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

3 files changed, 131 insertions, 0 deletions

diff --git a/samples/pfsm/.gitignore b/samples/pfsm/.gitignore
new file mode 100644
index 000000000..f350a030a
--- /dev/null
+++ b/samples/pfsm/.gitignore
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0
+/pfsm-wakeup
diff --git a/samples/pfsm/Makefile b/samples/pfsm/Makefile
new file mode 100644
index 000000000..213e8d9f5
--- /dev/null
+++ b/samples/pfsm/Makefile
@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: GPL-2.0
+userprogs-always-y += pfsm-wakeup
+
+userccflags += -I usr/include
diff --git a/samples/pfsm/pfsm-wakeup.c b/samples/pfsm/pfsm-wakeup.c
new file mode 100644
index 000000000..299dd9e1f
--- /dev/null
+++ b/samples/pfsm/pfsm-wakeup.c
@@ -0,0 +1,125 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * TPS6594 PFSM userspace example
+ *
+ * Copyright (C) 2023 BayLibre Incorporated - https://www.baylibre.com/
+ *
+ * This example shows how to use PFSMs from a userspace application,
+ * on TI j721s2 platform. The PMIC is armed to be triggered by a RTC
+ * alarm to execute state transition (RETENTION to ACTIVE).
+ */
+
+#include <fcntl.h>
+#include <stdio.h>
+#include <sys/ioctl.h>
+#include <unistd.h>
+
+#include <linux/rtc.h>
+#include <linux/tps6594_pfsm.h>
+
+#define ALARM_DELTA_SEC 30
+
+#define RTC_A "/dev/rtc0"
+
+#define PMIC_NB 3
+#define PMIC_A "/dev/pfsm-0-0x48"
+#define PMIC_B "/dev/pfsm-0-0x4c"
+#define PMIC_C "/dev/pfsm-2-0x58"
+
+static const char * const dev_pfsm[] = {PMIC_A, PMIC_B, PMIC_C};
+
+int main(int argc, char *argv[])
+{
+	int i, ret, fd_rtc, fd_pfsm[PMIC_NB] = { 0 };
+	struct rtc_time rtc_tm;
+	struct pmic_state_opt pmic_opt = { 0 };
+	unsigned long data;
+
+	fd_rtc = open(RTC_A, O_RDONLY);
+	if (fd_rtc < 0) {
+		perror("Failed to open RTC device.");
+		goto out;
+	}
+
+	for (i = 0 ; i < PMIC_NB ; i++) {
+		fd_pfsm[i] = open(dev_pfsm[i], O_RDWR);
+		if (fd_pfsm[i] < 0) {
+			perror("Failed to open PFSM device.");
+			goto out;
+		}
+	}
+
+	/* Read RTC date/time */
+	ret = ioctl(fd_rtc, RTC_RD_TIME, &rtc_tm);
+	if (ret < 0) {
+		perror("Failed to read RTC date/time.");
+		goto out;
+	}
+	printf("Current RTC date/time is %d-%d-%d, %02d:%02d:%02d.\n",
+	       rtc_tm.tm_mday, rtc_tm.tm_mon + 1, rtc_tm.tm_year + 1900,
+	       rtc_tm.tm_hour, rtc_tm.tm_min, rtc_tm.tm_sec);
+
+	/* Set RTC alarm to ALARM_DELTA_SEC sec in the future, and check for rollover */
+	rtc_tm.tm_sec += ALARM_DELTA_SEC;
+	if (rtc_tm.tm_sec >= 60) {
+		rtc_tm.tm_sec %= 60;
+		rtc_tm.tm_min++;
+	}
+	if (rtc_tm.tm_min == 60) {
+		rtc_tm.tm_min = 0;
+		rtc_tm.tm_hour++;
+	}
+	if (rtc_tm.tm_hour == 24)
+		rtc_tm.tm_hour = 0;
+	ret = ioctl(fd_rtc, RTC_ALM_SET, &rtc_tm);
+	if (ret < 0) {
+		perror("Failed to set RTC alarm.");
+		goto out;
+	}
+
+	/* Enable alarm interrupts */
+	ret = ioctl(fd_rtc, RTC_AIE_ON, 0);
+	if (ret < 0) {
+		perror("Failed to enable alarm interrupts.");
+		goto out;
+	}
+	printf("Waiting %d seconds for alarm...\n", ALARM_DELTA_SEC);
+
+	/*
+	 * Set RETENTION state with options for PMIC_C/B/A respectively.
+	 * Since PMIC_A is master, it should be the last one to be configured.
+	 */
+	pmic_opt.ddr_retention = 1;
+	for (i = PMIC_NB - 1 ; i >= 0 ; i--) {
+		printf("Set RETENTION state for PMIC_%d.\n", i);
+		sleep(1);
+		ret = ioctl(fd_pfsm[i], PMIC_SET_RETENTION_STATE, &pmic_opt);
+		if (ret < 0) {
+			perror("Failed to set RETENTION state.");
+			goto out_reset;
+		}
+	}
+
+	/* This blocks until the alarm ring causes an interrupt */
+	ret = read(fd_rtc, &data, sizeof(unsigned long));
+	if (ret < 0)
+		perror("Failed to get RTC alarm.");
+	else
+		puts("Alarm rang.\n");
+
+out_reset:
+	ioctl(fd_rtc, RTC_AIE_OFF, 0);
+
+	/* Set ACTIVE state for PMIC_A */
+	ioctl(fd_pfsm[0], PMIC_SET_ACTIVE_STATE, 0);
+
+out:
+	for (i = 0 ; i < PMIC_NB ; i++)
+		if (fd_pfsm[i])
+			close(fd_pfsm[i]);
+
+	if (fd_rtc)
+		close(fd_rtc);
+
+	return 0;
+}