x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

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) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

1 files changed, 109 insertions, 0 deletions

diff --git a/net/l2tp/Kconfig b/net/l2tp/Kconfig
new file mode 100644
index 000000000..0de178d5b
--- /dev/null
+++ b/net/l2tp/Kconfig
@@ -0,0 +1,109 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Layer Two Tunneling Protocol (L2TP)
+#
+
+menuconfig L2TP
+	tristate "Layer Two Tunneling Protocol (L2TP)"
+	depends on INET
+	select NET_UDP_TUNNEL
+	help
+	  Layer Two Tunneling Protocol
+
+	  From RFC 2661 <http://www.ietf.org/rfc/rfc2661.txt>.
+
+	  L2TP facilitates the tunneling of packets across an
+	  intervening network in a way that is as transparent as
+	  possible to both end-users and applications.
+
+	  L2TP is often used to tunnel PPP traffic over IP
+	  tunnels. One IP tunnel may carry thousands of individual PPP
+	  connections. L2TP is also used as a VPN protocol, popular
+	  with home workers to connect to their offices.
+
+	  L2TPv3 allows other protocols as well as PPP to be carried
+	  over L2TP tunnels. L2TPv3 is defined in RFC 3931
+	  <http://www.ietf.org/rfc/rfc3931.txt>.
+
+	  The kernel component handles only L2TP data packets: a
+	  userland daemon handles L2TP the control protocol (tunnel
+	  and session setup). One such daemon is OpenL2TP
+	  (http://openl2tp.org/).
+
+	  If you don't need L2TP, say N. To compile all L2TP code as
+	  modules, choose M here.
+
+config L2TP_DEBUGFS
+	tristate "L2TP debugfs support"
+	depends on L2TP && DEBUG_FS
+	help
+	  Support for l2tp directory in debugfs filesystem. This may be
+	  used to dump internal state of the l2tp drivers for problem
+	  analysis.
+
+	  If unsure, say 'Y'.
+
+	  To compile this driver as a module, choose M here. The module
+	  will be called l2tp_debugfs.
+
+config L2TP_V3
+	bool "L2TPv3 support"
+	depends on L2TP
+	help
+	  Layer Two Tunneling Protocol Version 3
+
+	  From RFC 3931 <http://www.ietf.org/rfc/rfc3931.txt>.
+
+	  The Layer Two Tunneling Protocol (L2TP) provides a dynamic
+	  mechanism for tunneling Layer 2 (L2) "circuits" across a
+	  packet-oriented data network (e.g., over IP).  L2TP, as
+	  originally defined in RFC 2661, is a standard method for
+	  tunneling Point-to-Point Protocol (PPP) [RFC1661] sessions.
+	  L2TP has since been adopted for tunneling a number of other
+	  L2 protocols, including ATM, Frame Relay, HDLC and even raw
+	  ethernet frames.
+
+	  If you are connecting to L2TPv3 equipment, or you want to
+	  tunnel raw ethernet frames using L2TP, say Y here. If
+	  unsure, say N.
+
+config L2TP_IP
+	tristate "L2TP IP encapsulation for L2TPv3"
+	depends on L2TP_V3
+	help
+	  Support for L2TP-over-IP socket family.
+
+	  The L2TPv3 protocol defines two possible encapsulations for
+	  L2TP frames, namely UDP and plain IP (without UDP). This
+	  driver provides a new L2TPIP socket family with which
+	  userspace L2TPv3 daemons may create L2TP/IP tunnel sockets
+	  when UDP encapsulation is not required. When L2TP is carried
+	  in IP packets, it used IP protocol number 115, so this port
+	  must be enabled in firewalls.
+
+	  To compile this driver as a module, choose M here. The module
+	  will be called l2tp_ip.
+
+config L2TP_ETH
+	tristate "L2TP ethernet pseudowire support for L2TPv3"
+	depends on L2TP_V3
+	help
+	  Support for carrying raw ethernet frames over L2TPv3.
+
+	  From RFC 4719 <http://www.ietf.org/rfc/rfc4719.txt>.
+
+	  The Layer 2 Tunneling Protocol, Version 3 (L2TPv3) can be
+	  used as a control protocol and for data encapsulation to set
+	  up Pseudowires for transporting layer 2 Packet Data Units
+	  across an IP network [RFC3931].
+
+	  This driver provides an ethernet virtual interface for each
+	  L2TP ethernet pseudowire instance. Standard Linux tools may
+	  be used to assign an IP address to the local virtual
+	  interface, or add the interface to a bridge.
+
+	  If you are using L2TPv3, you will almost certainly want to
+	  enable this option.
+
+	  To compile this driver as a module, choose M here. The module
+	  will be called l2tp_eth.