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, 215 insertions, 0 deletions

diff --git a/drivers/clocksource/timer-gxp.c b/drivers/clocksource/timer-gxp.c
new file mode 100644
index 000000000..1becc9433
--- /dev/null
+++ b/drivers/clocksource/timer-gxp.c
@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (C) 2022 Hewlett-Packard Enterprise Development Company, L.P. */
+
+#include <linux/clk.h>
+#include <linux/clockchips.h>
+#include <linux/clocksource.h>
+#include <linux/interrupt.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/sched_clock.h>
+
+#define TIMER0_FREQ	1000000
+#define GXP_TIMER_CNT_OFS 0x00
+#define GXP_TIMESTAMP_OFS 0x08
+#define GXP_TIMER_CTRL_OFS 0x14
+
+/* TCS Stands for Timer Control/Status: these are masks to be used in */
+/* the Timer Count Registers */
+#define MASK_TCS_ENABLE	0x01
+#define MASK_TCS_PERIOD	0x02
+#define MASK_TCS_RELOAD	0x04
+#define MASK_TCS_TC	0x80
+
+struct gxp_timer {
+	void __iomem *counter;
+	void __iomem *control;
+	struct clock_event_device evt;
+};
+
+static struct gxp_timer *gxp_timer;
+
+static void __iomem *system_clock __ro_after_init;
+
+static inline struct gxp_timer *to_gxp_timer(struct clock_event_device *evt_dev)
+{
+	return container_of(evt_dev, struct gxp_timer, evt);
+}
+
+static u64 notrace gxp_sched_read(void)
+{
+	return readl_relaxed(system_clock);
+}
+
+static int gxp_time_set_next_event(unsigned long event, struct clock_event_device *evt_dev)
+{
+	struct gxp_timer *timer = to_gxp_timer(evt_dev);
+
+	/* Stop counting and disable interrupt before updating */
+	writeb_relaxed(MASK_TCS_TC, timer->control);
+	writel_relaxed(event, timer->counter);
+	writeb_relaxed(MASK_TCS_TC | MASK_TCS_ENABLE, timer->control);
+
+	return 0;
+}
+
+static irqreturn_t gxp_timer_interrupt(int irq, void *dev_id)
+{
+	struct gxp_timer *timer = (struct gxp_timer *)dev_id;
+
+	if (!(readb_relaxed(timer->control) & MASK_TCS_TC))
+		return IRQ_NONE;
+
+	writeb_relaxed(MASK_TCS_TC, timer->control);
+
+	timer->evt.event_handler(&timer->evt);
+
+	return IRQ_HANDLED;
+}
+
+static int __init gxp_timer_init(struct device_node *node)
+{
+	void __iomem *base;
+	struct clk *clk;
+	u32 freq;
+	int ret, irq;
+
+	gxp_timer = kzalloc_obj(*gxp_timer);
+	if (!gxp_timer) {
+		ret = -ENOMEM;
+		pr_err("Can't allocate gxp_timer");
+		return ret;
+	}
+
+	clk = of_clk_get(node, 0);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		pr_err("%pOFn clock not found: %d\n", node, ret);
+		goto err_free;
+	}
+
+	ret = clk_prepare_enable(clk);
+	if (ret) {
+		pr_err("%pOFn clock enable failed: %d\n", node, ret);
+		goto err_clk_enable;
+	}
+
+	base = of_iomap(node, 0);
+	if (!base) {
+		ret = -ENXIO;
+		pr_err("Can't map timer base registers");
+		goto err_iomap;
+	}
+
+	/* Set the offsets to the clock register and timer registers */
+	gxp_timer->counter = base + GXP_TIMER_CNT_OFS;
+	gxp_timer->control = base + GXP_TIMER_CTRL_OFS;
+	system_clock = base + GXP_TIMESTAMP_OFS;
+
+	gxp_timer->evt.name = node->name;
+	gxp_timer->evt.rating = 300;
+	gxp_timer->evt.features = CLOCK_EVT_FEAT_ONESHOT;
+	gxp_timer->evt.set_next_event = gxp_time_set_next_event;
+	gxp_timer->evt.cpumask = cpumask_of(0);
+
+	irq = irq_of_parse_and_map(node, 0);
+	if (irq <= 0) {
+		ret = -EINVAL;
+		pr_err("GXP Timer Can't parse IRQ %d", irq);
+		goto err_exit;
+	}
+
+	freq = clk_get_rate(clk);
+
+	ret = clocksource_mmio_init(system_clock, node->name, freq,
+				    300, 32, clocksource_mmio_readl_up);
+	if (ret) {
+		pr_err("%pOFn init clocksource failed: %d", node, ret);
+		goto err_exit;
+	}
+
+	sched_clock_register(gxp_sched_read, 32, freq);
+
+	irq = irq_of_parse_and_map(node, 0);
+	if (irq <= 0) {
+		ret = -EINVAL;
+		pr_err("%pOFn Can't parse IRQ %d", node, irq);
+		goto err_exit;
+	}
+
+	clockevents_config_and_register(&gxp_timer->evt, TIMER0_FREQ,
+					0xf, 0xffffffff);
+
+	ret = request_irq(irq, gxp_timer_interrupt, IRQF_TIMER | IRQF_SHARED,
+			  node->name, gxp_timer);
+	if (ret) {
+		pr_err("%pOFn request_irq() failed: %d", node, ret);
+		goto err_exit;
+	}
+
+	pr_debug("gxp: system timer (irq = %d)\n", irq);
+	return 0;
+
+err_exit:
+	iounmap(base);
+err_iomap:
+	clk_disable_unprepare(clk);
+err_clk_enable:
+	clk_put(clk);
+err_free:
+	kfree(gxp_timer);
+	return ret;
+}
+
+/*
+ * This probe gets called after the timer is already up and running. This will create
+ * the watchdog device as a child since the registers are shared.
+ */
+
+static int gxp_timer_probe(struct platform_device *pdev)
+{
+	struct platform_device *gxp_watchdog_device;
+	struct device *dev = &pdev->dev;
+	int ret;
+
+	if (!gxp_timer) {
+		pr_err("Gxp Timer not initialized, cannot create watchdog");
+		return -ENOMEM;
+	}
+
+	gxp_watchdog_device = platform_device_alloc("gxp-wdt", -1);
+	if (!gxp_watchdog_device) {
+		pr_err("Timer failed to allocate gxp-wdt");
+		return -ENOMEM;
+	}
+
+	/* Pass the base address (counter) as platform data and nothing else */
+	gxp_watchdog_device->dev.platform_data = gxp_timer->counter;
+	gxp_watchdog_device->dev.parent = dev;
+
+	ret = platform_device_add(gxp_watchdog_device);
+	if (ret)
+		platform_device_put(gxp_watchdog_device);
+
+	return ret;
+}
+
+static const struct of_device_id gxp_timer_of_match[] = {
+	{ .compatible = "hpe,gxp-timer", },
+	{},
+};
+
+static struct platform_driver gxp_timer_driver = {
+	.probe  = gxp_timer_probe,
+	.driver = {
+		.name = "gxp-timer",
+		.of_match_table = gxp_timer_of_match,
+		.suppress_bind_attrs = true,
+	},
+};
+
+builtin_platform_driver(gxp_timer_driver);
+
+TIMER_OF_DECLARE(gxp, "hpe,gxp-timer", gxp_timer_init);