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, 204 insertions, 0 deletions

diff --git a/fs/verity/read_metadata.c b/fs/verity/read_metadata.c
new file mode 100644
index 000000000..b4c089243
--- /dev/null
+++ b/fs/verity/read_metadata.c
@@ -0,0 +1,204 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Ioctl to read verity metadata
+ *
+ * Copyright 2021 Google LLC
+ */
+
+#include "fsverity_private.h"
+
+#include <linux/backing-dev.h>
+#include <linux/export.h>
+#include <linux/highmem.h>
+#include <linux/sched/signal.h>
+#include <linux/uaccess.h>
+
+static int fsverity_read_merkle_tree(struct inode *inode,
+				     const struct fsverity_info *vi,
+				     void __user *buf, u64 offset, int length)
+{
+	const struct fsverity_operations *vops = inode->i_sb->s_vop;
+	u64 end_offset;
+	unsigned int offs_in_page;
+	pgoff_t index, last_index;
+	int retval = 0;
+	int err = 0;
+
+	end_offset = min(offset + length, vi->tree_params.tree_size);
+	if (offset >= end_offset)
+		return 0;
+	offs_in_page = offset_in_page(offset);
+	index = offset >> PAGE_SHIFT;
+	last_index = (end_offset - 1) >> PAGE_SHIFT;
+
+	/*
+	 * Kick off readahead for the range we are going to read to ensure a
+	 * single large sequential read instead of lots of small ones.
+	 */
+	if (inode->i_sb->s_vop->readahead_merkle_tree) {
+		filemap_invalidate_lock_shared(inode->i_mapping);
+		inode->i_sb->s_vop->readahead_merkle_tree(
+			inode, index, last_index - index + 1);
+		filemap_invalidate_unlock_shared(inode->i_mapping);
+	}
+
+	/*
+	 * Iterate through each Merkle tree page in the requested range and copy
+	 * the requested portion to userspace.  Note that the Merkle tree block
+	 * size isn't important here, as we are returning a byte stream; i.e.,
+	 * we can just work with pages even if the tree block size != PAGE_SIZE.
+	 */
+	for (; index <= last_index; index++) {
+		unsigned int bytes_to_copy = min_t(u64, end_offset - offset,
+						   PAGE_SIZE - offs_in_page);
+		struct page *page;
+		const void *virt;
+
+		page = vops->read_merkle_tree_page(inode, index);
+		if (IS_ERR(page)) {
+			err = PTR_ERR(page);
+			fsverity_err(inode,
+				     "Error %d reading Merkle tree page %lu",
+				     err, index);
+			break;
+		}
+
+		virt = kmap_local_page(page);
+		if (copy_to_user(buf, virt + offs_in_page, bytes_to_copy)) {
+			kunmap_local(virt);
+			put_page(page);
+			err = -EFAULT;
+			break;
+		}
+		kunmap_local(virt);
+		put_page(page);
+
+		retval += bytes_to_copy;
+		buf += bytes_to_copy;
+		offset += bytes_to_copy;
+
+		if (fatal_signal_pending(current))  {
+			err = -EINTR;
+			break;
+		}
+		cond_resched();
+		offs_in_page = 0;
+	}
+	return retval ? retval : err;
+}
+
+/* Copy the requested portion of the buffer to userspace. */
+static int fsverity_read_buffer(void __user *dst, u64 offset, int length,
+				const void *src, size_t src_length)
+{
+	if (offset >= src_length)
+		return 0;
+	src += offset;
+	src_length -= offset;
+
+	length = min_t(size_t, length, src_length);
+
+	if (copy_to_user(dst, src, length))
+		return -EFAULT;
+
+	return length;
+}
+
+static int fsverity_read_descriptor(struct inode *inode,
+				    void __user *buf, u64 offset, int length)
+{
+	struct fsverity_descriptor *desc;
+	size_t desc_size;
+	int res;
+
+	res = fsverity_get_descriptor(inode, &desc);
+	if (res)
+		return res;
+
+	/* don't include the builtin signature */
+	desc_size = offsetof(struct fsverity_descriptor, signature);
+	desc->sig_size = 0;
+
+	res = fsverity_read_buffer(buf, offset, length, desc, desc_size);
+
+	kfree(desc);
+	return res;
+}
+
+static int fsverity_read_signature(struct inode *inode,
+				   void __user *buf, u64 offset, int length)
+{
+	struct fsverity_descriptor *desc;
+	int res;
+
+	res = fsverity_get_descriptor(inode, &desc);
+	if (res)
+		return res;
+
+	if (desc->sig_size == 0) {
+		res = -ENODATA;
+		goto out;
+	}
+
+	/*
+	 * Include only the builtin signature.  fsverity_get_descriptor()
+	 * already verified that sig_size is in-bounds.
+	 */
+	res = fsverity_read_buffer(buf, offset, length, desc->signature,
+				   le32_to_cpu(desc->sig_size));
+out:
+	kfree(desc);
+	return res;
+}
+
+/**
+ * fsverity_ioctl_read_metadata() - read verity metadata from a file
+ * @filp: file to read the metadata from
+ * @uarg: user pointer to fsverity_read_metadata_arg
+ *
+ * Return: length read on success, 0 on EOF, -errno on failure
+ */
+int fsverity_ioctl_read_metadata(struct file *filp, const void __user *uarg)
+{
+	struct inode *inode = file_inode(filp);
+	const struct fsverity_info *vi;
+	struct fsverity_read_metadata_arg arg;
+	int length;
+	void __user *buf;
+
+	vi = fsverity_get_info(inode);
+	if (!vi)
+		return -ENODATA; /* not a verity file */
+	/*
+	 * Note that we don't have to explicitly check that the file is open for
+	 * reading, since verity files can only be opened for reading.
+	 */
+
+	if (copy_from_user(&arg, uarg, sizeof(arg)))
+		return -EFAULT;
+
+	if (arg.__reserved)
+		return -EINVAL;
+
+	/* offset + length must not overflow. */
+	if (arg.offset + arg.length < arg.offset)
+		return -EINVAL;
+
+	/* Ensure that the return value will fit in INT_MAX. */
+	length = min_t(u64, arg.length, INT_MAX);
+
+	buf = u64_to_user_ptr(arg.buf_ptr);
+
+	switch (arg.metadata_type) {
+	case FS_VERITY_METADATA_TYPE_MERKLE_TREE:
+		return fsverity_read_merkle_tree(inode, vi, buf, arg.offset,
+						 length);
+	case FS_VERITY_METADATA_TYPE_DESCRIPTOR:
+		return fsverity_read_descriptor(inode, buf, arg.offset, length);
+	case FS_VERITY_METADATA_TYPE_SIGNATURE:
+		return fsverity_read_signature(inode, buf, arg.offset, length);
+	default:
+		return -EINVAL;
+	}
+}
+EXPORT_SYMBOL_GPL(fsverity_ioctl_read_metadata);