d/numam-dpdk
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

vhost: fix code style

Browse Source 
Fix alignment issues, lengthy lines, misordered type and other coding style issues.

Signed-off-by: Huawei Xie <huawei.xie@intel.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
This commit is contained in:
Huawei Xie 2014-11-06 07:31:41 +08:00 committed by Thomas Monjalon
parent ac718398f4
commit af4f2c5feb
6 changed files with 494 additions and 397 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

244
lib/librte_vhost/eventfd_link/eventfd_link.c
View File

@ -1,26 +1,26 @@
/*-
* * GPL LICENSE SUMMARY
* *
* * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
* *
* * This program is free software; you can redistribute it and/or modify
* * it under the terms of version 2 of the GNU General Public License as
* * published by the Free Software Foundation.
* *
* * This program is distributed in the hope that it will be useful, but
* * WITHOUT ANY WARRANTY; without even the implied warranty of
* * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* * General Public License 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., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
* * The full GNU General Public License is included in this distribution
* * in the file called LICENSE.GPL.
* *
* * Contact Information:
* * Intel Corporation
* */
* GPL LICENSE SUMMARY
*
* Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of version 2 of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License 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., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
* The full GNU General Public License is included in this distribution
* in the file called LICENSE.GPL.
*
* Contact Information:
* Intel Corporation
*/
#include <linux/eventfd.h>
#include <linux/miscdevice.h>
@ -42,15 +42,15 @@
* get_files_struct is copied from fs/file.c
*/
struct files_struct *
get_files_struct (struct task_struct *task)
get_files_struct(struct task_struct *task)
{
struct files_struct *files;
task_lock (task);
task_lock(task);
files = task->files;
if (files)
atomic_inc (&files->count);
task_unlock (task);
atomic_inc(&files->count);
task_unlock(task);
return files;
}
@ -59,17 +59,15 @@ get_files_struct (struct task_struct *task)
* put_files_struct is extracted from fs/file.c
*/
void
put_files_struct (struct files_struct *files)
put_files_struct(struct files_struct *files)
{
if (atomic_dec_and_test (&files->count))
{
BUG ();
}
if (atomic_dec_and_test(&files->count))
BUG();
}
static long
eventfd_link_ioctl (struct file *f, unsigned int ioctl, unsigned long arg)
eventfd_link_ioctl(struct file *f, unsigned int ioctl, unsigned long arg)
{
void __user *argp = (void __user *) arg;
struct task_struct *task_target = NULL;
@ -78,96 +76,88 @@ eventfd_link_ioctl (struct file *f, unsigned int ioctl, unsigned long arg)
struct fdtable *fdt;
struct eventfd_copy eventfd_copy;
switch (ioctl)
{
case EVENTFD_COPY:
if (copy_from_user (&eventfd_copy, argp, sizeof (struct eventfd_copy)))
return -EFAULT;
switch (ioctl) {
case EVENTFD_COPY:
if (copy_from_user(&eventfd_copy, argp,
sizeof(struct eventfd_copy)))
return -EFAULT;
/*
* Find the task struct for the target pid
*/
task_target =
pid_task (find_vpid (eventfd_copy.target_pid), PIDTYPE_PID);
if (task_target == NULL)
{
printk (KERN_DEBUG "Failed to get mem ctx for target pid\n");
return -EFAULT;
}
/*
* Find the task struct for the target pid
*/
task_target =
pid_task(find_vpid(eventfd_copy.target_pid), PIDTYPE_PID);
if (task_target == NULL) {
pr_debug("Failed to get mem ctx for target pid\n");
return -EFAULT;
}
files = get_files_struct (current);
if (files == NULL)
{
printk (KERN_DEBUG "Failed to get files struct\n");
return -EFAULT;
}
files = get_files_struct(current);
if (files == NULL) {
pr_debug("Failed to get files struct\n");
return -EFAULT;
}
rcu_read_lock ();
file = fcheck_files (files, eventfd_copy.source_fd);
if (file)
{
if (file->f_mode & FMODE_PATH
|| !atomic_long_inc_not_zero (&file->f_count))
file = NULL;
}
rcu_read_unlock ();
put_files_struct (files);
if (file == NULL)
{
printk (KERN_DEBUG "Failed to get file from source pid\n");
return 0;
}
/*
* Release the existing eventfd in the source process
*/
spin_lock (&files->file_lock);
filp_close (file, files);
fdt = files_fdtable (files);
fdt->fd[eventfd_copy.source_fd] = NULL;
spin_unlock (&files->file_lock);
/*
* Find the file struct associated with the target fd.
*/
files = get_files_struct (task_target);
if (files == NULL)
{
printk (KERN_DEBUG "Failed to get files struct\n");
return -EFAULT;
}
rcu_read_lock ();
file = fcheck_files (files, eventfd_copy.target_fd);
if (file)
{
if (file->f_mode & FMODE_PATH
|| !atomic_long_inc_not_zero (&file->f_count))
file = NULL;
}
rcu_read_unlock ();
put_files_struct (files);
if (file == NULL)
{
printk (KERN_DEBUG "Failed to get file from target pid\n");
return 0;
}
/*
* Install the file struct from the target process into the
* file desciptor of the source process,
*/
fd_install (eventfd_copy.source_fd, file);
rcu_read_lock();
file = fcheck_files(files, eventfd_copy.source_fd);
if (file) {
if (file->f_mode & FMODE_PATH ||
!atomic_long_inc_not_zero(&file->f_count))
file = NULL;
}
rcu_read_unlock();
put_files_struct(files);
if (file == NULL) {
pr_debug("Failed to get file from source pid\n");
return 0;
}
default:
return -ENOIOCTLCMD;
/*
* Release the existing eventfd in the source process
*/
spin_lock(&files->file_lock);
filp_close(file, files);
fdt = files_fdtable(files);
fdt->fd[eventfd_copy.source_fd] = NULL;
spin_unlock(&files->file_lock);
/*
* Find the file struct associated with the target fd.
*/
files = get_files_struct(task_target);
if (files == NULL) {
pr_debug("Failed to get files struct\n");
return -EFAULT;
}
rcu_read_lock();
file = fcheck_files(files, eventfd_copy.target_fd);
if (file) {
if (file->f_mode & FMODE_PATH ||
!atomic_long_inc_not_zero(&file->f_count))
file = NULL;
}
rcu_read_unlock();
put_files_struct(files);
if (file == NULL) {
pr_debug("Failed to get file from target pid\n");
return 0;
}
/*
* Install the file struct from the target process into the
* file desciptor of the source process,
*/
fd_install(eventfd_copy.source_fd, file);
return 0;
default:
return -ENOIOCTLCMD;
}
}
@ -183,23 +173,23 @@ static struct miscdevice eventfd_link_misc = {
};
static int __init
eventfd_link_init (void)
eventfd_link_init(void)
{
return misc_register (&eventfd_link_misc);
return misc_register(&eventfd_link_misc);
}
module_init (eventfd_link_init);
module_init(eventfd_link_init);
static void __exit
eventfd_link_exit (void)
eventfd_link_exit(void)
{
misc_deregister (&eventfd_link_misc);
misc_deregister(&eventfd_link_misc);
}
module_exit (eventfd_link_exit);
module_exit(eventfd_link_exit);
MODULE_VERSION ("0.0.1");
MODULE_LICENSE ("GPL v2");
MODULE_AUTHOR ("Anthony Fee");
MODULE_DESCRIPTION ("Link eventfd");
MODULE_ALIAS ("devname:eventfd-link");
MODULE_VERSION("0.0.1");
MODULE_LICENSE("GPL v2");
MODULE_AUTHOR("Anthony Fee");
MODULE_DESCRIPTION("Link eventfd");
MODULE_ALIAS("devname:eventfd-link");

127
lib/librte_vhost/eventfd_link/eventfd_link.h
View File

@ -1,79 +1,76 @@
/*-
* * This file is provided under a dual BSD/GPLv2 license. When using or
* * redistributing this file, you may do so under either license.
* *
* * GPL LICENSE SUMMARY
* *
* * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
* *
* * This program is free software; you can redistribute it and/or modify
* * it under the terms of version 2 of the GNU General Public License as
* * published by the Free Software Foundation.
* *
* * This program is distributed in the hope that it will be useful, but
* * WITHOUT ANY WARRANTY; without even the implied warranty of
* * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* * General Public License 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., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
* * The full GNU General Public License is included in this distribution
* * in the file called LICENSE.GPL.
* *
* * Contact Information:
* * Intel Corporation
* *
* * BSD LICENSE
* *
* * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
* * All rights reserved.
* *
* * Redistribution and use in source and binary forms, with or without
* * modification, are permitted provided that the following conditions
* * are met:
* *
* * * Redistributions of source code must retain the above copyright
* * notice, this list of conditions and the following disclaimer.
* * * Redistributions in binary form must reproduce the above copyright
* * notice, this list of conditions and the following disclaimer in
* * the documentation and/or other materials provided with the
* * distribution.
* * * Neither the name of Intel Corporation nor the names of its
* * contributors may be used to endorse or promote products derived
* * from this software without specific prior written permission.
* *
* * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
* *
* */
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
*
* GPL LICENSE SUMMARY
*
* Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of version 2 of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License 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., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
* The full GNU General Public License is included in this distribution
* in the file called LICENSE.GPL.
*
* Contact Information:
* Intel Corporation
*
* BSD LICENSE
*
* Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
* Neither the name of Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
*/
#ifndef _EVENTFD_LINK_H_
#define _EVENTFD_LINK_H_
/*
* ioctl to copy an fd entry in calling process to an fd in a target process
* ioctl to copy an fd entry in calling process to an fd in a target process
*/
#define EVENTFD_COPY 1
/*
* arguements for the EVENTFD_COPY ioctl
* arguements for the EVENTFD_COPY ioctl
*/
struct eventfd_copy {
// fd in the target pid
unsigned target_fd;
// fd in the calling pid
unsigned source_fd;
// pid of the target pid
pid_t target_pid;
unsigned target_fd; /* fd in the target pid */
unsigned source_fd; /* fd in the calling pid */
pid_t target_pid; /* pid of the target pid */
};
#endif /* _EVENTFD_LINK_H_ */

3
lib/librte_vhost/rte_virtio_net.h
View File

@ -90,8 +90,7 @@ struct vhost_virtqueue {
/**
* Device structure contains all configuration information relating to the device.
*/
struct virtio_net
{
struct virtio_net {
struct vhost_virtqueue *virtqueue[VIRTIO_QNUM]; /**< Contains all virtqueue information. */
struct virtio_memory *mem; /**< QEMU memory and memory region information. */
uint64_t features; /**< Negotiated feature set. */

187
lib/librte_vhost/vhost-net-cdev.c
View File

@ -46,21 +46,21 @@
#include "vhost-net-cdev.h"
#define FUSE_OPT_DUMMY "\0\0"
#define FUSE_OPT_FORE "-f\0\0"
#define FUSE_OPT_NOMULTI "-s\0\0"
#define FUSE_OPT_DUMMY "\0\0"
#define FUSE_OPT_FORE "-f\0\0"
#define FUSE_OPT_NOMULTI "-s\0\0"
static const uint32_t default_major = 231;
static const uint32_t default_minor = 1;
static const char cuse_device_name[] = "/dev/cuse";
static const char default_cdev[] = "vhost-net";
static const uint32_t default_major = 231;
static const uint32_t default_minor = 1;
static const char cuse_device_name[] = "/dev/cuse";
static const char default_cdev[] = "vhost-net";
static struct fuse_session *session;
static struct vhost_net_device_ops const *ops;
static struct fuse_session *session;
static struct vhost_net_device_ops const *ops;
/*
* Returns vhost_device_ctx from given fuse_req_t. The index is populated later when
* the device is added to the device linked list.
* Returns vhost_device_ctx from given fuse_req_t. The index is populated later
* when the device is added to the device linked list.
*/
static struct vhost_device_ctx
fuse_req_to_vhost_ctx(fuse_req_t req, struct fuse_file_info *fi)
@ -75,7 +75,8 @@ fuse_req_to_vhost_ctx(fuse_req_t req, struct fuse_file_info *fi)
}
/*
* When the device is created in QEMU it gets initialised here and added to the device linked list.
* When the device is created in QEMU it gets initialised here and
* added to the device linked list.
*/
static void
vhost_net_open(fuse_req_t req, struct fuse_file_info *fi)
@ -91,7 +92,8 @@ vhost_net_open(fuse_req_t req, struct fuse_file_info *fi)
fi->fh = err;
RTE_LOG(INFO, VHOST_CONFIG, "(%"PRIu64") Device configuration started\n", fi->fh);
RTE_LOG(INFO, VHOST_CONFIG,
"(%"PRIu64") Device configuration started\n", fi->fh);
fuse_reply_open(req, fi);
}
@ -113,8 +115,8 @@ vhost_net_release(fuse_req_t req, struct fuse_file_info *fi)
* Boilerplate code for CUSE IOCTL
* Implicit arguments: ctx, req, result.
*/
#define VHOST_IOCTL(func) do { \
result = (func)(ctx); \
#define VHOST_IOCTL(func) do { \
result = (func)(ctx); \
fuse_reply_ioctl(req, result, NULL, 0); \
} while (0)
@ -122,57 +124,58 @@ vhost_net_release(fuse_req_t req, struct fuse_file_info *fi)
* Boilerplate IOCTL RETRY
* Implicit arguments: req.
*/
#define VHOST_IOCTL_RETRY(size_r, size_w) do { \
struct iovec iov_r = { arg, (size_r) }; \
struct iovec iov_w = { arg, (size_w) }; \
fuse_reply_ioctl_retry(req, &iov_r, (size_r)?1:0, &iov_w, (size_w)?1:0); \
#define VHOST_IOCTL_RETRY(size_r, size_w) do { \
struct iovec iov_r = { arg, (size_r) }; \
struct iovec iov_w = { arg, (size_w) }; \
fuse_reply_ioctl_retry(req, &iov_r, \
(size_r) ? 1 : 0, &iov_w, (size_w) ? 1 : 0);\
} while (0)
/*
* Boilerplate code for CUSE Read IOCTL
* Implicit arguments: ctx, req, result, in_bufsz, in_buf.
*/
#define VHOST_IOCTL_R(type, var, func) do { \
if (!in_bufsz) { \
VHOST_IOCTL_RETRY(sizeof(type), 0); \
} else { \
(var) = *(const type*) in_buf; \
result = func(ctx, &(var)); \
fuse_reply_ioctl(req, result, NULL, 0); \
} \
#define VHOST_IOCTL_R(type, var, func) do { \
if (!in_bufsz) { \
VHOST_IOCTL_RETRY(sizeof(type), 0);\
} else { \
(var) = *(const type*)in_buf; \
result = func(ctx, &(var)); \
fuse_reply_ioctl(req, result, NULL, 0);\
} \
} while (0)
/*
* Boilerplate code for CUSE Write IOCTL
* Boilerplate code for CUSE Write IOCTL
* Implicit arguments: ctx, req, result, out_bufsz.
*/
#define VHOST_IOCTL_W(type, var, func) do { \
if (!out_bufsz) { \
VHOST_IOCTL_RETRY(0, sizeof(type)); \
} else { \
result = (func)(ctx, &(var)); \
fuse_reply_ioctl(req, result, &(var), sizeof(type)); \
} \
#define VHOST_IOCTL_W(type, var, func) do { \
if (!out_bufsz) { \
VHOST_IOCTL_RETRY(0, sizeof(type));\
} else { \
result = (func)(ctx, &(var));\
fuse_reply_ioctl(req, result, &(var), sizeof(type));\
} \
} while (0)
/*
* Boilerplate code for CUSE Read/Write IOCTL
* Implicit arguments: ctx, req, result, in_bufsz, in_buf.
*/
#define VHOST_IOCTL_RW(type1, var1, type2, var2, func) do { \
if (!in_bufsz) { \
VHOST_IOCTL_RETRY(sizeof(type1), sizeof(type2)); \
} else { \
(var1) = *(const type1*) (in_buf); \
result = (func)(ctx, (var1), &(var2)); \
fuse_reply_ioctl(req, result, &(var2), sizeof(type2)); \
} \
#define VHOST_IOCTL_RW(type1, var1, type2, var2, func) do { \
if (!in_bufsz) { \
VHOST_IOCTL_RETRY(sizeof(type1), sizeof(type2));\
} else { \
(var1) = *(const type1*) (in_buf); \
result = (func)(ctx, (var1), &(var2)); \
fuse_reply_ioctl(req, result, &(var2), sizeof(type2));\
} \
} while (0)
/*
* The IOCTLs are handled using CUSE/FUSE in userspace. Depending on
* the type of IOCTL a buffer is requested to read or to write. This
* request is handled by FUSE and the buffer is then given to CUSE.
* The IOCTLs are handled using CUSE/FUSE in userspace. Depending on the type
* of IOCTL a buffer is requested to read or to write. This request is handled
* by FUSE and the buffer is then given to CUSE.
*/
static void
vhost_net_ioctl(fuse_req_t req, int cmd, void *arg,
@ -189,33 +192,39 @@ vhost_net_ioctl(fuse_req_t req, int cmd, void *arg,
switch (cmd) {
case VHOST_NET_SET_BACKEND:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_NET_SET_BACKEND\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_NET_SET_BACKEND\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_file, file, ops->set_backend);
break;
case VHOST_GET_FEATURES:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_GET_FEATURES\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_GET_FEATURES\n", ctx.fh);
VHOST_IOCTL_W(uint64_t, features, ops->get_features);
break;
case VHOST_SET_FEATURES:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_FEATURES\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_FEATURES\n", ctx.fh);
VHOST_IOCTL_R(uint64_t, features, ops->set_features);
break;
case VHOST_RESET_OWNER:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_RESET_OWNER\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_RESET_OWNER\n", ctx.fh);
VHOST_IOCTL(ops->reset_owner);
break;
case VHOST_SET_OWNER:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_OWNER\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_OWNER\n", ctx.fh);
VHOST_IOCTL(ops->set_owner);
break;
case VHOST_SET_MEM_TABLE:
/*TODO fix race condition.*/
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_MEM_TABLE\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_MEM_TABLE\n", ctx.fh);
static struct vhost_memory mem_temp;
switch (in_bufsz) {
@ -227,7 +236,9 @@ vhost_net_ioctl(fuse_req_t req, int cmd, void *arg,
mem_temp = *(const struct vhost_memory *) in_buf;
if (mem_temp.nregions > 0) {
VHOST_IOCTL_RETRY(sizeof(struct vhost_memory) + (sizeof(struct vhost_memory_region) * mem_temp.nregions), 0);
VHOST_IOCTL_RETRY(sizeof(struct vhost_memory) +
(sizeof(struct vhost_memory_region) *
mem_temp.nregions), 0);
} else {
result = -1;
fuse_reply_ioctl(req, result, NULL, 0);
@ -235,56 +246,70 @@ vhost_net_ioctl(fuse_req_t req, int cmd, void *arg,
break;
default:
result = ops->set_mem_table(ctx, in_buf, mem_temp.nregions);
result = ops->set_mem_table(ctx,
in_buf, mem_temp.nregions);
if (result)
fuse_reply_err(req, EINVAL);
else
fuse_reply_ioctl(req, result, NULL, 0);
}
break;
case VHOST_SET_VRING_NUM:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_VRING_NUM\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_state, state, ops->set_vring_num);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_VRING_NUM\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_state, state,
ops->set_vring_num);
break;
case VHOST_SET_VRING_BASE:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_VRING_BASE\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_state, state, ops->set_vring_base);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_VRING_BASE\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_state, state,
ops->set_vring_base);
break;
case VHOST_GET_VRING_BASE:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_GET_VRING_BASE\n", ctx.fh);
VHOST_IOCTL_RW(uint32_t, index, struct vhost_vring_state, state, ops->get_vring_base);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_GET_VRING_BASE\n", ctx.fh);
VHOST_IOCTL_RW(uint32_t, index,
struct vhost_vring_state, state, ops->get_vring_base);
break;
case VHOST_SET_VRING_ADDR:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_VRING_ADDR\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_addr, addr, ops->set_vring_addr);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_VRING_ADDR\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_addr, addr,
ops->set_vring_addr);
break;
case VHOST_SET_VRING_KICK:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_VRING_KICK\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_file, file, ops->set_vring_kick);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_VRING_KICK\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_file, file,
ops->set_vring_kick);
break;
case VHOST_SET_VRING_CALL:
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: VHOST_SET_VRING_CALL\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_file, file, ops->set_vring_call);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: VHOST_SET_VRING_CALL\n", ctx.fh);
VHOST_IOCTL_R(struct vhost_vring_file, file,
ops->set_vring_call);
break;
default:
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") IOCTL: DOESN NOT EXIST\n", ctx.fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") IOCTL: DOESN NOT EXIST\n", ctx.fh);
result = -1;
fuse_reply_ioctl(req, result, NULL, 0);
}
if (result < 0)
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: FAIL\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: FAIL\n", ctx.fh);
else
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") IOCTL: SUCCESS\n", ctx.fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") IOCTL: SUCCESS\n", ctx.fh);
}
/*
@ -297,8 +322,8 @@ static const struct cuse_lowlevel_ops vhost_net_ops = {
};
/*
* cuse_info is populated and used to register the cuse device. vhost_net_device_ops are
* also passed when the device is registered in main.c.
* cuse_info is populated and used to register the cuse device.
* vhost_net_device_ops are also passed when the device is registered in app.
*/
int
rte_vhost_driver_register(const char *dev_name)
@ -314,20 +339,23 @@ rte_vhost_driver_register(const char *dev_name)
char *fuse_argv[] = {fuse_opt_dummy, fuse_opt_fore, fuse_opt_nomulti};
if (access(cuse_device_name, R_OK | W_OK) < 0) {
RTE_LOG(ERR, VHOST_CONFIG, "Character device %s can't be accessed, maybe not exist\n", cuse_device_name);
RTE_LOG(ERR, VHOST_CONFIG,
"char device %s can't be accessed, maybe not exist\n",
cuse_device_name);
return -1;
}
/*
* The device name is created. This is passed to QEMU so that it can register
* the device with our application.
* The device name is created. This is passed to QEMU so that it can
* register the device with our application.
*/
snprintf(device_name, PATH_MAX, "DEVNAME=%s", dev_name);
snprintf(char_device_name, PATH_MAX, "/dev/%s", dev_name);
/* Check if device already exists. */
if (access(char_device_name, F_OK) != -1) {
RTE_LOG(ERR, VHOST_CONFIG, "Character device %s already exists\n", char_device_name);
RTE_LOG(ERR, VHOST_CONFIG,
"char device %s already exists\n", char_device_name);
return -1;
}
@ -341,7 +369,7 @@ rte_vhost_driver_register(const char *dev_name)
ops = get_virtio_net_callbacks();
session = cuse_lowlevel_setup(3, fuse_argv,
&cuse_info, &vhost_net_ops, 0, NULL);
&cuse_info, &vhost_net_ops, 0, NULL);
if (session == NULL)
return -1;
@ -349,7 +377,8 @@ rte_vhost_driver_register(const char *dev_name)
}
/**
* The CUSE session is launched allowing the application to receive open, release and ioctl calls.
* The CUSE session is launched allowing the application to receive open,
* release and ioctl calls.
*/
int
rte_vhost_driver_session_start(void)

13
lib/librte_vhost/vhost_rxtx.c
View File

@ -45,7 +45,7 @@
/**
* This function adds buffers to the virtio devices RX virtqueue. Buffers can
* be received from the physical port or from another virtio device. A packet
* count is returned to indicate the number of packets that were succesfully
* count is returned to indicate the number of packets that are succesfully
* added to the RX queue. This function works when mergeable is disabled.
*/
static inline uint32_t __attribute__((always_inline))
@ -76,7 +76,7 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id,
count = (count > MAX_PKT_BURST) ? MAX_PKT_BURST : count;
/*
* As many data cores may want access to available buffers,
* As many data cores may want access to available buffers,
* they need to be reserved.
*/
do {
@ -143,7 +143,8 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id,
}
/* Update used ring with desc information */
vq->used->ring[res_cur_idx & (vq->size - 1)].id = head[packet_success];
vq->used->ring[res_cur_idx & (vq->size - 1)].id =
head[packet_success];
vq->used->ring[res_cur_idx & (vq->size - 1)].len = packet_len;
/* Copy mbuf data to buffer */
@ -389,10 +390,7 @@ copy_from_mbuf_to_vring(struct virtio_net *dev, uint16_t res_base_idx,
}
/*
* This function adds buffers to the virtio devices RX virtqueue. Buffers can
* be received from the physical port or from another virtio device. A packet
* count is returned to indicate the number of packets that were succesfully
* added to the RX queue. This function works for mergeable RX.
* This function works for mergeable RX.
*/
static inline uint32_t __attribute__((always_inline))
virtio_dev_merge_rx(struct virtio_net *dev, uint16_t queue_id,
@ -729,5 +727,4 @@ rte_vhost_dequeue_burst(struct virtio_net *dev, uint16_t queue_id,
if (!(vq->avail->flags & VRING_AVAIL_F_NO_INTERRUPT))
eventfd_write((int)vq->kickfd, 1);
return entry_success;
}

317
lib/librte_vhost/virtio-net.c
View File

@ -52,27 +52,27 @@
#include "vhost-net-cdev.h"
#include "eventfd_link/eventfd_link.h"
/**
/*
* Device linked list structure for configuration.
*/
struct virtio_net_config_ll {
struct virtio_net dev; /* Virtio device.*/
struct virtio_net_config_ll *next; /* Next entry on linked list.*/
struct virtio_net dev; /* Virtio device.*/
struct virtio_net_config_ll *next; /* Next dev on linked list.*/
};
const char eventfd_cdev[] = "/dev/eventfd-link";
/* device ops to add/remove device to data core. */
/* device ops to add/remove device to/from data core. */
static struct virtio_net_device_ops const *notify_ops;
/* Root address of the linked list in the configuration core. */
static struct virtio_net_config_ll *ll_root;
/* root address of the linked list of managed virtio devices */
static struct virtio_net_config_ll *ll_root;
/* Features supported by this application. RX merge buffers are enabled by default. */
/* Features supported by this lib. */
#define VHOST_SUPPORTED_FEATURES (1ULL << VIRTIO_NET_F_MRG_RXBUF)
static uint64_t VHOST_FEATURES = VHOST_SUPPORTED_FEATURES;
/* Line size for reading maps file. */
const uint32_t BUFSIZE = PATH_MAX;
static const uint32_t BUFSIZE = PATH_MAX;
/* Size of prot char array in procmap. */
#define PROT_SZ 5
@ -82,19 +82,19 @@ const uint32_t BUFSIZE = PATH_MAX;
/* Structure containing information gathered from maps file. */
struct procmap {
uint64_t va_start; /* Start virtual address in file. */
uint64_t len; /* Size of file. */
uint64_t pgoff; /* Not used. */
uint32_t maj; /* Not used. */
uint32_t min; /* Not used. */
uint32_t ino; /* Not used. */
char prot[PROT_SZ]; /* Not used. */
char fname[PATH_MAX];/* File name. */
uint64_t va_start; /* Start virtual address in file. */
uint64_t len; /* Size of file. */
uint64_t pgoff; /* Not used. */
uint32_t maj; /* Not used. */
uint32_t min; /* Not used. */
uint32_t ino; /* Not used. */
char prot[PROT_SZ]; /* Not used. */
char fname[PATH_MAX]; /* File name. */
};
/*
* Converts QEMU virtual address to Vhost virtual address. This function is used
* to convert the ring addresses to our address space.
* Converts QEMU virtual address to Vhost virtual address. This function is
* used to convert the ring addresses to our address space.
*/
static uint64_t
qva_to_vva(struct virtio_net *dev, uint64_t qemu_va)
@ -107,8 +107,8 @@ qva_to_vva(struct virtio_net *dev, uint64_t qemu_va)
for (regionidx = 0; regionidx < dev->mem->nregions; regionidx++) {
region = &dev->mem->regions[regionidx];
if ((qemu_va >= region->userspace_address) &&
(qemu_va <= region->userspace_address +
region->memory_size)) {
(qemu_va <= region->userspace_address +
region->memory_size)) {
vhost_va = dev->mem->mapped_address + qemu_va -
dev->mem->base_address;
break;
@ -118,7 +118,8 @@ qva_to_vva(struct virtio_net *dev, uint64_t qemu_va)
}
/*
* Locate the file containing QEMU's memory space and map it to our address space.
* Locate the file containing QEMU's memory space and
* map it to our address space.
*/
static int
host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
@ -134,10 +135,10 @@ host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
char procdir[PATH_MAX];
char resolved_path[PATH_MAX];
char *path = NULL;
FILE *fmap;
void *map;
uint8_t found = 0;
char line[BUFSIZE];
FILE *fmap;
void *map;
uint8_t found = 0;
char line[BUFSIZE];
char dlm[] = "- : ";
char *str, *sp, *in[PROCMAP_SZ];
char *end = NULL;
@ -159,7 +160,7 @@ host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
while (fgets(line, BUFSIZE, fmap) != 0) {
str = line;
errno = 0;
/* Split line in to fields. */
/* Split line into fields. */
for (i = 0; i < PROCMAP_SZ; i++) {
in[i] = strtok_r(str, &dlm[i], &sp);
if ((in[i] == NULL) || (errno != 0)) {
@ -171,37 +172,43 @@ host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
/* Convert/Copy each field as needed. */
procmap.va_start = strtoull(in[0], &end, 16);
if ((in[0] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) {
if ((in[0] == '\0') || (end == NULL) || (*end != '\0') ||
(errno != 0)) {
fclose(fmap);
return -1;
}
procmap.len = strtoull(in[1], &end, 16);
if ((in[1] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) {
if ((in[1] == '\0') || (end == NULL) || (*end != '\0') ||
(errno != 0)) {
fclose(fmap);
return -1;
}
procmap.pgoff = strtoull(in[3], &end, 16);
if ((in[3] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) {
if ((in[3] == '\0') || (end == NULL) || (*end != '\0') ||
(errno != 0)) {
fclose(fmap);
return -1;
}
procmap.maj = strtoul(in[4], &end, 16);
if ((in[4] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) {
if ((in[4] == '\0') || (end == NULL) || (*end != '\0') ||
(errno != 0)) {
fclose(fmap);
return -1;
}
procmap.min = strtoul(in[5], &end, 16);
if ((in[5] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) {
if ((in[5] == '\0') || (end == NULL) || (*end != '\0') ||
(errno != 0)) {
fclose(fmap);
return -1;
}
procmap.ino = strtoul(in[6], &end, 16);
if ((in[6] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) {
if ((in[6] == '\0') || (end == NULL) || (*end != '\0') ||
(errno != 0)) {
fclose(fmap);
return -1;
}
@ -218,16 +225,19 @@ host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
fclose(fmap);
if (!found) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to find memory file in pid %d maps file\n", dev->device_fh, pid);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to find memory file in pid %d maps file\n",
dev->device_fh, pid);
return -1;
}
/* Find the guest memory file among the process fds. */
dp = opendir(procdir);
if (dp == NULL) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Cannot open pid %d process directory\n", dev->device_fh, pid);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Cannot open pid %d process directory\n",
dev->device_fh, pid);
return -1;
}
found = 0;
@ -254,23 +264,29 @@ host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
closedir(dp);
if (found == 0) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to find memory file for pid %d\n", dev->device_fh, pid);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to find memory file for pid %d\n",
dev->device_fh, pid);
return -1;
}
/* Open the shared memory file and map the memory into this process. */
fd = open(memfile, O_RDWR);
if (fd == -1) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to open %s for pid %d\n", dev->device_fh, memfile, pid);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to open %s for pid %d\n",
dev->device_fh, memfile, pid);
return -1;
}
map = mmap(0, (size_t)procmap.len, PROT_READ|PROT_WRITE ,
MAP_POPULATE|MAP_SHARED, fd, 0);
map = mmap(0, (size_t)procmap.len, PROT_READ|PROT_WRITE,
MAP_POPULATE|MAP_SHARED, fd, 0);
close(fd);
if (map == MAP_FAILED) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Error mapping the file %s for pid %d\n", dev->device_fh, memfile, pid);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Error mapping the file %s for pid %d\n",
dev->device_fh, memfile, pid);
return -1;
}
@ -278,8 +294,11 @@ host_memory_map(struct virtio_net *dev, struct virtio_memory *mem,
mem->mapped_address = (uint64_t)(uintptr_t)map;
mem->mapped_size = procmap.len;
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") Mem File: %s->%s - Size: %llu - VA: %p\n", dev->device_fh,
memfile, resolved_path, (long long unsigned)mem->mapped_size, map);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") Mem File: %s->%s - Size: %llu - VA: %p\n",
dev->device_fh,
memfile, resolved_path,
(unsigned long long)mem->mapped_size, map);
return 0;
}
@ -303,7 +322,8 @@ get_config_ll_entry(struct vhost_device_ctx ctx)
}
/*
* Searches the configuration core linked list and retrieves the device if it exists.
* Searches the configuration core linked list and
* retrieves the device if it exists.
*/
static struct virtio_net *
get_device(struct vhost_device_ctx ctx)
@ -312,11 +332,11 @@ get_device(struct vhost_device_ctx ctx)
ll_dev = get_config_ll_entry(ctx);
/* If a matching entry is found in the linked list, return the device in that entry. */
if (ll_dev)
return &ll_dev->dev;
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Device not found in linked list.\n", ctx.fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Device not found in linked list.\n", ctx.fh);
return NULL;
}
@ -331,13 +351,18 @@ add_config_ll_entry(struct virtio_net_config_ll *new_ll_dev)
/* If ll_dev == NULL then this is the first device so go to else */
if (ll_dev) {
/* If the 1st device_fh != 0 then we insert our device here. */
if (ll_dev->dev.device_fh != 0) {
if (ll_dev->dev.device_fh != 0) {
new_ll_dev->dev.device_fh = 0;
new_ll_dev->next = ll_dev;
ll_root = new_ll_dev;
} else {
/* Increment through the ll until we find un unused device_fh. Insert the device at that entry*/
while ((ll_dev->next != NULL) && (ll_dev->dev.device_fh == (ll_dev->next->dev.device_fh - 1)))
/*
* Increment through the ll until we find un unused
* device_fh. Insert the device at that entry.
*/
while ((ll_dev->next != NULL) &&
(ll_dev->dev.device_fh ==
(ll_dev->next->dev.device_fh - 1)))
ll_dev = ll_dev->next;
new_ll_dev->dev.device_fh = ll_dev->dev.device_fh + 1;
@ -352,7 +377,8 @@ add_config_ll_entry(struct virtio_net_config_ll *new_ll_dev)
}
/*
* Unmap any memory, close any file descriptors and free any memory owned by a device.
* Unmap any memory, close any file descriptors and
* free any memory owned by a device.
*/
static void
cleanup_device(struct virtio_net *dev)
@ -386,6 +412,7 @@ free_device(struct virtio_net_config_ll *ll_dev)
free(ll_dev->dev.virtqueue[VIRTIO_TXQ]);
free(ll_dev);
}
/*
* Remove an entry from the device configuration linked list.
*/
@ -423,7 +450,10 @@ init_device(struct virtio_net *dev)
{
uint64_t vq_offset;
/* Virtqueues have already been malloced so we don't want to set them to NULL. */
/*
* Virtqueues have already been malloced so
* we don't want to set them to NULL.
*/
vq_offset = offsetof(struct virtio_net, mem);
/* Set everything to 0. */
@ -491,8 +521,8 @@ new_device(struct vhost_device_ctx ctx)
}
/*
* Function is called from the CUSE release function. This function will cleanup
* the device and remove it from device configuration linked list.
* Function is called from the CUSE release function. This function will
* cleanup the device and remove it from device configuration linked list.
*/
static void
destroy_device(struct vhost_device_ctx ctx)
@ -503,15 +533,19 @@ destroy_device(struct vhost_device_ctx ctx)
/* Find the linked list entry for the device to be removed. */
ll_dev_cur_ctx = get_config_ll_entry(ctx);
while (ll_dev_cur != NULL) {
/* If the device is found or a device that doesn't exist is found then it is removed. */
/*
* If the device is found or
* a device that doesn't exist is found then it is removed.
*/
if (ll_dev_cur == ll_dev_cur_ctx) {
/*
* If the device is running on a data core then call the function to remove it from
* the data core.
* If the device is running on a data core then call
* the function to remove it from the data core.
*/
if ((ll_dev_cur->dev.flags & VIRTIO_DEV_RUNNING))
notify_ops->destroy_device(&(ll_dev_cur->dev));
ll_dev_cur = rm_config_ll_entry(ll_dev_cur, ll_dev_last);
ll_dev_cur = rm_config_ll_entry(ll_dev_cur,
ll_dev_last);
} else {
ll_dev_last = ll_dev_cur;
ll_dev_cur = ll_dev_cur->next;
@ -521,7 +555,8 @@ destroy_device(struct vhost_device_ctx ctx)
/*
* Called from CUSE IOCTL: VHOST_SET_OWNER
* This function just returns success at the moment unless the device hasn't been initialised.
* This function just returns success at the moment unless
* the device hasn't been initialised.
*/
static int
set_owner(struct vhost_device_ctx ctx)
@ -571,7 +606,7 @@ get_features(struct vhost_device_ctx ctx, uint64_t *pu)
/*
* Called from CUSE IOCTL: VHOST_SET_FEATURES
* We receive the negotiated set of features supported by us and the virtio device.
* We receive the negotiated features supported by us and the virtio device.
*/
static int
set_features(struct vhost_device_ctx ctx, uint64_t *pu)
@ -589,13 +624,17 @@ set_features(struct vhost_device_ctx ctx, uint64_t *pu)
/* Set the vhost_hlen depending on if VIRTIO_NET_F_MRG_RXBUF is set. */
if (dev->features & (1 << VIRTIO_NET_F_MRG_RXBUF)) {
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") Mergeable RX buffers enabled\n", dev->device_fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") Mergeable RX buffers enabled\n",
dev->device_fh);
dev->virtqueue[VIRTIO_RXQ]->vhost_hlen =
sizeof(struct virtio_net_hdr_mrg_rxbuf);
dev->virtqueue[VIRTIO_TXQ]->vhost_hlen =
sizeof(struct virtio_net_hdr_mrg_rxbuf);
} else {
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") Mergeable RX buffers disabled\n", dev->device_fh);
LOG_DEBUG(VHOST_CONFIG,
"(%"PRIu64") Mergeable RX buffers disabled\n",
dev->device_fh);
dev->virtqueue[VIRTIO_RXQ]->vhost_hlen =
sizeof(struct virtio_net_hdr);
dev->virtqueue[VIRTIO_TXQ]->vhost_hlen =
@ -607,8 +646,8 @@ set_features(struct vhost_device_ctx ctx, uint64_t *pu)
/*
* Called from CUSE IOCTL: VHOST_SET_MEM_TABLE
* This function creates and populates the memory structure for the device. This includes
* storing offsets used to translate buffer addresses.
* This function creates and populates the memory structure for the device.
* This includes storing offsets used to translate buffer addresses.
*/
static int
set_mem_table(struct vhost_device_ctx ctx, const void *mem_regions_addr,
@ -634,7 +673,9 @@ set_mem_table(struct vhost_device_ctx ctx, const void *mem_regions_addr,
mem = calloc(1, sizeof(struct virtio_memory) +
(sizeof(struct virtio_memory_regions) * nregions));
if (mem == NULL) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to allocate memory for dev->mem.\n", dev->device_fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to allocate memory for dev->mem.\n",
dev->device_fh);
return -1;
}
@ -656,15 +697,18 @@ set_mem_table(struct vhost_device_ctx ctx, const void *mem_regions_addr,
mem_regions[regionidx].userspace_addr;
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") REGION: %u - GPA: %p - QEMU VA: %p - SIZE (%"PRIu64")\n", dev->device_fh,
regionidx, (void *)(uintptr_t)mem->regions[regionidx].guest_phys_address,
(void *)(uintptr_t)mem->regions[regionidx].userspace_address,
mem->regions[regionidx].memory_size);
regionidx,
(void *)(uintptr_t)mem->regions[regionidx].guest_phys_address,
(void *)(uintptr_t)mem->regions[regionidx].userspace_address,
mem->regions[regionidx].memory_size);
/*set the base address mapping*/
if (mem->regions[regionidx].guest_phys_address == 0x0) {
mem->base_address = mem->regions[regionidx].userspace_address;
mem->base_address =
mem->regions[regionidx].userspace_address;
/* Map VM memory file */
if (host_memory_map(dev, mem, ctx.pid, mem->base_address) != 0) {
if (host_memory_map(dev, mem, ctx.pid,
mem->base_address) != 0) {
free(mem);
return -1;
}
@ -678,27 +722,42 @@ set_mem_table(struct vhost_device_ctx ctx, const void *mem_regions_addr,
return -1;
}
/* Check if all of our regions have valid mappings. Usually one does not exist in the QEMU memory file. */
/*
* Check if all of our regions have valid mappings.
* Usually one does not exist in the QEMU memory file.
*/
valid_regions = mem->nregions;
for (regionidx = 0; regionidx < mem->nregions; regionidx++) {
if ((mem->regions[regionidx].userspace_address < mem->base_address) ||
(mem->regions[regionidx].userspace_address > (mem->base_address + mem->mapped_size)))
if ((mem->regions[regionidx].userspace_address <
mem->base_address) ||
(mem->regions[regionidx].userspace_address >
(mem->base_address + mem->mapped_size)))
valid_regions--;
}
/* If a region does not have a valid mapping we rebuild our memory struct to contain only valid entries. */
/*
* If a region does not have a valid mapping,
* we rebuild our memory struct to contain only valid entries.
*/
if (valid_regions != mem->nregions) {
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") Not all memory regions exist in the QEMU mem file. Re-populating mem structure\n",
dev->device_fh);
/* Re-populate the memory structure with only valid regions. Invalid regions are over-written with memmove. */
/*
* Re-populate the memory structure with only valid regions.
* Invalid regions are over-written with memmove.
*/
valid_regions = 0;
for (regionidx = mem->nregions; 0 != regionidx--;) {
if ((mem->regions[regionidx].userspace_address < mem->base_address) ||
(mem->regions[regionidx].userspace_address > (mem->base_address + mem->mapped_size))) {
memmove(&mem->regions[regionidx], &mem->regions[regionidx + 1],
sizeof(struct virtio_memory_regions) * valid_regions);
if ((mem->regions[regionidx].userspace_address <
mem->base_address) ||
(mem->regions[regionidx].userspace_address >
(mem->base_address + mem->mapped_size))) {
memmove(&mem->regions[regionidx],
&mem->regions[regionidx + 1],
sizeof(struct virtio_memory_regions) *
valid_regions);
} else {
valid_regions++;
}
@ -708,12 +767,16 @@ set_mem_table(struct vhost_device_ctx ctx, const void *mem_regions_addr,
dev->mem = mem;
/*
* Calculate the address offset for each region. This offset is used to identify the vhost virtual address
* Calculate the address offset for each region.
* This offset is used to identify the vhost virtual address
* corresponding to a QEMU guest physical address.
*/
for (regionidx = 0; regionidx < dev->mem->nregions; regionidx++) {
dev->mem->regions[regionidx].address_offset = dev->mem->regions[regionidx].userspace_address - dev->mem->base_address
+ dev->mem->mapped_address - dev->mem->regions[regionidx].guest_phys_address;
dev->mem->regions[regionidx].address_offset =
dev->mem->regions[regionidx].userspace_address -
dev->mem->base_address +
dev->mem->mapped_address -
dev->mem->regions[regionidx].guest_phys_address;
}
return 0;
@ -732,7 +795,7 @@ set_vring_num(struct vhost_device_ctx ctx, struct vhost_vring_state *state)
if (dev == NULL)
return -1;
/* State->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* State->index refers to the queue index. The txq is 1, rxq is 0. */
dev->virtqueue[state->index]->size = state->num;
return 0;
@ -740,8 +803,8 @@ set_vring_num(struct vhost_device_ctx ctx, struct vhost_vring_state *state)
/*
* Called from CUSE IOCTL: VHOST_SET_VRING_ADDR
* The virtio device sends us the desc, used and avail ring addresses. This function
* then converts these to our address space.
* The virtio device sends us the desc, used and avail ring addresses.
* This function then converts these to our address space.
*/
static int
set_vring_addr(struct vhost_device_ctx ctx, struct vhost_vring_addr *addr)
@ -753,31 +816,43 @@ set_vring_addr(struct vhost_device_ctx ctx, struct vhost_vring_addr *addr)
if (dev == NULL)
return -1;
/* addr->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* addr->index refers to the queue index. The txq 1, rxq is 0. */
vq = dev->virtqueue[addr->index];
/* The addresses are converted from QEMU virtual to Vhost virtual. */
vq->desc = (struct vring_desc *)(uintptr_t)qva_to_vva(dev, addr->desc_user_addr);
vq->desc = (struct vring_desc *)(uintptr_t)qva_to_vva(dev,
addr->desc_user_addr);
if (vq->desc == 0) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to find descriptor ring address.\n", dev->device_fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to find desc ring address.\n",
dev->device_fh);
return -1;
}
vq->avail = (struct vring_avail *)(uintptr_t)qva_to_vva(dev, addr->avail_user_addr);
vq->avail = (struct vring_avail *)(uintptr_t)qva_to_vva(dev,
addr->avail_user_addr);
if (vq->avail == 0) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to find available ring address.\n", dev->device_fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to find avail ring address.\n",
dev->device_fh);
return -1;
}
vq->used = (struct vring_used *)(uintptr_t)qva_to_vva(dev, addr->used_user_addr);
vq->used = (struct vring_used *)(uintptr_t)qva_to_vva(dev,
addr->used_user_addr);
if (vq->used == 0) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") Failed to find used ring address.\n", dev->device_fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") Failed to find used ring address.\n",
dev->device_fh);
return -1;
}
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address desc: %p\n", dev->device_fh, vq->desc);
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address avail: %p\n", dev->device_fh, vq->avail);
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address used: %p\n", dev->device_fh, vq->used);
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address desc: %p\n",
dev->device_fh, vq->desc);
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address avail: %p\n",
dev->device_fh, vq->avail);
LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address used: %p\n",
dev->device_fh, vq->used);
return 0;
}
@ -795,7 +870,7 @@ set_vring_base(struct vhost_device_ctx ctx, struct vhost_vring_state *state)
if (dev == NULL)
return -1;
/* State->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* State->index refers to the queue index. The txq is 1, rxq is 0. */
dev->virtqueue[state->index]->last_used_idx = state->num;
dev->virtqueue[state->index]->last_used_idx_res = state->num;
@ -817,15 +892,15 @@ get_vring_base(struct vhost_device_ctx ctx, uint32_t index,
return -1;
state->index = index;
/* State->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* State->index refers to the queue index. The txq is 1, rxq is 0. */
state->num = dev->virtqueue[state->index]->last_used_idx;
return 0;
}
/*
* This function uses the eventfd_link kernel module to copy an eventfd file descriptor
* provided by QEMU in to our process space.
* This function uses the eventfd_link kernel module to copy an eventfd file
* descriptor provided by QEMU in to our process space.
*/
static int
eventfd_copy(struct virtio_net *dev, struct eventfd_copy *eventfd_copy)
@ -835,7 +910,9 @@ eventfd_copy(struct virtio_net *dev, struct eventfd_copy *eventfd_copy)
/* Open the character device to the kernel module. */
eventfd_link = open(eventfd_cdev, O_RDWR);
if (eventfd_link < 0) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") eventfd_link module is not loaded\n", dev->device_fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") eventfd_link module is not loaded\n",
dev->device_fh);
return -1;
}
@ -844,18 +921,19 @@ eventfd_copy(struct virtio_net *dev, struct eventfd_copy *eventfd_copy)
close(eventfd_link);
if (ret < 0) {
RTE_LOG(ERR, VHOST_CONFIG, "(%"PRIu64") EVENTFD_COPY ioctl failed\n", dev->device_fh);
RTE_LOG(ERR, VHOST_CONFIG,
"(%"PRIu64") EVENTFD_COPY ioctl failed\n",
dev->device_fh);
return -1;
}
return 0;
}
/*
* Called from CUSE IOCTL: VHOST_SET_VRING_CALL
* The virtio device sends an eventfd to interrupt the guest. This fd gets copied in
* to our process space.
* The virtio device sends an eventfd to interrupt the guest. This fd gets
* copied into our process space.
*/
static int
set_vring_call(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
@ -868,7 +946,7 @@ set_vring_call(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
if (dev == NULL)
return -1;
/* file->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* file->index refers to the queue index. The txq is 1, rxq is 0. */
vq = dev->virtqueue[file->index];
if (vq->kickfd)
@ -888,8 +966,8 @@ set_vring_call(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
/*
* Called from CUSE IOCTL: VHOST_SET_VRING_KICK
* The virtio device sends an eventfd that it can use to notify us. This fd gets copied in
* to our process space.
* The virtio device sends an eventfd that it can use to notify us.
* This fd gets copied into our process space.
*/
static int
set_vring_kick(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
@ -902,7 +980,7 @@ set_vring_kick(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
if (dev == NULL)
return -1;
/* file->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* file->index refers to the queue index. The txq is 1, rxq is 0. */
vq = dev->virtqueue[file->index];
if (vq->callfd)
@ -922,10 +1000,12 @@ set_vring_kick(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
/*
* Called from CUSE IOCTL: VHOST_NET_SET_BACKEND
* To complete device initialisation when the virtio driver is loaded we are provided with a
* valid fd for a tap device (not used by us). If this happens then we can add the device to a
* data core. When the virtio driver is removed we get fd=-1. At that point we remove the device
* from the data core. The device will still exist in the device configuration linked list.
* To complete device initialisation when the virtio driver is loaded,
* we are provided with a valid fd for a tap device (not used by us).
* If this happens then we can add the device to a data core.
* When the virtio driver is removed we get fd=-1.
* At that point we remove the device from the data core.
* The device will still exist in the device configuration linked list.
*/
static int
set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
@ -936,10 +1016,13 @@ set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
if (dev == NULL)
return -1;
/* file->index refers to the queue index. The TX queue is 1, RX queue is 0. */
/* file->index refers to the queue index. The txq is 1, rxq is 0. */
dev->virtqueue[file->index]->backend = file->fd;
/* If the device isn't already running and both backend fds are set we add the device. */
/*
* If the device isn't already running and both backend fds are set,
* we add the device.
*/
if (!(dev->flags & VIRTIO_DEV_RUNNING)) {
if (((int)dev->virtqueue[VIRTIO_TXQ]->backend != VIRTIO_DEV_STOPPED) &&
((int)dev->virtqueue[VIRTIO_RXQ]->backend != VIRTIO_DEV_STOPPED))
@ -952,8 +1035,8 @@ set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
}
/*
* Function pointers are set for the device operations to allow CUSE to call functions
* when an IOCTL, device_add or device_release is received.
* Function pointers are set for the device operations to allow CUSE to call
* functions when an IOCTL, device_add or device_release is received.
*/
static const struct vhost_net_device_ops vhost_device_ops = {
.new_device = new_device,
@ -991,11 +1074,13 @@ int rte_vhost_enable_guest_notification(struct virtio_net *dev,
uint16_t queue_id, int enable)
{
if (enable) {
RTE_LOG(ERR, VHOST_CONFIG, "guest notification isn't supported.\n");
RTE_LOG(ERR, VHOST_CONFIG,
"guest notification isn't supported.\n");
return -1;
}
dev->virtqueue[queue_id]->used->flags = enable ? 0 : VRING_USED_F_NO_NOTIFY;
dev->virtqueue[queue_id]->used->flags =
enable ? 0 : VRING_USED_F_NO_NOTIFY;
return 0;
}