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, 54 insertions, 0 deletions

diff --git a/Documentation/driver-api/fpga/intro.rst b/Documentation/driver-api/fpga/intro.rst
new file mode 100644
index 000000000..f54c7dabc
--- /dev/null
+++ b/Documentation/driver-api/fpga/intro.rst
@@ -0,0 +1,54 @@
+Introduction
+============
+
+The FPGA subsystem supports reprogramming FPGAs dynamically under
+Linux.  Some of the core intentions of the FPGA subsystems are:
+
+* The FPGA subsystem is vendor agnostic.
+
+* The FPGA subsystem separates upper layers (userspace interfaces and
+  enumeration) from lower layers that know how to program a specific
+  FPGA.
+
+* Code should not be shared between upper and lower layers.  This
+  should go without saying.  If that seems necessary, there's probably
+  framework functionality that can be added that will benefit
+  other users.  Write the linux-fpga mailing list and maintainers and
+  seek out a solution that expands the framework for broad reuse.
+
+* Generally, when adding code, think of the future.  Plan for reuse.
+
+The framework in the kernel is divided into:
+
+FPGA Manager
+------------
+
+If you are adding a new FPGA or a new method of programming an FPGA,
+this is the subsystem for you.  Low level FPGA manager drivers contain
+the knowledge of how to program a specific device.  This subsystem
+includes the framework in fpga-mgr.c and the low level drivers that
+are registered with it.
+
+FPGA Bridge
+-----------
+
+FPGA Bridges prevent spurious signals from going out of an FPGA or a
+region of an FPGA during programming.  They are disabled before
+programming begins and re-enabled afterwards.  An FPGA bridge may be
+actual hard hardware that gates a bus to a CPU or a soft ("freeze")
+bridge in FPGA fabric that surrounds a partial reconfiguration region
+of an FPGA.  This subsystem includes fpga-bridge.c and the low level
+drivers that are registered with it.
+
+FPGA Region
+-----------
+
+If you are adding a new interface to the FPGA framework, add it on top
+of an FPGA region.
+
+The FPGA Region framework (fpga-region.c) associates managers and
+bridges as reconfigurable regions.  A region may refer to the whole
+FPGA in full reconfiguration or to a partial reconfiguration region.
+
+The Device Tree FPGA Region support (of-fpga-region.c) handles
+reprogramming FPGAs when device tree overlays are applied.