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, 76 insertions, 0 deletions

diff --git a/Documentation/ABI/testing/sysfs-class-net-qmi b/Documentation/ABI/testing/sysfs-class-net-qmi
new file mode 100644
index 000000000..b028f5bc8
--- /dev/null
+++ b/Documentation/ABI/testing/sysfs-class-net-qmi
@@ -0,0 +1,76 @@
+What:		/sys/class/net/<iface>/qmi/raw_ip
+Date:		Dec 2015
+KernelVersion:	4.4
+Contact:	Bjørn Mork <bjorn@mork.no>
+Description:
+		Boolean.  Default: 'N'
+
+		Set this to 'Y' to change the network device link
+		framing from '802.3' to 'raw-ip'.
+
+		The netdev will change to reflect the link framing
+		mode.  The netdev is an ordinary ethernet device in
+		'802.3' mode, and the driver expects to exchange
+		frames with an ethernet header over the USB link. The
+		netdev is a headerless p-t-p device in 'raw-ip' mode,
+		and the driver expects to echange IPv4 or IPv6 packets
+		without any L2 header over the USB link.
+
+		Userspace is in full control of firmware configuration
+		through the delegation of the QMI protocol. Userspace
+		is responsible for coordination of driver and firmware
+		link framing mode, changing this setting to 'Y' if the
+		firmware is configured for 'raw-ip' mode.
+
+What:		/sys/class/net/<iface>/qmi/add_mux
+Date:		March 2017
+KernelVersion:	4.11
+Contact:	Bjørn Mork <bjorn@mork.no>
+Description:
+		Unsigned integer.
+
+		Write a number ranging from 1 to 254 to add a qmap mux
+		based network device, supported by recent Qualcomm based
+		modems.
+
+		The network device will be called qmimux.
+
+		Userspace is in charge of managing the qmux network device
+		activation and data stream setup on the modem side by
+		using the proper QMI protocol requests.
+
+What:		/sys/class/net/<iface>/qmi/del_mux
+Date:		March 2017
+KernelVersion:	4.11
+Contact:	Bjørn Mork <bjorn@mork.no>
+Description:
+		Unsigned integer.
+
+		Write a number ranging from 1 to 254 to delete a previously
+		created qmap mux based network device.
+
+What:		/sys/class/net/<qmimux iface>/qmap/mux_id
+Date:		January 2021
+KernelVersion:	5.12
+Contact:	Daniele Palmas <dnlplm@gmail.com>
+Description:
+		Unsigned integer
+
+		Indicates the mux id associated to the qmimux network interface
+		during its creation.
+
+What:		/sys/class/net/<iface>/qmi/pass_through
+Date:		January 2021
+KernelVersion:	5.12
+Contact:	Subash Abhinov Kasiviswanathan <quic_subashab@quicinc.com>
+Description:
+		Boolean.  Default: 'N'
+
+		Set this to 'Y' to enable 'pass-through' mode, allowing packets
+		in MAP format to be passed on to the stack.
+
+		Normally the rmnet driver (CONFIG_RMNET) is then used to process
+		and demultiplex these packets.
+
+		'Pass-through' mode can be enabled when the device is in
+		'raw-ip' mode only.