[linux-block.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 = 2;
static int three = 3;
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
 * 3 - Both inherit all settings from current netns
 */
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 *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 *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(buffer, 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';
		memcpy(buffer, kbuf, len);
		*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 *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 *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 *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 *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 *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 *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 *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[] = {
	{
		.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		= &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		= 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
	{
		.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		= &two,
	},
	{
		.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		= &three,
	},
	{
		.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 int __init fb_tunnels_only_for_init_net_sysctl_setup(char *str)
{
	/* fallback tunnels for initns only */
	if (!strncmp(str, "initns", 6))
		sysctl_fb_tunnels_only_for_init_net = 1;
	/* no fallback tunnels anywhere */
	else if (!strncmp(str, "none", 4))
		sysctl_fb_tunnels_only_for_init_net = 2;

	return 1;
}
__setup("fb_tunnels=", fb_tunnels_only_for_init_net_sysctl_setup);

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
316cdaa1	25	static int two = 2;
9efd6a3c	26	static int three = 3;
b1cb59cf AK	27	static int min_sndbuf = SOCK_MIN_SNDBUF;
b1cb59cf AK	28	static int min_rcvbuf = SOCK_MIN_RCVBUF;
5f74f82e	29	static int max_skb_frags = MAX_SKB_FRAGS;
fdadd049 DB	30	static long long_one __maybe_unused = 1;
fdadd049 DB	31	static long long_max __maybe_unused = LONG_MAX;
cdda8891	32
ba7a46f1 JP	33	static int net_msg_warn; /* Unused, but still a sysctl */
ba7a46f1 JP	34
79134e6c ED	35	int sysctl_fb_tunnels_only_for_init_net __read_mostly = 0;
	36	EXPORT_SYMBOL(sysctl_fb_tunnels_only_for_init_net);
	37
856c395c CW	38	/* 0 - Keep current behavior:
	39	* IPv4: inherit all current settings from init_net
	40	* IPv6: reset all settings to default
	41	* 1 - Both inherit all current settings from init_net
	42	* 2 - Both reset all settings to default
9efd6a3c	43	* 3 - Both inherit all settings from current netns
856c395c CW	44	*/
	45	int sysctl_devconf_inherit_init_net __read_mostly;
	46	EXPORT_SYMBOL(sysctl_devconf_inherit_init_net);
	47
fec5e652	48	#ifdef CONFIG_RPS
fe2c6338	49	static int rps_sock_flow_sysctl(struct ctl_table *table, int write,
32927393	50	void buffer, size_t lenp, loff_t *ppos)
fec5e652 TH	51	{
	52	unsigned int orig_size, size;
	53	int ret, i;
fe2c6338	54	struct ctl_table tmp = {
fec5e652 TH	55	.data = &size,
	56	.maxlen = sizeof(size),
	57	.mode = table->mode
	58	};
	59	struct rps_sock_flow_table orig_sock_table, sock_table;
	60	static DEFINE_MUTEX(sock_flow_mutex);
	61
	62	mutex_lock(&sock_flow_mutex);
	63
6e3f7faf ED	64	orig_sock_table = rcu_dereference_protected(rps_sock_flow_table,
6e3f7faf ED	65	lockdep_is_held(&sock_flow_mutex));
fec5e652 TH	66	size = orig_size = orig_sock_table ? orig_sock_table->mask + 1 : 0;
	67
	68	ret = proc_dointvec(&tmp, write, buffer, lenp, ppos);
	69
	70	if (write) {
	71	if (size) {
93c1af6c	72	if (size > 1<<29) {
fec5e652 TH	73	/* Enforce limit to prevent overflow */
	74	mutex_unlock(&sock_flow_mutex);
	75	return -EINVAL;
	76	}
	77	size = roundup_pow_of_two(size);
	78	if (size != orig_size) {
	79	sock_table =
	80	vmalloc(RPS_SOCK_FLOW_TABLE_SIZE(size));
	81	if (!sock_table) {
	82	mutex_unlock(&sock_flow_mutex);
	83	return -ENOMEM;
	84	}
567e4b79	85	rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1;
fec5e652 TH	86	sock_table->mask = size - 1;
	87	} else
	88	sock_table = orig_sock_table;
	89
	90	for (i = 0; i < size; i++)
	91	sock_table->ents[i] = RPS_NO_CPU;
	92	} else
	93	sock_table = NULL;
	94
	95	if (sock_table != orig_sock_table) {
	96	rcu_assign_pointer(rps_sock_flow_table, sock_table);
13bfff25	97	if (sock_table) {
dc05360f ED	98	static_branch_inc(&rps_needed);
dc05360f ED	99	static_branch_inc(&rfs_needed);
13bfff25	100	}
adc9300e	101	if (orig_sock_table) {
dc05360f ED	102	static_branch_dec(&rps_needed);
dc05360f ED	103	static_branch_dec(&rfs_needed);
adc9300e ED	104	synchronize_rcu();
	105	vfree(orig_sock_table);
	106	}
fec5e652 TH	107	}
	108	}
	109
	110	mutex_unlock(&sock_flow_mutex);
	111
	112	return ret;
	113	}
	114	#endif /* CONFIG_RPS */
	115
99bbc707 WB	116	#ifdef CONFIG_NET_FLOW_LIMIT
	117	static DEFINE_MUTEX(flow_limit_update_mutex);
	118
fe2c6338	119	static int flow_limit_cpu_sysctl(struct ctl_table *table, int write,
32927393	120	void buffer, size_t lenp, loff_t *ppos)
99bbc707 WB	121	{
	122	struct sd_flow_limit *cur;
	123	struct softnet_data *sd;
	124	cpumask_var_t mask;
	125	int i, len, ret = 0;
	126
	127	if (!alloc_cpumask_var(&mask, GFP_KERNEL))
	128	return -ENOMEM;
	129
	130	if (write) {
56965ac7	131	ret = cpumask_parse(buffer, mask);
99bbc707 WB	132	if (ret)
	133	goto done;
	134
	135	mutex_lock(&flow_limit_update_mutex);
	136	len = sizeof(*cur) + netdev_flow_limit_table_len;
	137	for_each_possible_cpu(i) {
	138	sd = &per_cpu(softnet_data, i);
	139	cur = rcu_dereference_protected(sd->flow_limit,
	140	lockdep_is_held(&flow_limit_update_mutex));
	141	if (cur && !cpumask_test_cpu(i, mask)) {
	142	RCU_INIT_POINTER(sd->flow_limit, NULL);
	143	synchronize_rcu();
	144	kfree(cur);
	145	} else if (!cur && cpumask_test_cpu(i, mask)) {
5b59d467 ED	146	cur = kzalloc_node(len, GFP_KERNEL,
5b59d467 ED	147	cpu_to_node(i));
99bbc707 WB	148	if (!cur) {
	149	/* not unwinding previous changes */
	150	ret = -ENOMEM;
	151	goto write_unlock;
	152	}
	153	cur->num_buckets = netdev_flow_limit_table_len;
	154	rcu_assign_pointer(sd->flow_limit, cur);
	155	}
	156	}
	157	write_unlock:
	158	mutex_unlock(&flow_limit_update_mutex);
	159	} else {
5f121b9a WB	160	char kbuf[128];
5f121b9a WB	161
99bbc707 WB	162	if (ppos \|\| !lenp) {
	163	*lenp = 0;
	164	goto done;
	165	}
	166
	167	cpumask_clear(mask);
	168	rcu_read_lock();
	169	for_each_possible_cpu(i) {
	170	sd = &per_cpu(softnet_data, i);
	171	if (rcu_dereference(sd->flow_limit))
	172	cpumask_set_cpu(i, mask);
	173	}
	174	rcu_read_unlock();
	175
5f121b9a	176	len = min(sizeof(kbuf) - 1, *lenp);
f0906827	177	len = scnprintf(kbuf, len, "%*pb", cpumask_pr_args(mask));
5f121b9a WB	178	if (!len) {
	179	*lenp = 0;
	180	goto done;
	181	}
	182	if (len < *lenp)
	183	kbuf[len++] = '\n';
32927393	184	memcpy(buffer, kbuf, len);
5f121b9a WB	185	*lenp = len;
5f121b9a WB	186	*ppos += len;
99bbc707 WB	187	}
	188
	189	done:
	190	free_cpumask_var(mask);
	191	return ret;
	192	}
	193
fe2c6338	194	static int flow_limit_table_len_sysctl(struct ctl_table *table, int write,
32927393	195	void buffer, size_t lenp, loff_t *ppos)
99bbc707 WB	196	{
	197	unsigned int old, *ptr;
	198	int ret;
	199
	200	mutex_lock(&flow_limit_update_mutex);
	201
	202	ptr = table->data;
	203	old = *ptr;
	204	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	205	if (!ret && write && !is_power_of_2(*ptr)) {
	206	*ptr = old;
	207	ret = -EINVAL;
	208	}
	209
	210	mutex_unlock(&flow_limit_update_mutex);
	211	return ret;
	212	}
	213	#endif /* CONFIG_NET_FLOW_LIMIT */
	214
6da7c8fc	215	#ifdef CONFIG_NET_SCHED
6da7c8fc	216	static int set_default_qdisc(struct ctl_table *table, int write,
32927393	217	void buffer, size_t lenp, loff_t *ppos)
6da7c8fc	218	{
	219	char id[IFNAMSIZ];
	220	struct ctl_table tbl = {
	221	.data = id,
	222	.maxlen = IFNAMSIZ,
	223	};
	224	int ret;
	225
	226	qdisc_get_default(id, IFNAMSIZ);
	227
	228	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
	229	if (write && ret == 0)
	230	ret = qdisc_set_default(id);
	231	return ret;
	232	}
	233	#endif
	234
3d48b53f	235	static int proc_do_dev_weight(struct ctl_table *table, int write,
32927393	236	void buffer, size_t lenp, loff_t *ppos)
3d48b53f MT	237	{
	238	int ret;
	239
	240	ret = proc_dointvec(table, write, buffer, lenp, ppos);
	241	if (ret != 0)
	242	return ret;
	243
	244	dev_rx_weight = weight_p * dev_weight_rx_bias;
	245	dev_tx_weight = weight_p * dev_weight_tx_bias;
	246
	247	return ret;
	248	}
	249
960fb622	250	static int proc_do_rss_key(struct ctl_table *table, int write,
32927393	251	void buffer, size_t lenp, loff_t *ppos)
960fb622 ED	252	{
	253	struct ctl_table fake_table;
	254	char buf[NETDEV_RSS_KEY_LEN * 3];
	255
	256	snprintf(buf, sizeof(buf), "%*phC", NETDEV_RSS_KEY_LEN, netdev_rss_key);
	257	fake_table.data = buf;
	258	fake_table.maxlen = sizeof(buf);
	259	return proc_dostring(&fake_table, write, buffer, lenp, ppos);
	260	}
	261
2e4a3098 DB	262	#ifdef CONFIG_BPF_JIT
2e4a3098 DB	263	static int proc_dointvec_minmax_bpf_enable(struct ctl_table *table, int write,
32927393	264	void buffer, size_t lenp,
2e4a3098 DB	265	loff_t *ppos)
	266	{
	267	int ret, jit_enable = (int )table->data;
	268	struct ctl_table tmp = *table;
	269
	270	if (write && !capable(CAP_SYS_ADMIN))
	271	return -EPERM;
	272
	273	tmp.data = &jit_enable;
	274	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
	275	if (write && !ret) {
	276	if (jit_enable < 2 \|\|
63960260	277	(jit_enable == 2 && bpf_dump_raw_ok(current_cred()))) {
2e4a3098 DB	278	(int )table->data = jit_enable;
	279	if (jit_enable == 2)
	280	pr_warn("bpf_jit_enable = 2 was set! NEVER use this in production, only for JIT debugging!\n");
	281	} else {
	282	ret = -EPERM;
	283	}
	284	}
	285	return ret;
	286	}
	287
1148f9ad	288	# ifdef CONFIG_HAVE_EBPF_JIT
2e4a3098 DB	289	static int
2e4a3098 DB	290	proc_dointvec_minmax_bpf_restricted(struct ctl_table *table, int write,
32927393	291	void buffer, size_t lenp, loff_t *ppos)
2e4a3098 DB	292	{
	293	if (!capable(CAP_SYS_ADMIN))
	294	return -EPERM;
	295
	296	return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
	297	}
1148f9ad	298	# endif /* CONFIG_HAVE_EBPF_JIT */
fdadd049 DB	299
	300	static int
	301	proc_dolongvec_minmax_bpf_restricted(struct ctl_table *table, int write,
32927393	302	void buffer, size_t lenp, loff_t *ppos)
fdadd049 DB	303	{
	304	if (!capable(CAP_SYS_ADMIN))
	305	return -EPERM;
	306
	307	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
	308	}
2e4a3098 DB	309	#endif
2e4a3098 DB	310
33eb9cfc	311	static struct ctl_table net_core_table[] = {
1da177e4	312	{
1da177e4 LT	313	.procname = "wmem_max",
	314	.data = &sysctl_wmem_max,
	315	.maxlen = sizeof(int),
	316	.mode = 0644,
cdda8891	317	.proc_handler = proc_dointvec_minmax,
b1cb59cf	318	.extra1 = &min_sndbuf,
1da177e4 LT	319	},
1da177e4 LT	320	{
1da177e4 LT	321	.procname = "rmem_max",
	322	.data = &sysctl_rmem_max,
	323	.maxlen = sizeof(int),
	324	.mode = 0644,
cdda8891	325	.proc_handler = proc_dointvec_minmax,
b1cb59cf	326	.extra1 = &min_rcvbuf,
1da177e4 LT	327	},
1da177e4 LT	328	{
1da177e4 LT	329	.procname = "wmem_default",
	330	.data = &sysctl_wmem_default,
	331	.maxlen = sizeof(int),
	332	.mode = 0644,
cdda8891	333	.proc_handler = proc_dointvec_minmax,
b1cb59cf	334	.extra1 = &min_sndbuf,
1da177e4 LT	335	},
1da177e4 LT	336	{
1da177e4 LT	337	.procname = "rmem_default",
	338	.data = &sysctl_rmem_default,
	339	.maxlen = sizeof(int),
	340	.mode = 0644,
cdda8891	341	.proc_handler = proc_dointvec_minmax,
b1cb59cf	342	.extra1 = &min_rcvbuf,
1da177e4 LT	343	},
1da177e4 LT	344	{
1da177e4 LT	345	.procname = "dev_weight",
	346	.data = &weight_p,
	347	.maxlen = sizeof(int),
	348	.mode = 0644,
3d48b53f MT	349	.proc_handler = proc_do_dev_weight,
	350	},
	351	{
	352	.procname = "dev_weight_rx_bias",
	353	.data = &dev_weight_rx_bias,
	354	.maxlen = sizeof(int),
	355	.mode = 0644,
	356	.proc_handler = proc_do_dev_weight,
	357	},
	358	{
	359	.procname = "dev_weight_tx_bias",
	360	.data = &dev_weight_tx_bias,
	361	.maxlen = sizeof(int),
	362	.mode = 0644,
	363	.proc_handler = proc_do_dev_weight,
1da177e4 LT	364	},
1da177e4 LT	365	{
1da177e4 LT	366	.procname = "netdev_max_backlog",
	367	.data = &netdev_max_backlog,
	368	.maxlen = sizeof(int),
	369	.mode = 0644,
6d9f239a	370	.proc_handler = proc_dointvec
1da177e4	371	},
960fb622 ED	372	{
	373	.procname = "netdev_rss_key",
	374	.data = &netdev_rss_key,
	375	.maxlen = sizeof(int),
	376	.mode = 0444,
	377	.proc_handler = proc_do_rss_key,
	378	},
0a14842f ED	379	#ifdef CONFIG_BPF_JIT
	380	{
	381	.procname = "bpf_jit_enable",
	382	.data = &bpf_jit_enable,
	383	.maxlen = sizeof(int),
	384	.mode = 0644,
2e4a3098	385	.proc_handler = proc_dointvec_minmax_bpf_enable,
fa9dd599	386	# ifdef CONFIG_BPF_JIT_ALWAYS_ON
eec4844f MC	387	.extra1 = SYSCTL_ONE,
eec4844f MC	388	.extra2 = SYSCTL_ONE,
fa9dd599	389	# else
eec4844f	390	.extra1 = SYSCTL_ZERO,
fa9dd599 DB	391	.extra2 = &two,
fa9dd599 DB	392	# endif
0a14842f	393	},
4f3446bb DB	394	# ifdef CONFIG_HAVE_EBPF_JIT
	395	{
	396	.procname = "bpf_jit_harden",
	397	.data = &bpf_jit_harden,
	398	.maxlen = sizeof(int),
	399	.mode = 0600,
2e4a3098	400	.proc_handler = proc_dointvec_minmax_bpf_restricted,
eec4844f	401	.extra1 = SYSCTL_ZERO,
fa9dd599	402	.extra2 = &two,
4f3446bb	403	},
74451e66 DB	404	{
	405	.procname = "bpf_jit_kallsyms",
	406	.data = &bpf_jit_kallsyms,
	407	.maxlen = sizeof(int),
	408	.mode = 0600,
2e4a3098	409	.proc_handler = proc_dointvec_minmax_bpf_restricted,
eec4844f MC	410	.extra1 = SYSCTL_ZERO,
eec4844f MC	411	.extra2 = SYSCTL_ONE,
74451e66	412	},
4f3446bb	413	# endif
ede95a63 DB	414	{
	415	.procname = "bpf_jit_limit",
	416	.data = &bpf_jit_limit,
fdadd049	417	.maxlen = sizeof(long),
ede95a63	418	.mode = 0600,
fdadd049 DB	419	.proc_handler = proc_dolongvec_minmax_bpf_restricted,
	420	.extra1 = &long_one,
	421	.extra2 = &long_max,
ede95a63	422	},
0a14842f	423	#endif
3b098e2d ED	424	{
	425	.procname = "netdev_tstamp_prequeue",
	426	.data = &netdev_tstamp_prequeue,
	427	.maxlen = sizeof(int),
	428	.mode = 0644,
	429	.proc_handler = proc_dointvec
	430	},
1da177e4	431	{
1da177e4	432	.procname = "message_cost",
717115e1	433	.data = &net_ratelimit_state.interval,
1da177e4 LT	434	.maxlen = sizeof(int),
1da177e4 LT	435	.mode = 0644,
6d9f239a	436	.proc_handler = proc_dointvec_jiffies,
1da177e4 LT	437	},
1da177e4 LT	438	{
1da177e4	439	.procname = "message_burst",
717115e1	440	.data = &net_ratelimit_state.burst,
1da177e4 LT	441	.maxlen = sizeof(int),
1da177e4 LT	442	.mode = 0644,
6d9f239a	443	.proc_handler = proc_dointvec,
1da177e4 LT	444	},
1da177e4 LT	445	{
1da177e4 LT	446	.procname = "optmem_max",
	447	.data = &sysctl_optmem_max,
	448	.maxlen = sizeof(int),
	449	.mode = 0644,
6d9f239a	450	.proc_handler = proc_dointvec
1da177e4	451	},
b245be1f WB	452	{
	453	.procname = "tstamp_allow_data",
	454	.data = &sysctl_tstamp_allow_data,
	455	.maxlen = sizeof(int),
	456	.mode = 0644,
	457	.proc_handler = proc_dointvec_minmax,
eec4844f MC	458	.extra1 = SYSCTL_ZERO,
eec4844f MC	459	.extra2 = SYSCTL_ONE
b245be1f	460	},
fec5e652 TH	461	#ifdef CONFIG_RPS
	462	{
	463	.procname = "rps_sock_flow_entries",
	464	.maxlen = sizeof(int),
	465	.mode = 0644,
	466	.proc_handler = rps_sock_flow_sysctl
	467	},
	468	#endif
99bbc707 WB	469	#ifdef CONFIG_NET_FLOW_LIMIT
	470	{
	471	.procname = "flow_limit_cpu_bitmap",
	472	.mode = 0644,
	473	.proc_handler = flow_limit_cpu_sysctl
	474	},
	475	{
	476	.procname = "flow_limit_table_len",
	477	.data = &netdev_flow_limit_table_len,
	478	.maxlen = sizeof(int),
	479	.mode = 0644,
	480	.proc_handler = flow_limit_table_len_sysctl
	481	},
	482	#endif /* CONFIG_NET_FLOW_LIMIT */
e0d1095a	483	#ifdef CONFIG_NET_RX_BUSY_POLL
06021292	484	{
64b0dc51 ET	485	.procname = "busy_poll",
64b0dc51 ET	486	.data = &sysctl_net_busy_poll,
eb6db622	487	.maxlen = sizeof(unsigned int),
06021292	488	.mode = 0644,
95f25521	489	.proc_handler = proc_dointvec_minmax,
eec4844f	490	.extra1 = SYSCTL_ZERO,
06021292	491	},
2d48d67f	492	{
64b0dc51 ET	493	.procname = "busy_read",
64b0dc51 ET	494	.data = &sysctl_net_busy_read,
2d48d67f ET	495	.maxlen = sizeof(unsigned int),
2d48d67f ET	496	.mode = 0644,
95f25521	497	.proc_handler = proc_dointvec_minmax,
eec4844f	498	.extra1 = SYSCTL_ZERO,
2d48d67f	499	},
6da7c8fc	500	#endif
	501	#ifdef CONFIG_NET_SCHED
	502	{
	503	.procname = "default_qdisc",
	504	.mode = 0644,
	505	.maxlen = IFNAMSIZ,
	506	.proc_handler = set_default_qdisc
	507	},
06021292	508	#endif
51b0bded	509	{
51b0bded SH	510	.procname = "netdev_budget",
	511	.data = &netdev_budget,
	512	.maxlen = sizeof(int),
	513	.mode = 0644,
6d9f239a	514	.proc_handler = proc_dointvec
51b0bded	515	},
a2a316fd	516	{
a2a316fd SH	517	.procname = "warnings",
	518	.data = &net_msg_warn,
	519	.maxlen = sizeof(int),
	520	.mode = 0644,
6d9f239a	521	.proc_handler = proc_dointvec
a2a316fd	522	},
5f74f82e HWR	523	{
	524	.procname = "max_skb_frags",
	525	.data = &sysctl_max_skb_frags,
	526	.maxlen = sizeof(int),
	527	.mode = 0644,
	528	.proc_handler = proc_dointvec_minmax,
eec4844f	529	.extra1 = SYSCTL_ONE,
5f74f82e HWR	530	.extra2 = &max_skb_frags,
5f74f82e HWR	531	},
7acf8a1e MW	532	{
	533	.procname = "netdev_budget_usecs",
	534	.data = &netdev_budget_usecs,
	535	.maxlen = sizeof(unsigned int),
	536	.mode = 0644,
	537	.proc_handler = proc_dointvec_minmax,
eec4844f	538	.extra1 = SYSCTL_ZERO,
7acf8a1e	539	},
79134e6c ED	540	{
	541	.procname = "fb_tunnels_only_for_init_net",
	542	.data = &sysctl_fb_tunnels_only_for_init_net,
	543	.maxlen = sizeof(int),
	544	.mode = 0644,
	545	.proc_handler = proc_dointvec_minmax,
eec4844f	546	.extra1 = SYSCTL_ZERO,
316cdaa1	547	.extra2 = &two,
79134e6c	548	},
856c395c CW	549	{
	550	.procname = "devconf_inherit_init_net",
	551	.data = &sysctl_devconf_inherit_init_net,
	552	.maxlen = sizeof(int),
	553	.mode = 0644,
	554	.proc_handler = proc_dointvec_minmax,
eec4844f	555	.extra1 = SYSCTL_ZERO,
9efd6a3c	556	.extra2 = &three,
856c395c	557	},
ce27ec60 ED	558	{
	559	.procname = "high_order_alloc_disable",
	560	.data = &net_high_order_alloc_disable_key.key,
	561	.maxlen = sizeof(net_high_order_alloc_disable_key),
	562	.mode = 0644,
	563	.proc_handler = proc_do_static_key,
	564	},
323ebb61 EC	565	{
	566	.procname = "gro_normal_batch",
	567	.data = &gro_normal_batch,
	568	.maxlen = sizeof(unsigned int),
	569	.mode = 0644,
	570	.proc_handler = proc_dointvec_minmax,
	571	.extra1 = SYSCTL_ONE,
	572	},
f8572d8f	573	{ }
1da177e4	574	};
33eb9cfc	575
d5a4502e PE	576	static struct ctl_table netns_core_table[] = {
d5a4502e PE	577	{
d5a4502e PE	578	.procname = "somaxconn",
	579	.data = &init_net.core.sysctl_somaxconn,
	580	.maxlen = sizeof(int),
	581	.mode = 0644,
eec4844f	582	.extra1 = SYSCTL_ZERO,
5f671d6b	583	.proc_handler = proc_dointvec_minmax
d5a4502e	584	},
f8572d8f	585	{ }
d5a4502e PE	586	};
d5a4502e PE	587
316cdaa1 MB	588	static int __init fb_tunnels_only_for_init_net_sysctl_setup(char *str)
	589	{
	590	/* fallback tunnels for initns only */
	591	if (!strncmp(str, "initns", 6))
	592	sysctl_fb_tunnels_only_for_init_net = 1;
	593	/* no fallback tunnels anywhere */
	594	else if (!strncmp(str, "none", 4))
	595	sysctl_fb_tunnels_only_for_init_net = 2;
	596
	597	return 1;
	598	}
	599	__setup("fb_tunnels=", fb_tunnels_only_for_init_net_sysctl_setup);
	600
024626e3	601	static __net_init int sysctl_core_net_init(struct net *net)
33eb9cfc	602	{
d5a4502e	603	struct ctl_table *tbl;
024626e3	604
d5a4502e	605	tbl = netns_core_table;
09ad9bc7	606	if (!net_eq(net, &init_net)) {
d5a4502e	607	tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL);
024626e3 PE	608	if (tbl == NULL)
	609	goto err_dup;
	610
d5a4502e	611	tbl[0].data = &net->core.sysctl_somaxconn;
464dc801 EB	612
	613	/* Don't export any sysctls to unprivileged users */
	614	if (net->user_ns != &init_user_ns) {
	615	tbl[0].procname = NULL;
	616	}
024626e3 PE	617	}
024626e3 PE	618
ec8f23ce	619	net->core.sysctl_hdr = register_net_sysctl(net, "net/core", tbl);
8efa6e93	620	if (net->core.sysctl_hdr == NULL)
024626e3	621	goto err_reg;
33eb9cfc	622
024626e3 PE	623	return 0;
	624
	625	err_reg:
d5a4502e	626	if (tbl != netns_core_table)
024626e3 PE	627	kfree(tbl);
	628	err_dup:
	629	return -ENOMEM;
	630	}
	631
	632	static __net_exit void sysctl_core_net_exit(struct net *net)
	633	{
	634	struct ctl_table *tbl;
	635
8efa6e93 PE	636	tbl = net->core.sysctl_hdr->ctl_table_arg;
8efa6e93 PE	637	unregister_net_sysctl_table(net->core.sysctl_hdr);
d5a4502e	638	BUG_ON(tbl == netns_core_table);
024626e3 PE	639	kfree(tbl);
	640	}
	641
	642	static __net_initdata struct pernet_operations sysctl_core_ops = {
	643	.init = sysctl_core_net_init,
	644	.exit = sysctl_core_net_exit,
	645	};
	646
	647	static __init int sysctl_core_init(void)
	648	{
43444757	649	register_net_sysctl(&init_net, "net/core", net_core_table);
024626e3	650	return register_pernet_subsys(&sysctl_core_ops);
33eb9cfc PE	651	}
33eb9cfc PE	652
b27aeadb	653	fs_initcall(sysctl_core_init);