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, 46 insertions, 0 deletions

diff --git a/Documentation/networking/x25.rst b/Documentation/networking/x25.rst
new file mode 100644
index 000000000..e11d9ebdf
--- /dev/null
+++ b/Documentation/networking/x25.rst
@@ -0,0 +1,46 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+==================
+Linux X.25 Project
+==================
+
+As my third year dissertation at University I have taken it upon myself to
+write an X.25 implementation for Linux. My aim is to provide a complete X.25
+Packet Layer and a LAPB module to allow for "normal" X.25 to be run using
+Linux. There are two sorts of X.25 cards available, intelligent ones that
+implement LAPB on the card itself, and unintelligent ones that simply do
+framing, bit-stuffing and checksumming. These both need to be handled by the
+system.
+
+I therefore decided to write the implementation such that as far as the
+Packet Layer is concerned, the link layer was being performed by a lower
+layer of the Linux kernel and therefore it did not concern itself with
+implementation of LAPB. Therefore the LAPB modules would be called by
+unintelligent X.25 card drivers and not by intelligent ones, this would
+provide a uniform device driver interface, and simplify configuration.
+
+To confuse matters a little, an 802.2 LLC implementation is also possible
+which could allow X.25 to be run over an Ethernet (or Token Ring) and
+conform with the JNT "Pink Book", this would have a different interface to
+the Packet Layer but there would be no confusion since the class of device
+being served by the LLC would be completely separate from LAPB.
+
+Just when you thought that it could not become more confusing, another
+option appeared, XOT. This allows X.25 Packet Layer frames to operate over
+the Internet using TCP/IP as a reliable link layer. RFC1613 specifies the
+format and behaviour of the protocol. If time permits this option will also
+be actively considered.
+
+A linux-x25 mailing list has been created at vger.kernel.org to support the
+development and use of Linux X.25. It is early days yet, but interested
+parties are welcome to subscribe to it. Just send a message to
+majordomo@vger.kernel.org with the following in the message body:
+
+subscribe linux-x25
+end
+
+The contents of the Subject line are ignored.
+
+Jonathan
+
+g4klx@g4klx.demon.co.uk