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, 154 insertions, 0 deletions

diff --git a/drivers/md/persistent-data/dm-transaction-manager.h b/drivers/md/persistent-data/dm-transaction-manager.h
new file mode 100644
index 000000000..61a8d1082
--- /dev/null
+++ b/drivers/md/persistent-data/dm-transaction-manager.h
@@ -0,0 +1,154 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2011 Red Hat, Inc.
+ *
+ * This file is released under the GPL.
+ */
+
+#ifndef _LINUX_DM_TRANSACTION_MANAGER_H
+#define _LINUX_DM_TRANSACTION_MANAGER_H
+
+#include "dm-block-manager.h"
+
+struct dm_transaction_manager;
+struct dm_space_map;
+
+/*----------------------------------------------------------------*/
+
+/*
+ * This manages the scope of a transaction.  It also enforces immutability
+ * of the on-disk data structures by limiting access to writeable blocks.
+ *
+ * Clients should not fiddle with the block manager directly.
+ */
+
+void dm_tm_destroy(struct dm_transaction_manager *tm);
+
+/*
+ * The non-blocking version of a transaction manager is intended for use in
+ * fast path code that needs to do lookups e.g. a dm mapping function.
+ * You create the non-blocking variant from a normal tm.  The interface is
+ * the same, except that most functions will just return -EWOULDBLOCK.
+ * Methods that return void yet may block should not be called on a clone
+ * viz. dm_tm_inc, dm_tm_dec.  Call dm_tm_destroy() as you would with a normal
+ * tm when you've finished with it.  You may not destroy the original prior
+ * to clones.
+ */
+struct dm_transaction_manager *dm_tm_create_non_blocking_clone(struct dm_transaction_manager *real);
+
+/*
+ * We use a 2-phase commit here.
+ *
+ * i) Make all changes for the transaction *except* for the superblock.
+ * Then call dm_tm_pre_commit() to flush them to disk.
+ *
+ * ii) Lock your superblock.  Update.  Then call dm_tm_commit() which will
+ * unlock the superblock and flush it.  No other blocks should be updated
+ * during this period.  Care should be taken to never unlock a partially
+ * updated superblock; perform any operations that could fail *before* you
+ * take the superblock lock.
+ */
+int dm_tm_pre_commit(struct dm_transaction_manager *tm);
+int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *superblock);
+
+/*
+ * These methods are the only way to get hold of a writeable block.
+ */
+
+/*
+ * dm_tm_new_block() is pretty self-explanatory.  Make sure you do actually
+ * write to the whole of @data before you unlock, otherwise you could get
+ * a data leak.  (The other option is for tm_new_block() to zero new blocks
+ * before handing them out, which will be redundant in most, if not all,
+ * cases).
+ * Zeroes the new block and returns with write lock held.
+ */
+int dm_tm_new_block(struct dm_transaction_manager *tm,
+		    const struct dm_block_validator *v,
+		    struct dm_block **result);
+
+/*
+ * dm_tm_shadow_block() allocates a new block and copies the data from @orig
+ * to it.  It then decrements the reference count on original block.  Use
+ * this to update the contents of a block in a data structure, don't
+ * confuse this with a clone - you shouldn't access the orig block after
+ * this operation.  Because the tm knows the scope of the transaction it
+ * can optimise requests for a shadow of a shadow to a no-op.  Don't forget
+ * to unlock when you've finished with the shadow.
+ *
+ * The @inc_children flag is used to tell the caller whether it needs to
+ * adjust reference counts for children.  (Data in the block may refer to
+ * other blocks.)
+ *
+ * Shadowing implicitly drops a reference on @orig so you must not have
+ * it locked when you call this.
+ */
+int dm_tm_shadow_block(struct dm_transaction_manager *tm, dm_block_t orig,
+		       const struct dm_block_validator *v,
+		       struct dm_block **result, int *inc_children);
+
+/*
+ * Read access.  You can lock any block you want.  If there's a write lock
+ * on it outstanding then it'll block.
+ */
+int dm_tm_read_lock(struct dm_transaction_manager *tm, dm_block_t b,
+		    const struct dm_block_validator *v,
+		    struct dm_block **result);
+
+void dm_tm_unlock(struct dm_transaction_manager *tm, struct dm_block *b);
+
+/*
+ * Functions for altering the reference count of a block directly.
+ */
+void dm_tm_inc(struct dm_transaction_manager *tm, dm_block_t b);
+void dm_tm_inc_range(struct dm_transaction_manager *tm, dm_block_t b, dm_block_t e);
+void dm_tm_dec(struct dm_transaction_manager *tm, dm_block_t b);
+void dm_tm_dec_range(struct dm_transaction_manager *tm, dm_block_t b, dm_block_t e);
+
+/*
+ * Builds up runs of adjacent blocks, and then calls the given fn
+ * (typically dm_tm_inc/dec).  Very useful when you have to perform
+ * the same tm operation on all values in a btree leaf.
+ */
+typedef void (*dm_tm_run_fn)(struct dm_transaction_manager *, dm_block_t, dm_block_t);
+void dm_tm_with_runs(struct dm_transaction_manager *tm,
+		     const __le64 *value_le, unsigned int count, dm_tm_run_fn fn);
+
+int dm_tm_ref(struct dm_transaction_manager *tm, dm_block_t b, uint32_t *result);
+
+/*
+ * Finds out if a given block is shared (ie. has a reference count higher
+ * than one).
+ */
+int dm_tm_block_is_shared(struct dm_transaction_manager *tm, dm_block_t b,
+			  int *result);
+
+struct dm_block_manager *dm_tm_get_bm(struct dm_transaction_manager *tm);
+
+/*
+ * If you're using a non-blocking clone the tm will build up a list of
+ * requested blocks that weren't in core.  This call will request those
+ * blocks to be prefetched.
+ */
+void dm_tm_issue_prefetches(struct dm_transaction_manager *tm);
+
+/*
+ * A little utility that ties the knot by producing a transaction manager
+ * that has a space map managed by the transaction manager...
+ *
+ * Returns a tm that has an open transaction to write the new disk sm.
+ * Caller should store the new sm root and commit.
+ *
+ * The superblock location is passed so the metadata space map knows it
+ * shouldn't be used.
+ */
+int dm_tm_create_with_sm(struct dm_block_manager *bm, dm_block_t sb_location,
+			 struct dm_transaction_manager **tm,
+			 struct dm_space_map **sm);
+
+int dm_tm_open_with_sm(struct dm_block_manager *bm, dm_block_t sb_location,
+		       void *sm_root, size_t root_len,
+		       struct dm_transaction_manager **tm,
+		       struct dm_space_map **sm);
+
+#endif	/* _LINUX_DM_TRANSACTION_MANAGER_H */