[mirror_ubuntu-bionic-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 <linux/kmemleak.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 zero = 0;
static int one = 1;
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 */

#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_key_slow_inc(&rps_needed);
				static_key_slow_inc(&rfs_needed);
			}
			if (orig_sock_table) {
				static_key_slow_dec(&rps_needed);
				static_key_slow_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) {
		*(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");
	}
	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		= &one,
		.extra2		= &one,
# else
		.extra1		= &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		= &zero,
		.extra2		= &two,
	},
	{
		.procname	= "bpf_jit_kallsyms",
		.data		= &bpf_jit_kallsyms,
		.maxlen		= sizeof(int),
		.mode		= 0600,
		.proc_handler	= proc_dointvec_minmax_bpf_restricted,
		.extra1		= &zero,
		.extra2		= &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		= &long_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		= &zero,
		.extra2		= &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		= &zero,
	},
	{
		.procname	= "busy_read",
		.data		= &sysctl_net_busy_read,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &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		= &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		= &zero,
	},
	{ }
};

static struct ctl_table netns_core_table[] = {
	{
		.procname	= "somaxconn",
		.data		= &init_net.core.sysctl_somaxconn,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.extra1		= &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>
7426a564	18	#include <linux/kmemleak.h>
3fff4c42	19
63d819ca	20	#include <net/ip.h>
20380731	21	#include <net/sock.h>
c5c177b4	22	#include <net/net_ratelimit.h>
076bb0c8	23	#include <net/busy_poll.h>
6da7c8fc	24	#include <net/pkt_sched.h>
1da177e4	25
5f671d6b	26	static int zero = 0;
cdda8891	27	static int one = 1;
69e97d45	28	static int two __maybe_unused = 2;
b1cb59cf AK	29	static int min_sndbuf = SOCK_MIN_SNDBUF;
b1cb59cf AK	30	static int min_rcvbuf = SOCK_MIN_RCVBUF;
5f74f82e	31	static int max_skb_frags = MAX_SKB_FRAGS;
ce873e76 DB	32	static long long_one __maybe_unused = 1;
ce873e76 DB	33	static long long_max __maybe_unused = LONG_MAX;
cdda8891	34
ba7a46f1 JP	35	static int net_msg_warn; /* Unused, but still a sysctl */
ba7a46f1 JP	36
fec5e652	37	#ifdef CONFIG_RPS
fe2c6338	38	static int rps_sock_flow_sysctl(struct ctl_table *table, int write,
fec5e652 TH	39	void __user buffer, size_t lenp, loff_t *ppos)
	40	{
	41	unsigned int orig_size, size;
	42	int ret, i;
fe2c6338	43	struct ctl_table tmp = {
fec5e652 TH	44	.data = &size,
	45	.maxlen = sizeof(size),
	46	.mode = table->mode
	47	};
	48	struct rps_sock_flow_table orig_sock_table, sock_table;
	49	static DEFINE_MUTEX(sock_flow_mutex);
	50
	51	mutex_lock(&sock_flow_mutex);
	52
6e3f7faf ED	53	orig_sock_table = rcu_dereference_protected(rps_sock_flow_table,
6e3f7faf ED	54	lockdep_is_held(&sock_flow_mutex));
fec5e652 TH	55	size = orig_size = orig_sock_table ? orig_sock_table->mask + 1 : 0;
	56
	57	ret = proc_dointvec(&tmp, write, buffer, lenp, ppos);
	58
	59	if (write) {
	60	if (size) {
93c1af6c	61	if (size > 1<<29) {
fec5e652 TH	62	/* Enforce limit to prevent overflow */
	63	mutex_unlock(&sock_flow_mutex);
	64	return -EINVAL;
	65	}
	66	size = roundup_pow_of_two(size);
	67	if (size != orig_size) {
	68	sock_table =
	69	vmalloc(RPS_SOCK_FLOW_TABLE_SIZE(size));
	70	if (!sock_table) {
	71	mutex_unlock(&sock_flow_mutex);
	72	return -ENOMEM;
	73	}
567e4b79	74	rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1;
fec5e652 TH	75	sock_table->mask = size - 1;
	76	} else
	77	sock_table = orig_sock_table;
	78
	79	for (i = 0; i < size; i++)
	80	sock_table->ents[i] = RPS_NO_CPU;
	81	} else
	82	sock_table = NULL;
	83
	84	if (sock_table != orig_sock_table) {
	85	rcu_assign_pointer(rps_sock_flow_table, sock_table);
13bfff25	86	if (sock_table) {
c5905afb	87	static_key_slow_inc(&rps_needed);
13bfff25 ED	88	static_key_slow_inc(&rfs_needed);
13bfff25 ED	89	}
adc9300e	90	if (orig_sock_table) {
c5905afb	91	static_key_slow_dec(&rps_needed);
13bfff25	92	static_key_slow_dec(&rfs_needed);
adc9300e ED	93	synchronize_rcu();
	94	vfree(orig_sock_table);
	95	}
fec5e652 TH	96	}
	97	}
	98
	99	mutex_unlock(&sock_flow_mutex);
	100
	101	return ret;
	102	}
	103	#endif /* CONFIG_RPS */
	104
99bbc707 WB	105	#ifdef CONFIG_NET_FLOW_LIMIT
	106	static DEFINE_MUTEX(flow_limit_update_mutex);
	107
fe2c6338	108	static int flow_limit_cpu_sysctl(struct ctl_table *table, int write,
99bbc707 WB	109	void __user buffer, size_t lenp,
	110	loff_t *ppos)
	111	{
	112	struct sd_flow_limit *cur;
	113	struct softnet_data *sd;
	114	cpumask_var_t mask;
	115	int i, len, ret = 0;
	116
	117	if (!alloc_cpumask_var(&mask, GFP_KERNEL))
	118	return -ENOMEM;
	119
	120	if (write) {
	121	ret = cpumask_parse_user(buffer, *lenp, mask);
	122	if (ret)
	123	goto done;
	124
	125	mutex_lock(&flow_limit_update_mutex);
	126	len = sizeof(*cur) + netdev_flow_limit_table_len;
	127	for_each_possible_cpu(i) {
	128	sd = &per_cpu(softnet_data, i);
	129	cur = rcu_dereference_protected(sd->flow_limit,
	130	lockdep_is_held(&flow_limit_update_mutex));
	131	if (cur && !cpumask_test_cpu(i, mask)) {
	132	RCU_INIT_POINTER(sd->flow_limit, NULL);
	133	synchronize_rcu();
	134	kfree(cur);
	135	} else if (!cur && cpumask_test_cpu(i, mask)) {
5b59d467 ED	136	cur = kzalloc_node(len, GFP_KERNEL,
5b59d467 ED	137	cpu_to_node(i));
99bbc707 WB	138	if (!cur) {
	139	/* not unwinding previous changes */
	140	ret = -ENOMEM;
	141	goto write_unlock;
	142	}
	143	cur->num_buckets = netdev_flow_limit_table_len;
	144	rcu_assign_pointer(sd->flow_limit, cur);
	145	}
	146	}
	147	write_unlock:
	148	mutex_unlock(&flow_limit_update_mutex);
	149	} else {
5f121b9a WB	150	char kbuf[128];
5f121b9a WB	151
99bbc707 WB	152	if (ppos \|\| !lenp) {
	153	*lenp = 0;
	154	goto done;
	155	}
	156
	157	cpumask_clear(mask);
	158	rcu_read_lock();
	159	for_each_possible_cpu(i) {
	160	sd = &per_cpu(softnet_data, i);
	161	if (rcu_dereference(sd->flow_limit))
	162	cpumask_set_cpu(i, mask);
	163	}
	164	rcu_read_unlock();
	165
5f121b9a	166	len = min(sizeof(kbuf) - 1, *lenp);
f0906827	167	len = scnprintf(kbuf, len, "%*pb", cpumask_pr_args(mask));
5f121b9a WB	168	if (!len) {
	169	*lenp = 0;
	170	goto done;
	171	}
	172	if (len < *lenp)
	173	kbuf[len++] = '\n';
	174	if (copy_to_user(buffer, kbuf, len)) {
	175	ret = -EFAULT;
	176	goto done;
	177	}
	178	*lenp = len;
	179	*ppos += len;
99bbc707 WB	180	}
	181
	182	done:
	183	free_cpumask_var(mask);
	184	return ret;
	185	}
	186
fe2c6338	187	static int flow_limit_table_len_sysctl(struct ctl_table *table, int write,
99bbc707 WB	188	void __user buffer, size_t lenp,
	189	loff_t *ppos)
	190	{
	191	unsigned int old, *ptr;
	192	int ret;
	193
	194	mutex_lock(&flow_limit_update_mutex);
	195
	196	ptr = table->data;
	197	old = *ptr;
	198	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	199	if (!ret && write && !is_power_of_2(*ptr)) {
	200	*ptr = old;
	201	ret = -EINVAL;
	202	}
	203
	204	mutex_unlock(&flow_limit_update_mutex);
	205	return ret;
	206	}
	207	#endif /* CONFIG_NET_FLOW_LIMIT */
	208
6da7c8fc	209	#ifdef CONFIG_NET_SCHED
	210	static int set_default_qdisc(struct ctl_table *table, int write,
	211	void __user buffer, size_t lenp, loff_t *ppos)
	212	{
	213	char id[IFNAMSIZ];
	214	struct ctl_table tbl = {
	215	.data = id,
	216	.maxlen = IFNAMSIZ,
	217	};
	218	int ret;
	219
	220	qdisc_get_default(id, IFNAMSIZ);
	221
	222	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
	223	if (write && ret == 0)
	224	ret = qdisc_set_default(id);
	225	return ret;
	226	}
	227	#endif
	228
3d48b53f MT	229	static int proc_do_dev_weight(struct ctl_table *table, int write,
	230	void __user buffer, size_t lenp, loff_t *ppos)
	231	{
	232	int ret;
	233
	234	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	235	if (ret != 0)
	236	return ret;
	237
	238	dev_rx_weight = weight_p * dev_weight_rx_bias;
	239	dev_tx_weight = weight_p * dev_weight_tx_bias;
	240
	241	return ret;
	242	}
	243
960fb622 ED	244	static int proc_do_rss_key(struct ctl_table *table, int write,
	245	void __user buffer, size_t lenp, loff_t *ppos)
	246	{
	247	struct ctl_table fake_table;
	248	char buf[NETDEV_RSS_KEY_LEN * 3];
	249
	250	snprintf(buf, sizeof(buf), "%*phC", NETDEV_RSS_KEY_LEN, netdev_rss_key);
	251	fake_table.data = buf;
	252	fake_table.maxlen = sizeof(buf);
	253	return proc_dostring(&fake_table, write, buffer, lenp, ppos);
	254	}
	255
c53c46f8 DB	256	#ifdef CONFIG_BPF_JIT
	257	static int proc_dointvec_minmax_bpf_enable(struct ctl_table *table, int write,
	258	void __user buffer, size_t lenp,
	259	loff_t *ppos)
	260	{
	261	int ret, jit_enable = (int )table->data;
	262	struct ctl_table tmp = *table;
	263
	264	if (write && !capable(CAP_SYS_ADMIN))
	265	return -EPERM;
	266
	267	tmp.data = &jit_enable;
	268	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
	269	if (write && !ret) {
	270	(int )table->data = jit_enable;
	271	if (jit_enable == 2)
	272	pr_warn("bpf_jit_enable = 2 was set! NEVER use this in production, only for JIT debugging!\n");
	273	}
	274	return ret;
	275	}
	276
94b1df32	277	# ifdef CONFIG_HAVE_EBPF_JIT
c53c46f8 DB	278	static int
	279	proc_dointvec_minmax_bpf_restricted(struct ctl_table *table, int write,
	280	void __user buffer, size_t lenp,
	281	loff_t *ppos)
	282	{
	283	if (!capable(CAP_SYS_ADMIN))
	284	return -EPERM;
	285
	286	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
	287	}
94b1df32	288	# endif /* CONFIG_HAVE_EBPF_JIT */
ce873e76 DB	289
	290	static int
	291	proc_dolongvec_minmax_bpf_restricted(struct ctl_table *table, int write,
	292	void __user buffer, size_t lenp,
	293	loff_t *ppos)
	294	{
	295	if (!capable(CAP_SYS_ADMIN))
	296	return -EPERM;
	297
	298	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
	299	}
c53c46f8 DB	300	#endif
c53c46f8 DB	301
33eb9cfc	302	static struct ctl_table net_core_table[] = {
1da177e4 LT	303	#ifdef CONFIG_NET
1da177e4 LT	304	{
1da177e4 LT	305	.procname = "wmem_max",
	306	.data = &sysctl_wmem_max,
	307	.maxlen = sizeof(int),
	308	.mode = 0644,
cdda8891	309	.proc_handler = proc_dointvec_minmax,
b1cb59cf	310	.extra1 = &min_sndbuf,
1da177e4 LT	311	},
1da177e4 LT	312	{
1da177e4 LT	313	.procname = "rmem_max",
	314	.data = &sysctl_rmem_max,
	315	.maxlen = sizeof(int),
	316	.mode = 0644,
cdda8891	317	.proc_handler = proc_dointvec_minmax,
b1cb59cf	318	.extra1 = &min_rcvbuf,
1da177e4 LT	319	},
1da177e4 LT	320	{
1da177e4 LT	321	.procname = "wmem_default",
	322	.data = &sysctl_wmem_default,
	323	.maxlen = sizeof(int),
	324	.mode = 0644,
cdda8891	325	.proc_handler = proc_dointvec_minmax,
b1cb59cf	326	.extra1 = &min_sndbuf,
1da177e4 LT	327	},
1da177e4 LT	328	{
1da177e4 LT	329	.procname = "rmem_default",
	330	.data = &sysctl_rmem_default,
	331	.maxlen = sizeof(int),
	332	.mode = 0644,
cdda8891	333	.proc_handler = proc_dointvec_minmax,
b1cb59cf	334	.extra1 = &min_rcvbuf,
1da177e4 LT	335	},
1da177e4 LT	336	{
1da177e4 LT	337	.procname = "dev_weight",
	338	.data = &weight_p,
	339	.maxlen = sizeof(int),
	340	.mode = 0644,
3d48b53f MT	341	.proc_handler = proc_do_dev_weight,
	342	},
	343	{
	344	.procname = "dev_weight_rx_bias",
	345	.data = &dev_weight_rx_bias,
	346	.maxlen = sizeof(int),
	347	.mode = 0644,
	348	.proc_handler = proc_do_dev_weight,
	349	},
	350	{
	351	.procname = "dev_weight_tx_bias",
	352	.data = &dev_weight_tx_bias,
	353	.maxlen = sizeof(int),
	354	.mode = 0644,
	355	.proc_handler = proc_do_dev_weight,
1da177e4 LT	356	},
1da177e4 LT	357	{
1da177e4 LT	358	.procname = "netdev_max_backlog",
	359	.data = &netdev_max_backlog,
	360	.maxlen = sizeof(int),
	361	.mode = 0644,
6d9f239a	362	.proc_handler = proc_dointvec
1da177e4	363	},
960fb622 ED	364	{
	365	.procname = "netdev_rss_key",
	366	.data = &netdev_rss_key,
	367	.maxlen = sizeof(int),
	368	.mode = 0444,
	369	.proc_handler = proc_do_rss_key,
	370	},
0a14842f ED	371	#ifdef CONFIG_BPF_JIT
	372	{
	373	.procname = "bpf_jit_enable",
	374	.data = &bpf_jit_enable,
	375	.maxlen = sizeof(int),
	376	.mode = 0644,
c53c46f8	377	.proc_handler = proc_dointvec_minmax_bpf_enable,
69e97d45	378	# ifdef CONFIG_BPF_JIT_ALWAYS_ON
290af866 AS	379	.extra1 = &one,
290af866 AS	380	.extra2 = &one,
69e97d45 DB	381	# else
	382	.extra1 = &zero,
	383	.extra2 = &two,
	384	# endif
0a14842f	385	},
4f3446bb DB	386	# ifdef CONFIG_HAVE_EBPF_JIT
	387	{
	388	.procname = "bpf_jit_harden",
	389	.data = &bpf_jit_harden,
	390	.maxlen = sizeof(int),
	391	.mode = 0600,
c53c46f8	392	.proc_handler = proc_dointvec_minmax_bpf_restricted,
69e97d45 DB	393	.extra1 = &zero,
69e97d45 DB	394	.extra2 = &two,
4f3446bb	395	},
74451e66 DB	396	{
	397	.procname = "bpf_jit_kallsyms",
	398	.data = &bpf_jit_kallsyms,
	399	.maxlen = sizeof(int),
	400	.mode = 0600,
c53c46f8	401	.proc_handler = proc_dointvec_minmax_bpf_restricted,
69e97d45 DB	402	.extra1 = &zero,
69e97d45 DB	403	.extra2 = &one,
74451e66	404	},
4f3446bb	405	# endif
409a37eb DB	406	{
	407	.procname = "bpf_jit_limit",
	408	.data = &bpf_jit_limit,
ce873e76	409	.maxlen = sizeof(long),
409a37eb	410	.mode = 0600,
ce873e76 DB	411	.proc_handler = proc_dolongvec_minmax_bpf_restricted,
	412	.extra1 = &long_one,
	413	.extra2 = &long_max,
409a37eb	414	},
0a14842f	415	#endif
3b098e2d ED	416	{
	417	.procname = "netdev_tstamp_prequeue",
	418	.data = &netdev_tstamp_prequeue,
	419	.maxlen = sizeof(int),
	420	.mode = 0644,
	421	.proc_handler = proc_dointvec
	422	},
1da177e4	423	{
1da177e4	424	.procname = "message_cost",
717115e1	425	.data = &net_ratelimit_state.interval,
1da177e4 LT	426	.maxlen = sizeof(int),
1da177e4 LT	427	.mode = 0644,
6d9f239a	428	.proc_handler = proc_dointvec_jiffies,
1da177e4 LT	429	},
1da177e4 LT	430	{
1da177e4	431	.procname = "message_burst",
717115e1	432	.data = &net_ratelimit_state.burst,
1da177e4 LT	433	.maxlen = sizeof(int),
1da177e4 LT	434	.mode = 0644,
6d9f239a	435	.proc_handler = proc_dointvec,
1da177e4 LT	436	},
1da177e4 LT	437	{
1da177e4 LT	438	.procname = "optmem_max",
	439	.data = &sysctl_optmem_max,
	440	.maxlen = sizeof(int),
	441	.mode = 0644,
6d9f239a	442	.proc_handler = proc_dointvec
1da177e4	443	},
b245be1f WB	444	{
	445	.procname = "tstamp_allow_data",
	446	.data = &sysctl_tstamp_allow_data,
	447	.maxlen = sizeof(int),
	448	.mode = 0644,
	449	.proc_handler = proc_dointvec_minmax,
	450	.extra1 = &zero,
	451	.extra2 = &one
	452	},
fec5e652 TH	453	#ifdef CONFIG_RPS
	454	{
	455	.procname = "rps_sock_flow_entries",
	456	.maxlen = sizeof(int),
	457	.mode = 0644,
	458	.proc_handler = rps_sock_flow_sysctl
	459	},
	460	#endif
99bbc707 WB	461	#ifdef CONFIG_NET_FLOW_LIMIT
	462	{
	463	.procname = "flow_limit_cpu_bitmap",
	464	.mode = 0644,
	465	.proc_handler = flow_limit_cpu_sysctl
	466	},
	467	{
	468	.procname = "flow_limit_table_len",
	469	.data = &netdev_flow_limit_table_len,
	470	.maxlen = sizeof(int),
	471	.mode = 0644,
	472	.proc_handler = flow_limit_table_len_sysctl
	473	},
	474	#endif /* CONFIG_NET_FLOW_LIMIT */
e0d1095a	475	#ifdef CONFIG_NET_RX_BUSY_POLL
06021292	476	{
64b0dc51 ET	477	.procname = "busy_poll",
64b0dc51 ET	478	.data = &sysctl_net_busy_poll,
eb6db622	479	.maxlen = sizeof(unsigned int),
06021292	480	.mode = 0644,
95f25521 AD	481	.proc_handler = proc_dointvec_minmax,
95f25521 AD	482	.extra1 = &zero,
06021292	483	},
2d48d67f	484	{
64b0dc51 ET	485	.procname = "busy_read",
64b0dc51 ET	486	.data = &sysctl_net_busy_read,
2d48d67f ET	487	.maxlen = sizeof(unsigned int),
2d48d67f ET	488	.mode = 0644,
95f25521 AD	489	.proc_handler = proc_dointvec_minmax,
95f25521 AD	490	.extra1 = &zero,
2d48d67f	491	},
6da7c8fc	492	#endif
	493	#ifdef CONFIG_NET_SCHED
	494	{
	495	.procname = "default_qdisc",
	496	.mode = 0644,
	497	.maxlen = IFNAMSIZ,
	498	.proc_handler = set_default_qdisc
	499	},
06021292	500	#endif
1da177e4	501	#endif /* CONFIG_NET */
51b0bded	502	{
51b0bded SH	503	.procname = "netdev_budget",
	504	.data = &netdev_budget,
	505	.maxlen = sizeof(int),
	506	.mode = 0644,
6d9f239a	507	.proc_handler = proc_dointvec
51b0bded	508	},
a2a316fd	509	{
a2a316fd SH	510	.procname = "warnings",
	511	.data = &net_msg_warn,
	512	.maxlen = sizeof(int),
	513	.mode = 0644,
6d9f239a	514	.proc_handler = proc_dointvec
a2a316fd	515	},
5f74f82e HWR	516	{
	517	.procname = "max_skb_frags",
	518	.data = &sysctl_max_skb_frags,
	519	.maxlen = sizeof(int),
	520	.mode = 0644,
	521	.proc_handler = proc_dointvec_minmax,
	522	.extra1 = &one,
	523	.extra2 = &max_skb_frags,
	524	},
7acf8a1e MW	525	{
	526	.procname = "netdev_budget_usecs",
	527	.data = &netdev_budget_usecs,
	528	.maxlen = sizeof(unsigned int),
	529	.mode = 0644,
	530	.proc_handler = proc_dointvec_minmax,
	531	.extra1 = &zero,
	532	},
f8572d8f	533	{ }
1da177e4	534	};
33eb9cfc	535
d5a4502e PE	536	static struct ctl_table netns_core_table[] = {
d5a4502e PE	537	{
d5a4502e PE	538	.procname = "somaxconn",
	539	.data = &init_net.core.sysctl_somaxconn,
	540	.maxlen = sizeof(int),
	541	.mode = 0644,
5f671d6b	542	.extra1 = &zero,
5f671d6b	543	.proc_handler = proc_dointvec_minmax
d5a4502e	544	},
f8572d8f	545	{ }
d5a4502e PE	546	};
d5a4502e PE	547
024626e3	548	static __net_init int sysctl_core_net_init(struct net *net)
33eb9cfc	549	{
d5a4502e	550	struct ctl_table *tbl;
024626e3	551
d5a4502e	552	tbl = netns_core_table;
09ad9bc7	553	if (!net_eq(net, &init_net)) {
d5a4502e	554	tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL);
024626e3 PE	555	if (tbl == NULL)
	556	goto err_dup;
	557
d5a4502e	558	tbl[0].data = &net->core.sysctl_somaxconn;
464dc801 EB	559
	560	/* Don't export any sysctls to unprivileged users */
	561	if (net->user_ns != &init_user_ns) {
	562	tbl[0].procname = NULL;
	563	}
024626e3 PE	564	}
024626e3 PE	565
ec8f23ce	566	net->core.sysctl_hdr = register_net_sysctl(net, "net/core", tbl);
8efa6e93	567	if (net->core.sysctl_hdr == NULL)
024626e3	568	goto err_reg;
33eb9cfc	569
024626e3 PE	570	return 0;
	571
	572	err_reg:
d5a4502e	573	if (tbl != netns_core_table)
024626e3 PE	574	kfree(tbl);
	575	err_dup:
	576	return -ENOMEM;
	577	}
	578
	579	static __net_exit void sysctl_core_net_exit(struct net *net)
	580	{
	581	struct ctl_table *tbl;
	582
8efa6e93 PE	583	tbl = net->core.sysctl_hdr->ctl_table_arg;
8efa6e93 PE	584	unregister_net_sysctl_table(net->core.sysctl_hdr);
d5a4502e	585	BUG_ON(tbl == netns_core_table);
024626e3 PE	586	kfree(tbl);
	587	}
	588
	589	static __net_initdata struct pernet_operations sysctl_core_ops = {
	590	.init = sysctl_core_net_init,
	591	.exit = sysctl_core_net_exit,
	592	};
	593
	594	static __init int sysctl_core_init(void)
	595	{
43444757	596	register_net_sysctl(&init_net, "net/core", net_core_table);
024626e3	597	return register_pernet_subsys(&sysctl_core_ops);
33eb9cfc PE	598	}
33eb9cfc PE	599
b27aeadb	600	fs_initcall(sysctl_core_init);