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, 186 insertions, 0 deletions

diff --git a/drivers/misc/ocxl/sysfs.c b/drivers/misc/ocxl/sysfs.c
new file mode 100644
index 000000000..1b6a86f17
--- /dev/null
+++ b/drivers/misc/ocxl/sysfs.c
@@ -0,0 +1,186 @@
+// SPDX-License-Identifier: GPL-2.0+
+// Copyright 2017 IBM Corp.
+#include <linux/sysfs.h>
+#include "ocxl_internal.h"
+
+static inline struct ocxl_afu *to_afu(struct device *device)
+{
+	struct ocxl_file_info *info = container_of(device, struct ocxl_file_info, dev);
+
+	return info->afu;
+}
+
+static ssize_t global_mmio_size_show(struct device *device,
+				struct device_attribute *attr,
+				char *buf)
+{
+	struct ocxl_afu *afu = to_afu(device);
+
+	return sysfs_emit(buf, "%d\n",
+			afu->config.global_mmio_size);
+}
+
+static ssize_t pp_mmio_size_show(struct device *device,
+				struct device_attribute *attr,
+				char *buf)
+{
+	struct ocxl_afu *afu = to_afu(device);
+
+	return sysfs_emit(buf, "%d\n",
+			afu->config.pp_mmio_stride);
+}
+
+static ssize_t afu_version_show(struct device *device,
+				struct device_attribute *attr,
+				char *buf)
+{
+	struct ocxl_afu *afu = to_afu(device);
+
+	return sysfs_emit(buf, "%hhu:%hhu\n",
+			afu->config.version_major,
+			afu->config.version_minor);
+}
+
+static ssize_t contexts_show(struct device *device,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct ocxl_afu *afu = to_afu(device);
+
+	return sysfs_emit(buf, "%d/%d\n",
+			afu->pasid_count, afu->pasid_max);
+}
+
+static ssize_t reload_on_reset_show(struct device *device,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct ocxl_afu *afu = to_afu(device);
+	struct ocxl_fn *fn = afu->fn;
+	struct pci_dev *pci_dev = to_pci_dev(fn->dev.parent);
+	int val;
+
+	if (ocxl_config_get_reset_reload(pci_dev, &val))
+		return sysfs_emit(buf, "unavailable\n");
+
+	return sysfs_emit(buf, "%d\n", val);
+}
+
+static ssize_t reload_on_reset_store(struct device *device,
+				     struct device_attribute *attr,
+				     const char *buf, size_t count)
+{
+	struct ocxl_afu *afu = to_afu(device);
+	struct ocxl_fn *fn = afu->fn;
+	struct pci_dev *pci_dev = to_pci_dev(fn->dev.parent);
+	int rc, val;
+
+	rc = kstrtoint(buf, 0, &val);
+	if (rc || (val != 0 && val != 1))
+		return -EINVAL;
+
+	if (ocxl_config_set_reset_reload(pci_dev, val))
+		return -ENODEV;
+
+	return count;
+}
+
+static struct device_attribute afu_attrs[] = {
+	__ATTR_RO(global_mmio_size),
+	__ATTR_RO(pp_mmio_size),
+	__ATTR_RO(afu_version),
+	__ATTR_RO(contexts),
+	__ATTR_RW(reload_on_reset),
+};
+
+static ssize_t global_mmio_read(struct file *filp, struct kobject *kobj,
+				const struct bin_attribute *bin_attr, char *buf,
+				loff_t off, size_t count)
+{
+	struct ocxl_afu *afu = to_afu(kobj_to_dev(kobj));
+
+	if (count == 0 || off < 0 ||
+		off >= afu->config.global_mmio_size)
+		return 0;
+	memcpy_fromio(buf, afu->global_mmio_ptr + off, count);
+	return count;
+}
+
+static vm_fault_t global_mmio_fault(struct vm_fault *vmf)
+{
+	struct vm_area_struct *vma = vmf->vma;
+	struct ocxl_afu *afu = vma->vm_private_data;
+	unsigned long offset;
+
+	if (vmf->pgoff >= (afu->config.global_mmio_size >> PAGE_SHIFT))
+		return VM_FAULT_SIGBUS;
+
+	offset = vmf->pgoff;
+	offset += (afu->global_mmio_start >> PAGE_SHIFT);
+	return vmf_insert_pfn(vma, vmf->address, offset);
+}
+
+static const struct vm_operations_struct global_mmio_vmops = {
+	.fault = global_mmio_fault,
+};
+
+static int global_mmio_mmap(struct file *filp, struct kobject *kobj,
+			const struct bin_attribute *bin_attr,
+			struct vm_area_struct *vma)
+{
+	struct ocxl_afu *afu = to_afu(kobj_to_dev(kobj));
+
+	if ((vma_pages(vma) + vma->vm_pgoff) >
+		(afu->config.global_mmio_size >> PAGE_SHIFT))
+		return -EINVAL;
+
+	vm_flags_set(vma, VM_IO | VM_PFNMAP);
+	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
+	vma->vm_ops = &global_mmio_vmops;
+	vma->vm_private_data = afu;
+	return 0;
+}
+
+int ocxl_sysfs_register_afu(struct ocxl_file_info *info)
+{
+	int i, rc;
+
+	for (i = 0; i < ARRAY_SIZE(afu_attrs); i++) {
+		rc = device_create_file(&info->dev, &afu_attrs[i]);
+		if (rc)
+			goto err;
+	}
+
+	sysfs_attr_init(&info->attr_global_mmio.attr);
+	info->attr_global_mmio.attr.name = "global_mmio_area";
+	info->attr_global_mmio.attr.mode = 0600;
+	info->attr_global_mmio.size = info->afu->config.global_mmio_size;
+	info->attr_global_mmio.read = global_mmio_read;
+	info->attr_global_mmio.mmap = global_mmio_mmap;
+	rc = device_create_bin_file(&info->dev, &info->attr_global_mmio);
+	if (rc) {
+		dev_err(&info->dev, "Unable to create global mmio attr for afu: %d\n", rc);
+		goto err;
+	}
+
+	return 0;
+
+err:
+	for (i--; i >= 0; i--)
+		device_remove_file(&info->dev, &afu_attrs[i]);
+
+	return rc;
+}
+
+void ocxl_sysfs_unregister_afu(struct ocxl_file_info *info)
+{
+	int i;
+
+	/*
+	 * device_remove_bin_file is safe to call if the file is not added as
+	 * the files are removed by name, and early exit if not found
+	 */
+	for (i = 0; i < ARRAY_SIZE(afu_attrs); i++)
+		device_remove_file(&info->dev, &afu_attrs[i]);
+	device_remove_bin_file(&info->dev, &info->attr_global_mmio);
+}