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, 124 insertions, 0 deletions

diff --git a/drivers/rpmsg/rpmsg_ns.c b/drivers/rpmsg/rpmsg_ns.c
new file mode 100644
index 000000000..bde8c8d43
--- /dev/null
+++ b/drivers/rpmsg/rpmsg_ns.c
@@ -0,0 +1,124 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) STMicroelectronics 2020 - All Rights Reserved
+ */
+#include <linux/device.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/rpmsg.h>
+#include <linux/rpmsg/ns.h>
+#include <linux/slab.h>
+
+#include "rpmsg_internal.h"
+
+/**
+ * rpmsg_ns_register_device() - register name service device based on rpdev
+ * @rpdev: prepared rpdev to be used for creating endpoints
+ *
+ * This function wraps rpmsg_register_device() preparing the rpdev for use as
+ * basis for the rpmsg name service device.
+ */
+int rpmsg_ns_register_device(struct rpmsg_device *rpdev)
+{
+	rpdev->src = RPMSG_NS_ADDR;
+	rpdev->dst = RPMSG_NS_ADDR;
+
+	return rpmsg_register_device_override(rpdev, "rpmsg_ns");
+}
+EXPORT_SYMBOL(rpmsg_ns_register_device);
+
+/* invoked when a name service announcement arrives */
+static int rpmsg_ns_cb(struct rpmsg_device *rpdev, void *data, int len,
+		       void *priv, u32 src)
+{
+	struct rpmsg_ns_msg *msg = data;
+	struct rpmsg_device *newch;
+	struct rpmsg_channel_info chinfo;
+	struct device *dev = rpdev->dev.parent;
+	int ret;
+
+#if defined(CONFIG_DYNAMIC_DEBUG)
+	dynamic_hex_dump("NS announcement: ", DUMP_PREFIX_NONE, 16, 1,
+			 data, len, true);
+#endif
+
+	if (len != sizeof(*msg)) {
+		dev_err(dev, "malformed ns msg (%d)\n", len);
+		return -EINVAL;
+	}
+
+	/* don't trust the remote processor for null terminating the name */
+	msg->name[RPMSG_NAME_SIZE - 1] = '\0';
+
+	strscpy_pad(chinfo.name, msg->name, sizeof(chinfo.name));
+	chinfo.src = RPMSG_ADDR_ANY;
+	chinfo.dst = rpmsg32_to_cpu(rpdev, msg->addr);
+
+	dev_info(dev, "%sing channel %s addr 0x%x\n",
+		 rpmsg32_to_cpu(rpdev, msg->flags) & RPMSG_NS_DESTROY ?
+		 "destroy" : "creat", msg->name, chinfo.dst);
+
+	if (rpmsg32_to_cpu(rpdev, msg->flags) & RPMSG_NS_DESTROY) {
+		ret = rpmsg_release_channel(rpdev, &chinfo);
+		if (ret)
+			dev_err(dev, "rpmsg_destroy_channel failed: %d\n", ret);
+	} else {
+		newch = rpmsg_create_channel(rpdev, &chinfo);
+		if (!newch)
+			dev_err(dev, "rpmsg_create_channel failed\n");
+	}
+
+	return 0;
+}
+
+static int rpmsg_ns_probe(struct rpmsg_device *rpdev)
+{
+	struct rpmsg_endpoint *ns_ept;
+	struct rpmsg_channel_info ns_chinfo = {
+		.src = RPMSG_NS_ADDR,
+		.dst = RPMSG_NS_ADDR,
+		.name = "name_service",
+	};
+
+	/*
+	 * Create the NS announcement service endpoint associated to the RPMsg
+	 * device. The endpoint will be automatically destroyed when the RPMsg
+	 * device will be deleted.
+	 */
+	ns_ept = rpmsg_create_ept(rpdev, rpmsg_ns_cb, NULL, ns_chinfo);
+	if (!ns_ept) {
+		dev_err(&rpdev->dev, "failed to create the ns ept\n");
+		return -ENOMEM;
+	}
+	rpdev->ept = ns_ept;
+
+	return 0;
+}
+
+static struct rpmsg_driver rpmsg_ns_driver = {
+	.drv.name = KBUILD_MODNAME,
+	.probe = rpmsg_ns_probe,
+};
+
+static int rpmsg_ns_init(void)
+{
+	int ret;
+
+	ret = register_rpmsg_driver(&rpmsg_ns_driver);
+	if (ret < 0)
+		pr_err("%s: Failed to register rpmsg driver\n", __func__);
+
+	return ret;
+}
+postcore_initcall(rpmsg_ns_init);
+
+static void rpmsg_ns_exit(void)
+{
+	unregister_rpmsg_driver(&rpmsg_ns_driver);
+}
+module_exit(rpmsg_ns_exit);
+
+MODULE_DESCRIPTION("Name service announcement rpmsg driver");
+MODULE_AUTHOR("Arnaud Pouliquen <arnaud.pouliquen@st.com>");
+MODULE_ALIAS("rpmsg:" KBUILD_MODNAME);
+MODULE_LICENSE("GPL v2");