Overview
How net-based IOCTL calls are made – prerequisite for other wiki pages.
UAPI links:
Kernel links:
-
- socket.c (
struct file_operations socket_file_ops) -
Sample userspace call: ethtool
Git repo is here.
send_ioctl()
From ethtool.c, a simple wrapper around the ioctl() call:
#ifndef TEST_ETHTOOL
int send_ioctl(struct cmd_context *ctx, void *cmd)
{
ctx->ifr.ifr_data = cmd; // command-specific structure
return ioctl(ctx->fd, SIOCETHTOOL, &ctx->ifr);
}
#endif
QUESTION: What is returned in ifreq->cmd?
struct cmd_context
From internal.h, struct cmd_context showing the embedded struct ifreq:
/* Context for sub-commands */
struct cmd_context {
const char *devname; /* net device name */
int fd; /* socket suitable for ethtool ioctl */
struct ifreq ifr; /* ifreq suitable for ethtool ioctl */
int argc; /* number of arguments to the sub-command */
char **argp; /* arguments to the sub-command */
};
struct ifreq
From include/uapi/linux/if.h, the unsigned long arg passed with ioctl():
#if __UAPI_DEF_IF_IFREQ
struct ifreq {
#define IFHWADDRLEN 6
union
{
char ifrn_name[IFNAMSIZ]; /* if name, e.g. "en0" */
} ifr_ifrn;
union {
struct sockaddr ifru_addr;
struct sockaddr ifru_dstaddr;
struct sockaddr ifru_broadaddr;
struct sockaddr ifru_netmask;
struct sockaddr ifru_hwaddr;
short ifru_flags;
int ifru_ivalue;
int ifru_mtu;
struct ifmap ifru_map;
char ifru_slave[IFNAMSIZ]; /* Just fits the size */
char ifru_newname[IFNAMSIZ];
void __user * ifru_data;
struct if_settings ifru_settings;
} ifr_ifru;
};
#endif /* __UAPI_DEF_IF_IFREQ */
#define ifr_name ifr_ifrn.ifrn_name /* interface name */
#define ifr_hwaddr ifr_ifru.ifru_hwaddr /* MAC address */
#define ifr_addr ifr_ifru.ifru_addr /* address */
#define ifr_dstaddr ifr_ifru.ifru_dstaddr /* other end of p-p lnk */
#define ifr_broadaddr ifr_ifru.ifru_broadaddr /* broadcast address */
#define ifr_netmask ifr_ifru.ifru_netmask /* interface net mask */
#define ifr_flags ifr_ifru.ifru_flags /* flags */
#define ifr_metric ifr_ifru.ifru_ivalue /* metric */
#define ifr_mtu ifr_ifru.ifru_mtu /* mtu */
#define ifr_map ifr_ifru.ifru_map /* device map */
#define ifr_slave ifr_ifru.ifru_slave /* slave device */
#define ifr_data ifr_ifru.ifru_data /* for use by interface */
#define ifr_ifindex ifr_ifru.ifru_ivalue /* interface index */
#define ifr_bandwidth ifr_ifru.ifru_ivalue /* link bandwidth */
#define ifr_qlen ifr_ifru.ifru_ivalue /* Queue length */
#define ifr_newname ifr_ifru.ifru_newname /* New name */
#define ifr_settings ifr_ifru.ifru_settings /* Device/proto settings*/
Kernel workflow
socket_file_ops
From net/socket.c:
static const struct file_operations socket_file_ops = {
.owner = THIS_MODULE,
.llseek = no_llseek,
.read_iter = sock_read_iter,
.write_iter = sock_write_iter,
.poll = sock_poll,
.unlocked_ioctl = sock_ioctl, <-- this one
#ifdef CONFIG_COMPAT
.compat_ioctl = compat_sock_ioctl,
#endif
.mmap = sock_mmap,
.release = sock_close,
.fasync = sock_fasync,
.sendpage = sock_sendpage,
.splice_write = generic_splice_sendpage,
.splice_read = sock_splice_read,
};
sock_ioctl()
Also from net/socket.c – anything related to SIOCDEVPRIVATE is seriously deprecated (see here):
static long sock_ioctl(struct file *file, unsigned cmd, unsigned long arg)
{
struct socket *sock;
struct sock *sk;
void __user *argp = (void __user *)arg;
int pid, err;
struct net *net;
sock = file->private_data;
sk = sock->sk;
net = sock_net(sk);
if (unlikely(cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))) {
struct ifreq ifr;
bool need_copyout;
if (copy_from_user(&ifr, argp, sizeof(struct ifreq)))
return -EFAULT;
err = dev_ioctl(net, cmd, &ifr, &need_copyout);
if (!err && need_copyout)
if (copy_to_user(argp, &ifr, sizeof(struct ifreq)))
return -EFAULT;
} else
This reflects wireless extensions, which should not be relevant:
#ifdef CONFIG_WEXT_CORE
if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
err = wext_handle_ioctl(net, cmd, argp);
} else
#endif
Finally, this is where we should end up:
switch (cmd) {
case FIOSETOWN:
case SIOCSPGRP:
err = -EFAULT;
if (get_user(pid, (int __user *)argp))
break;
err = f_setown(sock->file, pid, 1);
break;
case FIOGETOWN:
case SIOCGPGRP:
err = put_user(f_getown(sock->file),
(int __user *)argp);
break;
case SIOCGIFBR:
case SIOCSIFBR:
case SIOCBRADDBR:
case SIOCBRDELBR:
err = -ENOPKG;
if (!br_ioctl_hook)
request_module("bridge");
mutex_lock(&br_ioctl_mutex);
if (br_ioctl_hook)
err = br_ioctl_hook(net, cmd, argp);
mutex_unlock(&br_ioctl_mutex);
break;
case SIOCGIFVLAN:
case SIOCSIFVLAN:
err = -ENOPKG;
if (!vlan_ioctl_hook)
request_module("8021q");
mutex_lock(&vlan_ioctl_mutex);
if (vlan_ioctl_hook)
err = vlan_ioctl_hook(net, argp);
mutex_unlock(&vlan_ioctl_mutex);
break;
case SIOCADDDLCI:
case SIOCDELDLCI:
err = -ENOPKG;
if (!dlci_ioctl_hook)
request_module("dlci");
mutex_lock(&dlci_ioctl_mutex);
if (dlci_ioctl_hook)
err = dlci_ioctl_hook(cmd, argp);
mutex_unlock(&dlci_ioctl_mutex);
break;
case SIOCGSKNS:
err = -EPERM;
if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
break;
err = open_related_ns(&net->ns, get_net_ns);
break;
default:
err = sock_do_ioctl(net, sock, cmd, arg); <-- there
break;
}
return err;
}
sock_do_ioctl()
Still in net/socket.c:
static long sock_do_ioctl(struct net *net, struct socket *sock,
unsigned int cmd, unsigned long arg)
{
int err;
void __user *argp = (void __user *)arg;
err = sock->ops->ioctl(sock, cmd, arg);
/*
* If this ioctl is unknown try to hand it down
* to the NIC driver.
*/
if (err != -ENOIOCTLCMD)
return err;
if (cmd == SIOCGIFCONF) {
struct ifconf ifc;
if (copy_from_user(&ifc, argp, sizeof(struct ifconf)))
return -EFAULT;
rtnl_lock();
err = dev_ifconf(net, &ifc, sizeof(struct ifreq));
rtnl_unlock();
if (!err && copy_to_user(argp, &ifc, sizeof(struct ifconf)))
err = -EFAULT;
} else {
struct ifreq ifr;
bool need_copyout;
if (copy_from_user(&ifr, argp, sizeof(struct ifreq)))
return -EFAULT;
err = dev_ioctl(net, cmd, &ifr, &need_copyout); <-- here???
if (!err && need_copyout)
if (copy_to_user(argp, &ifr, sizeof(struct ifreq)))
return -EFAULT;
}
return err;
}