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, 141 insertions, 0 deletions

diff --git a/drivers/pps/clients/pps-ldisc.c b/drivers/pps/clients/pps-ldisc.c
new file mode 100644
index 000000000..fa5660f3c
--- /dev/null
+++ b/drivers/pps/clients/pps-ldisc.c
@@ -0,0 +1,141 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * pps-ldisc.c -- PPS line discipline
+ *
+ * Copyright (C) 2008	Rodolfo Giometti <giometti@linux.it>
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/serial_core.h>
+#include <linux/tty.h>
+#include <linux/pps_kernel.h>
+#include <linux/bug.h>
+
+static void pps_tty_dcd_change(struct tty_struct *tty, bool active)
+{
+	struct pps_device *pps;
+	struct pps_event_time ts;
+
+	pps_get_ts(&ts);
+
+	pps = pps_lookup_dev(tty);
+	/*
+	 * This should never fail, but the ldisc locking is very
+	 * convoluted, so don't crash just in case.
+	 */
+	if (WARN_ON_ONCE(pps == NULL))
+		return;
+
+	/* Now do the PPS event report */
+	pps_event(pps, &ts, active ? PPS_CAPTUREASSERT :
+			PPS_CAPTURECLEAR, NULL);
+
+	dev_dbg(&pps->dev, "PPS %s at %lu\n",
+			active ? "assert" : "clear", jiffies);
+}
+
+static int (*alias_n_tty_open)(struct tty_struct *tty);
+
+static int pps_tty_open(struct tty_struct *tty)
+{
+	struct pps_source_info info;
+	struct tty_driver *drv = tty->driver;
+	int index = tty->index + drv->name_base;
+	struct pps_device *pps;
+	int ret;
+
+	info.owner = THIS_MODULE;
+	info.dev = NULL;
+	snprintf(info.name, PPS_MAX_NAME_LEN, "%s%d", drv->driver_name, index);
+	snprintf(info.path, PPS_MAX_NAME_LEN, "/dev/%s%d", drv->name, index);
+	info.mode = PPS_CAPTUREBOTH | \
+			PPS_OFFSETASSERT | PPS_OFFSETCLEAR | \
+			PPS_CANWAIT | PPS_TSFMT_TSPEC;
+
+	pps = pps_register_source(&info, PPS_CAPTUREBOTH | \
+				PPS_OFFSETASSERT | PPS_OFFSETCLEAR);
+	if (IS_ERR(pps)) {
+		pr_err("cannot register PPS source \"%s\"\n", info.path);
+		return PTR_ERR(pps);
+	}
+	pps->lookup_cookie = tty;
+
+	/* Now open the base class N_TTY ldisc */
+	ret = alias_n_tty_open(tty);
+	if (ret < 0) {
+		pr_err("cannot open tty ldisc \"%s\"\n", info.path);
+		goto err_unregister;
+	}
+
+	dev_dbg(&pps->dev, "source \"%s\" added\n", info.path);
+
+	return 0;
+
+err_unregister:
+	pps_unregister_source(pps);
+	return ret;
+}
+
+static void (*alias_n_tty_close)(struct tty_struct *tty);
+
+static void pps_tty_close(struct tty_struct *tty)
+{
+	struct pps_device *pps = pps_lookup_dev(tty);
+
+	alias_n_tty_close(tty);
+
+	if (WARN_ON(!pps))
+		return;
+
+	dev_info(&pps->dev, "removed\n");
+	pps_unregister_source(pps);
+}
+
+static struct tty_ldisc_ops pps_ldisc_ops;
+
+/*
+ * Module stuff
+ */
+
+static int __init pps_tty_init(void)
+{
+	int err;
+
+	/* Inherit the N_TTY's ops */
+	n_tty_inherit_ops(&pps_ldisc_ops);
+
+	/* Save N_TTY's open()/close() methods */
+	alias_n_tty_open = pps_ldisc_ops.open;
+	alias_n_tty_close = pps_ldisc_ops.close;
+
+	/* Init PPS_TTY data */
+	pps_ldisc_ops.owner = THIS_MODULE;
+	pps_ldisc_ops.num = N_PPS;
+	pps_ldisc_ops.name = "pps_tty";
+	pps_ldisc_ops.dcd_change = pps_tty_dcd_change;
+	pps_ldisc_ops.open = pps_tty_open;
+	pps_ldisc_ops.close = pps_tty_close;
+
+	err = tty_register_ldisc(&pps_ldisc_ops);
+	if (err)
+		pr_err("can't register PPS line discipline\n");
+	else
+		pr_info("PPS line discipline registered\n");
+
+	return err;
+}
+
+static void __exit pps_tty_cleanup(void)
+{
+	tty_unregister_ldisc(&pps_ldisc_ops);
+}
+
+module_init(pps_tty_init);
+module_exit(pps_tty_cleanup);
+
+MODULE_ALIAS_LDISC(N_PPS);
+MODULE_AUTHOR("Rodolfo Giometti <giometti@linux.it>");
+MODULE_DESCRIPTION("PPS TTY device driver");
+MODULE_LICENSE("GPL");