diff options
Diffstat (limited to 'drivers/net/hyperv/netvsc.c')
| -rw-r--r-- | drivers/net/hyperv/netvsc.c | 2042 |
1 files changed, 1471 insertions, 571 deletions
diff --git a/drivers/net/hyperv/netvsc.c b/drivers/net/hyperv/netvsc.c index 2b0480416b31..60a4629fe6ba 100644 --- a/drivers/net/hyperv/netvsc.c +++ b/drivers/net/hyperv/netvsc.c @@ -1,19 +1,7 @@ +// SPDX-License-Identifier: GPL-2.0-only /* * Copyright (c) 2009, Microsoft Corporation. * - * This program is free software; you can redistribute it and/or modify it - * under the terms and conditions of the GNU General Public License, - * version 2, as published by the Free Software Foundation. - * - * This program is distributed in the hope 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 for - * more details. - * - * You should have received a copy of the GNU General Public License along with - * this program; if not, write to the Free Software Foundation, Inc., 59 Temple - * Place - Suite 330, Boston, MA 02111-1307 USA. - * * Authors: * Haiyang Zhang <haiyangz@microsoft.com> * Hank Janssen <hjanssen@microsoft.com> @@ -29,63 +17,170 @@ #include <linux/slab.h> #include <linux/netdevice.h> #include <linux/if_ether.h> +#include <linux/vmalloc.h> +#include <linux/rtnetlink.h> +#include <linux/prefetch.h> +#include <linux/filter.h> + +#include <asm/sync_bitops.h> +#include <asm/mshyperv.h> #include "hyperv_net.h" +#include "netvsc_trace.h" + +/* + * Switch the data path from the synthetic interface to the VF + * interface. + */ +int netvsc_switch_datapath(struct net_device *ndev, bool vf) +{ + struct net_device_context *net_device_ctx = netdev_priv(ndev); + struct hv_device *dev = net_device_ctx->device_ctx; + struct netvsc_device *nv_dev = rtnl_dereference(net_device_ctx->nvdev); + struct nvsp_message *init_pkt = &nv_dev->channel_init_pkt; + int ret, retry = 0; + + /* Block sending traffic to VF if it's about to be gone */ + if (!vf) + net_device_ctx->data_path_is_vf = vf; + + memset(init_pkt, 0, sizeof(struct nvsp_message)); + init_pkt->hdr.msg_type = NVSP_MSG4_TYPE_SWITCH_DATA_PATH; + if (vf) + init_pkt->msg.v4_msg.active_dp.active_datapath = + NVSP_DATAPATH_VF; + else + init_pkt->msg.v4_msg.active_dp.active_datapath = + NVSP_DATAPATH_SYNTHETIC; +again: + trace_nvsp_send(ndev, init_pkt); + + ret = vmbus_sendpacket(dev->channel, init_pkt, + sizeof(struct nvsp_message), + (unsigned long)init_pkt, VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + + /* If failed to switch to/from VF, let data_path_is_vf stay false, + * so we use synthetic path to send data. + */ + if (ret) { + if (ret != -EAGAIN) { + netdev_err(ndev, + "Unable to send sw datapath msg, err: %d\n", + ret); + return ret; + } -static struct netvsc_device *alloc_net_device(struct hv_device *device) + if (retry++ < RETRY_MAX) { + usleep_range(RETRY_US_LO, RETRY_US_HI); + goto again; + } else { + netdev_err( + ndev, + "Retry failed to send sw datapath msg, err: %d\n", + ret); + return ret; + } + } + + wait_for_completion(&nv_dev->channel_init_wait); + net_device_ctx->data_path_is_vf = vf; + + return 0; +} + +/* Worker to setup sub channels on initial setup + * Initial hotplug event occurs in softirq context + * and can't wait for channels. + */ +static void netvsc_subchan_work(struct work_struct *w) +{ + struct netvsc_device *nvdev = + container_of(w, struct netvsc_device, subchan_work); + struct rndis_device *rdev; + int i, ret; + + /* Avoid deadlock with device removal already under RTNL */ + if (!rtnl_trylock()) { + schedule_work(w); + return; + } + + rdev = nvdev->extension; + if (rdev) { + ret = rndis_set_subchannel(rdev->ndev, nvdev, NULL); + if (ret == 0) { + netif_device_attach(rdev->ndev); + } else { + /* fallback to only primary channel */ + for (i = 1; i < nvdev->num_chn; i++) + netif_napi_del(&nvdev->chan_table[i].napi); + + nvdev->max_chn = 1; + nvdev->num_chn = 1; + } + } + + rtnl_unlock(); +} + +static struct netvsc_device *alloc_net_device(void) { struct netvsc_device *net_device; - struct net_device *ndev = hv_get_drvdata(device); net_device = kzalloc(sizeof(struct netvsc_device), GFP_KERNEL); if (!net_device) return NULL; init_waitqueue_head(&net_device->wait_drain); - net_device->start_remove = false; net_device->destroy = false; - net_device->dev = device; - net_device->ndev = ndev; - - hv_set_drvdata(device, net_device); - return net_device; -} + net_device->tx_disable = true; -static struct netvsc_device *get_outbound_net_device(struct hv_device *device) -{ - struct netvsc_device *net_device; + net_device->max_pkt = RNDIS_MAX_PKT_DEFAULT; + net_device->pkt_align = RNDIS_PKT_ALIGN_DEFAULT; - net_device = hv_get_drvdata(device); - if (net_device && net_device->destroy) - net_device = NULL; + init_completion(&net_device->channel_init_wait); + init_waitqueue_head(&net_device->subchan_open); + INIT_WORK(&net_device->subchan_work, netvsc_subchan_work); return net_device; } -static struct netvsc_device *get_inbound_net_device(struct hv_device *device) +static void free_netvsc_device(struct rcu_head *head) { - struct netvsc_device *net_device; + struct netvsc_device *nvdev + = container_of(head, struct netvsc_device, rcu); + int i; - net_device = hv_get_drvdata(device); + kfree(nvdev->extension); - if (!net_device) - goto get_in_err; + if (!nvdev->recv_buf_gpadl_handle.decrypted) + vfree(nvdev->recv_buf); + if (!nvdev->send_buf_gpadl_handle.decrypted) + vfree(nvdev->send_buf); + bitmap_free(nvdev->send_section_map); - if (net_device->destroy && - atomic_read(&net_device->num_outstanding_sends) == 0) - net_device = NULL; + for (i = 0; i < VRSS_CHANNEL_MAX; i++) { + xdp_rxq_info_unreg(&nvdev->chan_table[i].xdp_rxq); + kfree(nvdev->chan_table[i].recv_buf); + vfree(nvdev->chan_table[i].mrc.slots); + } -get_in_err: - return net_device; + kfree(nvdev); } +static void free_netvsc_device_rcu(struct netvsc_device *nvdev) +{ + call_rcu(&nvdev->rcu, free_netvsc_device); +} -static int netvsc_destroy_recv_buf(struct netvsc_device *net_device) +static void netvsc_revoke_recv_buf(struct hv_device *device, + struct netvsc_device *net_device, + struct net_device *ndev) { struct nvsp_message *revoke_packet; - int ret = 0; - struct net_device *ndev = net_device->ndev; + int ret; /* * If we got a section count, it means we received a @@ -103,11 +198,20 @@ static int netvsc_destroy_recv_buf(struct netvsc_device *net_device) revoke_packet->msg.v1_msg. revoke_recv_buf.id = NETVSC_RECEIVE_BUFFER_ID; - ret = vmbus_sendpacket(net_device->dev->channel, + trace_nvsp_send(ndev, revoke_packet); + + ret = vmbus_sendpacket(device->channel, revoke_packet, sizeof(struct nvsp_message), - (unsigned long)revoke_packet, + VMBUS_RQST_ID_NO_RESPONSE, VM_PKT_DATA_INBAND, 0); + /* If the failure is because the channel is rescinded; + * ignore the failure since we cannot send on a rescinded + * channel. This would allow us to properly cleanup + * even when the channel is rescinded. + */ + if (device->channel->rescind) + ret = 0; /* * If we failed here, we might as well return and * have a leak rather than continue and a bugchk @@ -115,14 +219,72 @@ static int netvsc_destroy_recv_buf(struct netvsc_device *net_device) if (ret != 0) { netdev_err(ndev, "unable to send " "revoke receive buffer to netvsp\n"); - return ret; + return; } + net_device->recv_section_cnt = 0; } +} - /* Teardown the gpadl on the vsp end */ - if (net_device->recv_buf_gpadl_handle) { - ret = vmbus_teardown_gpadl(net_device->dev->channel, - net_device->recv_buf_gpadl_handle); +static void netvsc_revoke_send_buf(struct hv_device *device, + struct netvsc_device *net_device, + struct net_device *ndev) +{ + struct nvsp_message *revoke_packet; + int ret; + + /* Deal with the send buffer we may have setup. + * If we got a send section size, it means we received a + * NVSP_MSG1_TYPE_SEND_SEND_BUF_COMPLETE msg (ie sent + * NVSP_MSG1_TYPE_SEND_SEND_BUF msg) therefore, we need + * to send a revoke msg here + */ + if (net_device->send_section_cnt) { + /* Send the revoke receive buffer */ + revoke_packet = &net_device->revoke_packet; + memset(revoke_packet, 0, sizeof(struct nvsp_message)); + + revoke_packet->hdr.msg_type = + NVSP_MSG1_TYPE_REVOKE_SEND_BUF; + revoke_packet->msg.v1_msg.revoke_send_buf.id = + NETVSC_SEND_BUFFER_ID; + + trace_nvsp_send(ndev, revoke_packet); + + ret = vmbus_sendpacket(device->channel, + revoke_packet, + sizeof(struct nvsp_message), + VMBUS_RQST_ID_NO_RESPONSE, + VM_PKT_DATA_INBAND, 0); + + /* If the failure is because the channel is rescinded; + * ignore the failure since we cannot send on a rescinded + * channel. This would allow us to properly cleanup + * even when the channel is rescinded. + */ + if (device->channel->rescind) + ret = 0; + + /* If we failed here, we might as well return and + * have a leak rather than continue and a bugchk + */ + if (ret != 0) { + netdev_err(ndev, "unable to send " + "revoke send buffer to netvsp\n"); + return; + } + net_device->send_section_cnt = 0; + } +} + +static void netvsc_teardown_recv_gpadl(struct hv_device *device, + struct netvsc_device *net_device, + struct net_device *ndev) +{ + int ret; + + if (net_device->recv_buf_gpadl_handle.gpadl_handle) { + ret = vmbus_teardown_gpadl(device->channel, + &net_device->recv_buf_gpadl_handle); /* If we failed here, we might as well return and have a leak * rather than continue and a bugchk @@ -130,57 +292,83 @@ static int netvsc_destroy_recv_buf(struct netvsc_device *net_device) if (ret != 0) { netdev_err(ndev, "unable to teardown receive buffer's gpadl\n"); - return ret; + return; } - net_device->recv_buf_gpadl_handle = 0; } +} - if (net_device->recv_buf) { - /* Free up the receive buffer */ - free_pages((unsigned long)net_device->recv_buf, - get_order(net_device->recv_buf_size)); - net_device->recv_buf = NULL; - } +static void netvsc_teardown_send_gpadl(struct hv_device *device, + struct netvsc_device *net_device, + struct net_device *ndev) +{ + int ret; - if (net_device->recv_section) { - net_device->recv_section_cnt = 0; - kfree(net_device->recv_section); - net_device->recv_section = NULL; + if (net_device->send_buf_gpadl_handle.gpadl_handle) { + ret = vmbus_teardown_gpadl(device->channel, + &net_device->send_buf_gpadl_handle); + + /* If we failed here, we might as well return and have a leak + * rather than continue and a bugchk + */ + if (ret != 0) { + netdev_err(ndev, + "unable to teardown send buffer's gpadl\n"); + return; + } } +} - return ret; +int netvsc_alloc_recv_comp_ring(struct netvsc_device *net_device, u32 q_idx) +{ + struct netvsc_channel *nvchan = &net_device->chan_table[q_idx]; + int node = cpu_to_node(nvchan->channel->target_cpu); + size_t size; + + size = net_device->recv_completion_cnt * sizeof(struct recv_comp_data); + nvchan->mrc.slots = vzalloc_node(size, node); + if (!nvchan->mrc.slots) + nvchan->mrc.slots = vzalloc(size); + + return nvchan->mrc.slots ? 0 : -ENOMEM; } -static int netvsc_init_recv_buf(struct hv_device *device) +static int netvsc_init_buf(struct hv_device *device, + struct netvsc_device *net_device, + const struct netvsc_device_info *device_info) { - int ret = 0; - int t; - struct netvsc_device *net_device; + struct nvsp_1_message_send_receive_buffer_complete *resp; + struct net_device *ndev = hv_get_drvdata(device); struct nvsp_message *init_packet; - struct net_device *ndev; + unsigned int buf_size; + int i, ret = 0; - net_device = get_outbound_net_device(device); - if (!net_device) - return -ENODEV; - ndev = net_device->ndev; + /* Get receive buffer area. */ + buf_size = device_info->recv_sections * device_info->recv_section_size; + buf_size = roundup(buf_size, PAGE_SIZE); - net_device->recv_buf = - (void *)__get_free_pages(GFP_KERNEL|__GFP_ZERO, - get_order(net_device->recv_buf_size)); + /* Legacy hosts only allow smaller receive buffer */ + if (net_device->nvsp_version <= NVSP_PROTOCOL_VERSION_2) + buf_size = min_t(unsigned int, buf_size, + NETVSC_RECEIVE_BUFFER_SIZE_LEGACY); + + net_device->recv_buf = vzalloc(buf_size); if (!net_device->recv_buf) { - netdev_err(ndev, "unable to allocate receive " - "buffer of size %d\n", net_device->recv_buf_size); + netdev_err(ndev, + "unable to allocate receive buffer of size %u\n", + buf_size); ret = -ENOMEM; goto cleanup; } + net_device->recv_buf_size = buf_size; + /* * Establish the gpadl handle for this buffer on this * channel. Note: This call uses the vmbus connection rather * than the channel to establish the gpadl handle. */ ret = vmbus_establish_gpadl(device->channel, net_device->recv_buf, - net_device->recv_buf_size, + buf_size, &net_device->recv_buf_gpadl_handle); if (ret != 0) { netdev_err(ndev, @@ -188,18 +376,17 @@ static int netvsc_init_recv_buf(struct hv_device *device) goto cleanup; } - /* Notify the NetVsp of the gpadl handle */ init_packet = &net_device->channel_init_pkt; - memset(init_packet, 0, sizeof(struct nvsp_message)); - init_packet->hdr.msg_type = NVSP_MSG1_TYPE_SEND_RECV_BUF; init_packet->msg.v1_msg.send_recv_buf. - gpadl_handle = net_device->recv_buf_gpadl_handle; + gpadl_handle = net_device->recv_buf_gpadl_handle.gpadl_handle; init_packet->msg.v1_msg. send_recv_buf.id = NETVSC_RECEIVE_BUFFER_ID; + trace_nvsp_send(ndev, init_packet); + /* Send the gpadl notification request */ ret = vmbus_sendpacket(device->channel, init_packet, sizeof(struct nvsp_message), @@ -212,68 +399,171 @@ static int netvsc_init_recv_buf(struct hv_device *device) goto cleanup; } - t = wait_for_completion_timeout(&net_device->channel_init_wait, 5*HZ); - BUG_ON(t == 0); - + wait_for_completion(&net_device->channel_init_wait); /* Check the response */ - if (init_packet->msg.v1_msg. - send_recv_buf_complete.status != NVSP_STAT_SUCCESS) { - netdev_err(ndev, "Unable to complete receive buffer " - "initialization with NetVsp - status %d\n", - init_packet->msg.v1_msg. - send_recv_buf_complete.status); + resp = &init_packet->msg.v1_msg.send_recv_buf_complete; + if (resp->status != NVSP_STAT_SUCCESS) { + netdev_err(ndev, + "Unable to complete receive buffer initialization with NetVsp - status %d\n", + resp->status); ret = -EINVAL; goto cleanup; } /* Parse the response */ + netdev_dbg(ndev, "Receive sections: %u sub_allocs: size %u count: %u\n", + resp->num_sections, resp->sections[0].sub_alloc_size, + resp->sections[0].num_sub_allocs); + + /* There should only be one section for the entire receive buffer */ + if (resp->num_sections != 1 || resp->sections[0].offset != 0) { + ret = -EINVAL; + goto cleanup; + } - net_device->recv_section_cnt = init_packet->msg. - v1_msg.send_recv_buf_complete.num_sections; + net_device->recv_section_size = resp->sections[0].sub_alloc_size; + net_device->recv_section_cnt = resp->sections[0].num_sub_allocs; - net_device->recv_section = kmemdup( - init_packet->msg.v1_msg.send_recv_buf_complete.sections, - net_device->recv_section_cnt * - sizeof(struct nvsp_1_receive_buffer_section), - GFP_KERNEL); - if (net_device->recv_section == NULL) { + /* Ensure buffer will not overflow */ + if (net_device->recv_section_size < NETVSC_MTU_MIN || (u64)net_device->recv_section_size * + (u64)net_device->recv_section_cnt > (u64)buf_size) { + netdev_err(ndev, "invalid recv_section_size %u\n", + net_device->recv_section_size); ret = -EINVAL; goto cleanup; } - /* - * For 1st release, there should only be 1 section that represents the - * entire receive buffer + for (i = 0; i < VRSS_CHANNEL_MAX; i++) { + struct netvsc_channel *nvchan = &net_device->chan_table[i]; + + nvchan->recv_buf = kzalloc(net_device->recv_section_size, GFP_KERNEL); + if (nvchan->recv_buf == NULL) { + ret = -ENOMEM; + goto cleanup; + } + } + + /* Setup receive completion ring. + * Add 1 to the recv_section_cnt because at least one entry in a + * ring buffer has to be empty. */ - if (net_device->recv_section_cnt != 1 || - net_device->recv_section->offset != 0) { + net_device->recv_completion_cnt = net_device->recv_section_cnt + 1; + ret = netvsc_alloc_recv_comp_ring(net_device, 0); + if (ret) + goto cleanup; + + /* Now setup the send buffer. */ + buf_size = device_info->send_sections * device_info->send_section_size; + buf_size = round_up(buf_size, PAGE_SIZE); + + net_device->send_buf = vzalloc(buf_size); + if (!net_device->send_buf) { + netdev_err(ndev, "unable to allocate send buffer of size %u\n", + buf_size); + ret = -ENOMEM; + goto cleanup; + } + net_device->send_buf_size = buf_size; + + /* Establish the gpadl handle for this buffer on this + * channel. Note: This call uses the vmbus connection rather + * than the channel to establish the gpadl handle. + */ + ret = vmbus_establish_gpadl(device->channel, net_device->send_buf, + buf_size, + &net_device->send_buf_gpadl_handle); + if (ret != 0) { + netdev_err(ndev, + "unable to establish send buffer's gpadl\n"); + goto cleanup; + } + + /* Notify the NetVsp of the gpadl handle */ + init_packet = &net_device->channel_init_pkt; + memset(init_packet, 0, sizeof(struct nvsp_message)); + init_packet->hdr.msg_type = NVSP_MSG1_TYPE_SEND_SEND_BUF; + init_packet->msg.v1_msg.send_send_buf.gpadl_handle = + net_device->send_buf_gpadl_handle.gpadl_handle; + init_packet->msg.v1_msg.send_send_buf.id = NETVSC_SEND_BUFFER_ID; + + trace_nvsp_send(ndev, init_packet); + + /* Send the gpadl notification request */ + ret = vmbus_sendpacket(device->channel, init_packet, + sizeof(struct nvsp_message), + (unsigned long)init_packet, + VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + if (ret != 0) { + netdev_err(ndev, + "unable to send send buffer's gpadl to netvsp\n"); + goto cleanup; + } + + wait_for_completion(&net_device->channel_init_wait); + + /* Check the response */ + if (init_packet->msg.v1_msg. + send_send_buf_complete.status != NVSP_STAT_SUCCESS) { + netdev_err(ndev, "Unable to complete send buffer " + "initialization with NetVsp - status %d\n", + init_packet->msg.v1_msg. + send_send_buf_complete.status); ret = -EINVAL; goto cleanup; } + /* Parse the response */ + net_device->send_section_size = init_packet->msg. + v1_msg.send_send_buf_complete.section_size; + if (net_device->send_section_size < NETVSC_MTU_MIN) { + netdev_err(ndev, "invalid send_section_size %u\n", + net_device->send_section_size); + ret = -EINVAL; + goto cleanup; + } + + /* Section count is simply the size divided by the section size. */ + net_device->send_section_cnt = buf_size / net_device->send_section_size; + + netdev_dbg(ndev, "Send section size: %d, Section count:%d\n", + net_device->send_section_size, net_device->send_section_cnt); + + /* Setup state for managing the send buffer. */ + net_device->send_section_map = bitmap_zalloc(net_device->send_section_cnt, + GFP_KERNEL); + if (!net_device->send_section_map) { + ret = -ENOMEM; + goto cleanup; + } + goto exit; cleanup: - netvsc_destroy_recv_buf(net_device); + netvsc_revoke_recv_buf(device, net_device, ndev); + netvsc_revoke_send_buf(device, net_device, ndev); + netvsc_teardown_recv_gpadl(device, net_device, ndev); + netvsc_teardown_send_gpadl(device, net_device, ndev); exit: return ret; } - /* Negotiate NVSP protocol version */ static int negotiate_nvsp_ver(struct hv_device *device, struct netvsc_device *net_device, struct nvsp_message *init_packet, u32 nvsp_ver) { - int ret, t; + struct net_device *ndev = hv_get_drvdata(device); + int ret; memset(init_packet, 0, sizeof(struct nvsp_message)); init_packet->hdr.msg_type = NVSP_MSG_TYPE_INIT; init_packet->msg.init_msg.init.min_protocol_ver = nvsp_ver; init_packet->msg.init_msg.init.max_protocol_ver = nvsp_ver; + trace_nvsp_send(ndev, init_packet); /* Send the init request */ ret = vmbus_sendpacket(device->channel, init_packet, @@ -285,55 +575,75 @@ static int negotiate_nvsp_ver(struct hv_device *device, if (ret != 0) return ret; - t = wait_for_completion_timeout(&net_device->channel_init_wait, 5*HZ); - - if (t == 0) - return -ETIMEDOUT; + wait_for_completion(&net_device->channel_init_wait); if (init_packet->msg.init_msg.init_complete.status != NVSP_STAT_SUCCESS) return -EINVAL; - if (nvsp_ver != NVSP_PROTOCOL_VERSION_2) + if (nvsp_ver == NVSP_PROTOCOL_VERSION_1) return 0; - /* NVSPv2 only: Send NDIS config */ + /* NVSPv2 or later: Send NDIS config */ memset(init_packet, 0, sizeof(struct nvsp_message)); init_packet->hdr.msg_type = NVSP_MSG2_TYPE_SEND_NDIS_CONFIG; - init_packet->msg.v2_msg.send_ndis_config.mtu = net_device->ndev->mtu; + init_packet->msg.v2_msg.send_ndis_config.mtu = ndev->mtu + ETH_HLEN; init_packet->msg.v2_msg.send_ndis_config.capability.ieee8021q = 1; + if (nvsp_ver >= NVSP_PROTOCOL_VERSION_5) { + if (hv_is_isolation_supported()) + netdev_info(ndev, "SR-IOV not advertised by guests on the host supporting isolation\n"); + else + init_packet->msg.v2_msg.send_ndis_config.capability.sriov = 1; + + /* Teaming bit is needed to receive link speed updates */ + init_packet->msg.v2_msg.send_ndis_config.capability.teaming = 1; + } + + if (nvsp_ver >= NVSP_PROTOCOL_VERSION_61) + init_packet->msg.v2_msg.send_ndis_config.capability.rsc = 1; + + trace_nvsp_send(ndev, init_packet); + ret = vmbus_sendpacket(device->channel, init_packet, sizeof(struct nvsp_message), - (unsigned long)init_packet, + VMBUS_RQST_ID_NO_RESPONSE, VM_PKT_DATA_INBAND, 0); return ret; } -static int netvsc_connect_vsp(struct hv_device *device) +static int netvsc_connect_vsp(struct hv_device *device, + struct netvsc_device *net_device, + const struct netvsc_device_info *device_info) { - int ret; - struct netvsc_device *net_device; + struct net_device *ndev = hv_get_drvdata(device); + static const u32 ver_list[] = { + NVSP_PROTOCOL_VERSION_1, NVSP_PROTOCOL_VERSION_2, + NVSP_PROTOCOL_VERSION_4, NVSP_PROTOCOL_VERSION_5, + NVSP_PROTOCOL_VERSION_6, NVSP_PROTOCOL_VERSION_61 + }; struct nvsp_message *init_packet; - int ndis_version; - struct net_device *ndev; - - net_device = get_outbound_net_device(device); - if (!net_device) - return -ENODEV; - ndev = net_device->ndev; + int ndis_version, i, ret; init_packet = &net_device->channel_init_pkt; /* Negotiate the latest NVSP protocol supported */ - if (negotiate_nvsp_ver(device, net_device, init_packet, - NVSP_PROTOCOL_VERSION_2) == 0) { - net_device->nvsp_version = NVSP_PROTOCOL_VERSION_2; - } else if (negotiate_nvsp_ver(device, net_device, init_packet, - NVSP_PROTOCOL_VERSION_1) == 0) { - net_device->nvsp_version = NVSP_PROTOCOL_VERSION_1; - } else { + for (i = ARRAY_SIZE(ver_list) - 1; i >= 0; i--) + if (negotiate_nvsp_ver(device, net_device, init_packet, + ver_list[i]) == 0) { + net_device->nvsp_version = ver_list[i]; + break; + } + + if (i < 0) { + ret = -EPROTO; + goto cleanup; + } + + if (hv_is_isolation_supported() && net_device->nvsp_version < NVSP_PROTOCOL_VERSION_61) { + netdev_err(ndev, "Invalid NVSP version 0x%x (expected >= 0x%x) from the host supporting isolation\n", + net_device->nvsp_version, NVSP_PROTOCOL_VERSION_61); ret = -EPROTO; goto cleanup; } @@ -343,7 +653,10 @@ static int netvsc_connect_vsp(struct hv_device *device) /* Send the ndis version */ memset(init_packet, 0, sizeof(struct nvsp_message)); - ndis_version = 0x00050001; + if (net_device->nvsp_version <= NVSP_PROTOCOL_VERSION_4) + ndis_version = 0x00060001; + else + ndis_version = 0x0006001e; init_packet->hdr.msg_type = NVSP_MSG1_TYPE_SEND_NDIS_VER; init_packet->msg.v1_msg. @@ -353,583 +666,1167 @@ static int netvsc_connect_vsp(struct hv_device *device) send_ndis_ver.ndis_minor_ver = ndis_version & 0xFFFF; + trace_nvsp_send(ndev, init_packet); + /* Send the init request */ ret = vmbus_sendpacket(device->channel, init_packet, sizeof(struct nvsp_message), - (unsigned long)init_packet, + VMBUS_RQST_ID_NO_RESPONSE, VM_PKT_DATA_INBAND, 0); if (ret != 0) goto cleanup; - /* Post the big receive buffer to NetVSP */ - ret = netvsc_init_recv_buf(device); + + ret = netvsc_init_buf(device, net_device, device_info); cleanup: return ret; } -static void netvsc_disconnect_vsp(struct netvsc_device *net_device) -{ - netvsc_destroy_recv_buf(net_device); -} - /* * netvsc_device_remove - Callback when the root bus device is removed */ -int netvsc_device_remove(struct hv_device *device) +void netvsc_device_remove(struct hv_device *device) { - struct netvsc_device *net_device; - struct hv_netvsc_packet *netvsc_packet, *pos; - unsigned long flags; - - net_device = hv_get_drvdata(device); - - netvsc_disconnect_vsp(net_device); + struct net_device *ndev = hv_get_drvdata(device); + struct net_device_context *net_device_ctx = netdev_priv(ndev); + struct netvsc_device *net_device + = rtnl_dereference(net_device_ctx->nvdev); + int i; /* - * Since we have already drained, we don't need to busy wait - * as was done in final_release_stor_device() - * Note that we cannot set the ext pointer to NULL until - * we have drained - to drain the outgoing packets, we need to - * allow incoming packets. + * Revoke receive buffer. If host is pre-Win2016 then tear down + * receive buffer GPADL. Do the same for send buffer. */ + netvsc_revoke_recv_buf(device, net_device, ndev); + if (vmbus_proto_version < VERSION_WIN10) + netvsc_teardown_recv_gpadl(device, net_device, ndev); + + netvsc_revoke_send_buf(device, net_device, ndev); + if (vmbus_proto_version < VERSION_WIN10) + netvsc_teardown_send_gpadl(device, net_device, ndev); + + RCU_INIT_POINTER(net_device_ctx->nvdev, NULL); + + /* Disable NAPI and disassociate its context from the device. */ + for (i = 0; i < net_device->num_chn; i++) { + /* See also vmbus_reset_channel_cb(). */ + /* only disable enabled NAPI channel */ + if (i < ndev->real_num_rx_queues) { + netif_queue_set_napi(ndev, i, NETDEV_QUEUE_TYPE_TX, + NULL); + netif_queue_set_napi(ndev, i, NETDEV_QUEUE_TYPE_RX, + NULL); + napi_disable(&net_device->chan_table[i].napi); + } - spin_lock_irqsave(&device->channel->inbound_lock, flags); - hv_set_drvdata(device, NULL); - spin_unlock_irqrestore(&device->channel->inbound_lock, flags); + netif_napi_del(&net_device->chan_table[i].napi); + } /* * At this point, no one should be accessing net_device * except in here */ - dev_notice(&device->device, "net device safe to remove\n"); + netdev_dbg(ndev, "net device safe to remove\n"); /* Now, we can close the channel safely */ vmbus_close(device->channel); - /* Release all resources */ - list_for_each_entry_safe(netvsc_packet, pos, - &net_device->recv_pkt_list, list_ent) { - list_del(&netvsc_packet->list_ent); - kfree(netvsc_packet); + /* + * If host is Win2016 or higher then we do the GPADL tear down + * here after VMBus is closed. + */ + if (vmbus_proto_version >= VERSION_WIN10) { + netvsc_teardown_recv_gpadl(device, net_device, ndev); + netvsc_teardown_send_gpadl(device, net_device, ndev); } - kfree(net_device); - return 0; + /* Release all resources */ + free_netvsc_device_rcu(net_device); } - #define RING_AVAIL_PERCENT_HIWATER 20 #define RING_AVAIL_PERCENT_LOWATER 10 -/* - * Get the percentage of available bytes to write in the ring. - * The return value is in range from 0 to 100. - */ -static inline u32 hv_ringbuf_avail_percent( - struct hv_ring_buffer_info *ring_info) +static inline void netvsc_free_send_slot(struct netvsc_device *net_device, + u32 index) +{ + sync_change_bit(index, net_device->send_section_map); +} + +static void netvsc_send_tx_complete(struct net_device *ndev, + struct netvsc_device *net_device, + struct vmbus_channel *channel, + const struct vmpacket_descriptor *desc, + int budget) { - u32 avail_read, avail_write; + struct net_device_context *ndev_ctx = netdev_priv(ndev); + struct sk_buff *skb; + u16 q_idx = 0; + int queue_sends; + u64 cmd_rqst; + + cmd_rqst = channel->request_addr_callback(channel, desc->trans_id); + if (cmd_rqst == VMBUS_RQST_ERROR) { + netdev_err(ndev, "Invalid transaction ID %llx\n", desc->trans_id); + return; + } + + skb = (struct sk_buff *)(unsigned long)cmd_rqst; + + /* Notify the layer above us */ + if (likely(skb)) { + struct hv_netvsc_packet *packet + = (struct hv_netvsc_packet *)skb->cb; + u32 send_index = packet->send_buf_index; + struct netvsc_stats_tx *tx_stats; + + if (send_index != NETVSC_INVALID_INDEX) + netvsc_free_send_slot(net_device, send_index); + q_idx = packet->q_idx; + + tx_stats = &net_device->chan_table[q_idx].tx_stats; - hv_get_ringbuffer_availbytes(ring_info, &avail_read, &avail_write); + u64_stats_update_begin(&tx_stats->syncp); + tx_stats->packets += packet->total_packets; + tx_stats->bytes += packet->total_bytes; + u64_stats_update_end(&tx_stats->syncp); - return avail_write * 100 / ring_info->ring_datasize; + netvsc_dma_unmap(ndev_ctx->device_ctx, packet); + napi_consume_skb(skb, budget); + } + + queue_sends = + atomic_dec_return(&net_device->chan_table[q_idx].queue_sends); + + if (unlikely(net_device->destroy)) { + if (queue_sends == 0) + wake_up(&net_device->wait_drain); + } else { + struct netdev_queue *txq = netdev_get_tx_queue(ndev, q_idx); + + if (netif_tx_queue_stopped(txq) && !net_device->tx_disable && + (hv_get_avail_to_write_percent(&channel->outbound) > + RING_AVAIL_PERCENT_HIWATER || queue_sends < 1)) { + netif_tx_wake_queue(txq); + ndev_ctx->eth_stats.wake_queue++; + } + } } -static void netvsc_send_completion(struct hv_device *device, - struct vmpacket_descriptor *packet) +static void netvsc_send_completion(struct net_device *ndev, + struct netvsc_device *net_device, + struct vmbus_channel *incoming_channel, + const struct vmpacket_descriptor *desc, + int budget) { - struct netvsc_device *net_device; - struct nvsp_message *nvsp_packet; - struct hv_netvsc_packet *nvsc_packet; - struct net_device *ndev; + const struct nvsp_message *nvsp_packet; + u32 msglen = hv_pkt_datalen(desc); + struct nvsp_message *pkt_rqst; + u64 cmd_rqst; + u32 status; + + /* First check if this is a VMBUS completion without data payload */ + if (!msglen) { + cmd_rqst = incoming_channel->request_addr_callback(incoming_channel, + desc->trans_id); + if (cmd_rqst == VMBUS_RQST_ERROR) { + netdev_err(ndev, "Invalid transaction ID %llx\n", desc->trans_id); + return; + } - net_device = get_inbound_net_device(device); - if (!net_device) + pkt_rqst = (struct nvsp_message *)(uintptr_t)cmd_rqst; + switch (pkt_rqst->hdr.msg_type) { + case NVSP_MSG4_TYPE_SWITCH_DATA_PATH: + complete(&net_device->channel_init_wait); + break; + + default: + netdev_err(ndev, "Unexpected VMBUS completion!!\n"); + } return; - ndev = net_device->ndev; - - nvsp_packet = (struct nvsp_message *)((unsigned long)packet + - (packet->offset8 << 3)); - - if ((nvsp_packet->hdr.msg_type == NVSP_MSG_TYPE_INIT_COMPLETE) || - (nvsp_packet->hdr.msg_type == - NVSP_MSG1_TYPE_SEND_RECV_BUF_COMPLETE) || - (nvsp_packet->hdr.msg_type == - NVSP_MSG1_TYPE_SEND_SEND_BUF_COMPLETE)) { - /* Copy the response back */ - memcpy(&net_device->channel_init_pkt, nvsp_packet, - sizeof(struct nvsp_message)); - complete(&net_device->channel_init_wait); - } else if (nvsp_packet->hdr.msg_type == - NVSP_MSG1_TYPE_SEND_RNDIS_PKT_COMPLETE) { - int num_outstanding_sends; - - /* Get the send context */ - nvsc_packet = (struct hv_netvsc_packet *)(unsigned long) - packet->trans_id; - - /* Notify the layer above us */ - if (nvsc_packet) - nvsc_packet->completion.send.send_completion( - nvsc_packet->completion.send. - send_completion_ctx); - - num_outstanding_sends = - atomic_dec_return(&net_device->num_outstanding_sends); - - if (net_device->destroy && num_outstanding_sends == 0) - wake_up(&net_device->wait_drain); + } - if (netif_queue_stopped(ndev) && !net_device->start_remove && - (hv_ringbuf_avail_percent(&device->channel->outbound) - > RING_AVAIL_PERCENT_HIWATER || - num_outstanding_sends < 1)) - netif_wake_queue(ndev); - } else { - netdev_err(ndev, "Unknown send completion packet type- " - "%d received!!\n", nvsp_packet->hdr.msg_type); + /* Ensure packet is big enough to read header fields */ + if (msglen < sizeof(struct nvsp_message_header)) { + netdev_err(ndev, "nvsp_message length too small: %u\n", msglen); + return; } + nvsp_packet = hv_pkt_data(desc); + switch (nvsp_packet->hdr.msg_type) { + case NVSP_MSG_TYPE_INIT_COMPLETE: + if (msglen < sizeof(struct nvsp_message_header) + + sizeof(struct nvsp_message_init_complete)) { + netdev_err(ndev, "nvsp_msg length too small: %u\n", + msglen); + return; + } + break; + + case NVSP_MSG1_TYPE_SEND_RECV_BUF_COMPLETE: + if (msglen < sizeof(struct nvsp_message_header) + + struct_size_t(struct nvsp_1_message_send_receive_buffer_complete, + sections, 1)) { + netdev_err(ndev, "nvsp_msg1 length too small: %u\n", + msglen); + return; + } + break; + + case NVSP_MSG1_TYPE_SEND_SEND_BUF_COMPLETE: + if (msglen < sizeof(struct nvsp_message_header) + + sizeof(struct nvsp_1_message_send_send_buffer_complete)) { + netdev_err(ndev, "nvsp_msg1 length too small: %u\n", + msglen); + return; + } + break; + + case NVSP_MSG5_TYPE_SUBCHANNEL: + if (msglen < sizeof(struct nvsp_message_header) + + sizeof(struct nvsp_5_subchannel_complete)) { + netdev_err(ndev, "nvsp_msg5 length too small: %u\n", + msglen); + return; + } + break; + + case NVSP_MSG1_TYPE_SEND_RNDIS_PKT_COMPLETE: + if (msglen < sizeof(struct nvsp_message_header) + + sizeof(struct nvsp_1_message_send_rndis_packet_complete)) { + if (net_ratelimit()) + netdev_err(ndev, "nvsp_rndis_pkt_complete length too small: %u\n", + msglen); + return; + } + + /* If status indicates an error, output a message so we know + * there's a problem. But process the completion anyway so the + * resources are released. + */ + status = nvsp_packet->msg.v1_msg.send_rndis_pkt_complete.status; + if (status != NVSP_STAT_SUCCESS && net_ratelimit()) + netdev_err(ndev, "nvsp_rndis_pkt_complete error status: %x\n", + status); + + netvsc_send_tx_complete(ndev, net_device, incoming_channel, + desc, budget); + return; + + default: + netdev_err(ndev, + "Unknown send completion type %d received!!\n", + nvsp_packet->hdr.msg_type); + return; + } + + /* Copy the response back */ + memcpy(&net_device->channel_init_pkt, nvsp_packet, + sizeof(struct nvsp_message)); + complete(&net_device->channel_init_wait); } -int netvsc_send(struct hv_device *device, - struct hv_netvsc_packet *packet) +static u32 netvsc_get_next_send_section(struct netvsc_device *net_device) { - struct netvsc_device *net_device; - int ret = 0; - struct nvsp_message sendMessage; - struct net_device *ndev; - u64 req_id; + unsigned long *map_addr = net_device->send_section_map; + unsigned int i; - net_device = get_outbound_net_device(device); - if (!net_device) - return -ENODEV; - ndev = net_device->ndev; + for_each_clear_bit(i, map_addr, net_device->send_section_cnt) { + if (sync_test_and_set_bit(i, map_addr) == 0) + return i; + } - sendMessage.hdr.msg_type = NVSP_MSG1_TYPE_SEND_RNDIS_PKT; - if (packet->is_data_pkt) { - /* 0 is RMC_DATA; */ - sendMessage.msg.v1_msg.send_rndis_pkt.channel_type = 0; - } else { - /* 1 is RMC_CONTROL; */ - sendMessage.msg.v1_msg.send_rndis_pkt.channel_type = 1; + return NETVSC_INVALID_INDEX; +} + +static void netvsc_copy_to_send_buf(struct netvsc_device *net_device, + unsigned int section_index, + u32 pend_size, + struct hv_netvsc_packet *packet, + struct rndis_message *rndis_msg, + struct hv_page_buffer *pb, + bool xmit_more) +{ + char *start = net_device->send_buf; + char *dest = start + (section_index * net_device->send_section_size) + + pend_size; + int i; + u32 padding = 0; + u32 page_count = packet->cp_partial ? 1 : packet->page_buf_cnt; + u32 remain; + + /* Add padding */ + remain = packet->total_data_buflen & (net_device->pkt_align - 1); + if (xmit_more && remain) { + padding = net_device->pkt_align - remain; + rndis_msg->msg_len += padding; + packet->total_data_buflen += padding; + } + + for (i = 0; i < page_count; i++) { + char *src = phys_to_virt(pb[i].pfn << HV_HYP_PAGE_SHIFT); + u32 offset = pb[i].offset; + u32 len = pb[i].len; + + memcpy(dest, (src + offset), len); + dest += len; } - /* Not using send buffer section */ - sendMessage.msg.v1_msg.send_rndis_pkt.send_buf_section_index = - 0xFFFFFFFF; - sendMessage.msg.v1_msg.send_rndis_pkt.send_buf_section_size = 0; + if (padding) + memset(dest, 0, padding); +} + +void netvsc_dma_unmap(struct hv_device *hv_dev, + struct hv_netvsc_packet *packet) +{ + int i; + + if (!hv_is_isolation_supported()) + return; + + if (!packet->dma_range) + return; + + for (i = 0; i < packet->page_buf_cnt; i++) + dma_unmap_single(&hv_dev->device, packet->dma_range[i].dma, + packet->dma_range[i].mapping_size, + DMA_TO_DEVICE); + + kfree(packet->dma_range); +} - if (packet->completion.send.send_completion) - req_id = (ulong)packet; +/* netvsc_dma_map - Map swiotlb bounce buffer with data page of + * packet sent by vmbus_sendpacket_pagebuffer() in the Isolation + * VM. + * + * In isolation VM, netvsc send buffer has been marked visible to + * host and so the data copied to send buffer doesn't need to use + * bounce buffer. The data pages handled by vmbus_sendpacket_pagebuffer() + * may not be copied to send buffer and so these pages need to be + * mapped with swiotlb bounce buffer. netvsc_dma_map() is to do + * that. The pfns in the struct hv_page_buffer need to be converted + * to bounce buffer's pfn. The loop here is necessary because the + * entries in the page buffer array are not necessarily full + * pages of data. Each entry in the array has a separate offset and + * len that may be non-zero, even for entries in the middle of the + * array. And the entries are not physically contiguous. So each + * entry must be individually mapped rather than as a contiguous unit. + * So not use dma_map_sg() here. + */ +static int netvsc_dma_map(struct hv_device *hv_dev, + struct hv_netvsc_packet *packet, + struct hv_page_buffer *pb) +{ + u32 page_count = packet->page_buf_cnt; + dma_addr_t dma; + int i; + + if (!hv_is_isolation_supported()) + return 0; + + packet->dma_range = kcalloc(page_count, + sizeof(*packet->dma_range), + GFP_ATOMIC); + if (!packet->dma_range) + return -ENOMEM; + + for (i = 0; i < page_count; i++) { + char *src = phys_to_virt((pb[i].pfn << HV_HYP_PAGE_SHIFT) + + pb[i].offset); + u32 len = pb[i].len; + + dma = dma_map_single(&hv_dev->device, src, len, + DMA_TO_DEVICE); + if (dma_mapping_error(&hv_dev->device, dma)) { + kfree(packet->dma_range); + return -ENOMEM; + } + + /* pb[].offset and pb[].len are not changed during dma mapping + * and so not reassign. + */ + packet->dma_range[i].dma = dma; + packet->dma_range[i].mapping_size = len; + pb[i].pfn = dma >> HV_HYP_PAGE_SHIFT; + } + + return 0; +} + +/* Build an "array" of mpb entries describing the data to be transferred + * over VMBus. After the desc header fields, each "array" entry is variable + * size, and each entry starts after the end of the previous entry. The + * "offset" and "len" fields for each entry imply the size of the entry. + * + * The pfns are in HV_HYP_PAGE_SIZE, because all communication with Hyper-V + * uses that granularity, even if the system page size of the guest is larger. + * Each entry in the input "pb" array must describe a contiguous range of + * guest physical memory so that the pfns are sequential if the range crosses + * a page boundary. The offset field must be < HV_HYP_PAGE_SIZE. + */ +static inline void netvsc_build_mpb_array(struct hv_page_buffer *pb, + u32 page_buffer_count, + struct vmbus_packet_mpb_array *desc, + u32 *desc_size) +{ + struct hv_mpb_array *mpb_entry = &desc->range; + int i, j; + + for (i = 0; i < page_buffer_count; i++) { + u32 offset = pb[i].offset; + u32 len = pb[i].len; + + mpb_entry->offset = offset; + mpb_entry->len = len; + + for (j = 0; j < HVPFN_UP(offset + len); j++) + mpb_entry->pfn_array[j] = pb[i].pfn + j; + + mpb_entry = (struct hv_mpb_array *)&mpb_entry->pfn_array[j]; + } + + desc->rangecount = page_buffer_count; + *desc_size = (char *)mpb_entry - (char *)desc; +} + +static inline int netvsc_send_pkt( + struct hv_device *device, + struct hv_netvsc_packet *packet, + struct netvsc_device *net_device, + struct hv_page_buffer *pb, + struct sk_buff *skb) +{ + struct nvsp_message nvmsg; + struct nvsp_1_message_send_rndis_packet *rpkt = + &nvmsg.msg.v1_msg.send_rndis_pkt; + struct netvsc_channel * const nvchan = + &net_device->chan_table[packet->q_idx]; + struct vmbus_channel *out_channel = nvchan->channel; + struct net_device *ndev = hv_get_drvdata(device); + struct net_device_context *ndev_ctx = netdev_priv(ndev); + struct netdev_queue *txq = netdev_get_tx_queue(ndev, packet->q_idx); + u64 req_id; + int ret; + u32 ring_avail = hv_get_avail_to_write_percent(&out_channel->outbound); + + memset(&nvmsg, 0, sizeof(struct nvsp_message)); + nvmsg.hdr.msg_type = NVSP_MSG1_TYPE_SEND_RNDIS_PKT; + if (skb) + rpkt->channel_type = 0; /* 0 is RMC_DATA */ else - req_id = 0; + rpkt->channel_type = 1; /* 1 is RMC_CONTROL */ + rpkt->send_buf_section_index = packet->send_buf_index; + if (packet->send_buf_index == NETVSC_INVALID_INDEX) + rpkt->send_buf_section_size = 0; + else + rpkt->send_buf_section_size = packet->total_data_buflen; + + req_id = (ulong)skb; + + if (out_channel->rescind) + return -ENODEV; + + trace_nvsp_send_pkt(ndev, out_channel, rpkt); + + packet->dma_range = NULL; if (packet->page_buf_cnt) { - ret = vmbus_sendpacket_pagebuffer(device->channel, - packet->page_buf, - packet->page_buf_cnt, - &sendMessage, - sizeof(struct nvsp_message), - req_id); + struct vmbus_channel_packet_page_buffer desc; + u32 desc_size; + + if (packet->cp_partial) + pb++; + + ret = netvsc_dma_map(ndev_ctx->device_ctx, packet, pb); + if (ret) { + ret = -EAGAIN; + goto exit; + } + + netvsc_build_mpb_array(pb, packet->page_buf_cnt, + (struct vmbus_packet_mpb_array *)&desc, + &desc_size); + ret = vmbus_sendpacket_mpb_desc(out_channel, + (struct vmbus_packet_mpb_array *)&desc, + desc_size, &nvmsg, sizeof(nvmsg), req_id); + if (ret) + netvsc_dma_unmap(ndev_ctx->device_ctx, packet); } else { - ret = vmbus_sendpacket(device->channel, &sendMessage, - sizeof(struct nvsp_message), - req_id, - VM_PKT_DATA_INBAND, - VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + ret = vmbus_sendpacket(out_channel, + &nvmsg, sizeof(nvmsg), + req_id, VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); } +exit: if (ret == 0) { - atomic_inc(&net_device->num_outstanding_sends); - if (hv_ringbuf_avail_percent(&device->channel->outbound) < - RING_AVAIL_PERCENT_LOWATER) { - netif_stop_queue(ndev); - if (atomic_read(&net_device-> - num_outstanding_sends) < 1) - netif_wake_queue(ndev); + atomic_inc_return(&nvchan->queue_sends); + + if (ring_avail < RING_AVAIL_PERCENT_LOWATER) { + netif_tx_stop_queue(txq); + ndev_ctx->eth_stats.stop_queue++; } } else if (ret == -EAGAIN) { - netif_stop_queue(ndev); - if (atomic_read(&net_device->num_outstanding_sends) < 1) { - netif_wake_queue(ndev); - ret = -ENOSPC; - } + netif_tx_stop_queue(txq); + ndev_ctx->eth_stats.stop_queue++; } else { - netdev_err(ndev, "Unable to send packet %p ret %d\n", - packet, ret); + netdev_err(ndev, + "Unable to send packet pages %u len %u, ret %d\n", + packet->page_buf_cnt, packet->total_data_buflen, + ret); + } + + if (netif_tx_queue_stopped(txq) && + atomic_read(&nvchan->queue_sends) < 1 && + !net_device->tx_disable) { + netif_tx_wake_queue(txq); + ndev_ctx->eth_stats.wake_queue++; + if (ret == -EAGAIN) + ret = -ENOSPC; } return ret; } -static void netvsc_send_recv_completion(struct hv_device *device, - u64 transaction_id, u32 status) +/* Move packet out of multi send data (msd), and clear msd */ +static inline void move_pkt_msd(struct hv_netvsc_packet **msd_send, + struct sk_buff **msd_skb, + struct multi_send_data *msdp) { - struct nvsp_message recvcompMessage; - int retries = 0; - int ret; - struct net_device *ndev; - struct netvsc_device *net_device = hv_get_drvdata(device); + *msd_skb = msdp->skb; + *msd_send = msdp->pkt; + msdp->skb = NULL; + msdp->pkt = NULL; + msdp->count = 0; +} - ndev = net_device->ndev; +/* RCU already held by caller */ +/* Batching/bouncing logic is designed to attempt to optimize + * performance. + * + * For small, non-LSO packets we copy the packet to a send buffer + * which is pre-registered with the Hyper-V side. This enables the + * hypervisor to avoid remapping the aperture to access the packet + * descriptor and data. + * + * If we already started using a buffer and the netdev is transmitting + * a burst of packets, keep on copying into the buffer until it is + * full or we are done collecting a burst. If there is an existing + * buffer with space for the RNDIS descriptor but not the packet, copy + * the RNDIS descriptor to the buffer, keeping the packet in place. + * + * If we do batching and send more than one packet using a single + * NetVSC message, free the SKBs of the packets copied, except for the + * last packet. This is done to streamline the handling of the case + * where the last packet only had the RNDIS descriptor copied to the + * send buffer, with the data pointers included in the NetVSC message. + */ +int netvsc_send(struct net_device *ndev, + struct hv_netvsc_packet *packet, + struct rndis_message *rndis_msg, + struct hv_page_buffer *pb, + struct sk_buff *skb, + bool xdp_tx) +{ + struct net_device_context *ndev_ctx = netdev_priv(ndev); + struct netvsc_device *net_device + = rcu_dereference_bh(ndev_ctx->nvdev); + struct hv_device *device = ndev_ctx->device_ctx; + int ret = 0; + struct netvsc_channel *nvchan; + u32 pktlen = packet->total_data_buflen, msd_len = 0; + unsigned int section_index = NETVSC_INVALID_INDEX; + struct multi_send_data *msdp; + struct hv_netvsc_packet *msd_send = NULL, *cur_send = NULL; + struct sk_buff *msd_skb = NULL; + bool try_batch, xmit_more; + + /* If device is rescinded, return error and packet will get dropped. */ + if (unlikely(!net_device || net_device->destroy)) + return -ENODEV; - recvcompMessage.hdr.msg_type = - NVSP_MSG1_TYPE_SEND_RNDIS_PKT_COMPLETE; + nvchan = &net_device->chan_table[packet->q_idx]; + packet->send_buf_index = NETVSC_INVALID_INDEX; + packet->cp_partial = false; - recvcompMessage.msg.v1_msg.send_rndis_pkt_complete.status = status; + /* Send a control message or XDP packet directly without accessing + * msd (Multi-Send Data) field which may be changed during data packet + * processing. + */ + if (!skb || xdp_tx) + return netvsc_send_pkt(device, packet, net_device, pb, skb); + + /* batch packets in send buffer if possible */ + msdp = &nvchan->msd; + if (msdp->pkt) + msd_len = msdp->pkt->total_data_buflen; + + try_batch = msd_len > 0 && msdp->count < net_device->max_pkt; + if (try_batch && msd_len + pktlen + net_device->pkt_align < + net_device->send_section_size) { + section_index = msdp->pkt->send_buf_index; + + } else if (try_batch && msd_len + packet->rmsg_size < + net_device->send_section_size) { + section_index = msdp->pkt->send_buf_index; + packet->cp_partial = true; + + } else if (pktlen + net_device->pkt_align < + net_device->send_section_size) { + section_index = netvsc_get_next_send_section(net_device); + if (unlikely(section_index == NETVSC_INVALID_INDEX)) { + ++ndev_ctx->eth_stats.tx_send_full; + } else { + move_pkt_msd(&msd_send, &msd_skb, msdp); + msd_len = 0; + } + } -retry_send_cmplt: - /* Send the completion */ - ret = vmbus_sendpacket(device->channel, &recvcompMessage, - sizeof(struct nvsp_message), transaction_id, - VM_PKT_COMP, 0); - if (ret == 0) { - /* success */ - /* no-op */ - } else if (ret == -EAGAIN) { - /* no more room...wait a bit and attempt to retry 3 times */ - retries++; - netdev_err(ndev, "unable to send receive completion pkt" - " (tid %llx)...retrying %d\n", transaction_id, retries); - - if (retries < 4) { - udelay(100); - goto retry_send_cmplt; + /* Keep aggregating only if stack says more data is coming + * and not doing mixed modes send and not flow blocked + */ + xmit_more = netdev_xmit_more() && + !packet->cp_partial && + !netif_xmit_stopped(netdev_get_tx_queue(ndev, packet->q_idx)); + + if (section_index != NETVSC_INVALID_INDEX) { + netvsc_copy_to_send_buf(net_device, + section_index, msd_len, + packet, rndis_msg, pb, xmit_more); + + packet->send_buf_index = section_index; + + if (packet->cp_partial) { + packet->page_buf_cnt--; + packet->total_data_buflen = msd_len + packet->rmsg_size; + } else { + packet->page_buf_cnt = 0; + packet->total_data_buflen += msd_len; + } + + if (msdp->pkt) { + packet->total_packets += msdp->pkt->total_packets; + packet->total_bytes += msdp->pkt->total_bytes; + } + + if (msdp->skb) + dev_consume_skb_any(msdp->skb); + + if (xmit_more) { + msdp->skb = skb; + msdp->pkt = packet; + msdp->count++; } else { - netdev_err(ndev, "unable to send receive " - "completion pkt (tid %llx)...give up retrying\n", - transaction_id); + cur_send = packet; + msdp->skb = NULL; + msdp->pkt = NULL; + msdp->count = 0; } } else { - netdev_err(ndev, "unable to send receive " - "completion pkt - %llx\n", transaction_id); + move_pkt_msd(&msd_send, &msd_skb, msdp); + cur_send = packet; + } + + if (msd_send) { + int m_ret = netvsc_send_pkt(device, msd_send, net_device, + NULL, msd_skb); + + if (m_ret != 0) { + netvsc_free_send_slot(net_device, + msd_send->send_buf_index); + dev_kfree_skb_any(msd_skb); + } } + + if (cur_send) + ret = netvsc_send_pkt(device, cur_send, net_device, pb, skb); + + if (ret != 0 && section_index != NETVSC_INVALID_INDEX) + netvsc_free_send_slot(net_device, section_index); + + return ret; } -/* Send a receive completion packet to RNDIS device (ie NetVsp) */ -static void netvsc_receive_completion(void *context) +/* Send pending recv completions */ +static int send_recv_completions(struct net_device *ndev, + struct netvsc_device *nvdev, + struct netvsc_channel *nvchan) { - struct hv_netvsc_packet *packet = context; - struct hv_device *device = packet->device; - struct netvsc_device *net_device; - u64 transaction_id = 0; - bool fsend_receive_comp = false; - unsigned long flags; - struct net_device *ndev; - u32 status = NVSP_STAT_NONE; + struct multi_recv_comp *mrc = &nvchan->mrc; + struct recv_comp_msg { + struct nvsp_message_header hdr; + u32 status; + } __packed; + struct recv_comp_msg msg = { + .hdr.msg_type = NVSP_MSG1_TYPE_SEND_RNDIS_PKT_COMPLETE, + }; + int ret; - /* - * Even though it seems logical to do a GetOutboundNetDevice() here to - * send out receive completion, we are using GetInboundNetDevice() - * since we may have disable outbound traffic already. - */ - net_device = get_inbound_net_device(device); - if (!net_device) - return; - ndev = net_device->ndev; + while (mrc->first != mrc->next) { + const struct recv_comp_data *rcd + = mrc->slots + mrc->first; - /* Overloading use of the lock. */ - spin_lock_irqsave(&net_device->recv_pkt_list_lock, flags); + msg.status = rcd->status; + ret = vmbus_sendpacket(nvchan->channel, &msg, sizeof(msg), + rcd->tid, VM_PKT_COMP, 0); + if (unlikely(ret)) { + struct net_device_context *ndev_ctx = netdev_priv(ndev); - if (packet->status != NVSP_STAT_SUCCESS) - packet->xfer_page_pkt->status = NVSP_STAT_FAIL; + ++ndev_ctx->eth_stats.rx_comp_busy; + return ret; + } - packet->xfer_page_pkt->count--; + if (++mrc->first == nvdev->recv_completion_cnt) + mrc->first = 0; + } - /* - * Last one in the line that represent 1 xfer page packet. - * Return the xfer page packet itself to the freelist - */ - if (packet->xfer_page_pkt->count == 0) { - fsend_receive_comp = true; - transaction_id = packet->completion.recv.recv_completion_tid; - status = packet->xfer_page_pkt->status; - list_add_tail(&packet->xfer_page_pkt->list_ent, - &net_device->recv_pkt_list); + /* receive completion ring has been emptied */ + if (unlikely(nvdev->destroy)) + wake_up(&nvdev->wait_drain); + return 0; +} + +/* Count how many receive completions are outstanding */ +static void recv_comp_slot_avail(const struct netvsc_device *nvdev, + const struct multi_recv_comp *mrc, + u32 *filled, u32 *avail) +{ + u32 count = nvdev->recv_completion_cnt; + + if (mrc->next >= mrc->first) + *filled = mrc->next - mrc->first; + else + *filled = (count - mrc->first) + mrc->next; + + *avail = count - *filled - 1; +} + +/* Add receive complete to ring to send to host. */ +static void enq_receive_complete(struct net_device *ndev, + struct netvsc_device *nvdev, u16 q_idx, + u64 tid, u32 status) +{ + struct netvsc_channel *nvchan = &nvdev->chan_table[q_idx]; + struct multi_recv_comp *mrc = &nvchan->mrc; + struct recv_comp_data *rcd; + u32 filled, avail; + + recv_comp_slot_avail(nvdev, mrc, &filled, &avail); + + if (unlikely(filled > NAPI_POLL_WEIGHT)) { + send_recv_completions(ndev, nvdev, nvchan); + recv_comp_slot_avail(nvdev, mrc, &filled, &avail); } - /* Put the packet back */ - list_add_tail(&packet->list_ent, &net_device->recv_pkt_list); - spin_unlock_irqrestore(&net_device->recv_pkt_list_lock, flags); + if (unlikely(!avail)) { + netdev_err(ndev, "Recv_comp full buf q:%hd, tid:%llx\n", + q_idx, tid); + return; + } - /* Send a receive completion for the xfer page packet */ - if (fsend_receive_comp) - netvsc_send_recv_completion(device, transaction_id, status); + rcd = mrc->slots + mrc->next; + rcd->tid = tid; + rcd->status = status; + if (++mrc->next == nvdev->recv_completion_cnt) + mrc->next = 0; } -static void netvsc_receive(struct hv_device *device, - struct vmpacket_descriptor *packet) +static int netvsc_receive(struct net_device *ndev, + struct netvsc_device *net_device, + struct netvsc_channel *nvchan, + const struct vmpacket_descriptor *desc) { - struct netvsc_device *net_device; - struct vmtransfer_page_packet_header *vmxferpage_packet; - struct nvsp_message *nvsp_packet; - struct hv_netvsc_packet *netvsc_packet = NULL; - /* struct netvsc_driver *netvscDriver; */ - struct xferpage_packet *xferpage_packet = NULL; + struct net_device_context *net_device_ctx = netdev_priv(ndev); + struct vmbus_channel *channel = nvchan->channel; + const struct vmtransfer_page_packet_header *vmxferpage_packet + = container_of(desc, const struct vmtransfer_page_packet_header, d); + const struct nvsp_message *nvsp = hv_pkt_data(desc); + u32 msglen = hv_pkt_datalen(desc); + u16 q_idx = channel->offermsg.offer.sub_channel_index; + char *recv_buf = net_device->recv_buf; + u32 status = NVSP_STAT_SUCCESS; int i; int count = 0; - unsigned long flags; - struct net_device *ndev; - LIST_HEAD(listHead); + /* Ensure packet is big enough to read header fields */ + if (msglen < sizeof(struct nvsp_message_header)) { + netif_err(net_device_ctx, rx_err, ndev, + "invalid nvsp header, length too small: %u\n", + msglen); + return 0; + } - net_device = get_inbound_net_device(device); - if (!net_device) - return; - ndev = net_device->ndev; + /* Make sure this is a valid nvsp packet */ + if (unlikely(nvsp->hdr.msg_type != NVSP_MSG1_TYPE_SEND_RNDIS_PKT)) { + netif_err(net_device_ctx, rx_err, ndev, + "Unknown nvsp packet type received %u\n", + nvsp->hdr.msg_type); + return 0; + } - /* - * All inbound packets other than send completion should be xfer page - * packet - */ - if (packet->type != VM_PKT_DATA_USING_XFER_PAGES) { - netdev_err(ndev, "Unknown packet type received - %d\n", - packet->type); - return; + /* Validate xfer page pkt header */ + if ((desc->offset8 << 3) < sizeof(struct vmtransfer_page_packet_header)) { + netif_err(net_device_ctx, rx_err, ndev, + "Invalid xfer page pkt, offset too small: %u\n", + desc->offset8 << 3); + return 0; } - nvsp_packet = (struct nvsp_message *)((unsigned long)packet + - (packet->offset8 << 3)); + if (unlikely(vmxferpage_packet->xfer_pageset_id != NETVSC_RECEIVE_BUFFER_ID)) { + netif_err(net_device_ctx, rx_err, ndev, + "Invalid xfer page set id - expecting %x got %x\n", + NETVSC_RECEIVE_BUFFER_ID, + vmxferpage_packet->xfer_pageset_id); + return 0; + } - /* Make sure this is a valid nvsp packet */ - if (nvsp_packet->hdr.msg_type != - NVSP_MSG1_TYPE_SEND_RNDIS_PKT) { - netdev_err(ndev, "Unknown nvsp packet type received-" - " %d\n", nvsp_packet->hdr.msg_type); + count = vmxferpage_packet->range_cnt; + + /* Check count for a valid value */ + if (NETVSC_XFER_HEADER_SIZE(count) > desc->offset8 << 3) { + netif_err(net_device_ctx, rx_err, ndev, + "Range count is not valid: %d\n", + count); + return 0; + } + + /* Each range represents 1 RNDIS pkt that contains 1 ethernet frame */ + for (i = 0; i < count; i++) { + u32 offset = vmxferpage_packet->ranges[i].byte_offset; + u32 buflen = vmxferpage_packet->ranges[i].byte_count; + void *data; + int ret; + + if (unlikely(offset > net_device->recv_buf_size || + buflen > net_device->recv_buf_size - offset)) { + nvchan->rsc.cnt = 0; + status = NVSP_STAT_FAIL; + netif_err(net_device_ctx, rx_err, ndev, + "Packet offset:%u + len:%u too big\n", + offset, buflen); + + continue; + } + + /* We're going to copy (sections of) the packet into nvchan->recv_buf; + * make sure that nvchan->recv_buf is large enough to hold the packet. + */ + if (unlikely(buflen > net_device->recv_section_size)) { + nvchan->rsc.cnt = 0; + status = NVSP_STAT_FAIL; + netif_err(net_device_ctx, rx_err, ndev, + "Packet too big: buflen=%u recv_section_size=%u\n", + buflen, net_device->recv_section_size); + + continue; + } + + data = recv_buf + offset; + + nvchan->rsc.is_last = (i == count - 1); + + trace_rndis_recv(ndev, q_idx, data); + + /* Pass it to the upper layer */ + ret = rndis_filter_receive(ndev, net_device, + nvchan, data, buflen); + + if (unlikely(ret != NVSP_STAT_SUCCESS)) { + /* Drop incomplete packet */ + nvchan->rsc.cnt = 0; + status = NVSP_STAT_FAIL; + } + } + + enq_receive_complete(ndev, net_device, q_idx, + vmxferpage_packet->d.trans_id, status); + + return count; +} + +static void netvsc_send_table(struct net_device *ndev, + struct netvsc_device *nvscdev, + const struct nvsp_message *nvmsg, + u32 msglen) +{ + struct net_device_context *net_device_ctx = netdev_priv(ndev); + u32 count, offset, *tab; + int i; + + /* Ensure packet is big enough to read send_table fields */ + if (msglen < sizeof(struct nvsp_message_header) + + sizeof(struct nvsp_5_send_indirect_table)) { + netdev_err(ndev, "nvsp_v5_msg length too small: %u\n", msglen); return; } - vmxferpage_packet = (struct vmtransfer_page_packet_header *)packet; + count = nvmsg->msg.v5_msg.send_table.count; + offset = nvmsg->msg.v5_msg.send_table.offset; - if (vmxferpage_packet->xfer_pageset_id != NETVSC_RECEIVE_BUFFER_ID) { - netdev_err(ndev, "Invalid xfer page set id - " - "expecting %x got %x\n", NETVSC_RECEIVE_BUFFER_ID, - vmxferpage_packet->xfer_pageset_id); + if (count != VRSS_SEND_TAB_SIZE) { + netdev_err(ndev, "Received wrong send-table size:%u\n", count); return; } - /* - * Grab free packets (range count + 1) to represent this xfer - * page packet. +1 to represent the xfer page packet itself. - * We grab it here so that we know exactly how many we can - * fulfil + /* If negotiated version <= NVSP_PROTOCOL_VERSION_6, the offset may be + * wrong due to a host bug. So fix the offset here. */ - spin_lock_irqsave(&net_device->recv_pkt_list_lock, flags); - while (!list_empty(&net_device->recv_pkt_list)) { - list_move_tail(net_device->recv_pkt_list.next, &listHead); - if (++count == vmxferpage_packet->range_cnt + 1) - break; + if (nvscdev->nvsp_version <= NVSP_PROTOCOL_VERSION_6 && + msglen >= sizeof(struct nvsp_message_header) + + sizeof(union nvsp_6_message_uber) + count * sizeof(u32)) + offset = sizeof(struct nvsp_message_header) + + sizeof(union nvsp_6_message_uber); + + /* Boundary check for all versions */ + if (msglen < count * sizeof(u32) || offset > msglen - count * sizeof(u32)) { + netdev_err(ndev, "Received send-table offset too big:%u\n", + offset); + return; } - spin_unlock_irqrestore(&net_device->recv_pkt_list_lock, flags); - /* - * We need at least 2 netvsc pkts (1 to represent the xfer - * page and at least 1 for the range) i.e. we can handled - * some of the xfer page packet ranges... - */ - if (count < 2) { - netdev_err(ndev, "Got only %d netvsc pkt...needed " - "%d pkts. Dropping this xfer page packet completely!\n", - count, vmxferpage_packet->range_cnt + 1); + tab = (void *)nvmsg + offset; - /* Return it to the freelist */ - spin_lock_irqsave(&net_device->recv_pkt_list_lock, flags); - for (i = count; i != 0; i--) { - list_move_tail(listHead.next, - &net_device->recv_pkt_list); - } - spin_unlock_irqrestore(&net_device->recv_pkt_list_lock, - flags); + for (i = 0; i < count; i++) + net_device_ctx->tx_table[i] = tab[i]; +} - netvsc_send_recv_completion(device, - vmxferpage_packet->d.trans_id, - NVSP_STAT_FAIL); +static void netvsc_send_vf(struct net_device *ndev, + const struct nvsp_message *nvmsg, + u32 msglen) +{ + struct net_device_context *net_device_ctx = netdev_priv(ndev); + /* Ensure packet is big enough to read its fields */ + if (msglen < sizeof(struct nvsp_message_header) + + sizeof(struct nvsp_4_send_vf_association)) { + netdev_err(ndev, "nvsp_v4_msg length too small: %u\n", msglen); return; } - /* Remove the 1st packet to represent the xfer page packet itself */ - xferpage_packet = (struct xferpage_packet *)listHead.next; - list_del(&xferpage_packet->list_ent); - xferpage_packet->status = NVSP_STAT_SUCCESS; + net_device_ctx->vf_alloc = nvmsg->msg.v4_msg.vf_assoc.allocated; + net_device_ctx->vf_serial = nvmsg->msg.v4_msg.vf_assoc.serial; + + if (net_device_ctx->vf_alloc) + complete(&net_device_ctx->vf_add); + + netdev_info(ndev, "VF slot %u %s\n", + net_device_ctx->vf_serial, + net_device_ctx->vf_alloc ? "added" : "removed"); +} + +static void netvsc_receive_inband(struct net_device *ndev, + struct netvsc_device *nvscdev, + const struct vmpacket_descriptor *desc) +{ + const struct nvsp_message *nvmsg = hv_pkt_data(desc); + u32 msglen = hv_pkt_datalen(desc); - /* This is how much we can satisfy */ - xferpage_packet->count = count - 1; + /* Ensure packet is big enough to read header fields */ + if (msglen < sizeof(struct nvsp_message_header)) { + netdev_err(ndev, "inband nvsp_message length too small: %u\n", msglen); + return; + } - if (xferpage_packet->count != vmxferpage_packet->range_cnt) { - netdev_err(ndev, "Needed %d netvsc pkts to satisfy " - "this xfer page...got %d\n", - vmxferpage_packet->range_cnt, xferpage_packet->count); + switch (nvmsg->hdr.msg_type) { + case NVSP_MSG5_TYPE_SEND_INDIRECTION_TABLE: + netvsc_send_table(ndev, nvscdev, nvmsg, msglen); + break; + + case NVSP_MSG4_TYPE_SEND_VF_ASSOCIATION: + if (hv_is_isolation_supported()) + netdev_err(ndev, "Ignore VF_ASSOCIATION msg from the host supporting isolation\n"); + else + netvsc_send_vf(ndev, nvmsg, msglen); + break; } +} - /* Each range represents 1 RNDIS pkt that contains 1 ethernet frame */ - for (i = 0; i < (count - 1); i++) { - netvsc_packet = (struct hv_netvsc_packet *)listHead.next; - list_del(&netvsc_packet->list_ent); - - /* Initialize the netvsc packet */ - netvsc_packet->status = NVSP_STAT_SUCCESS; - netvsc_packet->xfer_page_pkt = xferpage_packet; - netvsc_packet->completion.recv.recv_completion = - netvsc_receive_completion; - netvsc_packet->completion.recv.recv_completion_ctx = - netvsc_packet; - netvsc_packet->device = device; - /* Save this so that we can send it back */ - netvsc_packet->completion.recv.recv_completion_tid = - vmxferpage_packet->d.trans_id; - - netvsc_packet->data = (void *)((unsigned long)net_device-> - recv_buf + vmxferpage_packet->ranges[i].byte_offset); - netvsc_packet->total_data_buflen = - vmxferpage_packet->ranges[i].byte_count; +static int netvsc_process_raw_pkt(struct hv_device *device, + struct netvsc_channel *nvchan, + struct netvsc_device *net_device, + struct net_device *ndev, + const struct vmpacket_descriptor *desc, + int budget) +{ + struct vmbus_channel *channel = nvchan->channel; + const struct nvsp_message *nvmsg = hv_pkt_data(desc); - /* Pass it to the upper layer */ - rndis_filter_receive(device, netvsc_packet); + trace_nvsp_recv(ndev, channel, nvmsg); + + switch (desc->type) { + case VM_PKT_COMP: + netvsc_send_completion(ndev, net_device, channel, desc, budget); + break; + + case VM_PKT_DATA_USING_XFER_PAGES: + return netvsc_receive(ndev, net_device, nvchan, desc); - netvsc_receive_completion(netvsc_packet-> - completion.recv.recv_completion_ctx); + case VM_PKT_DATA_INBAND: + netvsc_receive_inband(ndev, net_device, desc); + break; + + default: + netdev_err(ndev, "unhandled packet type %d, tid %llx\n", + desc->type, desc->trans_id); + break; } + return 0; +} + +static struct hv_device *netvsc_channel_to_device(struct vmbus_channel *channel) +{ + struct vmbus_channel *primary = channel->primary_channel; + + return primary ? primary->device_obj : channel->device_obj; } -static void netvsc_channel_cb(void *context) +/* Network processing softirq + * Process data in incoming ring buffer from host + * Stops when ring is empty or budget is met or exceeded. + */ +int netvsc_poll(struct napi_struct *napi, int budget) { + struct netvsc_channel *nvchan + = container_of(napi, struct netvsc_channel, napi); + struct netvsc_device *net_device = nvchan->net_device; + struct vmbus_channel *channel = nvchan->channel; + struct hv_device *device = netvsc_channel_to_device(channel); + struct net_device *ndev = hv_get_drvdata(device); + int work_done = 0; int ret; - struct hv_device *device = context; - struct netvsc_device *net_device; - u32 bytes_recvd; - u64 request_id; - unsigned char *packet; - struct vmpacket_descriptor *desc; - unsigned char *buffer; - int bufferlen = NETVSC_PACKET_SIZE; - struct net_device *ndev; - - packet = kzalloc(NETVSC_PACKET_SIZE * sizeof(unsigned char), - GFP_ATOMIC); - if (!packet) - return; - buffer = packet; - net_device = get_inbound_net_device(device); - if (!net_device) - goto out; - ndev = net_device->ndev; + /* If starting a new interval */ + if (!nvchan->desc) + nvchan->desc = hv_pkt_iter_first(channel); - do { - ret = vmbus_recvpacket_raw(device->channel, buffer, bufferlen, - &bytes_recvd, &request_id); - if (ret == 0) { - if (bytes_recvd > 0) { - desc = (struct vmpacket_descriptor *)buffer; - switch (desc->type) { - case VM_PKT_COMP: - netvsc_send_completion(device, desc); - break; - - case VM_PKT_DATA_USING_XFER_PAGES: - netvsc_receive(device, desc); - break; - - default: - netdev_err(ndev, - "unhandled packet type %d, " - "tid %llx len %d\n", - desc->type, request_id, - bytes_recvd); - break; - } - - /* reset */ - if (bufferlen > NETVSC_PACKET_SIZE) { - kfree(buffer); - buffer = packet; - bufferlen = NETVSC_PACKET_SIZE; - } - } else { - /* reset */ - if (bufferlen > NETVSC_PACKET_SIZE) { - kfree(buffer); - buffer = packet; - bufferlen = NETVSC_PACKET_SIZE; - } - - break; - } - } else if (ret == -ENOBUFS) { - /* Handle large packet */ - buffer = kmalloc(bytes_recvd, GFP_ATOMIC); - if (buffer == NULL) { - /* Try again next time around */ - netdev_err(ndev, - "unable to allocate buffer of size " - "(%d)!!\n", bytes_recvd); - break; - } - - bufferlen = bytes_recvd; - } - } while (1); - -out: - kfree(buffer); - return; + nvchan->xdp_flush = false; + + while (nvchan->desc && work_done < budget) { + work_done += netvsc_process_raw_pkt(device, nvchan, net_device, + ndev, nvchan->desc, budget); + nvchan->desc = hv_pkt_iter_next(channel, nvchan->desc); + } + + if (nvchan->xdp_flush) + xdp_do_flush(); + + /* Send any pending receive completions */ + ret = send_recv_completions(ndev, net_device, nvchan); + + /* If it did not exhaust NAPI budget this time + * and not doing busy poll + * then re-enable host interrupts + * and reschedule if ring is not empty + * or sending receive completion failed. + */ + if (work_done < budget && + napi_complete_done(napi, work_done) && + (ret || hv_end_read(&channel->inbound)) && + napi_schedule_prep(napi)) { + hv_begin_read(&channel->inbound); + __napi_schedule(napi); + } + + /* Driver may overshoot since multiple packets per descriptor */ + return min(work_done, budget); +} + +/* Call back when data is available in host ring buffer. + * Processing is deferred until network softirq (NAPI) + */ +void netvsc_channel_cb(void *context) +{ + struct netvsc_channel *nvchan = context; + struct vmbus_channel *channel = nvchan->channel; + struct hv_ring_buffer_info *rbi = &channel->inbound; + + /* preload first vmpacket descriptor */ + prefetch(hv_get_ring_buffer(rbi) + rbi->priv_read_index); + + if (napi_schedule_prep(&nvchan->napi)) { + /* disable interrupts from host */ + hv_begin_read(rbi); + + __napi_schedule_irqoff(&nvchan->napi); + } } /* * netvsc_device_add - Callback when the device belonging to this * driver is added */ -int netvsc_device_add(struct hv_device *device, void *additional_info) +struct netvsc_device *netvsc_device_add(struct hv_device *device, + const struct netvsc_device_info *device_info) { - int ret = 0; - int i; - int ring_size = - ((struct netvsc_device_info *)additional_info)->ring_size; + int i, ret = 0; struct netvsc_device *net_device; - struct hv_netvsc_packet *packet, *pos; - struct net_device *ndev; + struct net_device *ndev = hv_get_drvdata(device); + struct net_device_context *net_device_ctx = netdev_priv(ndev); - net_device = alloc_net_device(device); - if (!net_device) { - ret = -ENOMEM; - goto cleanup; - } + net_device = alloc_net_device(); + if (!net_device) + return ERR_PTR(-ENOMEM); - /* - * Coming into this function, struct net_device * is - * registered as the driver private data. - * In alloc_net_device(), we register struct netvsc_device * - * as the driver private data and stash away struct net_device * - * in struct netvsc_device *. + for (i = 0; i < VRSS_SEND_TAB_SIZE; i++) + net_device_ctx->tx_table[i] = 0; + + /* Because the device uses NAPI, all the interrupt batching and + * control is done via Net softirq, not the channel handling + */ + set_channel_read_mode(device->channel, HV_CALL_ISR); + + /* If we're reopening the device we may have multiple queues, fill the + * chn_table with the default channel to use it before subchannels are + * opened. + * Initialize the channel state before we open; + * we can be interrupted as soon as we open the channel. */ - ndev = net_device->ndev; - /* Initialize the NetVSC channel extension */ - net_device->recv_buf_size = NETVSC_RECEIVE_BUFFER_SIZE; - spin_lock_init(&net_device->recv_pkt_list_lock); + for (i = 0; i < VRSS_CHANNEL_MAX; i++) { + struct netvsc_channel *nvchan = &net_device->chan_table[i]; - INIT_LIST_HEAD(&net_device->recv_pkt_list); + nvchan->channel = device->channel; + nvchan->net_device = net_device; + u64_stats_init(&nvchan->tx_stats.syncp); + u64_stats_init(&nvchan->rx_stats.syncp); - for (i = 0; i < NETVSC_RECEIVE_PACKETLIST_COUNT; i++) { - packet = kzalloc(sizeof(struct hv_netvsc_packet), GFP_KERNEL); - if (!packet) - break; + ret = xdp_rxq_info_reg(&nvchan->xdp_rxq, ndev, i, 0); - list_add_tail(&packet->list_ent, - &net_device->recv_pkt_list); + if (ret) { + netdev_err(ndev, "xdp_rxq_info_reg fail: %d\n", ret); + goto cleanup2; + } + + ret = xdp_rxq_info_reg_mem_model(&nvchan->xdp_rxq, + MEM_TYPE_PAGE_SHARED, NULL); + + if (ret) { + netdev_err(ndev, "xdp reg_mem_model fail: %d\n", ret); + goto cleanup2; + } } - init_completion(&net_device->channel_init_wait); + + /* Enable NAPI handler before init callbacks */ + netif_napi_add(ndev, &net_device->chan_table[0].napi, netvsc_poll); + napi_enable(&net_device->chan_table[0].napi); + netif_queue_set_napi(ndev, 0, NETDEV_QUEUE_TYPE_RX, + &net_device->chan_table[0].napi); + netif_queue_set_napi(ndev, 0, NETDEV_QUEUE_TYPE_TX, + &net_device->chan_table[0].napi); /* Open the channel */ - ret = vmbus_open(device->channel, ring_size * PAGE_SIZE, - ring_size * PAGE_SIZE, NULL, 0, - netvsc_channel_cb, device); + device->channel->next_request_id_callback = vmbus_next_request_id; + device->channel->request_addr_callback = vmbus_request_addr; + device->channel->rqstor_size = netvsc_rqstor_size(netvsc_ring_bytes); + device->channel->max_pkt_size = NETVSC_MAX_PKT_SIZE; + + ret = vmbus_open(device->channel, netvsc_ring_bytes, + netvsc_ring_bytes, NULL, 0, + netvsc_channel_cb, net_device->chan_table); if (ret != 0) { netdev_err(ndev, "unable to open channel: %d\n", ret); @@ -937,34 +1834,37 @@ int netvsc_device_add(struct hv_device *device, void *additional_info) } /* Channel is opened */ - pr_info("hv_netvsc channel opened successfully\n"); + netdev_dbg(ndev, "hv_netvsc channel opened successfully\n"); /* Connect with the NetVsp */ - ret = netvsc_connect_vsp(device); + ret = netvsc_connect_vsp(device, net_device, device_info); if (ret != 0) { netdev_err(ndev, "unable to connect to NetVSP - %d\n", ret); goto close; } - return ret; + /* Writing nvdev pointer unlocks netvsc_send(), make sure chn_table is + * populated. + */ + rcu_assign_pointer(net_device_ctx->nvdev, net_device); + + return net_device; close: + RCU_INIT_POINTER(net_device_ctx->nvdev, NULL); + /* Now, we can close the channel safely */ vmbus_close(device->channel); cleanup: + netif_queue_set_napi(ndev, 0, NETDEV_QUEUE_TYPE_TX, NULL); + netif_queue_set_napi(ndev, 0, NETDEV_QUEUE_TYPE_RX, NULL); + napi_disable(&net_device->chan_table[0].napi); + netif_napi_del(&net_device->chan_table[0].napi); - if (net_device) { - list_for_each_entry_safe(packet, pos, - &net_device->recv_pkt_list, - list_ent) { - list_del(&packet->list_ent); - kfree(packet); - } - - kfree(net_device); - } +cleanup2: + free_netvsc_device(&net_device->rcu); - return ret; + return ERR_PTR(ret); } |