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, 216 insertions, 0 deletions

diff --git a/drivers/clocksource/timer-pistachio.c b/drivers/clocksource/timer-pistachio.c
new file mode 100644
index 000000000..57b2197a0
--- /dev/null
+++ b/drivers/clocksource/timer-pistachio.c
@@ -0,0 +1,216 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Pistachio clocksource based on general-purpose timers
+ *
+ * Copyright (C) 2015 Imagination Technologies
+ */
+
+#define pr_fmt(fmt) "%s: " fmt, __func__
+
+#include <linux/clk.h>
+#include <linux/clocksource.h>
+#include <linux/clockchips.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/spinlock.h>
+#include <linux/mfd/syscon.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/sched_clock.h>
+#include <linux/time.h>
+
+/* Top level reg */
+#define CR_TIMER_CTRL_CFG		0x00
+#define TIMER_ME_GLOBAL			BIT(0)
+#define CR_TIMER_REV			0x10
+
+/* Timer specific registers */
+#define TIMER_CFG			0x20
+#define TIMER_ME_LOCAL			BIT(0)
+#define TIMER_RELOAD_VALUE		0x24
+#define TIMER_CURRENT_VALUE		0x28
+#define TIMER_CURRENT_OVERFLOW_VALUE	0x2C
+#define TIMER_IRQ_STATUS		0x30
+#define TIMER_IRQ_CLEAR			0x34
+#define TIMER_IRQ_MASK			0x38
+
+#define PERIP_TIMER_CONTROL		0x90
+
+/* Timer specific configuration Values */
+#define RELOAD_VALUE			0xffffffff
+
+struct pistachio_clocksource {
+	void __iomem *base;
+	raw_spinlock_t lock;
+	struct clocksource cs;
+};
+
+static struct pistachio_clocksource pcs_gpt;
+
+#define to_pistachio_clocksource(cs)	\
+	container_of(cs, struct pistachio_clocksource, cs)
+
+static inline u32 gpt_readl(void __iomem *base, u32 offset, u32 gpt_id)
+{
+	return readl(base + 0x20 * gpt_id + offset);
+}
+
+static inline void gpt_writel(void __iomem *base, u32 value, u32 offset,
+		u32 gpt_id)
+{
+	writel(value, base + 0x20 * gpt_id + offset);
+}
+
+static u64 notrace
+pistachio_clocksource_read_cycles(struct clocksource *cs)
+{
+	struct pistachio_clocksource *pcs = to_pistachio_clocksource(cs);
+	__maybe_unused u32 overflow;
+	u32 counter;
+	unsigned long flags;
+
+	/*
+	 * The counter value is only refreshed after the overflow value is read.
+	 * And they must be read in strict order, hence raw spin lock added.
+	 */
+
+	raw_spin_lock_irqsave(&pcs->lock, flags);
+	overflow = gpt_readl(pcs->base, TIMER_CURRENT_OVERFLOW_VALUE, 0);
+	counter = gpt_readl(pcs->base, TIMER_CURRENT_VALUE, 0);
+	raw_spin_unlock_irqrestore(&pcs->lock, flags);
+
+	return (u64)~counter;
+}
+
+static u64 notrace pistachio_read_sched_clock(void)
+{
+	return pistachio_clocksource_read_cycles(&pcs_gpt.cs);
+}
+
+static void pistachio_clksrc_set_mode(struct clocksource *cs, int timeridx,
+			int enable)
+{
+	struct pistachio_clocksource *pcs = to_pistachio_clocksource(cs);
+	u32 val;
+
+	val = gpt_readl(pcs->base, TIMER_CFG, timeridx);
+	if (enable)
+		val |= TIMER_ME_LOCAL;
+	else
+		val &= ~TIMER_ME_LOCAL;
+
+	gpt_writel(pcs->base, val, TIMER_CFG, timeridx);
+}
+
+static void pistachio_clksrc_enable(struct clocksource *cs, int timeridx)
+{
+	struct pistachio_clocksource *pcs = to_pistachio_clocksource(cs);
+
+	/* Disable GPT local before loading reload value */
+	pistachio_clksrc_set_mode(cs, timeridx, false);
+	gpt_writel(pcs->base, RELOAD_VALUE, TIMER_RELOAD_VALUE, timeridx);
+	pistachio_clksrc_set_mode(cs, timeridx, true);
+}
+
+static void pistachio_clksrc_disable(struct clocksource *cs, int timeridx)
+{
+	/* Disable GPT local */
+	pistachio_clksrc_set_mode(cs, timeridx, false);
+}
+
+static int pistachio_clocksource_enable(struct clocksource *cs)
+{
+	pistachio_clksrc_enable(cs, 0);
+	return 0;
+}
+
+static void pistachio_clocksource_disable(struct clocksource *cs)
+{
+	pistachio_clksrc_disable(cs, 0);
+}
+
+/* Desirable clock source for pistachio platform */
+static struct pistachio_clocksource pcs_gpt = {
+	.cs =	{
+		.name		= "gptimer",
+		.rating		= 300,
+		.enable		= pistachio_clocksource_enable,
+		.disable	= pistachio_clocksource_disable,
+		.read		= pistachio_clocksource_read_cycles,
+		.mask		= CLOCKSOURCE_MASK(32),
+		.flags		= CLOCK_SOURCE_IS_CONTINUOUS |
+				  CLOCK_SOURCE_SUSPEND_NONSTOP,
+		},
+};
+
+static int __init pistachio_clksrc_of_init(struct device_node *node)
+{
+	struct clk *sys_clk, *fast_clk;
+	struct regmap *periph_regs;
+	unsigned long rate;
+	int ret;
+
+	pcs_gpt.base = of_iomap(node, 0);
+	if (!pcs_gpt.base) {
+		pr_err("cannot iomap\n");
+		return -ENXIO;
+	}
+
+	periph_regs = syscon_regmap_lookup_by_phandle(node, "img,cr-periph");
+	if (IS_ERR(periph_regs)) {
+		pr_err("cannot get peripheral regmap (%ld)\n",
+		       PTR_ERR(periph_regs));
+		return PTR_ERR(periph_regs);
+	}
+
+	/* Switch to using the fast counter clock */
+	ret = regmap_update_bits(periph_regs, PERIP_TIMER_CONTROL,
+				 0xf, 0x0);
+	if (ret)
+		return ret;
+
+	sys_clk = of_clk_get_by_name(node, "sys");
+	if (IS_ERR(sys_clk)) {
+		pr_err("clock get failed (%ld)\n", PTR_ERR(sys_clk));
+		return PTR_ERR(sys_clk);
+	}
+
+	fast_clk = of_clk_get_by_name(node, "fast");
+	if (IS_ERR(fast_clk)) {
+		pr_err("clock get failed (%lu)\n", PTR_ERR(fast_clk));
+		return PTR_ERR(fast_clk);
+	}
+
+	ret = clk_prepare_enable(sys_clk);
+	if (ret < 0) {
+		pr_err("failed to enable clock (%d)\n", ret);
+		return ret;
+	}
+
+	ret = clk_prepare_enable(fast_clk);
+	if (ret < 0) {
+		pr_err("failed to enable clock (%d)\n", ret);
+		clk_disable_unprepare(sys_clk);
+		return ret;
+	}
+
+	rate = clk_get_rate(fast_clk);
+
+	/* Disable irq's for clocksource usage */
+	gpt_writel(pcs_gpt.base, 0, TIMER_IRQ_MASK, 0);
+	gpt_writel(pcs_gpt.base, 0, TIMER_IRQ_MASK, 1);
+	gpt_writel(pcs_gpt.base, 0, TIMER_IRQ_MASK, 2);
+	gpt_writel(pcs_gpt.base, 0, TIMER_IRQ_MASK, 3);
+
+	/* Enable timer block */
+	writel(TIMER_ME_GLOBAL, pcs_gpt.base);
+
+	raw_spin_lock_init(&pcs_gpt.lock);
+	sched_clock_register(pistachio_read_sched_clock, 32, rate);
+	return clocksource_register_hz(&pcs_gpt.cs, rate);
+}
+TIMER_OF_DECLARE(pistachio_gptimer, "img,pistachio-gptimer",
+		       pistachio_clksrc_of_init);