[mirror_ubuntu-focal-kernel.git] / net / core / sysctl_net_core.c

// SPDX-License-Identifier: GPL-2.0
/* -*- linux-c -*-
 * sysctl_net_core.c: sysctl interface to net core subsystem.
 *
 * Begun April 1, 1996, Mike Shaver.
 * Added /proc/sys/net/core directory entry (empty =) ). [MS]
 */

#include <linux/mm.h>
#include <linux/sysctl.h>
#include <linux/module.h>
#include <linux/socket.h>
#include <linux/netdevice.h>
#include <linux/ratelimit.h>
#include <linux/vmalloc.h>
#include <linux/init.h>
#include <linux/slab.h>

#include <net/ip.h>
#include <net/sock.h>
#include <net/net_ratelimit.h>
#include <net/busy_poll.h>
#include <net/pkt_sched.h>

static int two __maybe_unused = 2;
static int min_sndbuf = SOCK_MIN_SNDBUF;
static int min_rcvbuf = SOCK_MIN_RCVBUF;
static int max_skb_frags = MAX_SKB_FRAGS;
static long long_one __maybe_unused = 1;
static long long_max __maybe_unused = LONG_MAX;

static int net_msg_warn;	/* Unused, but still a sysctl */

int sysctl_fb_tunnels_only_for_init_net __read_mostly = 0;
EXPORT_SYMBOL(sysctl_fb_tunnels_only_for_init_net);

/* 0 - Keep current behavior:
 *     IPv4: inherit all current settings from init_net
 *     IPv6: reset all settings to default
 * 1 - Both inherit all current settings from init_net
 * 2 - Both reset all settings to default
 */
int sysctl_devconf_inherit_init_net __read_mostly;
EXPORT_SYMBOL(sysctl_devconf_inherit_init_net);

#ifdef CONFIG_RPS
static int rps_sock_flow_sysctl(struct ctl_table *table, int write,
				void __user *buffer, size_t *lenp, loff_t *ppos)
{
	unsigned int orig_size, size;
	int ret, i;
	struct ctl_table tmp = {
		.data = &size,
		.maxlen = sizeof(size),
		.mode = table->mode
	};
	struct rps_sock_flow_table *orig_sock_table, *sock_table;
	static DEFINE_MUTEX(sock_flow_mutex);

	mutex_lock(&sock_flow_mutex);

	orig_sock_table = rcu_dereference_protected(rps_sock_flow_table,
					lockdep_is_held(&sock_flow_mutex));
	size = orig_size = orig_sock_table ? orig_sock_table->mask + 1 : 0;

	ret = proc_dointvec(&tmp, write, buffer, lenp, ppos);

	if (write) {
		if (size) {
			if (size > 1<<29) {
				/* Enforce limit to prevent overflow */
				mutex_unlock(&sock_flow_mutex);
				return -EINVAL;
			}
			size = roundup_pow_of_two(size);
			if (size != orig_size) {
				sock_table =
				    vmalloc(RPS_SOCK_FLOW_TABLE_SIZE(size));
				if (!sock_table) {
					mutex_unlock(&sock_flow_mutex);
					return -ENOMEM;
				}
				rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1;
				sock_table->mask = size - 1;
			} else
				sock_table = orig_sock_table;

			for (i = 0; i < size; i++)
				sock_table->ents[i] = RPS_NO_CPU;
		} else
			sock_table = NULL;

		if (sock_table != orig_sock_table) {
			rcu_assign_pointer(rps_sock_flow_table, sock_table);
			if (sock_table) {
				static_branch_inc(&rps_needed);
				static_branch_inc(&rfs_needed);
			}
			if (orig_sock_table) {
				static_branch_dec(&rps_needed);
				static_branch_dec(&rfs_needed);
				synchronize_rcu();
				vfree(orig_sock_table);
			}
		}
	}

	mutex_unlock(&sock_flow_mutex);

	return ret;
}
#endif /* CONFIG_RPS */

#ifdef CONFIG_NET_FLOW_LIMIT
static DEFINE_MUTEX(flow_limit_update_mutex);

static int flow_limit_cpu_sysctl(struct ctl_table *table, int write,
				 void __user *buffer, size_t *lenp,
				 loff_t *ppos)
{
	struct sd_flow_limit *cur;
	struct softnet_data *sd;
	cpumask_var_t mask;
	int i, len, ret = 0;

	if (!alloc_cpumask_var(&mask, GFP_KERNEL))
		return -ENOMEM;

	if (write) {
		ret = cpumask_parse_user(buffer, *lenp, mask);
		if (ret)
			goto done;

		mutex_lock(&flow_limit_update_mutex);
		len = sizeof(*cur) + netdev_flow_limit_table_len;
		for_each_possible_cpu(i) {
			sd = &per_cpu(softnet_data, i);
			cur = rcu_dereference_protected(sd->flow_limit,
				     lockdep_is_held(&flow_limit_update_mutex));
			if (cur && !cpumask_test_cpu(i, mask)) {
				RCU_INIT_POINTER(sd->flow_limit, NULL);
				synchronize_rcu();
				kfree(cur);
			} else if (!cur && cpumask_test_cpu(i, mask)) {
				cur = kzalloc_node(len, GFP_KERNEL,
						   cpu_to_node(i));
				if (!cur) {
					/* not unwinding previous changes */
					ret = -ENOMEM;
					goto write_unlock;
				}
				cur->num_buckets = netdev_flow_limit_table_len;
				rcu_assign_pointer(sd->flow_limit, cur);
			}
		}
write_unlock:
		mutex_unlock(&flow_limit_update_mutex);
	} else {
		char kbuf[128];

		if (*ppos || !*lenp) {
			*lenp = 0;
			goto done;
		}

		cpumask_clear(mask);
		rcu_read_lock();
		for_each_possible_cpu(i) {
			sd = &per_cpu(softnet_data, i);
			if (rcu_dereference(sd->flow_limit))
				cpumask_set_cpu(i, mask);
		}
		rcu_read_unlock();

		len = min(sizeof(kbuf) - 1, *lenp);
		len = scnprintf(kbuf, len, "%*pb", cpumask_pr_args(mask));
		if (!len) {
			*lenp = 0;
			goto done;
		}
		if (len < *lenp)
			kbuf[len++] = '\n';
		if (copy_to_user(buffer, kbuf, len)) {
			ret = -EFAULT;
			goto done;
		}
		*lenp = len;
		*ppos += len;
	}

done:
	free_cpumask_var(mask);
	return ret;
}

static int flow_limit_table_len_sysctl(struct ctl_table *table, int write,
				       void __user *buffer, size_t *lenp,
				       loff_t *ppos)
{
	unsigned int old, *ptr;
	int ret;

	mutex_lock(&flow_limit_update_mutex);

	ptr = table->data;
	old = *ptr;
	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	if (!ret && write && !is_power_of_2(*ptr)) {
		*ptr = old;
		ret = -EINVAL;
	}

	mutex_unlock(&flow_limit_update_mutex);
	return ret;
}
#endif /* CONFIG_NET_FLOW_LIMIT */

#ifdef CONFIG_NET_SCHED
static int set_default_qdisc(struct ctl_table *table, int write,
			     void __user *buffer, size_t *lenp, loff_t *ppos)
{
	char id[IFNAMSIZ];
	struct ctl_table tbl = {
		.data = id,
		.maxlen = IFNAMSIZ,
	};
	int ret;

	qdisc_get_default(id, IFNAMSIZ);

	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
	if (write && ret == 0)
		ret = qdisc_set_default(id);
	return ret;
}
#endif

static int proc_do_dev_weight(struct ctl_table *table, int write,
			   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int ret;

	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	if (ret != 0)
		return ret;

	dev_rx_weight = weight_p * dev_weight_rx_bias;
	dev_tx_weight = weight_p * dev_weight_tx_bias;

	return ret;
}

static int proc_do_rss_key(struct ctl_table *table, int write,
			   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	struct ctl_table fake_table;
	char buf[NETDEV_RSS_KEY_LEN * 3];

	snprintf(buf, sizeof(buf), "%*phC", NETDEV_RSS_KEY_LEN, netdev_rss_key);
	fake_table.data = buf;
	fake_table.maxlen = sizeof(buf);
	return proc_dostring(&fake_table, write, buffer, lenp, ppos);
}

#ifdef CONFIG_BPF_JIT
static int proc_dointvec_minmax_bpf_enable(struct ctl_table *table, int write,
					   void __user *buffer, size_t *lenp,
					   loff_t *ppos)
{
	int ret, jit_enable = *(int *)table->data;
	struct ctl_table tmp = *table;

	if (write && !capable(CAP_SYS_ADMIN))
		return -EPERM;

	tmp.data = &jit_enable;
	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
	if (write && !ret) {
		if (jit_enable < 2 ||
		    (jit_enable == 2 && bpf_dump_raw_ok(current_cred()))) {
			*(int *)table->data = jit_enable;
			if (jit_enable == 2)
				pr_warn("bpf_jit_enable = 2 was set! NEVER use this in production, only for JIT debugging!\n");
		} else {
			ret = -EPERM;
		}
	}
	return ret;
}

# ifdef CONFIG_HAVE_EBPF_JIT
static int
proc_dointvec_minmax_bpf_restricted(struct ctl_table *table, int write,
				    void __user *buffer, size_t *lenp,
				    loff_t *ppos)
{
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
}
# endif /* CONFIG_HAVE_EBPF_JIT */

static int
proc_dolongvec_minmax_bpf_restricted(struct ctl_table *table, int write,
				     void __user *buffer, size_t *lenp,
				     loff_t *ppos)
{
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
}
#endif

static struct ctl_table net_core_table[] = {
#ifdef CONFIG_NET
	{
		.procname	= "wmem_max",
		.data		= &sysctl_wmem_max,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &min_sndbuf,
	},
	{
		.procname	= "rmem_max",
		.data		= &sysctl_rmem_max,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &min_rcvbuf,
	},
	{
		.procname	= "wmem_default",
		.data		= &sysctl_wmem_default,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &min_sndbuf,
	},
	{
		.procname	= "rmem_default",
		.data		= &sysctl_rmem_default,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &min_rcvbuf,
	},
	{
		.procname	= "dev_weight",
		.data		= &weight_p,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_do_dev_weight,
	},
	{
		.procname	= "dev_weight_rx_bias",
		.data		= &dev_weight_rx_bias,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_do_dev_weight,
	},
	{
		.procname	= "dev_weight_tx_bias",
		.data		= &dev_weight_tx_bias,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_do_dev_weight,
	},
	{
		.procname	= "netdev_max_backlog",
		.data		= &netdev_max_backlog,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "netdev_rss_key",
		.data		= &netdev_rss_key,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= proc_do_rss_key,
	},
#ifdef CONFIG_BPF_JIT
	{
		.procname	= "bpf_jit_enable",
		.data		= &bpf_jit_enable,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax_bpf_enable,
# ifdef CONFIG_BPF_JIT_ALWAYS_ON
		.extra1		= SYSCTL_ONE,
		.extra2		= SYSCTL_ONE,
# else
		.extra1		= SYSCTL_ZERO,
		.extra2		= &two,
# endif
	},
# ifdef CONFIG_HAVE_EBPF_JIT
	{
		.procname	= "bpf_jit_harden",
		.data		= &bpf_jit_harden,
		.maxlen		= sizeof(int),
		.mode		= 0600,
		.proc_handler	= proc_dointvec_minmax_bpf_restricted,
		.extra1		= SYSCTL_ZERO,
		.extra2		= &two,
	},
	{
		.procname	= "bpf_jit_kallsyms",
		.data		= &bpf_jit_kallsyms,
		.maxlen		= sizeof(int),
		.mode		= 0600,
		.proc_handler	= proc_dointvec_minmax_bpf_restricted,
		.extra1		= SYSCTL_ZERO,
		.extra2		= SYSCTL_ONE,
	},
# endif
	{
		.procname	= "bpf_jit_limit",
		.data		= &bpf_jit_limit,
		.maxlen		= sizeof(long),
		.mode		= 0600,
		.proc_handler	= proc_dolongvec_minmax_bpf_restricted,
		.extra1		= &long_one,
		.extra2		= &bpf_jit_limit_max,
	},
#endif
	{
		.procname	= "netdev_tstamp_prequeue",
		.data		= &netdev_tstamp_prequeue,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "message_cost",
		.data		= &net_ratelimit_state.interval,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
	{
		.procname	= "message_burst",
		.data		= &net_ratelimit_state.burst,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "optmem_max",
		.data		= &sysctl_optmem_max,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "tstamp_allow_data",
		.data		= &sysctl_tstamp_allow_data,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
		.extra2		= SYSCTL_ONE
	},
#ifdef CONFIG_RPS
	{
		.procname	= "rps_sock_flow_entries",
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= rps_sock_flow_sysctl
	},
#endif
#ifdef CONFIG_NET_FLOW_LIMIT
	{
		.procname	= "flow_limit_cpu_bitmap",
		.mode		= 0644,
		.proc_handler	= flow_limit_cpu_sysctl
	},
	{
		.procname	= "flow_limit_table_len",
		.data		= &netdev_flow_limit_table_len,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= flow_limit_table_len_sysctl
	},
#endif /* CONFIG_NET_FLOW_LIMIT */
#ifdef CONFIG_NET_RX_BUSY_POLL
	{
		.procname	= "busy_poll",
		.data		= &sysctl_net_busy_poll,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
	},
	{
		.procname	= "busy_read",
		.data		= &sysctl_net_busy_read,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
	},
#endif
#ifdef CONFIG_NET_SCHED
	{
		.procname	= "default_qdisc",
		.mode		= 0644,
		.maxlen		= IFNAMSIZ,
		.proc_handler	= set_default_qdisc
	},
#endif
#endif /* CONFIG_NET */
	{
		.procname	= "netdev_budget",
		.data		= &netdev_budget,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "warnings",
		.data		= &net_msg_warn,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "max_skb_frags",
		.data		= &sysctl_max_skb_frags,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ONE,
		.extra2		= &max_skb_frags,
	},
	{
		.procname	= "netdev_budget_usecs",
		.data		= &netdev_budget_usecs,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
	},
	{
		.procname	= "fb_tunnels_only_for_init_net",
		.data		= &sysctl_fb_tunnels_only_for_init_net,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
		.extra2		= SYSCTL_ONE,
	},
	{
		.procname	= "devconf_inherit_init_net",
		.data		= &sysctl_devconf_inherit_init_net,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ZERO,
		.extra2		= &two,
	},
	{
		.procname	= "high_order_alloc_disable",
		.data		= &net_high_order_alloc_disable_key.key,
		.maxlen         = sizeof(net_high_order_alloc_disable_key),
		.mode		= 0644,
		.proc_handler	= proc_do_static_key,
	},
	{
		.procname	= "gro_normal_batch",
		.data		= &gro_normal_batch,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= SYSCTL_ONE,
	},
	{ }
};

static struct ctl_table netns_core_table[] = {
	{
		.procname	= "somaxconn",
		.data		= &init_net.core.sysctl_somaxconn,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.extra1		= SYSCTL_ZERO,
		.proc_handler	= proc_dointvec_minmax
	},
	{ }
};

static __net_init int sysctl_core_net_init(struct net *net)
{
	struct ctl_table *tbl;

	tbl = netns_core_table;
	if (!net_eq(net, &init_net)) {
		tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL);
		if (tbl == NULL)
			goto err_dup;

		tbl[0].data = &net->core.sysctl_somaxconn;

		/* Don't export any sysctls to unprivileged users */
		if (net->user_ns != &init_user_ns) {
			tbl[0].procname = NULL;
		}
	}

	net->core.sysctl_hdr = register_net_sysctl(net, "net/core", tbl);
	if (net->core.sysctl_hdr == NULL)
		goto err_reg;

	return 0;

err_reg:
	if (tbl != netns_core_table)
		kfree(tbl);
err_dup:
	return -ENOMEM;
}

static __net_exit void sysctl_core_net_exit(struct net *net)
{
	struct ctl_table *tbl;

	tbl = net->core.sysctl_hdr->ctl_table_arg;
	unregister_net_sysctl_table(net->core.sysctl_hdr);
	BUG_ON(tbl == netns_core_table);
	kfree(tbl);
}

static __net_initdata struct pernet_operations sysctl_core_ops = {
	.init = sysctl_core_net_init,
	.exit = sysctl_core_net_exit,
};

static __init int sysctl_core_init(void)
{
	register_net_sysctl(&init_net, "net/core", net_core_table);
	return register_pernet_subsys(&sysctl_core_ops);
}

fs_initcall(sysctl_core_init);
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/* -- linux-c --
	3	* sysctl_net_core.c: sysctl interface to net core subsystem.
	4	*
	5	* Begun April 1, 1996, Mike Shaver.
	6	* Added /proc/sys/net/core directory entry (empty =) ). [MS]
	7	*/
	8
	9	#include <linux/mm.h>
	10	#include <linux/sysctl.h>
1da177e4	11	#include <linux/module.h>
20380731	12	#include <linux/socket.h>
a37ae408	13	#include <linux/netdevice.h>
3fff4c42	14	#include <linux/ratelimit.h>
fec5e652	15	#include <linux/vmalloc.h>
33eb9cfc	16	#include <linux/init.h>
5a0e3ad6	17	#include <linux/slab.h>
3fff4c42	18
63d819ca	19	#include <net/ip.h>
20380731	20	#include <net/sock.h>
c5c177b4	21	#include <net/net_ratelimit.h>
076bb0c8	22	#include <net/busy_poll.h>
6da7c8fc	23	#include <net/pkt_sched.h>
1da177e4	24
fa9dd599	25	static int two __maybe_unused = 2;
b1cb59cf AK	26	static int min_sndbuf = SOCK_MIN_SNDBUF;
b1cb59cf AK	27	static int min_rcvbuf = SOCK_MIN_RCVBUF;
5f74f82e	28	static int max_skb_frags = MAX_SKB_FRAGS;
fdadd049 DB	29	static long long_one __maybe_unused = 1;
fdadd049 DB	30	static long long_max __maybe_unused = LONG_MAX;
cdda8891	31
ba7a46f1 JP	32	static int net_msg_warn; /* Unused, but still a sysctl */
ba7a46f1 JP	33
79134e6c ED	34	int sysctl_fb_tunnels_only_for_init_net __read_mostly = 0;
	35	EXPORT_SYMBOL(sysctl_fb_tunnels_only_for_init_net);
	36
856c395c CW	37	/* 0 - Keep current behavior:
	38	* IPv4: inherit all current settings from init_net
	39	* IPv6: reset all settings to default
	40	* 1 - Both inherit all current settings from init_net
	41	* 2 - Both reset all settings to default
	42	*/
	43	int sysctl_devconf_inherit_init_net __read_mostly;
	44	EXPORT_SYMBOL(sysctl_devconf_inherit_init_net);
	45
fec5e652	46	#ifdef CONFIG_RPS
fe2c6338	47	static int rps_sock_flow_sysctl(struct ctl_table *table, int write,
fec5e652 TH	48	void __user buffer, size_t lenp, loff_t *ppos)
	49	{
	50	unsigned int orig_size, size;
	51	int ret, i;
fe2c6338	52	struct ctl_table tmp = {
fec5e652 TH	53	.data = &size,
	54	.maxlen = sizeof(size),
	55	.mode = table->mode
	56	};
	57	struct rps_sock_flow_table orig_sock_table, sock_table;
	58	static DEFINE_MUTEX(sock_flow_mutex);
	59
	60	mutex_lock(&sock_flow_mutex);
	61
6e3f7faf ED	62	orig_sock_table = rcu_dereference_protected(rps_sock_flow_table,
6e3f7faf ED	63	lockdep_is_held(&sock_flow_mutex));
fec5e652 TH	64	size = orig_size = orig_sock_table ? orig_sock_table->mask + 1 : 0;
	65
	66	ret = proc_dointvec(&tmp, write, buffer, lenp, ppos);
	67
	68	if (write) {
	69	if (size) {
93c1af6c	70	if (size > 1<<29) {
fec5e652 TH	71	/* Enforce limit to prevent overflow */
	72	mutex_unlock(&sock_flow_mutex);
	73	return -EINVAL;
	74	}
	75	size = roundup_pow_of_two(size);
	76	if (size != orig_size) {
	77	sock_table =
	78	vmalloc(RPS_SOCK_FLOW_TABLE_SIZE(size));
	79	if (!sock_table) {
	80	mutex_unlock(&sock_flow_mutex);
	81	return -ENOMEM;
	82	}
567e4b79	83	rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1;
fec5e652 TH	84	sock_table->mask = size - 1;
	85	} else
	86	sock_table = orig_sock_table;
	87
	88	for (i = 0; i < size; i++)
	89	sock_table->ents[i] = RPS_NO_CPU;
	90	} else
	91	sock_table = NULL;
	92
	93	if (sock_table != orig_sock_table) {
	94	rcu_assign_pointer(rps_sock_flow_table, sock_table);
13bfff25	95	if (sock_table) {
dc05360f ED	96	static_branch_inc(&rps_needed);
dc05360f ED	97	static_branch_inc(&rfs_needed);
13bfff25	98	}
adc9300e	99	if (orig_sock_table) {
dc05360f ED	100	static_branch_dec(&rps_needed);
dc05360f ED	101	static_branch_dec(&rfs_needed);
adc9300e ED	102	synchronize_rcu();
	103	vfree(orig_sock_table);
	104	}
fec5e652 TH	105	}
	106	}
	107
	108	mutex_unlock(&sock_flow_mutex);
	109
	110	return ret;
	111	}
	112	#endif /* CONFIG_RPS */
	113
99bbc707 WB	114	#ifdef CONFIG_NET_FLOW_LIMIT
	115	static DEFINE_MUTEX(flow_limit_update_mutex);
	116
fe2c6338	117	static int flow_limit_cpu_sysctl(struct ctl_table *table, int write,
99bbc707 WB	118	void __user buffer, size_t lenp,
	119	loff_t *ppos)
	120	{
	121	struct sd_flow_limit *cur;
	122	struct softnet_data *sd;
	123	cpumask_var_t mask;
	124	int i, len, ret = 0;
	125
	126	if (!alloc_cpumask_var(&mask, GFP_KERNEL))
	127	return -ENOMEM;
	128
	129	if (write) {
	130	ret = cpumask_parse_user(buffer, *lenp, mask);
	131	if (ret)
	132	goto done;
	133
	134	mutex_lock(&flow_limit_update_mutex);
	135	len = sizeof(*cur) + netdev_flow_limit_table_len;
	136	for_each_possible_cpu(i) {
	137	sd = &per_cpu(softnet_data, i);
	138	cur = rcu_dereference_protected(sd->flow_limit,
	139	lockdep_is_held(&flow_limit_update_mutex));
	140	if (cur && !cpumask_test_cpu(i, mask)) {
	141	RCU_INIT_POINTER(sd->flow_limit, NULL);
	142	synchronize_rcu();
	143	kfree(cur);
	144	} else if (!cur && cpumask_test_cpu(i, mask)) {
5b59d467 ED	145	cur = kzalloc_node(len, GFP_KERNEL,
5b59d467 ED	146	cpu_to_node(i));
99bbc707 WB	147	if (!cur) {
	148	/* not unwinding previous changes */
	149	ret = -ENOMEM;
	150	goto write_unlock;
	151	}
	152	cur->num_buckets = netdev_flow_limit_table_len;
	153	rcu_assign_pointer(sd->flow_limit, cur);
	154	}
	155	}
	156	write_unlock:
	157	mutex_unlock(&flow_limit_update_mutex);
	158	} else {
5f121b9a WB	159	char kbuf[128];
5f121b9a WB	160
99bbc707 WB	161	if (ppos \|\| !lenp) {
	162	*lenp = 0;
	163	goto done;
	164	}
	165
	166	cpumask_clear(mask);
	167	rcu_read_lock();
	168	for_each_possible_cpu(i) {
	169	sd = &per_cpu(softnet_data, i);
	170	if (rcu_dereference(sd->flow_limit))
	171	cpumask_set_cpu(i, mask);
	172	}
	173	rcu_read_unlock();
	174
5f121b9a	175	len = min(sizeof(kbuf) - 1, *lenp);
f0906827	176	len = scnprintf(kbuf, len, "%*pb", cpumask_pr_args(mask));
5f121b9a WB	177	if (!len) {
	178	*lenp = 0;
	179	goto done;
	180	}
	181	if (len < *lenp)
	182	kbuf[len++] = '\n';
	183	if (copy_to_user(buffer, kbuf, len)) {
	184	ret = -EFAULT;
	185	goto done;
	186	}
	187	*lenp = len;
	188	*ppos += len;
99bbc707 WB	189	}
	190
	191	done:
	192	free_cpumask_var(mask);
	193	return ret;
	194	}
	195
fe2c6338	196	static int flow_limit_table_len_sysctl(struct ctl_table *table, int write,
99bbc707 WB	197	void __user buffer, size_t lenp,
	198	loff_t *ppos)
	199	{
	200	unsigned int old, *ptr;
	201	int ret;
	202
	203	mutex_lock(&flow_limit_update_mutex);
	204
	205	ptr = table->data;
	206	old = *ptr;
	207	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	208	if (!ret && write && !is_power_of_2(*ptr)) {
	209	*ptr = old;
	210	ret = -EINVAL;
	211	}
	212
	213	mutex_unlock(&flow_limit_update_mutex);
	214	return ret;
	215	}
	216	#endif /* CONFIG_NET_FLOW_LIMIT */
	217
6da7c8fc	218	#ifdef CONFIG_NET_SCHED
	219	static int set_default_qdisc(struct ctl_table *table, int write,
	220	void __user buffer, size_t lenp, loff_t *ppos)
	221	{
	222	char id[IFNAMSIZ];
	223	struct ctl_table tbl = {
	224	.data = id,
	225	.maxlen = IFNAMSIZ,
	226	};
	227	int ret;
	228
	229	qdisc_get_default(id, IFNAMSIZ);
	230
	231	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
	232	if (write && ret == 0)
	233	ret = qdisc_set_default(id);
	234	return ret;
	235	}
	236	#endif
	237
3d48b53f MT	238	static int proc_do_dev_weight(struct ctl_table *table, int write,
	239	void __user buffer, size_t lenp, loff_t *ppos)
	240	{
	241	int ret;
	242
	243	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	244	if (ret != 0)
	245	return ret;
	246
	247	dev_rx_weight = weight_p * dev_weight_rx_bias;
	248	dev_tx_weight = weight_p * dev_weight_tx_bias;
	249
	250	return ret;
	251	}
	252
960fb622 ED	253	static int proc_do_rss_key(struct ctl_table *table, int write,
	254	void __user buffer, size_t lenp, loff_t *ppos)
	255	{
	256	struct ctl_table fake_table;
	257	char buf[NETDEV_RSS_KEY_LEN * 3];
	258
	259	snprintf(buf, sizeof(buf), "%*phC", NETDEV_RSS_KEY_LEN, netdev_rss_key);
	260	fake_table.data = buf;
	261	fake_table.maxlen = sizeof(buf);
	262	return proc_dostring(&fake_table, write, buffer, lenp, ppos);
	263	}
	264
2e4a3098 DB	265	#ifdef CONFIG_BPF_JIT
	266	static int proc_dointvec_minmax_bpf_enable(struct ctl_table *table, int write,
	267	void __user buffer, size_t lenp,
	268	loff_t *ppos)
	269	{
	270	int ret, jit_enable = (int )table->data;
	271	struct ctl_table tmp = *table;
	272
	273	if (write && !capable(CAP_SYS_ADMIN))
	274	return -EPERM;
	275
	276	tmp.data = &jit_enable;
	277	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
	278	if (write && !ret) {
	279	if (jit_enable < 2 \|\|
393c79c6	280	(jit_enable == 2 && bpf_dump_raw_ok(current_cred()))) {
2e4a3098 DB	281	(int )table->data = jit_enable;
	282	if (jit_enable == 2)
	283	pr_warn("bpf_jit_enable = 2 was set! NEVER use this in production, only for JIT debugging!\n");
	284	} else {
	285	ret = -EPERM;
	286	}
	287	}
	288	return ret;
	289	}
	290
48c35560	291	# ifdef CONFIG_HAVE_EBPF_JIT
2e4a3098 DB	292	static int
	293	proc_dointvec_minmax_bpf_restricted(struct ctl_table *table, int write,
	294	void __user buffer, size_t lenp,
	295	loff_t *ppos)
	296	{
	297	if (!capable(CAP_SYS_ADMIN))
	298	return -EPERM;
	299
	300	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
	301	}
48c35560	302	# endif /* CONFIG_HAVE_EBPF_JIT */
fdadd049 DB	303
	304	static int
	305	proc_dolongvec_minmax_bpf_restricted(struct ctl_table *table, int write,
	306	void __user buffer, size_t lenp,
	307	loff_t *ppos)
	308	{
	309	if (!capable(CAP_SYS_ADMIN))
	310	return -EPERM;
	311
	312	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
	313	}
2e4a3098 DB	314	#endif
2e4a3098 DB	315
33eb9cfc	316	static struct ctl_table net_core_table[] = {
1da177e4 LT	317	#ifdef CONFIG_NET
1da177e4 LT	318	{
1da177e4 LT	319	.procname = "wmem_max",
	320	.data = &sysctl_wmem_max,
	321	.maxlen = sizeof(int),
	322	.mode = 0644,
cdda8891	323	.proc_handler = proc_dointvec_minmax,
b1cb59cf	324	.extra1 = &min_sndbuf,
1da177e4 LT	325	},
1da177e4 LT	326	{
1da177e4 LT	327	.procname = "rmem_max",
	328	.data = &sysctl_rmem_max,
	329	.maxlen = sizeof(int),
	330	.mode = 0644,
cdda8891	331	.proc_handler = proc_dointvec_minmax,
b1cb59cf	332	.extra1 = &min_rcvbuf,
1da177e4 LT	333	},
1da177e4 LT	334	{
1da177e4 LT	335	.procname = "wmem_default",
	336	.data = &sysctl_wmem_default,
	337	.maxlen = sizeof(int),
	338	.mode = 0644,
cdda8891	339	.proc_handler = proc_dointvec_minmax,
b1cb59cf	340	.extra1 = &min_sndbuf,
1da177e4 LT	341	},
1da177e4 LT	342	{
1da177e4 LT	343	.procname = "rmem_default",
	344	.data = &sysctl_rmem_default,
	345	.maxlen = sizeof(int),
	346	.mode = 0644,
cdda8891	347	.proc_handler = proc_dointvec_minmax,
b1cb59cf	348	.extra1 = &min_rcvbuf,
1da177e4 LT	349	},
1da177e4 LT	350	{
1da177e4 LT	351	.procname = "dev_weight",
	352	.data = &weight_p,
	353	.maxlen = sizeof(int),
	354	.mode = 0644,
3d48b53f MT	355	.proc_handler = proc_do_dev_weight,
	356	},
	357	{
	358	.procname = "dev_weight_rx_bias",
	359	.data = &dev_weight_rx_bias,
	360	.maxlen = sizeof(int),
	361	.mode = 0644,
	362	.proc_handler = proc_do_dev_weight,
	363	},
	364	{
	365	.procname = "dev_weight_tx_bias",
	366	.data = &dev_weight_tx_bias,
	367	.maxlen = sizeof(int),
	368	.mode = 0644,
	369	.proc_handler = proc_do_dev_weight,
1da177e4 LT	370	},
1da177e4 LT	371	{
1da177e4 LT	372	.procname = "netdev_max_backlog",
	373	.data = &netdev_max_backlog,
	374	.maxlen = sizeof(int),
	375	.mode = 0644,
6d9f239a	376	.proc_handler = proc_dointvec
1da177e4	377	},
960fb622 ED	378	{
	379	.procname = "netdev_rss_key",
	380	.data = &netdev_rss_key,
	381	.maxlen = sizeof(int),
	382	.mode = 0444,
	383	.proc_handler = proc_do_rss_key,
	384	},
0a14842f ED	385	#ifdef CONFIG_BPF_JIT
	386	{
	387	.procname = "bpf_jit_enable",
	388	.data = &bpf_jit_enable,
	389	.maxlen = sizeof(int),
	390	.mode = 0644,
2e4a3098	391	.proc_handler = proc_dointvec_minmax_bpf_enable,
fa9dd599	392	# ifdef CONFIG_BPF_JIT_ALWAYS_ON
eec4844f MC	393	.extra1 = SYSCTL_ONE,
eec4844f MC	394	.extra2 = SYSCTL_ONE,
fa9dd599	395	# else
eec4844f	396	.extra1 = SYSCTL_ZERO,
fa9dd599 DB	397	.extra2 = &two,
fa9dd599 DB	398	# endif
0a14842f	399	},
4f3446bb DB	400	# ifdef CONFIG_HAVE_EBPF_JIT
	401	{
	402	.procname = "bpf_jit_harden",
	403	.data = &bpf_jit_harden,
	404	.maxlen = sizeof(int),
	405	.mode = 0600,
2e4a3098	406	.proc_handler = proc_dointvec_minmax_bpf_restricted,
eec4844f	407	.extra1 = SYSCTL_ZERO,
fa9dd599	408	.extra2 = &two,
4f3446bb	409	},
74451e66 DB	410	{
	411	.procname = "bpf_jit_kallsyms",
	412	.data = &bpf_jit_kallsyms,
	413	.maxlen = sizeof(int),
	414	.mode = 0600,
2e4a3098	415	.proc_handler = proc_dointvec_minmax_bpf_restricted,
eec4844f MC	416	.extra1 = SYSCTL_ZERO,
eec4844f MC	417	.extra2 = SYSCTL_ONE,
74451e66	418	},
4f3446bb	419	# endif
ede95a63 DB	420	{
	421	.procname = "bpf_jit_limit",
	422	.data = &bpf_jit_limit,
fdadd049	423	.maxlen = sizeof(long),
ede95a63	424	.mode = 0600,
fdadd049 DB	425	.proc_handler = proc_dolongvec_minmax_bpf_restricted,
fdadd049 DB	426	.extra1 = &long_one,
22bca93f	427	.extra2 = &bpf_jit_limit_max,
ede95a63	428	},
0a14842f	429	#endif
3b098e2d ED	430	{
	431	.procname = "netdev_tstamp_prequeue",
	432	.data = &netdev_tstamp_prequeue,
	433	.maxlen = sizeof(int),
	434	.mode = 0644,
	435	.proc_handler = proc_dointvec
	436	},
1da177e4	437	{
1da177e4	438	.procname = "message_cost",
717115e1	439	.data = &net_ratelimit_state.interval,
1da177e4 LT	440	.maxlen = sizeof(int),
1da177e4 LT	441	.mode = 0644,
6d9f239a	442	.proc_handler = proc_dointvec_jiffies,
1da177e4 LT	443	},
1da177e4 LT	444	{
1da177e4	445	.procname = "message_burst",
717115e1	446	.data = &net_ratelimit_state.burst,
1da177e4 LT	447	.maxlen = sizeof(int),
1da177e4 LT	448	.mode = 0644,
6d9f239a	449	.proc_handler = proc_dointvec,
1da177e4 LT	450	},
1da177e4 LT	451	{
1da177e4 LT	452	.procname = "optmem_max",
	453	.data = &sysctl_optmem_max,
	454	.maxlen = sizeof(int),
	455	.mode = 0644,
6d9f239a	456	.proc_handler = proc_dointvec
1da177e4	457	},
b245be1f WB	458	{
	459	.procname = "tstamp_allow_data",
	460	.data = &sysctl_tstamp_allow_data,
	461	.maxlen = sizeof(int),
	462	.mode = 0644,
	463	.proc_handler = proc_dointvec_minmax,
eec4844f MC	464	.extra1 = SYSCTL_ZERO,
eec4844f MC	465	.extra2 = SYSCTL_ONE
b245be1f	466	},
fec5e652 TH	467	#ifdef CONFIG_RPS
	468	{
	469	.procname = "rps_sock_flow_entries",
	470	.maxlen = sizeof(int),
	471	.mode = 0644,
	472	.proc_handler = rps_sock_flow_sysctl
	473	},
	474	#endif
99bbc707 WB	475	#ifdef CONFIG_NET_FLOW_LIMIT
	476	{
	477	.procname = "flow_limit_cpu_bitmap",
	478	.mode = 0644,
	479	.proc_handler = flow_limit_cpu_sysctl
	480	},
	481	{
	482	.procname = "flow_limit_table_len",
	483	.data = &netdev_flow_limit_table_len,
	484	.maxlen = sizeof(int),
	485	.mode = 0644,
	486	.proc_handler = flow_limit_table_len_sysctl
	487	},
	488	#endif /* CONFIG_NET_FLOW_LIMIT */
e0d1095a	489	#ifdef CONFIG_NET_RX_BUSY_POLL
06021292	490	{
64b0dc51 ET	491	.procname = "busy_poll",
64b0dc51 ET	492	.data = &sysctl_net_busy_poll,
eb6db622	493	.maxlen = sizeof(unsigned int),
06021292	494	.mode = 0644,
95f25521	495	.proc_handler = proc_dointvec_minmax,
eec4844f	496	.extra1 = SYSCTL_ZERO,
06021292	497	},
2d48d67f	498	{
64b0dc51 ET	499	.procname = "busy_read",
64b0dc51 ET	500	.data = &sysctl_net_busy_read,
2d48d67f ET	501	.maxlen = sizeof(unsigned int),
2d48d67f ET	502	.mode = 0644,
95f25521	503	.proc_handler = proc_dointvec_minmax,
eec4844f	504	.extra1 = SYSCTL_ZERO,
2d48d67f	505	},
6da7c8fc	506	#endif
	507	#ifdef CONFIG_NET_SCHED
	508	{
	509	.procname = "default_qdisc",
	510	.mode = 0644,
	511	.maxlen = IFNAMSIZ,
	512	.proc_handler = set_default_qdisc
	513	},
06021292	514	#endif
1da177e4	515	#endif /* CONFIG_NET */
51b0bded	516	{
51b0bded SH	517	.procname = "netdev_budget",
	518	.data = &netdev_budget,
	519	.maxlen = sizeof(int),
	520	.mode = 0644,
6d9f239a	521	.proc_handler = proc_dointvec
51b0bded	522	},
a2a316fd	523	{
a2a316fd SH	524	.procname = "warnings",
	525	.data = &net_msg_warn,
	526	.maxlen = sizeof(int),
	527	.mode = 0644,
6d9f239a	528	.proc_handler = proc_dointvec
a2a316fd	529	},
5f74f82e HWR	530	{
	531	.procname = "max_skb_frags",
	532	.data = &sysctl_max_skb_frags,
	533	.maxlen = sizeof(int),
	534	.mode = 0644,
	535	.proc_handler = proc_dointvec_minmax,
eec4844f	536	.extra1 = SYSCTL_ONE,
5f74f82e HWR	537	.extra2 = &max_skb_frags,
5f74f82e HWR	538	},
7acf8a1e MW	539	{
	540	.procname = "netdev_budget_usecs",
	541	.data = &netdev_budget_usecs,
	542	.maxlen = sizeof(unsigned int),
	543	.mode = 0644,
	544	.proc_handler = proc_dointvec_minmax,
eec4844f	545	.extra1 = SYSCTL_ZERO,
7acf8a1e	546	},
79134e6c ED	547	{
	548	.procname = "fb_tunnels_only_for_init_net",
	549	.data = &sysctl_fb_tunnels_only_for_init_net,
	550	.maxlen = sizeof(int),
	551	.mode = 0644,
	552	.proc_handler = proc_dointvec_minmax,
eec4844f MC	553	.extra1 = SYSCTL_ZERO,
eec4844f MC	554	.extra2 = SYSCTL_ONE,
79134e6c	555	},
856c395c CW	556	{
	557	.procname = "devconf_inherit_init_net",
	558	.data = &sysctl_devconf_inherit_init_net,
	559	.maxlen = sizeof(int),
	560	.mode = 0644,
	561	.proc_handler = proc_dointvec_minmax,
eec4844f	562	.extra1 = SYSCTL_ZERO,
856c395c CW	563	.extra2 = &two,
856c395c CW	564	},
ce27ec60 ED	565	{
	566	.procname = "high_order_alloc_disable",
	567	.data = &net_high_order_alloc_disable_key.key,
	568	.maxlen = sizeof(net_high_order_alloc_disable_key),
	569	.mode = 0644,
	570	.proc_handler = proc_do_static_key,
	571	},
323ebb61 EC	572	{
	573	.procname = "gro_normal_batch",
	574	.data = &gro_normal_batch,
	575	.maxlen = sizeof(unsigned int),
	576	.mode = 0644,
	577	.proc_handler = proc_dointvec_minmax,
	578	.extra1 = SYSCTL_ONE,
	579	},
f8572d8f	580	{ }
1da177e4	581	};
33eb9cfc	582
d5a4502e PE	583	static struct ctl_table netns_core_table[] = {
d5a4502e PE	584	{
d5a4502e PE	585	.procname = "somaxconn",
	586	.data = &init_net.core.sysctl_somaxconn,
	587	.maxlen = sizeof(int),
	588	.mode = 0644,
eec4844f	589	.extra1 = SYSCTL_ZERO,
5f671d6b	590	.proc_handler = proc_dointvec_minmax
d5a4502e	591	},
f8572d8f	592	{ }
d5a4502e PE	593	};
d5a4502e PE	594
024626e3	595	static __net_init int sysctl_core_net_init(struct net *net)
33eb9cfc	596	{
d5a4502e	597	struct ctl_table *tbl;
024626e3	598
d5a4502e	599	tbl = netns_core_table;
09ad9bc7	600	if (!net_eq(net, &init_net)) {
d5a4502e	601	tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL);
024626e3 PE	602	if (tbl == NULL)
	603	goto err_dup;
	604
d5a4502e	605	tbl[0].data = &net->core.sysctl_somaxconn;
464dc801 EB	606
	607	/* Don't export any sysctls to unprivileged users */
	608	if (net->user_ns != &init_user_ns) {
	609	tbl[0].procname = NULL;
	610	}
024626e3 PE	611	}
024626e3 PE	612
ec8f23ce	613	net->core.sysctl_hdr = register_net_sysctl(net, "net/core", tbl);
8efa6e93	614	if (net->core.sysctl_hdr == NULL)
024626e3	615	goto err_reg;
33eb9cfc	616
024626e3 PE	617	return 0;
	618
	619	err_reg:
d5a4502e	620	if (tbl != netns_core_table)
024626e3 PE	621	kfree(tbl);
	622	err_dup:
	623	return -ENOMEM;
	624	}
	625
	626	static __net_exit void sysctl_core_net_exit(struct net *net)
	627	{
	628	struct ctl_table *tbl;
	629
8efa6e93 PE	630	tbl = net->core.sysctl_hdr->ctl_table_arg;
8efa6e93 PE	631	unregister_net_sysctl_table(net->core.sysctl_hdr);
d5a4502e	632	BUG_ON(tbl == netns_core_table);
024626e3 PE	633	kfree(tbl);
	634	}
	635
	636	static __net_initdata struct pernet_operations sysctl_core_ops = {
	637	.init = sysctl_core_net_init,
	638	.exit = sysctl_core_net_exit,
	639	};
	640
	641	static __init int sysctl_core_init(void)
	642	{
43444757	643	register_net_sysctl(&init_net, "net/core", net_core_table);
024626e3	644	return register_pernet_subsys(&sysctl_core_ops);
33eb9cfc PE	645	}
33eb9cfc PE	646
b27aeadb	647	fs_initcall(sysctl_core_init);