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, 335 insertions, 0 deletions

diff --git a/arch/arm64/hyperv/Makefile b/arch/arm64/hyperv/Makefile
new file mode 100644
index 000000000..87c31c001
--- /dev/null
+++ b/arch/arm64/hyperv/Makefile
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-y		:= hv_core.o mshyperv.o
diff --git a/arch/arm64/hyperv/hv_core.c b/arch/arm64/hyperv/hv_core.c
new file mode 100644
index 000000000..e33a9e3c3
--- /dev/null
+++ b/arch/arm64/hyperv/hv_core.c
@@ -0,0 +1,197 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Low level utility routines for interacting with Hyper-V.
+ *
+ * Copyright (C) 2021, Microsoft, Inc.
+ *
+ * Author : Michael Kelley <mikelley@microsoft.com>
+ */
+
+#include <linux/types.h>
+#include <linux/export.h>
+#include <linux/mm.h>
+#include <linux/arm-smccc.h>
+#include <linux/module.h>
+#include <asm-generic/bug.h>
+#include <hyperv/hvhdk.h>
+#include <asm/mshyperv.h>
+
+/*
+ * hv_do_hypercall- Invoke the specified hypercall
+ */
+u64 hv_do_hypercall(u64 control, void *input, void *output)
+{
+	struct arm_smccc_res	res;
+	u64			input_address;
+	u64			output_address;
+
+	input_address = input ? virt_to_phys(input) : 0;
+	output_address = output ? virt_to_phys(output) : 0;
+
+	arm_smccc_1_1_hvc(HV_FUNC_ID, control,
+			  input_address, output_address, &res);
+	return res.a0;
+}
+EXPORT_SYMBOL_GPL(hv_do_hypercall);
+
+/*
+ * hv_do_fast_hypercall8 -- Invoke the specified hypercall
+ * with arguments in registers instead of physical memory.
+ * Avoids the overhead of virt_to_phys for simple hypercalls.
+ */
+
+u64 hv_do_fast_hypercall8(u16 code, u64 input)
+{
+	struct arm_smccc_res	res;
+	u64			control;
+
+	control = (u64)code | HV_HYPERCALL_FAST_BIT;
+
+	arm_smccc_1_1_hvc(HV_FUNC_ID, control, input, &res);
+	return res.a0;
+}
+EXPORT_SYMBOL_GPL(hv_do_fast_hypercall8);
+
+/*
+ * hv_do_fast_hypercall16 -- Invoke the specified hypercall
+ * with arguments in registers instead of physical memory.
+ * Avoids the overhead of virt_to_phys for simple hypercalls.
+ */
+u64 hv_do_fast_hypercall16(u16 code, u64 input1, u64 input2)
+{
+	struct arm_smccc_res	res;
+	u64			control;
+
+	control = (u64)code | HV_HYPERCALL_FAST_BIT;
+
+	arm_smccc_1_1_hvc(HV_FUNC_ID, control, input1, input2, &res);
+	return res.a0;
+}
+EXPORT_SYMBOL_GPL(hv_do_fast_hypercall16);
+
+/*
+ * Set a single VP register to a 64-bit value.
+ */
+void hv_set_vpreg(u32 msr, u64 value)
+{
+	struct arm_smccc_res res;
+
+	arm_smccc_1_1_hvc(HV_FUNC_ID,
+		HVCALL_SET_VP_REGISTERS | HV_HYPERCALL_FAST_BIT |
+			HV_HYPERCALL_REP_COMP_1,
+		HV_PARTITION_ID_SELF,
+		HV_VP_INDEX_SELF,
+		msr,
+		0,
+		value,
+		0,
+		&res);
+
+	/*
+	 * Something is fundamentally broken in the hypervisor if
+	 * setting a VP register fails. There's really no way to
+	 * continue as a guest VM, so panic.
+	 */
+	BUG_ON(!hv_result_success(res.a0));
+}
+EXPORT_SYMBOL_GPL(hv_set_vpreg);
+
+/*
+ * Get the value of a single VP register.  One version
+ * returns just 64 bits and another returns the full 128 bits.
+ * The two versions are separate to avoid complicating the
+ * calling sequence for the more frequently used 64 bit version.
+ */
+
+void hv_get_vpreg_128(u32 msr, struct hv_get_vp_registers_output *result)
+{
+	struct arm_smccc_1_2_regs args;
+	struct arm_smccc_1_2_regs res;
+
+	args.a0 = HV_FUNC_ID;
+	args.a1 = HVCALL_GET_VP_REGISTERS | HV_HYPERCALL_FAST_BIT |
+			HV_HYPERCALL_REP_COMP_1;
+	args.a2 = HV_PARTITION_ID_SELF;
+	args.a3 = HV_VP_INDEX_SELF;
+	args.a4 = msr;
+
+	/*
+	 * Use the SMCCC 1.2 interface because the results are in registers
+	 * beyond X0-X3.
+	 */
+	arm_smccc_1_2_hvc(&args, &res);
+
+	/*
+	 * Something is fundamentally broken in the hypervisor if
+	 * getting a VP register fails. There's really no way to
+	 * continue as a guest VM, so panic.
+	 */
+	BUG_ON(!hv_result_success(res.a0));
+
+	result->as64.low = res.a6;
+	result->as64.high = res.a7;
+}
+EXPORT_SYMBOL_GPL(hv_get_vpreg_128);
+
+u64 hv_get_vpreg(u32 msr)
+{
+	struct hv_get_vp_registers_output output;
+
+	hv_get_vpreg_128(msr, &output);
+
+	return output.as64.low;
+}
+EXPORT_SYMBOL_GPL(hv_get_vpreg);
+
+/*
+ * hyperv_report_panic - report a panic to Hyper-V.  This function uses
+ * the older version of the Hyper-V interface that admittedly doesn't
+ * pass enough information to be useful beyond just recording the
+ * occurrence of a panic. The parallel hv_kmsg_dump() uses the
+ * new interface that allows reporting 4 Kbytes of data, which is much
+ * more useful. Hyper-V on ARM64 always supports the newer interface, but
+ * we retain support for the older version because the sysadmin is allowed
+ * to disable the newer version via sysctl in case of information security
+ * concerns about the more verbose version.
+ */
+void hyperv_report_panic(struct pt_regs *regs, long err, bool in_die)
+{
+	static bool	panic_reported;
+	u64		guest_id;
+
+	/* Don't report a panic to Hyper-V if we're not going to panic */
+	if (in_die && !panic_on_oops)
+		return;
+
+	/*
+	 * We prefer to report panic on 'die' chain as we have proper
+	 * registers to report, but if we miss it (e.g. on BUG()) we need
+	 * to report it on 'panic'.
+	 *
+	 * Calling code in the 'die' and 'panic' paths ensures that only
+	 * one CPU is running this code, so no atomicity is needed.
+	 */
+	if (panic_reported)
+		return;
+	panic_reported = true;
+
+	guest_id = hv_get_vpreg(HV_REGISTER_GUEST_OS_ID);
+
+	/*
+	 * Hyper-V provides the ability to store only 5 values.
+	 * Pick the passed in error value, the guest_id, the PC,
+	 * and the SP.
+	 */
+	hv_set_vpreg(HV_REGISTER_GUEST_CRASH_P0, err);
+	hv_set_vpreg(HV_REGISTER_GUEST_CRASH_P1, guest_id);
+	hv_set_vpreg(HV_REGISTER_GUEST_CRASH_P2, regs->pc);
+	hv_set_vpreg(HV_REGISTER_GUEST_CRASH_P3, regs->sp);
+	hv_set_vpreg(HV_REGISTER_GUEST_CRASH_P4, 0);
+
+	/*
+	 * Let Hyper-V know there is crash data available
+	 */
+	hv_set_vpreg(HV_REGISTER_GUEST_CRASH_CTL, HV_CRASH_CTL_CRASH_NOTIFY);
+}
+EXPORT_SYMBOL_GPL(hyperv_report_panic);
diff --git a/arch/arm64/hyperv/mshyperv.c b/arch/arm64/hyperv/mshyperv.c
new file mode 100644
index 000000000..4fdc26ade
--- /dev/null
+++ b/arch/arm64/hyperv/mshyperv.c
@@ -0,0 +1,136 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Core routines for interacting with Microsoft's Hyper-V hypervisor,
+ * including hypervisor initialization.
+ *
+ * Copyright (C) 2021, Microsoft, Inc.
+ *
+ * Author : Michael Kelley <mikelley@microsoft.com>
+ */
+
+#include <linux/types.h>
+#include <linux/acpi.h>
+#include <linux/export.h>
+#include <linux/errno.h>
+#include <linux/version.h>
+#include <linux/cpuhotplug.h>
+#include <asm/mshyperv.h>
+
+static bool hyperv_initialized;
+
+int hv_get_hypervisor_version(union hv_hypervisor_version_info *info)
+{
+	hv_get_vpreg_128(HV_REGISTER_HYPERVISOR_VERSION,
+			 (struct hv_get_vp_registers_output *)info);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(hv_get_hypervisor_version);
+
+#ifdef CONFIG_ACPI
+
+static bool __init hyperv_detect_via_acpi(void)
+{
+	if (acpi_disabled)
+		return false;
+	/*
+	 * Hypervisor ID is only available in ACPI v6+, and the
+	 * structure layout was extended in v6 to accommodate that
+	 * new field.
+	 *
+	 * At the very minimum, this check makes sure not to read
+	 * past the FADT structure.
+	 *
+	 * It is also needed to catch running in some unknown
+	 * non-Hyper-V environment that has ACPI 5.x or less.
+	 * In such a case, it can't be Hyper-V.
+	 */
+	if (acpi_gbl_FADT.header.revision < 6)
+		return false;
+	return strncmp((char *)&acpi_gbl_FADT.hypervisor_id, "MsHyperV", 8) == 0;
+}
+
+#else
+
+static bool __init hyperv_detect_via_acpi(void)
+{
+	return false;
+}
+
+#endif
+
+static bool __init hyperv_detect_via_smccc(void)
+{
+	uuid_t hyperv_uuid = UUID_INIT(
+		0x58ba324d, 0x6447, 0x24cd,
+		0x75, 0x6c, 0xef, 0x8e,
+		0x24, 0x70, 0x59, 0x16);
+
+	return arm_smccc_hypervisor_has_uuid(&hyperv_uuid);
+}
+
+static int __init hyperv_init(void)
+{
+	struct hv_get_vp_registers_output	result;
+	u64	guest_id;
+	int	ret;
+
+	/*
+	 * Allow for a kernel built with CONFIG_HYPERV to be running in
+	 * a non-Hyper-V environment.
+	 *
+	 * In such cases, do nothing and return success.
+	 */
+	if (!hyperv_detect_via_acpi() && !hyperv_detect_via_smccc())
+		return 0;
+
+	/* Setup the guest ID */
+	guest_id = hv_generate_guest_id(LINUX_VERSION_CODE);
+	hv_set_vpreg(HV_REGISTER_GUEST_OS_ID, guest_id);
+
+	/* Get the features and hints from Hyper-V */
+	hv_get_vpreg_128(HV_REGISTER_PRIVILEGES_AND_FEATURES_INFO, &result);
+	ms_hyperv.features = result.as32.a;
+	ms_hyperv.priv_high = result.as32.b;
+	ms_hyperv.misc_features = result.as32.c;
+
+	hv_get_vpreg_128(HV_REGISTER_FEATURES_INFO, &result);
+	ms_hyperv.hints = result.as32.a;
+
+	pr_info("Hyper-V: privilege flags low 0x%x, high 0x%x, hints 0x%x, misc 0x%x\n",
+		ms_hyperv.features, ms_hyperv.priv_high, ms_hyperv.hints,
+		ms_hyperv.misc_features);
+
+	hv_identify_partition_type();
+
+	ret = hv_common_init();
+	if (ret)
+		return ret;
+
+	ret = cpuhp_setup_state(CPUHP_AP_HYPERV_ONLINE, "arm64/hyperv_init:online",
+				hv_common_cpu_init, hv_common_cpu_die);
+	if (ret < 0) {
+		hv_common_free();
+		return ret;
+	}
+
+	if (ms_hyperv.priv_high & HV_ACCESS_PARTITION_ID)
+		hv_get_partition_id();
+	ms_hyperv.vtl = get_vtl();
+	if (ms_hyperv.vtl > 0) /* non default VTL */
+		pr_info("Linux runs in Hyper-V Virtual Trust Level %d\n", ms_hyperv.vtl);
+
+	ms_hyperv_late_init();
+
+	hyperv_initialized = true;
+	return 0;
+}
+
+early_initcall(hyperv_init);
+
+bool hv_is_hyperv_initialized(void)
+{
+	return hyperv_initialized;
+}
+EXPORT_SYMBOL_GPL(hv_is_hyperv_initialized);