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, 331 insertions, 0 deletions

diff --git a/fs/hfs/dir.c b/fs/hfs/dir.c
new file mode 100644
index 000000000..e1f1fb351
--- /dev/null
+++ b/fs/hfs/dir.c
@@ -0,0 +1,331 @@
+/*
+ *  linux/fs/hfs/dir.c
+ *
+ * Copyright (C) 1995-1997  Paul H. Hargrove
+ * (C) 2003 Ardis Technologies <roman@ardistech.com>
+ * This file may be distributed under the terms of the GNU General Public License.
+ *
+ * This file contains directory-related functions independent of which
+ * scheme is being used to represent forks.
+ *
+ * Based on the minix file system code, (C) 1991, 1992 by Linus Torvalds
+ */
+
+#include "hfs_fs.h"
+#include "btree.h"
+
+/*
+ * hfs_lookup()
+ */
+static struct dentry *hfs_lookup(struct inode *dir, struct dentry *dentry,
+				 unsigned int flags)
+{
+	hfs_cat_rec rec;
+	struct hfs_find_data fd;
+	struct inode *inode = NULL;
+	int res;
+
+	res = hfs_find_init(HFS_SB(dir->i_sb)->cat_tree, &fd);
+	if (res)
+		return ERR_PTR(res);
+	hfs_cat_build_key(dir->i_sb, fd.search_key, dir->i_ino, &dentry->d_name);
+	res = hfs_brec_read(&fd, &rec, sizeof(rec));
+	if (res) {
+		if (res != -ENOENT)
+			inode = ERR_PTR(res);
+	} else {
+		inode = hfs_iget(dir->i_sb, &fd.search_key->cat, &rec);
+		if (!inode)
+			inode = ERR_PTR(-EACCES);
+	}
+	hfs_find_exit(&fd);
+	return d_splice_alias(inode, dentry);
+}
+
+/*
+ * hfs_readdir
+ */
+static int hfs_readdir(struct file *file, struct dir_context *ctx)
+{
+	struct inode *inode = file_inode(file);
+	struct super_block *sb = inode->i_sb;
+	int len, err;
+	char strbuf[HFS_MAX_NAMELEN];
+	union hfs_cat_rec entry;
+	struct hfs_find_data fd;
+	struct hfs_readdir_data *rd;
+	u16 type;
+
+	if (ctx->pos >= inode->i_size)
+		return 0;
+
+	err = hfs_find_init(HFS_SB(sb)->cat_tree, &fd);
+	if (err)
+		return err;
+	hfs_cat_build_key(sb, fd.search_key, inode->i_ino, NULL);
+	err = hfs_brec_find(&fd);
+	if (err)
+		goto out;
+
+	if (ctx->pos == 0) {
+		/* This is completely artificial... */
+		if (!dir_emit_dot(file, ctx))
+			goto out;
+		ctx->pos = 1;
+	}
+	if (ctx->pos == 1) {
+		if (fd.entrylength > sizeof(entry) || fd.entrylength < 0) {
+			err = -EIO;
+			goto out;
+		}
+
+		hfs_bnode_read(fd.bnode, &entry, fd.entryoffset, fd.entrylength);
+		if (entry.type != HFS_CDR_THD) {
+			pr_err("bad catalog folder thread\n");
+			err = -EIO;
+			goto out;
+		}
+		//if (fd.entrylength < HFS_MIN_THREAD_SZ) {
+		//	pr_err("truncated catalog thread\n");
+		//	err = -EIO;
+		//	goto out;
+		//}
+		if (!dir_emit(ctx, "..", 2,
+			    be32_to_cpu(entry.thread.ParID), DT_DIR))
+			goto out;
+		ctx->pos = 2;
+	}
+	if (ctx->pos >= inode->i_size)
+		goto out;
+	rd = file->private_data;
+	if (rd && rd->pos == ctx->pos) {
+		memcpy(fd.search_key, &rd->key, sizeof(struct hfs_cat_key));
+		err = hfs_brec_find(&fd);
+		if (err == -ENOENT)
+			err = hfs_brec_goto(&fd, 1);
+	} else {
+		err = hfs_brec_goto(&fd, ctx->pos - 1);
+	}
+	if (err)
+		goto out;
+
+	for (;;) {
+		if (be32_to_cpu(fd.key->cat.ParID) != inode->i_ino) {
+			pr_err("walked past end of dir\n");
+			err = -EIO;
+			goto out;
+		}
+
+		if (fd.entrylength > sizeof(entry) || fd.entrylength < 0) {
+			err = -EIO;
+			goto out;
+		}
+
+		hfs_bnode_read(fd.bnode, &entry, fd.entryoffset, fd.entrylength);
+		type = entry.type;
+		len = hfs_mac2asc(sb, strbuf, &fd.key->cat.CName);
+		if (type == HFS_CDR_DIR) {
+			if (fd.entrylength < sizeof(struct hfs_cat_dir)) {
+				pr_err("small dir entry\n");
+				err = -EIO;
+				goto out;
+			}
+			if (!dir_emit(ctx, strbuf, len,
+				    be32_to_cpu(entry.dir.DirID), DT_DIR))
+				break;
+		} else if (type == HFS_CDR_FIL) {
+			if (fd.entrylength < sizeof(struct hfs_cat_file)) {
+				pr_err("small file entry\n");
+				err = -EIO;
+				goto out;
+			}
+			if (!dir_emit(ctx, strbuf, len,
+				    be32_to_cpu(entry.file.FlNum), DT_REG))
+				break;
+		} else {
+			pr_err("bad catalog entry type %d\n", type);
+			err = -EIO;
+			goto out;
+		}
+		ctx->pos++;
+		if (ctx->pos >= inode->i_size)
+			goto out;
+		err = hfs_brec_goto(&fd, 1);
+		if (err)
+			goto out;
+	}
+	if (!rd) {
+		rd = kmalloc_obj(struct hfs_readdir_data);
+		if (!rd) {
+			err = -ENOMEM;
+			goto out;
+		}
+		file->private_data = rd;
+	}
+	rd->pos = ctx->pos;
+	memcpy(&rd->key, &fd.key->cat, sizeof(struct hfs_cat_key));
+out:
+	hfs_find_exit(&fd);
+	return err;
+}
+
+static int hfs_dir_release(struct inode *inode, struct file *file)
+{
+	kfree(file->private_data);
+	return 0;
+}
+
+/*
+ * hfs_create()
+ *
+ * This is the create() entry in the inode_operations structure for
+ * regular HFS directories.  The purpose is to create a new file in
+ * a directory and return a corresponding inode, given the inode for
+ * the directory and the name (and its length) of the new file.
+ */
+static int hfs_create(struct mnt_idmap *idmap, struct inode *dir,
+		      struct dentry *dentry, umode_t mode)
+{
+	struct inode *inode;
+	int res;
+
+	inode = hfs_new_inode(dir, &dentry->d_name, mode);
+	if (IS_ERR(inode))
+		return PTR_ERR(inode);
+
+	res = hfs_cat_create(inode->i_ino, dir, &dentry->d_name, inode);
+	if (res) {
+		clear_nlink(inode);
+		hfs_delete_inode(inode);
+		iput(inode);
+		return res;
+	}
+	d_instantiate(dentry, inode);
+	mark_inode_dirty(inode);
+	return 0;
+}
+
+/*
+ * hfs_mkdir()
+ *
+ * This is the mkdir() entry in the inode_operations structure for
+ * regular HFS directories.  The purpose is to create a new directory
+ * in a directory, given the inode for the parent directory and the
+ * name (and its length) of the new directory.
+ */
+static struct dentry *hfs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
+				struct dentry *dentry, umode_t mode)
+{
+	struct inode *inode;
+	int res;
+
+	inode = hfs_new_inode(dir, &dentry->d_name, mode);
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	res = hfs_cat_create(inode->i_ino, dir, &dentry->d_name, inode);
+	if (res) {
+		clear_nlink(inode);
+		hfs_delete_inode(inode);
+		iput(inode);
+		return ERR_PTR(res);
+	}
+	d_instantiate(dentry, inode);
+	mark_inode_dirty(inode);
+	return NULL;
+}
+
+/*
+ * hfs_remove()
+ *
+ * This serves as both unlink() and rmdir() in the inode_operations
+ * structure for regular HFS directories.  The purpose is to delete
+ * an existing child, given the inode for the parent directory and
+ * the name (and its length) of the existing directory.
+ *
+ * HFS does not have hardlinks, so both rmdir and unlink set the
+ * link count to 0.  The only difference is the emptiness check.
+ */
+static int hfs_remove(struct inode *dir, struct dentry *dentry)
+{
+	struct super_block *sb = dir->i_sb;
+	struct inode *inode = d_inode(dentry);
+	int res;
+
+	if (S_ISDIR(inode->i_mode) && inode->i_size != 2)
+		return -ENOTEMPTY;
+
+	if (unlikely(!is_hfs_cnid_counts_valid(sb))) {
+	    pr_err("cannot remove file/folder\n");
+	    return -ERANGE;
+	}
+
+	res = hfs_cat_delete(inode->i_ino, dir, &dentry->d_name);
+	if (res)
+		return res;
+	clear_nlink(inode);
+	inode_set_ctime_current(inode);
+	hfs_delete_inode(inode);
+	mark_inode_dirty(inode);
+	return 0;
+}
+
+/*
+ * hfs_rename()
+ *
+ * This is the rename() entry in the inode_operations structure for
+ * regular HFS directories.  The purpose is to rename an existing
+ * file or directory, given the inode for the current directory and
+ * the name (and its length) of the existing file/directory and the
+ * inode for the new directory and the name (and its length) of the
+ * new file/directory.
+ * XXX: how do you handle must_be dir?
+ */
+static int hfs_rename(struct mnt_idmap *idmap, struct inode *old_dir,
+		      struct dentry *old_dentry, struct inode *new_dir,
+		      struct dentry *new_dentry, unsigned int flags)
+{
+	int res;
+
+	if (flags & ~RENAME_NOREPLACE)
+		return -EINVAL;
+
+	/* Unlink destination if it already exists */
+	if (d_really_is_positive(new_dentry)) {
+		res = hfs_remove(new_dir, new_dentry);
+		if (res)
+			return res;
+	}
+
+	res = hfs_cat_move(d_inode(old_dentry)->i_ino,
+			   old_dir, &old_dentry->d_name,
+			   new_dir, &new_dentry->d_name);
+	if (!res) {
+		struct inode *inode = d_inode(old_dentry);
+
+		hfs_cat_build_key(old_dir->i_sb,
+				  (btree_key *)&HFS_I(inode)->cat_key,
+				  new_dir->i_ino, &new_dentry->d_name);
+		inode_set_ctime_current(inode);
+		mark_inode_dirty(inode);
+	}
+	return res;
+}
+
+const struct file_operations hfs_dir_operations = {
+	.read		= generic_read_dir,
+	.iterate_shared	= hfs_readdir,
+	.llseek		= generic_file_llseek,
+	.release	= hfs_dir_release,
+};
+
+const struct inode_operations hfs_dir_inode_operations = {
+	.create		= hfs_create,
+	.lookup		= hfs_lookup,
+	.unlink		= hfs_remove,
+	.mkdir		= hfs_mkdir,
+	.rmdir		= hfs_remove,
+	.rename		= hfs_rename,
+	.setattr	= hfs_inode_setattr,
+	.fileattr_get	= hfs_fileattr_get,
+};