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, 154 insertions, 0 deletions

diff --git a/arch/mips/sgi-ip27/ip27-timer.c b/arch/mips/sgi-ip27/ip27-timer.c
new file mode 100644
index 000000000..5f4da05cb
--- /dev/null
+++ b/arch/mips/sgi-ip27/ip27-timer.c
@@ -0,0 +1,154 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 1999, 2000, 05, 06 Ralf Baechle (ralf@linux-mips.org)
+ * Copyright (C) 1999, 2000 Silicon Graphics, Inc.
+ */
+#include <linux/bcd.h>
+#include <linux/clockchips.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/sched_clock.h>
+#include <linux/interrupt.h>
+#include <linux/kernel_stat.h>
+#include <linux/param.h>
+#include <linux/smp.h>
+#include <linux/time.h>
+#include <linux/timex.h>
+#include <linux/mm.h>
+#include <linux/platform_device.h>
+
+#include <asm/time.h>
+#include <asm/sgialib.h>
+#include <asm/sn/klconfig.h>
+#include <asm/sn/arch.h>
+#include <asm/sn/addrs.h>
+#include <asm/sn/agent.h>
+
+#include "ip27-common.h"
+
+static int rt_next_event(unsigned long delta, struct clock_event_device *evt)
+{
+	unsigned int cpu = smp_processor_id();
+	int slice = cputoslice(cpu);
+	unsigned long cnt;
+
+	cnt = LOCAL_HUB_L(PI_RT_COUNT);
+	cnt += delta;
+	LOCAL_HUB_S(PI_RT_COMPARE_A + PI_COUNT_OFFSET * slice, cnt);
+
+	return LOCAL_HUB_L(PI_RT_COUNT) >= cnt ? -ETIME : 0;
+}
+
+static DEFINE_PER_CPU(struct clock_event_device, hub_rt_clockevent);
+static DEFINE_PER_CPU(char [11], hub_rt_name);
+
+static irqreturn_t hub_rt_counter_handler(int irq, void *dev_id)
+{
+	unsigned int cpu = smp_processor_id();
+	struct clock_event_device *cd = &per_cpu(hub_rt_clockevent, cpu);
+	int slice = cputoslice(cpu);
+
+	/*
+	 * Ack
+	 */
+	LOCAL_HUB_S(PI_RT_PEND_A + PI_COUNT_OFFSET * slice, 0);
+	cd->event_handler(cd);
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * This is a hack; we really need to figure these values out dynamically
+ *
+ * Since 800 ns works very well with various HUB frequencies, such as
+ * 360, 380, 390 and 400 MHZ, we use 800 ns rtc cycle time.
+ *
+ * Ralf: which clock rate is used to feed the counter?
+ */
+#define NSEC_PER_CYCLE		800
+#define CYCLES_PER_SEC		(NSEC_PER_SEC / NSEC_PER_CYCLE)
+
+void hub_rt_clock_event_init(void)
+{
+	unsigned int cpu = smp_processor_id();
+	struct clock_event_device *cd = &per_cpu(hub_rt_clockevent, cpu);
+	unsigned char *name = per_cpu(hub_rt_name, cpu);
+
+	sprintf(name, "hub-rt %d", cpu);
+	cd->name		= name;
+	cd->features		= CLOCK_EVT_FEAT_ONESHOT;
+	clockevent_set_clock(cd, CYCLES_PER_SEC);
+	cd->max_delta_ns	= clockevent_delta2ns(0xfffffffffffff, cd);
+	cd->max_delta_ticks	= 0xfffffffffffff;
+	cd->min_delta_ns	= clockevent_delta2ns(0x300, cd);
+	cd->min_delta_ticks	= 0x300;
+	cd->rating		= 200;
+	cd->irq			= IP27_RT_TIMER_IRQ;
+	cd->cpumask		= cpumask_of(cpu);
+	cd->set_next_event	= rt_next_event;
+	clockevents_register_device(cd);
+
+	enable_percpu_irq(IP27_RT_TIMER_IRQ, IRQ_TYPE_NONE);
+}
+
+static void __init hub_rt_clock_event_global_init(void)
+{
+	irq_set_handler(IP27_RT_TIMER_IRQ, handle_percpu_devid_irq);
+	irq_set_percpu_devid(IP27_RT_TIMER_IRQ);
+	WARN_ON(request_percpu_irq(IP27_RT_TIMER_IRQ, hub_rt_counter_handler,
+				   "hub-rt", &hub_rt_clockevent));
+}
+
+static u64 hub_rt_read(struct clocksource *cs)
+{
+	return REMOTE_HUB_L(cputonasid(0), PI_RT_COUNT);
+}
+
+struct clocksource hub_rt_clocksource = {
+	.name	= "HUB-RT",
+	.rating = 200,
+	.read	= hub_rt_read,
+	.mask	= CLOCKSOURCE_MASK(52),
+	.flags	= CLOCK_SOURCE_IS_CONTINUOUS,
+};
+
+static u64 notrace hub_rt_read_sched_clock(void)
+{
+	return REMOTE_HUB_L(cputonasid(0), PI_RT_COUNT);
+}
+
+static void __init hub_rt_clocksource_init(void)
+{
+	struct clocksource *cs = &hub_rt_clocksource;
+
+	clocksource_register_hz(cs, CYCLES_PER_SEC);
+
+	sched_clock_register(hub_rt_read_sched_clock, 52, CYCLES_PER_SEC);
+}
+
+void __init plat_time_init(void)
+{
+	hub_rt_clocksource_init();
+	hub_rt_clock_event_global_init();
+	hub_rt_clock_event_init();
+}
+
+void hub_rtc_init(nasid_t nasid)
+{
+
+	/*
+	 * We only need to initialize the current node.
+	 * If this is not the current node then it is a cpuless
+	 * node and timeouts will not happen there.
+	 */
+	if (get_nasid() == nasid) {
+		LOCAL_HUB_S(PI_RT_EN_A, 1);
+		LOCAL_HUB_S(PI_RT_EN_B, 1);
+		LOCAL_HUB_S(PI_PROF_EN_A, 0);
+		LOCAL_HUB_S(PI_PROF_EN_B, 0);
+		LOCAL_HUB_S(PI_RT_COUNT, 0);
+		LOCAL_HUB_S(PI_RT_PEND_A, 0);
+		LOCAL_HUB_S(PI_RT_PEND_B, 0);
+	}
+}