Re: [BUG] general protection fault in __instance_destroy

Hi,

On Tue, Aug 25, 2026 at 12:25:25AM +0900, Jaeyoung Chung wrote:
> Hello,
> 
> We found a "general protection fault in __instance_destroy" on Linux v7.2.
> The issue was found by our own race fuzzer. We have not analyzed the root cause,
> so we do not have a proposed fix to offer.
> 
> To reproduce the race reliably, we applied the delay patch below to the
> kernel and ran the C reproducer as root inside an x86_64 QEMU guest. The
> crash log we observed, the delay patch and the reproducer are all included
> below.
> 
> The following kernel config options are required to reproduce the issue:
>     CONFIG_NETFILTER=y
>     CONFIG_NETFILTER_NETLINK=y
>     CONFIG_NETFILTER_NETLINK_LOG=y
>     CONFIG_NET=y
>     CONFIG_KASAN=y
> 
> We hope this report is useful. Please let us know if any further
> information would help.

It seems instance_destroy() lost race with nfulnl_rcv_nl_event().

instance_destroy() calls hlist_del_rcu() for an instance that was
already removed by nfulnl_rcv_nl_event().

nfulnl_recv_config() holds a reference on the instance but it is not
sufficient.

> Reported-by: Eulgyu Kim <[email protected]>
> Reported-by: Jaeyoung Chung <[email protected]>
> 
> Kernel delay patch:
> ==================================================================
> diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
> index 6c7fa2ed34f5..04d2250f992a 100644
> --- a/net/netfilter/nfnetlink_log.c
> +++ b/net/netfilter/nfnetlink_log.c
> @@ -37,6 +37,8 @@
>  
>  #include <linux/atomic.h>
>  #include <linux/refcount.h>
> +#include <linux/delay.h>
> +#include <linux/sched.h>
>  
>  
>  #if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
> @@ -982,6 +984,9 @@ static int nfulnl_recv_config(struct sk_buff *skb, const struct nfnl_info *info,
>  				goto out;
>  			}
>  
> +			if (strncmp(current->comm, "syzrepro1", 10) == 0) {
> +				mdelay(100);
> +			}
>  			instance_destroy(log, inst);
>  			goto out_put;
>  		default:
> diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
> index 5202fe0b0867..0ba2ab6c4547 100644
> --- a/net/netlink/af_netlink.c
> +++ b/net/netlink/af_netlink.c
> @@ -25,6 +25,7 @@
>  #include <linux/kernel.h>
>  #include <linux/filter.h>
>  #include <linux/init.h>
> +#include <linux/delay.h>
>  #include <linux/signal.h>
>  #include <linux/sched.h>
>  #include <linux/errno.h>
> @@ -724,6 +725,9 @@ static int netlink_release(struct socket *sock)
n>  		return 0;
>  
>  	netlink_remove(sk);
> +	if (strncmp(current->comm, "syzrepro0", 10) == 0) {
> +		mdelay(60);
> +	}
>  	sock_orphan(sk);
>  	nlk = nlk_sk(sk);
>  
> ==================================================================
> 
> C reproducer:
> ==================================================================
> #define _GNU_SOURCE
> #include <errno.h>
> #include <pthread.h>
> #include <sched.h>
> #include <stdio.h>
> #include <stdlib.h>
> #include <string.h>
> #include <sys/prctl.h>
> #include <sys/socket.h>
> #include <unistd.h>
> #include <linux/netlink.h>
> 
> #define SYSCHK(x) ({ long __r = (long)(x); if (__r == -1L) { perror(#x); exit(1); } __r; })
> 
> #define NETLINK_NETFILTER 12
> #define NFNL_SUBSYS_ULOG 4
> #define NFULNL_MSG_CONFIG 1
> #define NFULA_CFG_CMD 1
> #define BIND 1
> #define UNBIND 2
> 
> /* Two threads take the same portid in turn, overlapping BIND/UNBIND on the
>  * same group 0 instance
>  */
> #define PORTID 0x00524143u
> #define SPIN 4000000
> 
> static volatile unsigned g_go, g_done;
> static volatile int g_stop;
> 
> static int build_cfg(char *buf, unsigned char cmd)
> {
> 	struct nlmsghdr *nlh = (struct nlmsghdr *)buf;
> 	unsigned short nla[2] = { 5, NFULA_CFG_CMD };
> 	unsigned char *p = (unsigned char *)buf + NLMSG_HDRLEN;
> 
> 	memset(buf, 0, 64);
> 	nlh->nlmsg_len = NLMSG_HDRLEN + 12;
> 	nlh->nlmsg_type = (NFNL_SUBSYS_ULOG << 8) | NFULNL_MSG_CONFIG;
> 	nlh->nlmsg_flags = NLM_F_REQUEST;
> 	nlh->nlmsg_seq = 1;
> 	memcpy(p + 4, nla, 4);
> 	p[8] = cmd;
> 	return NLMSG_HDRLEN + 12;
> }
> 
> static int nl_send(int fd, const char *buf, int len)
> {
> 	struct sockaddr_nl dst = { .nl_family = AF_NETLINK };
> 	struct iovec iov = { (void *)buf, len };
> 	struct msghdr mh = { .msg_name = &dst, .msg_namelen = sizeof(dst),
> 			     .msg_iov = &iov, .msg_iovlen = 1 };
> 
> 	return sendmsg(fd, &mh, 0);
> }
> 
> static int nl_open_bind(void)
> {
> 	struct sockaddr_nl a = { .nl_family = AF_NETLINK, .nl_pid = PORTID };
> 	int fd = SYSCHK(socket(AF_NETLINK, SOCK_RAW, NETLINK_NETFILTER));
> 	long spin;
> 
> 	for (spin = 0; spin < SPIN && !g_stop; spin++) {
> 		if (bind(fd, (struct sockaddr *)&a, sizeof(a)) == 0)
> 			return fd;
> 		if (errno != EADDRINUSE)
> 			break;
> 		sched_yield();
> 	}
> 	close(fd);
> 	return -1;
> }
> 
> static void *binder(void *u)
> {
> 	char buf[64];
> 	int len, i;
> 
> 	(void)u;
> 	prctl(PR_SET_NAME, "syzrepro0", 0, 0, 0);
> 	len = build_cfg(buf, BIND);
> 	for (i = 1; i <= 120 && !g_stop; i++) {
> 		long spin;
> 		int fd = nl_open_bind();
> 
> 		if (fd < 0 || nl_send(fd, buf, len) < 0) {
> 			if (fd >= 0)
> 				close(fd);
> 			break;
> 		}
> 		__sync_synchronize();
> 		g_go = i;
> 		__sync_synchronize();
> 		close(fd);
> 		for (spin = 0; g_done < (unsigned)i && spin < SPIN; spin++)
> 			sched_yield();
> 	}
> 	g_stop = 1;
> 	return NULL;
> }
> 
> static void *unbinder(void *u)
> {
> 	char buf[64];
> 	unsigned last = 0;
> 	int len;
> 
> 	(void)u;
> 	prctl(PR_SET_NAME, "syzrepro1", 0, 0, 0);
> 	len = build_cfg(buf, UNBIND);
> 	while (!g_stop) {
> 		unsigned cur;
> 		int fd;
> 
> 		for (;;) {
> 			cur = g_go;
> 			if (cur != last)
> 				break;
> 			if (g_stop)
> 				return NULL;
> 			sched_yield();
> 		}
> 		last = cur;
> 		fd = nl_open_bind();
> 		if (fd >= 0) {
> 			nl_send(fd, buf, len);
> 			close(fd);
> 		}
> 		__sync_synchronize();
> 		g_done = cur;
> 		__sync_synchronize();
> 	}
> 	return NULL;
> }
> 
> int main(void)
> {
> 	pthread_t t0, t1;
> 
> 	pthread_create(&t1, NULL, unbinder, NULL);
> 	pthread_create(&t0, NULL, binder, NULL);
> 	pthread_join(t0, NULL);
> 	g_stop = 1;
> 	pthread_join(t1, NULL);
> 	return 0;
> }
> ==================================================================
> 
> Crash log:
> ==================================================================
> Oops: general protection fault, probably for non-canonical address 0xfbd59c0000000024: 0000 [#1] SMP KASAN PTI
> KASAN: maybe wild-memory-access in range [0xdead000000000120-0xdead000000000127]
> CPU: 0 UID: 0 PID: 401 Comm: syzrepro1 Not tainted 7.2.0-dirty #2 PREEMPT 
> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014
> RIP: 0010:__hlist_del include/linux/list.h:1029 [inline]
> RIP: 0010:hlist_del_rcu include/linux/rculist.h:599 [inline]
> RIP: 0010:__instance_destroy+0x5a/0x190 net/netfilter/nfnetlink_log.c:234
> Code: f2 fc 4c 8b 3b 4c 8d 73 08 4c 89 f5 48 c1 ed 03 42 80 7c 2d 00 00 74 08 4c 89 f7 e8 60 b5 f2 fc 4d 8b 26 4c 89 e0 48 c1 e8 03 <42> 80 3c 28 00 74 08 4c 89 e7 e8 37 b6 f2 fc 4d 89 3c 24 4d 85 ff
> RSP: 0018:ffff88811118f350 EFLAGS: 00010a06
> RAX: 1bd5a00000000024 RBX: ffff888107955000 RCX: 0000000000000001
> RDX: 0000000000000001 RSI: 0000000000000004 RDI: ffff888107955000
> RBP: 1ffff11020f2aa01 R08: ffff88811118f36f R09: 1ffff11022231e6d
> R10: dffffc0000000000 R11: ffffed1022231e6e R12: dead000000000122
> R13: dffffc0000000000 R14: ffff888107955008 R15: 0000000000000000
> FS:  00007b5300d0a6c0(0000) GS:ffff88817d75f000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00007b5300dfc010 CR3: 000000011127e000 CR4: 00000000000006f0
> Call Trace:
>  <TASK>
>  instance_destroy net/netfilter/nfnetlink_log.c:256 [inline]
>  nfulnl_recv_config+0x60e/0xf70 net/netfilter/nfnetlink_log.c:990
>  nfnetlink_rcv_msg+0x5d5/0x760 net/netfilter/nfnetlink.c:300
>  netlink_rcv_skb+0x168/0x310 net/netlink/af_netlink.c:2560
>  nfnetlink_rcv+0x370/0x2140 net/netfilter/nfnetlink.c:667
>  netlink_unicast_kernel net/netlink/af_netlink.c:1323 [inline]
>  netlink_unicast+0x652/0x880 net/netlink/af_netlink.c:1349
>  netlink_sendmsg+0x5a1/0x870 net/netlink/af_netlink.c:1904
>  sock_sendmsg_nosec net/socket.c:775 [inline]
>  __sock_sendmsg+0x18f/0x1a0 net/socket.c:790
>  ____sys_sendmsg+0x468/0x660 net/socket.c:2684
>  ___sys_sendmsg+0x15e/0x1a0 net/socket.c:2738
>  __sys_sendmsg net/socket.c:2770 [inline]
>  __do_sys_sendmsg net/socket.c:2775 [inline]
>  __se_sys_sendmsg net/socket.c:2773 [inline]
>  __x64_sys_sendmsg+0x11e/0x170 net/socket.c:2773
>  do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
>  do_syscall_64+0xf7/0x370 arch/x86/entry/syscall_64.c:94
>  entry_SYSCALL_64_after_hwframe+0x76/0x7e
> RIP: 0033:0x7b5300e18c4d
> Code: 28 89 54 24 1c 48 89 74 24 10 89 7c 24 08 e8 ea ab f7 ff 8b 54 24 1c 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 33 44 89 c7 48 89 44 24 08 e8 3e ac f7 ff 48
> RSP: 002b:00007b5300d09dc0 EFLAGS: 00000293 ORIG_RAX: 000000000000002e
> RAX: ffffffffffffffda RBX: 0000000000000028 RCX: 00007b5300e18c4d
> RDX: 0000000000000000 RSI: 00007b5300d09e10 RDI: 0000000000000003
> RBP: 0000000000000003 R08: 0000000000000000 R09: 00007ffd57a75617
> R10: 0000000000000000 R11: 0000000000000293 R12: 000000000000001c
> R13: 00007b5300d09e60 R14: 00007ffd57a75520 R15: 00007b530050a000
>  </TASK>
> Modules linked in:
> ---[ end trace 0000000000000000 ]---
> ==================================================================
> 
>