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, 149 insertions, 0 deletions

diff --git a/fs/dlm/requestqueue.c b/fs/dlm/requestqueue.c
new file mode 100644
index 000000000..719a5243a
--- /dev/null
+++ b/fs/dlm/requestqueue.c
@@ -0,0 +1,149 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "member.h"
+#include "lock.h"
+#include "dir.h"
+#include "config.h"
+#include "requestqueue.h"
+#include "util.h"
+
+struct rq_entry {
+	struct list_head list;
+	uint32_t recover_seq;
+	int nodeid;
+	struct dlm_message request;
+};
+
+/*
+ * Requests received while the lockspace is in recovery get added to the
+ * request queue and processed when recovery is complete.  This happens when
+ * the lockspace is suspended on some nodes before it is on others, or the
+ * lockspace is enabled on some while still suspended on others.
+ */
+
+void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
+			  const struct dlm_message *ms)
+{
+	struct rq_entry *e;
+	int length = le16_to_cpu(ms->m_header.h_length) -
+		sizeof(struct dlm_message);
+
+	e = kmalloc(sizeof(struct rq_entry) + length, GFP_ATOMIC);
+	if (!e) {
+		log_print("dlm_add_requestqueue: out of memory len %d", length);
+		return;
+	}
+
+	e->recover_seq = ls->ls_recover_seq & 0xFFFFFFFF;
+	e->nodeid = nodeid;
+	memcpy(&e->request, ms, sizeof(*ms));
+	memcpy(&e->request.m_extra, ms->m_extra, length);
+
+	list_add_tail(&e->list, &ls->ls_requestqueue);
+}
+
+/*
+ * Called by dlm_recoverd to process normal messages saved while recovery was
+ * happening.  Normal locking has been enabled before this is called.  dlm_recv
+ * upon receiving a message, will wait for all saved messages to be drained
+ * here before processing the message it got.  If a new dlm_ls_stop() arrives
+ * while we're processing these saved messages, it may block trying to suspend
+ * dlm_recv if dlm_recv is waiting for us in dlm_wait_requestqueue.  In that
+ * case, we don't abort since locking_stopped is still 0.  If dlm_recv is not
+ * waiting for us, then this processing may be aborted due to locking_stopped.
+ */
+
+int dlm_process_requestqueue(struct dlm_ls *ls)
+{
+	struct rq_entry *e;
+	struct dlm_message *ms;
+	int error = 0;
+
+	write_lock_bh(&ls->ls_requestqueue_lock);
+	for (;;) {
+		if (list_empty(&ls->ls_requestqueue)) {
+			clear_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags);
+			error = 0;
+			break;
+		}
+		e = list_first_entry(&ls->ls_requestqueue, struct rq_entry, list);
+
+		ms = &e->request;
+
+		log_limit(ls, "dlm_process_requestqueue msg %d from %d "
+			  "lkid %x remid %x result %d seq %u",
+			  le32_to_cpu(ms->m_type),
+			  le32_to_cpu(ms->m_header.h_nodeid),
+			  le32_to_cpu(ms->m_lkid), le32_to_cpu(ms->m_remid),
+			  from_dlm_errno(le32_to_cpu(ms->m_result)),
+			  e->recover_seq);
+
+		dlm_receive_message_saved(ls, &e->request, e->recover_seq);
+		list_del(&e->list);
+		kfree(e);
+
+		if (dlm_locking_stopped(ls)) {
+			log_debug(ls, "process_requestqueue abort running");
+			error = -EINTR;
+			break;
+		}
+		write_unlock_bh(&ls->ls_requestqueue_lock);
+		schedule();
+		write_lock_bh(&ls->ls_requestqueue_lock);
+	}
+	write_unlock_bh(&ls->ls_requestqueue_lock);
+
+	return error;
+}
+
+static int purge_request(struct dlm_ls *ls, struct dlm_message *ms, int nodeid)
+{
+	__le32 type = ms->m_type;
+
+	/* the ls is being cleaned up and freed by release_lockspace */
+	if (!atomic_read(&ls->ls_count))
+		return 1;
+
+	if (dlm_is_removed(ls, nodeid))
+		return 1;
+
+	/* directory operations are always purged because the directory is
+	   always rebuilt during recovery and the lookups resent */
+
+	if (type == cpu_to_le32(DLM_MSG_REMOVE) ||
+	    type == cpu_to_le32(DLM_MSG_LOOKUP) ||
+	    type == cpu_to_le32(DLM_MSG_LOOKUP_REPLY))
+		return 1;
+
+	if (!dlm_no_directory(ls))
+		return 0;
+
+	return 1;
+}
+
+void dlm_purge_requestqueue(struct dlm_ls *ls)
+{
+	struct dlm_message *ms;
+	struct rq_entry *e, *safe;
+
+	write_lock_bh(&ls->ls_requestqueue_lock);
+	list_for_each_entry_safe(e, safe, &ls->ls_requestqueue, list) {
+		ms =  &e->request;
+
+		if (purge_request(ls, ms, e->nodeid)) {
+			list_del(&e->list);
+			kfree(e);
+		}
+	}
+	write_unlock_bh(&ls->ls_requestqueue_lock);
+}
+