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, 157 insertions, 0 deletions

diff --git a/arch/um/os-Linux/time.c b/arch/um/os-Linux/time.c
new file mode 100644
index 000000000..13ebc8691
--- /dev/null
+++ b/arch/um/os-Linux/time.c
@@ -0,0 +1,157 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2015 Anton Ivanov (aivanov@{brocade.com,kot-begemot.co.uk})
+ * Copyright (C) 2015 Thomas Meyer (thomas@m3y3r.de)
+ * Copyright (C) 2012-2014 Cisco Systems
+ * Copyright (C) 2000 - 2007 Jeff Dike (jdike{addtoit,linux.intel}.com)
+ */
+
+#include <stddef.h>
+#include <unistd.h>
+#include <errno.h>
+#include <signal.h>
+#include <time.h>
+#include <sys/signalfd.h>
+#include <sys/time.h>
+#include <kern_util.h>
+#include <os.h>
+#include <smp.h>
+#include <string.h>
+#include "internal.h"
+
+static timer_t event_high_res_timer[CONFIG_NR_CPUS] = { 0 };
+
+static inline long long timespec_to_ns(const struct timespec *ts)
+{
+	return ((long long) ts->tv_sec * UM_NSEC_PER_SEC) + ts->tv_nsec;
+}
+
+long long os_persistent_clock_emulation(void)
+{
+	struct timespec realtime_tp;
+
+	clock_gettime(CLOCK_REALTIME, &realtime_tp);
+	return timespec_to_ns(&realtime_tp);
+}
+
+#ifndef sigev_notify_thread_id
+#define sigev_notify_thread_id _sigev_un._tid
+#endif
+
+/**
+ * os_timer_create() - create an new posix (interval) timer
+ */
+int os_timer_create(void)
+{
+	int cpu = uml_curr_cpu();
+	timer_t *t = &event_high_res_timer[cpu];
+	struct sigevent sev = {
+		.sigev_notify = SIGEV_THREAD_ID,
+		.sigev_signo = SIGALRM,
+		.sigev_value.sival_ptr = t,
+		.sigev_notify_thread_id = gettid(),
+	};
+
+	if (timer_create(CLOCK_MONOTONIC, &sev, t) == -1)
+		return -1;
+
+	return 0;
+}
+
+int os_timer_set_interval(int cpu, unsigned long long nsecs)
+{
+	struct itimerspec its;
+
+	its.it_value.tv_sec = nsecs / UM_NSEC_PER_SEC;
+	its.it_value.tv_nsec = nsecs % UM_NSEC_PER_SEC;
+
+	its.it_interval.tv_sec = nsecs / UM_NSEC_PER_SEC;
+	its.it_interval.tv_nsec = nsecs % UM_NSEC_PER_SEC;
+
+	if (timer_settime(event_high_res_timer[cpu], 0, &its, NULL) == -1)
+		return -errno;
+
+	return 0;
+}
+
+int os_timer_one_shot(int cpu, unsigned long long nsecs)
+{
+	struct itimerspec its = {
+		.it_value.tv_sec = nsecs / UM_NSEC_PER_SEC,
+		.it_value.tv_nsec = nsecs % UM_NSEC_PER_SEC,
+
+		.it_interval.tv_sec = 0,
+		.it_interval.tv_nsec = 0, // we cheat here
+	};
+
+	timer_settime(event_high_res_timer[cpu], 0, &its, NULL);
+	return 0;
+}
+
+/**
+ * os_timer_disable() - disable the posix (interval) timer
+ * @cpu: the CPU for which the timer is to be disabled
+ */
+void os_timer_disable(int cpu)
+{
+	struct itimerspec its;
+
+	memset(&its, 0, sizeof(struct itimerspec));
+	timer_settime(event_high_res_timer[cpu], 0, &its, NULL);
+}
+
+long long os_nsecs(void)
+{
+	struct timespec ts;
+
+	clock_gettime(CLOCK_MONOTONIC,&ts);
+	return timespec_to_ns(&ts);
+}
+
+static __thread int wake_signals;
+
+void os_idle_prepare(void)
+{
+	sigset_t set;
+
+	sigemptyset(&set);
+	sigaddset(&set, SIGALRM);
+	sigaddset(&set, IPI_SIGNAL);
+
+	/*
+	 * We need to use signalfd rather than sigsuspend in idle sleep
+	 * because the IPI signal is a real-time signal that carries data,
+	 * and unlike handling SIGALRM, we cannot simply flag it in
+	 * signals_pending.
+	 */
+	wake_signals = signalfd(-1, &set, SFD_CLOEXEC);
+	if (wake_signals < 0)
+		panic("Failed to create signal FD, errno = %d", errno);
+}
+
+/**
+ * os_idle_sleep() - sleep until interrupted
+ */
+void os_idle_sleep(void)
+{
+	sigset_t set;
+
+	/*
+	 * Block SIGALRM while performing the need_resched check.
+	 * Note that, because IRQs are disabled, the IPI signal is
+	 * already blocked.
+	 */
+	sigemptyset(&set);
+	sigaddset(&set, SIGALRM);
+	sigprocmask(SIG_BLOCK, &set, NULL);
+
+	/*
+	 * Because disabling IRQs does not block SIGALRM, it is also
+	 * necessary to check for any pending timer alarms.
+	 */
+	if (!uml_need_resched() && !timer_alarm_pending())
+		os_poll(1, &wake_signals);
+
+	/* Restore the signal mask. */
+	sigprocmask(SIG_UNBLOCK, &set, NULL);
+}