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, 288 insertions, 0 deletions

diff --git a/drivers/hv/hv_proc.c b/drivers/hv/hv_proc.c
new file mode 100644
index 000000000..57b2c6419
--- /dev/null
+++ b/drivers/hv/hv_proc.c
@@ -0,0 +1,288 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/types.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/clockchips.h>
+#include <linux/slab.h>
+#include <linux/cpuhotplug.h>
+#include <linux/minmax.h>
+#include <linux/export.h>
+#include <asm/mshyperv.h>
+
+/*
+ * See struct hv_deposit_memory. The first u64 is partition ID, the rest
+ * are GPAs.
+ */
+#define HV_DEPOSIT_MAX (HV_HYP_PAGE_SIZE / sizeof(u64) - 1)
+
+/* Deposits exact number of pages. Must be called with interrupts enabled.  */
+int hv_call_deposit_pages(int node, u64 partition_id, u32 num_pages)
+{
+	struct page **pages, *page;
+	int *counts;
+	int num_allocations;
+	int i, j, page_count;
+	int order;
+	u64 status;
+	int ret;
+	u64 base_pfn;
+	struct hv_deposit_memory *input_page;
+	unsigned long flags;
+
+	if (num_pages > HV_DEPOSIT_MAX)
+		return -E2BIG;
+	if (!num_pages)
+		return 0;
+
+	/* One buffer for page pointers and counts */
+	page = alloc_page(GFP_KERNEL);
+	if (!page)
+		return -ENOMEM;
+	pages = page_address(page);
+
+	counts = kzalloc_objs(int, HV_DEPOSIT_MAX);
+	if (!counts) {
+		free_page((unsigned long)pages);
+		return -ENOMEM;
+	}
+
+	/* Allocate all the pages before disabling interrupts */
+	i = 0;
+
+	while (num_pages) {
+		/* Find highest order we can actually allocate */
+		order = 31 - __builtin_clz(num_pages);
+
+		while (1) {
+			pages[i] = alloc_pages_node(node, GFP_KERNEL, order);
+			if (pages[i])
+				break;
+			if (!order) {
+				ret = -ENOMEM;
+				num_allocations = i;
+				goto err_free_allocations;
+			}
+			--order;
+		}
+
+		split_page(pages[i], order);
+		counts[i] = 1 << order;
+		num_pages -= counts[i];
+		i++;
+	}
+	num_allocations = i;
+
+	local_irq_save(flags);
+
+	input_page = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+	input_page->partition_id = partition_id;
+
+	/* Populate gpa_page_list - these will fit on the input page */
+	for (i = 0, page_count = 0; i < num_allocations; ++i) {
+		base_pfn = page_to_pfn(pages[i]);
+		for (j = 0; j < counts[i]; ++j, ++page_count)
+			input_page->gpa_page_list[page_count] = base_pfn + j;
+	}
+	status = hv_do_rep_hypercall(HVCALL_DEPOSIT_MEMORY,
+				     page_count, 0, input_page, NULL);
+	local_irq_restore(flags);
+	if (!hv_result_success(status)) {
+		hv_status_err(status, "\n");
+		ret = hv_result_to_errno(status);
+		goto err_free_allocations;
+	}
+
+	ret = 0;
+	goto free_buf;
+
+err_free_allocations:
+	for (i = 0; i < num_allocations; ++i) {
+		base_pfn = page_to_pfn(pages[i]);
+		for (j = 0; j < counts[i]; ++j)
+			__free_page(pfn_to_page(base_pfn + j));
+	}
+
+free_buf:
+	free_page((unsigned long)pages);
+	kfree(counts);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(hv_call_deposit_pages);
+
+int hv_deposit_memory_node(int node, u64 partition_id,
+			   u64 hv_status)
+{
+	u32 num_pages = 1;
+
+	switch (hv_result(hv_status)) {
+	case HV_STATUS_INSUFFICIENT_MEMORY:
+		break;
+	case HV_STATUS_INSUFFICIENT_CONTIGUOUS_MEMORY:
+		num_pages = HV_MAX_CONTIGUOUS_ALLOCATION_PAGES;
+		break;
+
+	case HV_STATUS_INSUFFICIENT_CONTIGUOUS_ROOT_MEMORY:
+		num_pages = HV_MAX_CONTIGUOUS_ALLOCATION_PAGES;
+		fallthrough;
+	case HV_STATUS_INSUFFICIENT_ROOT_MEMORY:
+		if (!hv_root_partition()) {
+			hv_status_err(hv_status, "Unexpected root memory deposit\n");
+			return -ENOMEM;
+		}
+		partition_id = HV_PARTITION_ID_SELF;
+		break;
+
+	default:
+		hv_status_err(hv_status, "Unexpected!\n");
+		return -ENOMEM;
+	}
+	return hv_call_deposit_pages(node, partition_id, num_pages);
+}
+EXPORT_SYMBOL_GPL(hv_deposit_memory_node);
+
+bool hv_result_needs_memory(u64 status)
+{
+	switch (hv_result(status)) {
+	case HV_STATUS_INSUFFICIENT_MEMORY:
+	case HV_STATUS_INSUFFICIENT_CONTIGUOUS_MEMORY:
+	case HV_STATUS_INSUFFICIENT_ROOT_MEMORY:
+	case HV_STATUS_INSUFFICIENT_CONTIGUOUS_ROOT_MEMORY:
+		return true;
+	}
+	return false;
+}
+EXPORT_SYMBOL_GPL(hv_result_needs_memory);
+
+int hv_call_add_logical_proc(int node, u32 lp_index, u32 apic_id)
+{
+	struct hv_input_add_logical_processor *input;
+	struct hv_output_add_logical_processor *output;
+	u64 status;
+	unsigned long flags;
+	int ret = 0;
+
+	/*
+	 * When adding a logical processor, the hypervisor may return
+	 * HV_STATUS_INSUFFICIENT_MEMORY. When that happens, we deposit more
+	 * pages and retry.
+	 */
+	do {
+		local_irq_save(flags);
+
+		input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+		/* We don't do anything with the output right now */
+		output = *this_cpu_ptr(hyperv_pcpu_output_arg);
+
+		input->lp_index = lp_index;
+		input->apic_id = apic_id;
+		input->proximity_domain_info = hv_numa_node_to_pxm_info(node);
+		status = hv_do_hypercall(HVCALL_ADD_LOGICAL_PROCESSOR,
+					 input, output);
+		local_irq_restore(flags);
+
+		if (!hv_result_needs_memory(status)) {
+			if (!hv_result_success(status)) {
+				hv_status_err(status, "cpu %u apic ID: %u\n",
+					      lp_index, apic_id);
+				ret = hv_result_to_errno(status);
+			}
+			break;
+		}
+		ret = hv_deposit_memory_node(node, hv_current_partition_id,
+					     status);
+	} while (!ret);
+
+	return ret;
+}
+
+int hv_call_create_vp(int node, u64 partition_id, u32 vp_index, u32 flags)
+{
+	struct hv_create_vp *input;
+	u64 status;
+	unsigned long irq_flags;
+	int ret = 0;
+
+	/* Root VPs don't seem to need pages deposited */
+	if (partition_id != hv_current_partition_id) {
+		/* The value 90 is empirically determined. It may change. */
+		ret = hv_call_deposit_pages(node, partition_id, 90);
+		if (ret)
+			return ret;
+	}
+
+	do {
+		local_irq_save(irq_flags);
+
+		input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+
+		input->partition_id = partition_id;
+		input->vp_index = vp_index;
+		input->flags = flags;
+		input->subnode_type = HV_SUBNODE_ANY;
+		input->proximity_domain_info = hv_numa_node_to_pxm_info(node);
+		status = hv_do_hypercall(HVCALL_CREATE_VP, input, NULL);
+		local_irq_restore(irq_flags);
+
+		if (!hv_result_needs_memory(status)) {
+			if (!hv_result_success(status)) {
+				hv_status_err(status, "vcpu: %u, lp: %u\n",
+					      vp_index, flags);
+				ret = hv_result_to_errno(status);
+			}
+			break;
+		}
+		ret = hv_deposit_memory_node(node, partition_id, status);
+
+	} while (!ret);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(hv_call_create_vp);
+
+int hv_call_notify_all_processors_started(void)
+{
+	struct hv_input_notify_partition_event *input;
+	u64 status;
+	unsigned long irq_flags;
+	int ret = 0;
+
+	local_irq_save(irq_flags);
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	memset(input, 0, sizeof(*input));
+	input->event = HV_PARTITION_ALL_LOGICAL_PROCESSORS_STARTED;
+	status = hv_do_hypercall(HVCALL_NOTIFY_PARTITION_EVENT,
+				 input, NULL);
+	local_irq_restore(irq_flags);
+
+	if (!hv_result_success(status)) {
+		hv_status_err(status, "\n");
+		ret = hv_result_to_errno(status);
+	}
+	return ret;
+}
+
+bool hv_lp_exists(u32 lp_index)
+{
+	struct hv_input_get_logical_processor_run_time *input;
+	struct hv_output_get_logical_processor_run_time *output;
+	unsigned long flags;
+	u64 status;
+
+	local_irq_save(flags);
+	input = *this_cpu_ptr(hyperv_pcpu_input_arg);
+	output = *this_cpu_ptr(hyperv_pcpu_output_arg);
+
+	input->lp_index = lp_index;
+	status = hv_do_hypercall(HVCALL_GET_LOGICAL_PROCESSOR_RUN_TIME,
+				 input, output);
+	local_irq_restore(flags);
+
+	if (!hv_result_success(status) &&
+	    hv_result(status) != HV_STATUS_INVALID_LP_INDEX) {
+		hv_status_err(status, "\n");
+		BUG();
+	}
+
+	return hv_result_success(status);
+}