_msg *fw_msg, void *msg, fw_msg->timeout = timeout; } +/* Query device config using common hwrm */ +static int bnxt_re_hwrm_qcfg(struct bnxt_re_dev *rdev, u32 *db_len, + u32 *offset) +{ + struct bnxt_en_dev *en_dev = rdev->en_dev; + struct hwrm_func_qcfg_output resp = {0}; + struct hwrm_func_qcfg_input req = {0}; + struct bnxt_fw_msg fw_msg; + int rc; + + memset(&fw_msg, 0, sizeof(fw_msg)); + bnxt_re_init_hwrm_hdr(rdev, (void *)&req, + HWRM_FUNC_QCFG, -1, -1); + req.fid = cpu_to_le16(0xffff); + bnxt_re_fill_fw_msg(&fw_msg, (void *)&req, sizeof(req), (void *)&resp, + sizeof(resp), DFLT_HWRM_CMD_TIMEOUT); + rc = bnxt_send_msg(en_dev, &fw_msg); + if (rc) { + *db_len = PAGE_ALIGN(le16_to_cpu(resp.l2_doorbell_bar_size_kb) * 1024); + *offset = PAGE_ALIGN(le16_to_cpu(resp.legacy_l2_db_size_kb) * 1024); + } + return rc; +} + /* Query function capabilities using common hwrm */ int bnxt_re_hwrm_qcaps(struct bnxt_re_dev *rdev) { @@ -851,7 +916,6 @@ static void bnxt_re_free_res(struct bnxt_re_dev *rdev) if (rdev->qplib_res.dpi_tbl.max) { bnxt_qplib_dealloc_dpi(&rdev->qplib_res, - &rdev->qplib_res.dpi_tbl, &rdev->dpi_privileged); } if (rdev->qplib_res.rcfw) { @@ -879,9 +943,9 @@ static int bnxt_re_alloc_res(struct bnxt_re_dev *rdev) if (rc) goto fail; - rc = bnxt_qplib_alloc_dpi(&rdev->qplib_res.dpi_tbl, + rc = bnxt_qplib_alloc_dpi(&rdev->qplib_res, &rdev->dpi_privileged, - rdev); + rdev, BNXT_QPLIB_DPI_TYPE_KERNEL); if (rc) goto dealloc_res; @@ -921,7 +985,6 @@ static int bnxt_re_alloc_res(struct bnxt_re_dev *rdev) bnxt_qplib_free_nq(&rdev->nq[i]); } bnxt_qplib_dealloc_dpi(&rdev->qplib_res, - &rdev->qplib_res.dpi_tbl, &rdev->dpi_privileged); dealloc_res: bnxt_qplib_free_res(&rdev->qplib_res); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index ade858f..719bdc6 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -655,7 +655,7 @@ int bnxt_qplib_create_srq(struct bnxt_qplib_res *res, srq->dbinfo.xid = srq->id; srq->dbinfo.db = srq->dpi->dbr; srq->dbinfo.max_slot = 1; - srq->dbinfo.priv_db = res->dpi_tbl.dbr_bar_reg_iomem; + srq->dbinfo.priv_db = res->dpi_tbl.priv_db; if (srq->threshold) bnxt_qplib_armen_db(&srq->dbinfo, DBC_DBC_TYPE_SRQ_ARMENA); srq->arm_req = false; @@ -2099,7 +2099,7 @@ int bnxt_qplib_create_cq(struct bnxt_qplib_res *res, struct bnxt_qplib_cq *cq) cq->dbinfo.hwq = &cq->hwq; cq->dbinfo.xid = cq->id; cq->dbinfo.db = cq->dpi->dbr; - cq->dbinfo.priv_db = res->dpi_tbl.dbr_bar_reg_iomem; + cq->dbinfo.priv_db = res->dpi_tbl.priv_db; bnxt_qplib_armen_db(&cq->dbinfo, DBC_DBC_TYPE_CQ_ARMENA); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h index dd56514..4d1ccf1 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.h @@ -111,6 +111,7 @@ static inline u32 bnxt_qplib_set_cmd_slots(struct cmdq_base *req) #define RCFW_BLOCKED_CMD_WAIT_COUNT 20000000UL /* 20 sec */ #define HWRM_VERSION_RCFW_CMDQ_DEPTH_CHECK 0x1000900020011ULL +#define HWRM_VERSION_DEV_ATTR_MAX_DPI 0x1000A0000000DULL /* Crsq buf is 1024-Byte */ struct bnxt_qplib_crsbe { diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.c b/drivers/infiniband/hw/bnxt_re/qplib_res.c index 920ab87..e1cbe59 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_res.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_res.c @@ -704,44 +704,73 @@ static int bnxt_qplib_alloc_pd_tbl(struct bnxt_qplib_res *res, } /* DPIs */ -int bnxt_qplib_alloc_dpi(struct bnxt_qplib_dpi_tbl *dpit, - struct bnxt_qplib_dpi *dpi, - void *app) +int bnxt_qplib_alloc_dpi(struct bnxt_qplib_res *res, + struct bnxt_qplib_dpi *dpi, + void *app, u8 type) { + struct bnxt_qplib_dpi_tbl *dpit = &res->dpi_tbl; + struct bnxt_qplib_reg_desc *reg; u32 bit_num; + u64 umaddr; + + reg = &dpit->wcreg; + mutex_lock(&res->dpi_tbl_lock); bit_num = find_first_bit(dpit->tbl, dpit->max); - if (bit_num == dpit->max) + if (bit_num == dpit->max) { + mutex_unlock(&res->dpi_tbl_lock); return -ENOMEM; + } /* Found unused DPI */ clear_bit(bit_num, dpit->tbl); dpit->app_tbl[bit_num] = app; - dpi->dpi = bit_num; - dpi->dbr = dpit->dbr_bar_reg_iomem + (bit_num * PAGE_SIZE); - dpi->umdbr = dpit->unmapped_dbr + (bit_num * PAGE_SIZE); + dpi->bit = bit_num; + dpi->dpi = bit_num + (reg->offset - dpit->ucreg.offset) / PAGE_SIZE; + + umaddr = reg->bar_base + reg->offset + bit_num * PAGE_SIZE; + dpi->umdbr = umaddr; + + switch (type) { + case BNXT_QPLIB_DPI_TYPE_KERNEL: + /* priviledged dbr was already mapped just initialize it. */ + dpi->umdbr = dpit->ucreg.bar_base + + dpit->ucreg.offset + bit_num * PAGE_SIZE; + dpi->dbr = dpit->priv_db; + dpi->dpi = dpi->bit; + break; + default: + dpi->dbr = ioremap(umaddr, PAGE_SIZE); + break; + } + dpi->type = type; + mutex_unlock(&res->dpi_tbl_lock); return 0; + } int bnxt_qplib_dealloc_dpi(struct bnxt_qplib_res *res, - struct bnxt_qplib_dpi_tbl *dpit, - struct bnxt_qplib_dpi *dpi) + struct bnxt_qplib_dpi *dpi) { - if (dpi->dpi >= dpit->max) { - dev_warn(&res->pdev->dev, "Invalid DPI? dpi = %d\n", dpi->dpi); - return -EINVAL; - } - if (test_and_set_bit(dpi->dpi, dpit->tbl)) { - dev_warn(&res->pdev->dev, "Freeing an unused DPI? dpi = %d\n", - dpi->dpi); + struct bnxt_qplib_dpi_tbl *dpit = &res->dpi_tbl; + + mutex_lock(&res->dpi_tbl_lock); + if (dpi->dpi && dpi->type != BNXT_QPLIB_DPI_TYPE_KERNEL) + pci_iounmap(res->pdev, dpi->dbr); + + if (test_and_set_bit(dpi->bit, dpit->tbl)) { + dev_warn(&res->pdev->dev, + "Freeing an unused DPI? dpi = %d, bit = %d\n", + dpi->dpi, dpi->bit); + mutex_unlock(&res->dpi_tbl_lock); return -EINVAL; } if (dpit->app_tbl) - dpit->app_tbl[dpi->dpi] = NULL; + dpit->app_tbl[dpi->bit] = NULL; memset(dpi, 0, sizeof(*dpi)); - + mutex_unlock(&res->dpi_tbl_lock); return 0; } @@ -750,52 +779,38 @@ static void bnxt_qplib_free_dpi_tbl(struct bnxt_qplib_res *res, { kfree(dpit->tbl); kfree(dpit->app_tbl); - if (dpit->dbr_bar_reg_iomem) - pci_iounmap(res->pdev, dpit->dbr_bar_reg_iomem); - memset(dpit, 0, sizeof(*dpit)); + dpit->tbl = NULL; + dpit->app_tbl = NULL; + dpit->max = 0; } -static int bnxt_qplib_alloc_dpi_tbl(struct bnxt_qplib_res *res, - struct bnxt_qplib_dpi_tbl *dpit, - u32 dbr_offset) +static int bnxt_qplib_alloc_dpi_tbl(struct bnxt_qplib_res *res, + struct bnxt_qplib_dev_attr *dev_attr) { - u32 dbr_bar_reg = RCFW_DBR_PCI_BAR_REGION; - resource_size_t bar_reg_base; - u32 dbr_len, bytes; - - if (dpit->dbr_bar_reg_iomem) { - dev_err(&res->pdev->dev, "DBR BAR region %d already mapped\n", - dbr_bar_reg); - return -EALREADY; - } - - bar_reg_base = pci_resource_start(res->pdev, dbr_bar_reg); - if (!bar_reg_base) { - dev_err(&res->pdev->dev, "BAR region %d resc start failed\n", - dbr_bar_reg); - return -ENOMEM; - } + struct bnxt_qplib_dpi_tbl *dpit; + struct bnxt_qplib_reg_desc *reg; + unsigned long bar_len; + u32 dbr_offset; + u32 bytes; - dbr_len = pci_resource_len(res->pdev, dbr_bar_reg) - dbr_offset; - if (!dbr_len || ((dbr_len & (PAGE_SIZE - 1)) != 0)) { - dev_err(&res->pdev->dev, "Invalid DBR length %d\n", dbr_len); - return -ENOMEM; - } + dpit = &res->dpi_tbl; + reg = &dpit->wcreg; - dpit->dbr_bar_reg_iomem = ioremap(bar_reg_base + dbr_offset, - dbr_len); - if (!dpit->dbr_bar_reg_iomem) { - dev_err(&res->pdev->dev, - "FP: DBR BAR region %d mapping failed\n", dbr_bar_reg); - return -ENOMEM; + if (!bnxt_qplib_is_chip_gen_p5(res->cctx)) { + /* Offest should come from L2 driver */ + dbr_offset = dev_attr->l2_db_size; + dpit->ucreg.offset = dbr_offset; + dpit->wcreg.offset = dbr_offset; } - dpit->unmapped_dbr = bar_reg_base + dbr_offset; - dpit->max = dbr_len / PAGE_SIZE; + bar_len = pci_resource_len(res->pdev, reg->bar_id); + dpit->max = (bar_len - reg->offset) / PAGE_SIZE; + if (dev_attr->max_dpi) + dpit->max = min_t(u32, dpit->max, dev_attr->max_dpi); - dpit->app_tbl = kcalloc(dpit->max, sizeof(void *), GFP_KERNEL); + dpit->app_tbl = kcalloc(dpit->max, sizeof(void *), GFP_KERNEL); if (!dpit->app_tbl) - goto unmap_io; + return -ENOMEM; bytes = dpit->max >> 3; if (!bytes) @@ -805,17 +820,14 @@ static int bnxt_qplib_alloc_dpi_tbl(struct bnxt_qplib_res *res, if (!dpit->tbl) { kfree(dpit->app_tbl); dpit->app_tbl = NULL; - goto unmap_io; + return -ENOMEM; } memset((u8 *)dpit->tbl, 0xFF, bytes); + dpit->priv_db = dpit->ucreg.bar_reg + dpit->ucreg.offset; return 0; -unmap_io: - iounmap(dpit->dbr_bar_reg_iomem); - dpit->dbr_bar_reg_iomem = NULL; - return -ENOMEM; } /* Stats */ @@ -882,7 +894,7 @@ int bnxt_qplib_alloc_res(struct bnxt_qplib_res *res, struct pci_dev *pdev, if (rc) goto fail; - rc = bnxt_qplib_alloc_dpi_tbl(res, &res->dpi_tbl, dev_attr->l2_db_size); + rc = bnxt_qplib_alloc_dpi_tbl(res, dev_attr); if (rc) goto fail; @@ -892,6 +904,46 @@ int bnxt_qplib_alloc_res(struct bnxt_qplib_res *res, struct pci_dev *pdev, return rc; } +void bnxt_qplib_unmap_db_bar(struct bnxt_qplib_res *res) +{ + struct bnxt_qplib_reg_desc *reg; + + reg = &res->dpi_tbl.ucreg; + if (reg->bar_reg) + pci_iounmap(res->pdev, reg->bar_reg); + reg->bar_reg = NULL; + reg->bar_base = 0; + reg->len = 0; + reg->bar_id = 0; +} + +int bnxt_qplib_map_db_bar(struct bnxt_qplib_res *res) +{ + struct bnxt_qplib_reg_desc *ucreg; + struct bnxt_qplib_reg_desc *wcreg; + + wcreg = &res->dpi_tbl.wcreg; + wcreg->bar_id = RCFW_DBR_PCI_BAR_REGION; + wcreg->bar_base = pci_resource_start(res->pdev, wcreg->bar_id); + + ucreg = &res->dpi_tbl.ucreg; + ucreg->bar_id = RCFW_DBR_PCI_BAR_REGION; + ucreg->bar_base = pci_resource_start(res->pdev, ucreg->bar_id); + ucreg->len = ucreg->offset + PAGE_SIZE; + if (!ucreg->len || ((ucreg->len & (PAGE_SIZE - 1)) != 0)) { + dev_err(&res->pdev->dev, "QPLIB: invalid dbr length %d", + (int)ucreg->len); + return -EINVAL; + } + ucreg->bar_reg = ioremap(ucreg->bar_base, ucreg->len); + if (!ucreg->bar_reg) { + dev_err(&res->pdev->dev, "priviledged dpi map failed!"); + return -ENOMEM; + } + + return 0; +} + int bnxt_qplib_determine_atomics(struct pci_dev *dev) { int comp; diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.h b/drivers/infiniband/hw/bnxt_re/qplib_res.h index 982e2c9..95b1d6c 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_res.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_res.h @@ -56,8 +56,12 @@ struct bnxt_qplib_chip_ctx { u8 chip_metal; u16 hw_stats_size; struct bnxt_qplib_drv_modes modes; + u64 hwrm_intf_ver; }; +#define BNXT_QPLIB_DBR_PF_DB_OFFSET 0x10000 +#define BNXT_QPLIB_DBR_VF_DB_OFFSET 0x4000 + #define PTR_CNT_PER_PG (PAGE_SIZE / sizeof(void *)) #define PTR_MAX_IDX_PER_PG (PTR_CNT_PER_PG - 1) #define PTR_PG(x) (((x) & ~PTR_MAX_IDX_PER_PG) / PTR_CNT_PER_PG) @@ -109,6 +113,7 @@ enum bnxt_qplib_hwrm_pg_size { struct bnxt_qplib_reg_desc { u8 bar_id; resource_size_t bar_base; + unsigned long offset; void __iomem *bar_reg; size_t len; }; @@ -185,18 +190,26 @@ struct bnxt_qplib_sgid_tbl { u8 *vlan; }; +enum { + BNXT_QPLIB_DPI_TYPE_KERNEL = 0, + BNXT_QPLIB_DPI_TYPE_UC = 1, +}; + struct bnxt_qplib_dpi { u32 dpi; + u32 bit; void __iomem *dbr; u64 umdbr; + u8 type; }; struct bnxt_qplib_dpi_tbl { void **app_tbl; unsigned long *tbl; u16 max; - void __iomem *dbr_bar_reg_iomem; - u64 unmapped_dbr; + struct bnxt_qplib_reg_desc ucreg; /* Hold entire DB bar. */ + struct bnxt_qplib_reg_desc wcreg; + void __iomem *priv_db; }; struct bnxt_qplib_stats { @@ -241,7 +254,6 @@ struct bnxt_qplib_ctx { struct bnxt_qplib_tqm_ctx tqm_ctx; struct bnxt_qplib_stats stats; struct bnxt_qplib_vf_res vf_res; - u64 hwrm_intf_ver; }; struct bnxt_qplib_res { @@ -253,6 +265,8 @@ struct bnxt_qplib_res { struct bnxt_qplib_pd_tbl pd_tbl; struct bnxt_qplib_sgid_tbl sgid_tbl; struct bnxt_qplib_dpi_tbl dpi_tbl; + /* To protect the dpi table bit map */ + struct mutex dpi_tbl_lock; bool prio; bool is_vf; }; @@ -344,11 +358,10 @@ int bnxt_qplib_alloc_pd(struct bnxt_qplib_pd_tbl *pd_tbl, int bnxt_qplib_dealloc_pd(struct bnxt_qplib_res *res, struct bnxt_qplib_pd_tbl *pd_tbl, struct bnxt_qplib_pd *pd); -int bnxt_qplib_alloc_dpi(struct bnxt_qplib_dpi_tbl *dpit, - struct bnxt_qplib_dpi *dpi, - void *app); +int bnxt_qplib_alloc_dpi(struct bnxt_qplib_res *res, + struct bnxt_qplib_dpi *dpi, + void *app, u8 type); int bnxt_qplib_dealloc_dpi(struct bnxt_qplib_res *res, - struct bnxt_qplib_dpi_tbl *dpi_tbl, struct bnxt_qplib_dpi *dpi); void bnxt_qplib_cleanup_res(struct bnxt_qplib_res *res); int bnxt_qplib_init_res(struct bnxt_qplib_res *res); @@ -361,6 +374,9 @@ void bnxt_qplib_free_ctx(struct bnxt_qplib_res *res, int bnxt_qplib_alloc_ctx(struct bnxt_qplib_res *res, struct bnxt_qplib_ctx *ctx, bool virt_fn, bool is_p5); +int bnxt_qplib_map_db_bar(struct bnxt_qplib_res *res); +void bnxt_qplib_unmap_db_bar(struct bnxt_qplib_res *res); + int bnxt_qplib_determine_atomics(struct pci_dev *dev); static inline void bnxt_qplib_hwq_incr_prod(struct bnxt_qplib_hwq *hwq, u32 cnt) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.c b/drivers/infiniband/hw/bnxt_re/qplib_sp.c index 1714a1e..fe3a22d 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.c @@ -170,6 +170,9 @@ int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, attr->tqm_alloc_reqs[i * 4 + 3] = *(++tqm_alloc); } + if (rcfw->res->cctx->hwrm_intf_ver >= HWRM_VERSION_DEV_ATTR_MAX_DPI) + attr->max_dpi = le32_to_cpu(sb->max_dpi); + attr->is_atomic = bnxt_qplib_is_atomic_cap(rcfw); bail: bnxt_qplib_rcfw_free_sbuf(rcfw, sbuf); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.h b/drivers/infiniband/hw/bnxt_re/qplib_sp.h index 5de87465..0efd464 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.h @@ -72,6 +72,7 @@ struct bnxt_qplib_dev_attr { u8 tqm_alloc_reqs[MAX_TQM_ALLOC_REQ]; bool is_atomic; u16 dev_cap_flags; + u32 max_dpi; }; struct bnxt_qplib_pd {