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, 220 insertions, 0 deletions

diff --git a/fs/afs/dir_search.c b/fs/afs/dir_search.c
new file mode 100644
index 000000000..11ebdfffc
--- /dev/null
+++ b/fs/afs/dir_search.c
@@ -0,0 +1,220 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Search a directory's hash table.
+ *
+ * Copyright (C) 2024 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ *
+ * https://tools.ietf.org/html/draft-keiser-afs3-directory-object-00
+ */
+
+#include <linux/kernel.h>
+#include <linux/fs.h>
+#include <linux/namei.h>
+#include <linux/iversion.h>
+#include "internal.h"
+#include "afs_fs.h"
+#include "xdr_fs.h"
+
+/*
+ * Calculate the name hash.
+ */
+unsigned int afs_dir_hash_name(const struct qstr *name)
+{
+	const unsigned char *p = name->name;
+	unsigned int hash = 0, i;
+	int bucket;
+
+	for (i = 0; i < name->len; i++)
+		hash = (hash * 173) + p[i];
+	bucket = hash & (AFS_DIR_HASHTBL_SIZE - 1);
+	if (hash > INT_MAX) {
+		bucket = AFS_DIR_HASHTBL_SIZE - bucket;
+		bucket &= (AFS_DIR_HASHTBL_SIZE - 1);
+	}
+	return bucket;
+}
+
+/*
+ * Reset a directory iterator.
+ */
+static bool afs_dir_reset_iter(struct afs_dir_iter *iter)
+{
+	unsigned long long i_size = i_size_read(&iter->dvnode->netfs.inode);
+	unsigned int nblocks;
+
+	/* Work out the maximum number of steps we can take. */
+	nblocks = umin(i_size / AFS_DIR_BLOCK_SIZE, AFS_DIR_MAX_BLOCKS);
+	if (!nblocks)
+		return false;
+	iter->loop_check = nblocks * (AFS_DIR_SLOTS_PER_BLOCK - AFS_DIR_RESV_BLOCKS);
+	iter->prev_entry = 0; /* Hash head is previous */
+	return true;
+}
+
+/*
+ * Initialise a directory iterator for looking up a name.
+ */
+bool afs_dir_init_iter(struct afs_dir_iter *iter, const struct qstr *name)
+{
+	iter->nr_slots = afs_dir_calc_slots(name->len);
+	iter->bucket = afs_dir_hash_name(name);
+	return afs_dir_reset_iter(iter);
+}
+
+/*
+ * Get a specific block.
+ */
+union afs_xdr_dir_block *afs_dir_find_block(struct afs_dir_iter *iter, size_t block)
+{
+	struct folio_queue *fq = iter->fq;
+	struct afs_vnode *dvnode = iter->dvnode;
+	struct folio *folio;
+	size_t blpos = block * AFS_DIR_BLOCK_SIZE;
+	size_t blend = (block + 1) * AFS_DIR_BLOCK_SIZE, fpos = iter->fpos;
+	int slot = iter->fq_slot;
+
+	_enter("%zx,%d", block, slot);
+
+	afs_dir_end_iter(iter);
+
+	if (dvnode->directory_size < blend)
+		goto fail;
+
+	if (!fq || blpos < fpos) {
+		fq = dvnode->directory;
+		slot = 0;
+		fpos = 0;
+	}
+
+	/* Search the folio queue for the folio containing the block... */
+	for (; fq; fq = fq->next) {
+		for (; slot < folioq_count(fq); slot++) {
+			size_t fsize = folioq_folio_size(fq, slot);
+
+			if (blend <= fpos + fsize) {
+				/* ... and then return the mapped block. */
+				folio = folioq_folio(fq, slot);
+				if (WARN_ON_ONCE(folio_pos(folio) != fpos))
+					goto fail;
+				iter->fq = fq;
+				iter->fq_slot = slot;
+				iter->fpos = fpos;
+				iter->block = kmap_local_folio(folio, blpos - fpos);
+				return iter->block;
+			}
+			fpos += fsize;
+		}
+		slot = 0;
+	}
+
+fail:
+	iter->fq = NULL;
+	iter->fq_slot = 0;
+	afs_invalidate_dir(dvnode, afs_dir_invalid_edit_get_block);
+	return NULL;
+}
+
+/*
+ * Search through a directory bucket.
+ */
+int afs_dir_search_bucket(struct afs_dir_iter *iter, const struct qstr *name,
+			  struct afs_fid *_fid)
+{
+	const union afs_xdr_dir_block *meta;
+	unsigned int entry;
+	int ret = -ESTALE;
+
+	meta = afs_dir_find_block(iter, 0);
+	if (!meta)
+		return -ESTALE;
+
+	entry = ntohs(meta->meta.hashtable[iter->bucket & (AFS_DIR_HASHTBL_SIZE - 1)]);
+	_enter("%x,%x", iter->bucket, entry);
+
+	while (entry) {
+		const union afs_xdr_dir_block *block;
+		const union afs_xdr_dirent *dire;
+		unsigned int blnum = entry / AFS_DIR_SLOTS_PER_BLOCK;
+		unsigned int slot = entry % AFS_DIR_SLOTS_PER_BLOCK;
+		unsigned int resv = (blnum == 0 ? AFS_DIR_RESV_BLOCKS0 : AFS_DIR_RESV_BLOCKS);
+
+		_debug("search %x", entry);
+
+		if (slot < resv) {
+			kdebug("slot out of range h=%x rs=%2x sl=%2x-%2x",
+			       iter->bucket, resv, slot, slot + iter->nr_slots - 1);
+			goto bad;
+		}
+
+		block = afs_dir_find_block(iter, blnum);
+		if (!block)
+			goto bad;
+		dire = &block->dirents[slot];
+
+		if (slot + iter->nr_slots <= AFS_DIR_SLOTS_PER_BLOCK &&
+		    memcmp(dire->u.name, name->name, name->len) == 0 &&
+		    dire->u.name[name->len] == '\0') {
+			_fid->vnode  = ntohl(dire->u.vnode);
+			_fid->unique = ntohl(dire->u.unique);
+			ret = entry;
+			goto found;
+		}
+
+		iter->prev_entry = entry;
+		entry = ntohs(dire->u.hash_next);
+		if (!--iter->loop_check) {
+			kdebug("dir chain loop h=%x", iter->bucket);
+			goto bad;
+		}
+	}
+
+	ret = -ENOENT;
+found:
+bad:
+	afs_dir_end_iter(iter);
+	if (ret == -ESTALE)
+		afs_invalidate_dir(iter->dvnode, afs_dir_invalid_iter_stale);
+	_leave(" = %d", ret);
+	return ret;
+}
+
+/*
+ * Search the appropriate hash chain in the contents of an AFS directory.
+ */
+int afs_dir_search(struct afs_vnode *dvnode, const struct qstr *name,
+		   struct afs_fid *_fid, afs_dataversion_t *_dir_version)
+{
+	struct afs_dir_iter iter = { .dvnode = dvnode, };
+	int ret, retry_limit = 3;
+
+	_enter("{%llu},,,", dvnode->netfs.inode.i_ino);
+
+	if (!afs_dir_init_iter(&iter, name))
+		return -ENOENT;
+	do {
+		if (--retry_limit < 0) {
+			pr_warn("afs_read_dir(): Too many retries\n");
+			ret = -ESTALE;
+			break;
+		}
+		ret = afs_read_dir(dvnode, NULL);
+		if (ret < 0) {
+			if (ret != -ESTALE)
+				break;
+			if (test_bit(AFS_VNODE_DELETED, &dvnode->flags)) {
+				ret = -ESTALE;
+				break;
+			}
+			continue;
+		}
+		*_dir_version = inode_peek_iversion_raw(&dvnode->netfs.inode);
+
+		ret = afs_dir_search_bucket(&iter, name, _fid);
+		up_read(&dvnode->validate_lock);
+		if (ret == -ESTALE)
+			afs_dir_reset_iter(&iter);
+	} while (ret == -ESTALE);
+
+	_leave(" = %d", ret);
+	return ret;
+}