diff options
author | Adit Ranadive <aditr@vmware.com> | 2016-10-02 19:10:22 -0700 |
---|---|---|
committer | Doug Ledford <dledford@redhat.com> | 2016-12-14 14:55:10 -0500 |
commit | 29c8d9eba550c6d73d17cc1618a9f5f2a7345aa1 (patch) | |
tree | e8cef3b3035f1f3c4ad12f0a25ec57b6c789bcab /drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c | |
parent | b1226c7db1d997fa6955cd3b54ba333bd0d8a29c (diff) |
IB: Add vmw_pvrdma driver
This patch series adds a driver for a paravirtual RDMA device. The
device is developed for VMware's Virtual Machines and allows existing RDMA
applications to continue to use existing Verbs API when deployed in VMs
on ESXi. We recently did a presentation in the OFA Workshop [1] regarding
this device.
Description and RDMA Support
============================
The virtual device is exposed as a dual function PCIe device. One part
is a virtual network device (VMXNet3) which provides networking properties
like MAC, IP addresses to the RDMA part of the device. The networking
properties are used to register GIDs required by RDMA applications to
communicate.
These patches add support and the all required infrastructure for
letting applications use such a device. We support the mandatory Verbs API as
well as the base memory management extensions (Local Inv, Send with Inv and
Fast Register Work Requests). We currently support both Reliable Connected
and Unreliable Datagram QPs but do not support Shared Receive Queues
(SRQs).
Also, we support the following types of Work Requests:
o Send/Receive (with or without Immediate Data)
o RDMA Write (with or without Immediate Data)
o RDMA Read
o Local Invalidate
o Send with Invalidate
o Fast Register Work Requests
This version only adds support for version 1 of RoCE. We will add RoCEv2
support in a future patch. We do support registration of both MAC-based
and IP-based GIDs. I have also created a git tree for our user-level driver
[2].
Testing
=======
We have tested this internally for various types of Guest OS - Red Hat,
Centos, Ubuntu 12.04/14.04/16.04, Oracle Enterprise Linux, SLES 12
using backported versions of this driver. The tests included several
runs of the performance tests (included with OFED), Intel MPI PingPong
benchmark on OpenMPI, krping for FRWRs. Mellanox has been kind enough
to test the backported version of the driver internally on their hardware
using a VMware provided ESX build. I have also applied and tested this
with Doug's k.o/for-4.9 branch (commit 5603910b). Note, that this patch
series should be applied all together. I split out the commits so that
it may be easier to review.
PVRDMA Resources
================
[1] OFA Workshop Presentation -
https://openfabrics.org/images/eventpresos/2016presentations/102parardma.pdf
[2] Libpvrdma User-level library -
http://git.openfabrics.org/?p=~aditr/libpvrdma.git;a=summary
Reviewed-by: Jorgen Hansen <jhansen@vmware.com>
Reviewed-by: George Zhang <georgezhang@vmware.com>
Reviewed-by: Aditya Sarwade <asarwade@vmware.com>
Reviewed-by: Bryan Tan <bryantan@vmware.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Adit Ranadive <aditr@vmware.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Diffstat (limited to 'drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c')
-rw-r--r-- | drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c | 425 |
1 files changed, 425 insertions, 0 deletions
diff --git a/drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c new file mode 100644 index 000000000000..e429ca5b16aa --- /dev/null +++ b/drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c @@ -0,0 +1,425 @@ +/* + * Copyright (c) 2012-2016 VMware, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of EITHER the GNU General Public License + * version 2 as published by the Free Software Foundation or the BSD + * 2-Clause License. This program is distributed in the hope that it + * will be useful, but WITHOUT ANY WARRANTY; WITHOUT EVEN THE IMPLIED + * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. + * See the GNU General Public License version 2 for more details at + * http://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html. + * + * You should have received a copy of the GNU General Public License + * along with this program available in the file COPYING in the main + * directory of this source tree. + * + * The BSD 2-Clause License + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS + * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE + * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, + * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES + * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR + * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED + * OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include <asm/page.h> +#include <linux/io.h> +#include <linux/wait.h> +#include <rdma/ib_addr.h> +#include <rdma/ib_smi.h> +#include <rdma/ib_user_verbs.h> + +#include "pvrdma.h" + +/** + * pvrdma_req_notify_cq - request notification for a completion queue + * @ibcq: the completion queue + * @notify_flags: notification flags + * + * @return: 0 for success. + */ +int pvrdma_req_notify_cq(struct ib_cq *ibcq, + enum ib_cq_notify_flags notify_flags) +{ + struct pvrdma_dev *dev = to_vdev(ibcq->device); + struct pvrdma_cq *cq = to_vcq(ibcq); + u32 val = cq->cq_handle; + + val |= (notify_flags & IB_CQ_SOLICITED_MASK) == IB_CQ_SOLICITED ? + PVRDMA_UAR_CQ_ARM_SOL : PVRDMA_UAR_CQ_ARM; + + pvrdma_write_uar_cq(dev, val); + + return 0; +} + +/** + * pvrdma_create_cq - create completion queue + * @ibdev: the device + * @attr: completion queue attributes + * @context: user context + * @udata: user data + * + * @return: ib_cq completion queue pointer on success, + * otherwise returns negative errno. + */ +struct ib_cq *pvrdma_create_cq(struct ib_device *ibdev, + const struct ib_cq_init_attr *attr, + struct ib_ucontext *context, + struct ib_udata *udata) +{ + int entries = attr->cqe; + struct pvrdma_dev *dev = to_vdev(ibdev); + struct pvrdma_cq *cq; + int ret; + int npages; + unsigned long flags; + union pvrdma_cmd_req req; + union pvrdma_cmd_resp rsp; + struct pvrdma_cmd_create_cq *cmd = &req.create_cq; + struct pvrdma_cmd_create_cq_resp *resp = &rsp.create_cq_resp; + struct pvrdma_create_cq ucmd; + + BUILD_BUG_ON(sizeof(struct pvrdma_cqe) != 64); + + entries = roundup_pow_of_two(entries); + if (entries < 1 || entries > dev->dsr->caps.max_cqe) + return ERR_PTR(-EINVAL); + + if (!atomic_add_unless(&dev->num_cqs, 1, dev->dsr->caps.max_cq)) + return ERR_PTR(-ENOMEM); + + cq = kzalloc(sizeof(*cq), GFP_KERNEL); + if (!cq) { + atomic_dec(&dev->num_cqs); + return ERR_PTR(-ENOMEM); + } + + cq->ibcq.cqe = entries; + + if (context) { + if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) { + ret = -EFAULT; + goto err_cq; + } + + cq->umem = ib_umem_get(context, ucmd.buf_addr, ucmd.buf_size, + IB_ACCESS_LOCAL_WRITE, 1); + if (IS_ERR(cq->umem)) { + ret = PTR_ERR(cq->umem); + goto err_cq; + } + + npages = ib_umem_page_count(cq->umem); + } else { + cq->is_kernel = true; + + /* One extra page for shared ring state */ + npages = 1 + (entries * sizeof(struct pvrdma_cqe) + + PAGE_SIZE - 1) / PAGE_SIZE; + + /* Skip header page. */ + cq->offset = PAGE_SIZE; + } + + if (npages < 0 || npages > PVRDMA_PAGE_DIR_MAX_PAGES) { + dev_warn(&dev->pdev->dev, + "overflow pages in completion queue\n"); + ret = -EINVAL; + goto err_umem; + } + + ret = pvrdma_page_dir_init(dev, &cq->pdir, npages, cq->is_kernel); + if (ret) { + dev_warn(&dev->pdev->dev, + "could not allocate page directory\n"); + goto err_umem; + } + + /* Ring state is always the first page. Set in library for user cq. */ + if (cq->is_kernel) + cq->ring_state = cq->pdir.pages[0]; + else + pvrdma_page_dir_insert_umem(&cq->pdir, cq->umem, 0); + + atomic_set(&cq->refcnt, 1); + init_waitqueue_head(&cq->wait); + spin_lock_init(&cq->cq_lock); + + memset(cmd, 0, sizeof(*cmd)); + cmd->hdr.cmd = PVRDMA_CMD_CREATE_CQ; + cmd->nchunks = npages; + cmd->ctx_handle = (context) ? + (u64)to_vucontext(context)->ctx_handle : 0; + cmd->cqe = entries; + cmd->pdir_dma = cq->pdir.dir_dma; + ret = pvrdma_cmd_post(dev, &req, &rsp, PVRDMA_CMD_CREATE_CQ_RESP); + if (ret < 0) { + dev_warn(&dev->pdev->dev, + "could not create completion queue, error: %d\n", ret); + goto err_page_dir; + } + + cq->ibcq.cqe = resp->cqe; + cq->cq_handle = resp->cq_handle; + spin_lock_irqsave(&dev->cq_tbl_lock, flags); + dev->cq_tbl[cq->cq_handle % dev->dsr->caps.max_cq] = cq; + spin_unlock_irqrestore(&dev->cq_tbl_lock, flags); + + if (context) { + cq->uar = &(to_vucontext(context)->uar); + + /* Copy udata back. */ + if (ib_copy_to_udata(udata, &cq->cq_handle, sizeof(__u32))) { + dev_warn(&dev->pdev->dev, + "failed to copy back udata\n"); + pvrdma_destroy_cq(&cq->ibcq); + return ERR_PTR(-EINVAL); + } + } + + return &cq->ibcq; + +err_page_dir: + pvrdma_page_dir_cleanup(dev, &cq->pdir); +err_umem: + if (context) + ib_umem_release(cq->umem); +err_cq: + atomic_dec(&dev->num_cqs); + kfree(cq); + + return ERR_PTR(ret); +} + +static void pvrdma_free_cq(struct pvrdma_dev *dev, struct pvrdma_cq *cq) +{ + atomic_dec(&cq->refcnt); + wait_event(cq->wait, !atomic_read(&cq->refcnt)); + + if (!cq->is_kernel) + ib_umem_release(cq->umem); + + pvrdma_page_dir_cleanup(dev, &cq->pdir); + kfree(cq); +} + +/** + * pvrdma_destroy_cq - destroy completion queue + * @cq: the completion queue to destroy. + * + * @return: 0 for success. + */ +int pvrdma_destroy_cq(struct ib_cq *cq) +{ + struct pvrdma_cq *vcq = to_vcq(cq); + union pvrdma_cmd_req req; + struct pvrdma_cmd_destroy_cq *cmd = &req.destroy_cq; + struct pvrdma_dev *dev = to_vdev(cq->device); + unsigned long flags; + int ret; + + memset(cmd, 0, sizeof(*cmd)); + cmd->hdr.cmd = PVRDMA_CMD_DESTROY_CQ; + cmd->cq_handle = vcq->cq_handle; + + ret = pvrdma_cmd_post(dev, &req, NULL, 0); + if (ret < 0) + dev_warn(&dev->pdev->dev, + "could not destroy completion queue, error: %d\n", + ret); + + /* free cq's resources */ + spin_lock_irqsave(&dev->cq_tbl_lock, flags); + dev->cq_tbl[vcq->cq_handle] = NULL; + spin_unlock_irqrestore(&dev->cq_tbl_lock, flags); + + pvrdma_free_cq(dev, vcq); + atomic_dec(&dev->num_cqs); + + return ret; +} + +/** + * pvrdma_modify_cq - modify the CQ moderation parameters + * @ibcq: the CQ to modify + * @cq_count: number of CQEs that will trigger an event + * @cq_period: max period of time in usec before triggering an event + * + * @return: -EOPNOTSUPP as CQ resize is not supported. + */ +int pvrdma_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period) +{ + return -EOPNOTSUPP; +} + +static inline struct pvrdma_cqe *get_cqe(struct pvrdma_cq *cq, int i) +{ + return (struct pvrdma_cqe *)pvrdma_page_dir_get_ptr( + &cq->pdir, + cq->offset + + sizeof(struct pvrdma_cqe) * i); +} + +void _pvrdma_flush_cqe(struct pvrdma_qp *qp, struct pvrdma_cq *cq) +{ + int head; + int has_data; + + if (!cq->is_kernel) + return; + + /* Lock held */ + has_data = pvrdma_idx_ring_has_data(&cq->ring_state->rx, + cq->ibcq.cqe, &head); + if (unlikely(has_data > 0)) { + int items; + int curr; + int tail = pvrdma_idx(&cq->ring_state->rx.prod_tail, + cq->ibcq.cqe); + struct pvrdma_cqe *cqe; + struct pvrdma_cqe *curr_cqe; + + items = (tail > head) ? (tail - head) : + (cq->ibcq.cqe - head + tail); + curr = --tail; + while (items-- > 0) { + if (curr < 0) + curr = cq->ibcq.cqe - 1; + if (tail < 0) + tail = cq->ibcq.cqe - 1; + curr_cqe = get_cqe(cq, curr); + if ((curr_cqe->qp & 0xFFFF) != qp->qp_handle) { + if (curr != tail) { + cqe = get_cqe(cq, tail); + *cqe = *curr_cqe; + } + tail--; + } else { + pvrdma_idx_ring_inc( + &cq->ring_state->rx.cons_head, + cq->ibcq.cqe); + } + curr--; + } + } +} + +static int pvrdma_poll_one(struct pvrdma_cq *cq, struct pvrdma_qp **cur_qp, + struct ib_wc *wc) +{ + struct pvrdma_dev *dev = to_vdev(cq->ibcq.device); + int has_data; + unsigned int head; + bool tried = false; + struct pvrdma_cqe *cqe; + +retry: + has_data = pvrdma_idx_ring_has_data(&cq->ring_state->rx, + cq->ibcq.cqe, &head); + if (has_data == 0) { + if (tried) + return -EAGAIN; + + pvrdma_write_uar_cq(dev, cq->cq_handle | PVRDMA_UAR_CQ_POLL); + + tried = true; + goto retry; + } else if (has_data == PVRDMA_INVALID_IDX) { + dev_err(&dev->pdev->dev, "CQ ring state invalid\n"); + return -EAGAIN; + } + + cqe = get_cqe(cq, head); + + /* Ensure cqe is valid. */ + rmb(); + if (dev->qp_tbl[cqe->qp & 0xffff]) + *cur_qp = (struct pvrdma_qp *)dev->qp_tbl[cqe->qp & 0xffff]; + else + return -EAGAIN; + + wc->opcode = pvrdma_wc_opcode_to_ib(cqe->opcode); + wc->status = pvrdma_wc_status_to_ib(cqe->status); + wc->wr_id = cqe->wr_id; + wc->qp = &(*cur_qp)->ibqp; + wc->byte_len = cqe->byte_len; + wc->ex.imm_data = cqe->imm_data; + wc->src_qp = cqe->src_qp; + wc->wc_flags = pvrdma_wc_flags_to_ib(cqe->wc_flags); + wc->pkey_index = cqe->pkey_index; + wc->slid = cqe->slid; + wc->sl = cqe->sl; + wc->dlid_path_bits = cqe->dlid_path_bits; + wc->port_num = cqe->port_num; + wc->vendor_err = 0; + + /* Update shared ring state */ + pvrdma_idx_ring_inc(&cq->ring_state->rx.cons_head, cq->ibcq.cqe); + + return 0; +} + +/** + * pvrdma_poll_cq - poll for work completion queue entries + * @ibcq: completion queue + * @num_entries: the maximum number of entries + * @entry: pointer to work completion array + * + * @return: number of polled completion entries + */ +int pvrdma_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) +{ + struct pvrdma_cq *cq = to_vcq(ibcq); + struct pvrdma_qp *cur_qp = NULL; + unsigned long flags; + int npolled; + + if (num_entries < 1 || wc == NULL) + return 0; + + spin_lock_irqsave(&cq->cq_lock, flags); + for (npolled = 0; npolled < num_entries; ++npolled) { + if (pvrdma_poll_one(cq, &cur_qp, wc + npolled)) + break; + } + + spin_unlock_irqrestore(&cq->cq_lock, flags); + + /* Ensure we do not return errors from poll_cq */ + return npolled; +} + +/** + * pvrdma_resize_cq - resize CQ + * @ibcq: the completion queue + * @entries: CQ entries + * @udata: user data + * + * @return: -EOPNOTSUPP as CQ resize is not supported. + */ +int pvrdma_resize_cq(struct ib_cq *ibcq, int entries, struct ib_udata *udata) +{ + return -EOPNOTSUPP; +} |