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, 307 insertions, 0 deletions

diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c
new file mode 100644
index 000000000..230d9326b
--- /dev/null
+++ b/fs/btrfs/export.c
@@ -0,0 +1,307 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/fs.h>
+#include <linux/types.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "btrfs_inode.h"
+#include "export.h"
+#include "accessors.h"
+#include "super.h"
+
+#define BTRFS_FID_SIZE_NON_CONNECTABLE (offsetof(struct btrfs_fid, \
+						 parent_objectid) / 4)
+#define BTRFS_FID_SIZE_CONNECTABLE (offsetof(struct btrfs_fid, \
+					     parent_root_objectid) / 4)
+#define BTRFS_FID_SIZE_CONNECTABLE_ROOT (sizeof(struct btrfs_fid) / 4)
+
+static int btrfs_encode_fh(struct inode *inode, u32 *fh, int *max_len,
+			   struct inode *parent)
+{
+	struct btrfs_fid *fid = (struct btrfs_fid *)fh;
+	int len = *max_len;
+	int type;
+
+	if (parent && (len < BTRFS_FID_SIZE_CONNECTABLE)) {
+		if (btrfs_root_id(BTRFS_I(inode)->root) !=
+		    btrfs_root_id(BTRFS_I(parent)->root))
+			*max_len = BTRFS_FID_SIZE_CONNECTABLE_ROOT;
+		else
+			*max_len = BTRFS_FID_SIZE_CONNECTABLE;
+		return FILEID_INVALID;
+	} else if (len < BTRFS_FID_SIZE_NON_CONNECTABLE) {
+		*max_len = BTRFS_FID_SIZE_NON_CONNECTABLE;
+		return FILEID_INVALID;
+	}
+
+	len  = BTRFS_FID_SIZE_NON_CONNECTABLE;
+	type = FILEID_BTRFS_WITHOUT_PARENT;
+
+	fid->objectid = btrfs_ino(BTRFS_I(inode));
+	fid->root_objectid = btrfs_root_id(BTRFS_I(inode)->root);
+	fid->gen = inode->i_generation;
+
+	if (parent) {
+		u64 parent_root_id;
+
+		fid->parent_objectid = btrfs_ino(BTRFS_I(parent));
+		fid->parent_gen = parent->i_generation;
+		parent_root_id = btrfs_root_id(BTRFS_I(parent)->root);
+
+		if (parent_root_id != fid->root_objectid) {
+			if (*max_len < BTRFS_FID_SIZE_CONNECTABLE_ROOT)
+				return FILEID_INVALID;
+			fid->parent_root_objectid = parent_root_id;
+			len = BTRFS_FID_SIZE_CONNECTABLE_ROOT;
+			type = FILEID_BTRFS_WITH_PARENT_ROOT;
+		} else {
+			len = BTRFS_FID_SIZE_CONNECTABLE;
+			type = FILEID_BTRFS_WITH_PARENT;
+		}
+	}
+
+	*max_len = len;
+	return type;
+}
+
+/*
+ * Read dentry of inode with @objectid from filesystem root @root_objectid.
+ *
+ * @sb:             the filesystem super block
+ * @objectid:       inode objectid
+ * @root_objectid:  object id of the subvolume root where to look up the inode
+ * @generation:     optional, if not zero, verify that the found inode
+ *                  generation matches
+ *
+ * Return dentry alias for the inode, otherwise an error. In case the
+ * generation does not match return ESTALE.
+ */
+struct dentry *btrfs_get_dentry(struct super_block *sb, u64 objectid,
+				u64 root_objectid, u64 generation)
+{
+	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
+	struct btrfs_root *root;
+	struct btrfs_inode *inode;
+
+	if (objectid < BTRFS_FIRST_FREE_OBJECTID)
+		return ERR_PTR(-ESTALE);
+
+	root = btrfs_get_fs_root(fs_info, root_objectid, true);
+	if (IS_ERR(root))
+		return ERR_CAST(root);
+
+	inode = btrfs_iget(objectid, root);
+	btrfs_put_root(root);
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	if (generation != 0 && generation != inode->vfs_inode.i_generation) {
+		iput(&inode->vfs_inode);
+		return ERR_PTR(-ESTALE);
+	}
+
+	return d_obtain_alias(&inode->vfs_inode);
+}
+
+static struct dentry *btrfs_fh_to_parent(struct super_block *sb, struct fid *fh,
+					 int fh_len, int fh_type)
+{
+	struct btrfs_fid *fid = (struct btrfs_fid *) fh;
+	u64 objectid, root_objectid;
+	u32 generation;
+
+	if (fh_type == FILEID_BTRFS_WITH_PARENT) {
+		if (fh_len <  BTRFS_FID_SIZE_CONNECTABLE)
+			return NULL;
+		root_objectid = fid->root_objectid;
+	} else if (fh_type == FILEID_BTRFS_WITH_PARENT_ROOT) {
+		if (fh_len < BTRFS_FID_SIZE_CONNECTABLE_ROOT)
+			return NULL;
+		root_objectid = fid->parent_root_objectid;
+	} else
+		return NULL;
+
+	objectid = fid->parent_objectid;
+	generation = fid->parent_gen;
+
+	return btrfs_get_dentry(sb, objectid, root_objectid, generation);
+}
+
+static struct dentry *btrfs_fh_to_dentry(struct super_block *sb, struct fid *fh,
+					 int fh_len, int fh_type)
+{
+	struct btrfs_fid *fid = (struct btrfs_fid *) fh;
+	u64 objectid, root_objectid;
+	u32 generation;
+
+	if ((fh_type != FILEID_BTRFS_WITH_PARENT ||
+	     fh_len < BTRFS_FID_SIZE_CONNECTABLE) &&
+	    (fh_type != FILEID_BTRFS_WITH_PARENT_ROOT ||
+	     fh_len < BTRFS_FID_SIZE_CONNECTABLE_ROOT) &&
+	    (fh_type != FILEID_BTRFS_WITHOUT_PARENT ||
+	     fh_len < BTRFS_FID_SIZE_NON_CONNECTABLE))
+		return NULL;
+
+	objectid = fid->objectid;
+	root_objectid = fid->root_objectid;
+	generation = fid->gen;
+
+	return btrfs_get_dentry(sb, objectid, root_objectid, generation);
+}
+
+struct dentry *btrfs_get_parent(struct dentry *child)
+{
+	struct btrfs_inode *dir = BTRFS_I(d_inode(child));
+	struct btrfs_inode *inode;
+	struct btrfs_root *root = dir->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_root_ref *ref;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return ERR_PTR(-ENOMEM);
+
+	if (btrfs_ino(dir) == BTRFS_FIRST_FREE_OBJECTID) {
+		key.objectid = btrfs_root_id(root);
+		key.type = BTRFS_ROOT_BACKREF_KEY;
+		key.offset = (u64)-1;
+		root = fs_info->tree_root;
+	} else {
+		key.objectid = btrfs_ino(dir);
+		key.type = BTRFS_INODE_REF_KEY;
+		key.offset = (u64)-1;
+	}
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto fail;
+	if (unlikely(ret == 0)) {
+		/*
+		 * Key with offset of -1 found, there would have to exist an
+		 * inode with such number or a root with such id.
+		 */
+		ret = -EUCLEAN;
+		goto fail;
+	}
+
+	if (path->slots[0] == 0) {
+		ret = -ENOENT;
+		goto fail;
+	}
+
+	path->slots[0]--;
+	leaf = path->nodes[0];
+
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+	if (found_key.objectid != key.objectid || found_key.type != key.type) {
+		ret = -ENOENT;
+		goto fail;
+	}
+
+	if (found_key.type == BTRFS_ROOT_BACKREF_KEY) {
+		ref = btrfs_item_ptr(leaf, path->slots[0],
+				     struct btrfs_root_ref);
+		key.objectid = btrfs_root_ref_dirid(leaf, ref);
+	} else {
+		key.objectid = found_key.offset;
+	}
+	btrfs_free_path(path);
+
+	if (found_key.type == BTRFS_ROOT_BACKREF_KEY) {
+		return btrfs_get_dentry(fs_info->sb, key.objectid,
+					found_key.offset, 0);
+	}
+
+	inode = btrfs_iget(key.objectid, root);
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	return d_obtain_alias(&inode->vfs_inode);
+fail:
+	btrfs_free_path(path);
+	return ERR_PTR(ret);
+}
+
+static int btrfs_get_name(struct dentry *parent, char *name,
+			  struct dentry *child)
+{
+	struct btrfs_inode *inode = BTRFS_I(d_inode(child));
+	struct btrfs_inode *dir = BTRFS_I(d_inode(parent));
+	struct btrfs_root *root = dir->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	BTRFS_PATH_AUTO_FREE(path);
+	struct btrfs_inode_ref *iref;
+	struct btrfs_root_ref *rref;
+	struct extent_buffer *leaf;
+	unsigned long name_ptr;
+	struct btrfs_key key;
+	int name_len;
+	int ret;
+	u64 ino;
+
+	if (!S_ISDIR(dir->vfs_inode.i_mode))
+		return -EINVAL;
+
+	ino = btrfs_ino(inode);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
+		key.objectid = btrfs_root_id(inode->root);
+		key.type = BTRFS_ROOT_BACKREF_KEY;
+		key.offset = (u64)-1;
+		root = fs_info->tree_root;
+	} else {
+		key.objectid = ino;
+		key.type = BTRFS_INODE_REF_KEY;
+		key.offset = btrfs_ino(dir);
+	}
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0) {
+		return ret;
+	} else if (ret > 0) {
+		if (ino == BTRFS_FIRST_FREE_OBJECTID)
+			path->slots[0]--;
+		else
+			return -ENOENT;
+	}
+	leaf = path->nodes[0];
+
+	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
+		rref = btrfs_item_ptr(leaf, path->slots[0],
+				     struct btrfs_root_ref);
+		name_ptr = (unsigned long)(rref + 1);
+		name_len = btrfs_root_ref_name_len(leaf, rref);
+	} else {
+		iref = btrfs_item_ptr(leaf, path->slots[0],
+				      struct btrfs_inode_ref);
+		name_ptr = (unsigned long)(iref + 1);
+		name_len = btrfs_inode_ref_name_len(leaf, iref);
+	}
+
+	read_extent_buffer(leaf, name, name_ptr, name_len);
+
+	/*
+	 * have to add the null termination to make sure that reconnect_path
+	 * gets the right len for strlen
+	 */
+	name[name_len] = '\0';
+
+	return 0;
+}
+
+const struct export_operations btrfs_export_ops = {
+	.encode_fh	= btrfs_encode_fh,
+	.fh_to_dentry	= btrfs_fh_to_dentry,
+	.fh_to_parent	= btrfs_fh_to_parent,
+	.get_parent	= btrfs_get_parent,
+	.get_name	= btrfs_get_name,
+};