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, 89 insertions, 0 deletions

diff --git a/drivers/macintosh/windfarm_pid.h b/drivers/macintosh/windfarm_pid.h
new file mode 100644
index 000000000..335613a20
--- /dev/null
+++ b/drivers/macintosh/windfarm_pid.h
@@ -0,0 +1,89 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+#ifndef _WINDFARM_PID_H
+#define _WINDFARM_PID_H
+
+/*
+ * Windfarm PowerMac thermal control. Generic PID helpers
+ *
+ * (c) Copyright 2005 Benjamin Herrenschmidt, IBM Corp.
+ *                    <benh@kernel.crashing.org>
+ *
+ * This is a pair of generic PID helpers that can be used by
+ * control loops. One is the basic PID implementation, the
+ * other one is more specifically tailored to the loops used
+ * for CPU control with 2 input sample types (temp and power)
+ */
+
+/*
+ * *** Simple PID ***
+ */
+
+#define WF_PID_MAX_HISTORY	32
+
+/* This parameter array is passed to the PID algorithm. Currently,
+ * we don't support changing parameters on the fly as it's not needed
+ * but could be implemented (with necessary adjustment of the history
+ * buffer
+ */
+struct wf_pid_param {
+	int	interval;	/* Interval between samples in seconds */
+	int	history_len;	/* Size of history buffer */
+	int	additive;	/* 1: target relative to previous value */
+	s32	gd, gp, gr;	/* PID gains */
+	s32	itarget;	/* PID input target */
+	s32	min,max;	/* min and max target values */
+};
+
+struct wf_pid_state {
+	int	first;				/* first run of the loop */
+	int	index; 				/* index of current sample */
+	s32	target;				/* current target value */
+	s32	samples[WF_PID_MAX_HISTORY];	/* samples history buffer */
+	s32	errors[WF_PID_MAX_HISTORY];	/* error history buffer */
+
+	struct wf_pid_param param;
+};
+
+extern void wf_pid_init(struct wf_pid_state *st, struct wf_pid_param *param);
+extern s32 wf_pid_run(struct wf_pid_state *st, s32 sample);
+
+
+/*
+ * *** CPU PID ***
+ */
+
+#define WF_CPU_PID_MAX_HISTORY	32
+
+/* This parameter array is passed to the CPU PID algorithm. Currently,
+ * we don't support changing parameters on the fly as it's not needed
+ * but could be implemented (with necessary adjustment of the history
+ * buffer
+ */
+struct wf_cpu_pid_param {
+	int	interval;	/* Interval between samples in seconds */
+	int	history_len;	/* Size of history buffer */
+	s32	gd, gp, gr;	/* PID gains */
+	s32	pmaxadj;	/* PID max power adjust */
+	s32	ttarget;	/* PID input target */
+	s32	tmax;		/* PID input max */
+	s32	min,max;	/* min and max target values */
+};
+
+struct wf_cpu_pid_state {
+	int	first;				/* first run of the loop */
+	int	index; 				/* index of current power */
+	int	tindex; 			/* index of current temp */
+	s32	target;				/* current target value */
+	s32	last_delta;			/* last Tactual - Ttarget */
+	s32	powers[WF_PID_MAX_HISTORY];	/* power history buffer */
+	s32	errors[WF_PID_MAX_HISTORY];	/* error history buffer */
+	s32	temps[2];			/* temp. history buffer */
+
+	struct wf_cpu_pid_param param;
+};
+
+extern void wf_cpu_pid_init(struct wf_cpu_pid_state *st,
+			    struct wf_cpu_pid_param *param);
+extern s32 wf_cpu_pid_run(struct wf_cpu_pid_state *st, s32 power, s32 temp);
+
+#endif /* _WINDFARM_PID_H */