forked from Minki/linux
4db428a7c9
reuseport_add_sock() needs to deal with attaching a socket having
its own sk_reuseport_cb, after a prior
setsockopt(SO_ATTACH_REUSEPORT_?BPF)
Without this fix, not only a WARN_ONCE() was issued, but we were also
leaking memory.
Thanks to sysbot and Eric Biggers for providing us nice C repros.
------------[ cut here ]------------
socket already in reuseport group
WARNING: CPU: 0 PID: 3496 at net/core/sock_reuseport.c:119
reuseport_add_sock+0x742/0x9b0 net/core/sock_reuseport.c:117
Kernel panic - not syncing: panic_on_warn set ...
CPU: 0 PID: 3496 Comm: syzkaller869503 Not tainted 4.15.0-rc6+ #245
Hardware name: Google Google Compute Engine/Google Compute Engine,
BIOS
Google 01/01/2011
Call Trace:
__dump_stack lib/dump_stack.c:17 [inline]
dump_stack+0x194/0x257 lib/dump_stack.c:53
panic+0x1e4/0x41c kernel/panic.c:183
__warn+0x1dc/0x200 kernel/panic.c:547
report_bug+0x211/0x2d0 lib/bug.c:184
fixup_bug.part.11+0x37/0x80 arch/x86/kernel/traps.c:178
fixup_bug arch/x86/kernel/traps.c:247 [inline]
do_error_trap+0x2d7/0x3e0 arch/x86/kernel/traps.c:296
do_invalid_op+0x1b/0x20 arch/x86/kernel/traps.c:315
invalid_op+0x22/0x40 arch/x86/entry/entry_64.S:1079
Fixes: ef456144da
("soreuseport: define reuseport groups")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot+c0ea2226f77a42936bf7@syzkaller.appspotmail.com
Acked-by: Craig Gallek <kraig@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
272 lines
6.8 KiB
C
272 lines
6.8 KiB
C
// SPDX-License-Identifier: GPL-2.0
|
|
/*
|
|
* To speed up listener socket lookup, create an array to store all sockets
|
|
* listening on the same port. This allows a decision to be made after finding
|
|
* the first socket. An optional BPF program can also be configured for
|
|
* selecting the socket index from the array of available sockets.
|
|
*/
|
|
|
|
#include <net/sock_reuseport.h>
|
|
#include <linux/bpf.h>
|
|
#include <linux/rcupdate.h>
|
|
|
|
#define INIT_SOCKS 128
|
|
|
|
static DEFINE_SPINLOCK(reuseport_lock);
|
|
|
|
static struct sock_reuseport *__reuseport_alloc(unsigned int max_socks)
|
|
{
|
|
unsigned int size = sizeof(struct sock_reuseport) +
|
|
sizeof(struct sock *) * max_socks;
|
|
struct sock_reuseport *reuse = kzalloc(size, GFP_ATOMIC);
|
|
|
|
if (!reuse)
|
|
return NULL;
|
|
|
|
reuse->max_socks = max_socks;
|
|
|
|
RCU_INIT_POINTER(reuse->prog, NULL);
|
|
return reuse;
|
|
}
|
|
|
|
int reuseport_alloc(struct sock *sk)
|
|
{
|
|
struct sock_reuseport *reuse;
|
|
|
|
/* bh lock used since this function call may precede hlist lock in
|
|
* soft irq of receive path or setsockopt from process context
|
|
*/
|
|
spin_lock_bh(&reuseport_lock);
|
|
|
|
/* Allocation attempts can occur concurrently via the setsockopt path
|
|
* and the bind/hash path. Nothing to do when we lose the race.
|
|
*/
|
|
if (rcu_dereference_protected(sk->sk_reuseport_cb,
|
|
lockdep_is_held(&reuseport_lock)))
|
|
goto out;
|
|
|
|
reuse = __reuseport_alloc(INIT_SOCKS);
|
|
if (!reuse) {
|
|
spin_unlock_bh(&reuseport_lock);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
reuse->socks[0] = sk;
|
|
reuse->num_socks = 1;
|
|
rcu_assign_pointer(sk->sk_reuseport_cb, reuse);
|
|
|
|
out:
|
|
spin_unlock_bh(&reuseport_lock);
|
|
|
|
return 0;
|
|
}
|
|
EXPORT_SYMBOL(reuseport_alloc);
|
|
|
|
static struct sock_reuseport *reuseport_grow(struct sock_reuseport *reuse)
|
|
{
|
|
struct sock_reuseport *more_reuse;
|
|
u32 more_socks_size, i;
|
|
|
|
more_socks_size = reuse->max_socks * 2U;
|
|
if (more_socks_size > U16_MAX)
|
|
return NULL;
|
|
|
|
more_reuse = __reuseport_alloc(more_socks_size);
|
|
if (!more_reuse)
|
|
return NULL;
|
|
|
|
more_reuse->max_socks = more_socks_size;
|
|
more_reuse->num_socks = reuse->num_socks;
|
|
more_reuse->prog = reuse->prog;
|
|
|
|
memcpy(more_reuse->socks, reuse->socks,
|
|
reuse->num_socks * sizeof(struct sock *));
|
|
|
|
for (i = 0; i < reuse->num_socks; ++i)
|
|
rcu_assign_pointer(reuse->socks[i]->sk_reuseport_cb,
|
|
more_reuse);
|
|
|
|
/* Note: we use kfree_rcu here instead of reuseport_free_rcu so
|
|
* that reuse and more_reuse can temporarily share a reference
|
|
* to prog.
|
|
*/
|
|
kfree_rcu(reuse, rcu);
|
|
return more_reuse;
|
|
}
|
|
|
|
static void reuseport_free_rcu(struct rcu_head *head)
|
|
{
|
|
struct sock_reuseport *reuse;
|
|
|
|
reuse = container_of(head, struct sock_reuseport, rcu);
|
|
if (reuse->prog)
|
|
bpf_prog_destroy(reuse->prog);
|
|
kfree(reuse);
|
|
}
|
|
|
|
/**
|
|
* reuseport_add_sock - Add a socket to the reuseport group of another.
|
|
* @sk: New socket to add to the group.
|
|
* @sk2: Socket belonging to the existing reuseport group.
|
|
* May return ENOMEM and not add socket to group under memory pressure.
|
|
*/
|
|
int reuseport_add_sock(struct sock *sk, struct sock *sk2)
|
|
{
|
|
struct sock_reuseport *old_reuse, *reuse;
|
|
|
|
if (!rcu_access_pointer(sk2->sk_reuseport_cb)) {
|
|
int err = reuseport_alloc(sk2);
|
|
|
|
if (err)
|
|
return err;
|
|
}
|
|
|
|
spin_lock_bh(&reuseport_lock);
|
|
reuse = rcu_dereference_protected(sk2->sk_reuseport_cb,
|
|
lockdep_is_held(&reuseport_lock));
|
|
old_reuse = rcu_dereference_protected(sk->sk_reuseport_cb,
|
|
lockdep_is_held(&reuseport_lock));
|
|
if (old_reuse && old_reuse->num_socks != 1) {
|
|
spin_unlock_bh(&reuseport_lock);
|
|
return -EBUSY;
|
|
}
|
|
|
|
if (reuse->num_socks == reuse->max_socks) {
|
|
reuse = reuseport_grow(reuse);
|
|
if (!reuse) {
|
|
spin_unlock_bh(&reuseport_lock);
|
|
return -ENOMEM;
|
|
}
|
|
}
|
|
|
|
reuse->socks[reuse->num_socks] = sk;
|
|
/* paired with smp_rmb() in reuseport_select_sock() */
|
|
smp_wmb();
|
|
reuse->num_socks++;
|
|
rcu_assign_pointer(sk->sk_reuseport_cb, reuse);
|
|
|
|
spin_unlock_bh(&reuseport_lock);
|
|
|
|
if (old_reuse)
|
|
call_rcu(&old_reuse->rcu, reuseport_free_rcu);
|
|
return 0;
|
|
}
|
|
|
|
void reuseport_detach_sock(struct sock *sk)
|
|
{
|
|
struct sock_reuseport *reuse;
|
|
int i;
|
|
|
|
spin_lock_bh(&reuseport_lock);
|
|
reuse = rcu_dereference_protected(sk->sk_reuseport_cb,
|
|
lockdep_is_held(&reuseport_lock));
|
|
rcu_assign_pointer(sk->sk_reuseport_cb, NULL);
|
|
|
|
for (i = 0; i < reuse->num_socks; i++) {
|
|
if (reuse->socks[i] == sk) {
|
|
reuse->socks[i] = reuse->socks[reuse->num_socks - 1];
|
|
reuse->num_socks--;
|
|
if (reuse->num_socks == 0)
|
|
call_rcu(&reuse->rcu, reuseport_free_rcu);
|
|
break;
|
|
}
|
|
}
|
|
spin_unlock_bh(&reuseport_lock);
|
|
}
|
|
EXPORT_SYMBOL(reuseport_detach_sock);
|
|
|
|
static struct sock *run_bpf(struct sock_reuseport *reuse, u16 socks,
|
|
struct bpf_prog *prog, struct sk_buff *skb,
|
|
int hdr_len)
|
|
{
|
|
struct sk_buff *nskb = NULL;
|
|
u32 index;
|
|
|
|
if (skb_shared(skb)) {
|
|
nskb = skb_clone(skb, GFP_ATOMIC);
|
|
if (!nskb)
|
|
return NULL;
|
|
skb = nskb;
|
|
}
|
|
|
|
/* temporarily advance data past protocol header */
|
|
if (!pskb_pull(skb, hdr_len)) {
|
|
kfree_skb(nskb);
|
|
return NULL;
|
|
}
|
|
index = bpf_prog_run_save_cb(prog, skb);
|
|
__skb_push(skb, hdr_len);
|
|
|
|
consume_skb(nskb);
|
|
|
|
if (index >= socks)
|
|
return NULL;
|
|
|
|
return reuse->socks[index];
|
|
}
|
|
|
|
/**
|
|
* reuseport_select_sock - Select a socket from an SO_REUSEPORT group.
|
|
* @sk: First socket in the group.
|
|
* @hash: When no BPF filter is available, use this hash to select.
|
|
* @skb: skb to run through BPF filter.
|
|
* @hdr_len: BPF filter expects skb data pointer at payload data. If
|
|
* the skb does not yet point at the payload, this parameter represents
|
|
* how far the pointer needs to advance to reach the payload.
|
|
* Returns a socket that should receive the packet (or NULL on error).
|
|
*/
|
|
struct sock *reuseport_select_sock(struct sock *sk,
|
|
u32 hash,
|
|
struct sk_buff *skb,
|
|
int hdr_len)
|
|
{
|
|
struct sock_reuseport *reuse;
|
|
struct bpf_prog *prog;
|
|
struct sock *sk2 = NULL;
|
|
u16 socks;
|
|
|
|
rcu_read_lock();
|
|
reuse = rcu_dereference(sk->sk_reuseport_cb);
|
|
|
|
/* if memory allocation failed or add call is not yet complete */
|
|
if (!reuse)
|
|
goto out;
|
|
|
|
prog = rcu_dereference(reuse->prog);
|
|
socks = READ_ONCE(reuse->num_socks);
|
|
if (likely(socks)) {
|
|
/* paired with smp_wmb() in reuseport_add_sock() */
|
|
smp_rmb();
|
|
|
|
if (prog && skb)
|
|
sk2 = run_bpf(reuse, socks, prog, skb, hdr_len);
|
|
|
|
/* no bpf or invalid bpf result: fall back to hash usage */
|
|
if (!sk2)
|
|
sk2 = reuse->socks[reciprocal_scale(hash, socks)];
|
|
}
|
|
|
|
out:
|
|
rcu_read_unlock();
|
|
return sk2;
|
|
}
|
|
EXPORT_SYMBOL(reuseport_select_sock);
|
|
|
|
struct bpf_prog *
|
|
reuseport_attach_prog(struct sock *sk, struct bpf_prog *prog)
|
|
{
|
|
struct sock_reuseport *reuse;
|
|
struct bpf_prog *old_prog;
|
|
|
|
spin_lock_bh(&reuseport_lock);
|
|
reuse = rcu_dereference_protected(sk->sk_reuseport_cb,
|
|
lockdep_is_held(&reuseport_lock));
|
|
old_prog = rcu_dereference_protected(reuse->prog,
|
|
lockdep_is_held(&reuseport_lock));
|
|
rcu_assign_pointer(reuse->prog, prog);
|
|
spin_unlock_bh(&reuseport_lock);
|
|
|
|
return old_prog;
|
|
}
|
|
EXPORT_SYMBOL(reuseport_attach_prog);
|