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, 99 insertions, 0 deletions

diff --git a/fs/nilfs2/alloc.h b/fs/nilfs2/alloc.h
new file mode 100644
index 000000000..046d876ea
--- /dev/null
+++ b/fs/nilfs2/alloc.h
@@ -0,0 +1,99 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Persistent object (dat entry/disk inode) allocator/deallocator
+ *
+ * Copyright (C) 2006-2008 Nippon Telegraph and Telephone Corporation.
+ *
+ * Originally written by Koji Sato.
+ * Two allocators were unified by Ryusuke Konishi and Amagai Yoshiji.
+ */
+
+#ifndef _NILFS_ALLOC_H
+#define _NILFS_ALLOC_H
+
+#include <linux/types.h>
+#include <linux/buffer_head.h>
+#include <linux/fs.h>
+
+/**
+ * nilfs_palloc_entries_per_group - get the number of entries per group
+ * @inode: inode of metadata file using this allocator
+ *
+ * The number of entries per group is defined by the number of bits
+ * that a bitmap block can maintain.
+ *
+ * Return: Number of entries per group.
+ */
+static inline unsigned long
+nilfs_palloc_entries_per_group(const struct inode *inode)
+{
+	return 1UL << (inode->i_blkbits + 3 /* log2(8 = CHAR_BITS) */);
+}
+
+int nilfs_palloc_init_blockgroup(struct inode *, unsigned int);
+int nilfs_palloc_get_entry_block(struct inode *, __u64, int,
+				 struct buffer_head **);
+size_t nilfs_palloc_entry_offset(const struct inode *inode, __u64 nr,
+				 const struct buffer_head *bh);
+
+int nilfs_palloc_count_max_entries(struct inode *, u64, u64 *);
+
+/**
+ * struct nilfs_palloc_req - persistent allocator request and reply
+ * @pr_entry_nr: entry number (vblocknr or inode number)
+ * @pr_desc_bh: buffer head of the buffer containing block group descriptors
+ * @pr_bitmap_bh: buffer head of the buffer containing a block group bitmap
+ * @pr_entry_bh: buffer head of the buffer containing translation entries
+ */
+struct nilfs_palloc_req {
+	__u64 pr_entry_nr;
+	struct buffer_head *pr_desc_bh;
+	struct buffer_head *pr_bitmap_bh;
+	struct buffer_head *pr_entry_bh;
+};
+
+int nilfs_palloc_prepare_alloc_entry(struct inode *inode,
+				     struct nilfs_palloc_req *req, bool wrap);
+void nilfs_palloc_commit_alloc_entry(struct inode *,
+				     struct nilfs_palloc_req *);
+void nilfs_palloc_abort_alloc_entry(struct inode *, struct nilfs_palloc_req *);
+void nilfs_palloc_commit_free_entry(struct inode *, struct nilfs_palloc_req *);
+int nilfs_palloc_prepare_free_entry(struct inode *, struct nilfs_palloc_req *);
+void nilfs_palloc_abort_free_entry(struct inode *, struct nilfs_palloc_req *);
+int nilfs_palloc_freev(struct inode *, __u64 *, size_t);
+
+#define nilfs_set_bit_atomic		ext2_set_bit_atomic
+#define nilfs_clear_bit_atomic		ext2_clear_bit_atomic
+#define nilfs_find_next_zero_bit	find_next_zero_bit_le
+#define nilfs_find_next_bit		find_next_bit_le
+
+/**
+ * struct nilfs_bh_assoc - block offset and buffer head association
+ * @blkoff: block offset
+ * @bh: buffer head
+ */
+struct nilfs_bh_assoc {
+	unsigned long blkoff;
+	struct buffer_head *bh;
+};
+
+/**
+ * struct nilfs_palloc_cache - persistent object allocator cache
+ * @lock: cache protecting lock
+ * @prev_desc: blockgroup descriptors cache
+ * @prev_bitmap: blockgroup bitmap cache
+ * @prev_entry: translation entries cache
+ */
+struct nilfs_palloc_cache {
+	spinlock_t lock;
+	struct nilfs_bh_assoc prev_desc;
+	struct nilfs_bh_assoc prev_bitmap;
+	struct nilfs_bh_assoc prev_entry;
+};
+
+void nilfs_palloc_setup_cache(struct inode *inode,
+			      struct nilfs_palloc_cache *cache);
+void nilfs_palloc_clear_cache(struct inode *inode);
+void nilfs_palloc_destroy_cache(struct inode *inode);
+
+#endif	/* _NILFS_ALLOC_H */