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, 177 insertions, 0 deletions

diff --git a/fs/proc/proc_tty.c b/fs/proc/proc_tty.c
new file mode 100644
index 000000000..5c6a5ceab
--- /dev/null
+++ b/fs/proc/proc_tty.c
@@ -0,0 +1,177 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * proc_tty.c -- handles /proc/tty
+ *
+ * Copyright 1997, Theodore Ts'o
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/time.h>
+#include <linux/proc_fs.h>
+#include <linux/stat.h>
+#include <linux/tty.h>
+#include <linux/seq_file.h>
+#include <linux/bitops.h>
+#include "internal.h"
+
+/*
+ * The /proc/tty directory inodes...
+ */
+static struct proc_dir_entry *proc_tty_driver;
+
+/*
+ * This is the handler for /proc/tty/drivers
+ */
+static void show_tty_range(struct seq_file *m, struct tty_driver *p,
+	dev_t from, int num)
+{
+	seq_printf(m, "%-20s ", p->driver_name ? p->driver_name : "unknown");
+	seq_printf(m, "/dev/%-8s ", p->name);
+	if (p->num > 1) {
+		seq_printf(m, "%3d %d-%d ", MAJOR(from), MINOR(from),
+			MINOR(from) + num - 1);
+	} else {
+		seq_printf(m, "%3d %7d ", MAJOR(from), MINOR(from));
+	}
+	switch (p->type) {
+	case TTY_DRIVER_TYPE_SYSTEM:
+		seq_puts(m, "system");
+		if (p->subtype == SYSTEM_TYPE_TTY)
+			seq_puts(m, ":/dev/tty");
+		else if (p->subtype == SYSTEM_TYPE_SYSCONS)
+			seq_puts(m, ":console");
+		else if (p->subtype == SYSTEM_TYPE_CONSOLE)
+			seq_puts(m, ":vtmaster");
+		break;
+	case TTY_DRIVER_TYPE_CONSOLE:
+		seq_puts(m, "console");
+		break;
+	case TTY_DRIVER_TYPE_SERIAL:
+		seq_puts(m, "serial");
+		break;
+	case TTY_DRIVER_TYPE_PTY:
+		if (p->subtype == PTY_TYPE_MASTER)
+			seq_puts(m, "pty:master");
+		else if (p->subtype == PTY_TYPE_SLAVE)
+			seq_puts(m, "pty:slave");
+		else
+			seq_puts(m, "pty");
+		break;
+	default:
+		seq_printf(m, "type:%d.%d", p->type, p->subtype);
+	}
+	seq_putc(m, '\n');
+}
+
+static int show_tty_driver(struct seq_file *m, void *v)
+{
+	struct tty_driver *p = list_entry(v, struct tty_driver, tty_drivers);
+	dev_t from = MKDEV(p->major, p->minor_start);
+	dev_t to = from + p->num;
+
+	if (&p->tty_drivers == tty_drivers.next) {
+		/* pseudo-drivers first */
+		seq_printf(m, "%-20s /dev/%-8s ", "/dev/tty", "tty");
+		seq_printf(m, "%3d %7d ", TTYAUX_MAJOR, 0);
+		seq_puts(m, "system:/dev/tty\n");
+		seq_printf(m, "%-20s /dev/%-8s ", "/dev/console", "console");
+		seq_printf(m, "%3d %7d ", TTYAUX_MAJOR, 1);
+		seq_puts(m, "system:console\n");
+#ifdef CONFIG_UNIX98_PTYS
+		seq_printf(m, "%-20s /dev/%-8s ", "/dev/ptmx", "ptmx");
+		seq_printf(m, "%3d %7d ", TTYAUX_MAJOR, 2);
+		seq_puts(m, "system\n");
+#endif
+#ifdef CONFIG_VT
+		seq_printf(m, "%-20s /dev/%-8s ", "/dev/vc/0", "vc/0");
+		seq_printf(m, "%3d %7d ", TTY_MAJOR, 0);
+		seq_puts(m, "system:vtmaster\n");
+#endif
+	}
+
+	while (MAJOR(from) < MAJOR(to)) {
+		dev_t next = MKDEV(MAJOR(from)+1, 0);
+		show_tty_range(m, p, from, next - from);
+		from = next;
+	}
+	if (from != to)
+		show_tty_range(m, p, from, to - from);
+	return 0;
+}
+
+/* iterator */
+static void *t_start(struct seq_file *m, loff_t *pos)
+{
+	mutex_lock(&tty_mutex);
+	return seq_list_start(&tty_drivers, *pos);
+}
+
+static void *t_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	return seq_list_next(v, &tty_drivers, pos);
+}
+
+static void t_stop(struct seq_file *m, void *v)
+{
+	mutex_unlock(&tty_mutex);
+}
+
+static const struct seq_operations tty_drivers_op = {
+	.start	= t_start,
+	.next	= t_next,
+	.stop	= t_stop,
+	.show	= show_tty_driver
+};
+
+/*
+ * This function is called by tty_register_driver() to handle
+ * registering the driver's /proc handler into /proc/tty/driver/<foo>
+ */
+void proc_tty_register_driver(struct tty_driver *driver)
+{
+	struct proc_dir_entry *ent;
+		
+	if (!driver->driver_name || driver->proc_entry ||
+	    !driver->ops->proc_show)
+		return;
+
+	ent = proc_create_single_data(driver->driver_name, 0, proc_tty_driver,
+			       driver->ops->proc_show, driver);
+	driver->proc_entry = ent;
+}
+
+/*
+ * This function is called by tty_unregister_driver()
+ */
+void proc_tty_unregister_driver(struct tty_driver *driver)
+{
+	struct proc_dir_entry *ent;
+
+	ent = driver->proc_entry;
+	if (!ent)
+		return;
+		
+	remove_proc_entry(ent->name, proc_tty_driver);
+	
+	driver->proc_entry = NULL;
+}
+
+/*
+ * Called by proc_root_init() to initialize the /proc/tty subtree
+ */
+void __init proc_tty_init(void)
+{
+	if (!proc_mkdir("tty", NULL))
+		return;
+	proc_mkdir("tty/ldisc", NULL);	/* Preserved: it's userspace visible */
+	/*
+	 * /proc/tty/driver/serial reveals the exact character counts for
+	 * serial links which is just too easy to abuse for inferring
+	 * password lengths and inter-keystroke timings during password
+	 * entry.
+	 */
+	proc_tty_driver = proc_mkdir_mode("tty/driver", S_IRUSR|S_IXUSR, NULL);
+	proc_create_seq("tty/ldiscs", 0, NULL, &tty_ldiscs_seq_ops);
+	proc_create_seq("tty/drivers", 0, NULL, &tty_drivers_op);
+}