From a5f7a5bb3b7f28ba7e4fa246775b29a0e5537255 Mon Sep 17 00:00:00 2001 From: Jinke Han Date: Tue, 8 Sep 2026 15:37:42 +0800 Subject: x86/kprobes: Fix crash when probing CS CALL instructions When using eBPF to probe CS CALL instructions within a function, a crash can be triggered. The eBPF tool probes offset 257 of the __hrtimer_run_queues() function: <__hrtimer_run_queues+249>: nopl 0x0(%rax,%rax,1) <__hrtimer_run_queues+254>: mov %r14,%rdi <__hrtimer_run_queues+257>: cs call <__x86_indirect_thunk_r12> <__hrtimer_run_queues+263>: mov %eax,%r12d <__hrtimer_run_queues+266>: xchg %ax,%ax <__hrtimer_run_queues+268>: mov %r13,%rdi Which triggers this crash: BUG: unable to handle page fault for address: 00000000000f41c9 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 0 P4D 0 Oops: 0002 [#1] SMP NOPTI CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P RIP: 0010:__hrtimer_run_queues+0x106/0x230 Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is at the 6th byte of the above CS CALL instruction. Since the CS CALL instruction occupies 6 bytes, the exception occurred in the middle of that call instruction. The root cause is that when using eBPF tools to probe in the middle of a function, a kprobe with INT3 is used as the underlying implementation. During single-step emulation of the original CALL instruction, int3_emulate_call() assumes that the probed CALL instruction is 5 bytes long. However, the actual CS-prefixed CALL instruction occupies 6 bytes, so it constructs an incorrect exception return address. When the CPU returns from the kprobe handler, the next instruction to be executed is at the address of the last byte of that CS CALL instruction. Coincidentally, starting from that address, the CPU fetches and decodes a completely different instruction, which ultimately triggers a kernel crash. Fix the issue by using the actual instruction length obtained from the instruction decoder when constructing the exception return address, rather than relying on the hardcoded CALL_INSN_SIZE macro. [ mingo: Refined the changelog ] Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step") Suggested-by: Masami Hiramatsu (Google) Signed-off-by: Jinke Han Signed-off-by: Ingo Molnar Reviewed-by: Masami Hiramatsu (Google) Acked-by: Yafang Shao Acked-by: Borislav Petkov Cc: Peter Zijlstra Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000 --- net/handshake/netlink.c | 287 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 287 insertions(+) create mode 100644 net/handshake/netlink.c (limited to 'net/handshake/netlink.c') diff --git a/net/handshake/netlink.c b/net/handshake/netlink.c new file mode 100644 index 000000000..3fd4fef9b --- /dev/null +++ b/net/handshake/netlink.c @@ -0,0 +1,287 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Generic netlink handshake service + * + * Author: Chuck Lever + * + * Copyright (c) 2023, Oracle and/or its affiliates. + */ + +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +#include + +#include +#include "handshake.h" +#include "genl.h" + +#include + +/** + * handshake_genl_notify - Notify handlers that a request is waiting + * @net: target network namespace + * @proto: handshake protocol + * @flags: memory allocation control flags + * + * Returns zero on success or a negative errno if notification failed. + */ +int handshake_genl_notify(struct net *net, const struct handshake_proto *proto, + gfp_t flags) +{ + struct sk_buff *msg; + void *hdr; + + /* Disable notifications during unit testing */ + if (!test_bit(HANDSHAKE_F_PROTO_NOTIFY, &proto->hp_flags)) + return 0; + + if (!genl_has_listeners(&handshake_nl_family, net, + proto->hp_handler_class)) + return -ESRCH; + + msg = genlmsg_new(GENLMSG_DEFAULT_SIZE, flags); + if (!msg) + return -ENOMEM; + + hdr = genlmsg_put(msg, 0, 0, &handshake_nl_family, 0, + HANDSHAKE_CMD_READY); + if (!hdr) + goto out_free; + + if (nla_put_u32(msg, HANDSHAKE_A_ACCEPT_HANDLER_CLASS, + proto->hp_handler_class) < 0) { + genlmsg_cancel(msg, hdr); + goto out_free; + } + + genlmsg_end(msg, hdr); + return genlmsg_multicast_netns(&handshake_nl_family, net, msg, + 0, proto->hp_handler_class, flags); + +out_free: + nlmsg_free(msg); + return -EMSGSIZE; +} + +/** + * handshake_genl_put - Create a generic netlink message header + * @msg: buffer in which to create the header + * @info: generic netlink message context + * + * Returns a ready-to-use header, or NULL. + */ +struct nlmsghdr *handshake_genl_put(struct sk_buff *msg, + struct genl_info *info) +{ + return genlmsg_put(msg, info->snd_portid, info->snd_seq, + &handshake_nl_family, 0, info->genlhdr->cmd); +} +EXPORT_SYMBOL(handshake_genl_put); + +int handshake_nl_accept_doit(struct sk_buff *skb, struct genl_info *info) +{ + struct net *net = sock_net(skb->sk); + struct handshake_net *hn = handshake_pernet(net); + struct handshake_req *req = NULL; + int class, err; + + err = -EOPNOTSUPP; + if (!hn) + goto out_status; + + err = -EINVAL; + if (GENL_REQ_ATTR_CHECK(info, HANDSHAKE_A_ACCEPT_HANDLER_CLASS)) + goto out_status; + class = nla_get_u32(info->attrs[HANDSHAKE_A_ACCEPT_HANDLER_CLASS]); + + err = -EAGAIN; + req = handshake_req_next(hn, class); + if (req) { + FD_PREPARE(fdf, O_CLOEXEC, req->hr_file); + if (fdf.err) { + fput(req->hr_file); /* drop ref from handshake_req_next() */ + err = fdf.err; + goto out_complete; + } + + err = req->hr_proto->hp_accept(req, info, fd_prepare_fd(fdf)); + if (err) + goto out_complete; /* Automatic cleanup handles fput */ + + trace_handshake_cmd_accept(net, req, req->hr_sk, fd_prepare_fd(fdf)); + fd_publish(fdf); + return 0; + } + +out_complete: + if (req) + handshake_complete(req, -EIO, NULL); +out_status: + trace_handshake_cmd_accept_err(net, req, NULL, err); + return err; +} + +int handshake_nl_done_doit(struct sk_buff *skb, struct genl_info *info) +{ + struct net *net = sock_net(skb->sk); + struct handshake_req *req; + struct socket *sock; + int fd, status, err; + + if (GENL_REQ_ATTR_CHECK(info, HANDSHAKE_A_DONE_SOCKFD)) + return -EINVAL; + fd = nla_get_s32(info->attrs[HANDSHAKE_A_DONE_SOCKFD]); + + sock = sockfd_lookup(fd, &err); + if (!sock) + return err; + + req = handshake_req_hash_lookup(sock->sk); + if (!req) { + err = -EBUSY; + trace_handshake_cmd_done_err(net, req, sock->sk, err); + sockfd_put(sock); + return err; + } + + trace_handshake_cmd_done(net, req, sock->sk, fd); + + status = -EIO; + if (info->attrs[HANDSHAKE_A_DONE_STATUS]) + status = -(int)nla_get_u32(info->attrs[HANDSHAKE_A_DONE_STATUS]); + + handshake_complete(req, status, info); + sockfd_put(sock); + return 0; +} + +static unsigned int handshake_net_id; + +static int __net_init handshake_net_init(struct net *net) +{ + struct handshake_net *hn = net_generic(net, handshake_net_id); + unsigned long tmp; + struct sysinfo si; + + /* + * Arbitrary limit to prevent handshakes that do not make + * progress from clogging up the system. The cap scales up + * with the amount of physical memory on the system. + */ + si_meminfo(&si); + tmp = si.totalram / (25 * si.mem_unit); + hn->hn_pending_max = clamp(tmp, 3UL, 50UL); + + spin_lock_init(&hn->hn_lock); + hn->hn_pending = 0; + hn->hn_flags = 0; + INIT_LIST_HEAD(&hn->hn_requests); + return 0; +} + +static void __net_exit handshake_net_exit(struct net *net) +{ + struct handshake_net *hn = net_generic(net, handshake_net_id); + struct handshake_req *req; + LIST_HEAD(requests); + + /* + * Drain the net's pending list. Requests that have been + * accepted and are in progress will be destroyed when + * the socket is closed. + */ + spin_lock_bh(&hn->hn_lock); + set_bit(HANDSHAKE_F_NET_DRAINING, &hn->hn_flags); + list_splice_init(&hn->hn_requests, &requests); + list_for_each_entry(req, &requests, hr_list) + get_file(req->hr_file); + spin_unlock_bh(&hn->hn_lock); + + while (!list_empty(&requests)) { + struct file *file; + + req = list_first_entry(&requests, struct handshake_req, hr_list); + file = req->hr_file; + list_del_init(&req->hr_list); + handshake_complete(req, -ETIMEDOUT, NULL); + fput(file); + } +} + +static struct pernet_operations handshake_genl_net_ops = { + .init = handshake_net_init, + .exit = handshake_net_exit, + .id = &handshake_net_id, + .size = sizeof(struct handshake_net), +}; + +/** + * handshake_pernet - Get the handshake private per-net structure + * @net: network namespace + * + * Returns a pointer to the net's private per-net structure for the + * handshake module, or NULL if handshake_init() failed. + */ +struct handshake_net *handshake_pernet(struct net *net) +{ + return handshake_net_id ? + net_generic(net, handshake_net_id) : NULL; +} +EXPORT_SYMBOL_IF_KUNIT(handshake_pernet); + +static int __init handshake_init(void) +{ + int ret; + + ret = handshake_req_hash_init(); + if (ret) { + pr_warn("handshake: hash initialization failed (%d)\n", ret); + return ret; + } + + ret = genl_register_family(&handshake_nl_family); + if (ret) { + pr_warn("handshake: netlink registration failed (%d)\n", ret); + handshake_req_hash_destroy(); + return ret; + } + + /* + * ORDER: register_pernet_subsys must be done last. + * + * If initialization does not make it past pernet_subsys + * registration, then handshake_net_id will remain 0. That + * shunts the handshake consumer API to return ENOTSUPP + * to prevent it from dereferencing something that hasn't + * been allocated. + */ + ret = register_pernet_subsys(&handshake_genl_net_ops); + if (ret) { + pr_warn("handshake: pernet registration failed (%d)\n", ret); + genl_unregister_family(&handshake_nl_family); + handshake_req_hash_destroy(); + } + + return ret; +} + +static void __exit handshake_exit(void) +{ + unregister_pernet_subsys(&handshake_genl_net_ops); + handshake_net_id = 0; + + handshake_req_hash_destroy(); + genl_unregister_family(&handshake_nl_family); +} + +module_init(handshake_init); +module_exit(handshake_exit); -- cgit v1.3.1