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, 145 insertions, 0 deletions

diff --git a/fs/nfs_common/grace.c b/fs/nfs_common/grace.c
new file mode 100644
index 000000000..27cd0d131
--- /dev/null
+++ b/fs/nfs_common/grace.c
@@ -0,0 +1,145 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Common code for control of lockd and nfsv4 grace periods.
+ *
+ * Transplanted from lockd code
+ */
+
+#include <linux/module.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
+#include <linux/fs.h>
+#include <linux/filelock.h>
+
+static unsigned int grace_net_id;
+static DEFINE_SPINLOCK(grace_lock);
+
+/**
+ * locks_start_grace
+ * @net: net namespace that this lock manager belongs to
+ * @lm: who this grace period is for
+ *
+ * A grace period is a period during which locks should not be given
+ * out.  Currently grace periods are only enforced by the two lock
+ * managers (lockd and nfsd), using the locks_in_grace() function to
+ * check when they are in a grace period.
+ *
+ * This function is called to start a grace period.
+ */
+void
+locks_start_grace(struct net *net, struct lock_manager *lm)
+{
+	struct list_head *grace_list = net_generic(net, grace_net_id);
+
+	spin_lock(&grace_lock);
+	if (list_empty(&lm->list))
+		list_add(&lm->list, grace_list);
+	else
+		WARN(1, "double list_add attempt detected in net %x %s\n",
+		     net->ns.inum, (net == &init_net) ? "(init_net)" : "");
+	spin_unlock(&grace_lock);
+}
+EXPORT_SYMBOL_GPL(locks_start_grace);
+
+/**
+ * locks_end_grace
+ * @lm: who this grace period is for
+ *
+ * Call this function to state that the given lock manager is ready to
+ * resume regular locking.  The grace period will not end until all lock
+ * managers that called locks_start_grace() also call locks_end_grace().
+ * Note that callers count on it being safe to call this more than once,
+ * and the second call should be a no-op.
+ */
+void
+locks_end_grace(struct lock_manager *lm)
+{
+	spin_lock(&grace_lock);
+	list_del_init(&lm->list);
+	spin_unlock(&grace_lock);
+}
+EXPORT_SYMBOL_GPL(locks_end_grace);
+
+static bool
+__state_in_grace(struct net *net, bool open)
+{
+	struct list_head *grace_list = net_generic(net, grace_net_id);
+	struct lock_manager *lm;
+
+	if (!open)
+		return !list_empty(grace_list);
+
+	spin_lock(&grace_lock);
+	list_for_each_entry(lm, grace_list, list) {
+		if (lm->block_opens) {
+			spin_unlock(&grace_lock);
+			return true;
+		}
+	}
+	spin_unlock(&grace_lock);
+	return false;
+}
+
+/**
+ * locks_in_grace
+ * @net: network namespace
+ *
+ * Lock managers call this function to determine when it is OK for them
+ * to answer ordinary lock requests, and when they should accept only
+ * lock reclaims.
+ */
+bool locks_in_grace(struct net *net)
+{
+	return __state_in_grace(net, false);
+}
+EXPORT_SYMBOL_GPL(locks_in_grace);
+
+bool opens_in_grace(struct net *net)
+{
+	return __state_in_grace(net, true);
+}
+EXPORT_SYMBOL_GPL(opens_in_grace);
+
+static int __net_init
+grace_init_net(struct net *net)
+{
+	struct list_head *grace_list = net_generic(net, grace_net_id);
+
+	INIT_LIST_HEAD(grace_list);
+	return 0;
+}
+
+static void __net_exit
+grace_exit_net(struct net *net)
+{
+	struct list_head *grace_list = net_generic(net, grace_net_id);
+
+	WARN_ONCE(!list_empty(grace_list),
+		  "net %x %s: grace_list is not empty\n",
+		  net->ns.inum, __func__);
+}
+
+static struct pernet_operations grace_net_ops = {
+	.init = grace_init_net,
+	.exit = grace_exit_net,
+	.id   = &grace_net_id,
+	.size = sizeof(struct list_head),
+};
+
+static int __init
+init_grace(void)
+{
+	return register_pernet_subsys(&grace_net_ops);
+}
+
+static void __exit
+exit_grace(void)
+{
+	unregister_pernet_subsys(&grace_net_ops);
+}
+
+MODULE_AUTHOR("Jeff Layton <jlayton@primarydata.com>");
+MODULE_DESCRIPTION("NFS client and server infrastructure");
+MODULE_LICENSE("GPL");
+module_init(init_grace)
+module_exit(exit_grace)