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, 43 insertions, 0 deletions

diff --git a/Documentation/input/devices/cs461x.rst b/Documentation/input/devices/cs461x.rst
new file mode 100644
index 000000000..b1e6d508a
--- /dev/null
+++ b/Documentation/input/devices/cs461x.rst
@@ -0,0 +1,43 @@
+Crystal SoundFusion CS4610/CS4612/CS461 joystick
+================================================
+
+This is a new low-level driver to support analog joystick attached to
+Crystal SoundFusion CS4610/CS4612/CS4615. This code is based upon
+Vortex/Solo drivers as an example of decoration style, and ALSA
+0.5.8a kernel drivers as an chipset documentation and samples.
+
+This version does not have cooked mode support; the basic code
+is present here, but have not tested completely. The button analysis
+is completed in this mode, but the axis movement is not.
+
+Raw mode works fine with analog joystick front-end driver and cs461x
+driver as a backend. I've tested this driver with CS4610, 4-axis and
+4-button joystick; I mean the jstest utility. Also I've tried to
+play in xracer game using joystick, and the result is better than
+keyboard only mode.
+
+The sensitivity and calibrate quality have not been tested; the two
+reasons are performed: the same hardware cannot work under Win95 (blue
+screen in VJOYD); I have no documentation on my chip; and the existing
+behavior in my case was not raised the requirement of joystick calibration.
+So the driver have no code to perform hardware related calibration.
+
+This driver have the basic support for PCI devices only; there is no
+ISA or PnP ISA cards supported.
+
+The driver works with ALSA drivers simultaneously. For example, the xracer
+uses joystick as input device and PCM device as sound output in one time.
+There are no sound or input collisions detected. The source code have
+comments about them; but I've found the joystick can be initialized
+separately of ALSA modules. So, you can use only one joystick driver
+without ALSA drivers. The ALSA drivers are not needed to compile or
+run this driver.
+
+There are no debug information print have been placed in source, and no
+specific options required to work this driver. The found chipset parameters
+are printed via printk(KERN_INFO "..."), see the /var/log/messages to
+inspect cs461x: prefixed messages to determine possible card detection
+errors.
+
+Regards,
+Viktor