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, 392 insertions, 0 deletions

diff --git a/drivers/cpufreq/speedstep-smi.c b/drivers/cpufreq/speedstep-smi.c
new file mode 100644
index 000000000..39265884c
--- /dev/null
+++ b/drivers/cpufreq/speedstep-smi.c
@@ -0,0 +1,392 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Intel SpeedStep SMI driver.
+ *
+ * (C) 2003  Hiroshi Miura <miura@da-cha.org>
+ */
+
+
+/*********************************************************************
+ *                        SPEEDSTEP - DEFINITIONS                    *
+ *********************************************************************/
+
+#define pr_fmt(fmt) "cpufreq: " fmt
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/cpufreq.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <asm/ist.h>
+#include <asm/cpu_device_id.h>
+
+#include "speedstep-lib.h"
+
+/* speedstep system management interface port/command.
+ *
+ * These parameters are got from IST-SMI BIOS call.
+ * If user gives it, these are used.
+ *
+ */
+static int smi_port;
+static int smi_cmd;
+static unsigned int smi_sig;
+
+/* info about the processor */
+static enum speedstep_processor speedstep_processor;
+
+/*
+ * There are only two frequency states for each processor. Values
+ * are in kHz for the time being.
+ */
+static struct cpufreq_frequency_table speedstep_freqs[] = {
+	{0, SPEEDSTEP_HIGH,	0},
+	{0, SPEEDSTEP_LOW,	0},
+	{0, 0,			CPUFREQ_TABLE_END},
+};
+
+#define GET_SPEEDSTEP_OWNER 0
+#define GET_SPEEDSTEP_STATE 1
+#define SET_SPEEDSTEP_STATE 2
+#define GET_SPEEDSTEP_FREQS 4
+
+/* how often shall the SMI call be tried if it failed, e.g. because
+ * of DMA activity going on? */
+#define SMI_TRIES 5
+
+/**
+ * speedstep_smi_ownership
+ */
+static int speedstep_smi_ownership(void)
+{
+	u32 command, result, magic, dummy;
+	u32 function = GET_SPEEDSTEP_OWNER;
+	unsigned char magic_data[] = "Copyright (c) 1999 Intel Corporation";
+
+	command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff);
+	magic = virt_to_phys(magic_data);
+
+	pr_debug("trying to obtain ownership with command %x at port %x\n",
+			command, smi_port);
+
+	__asm__ __volatile__(
+		"push %%ebp\n"
+		"out %%al, (%%dx)\n"
+		"pop %%ebp\n"
+		: "=D" (result),
+		  "=a" (dummy), "=b" (dummy), "=c" (dummy), "=d" (dummy),
+		  "=S" (dummy)
+		: "a" (command), "b" (function), "c" (0), "d" (smi_port),
+		  "D" (0), "S" (magic)
+		: "memory"
+	);
+
+	pr_debug("result is %x\n", result);
+
+	return result;
+}
+
+/**
+ * speedstep_smi_get_freqs - get SpeedStep preferred & current freq.
+ * @low: the low frequency value is placed here
+ * @high: the high frequency value is placed here
+ *
+ * Only available on later SpeedStep-enabled systems, returns false results or
+ * even hangs [cf. bugme.osdl.org # 1422] on earlier systems. Empirical testing
+ * shows that the latter occurs if !(ist_info.event & 0xFFFF).
+ */
+static int speedstep_smi_get_freqs(unsigned int *low, unsigned int *high)
+{
+	u32 command, result = 0, edi, high_mhz, low_mhz, dummy;
+	u32 state = 0;
+	u32 function = GET_SPEEDSTEP_FREQS;
+
+	if (!(ist_info.event & 0xFFFF)) {
+		pr_debug("bug #1422 -- can't read freqs from BIOS\n");
+		return -ENODEV;
+	}
+
+	command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff);
+
+	pr_debug("trying to determine frequencies with command %x at port %x\n",
+			command, smi_port);
+
+	__asm__ __volatile__(
+		"push %%ebp\n"
+		"out %%al, (%%dx)\n"
+		"pop %%ebp"
+		: "=a" (result),
+		  "=b" (high_mhz),
+		  "=c" (low_mhz),
+		  "=d" (state), "=D" (edi), "=S" (dummy)
+		: "a" (command),
+		  "b" (function),
+		  "c" (state),
+		  "d" (smi_port), "S" (0), "D" (0)
+	);
+
+	pr_debug("result %x, low_freq %u, high_freq %u\n",
+			result, low_mhz, high_mhz);
+
+	/* abort if results are obviously incorrect... */
+	if ((high_mhz + low_mhz) < 600)
+		return -EINVAL;
+
+	*high = high_mhz * 1000;
+	*low  = low_mhz  * 1000;
+
+	return result;
+}
+
+/**
+ * speedstep_set_state - set the SpeedStep state
+ * @state: new processor frequency state (SPEEDSTEP_LOW or SPEEDSTEP_HIGH)
+ *
+ */
+static void speedstep_set_state(unsigned int state)
+{
+	unsigned int result = 0, command, new_state, dummy;
+	unsigned long flags;
+	unsigned int function = SET_SPEEDSTEP_STATE;
+	unsigned int retry = 0;
+
+	if (state > 0x1)
+		return;
+
+	/* Disable IRQs */
+	preempt_disable();
+	local_irq_save(flags);
+
+	command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff);
+
+	pr_debug("trying to set frequency to state %u "
+		"with command %x at port %x\n",
+		state, command, smi_port);
+
+	do {
+		if (retry) {
+			/*
+			 * We need to enable interrupts, otherwise the blockage
+			 * won't resolve.
+			 *
+			 * We disable preemption so that other processes don't
+			 * run. If other processes were running, they could
+			 * submit more DMA requests, making the blockage worse.
+			 */
+			pr_debug("retry %u, previous result %u, waiting...\n",
+					retry, result);
+			local_irq_enable();
+			mdelay(retry * 50);
+			local_irq_disable();
+		}
+		retry++;
+		__asm__ __volatile__(
+			"push %%ebp\n"
+			"out %%al, (%%dx)\n"
+			"pop %%ebp"
+			: "=b" (new_state), "=D" (result),
+			  "=c" (dummy), "=a" (dummy),
+			  "=d" (dummy), "=S" (dummy)
+			: "a" (command), "b" (function), "c" (state),
+			  "d" (smi_port), "S" (0), "D" (0)
+			);
+	} while ((new_state != state) && (retry <= SMI_TRIES));
+
+	/* enable IRQs */
+	local_irq_restore(flags);
+	preempt_enable();
+
+	if (new_state == state)
+		pr_debug("change to %u MHz succeeded after %u tries "
+			"with result %u\n",
+			(speedstep_freqs[new_state].frequency / 1000),
+			retry, result);
+	else
+		pr_err("change to state %u failed with new_state %u and result %u\n",
+		       state, new_state, result);
+
+	return;
+}
+
+
+/**
+ * speedstep_target - set a new CPUFreq policy
+ * @policy: new policy
+ * @index: index of new freq
+ *
+ * Sets a new CPUFreq policy/freq.
+ */
+static int speedstep_target(struct cpufreq_policy *policy, unsigned int index)
+{
+	speedstep_set_state(index);
+
+	return 0;
+}
+
+
+static int speedstep_cpu_init(struct cpufreq_policy *policy)
+{
+	int result;
+	unsigned int *low, *high;
+
+	/* capability check */
+	if (policy->cpu != 0)
+		return -ENODEV;
+
+	result = speedstep_smi_ownership();
+	if (result) {
+		pr_debug("fails in acquiring ownership of a SMI interface.\n");
+		return -EINVAL;
+	}
+
+	/* detect low and high frequency */
+	low = &speedstep_freqs[SPEEDSTEP_LOW].frequency;
+	high = &speedstep_freqs[SPEEDSTEP_HIGH].frequency;
+
+	result = speedstep_smi_get_freqs(low, high);
+	if (result) {
+		/* fall back to speedstep_lib.c dection mechanism:
+		 * try both states out */
+		pr_debug("could not detect low and high frequencies "
+				"by SMI call.\n");
+		result = speedstep_get_freqs(speedstep_processor,
+				low, high,
+				NULL,
+				&speedstep_set_state);
+
+		if (result) {
+			pr_debug("could not detect two different speeds"
+					" -- aborting.\n");
+			return result;
+		} else
+			pr_debug("workaround worked.\n");
+	}
+
+	policy->freq_table = speedstep_freqs;
+
+	return 0;
+}
+
+static unsigned int speedstep_get(unsigned int cpu)
+{
+	if (cpu)
+		return -ENODEV;
+	return speedstep_get_frequency(speedstep_processor);
+}
+
+
+static int speedstep_resume(struct cpufreq_policy *policy)
+{
+	int result = speedstep_smi_ownership();
+
+	if (result)
+		pr_debug("fails in re-acquiring ownership of a SMI interface.\n");
+
+	return result;
+}
+
+static struct cpufreq_driver speedstep_driver = {
+	.name		= "speedstep-smi",
+	.flags		= CPUFREQ_NO_AUTO_DYNAMIC_SWITCHING,
+	.verify		= cpufreq_generic_frequency_table_verify,
+	.target_index	= speedstep_target,
+	.init		= speedstep_cpu_init,
+	.get		= speedstep_get,
+	.resume		= speedstep_resume,
+};
+
+static const struct x86_cpu_id ss_smi_ids[] = {
+	X86_MATCH_VENDOR_FAM_MODEL(INTEL,  6, 0x8, 0),
+	X86_MATCH_VENDOR_FAM_MODEL(INTEL,  6, 0xb, 0),
+	X86_MATCH_VENDOR_FAM_MODEL(INTEL, 15, 0x2, 0),
+	{}
+};
+
+/**
+ * speedstep_init - initializes the SpeedStep CPUFreq driver
+ *
+ *   Initializes the SpeedStep support. Returns -ENODEV on unsupported
+ * BIOS, -EINVAL on problems during initiatization, and zero on
+ * success.
+ */
+static int __init speedstep_init(void)
+{
+	if (!x86_match_cpu(ss_smi_ids))
+		return -ENODEV;
+
+	speedstep_processor = speedstep_detect_processor();
+
+	switch (speedstep_processor) {
+	case SPEEDSTEP_CPU_PIII_T:
+	case SPEEDSTEP_CPU_PIII_C:
+	case SPEEDSTEP_CPU_PIII_C_EARLY:
+		break;
+	default:
+		speedstep_processor = 0;
+	}
+
+	if (!speedstep_processor) {
+		pr_debug("No supported Intel CPU detected.\n");
+		return -ENODEV;
+	}
+
+	pr_debug("signature:0x%.8x, command:0x%.8x, "
+		"event:0x%.8x, perf_level:0x%.8x.\n",
+		ist_info.signature, ist_info.command,
+		ist_info.event, ist_info.perf_level);
+
+	/* Error if no IST-SMI BIOS or no PARM
+		 sig= 'ISGE' aka 'Intel Speedstep Gate E' */
+	if ((ist_info.signature !=  0x47534943) && (
+	    (smi_port == 0) || (smi_cmd == 0)))
+		return -ENODEV;
+
+	if (smi_sig == 1)
+		smi_sig = 0x47534943;
+	else
+		smi_sig = ist_info.signature;
+
+	/* setup smi_port from MODLULE_PARM or BIOS */
+	if ((smi_port > 0xff) || (smi_port < 0))
+		return -EINVAL;
+	else if (smi_port == 0)
+		smi_port = ist_info.command & 0xff;
+
+	if ((smi_cmd > 0xff) || (smi_cmd < 0))
+		return -EINVAL;
+	else if (smi_cmd == 0)
+		smi_cmd = (ist_info.command >> 16) & 0xff;
+
+	return cpufreq_register_driver(&speedstep_driver);
+}
+
+
+/**
+ * speedstep_exit - unregisters SpeedStep support
+ *
+ *   Unregisters SpeedStep support.
+ */
+static void __exit speedstep_exit(void)
+{
+	cpufreq_unregister_driver(&speedstep_driver);
+}
+
+module_param_hw(smi_port, int, ioport, 0444);
+module_param(smi_cmd,  int, 0444);
+module_param(smi_sig, uint, 0444);
+
+MODULE_PARM_DESC(smi_port, "Override the BIOS-given IST port with this value "
+		"-- Intel's default setting is 0xb2");
+MODULE_PARM_DESC(smi_cmd, "Override the BIOS-given IST command with this value "
+		"-- Intel's default setting is 0x82");
+MODULE_PARM_DESC(smi_sig, "Set to 1 to fake the IST signature when using the "
+		"SMI interface.");
+
+MODULE_AUTHOR("Hiroshi Miura");
+MODULE_DESCRIPTION("Speedstep driver for IST applet SMI interface.");
+MODULE_LICENSE("GPL");
+
+module_init(speedstep_init);
+module_exit(speedstep_exit);