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, 95 insertions, 0 deletions

diff --git a/Documentation/networking/netif-msg.rst b/Documentation/networking/netif-msg.rst
new file mode 100644
index 000000000..b20d265a7
--- /dev/null
+++ b/Documentation/networking/netif-msg.rst
@@ -0,0 +1,95 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+===============
+NETIF Msg Level
+===============
+
+The design of the network interface message level setting.
+
+History
+-------
+
+ The design of the debugging message interface was guided and
+ constrained by backwards compatibility previous practice.  It is useful
+ to understand the history and evolution in order to understand current
+ practice and relate it to older driver source code.
+
+ From the beginning of Linux, each network device driver has had a local
+ integer variable that controls the debug message level.  The message
+ level ranged from 0 to 7, and monotonically increased in verbosity.
+
+ The message level was not precisely defined past level 3, but were
+ always implemented within +-1 of the specified level.  Drivers tended
+ to shed the more verbose level messages as they matured.
+
+   - 0  Minimal messages, only essential information on fatal errors.
+   - 1  Standard messages, initialization status.  No run-time messages
+   - 2  Special media selection messages, generally timer-driver.
+   - 3  Interface starts and stops, including normal status messages
+   - 4  Tx and Rx frame error messages, and abnormal driver operation
+   - 5  Tx packet queue information, interrupt events.
+   - 6  Status on each completed Tx packet and received Rx packets
+   - 7  Initial contents of Tx and Rx packets
+
+ Initially this message level variable was uniquely named in each driver
+ e.g. "lance_debug", so that a kernel symbolic debugger could locate and
+ modify the setting.  When kernel modules became common, the variables
+ were consistently renamed to "debug" and allowed to be set as a module
+ parameter.
+
+ This approach worked well.  However there is always a demand for
+ additional features.  Over the years the following emerged as
+ reasonable and easily implemented enhancements
+
+   - Using an ioctl() call to modify the level.
+   - Per-interface rather than per-driver message level setting.
+   - More selective control over the type of messages emitted.
+
+ The netif_msg recommendation adds these features with only a minor
+ complexity and code size increase.
+
+ The recommendation is the following points
+
+  - Retaining the per-driver integer variable "debug" as a module
+    parameter with a default level of '1'.
+
+  - Adding a per-interface private variable named "msg_enable".  The
+    variable is a bit map rather than a level, and is initialized as::
+
+       1 << debug
+
+    Or more precisely::
+
+	debug < 0 ? 0 : 1 << min(sizeof(int)-1, debug)
+
+    Messages should changes from::
+
+      if (debug > 1)
+	   printk(MSG_DEBUG "%s: ...
+
+    to::
+
+      if (np->msg_enable & NETIF_MSG_LINK)
+	   printk(MSG_DEBUG "%s: ...
+
+
+The set of message levels is named
+
+
+  =========   ===================	============
+  Old level   Name			Bit position
+  =========   ===================	============
+    0         NETIF_MSG_DRV		0x0001
+    1         NETIF_MSG_PROBE		0x0002
+    2         NETIF_MSG_LINK		0x0004
+    2         NETIF_MSG_TIMER		0x0004
+    3         NETIF_MSG_IFDOWN		0x0008
+    3         NETIF_MSG_IFUP		0x0008
+    4         NETIF_MSG_RX_ERR		0x0010
+    4         NETIF_MSG_TX_ERR		0x0010
+    5         NETIF_MSG_TX_QUEUED	0x0020
+    5         NETIF_MSG_INTR		0x0020
+    6         NETIF_MSG_TX_DONE		0x0040
+    6         NETIF_MSG_RX_STATUS	0x0040
+    7         NETIF_MSG_PKTDATA		0x0080
+  =========   ===================	============