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/ipsec.rst b/Documentation/networking/ipsec.rst
new file mode 100644
index 000000000..afe9d7b48
--- /dev/null
+++ b/Documentation/networking/ipsec.rst
@@ -0,0 +1,46 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+=====
+IPsec
+=====
+
+
+Here documents known IPsec corner cases which need to be keep in mind when
+deploy various IPsec configuration in real world production environment.
+
+1. IPcomp:
+	   Small IP packet won't get compressed at sender, and failed on
+	   policy check on receiver.
+
+Quote from RFC3173::
+
+  2.2. Non-Expansion Policy
+
+   If the total size of a compressed payload and the IPComp header, as
+   defined in section 3, is not smaller than the size of the original
+   payload, the IP datagram MUST be sent in the original non-compressed
+   form.  To clarify: If an IP datagram is sent non-compressed, no
+
+   IPComp header is added to the datagram.  This policy ensures saving
+   the decompression processing cycles and avoiding incurring IP
+   datagram fragmentation when the expanded datagram is larger than the
+   MTU.
+
+   Small IP datagrams are likely to expand as a result of compression.
+   Therefore, a numeric threshold should be applied before compression,
+   where IP datagrams of size smaller than the threshold are sent in the
+   original form without attempting compression.  The numeric threshold
+   is implementation dependent.
+
+Current IPComp implementation is indeed by the book, while as in practice
+when sending non-compressed packet to the peer (whether or not packet len
+is smaller than the threshold or the compressed len is larger than original
+packet len), the packet is dropped when checking the policy as this packet
+matches the selector but not coming from any XFRM layer, i.e., with no
+security path. Such naked packet will not eventually make it to upper layer.
+The result is much more wired to the user when ping peer with different
+payload length.
+
+One workaround is try to set "level use" for each policy if user observed
+above scenario. The consequence of doing so is small packet(uncompressed)
+will skip policy checking on receiver side.