nuttx/drivers/virtio/virtio-net.c

599 lines
18 KiB
C
Raw Normal View History

virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/****************************************************************************
* drivers/virtio/virtio-net.c
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership. The
* ASF licenses this file to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance with the
* License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
* License for the specific language governing permissions and limitations
* under the License.
*
****************************************************************************/
/****************************************************************************
* Included Files
****************************************************************************/
#include <debug.h>
#include <errno.h>
#include <stdint.h>
#include <string.h>
#include <nuttx/compiler.h>
#include <nuttx/kmalloc.h>
#include <nuttx/net/ip.h>
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
#include <nuttx/net/netdev_lowerhalf.h>
#include <nuttx/virtio/virtio.h>
#include "virtio-net.h"
/****************************************************************************
* Pre-processor Definitions
****************************************************************************/
/* Virtio net header size and packet buffer size */
#define VIRTIO_NET_HDRSIZE (sizeof(struct virtio_net_hdr_s))
#define VIRTIO_NET_LLHDRSIZE (sizeof(struct virtio_net_llhdr_s))
#define VIRTIO_NET_BUFSIZE (CONFIG_NET_ETH_PKTSIZE + CONFIG_NET_GUARDSIZE)
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/* Virtio net virtqueue index and number */
#define VIRTIO_NET_RX 0
#define VIRTIO_NET_TX 1
#define VIRTIO_NET_NUM 2
#define VIRTIO_NET_MAX_PKT_SIZE \
((CONFIG_NET_LL_GUARDSIZE - ETH_HDRLEN) + VIRTIO_NET_BUFSIZE)
#define VIRTIO_NET_MAX_NIOB \
((VIRTIO_NET_MAX_PKT_SIZE + CONFIG_IOB_BUFSIZE - 1) / CONFIG_IOB_BUFSIZE)
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/****************************************************************************
* Private Types
****************************************************************************/
/* Virtio net header, just define it here for further, now only use it
* to calculate the virto net header size, see marco VIRTIO_NET_HDRSIZE
*/
begin_packed_struct struct virtio_net_hdr_s
{
uint8_t flags;
uint8_t gso_type;
uint16_t hdr_len;
uint16_t gso_size;
uint16_t csum_start;
uint16_t csum_offset;
} end_packed_struct;
struct virtio_net_priv_s
{
/* This holds the information visible to the NuttX network */
struct netdev_lowerhalf_s lower; /* The netdev lowerhalf */
/* Virtio device information */
FAR struct virtio_device *vdev; /* Virtio device pointer */
int bufnum; /* TX and RX Buffer number */
};
/* Virtio Link Layer Header, follow shows the iob buffer layout:
*
* |<-- CONFIG_NET_LL_GUARDSIZE -->|
* +---------------+---------------+------------+------+ +-------------+
* | Virtio Header | ETH Header | data | free | --> | next netpkt |
* +---------------+---------------+------------+------+ +-------------+
* | |<--------- datalen -------->|
* ^base ^data
*
* CONFIG_NET_LL_GUARDSIZE >= VIRTIO_NET_LLHDRSIZE + ETH_HDR_SIZE
* = sizeof(uintptr) + 10 + 14
* = 32 (64-Bit)
* = 28 (32-Bit)
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
*/
begin_packed_struct struct virtio_net_llhdr_s
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
{
FAR netpkt_t *pkt; /* Netpaket pointer */
struct virtio_net_hdr_s vhdr; /* Virtio net header */
} end_packed_struct;
static_assert(CONFIG_NET_LL_GUARDSIZE >= VIRTIO_NET_LLHDRSIZE + ETH_HDRLEN,
"CONFIG_NET_LL_GUARDSIZE cannot be less than ETH_HDRLEN"
" + VIRTIO_NET_LLHDRSIZE");
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/****************************************************************************
* Private Function Prototypes
****************************************************************************/
static int virtio_net_ifup(FAR struct netdev_lowerhalf_s *dev);
static int virtio_net_ifdown(FAR struct netdev_lowerhalf_s *dev);
static int virtio_net_send(FAR struct netdev_lowerhalf_s *dev,
FAR netpkt_t *pkt);
static netpkt_t *virtio_net_recv(FAR struct netdev_lowerhalf_s *dev);
#ifdef CONFIG_NET_MCASTGROUP
static int virtio_net_addmac(FAR struct netdev_lowerhalf_s *dev,
FAR const uint8_t *mac);
static int virtio_net_rmmac(FAR struct netdev_lowerhalf_s *dev,
FAR const uint8_t *mac);
#endif
#ifdef CONFIG_NETDEV_IOCTL
static int virtio_net_ioctl(FAR struct netdev_lowerhalf_s *dev);
#endif
static int virtio_net_probe(FAR struct virtio_device *vdev);
static void virtio_net_remove(FAR struct virtio_device *vdev);
/****************************************************************************
* Private Data
****************************************************************************/
static struct virtio_driver g_virtio_net_driver =
{
LIST_INITIAL_VALUE(g_virtio_net_driver.node), /* node */
VIRTIO_ID_NETWORK, /* device id */
virtio_net_probe, /* probe */
virtio_net_remove, /* remove */
};
static const struct netdev_ops_s g_virtio_net_ops =
{
virtio_net_ifup,
virtio_net_ifdown,
virtio_net_send,
virtio_net_recv,
#ifdef CONFIG_NET_MCASTGROUP
virtio_net_addmac,
virtio_net_rmmac,
#endif
#ifdef CONFIG_NETDEV_IOCTL
virtio_net_ioctl,
#endif
};
/****************************************************************************
* Private Functions
****************************************************************************/
/****************************************************************************
* Name: virtio_net_rxfill
****************************************************************************/
static void virtio_net_rxfill(FAR struct netdev_lowerhalf_s *dev)
{
FAR struct virtio_net_priv_s *priv = (FAR struct virtio_net_priv_s *)dev;
FAR struct virtqueue *vq = priv->vdev->vrings_info[VIRTIO_NET_RX].vq;
FAR struct virtio_net_llhdr_s *hdr;
struct virtqueue_buf vb[VIRTIO_NET_MAX_NIOB];
struct iovec iov[VIRTIO_NET_MAX_NIOB];
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
FAR netpkt_t *pkt;
int iov_cnt;
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
int i;
for (i = 0; i < priv->bufnum; i++)
{
/* IOB Offload, Alloc buffer from RX netpkt */
pkt = netpkt_alloc(dev, NETPKT_RX);
if (pkt == NULL)
{
vrtinfo("Has ran out of the RX buffer, i=%d\n", i);
break;
}
/* Preserve data length */
if (netpkt_setdatalen(dev, pkt, VIRTIO_NET_BUFSIZE) <
VIRTIO_NET_BUFSIZE)
{
vrtwarn("No enough buffer to prepare RX buffer, i=%d\n", i);
netpkt_free(dev, pkt, NETPKT_RX);
break;
}
/* Convert netpkt to virtqueue_buf */
iov_cnt = netpkt_to_iov(dev, pkt, iov, VIRTIO_NET_MAX_NIOB);
for (i = 0; i < iov_cnt; i++)
{
vb[i].buf = iov[i].iov_base;
vb[i].len = iov[i].iov_len;
}
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/* Alloc cookie and net header from transport layer */
hdr = (FAR struct virtio_net_llhdr_s *)
((FAR uint8_t *)vb[0].buf - VIRTIO_NET_LLHDRSIZE);
DEBUGASSERT((FAR uint8_t *)hdr >= netpkt_getbase(pkt));
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
memset(&hdr->vhdr, 0, sizeof(hdr->vhdr));
hdr->pkt = pkt;
/* Buffer 0, the virtio net header */
vb[0].buf = &hdr->vhdr;
vb[0].len += VIRTIO_NET_HDRSIZE;
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
vrtinfo("Fill rx, hdr=%p, count=%d\n", hdr, iov_cnt);
virtqueue_add_buffer(vq, vb, 0, iov_cnt, hdr);
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
}
if (i > 0)
{
virtqueue_kick(vq);
}
}
/****************************************************************************
* Name: virtio_net_txfree
****************************************************************************/
static void virtio_net_txfree(FAR struct netdev_lowerhalf_s *dev)
{
FAR struct virtio_net_priv_s *priv = (FAR struct virtio_net_priv_s *)dev;
FAR struct virtqueue *vq = priv->vdev->vrings_info[VIRTIO_NET_TX].vq;
FAR struct virtio_net_llhdr_s *hdr;
while (1)
{
/* Get buffer from tx virtqueue */
hdr = virtqueue_get_buffer(vq, NULL, NULL);
if (hdr == NULL)
{
break;
}
netpkt_free(dev, hdr->pkt, NETPKT_TX);
vrtinfo("Free, hdr: %p, pkt: %p\n", hdr, hdr->pkt);
}
}
/****************************************************************************
* Name: virtio_net_ifup
****************************************************************************/
static int virtio_net_ifup(FAR struct netdev_lowerhalf_s *dev)
{
FAR struct virtio_net_priv_s *priv = (FAR struct virtio_net_priv_s *)dev;
#ifdef CONFIG_NET_IPv4
vrtinfo("Bringing up: %u.%u.%u.%u\n",
ip4_addr1(dev->netdev.d_ipaddr), ip4_addr2(dev->netdev.d_ipaddr),
ip4_addr3(dev->netdev.d_ipaddr), ip4_addr4(dev->netdev.d_ipaddr));
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
#endif
#ifdef CONFIG_NET_IPv6
vrtinfo("Bringing up: %04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n",
dev->netdev.d_ipv6addr[0], dev->netdev.d_ipv6addr[1],
dev->netdev.d_ipv6addr[2], dev->netdev.d_ipv6addr[3],
dev->netdev.d_ipv6addr[4], dev->netdev.d_ipv6addr[5],
dev->netdev.d_ipv6addr[6], dev->netdev.d_ipv6addr[7]);
#endif
/* Prepare interrupt and packets for receiving */
virtqueue_enable_cb(priv->vdev->vrings_info[VIRTIO_NET_RX].vq);
virtio_net_rxfill(dev);
return netdev_lower_carrier_on(dev);
}
/****************************************************************************
* Name: virtio_net_ifdown
****************************************************************************/
static int virtio_net_ifdown(FAR struct netdev_lowerhalf_s *dev)
{
FAR struct virtio_net_priv_s *priv = (FAR struct virtio_net_priv_s *)dev;
int i;
/* Disable the Ethernet interrupt */
for (i = 0; i < VIRTIO_NET_NUM; i++)
{
virtqueue_disable_cb(priv->vdev->vrings_info[i].vq);
}
return netdev_lower_carrier_off(dev);
}
/****************************************************************************
* Name: virtio_net_send
****************************************************************************/
static int virtio_net_send(FAR struct netdev_lowerhalf_s *dev,
FAR netpkt_t *pkt)
{
FAR struct virtio_net_priv_s *priv = (FAR struct virtio_net_priv_s *)dev;
FAR struct virtqueue *vq = priv->vdev->vrings_info[VIRTIO_NET_TX].vq;
FAR struct virtio_net_llhdr_s *hdr;
struct virtqueue_buf vb[VIRTIO_NET_MAX_NIOB];
struct iovec iov[VIRTIO_NET_MAX_NIOB];
int iov_cnt;
int i;
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/* Check the send length */
if (netpkt_getdatalen(dev, pkt) > VIRTIO_NET_BUFSIZE)
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
{
vrterr("net send buffer too large\n");
return -EINVAL;
}
/* Convert netpkt to virtqueue_buf */
iov_cnt = netpkt_to_iov(dev, pkt, iov, VIRTIO_NET_MAX_NIOB);
for (i = 0; i < iov_cnt; i++)
{
vb[i].buf = iov[i].iov_base;
vb[i].len = iov[i].iov_len;
}
/* Prepare virtio net header */
hdr = (FAR struct virtio_net_llhdr_s *)
((FAR uint8_t *)vb[0].buf - VIRTIO_NET_LLHDRSIZE);
DEBUGASSERT((FAR uint8_t *)hdr >= netpkt_getbase(pkt));
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
hdr->pkt = pkt;
memset(&hdr->vhdr, 0, sizeof(hdr->vhdr));
/* Buffer 0 is the virtio net header */
vb[0].buf = &hdr->vhdr;
vb[0].len += VIRTIO_NET_HDRSIZE;
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
/* Add buffer to vq and notify the other side */
vrtinfo("Send, hdr=%p, count=%d\n", hdr, iov_cnt);
virtqueue_add_buffer(vq, vb, iov_cnt, 0, hdr);
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
virtqueue_kick(vq);
/* Try return Netpkt TX buffer to upper-half. */
virtio_net_txfree(dev);
/* If we have no buffer left, enable TX done callback. */
if (netdev_lower_quota_load(dev, NETPKT_TX) <= 0)
{
virtqueue_enable_cb(vq);
}
return OK;
}
/****************************************************************************
* Name: virtio_net_recv
****************************************************************************/
static netpkt_t *virtio_net_recv(FAR struct netdev_lowerhalf_s *dev)
{
FAR struct virtio_net_priv_s *priv = (FAR struct virtio_net_priv_s *)dev;
FAR struct virtqueue *vq = priv->vdev->vrings_info[VIRTIO_NET_RX].vq;
FAR struct virtio_net_llhdr_s *hdr;
uint32_t len;
/* Fill the free Netpkt RX buffer to the RX virtqueue */
virtio_net_rxfill(dev);
/* Get received buffer form RX virtqueue */
hdr = virtqueue_get_buffer(vq, &len, NULL);
if (hdr == NULL)
{
/* If we have no buffer left, enable RX callback. */
virtqueue_enable_cb(vq);
/* We do transmit after recv, now it's time to free TX buffer.
* Depends on upper-half order (Call TX after RX).
*
* TODO: Find a better way to free TX buffer.
*/
virtio_net_txfree(&priv->lower);
vrtinfo("get NULL buffer\n");
return NULL;
}
/* Set the received pkt length */
netpkt_setdatalen(dev, hdr->pkt, len - VIRTIO_NET_HDRSIZE);
vrtinfo("Recv, hdr=%p, pkt=%p, len=%" PRIu32 "\n", hdr, hdr->pkt, len);
return hdr->pkt;
}
#ifdef CONFIG_NET_MCASTGROUP
/****************************************************************************
* Name: virtio_net_addmac
****************************************************************************/
static int virtio_net_addmac(FAR struct netdev_lowerhalf_s *dev,
FAR const uint8_t *mac)
{
return -ENOSYS;
}
/****************************************************************************
* Name: virtio_net_rmmac
****************************************************************************/
static int virtio_net_rmmac(FAR struct netdev_lowerhalf_s *dev,
FAR const uint8_t *mac)
{
return -ENOSYS;
}
#endif
#ifdef CONFIG_NETDEV_IOCTL
/****************************************************************************
* Name: virtio_net_ioctl
****************************************************************************/
static int virtio_net_ioctl(FAR struct netdev_lowerhalf_s *dev)
{
return -ENOTTY;
}
#endif
/****************************************************************************
* Name: virtio_net_rxready
****************************************************************************/
static void virtio_net_rxready(FAR struct virtqueue *vq)
{
FAR struct virtio_net_priv_s *priv = vq->vq_dev->priv;
virtqueue_disable_cb(vq);
netdev_lower_rxready(&priv->lower);
}
/****************************************************************************
* Name: virtio_net_txdone
****************************************************************************/
static void virtio_net_txdone(FAR struct virtqueue *vq)
{
FAR struct virtio_net_priv_s *priv = vq->vq_dev->priv;
virtqueue_disable_cb(vq);
netdev_lower_txdone(&priv->lower);
}
/****************************************************************************
* Name: virtio_net_init
****************************************************************************/
static int virtio_net_init(FAR struct virtio_net_priv_s *priv,
FAR struct virtio_device *vdev)
{
FAR const char *vqnames[VIRTIO_NET_NUM];
vq_callback callbacks[VIRTIO_NET_NUM];
int ret;
priv->vdev = vdev;
vdev->priv = priv;
/* Initialize the virtio device */
virtio_set_status(vdev, VIRTIO_CONFIG_STATUS_DRIVER);
virtio_set_features(vdev, 0);
virtio_set_status(vdev, VIRTIO_CONFIG_FEATURES_OK);
vqnames[VIRTIO_NET_RX] = "virtio_net_rx";
vqnames[VIRTIO_NET_TX] = "virtio_net_tx";
callbacks[VIRTIO_NET_RX] = virtio_net_rxready;
callbacks[VIRTIO_NET_TX] = virtio_net_txdone;
ret = virtio_create_virtqueues(vdev, 0, VIRTIO_NET_NUM, vqnames,
callbacks);
if (ret < 0)
{
vrterr("virtio_device_create_virtqueue failed, ret=%d\n", ret);
return ret;
}
virtio_set_status(vdev, VIRTIO_CONFIG_STATUS_DRIVER_OK);
#if CONFIG_DRIVERS_VIRTIO_NET_BUFNUM > 0
priv->bufnum = CONFIG_DRIVERS_VIRTIO_NET_BUFNUM;
#else
/* Calculate the virtio network buffer number:
* 1/4 for the TX netpkts, 1/4 for the RX netpkts.
*/
priv->bufnum = CONFIG_IOB_NBUFFERS / VIRTIO_NET_MAX_NIOB / 4;
virtio: add virtio framework in NuttX 1. virtio devics/drivers match and probe/remote mechanism; 2. virtio mmio transport layer based on OpenAmp (Compatible with both virtio mmio version 1 and 2); 3. virtio-serial driver based on new virtio framework; 4. virtio-rng driver based on new virtio framework; 5. virtio-net driver based on new virtio framework (IOB Offload implementation); 6. virtio-blk driver based on new virtio framework; 7. Remove the old virtio mmio framework, the old framework only support mmio transport layer, and the new framwork support more transport layer and this commit has implemented all the old virtio drivers; 8. Refresh the the qemu-arm64 and qemu-riscv virtio related configs, and update its README.txt; New virtio-net driver has better performance Compared with previous virtio-mmio-net: | | master/-c | master/-s | this/-c | this/-s | | :--------------------: | :-------: | :-------: | :-----: | :-----: | | qemu-armv8a:netnsh | 539Mbps | 524Mbps | 906Mbps | 715Mbps | | qemu-armv8a:netnsh_smp | 401Mbps | 437Mbps | 583Mbps | 505Mbps | | rv-virt:netnsh | 487Mbps | 512Mbps | 760Mbps | 634Mbps | | rv-virt:netnsh_smp | 387Mbps | 455Mbps | 447Mbps | 502Mbps | | rv-virt:netnsh64 | 602Mbps | 595Mbps | 881Mbps | 769Mbps | | rv-virt:netnsh64_smp | 414Mbps | 515Mbps | 491Mbps | 525Mbps | | rv-virt:knetnsh64 | 515Mbps | 457Mbps | 606Mbps | 540Mbps | | rv-virt:knetnsh64_smp | 308Mbps | 389Mbps | 415Mbps | 474Mbps | Note: Both CONFIG_IOB_NBUFFERS=64, using iperf command, all in Mbits/sec Tested in QEMU 7.2.2 Signed-off-by: wangbowen6 <wangbowen6@xiaomi.com> Signed-off-by: Zhe Weng <wengzhe@xiaomi.com>
2023-03-22 04:49:43 +01:00
#endif
priv->bufnum = MIN(vdev->vrings_info[VIRTIO_NET_RX].info.num_descs,
priv->bufnum);
priv->bufnum = MIN(vdev->vrings_info[VIRTIO_NET_TX].info.num_descs,
priv->bufnum);
return OK;
}
/****************************************************************************
* Name: virtio_net_probe
****************************************************************************/
static int virtio_net_probe(FAR struct virtio_device *vdev)
{
FAR struct netdev_lowerhalf_s *netdev;
FAR struct virtio_net_priv_s *priv;
int ret;
priv = kmm_zalloc(sizeof(*priv));
if (priv == NULL)
{
vrterr("Virtio net driver priv alloc failed\n");
return -ENOMEM;
}
ret = virtio_net_init(priv, vdev);
if (ret < 0)
{
vrterr("virtio_net_init failed, ret=%d\n", ret);
goto err_with_priv;
}
/* Initialize the netdev lower half */
netdev = &priv->lower;
netdev->quota[NETPKT_RX] = priv->bufnum;
netdev->quota[NETPKT_TX] = priv->bufnum;
netdev->ops = &g_virtio_net_ops;
/* Register the net deivce */
ret = netdev_lower_register(netdev, NET_LL_ETHERNET);
if (ret < 0)
{
vrterr("netdev_lower_register failed, ret=%d\n", ret);
goto err_with_virtqueues;
}
return ret;
err_with_virtqueues:
virtio_reset_device(vdev);
virtio_delete_virtqueues(vdev);
err_with_priv:
kmm_free(priv);
return ret;
}
/****************************************************************************
* Name: virtio_net_remove
****************************************************************************/
static void virtio_net_remove(FAR struct virtio_device *vdev)
{
FAR struct virtio_net_priv_s *priv = vdev->priv;
netdev_lower_unregister(&priv->lower);
virtio_reset_device(vdev);
virtio_delete_virtqueues(vdev);
kmm_free(priv);
}
/****************************************************************************
* Public Functions
****************************************************************************/
/****************************************************************************
* Name: virtio_register_net_driver
****************************************************************************/
int virtio_register_net_driver(void)
{
return virtio_register_driver(&g_virtio_net_driver);
}