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, 273 insertions, 0 deletions

diff --git a/drivers/cpufreq/pasemi-cpufreq.c b/drivers/cpufreq/pasemi-cpufreq.c
new file mode 100644
index 000000000..a39313493
--- /dev/null
+++ b/drivers/cpufreq/pasemi-cpufreq.c
@@ -0,0 +1,273 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2007 PA Semi, Inc
+ *
+ * Authors: Egor Martovetsky <egor@pasemi.com>
+ *	    Olof Johansson <olof@lixom.net>
+ *
+ * Maintained by: Olof Johansson <olof@lixom.net>
+ *
+ * Based on arch/powerpc/platforms/cell/cbe_cpufreq.c:
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ */
+
+#include <linux/cpufreq.h>
+#include <linux/timer.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+
+#include <asm/hw_irq.h>
+#include <asm/io.h>
+#include <asm/time.h>
+#include <asm/smp.h>
+
+#include <platforms/pasemi/pasemi.h>
+
+#define SDCASR_REG		0x0100
+#define SDCASR_REG_STRIDE	0x1000
+#define SDCPWR_CFGA0_REG	0x0100
+#define SDCPWR_PWST0_REG	0x0000
+#define SDCPWR_GIZTIME_REG	0x0440
+
+/* SDCPWR_GIZTIME_REG fields */
+#define SDCPWR_GIZTIME_GR	0x80000000
+#define SDCPWR_GIZTIME_LONGLOCK	0x000000ff
+
+/* Offset of ASR registers from SDC base */
+#define SDCASR_OFFSET		0x120000
+
+static void __iomem *sdcpwr_mapbase;
+static void __iomem *sdcasr_mapbase;
+
+/* Current astate, is used when waking up from power savings on
+ * one core, in case the other core has switched states during
+ * the idle time.
+ */
+static int current_astate;
+
+/* We support 5(A0-A4) power states excluding turbo(A5-A6) modes */
+static struct cpufreq_frequency_table pas_freqs[] = {
+	{0, 0,	0},
+	{0, 1,	0},
+	{0, 2,	0},
+	{0, 3,	0},
+	{0, 4,	0},
+	{0, 0,	CPUFREQ_TABLE_END},
+};
+
+/*
+ * hardware specific functions
+ */
+
+static int get_astate_freq(int astate)
+{
+	u32 ret;
+	ret = in_le32(sdcpwr_mapbase + SDCPWR_CFGA0_REG + (astate * 0x10));
+
+	return ret & 0x3f;
+}
+
+static int get_cur_astate(int cpu)
+{
+	u32 ret;
+
+	ret = in_le32(sdcpwr_mapbase + SDCPWR_PWST0_REG);
+	ret = (ret >> (cpu * 4)) & 0x7;
+
+	return ret;
+}
+
+static int get_gizmo_latency(void)
+{
+	u32 giztime, ret;
+
+	giztime = in_le32(sdcpwr_mapbase + SDCPWR_GIZTIME_REG);
+
+	/* just provide the upper bound */
+	if (giztime & SDCPWR_GIZTIME_GR)
+		ret = (giztime & SDCPWR_GIZTIME_LONGLOCK) * 128000;
+	else
+		ret = (giztime & SDCPWR_GIZTIME_LONGLOCK) * 1000;
+
+	return ret;
+}
+
+static void set_astate(int cpu, unsigned int astate)
+{
+	unsigned long flags;
+
+	/* Return if called before init has run */
+	if (unlikely(!sdcasr_mapbase))
+		return;
+
+	local_irq_save(flags);
+
+	out_le32(sdcasr_mapbase + SDCASR_REG + SDCASR_REG_STRIDE*cpu, astate);
+
+	local_irq_restore(flags);
+}
+
+int check_astate(void)
+{
+	return get_cur_astate(hard_smp_processor_id());
+}
+
+void restore_astate(int cpu)
+{
+	set_astate(cpu, current_astate);
+}
+
+/*
+ * cpufreq functions
+ */
+
+static int pas_cpufreq_cpu_init(struct cpufreq_policy *policy)
+{
+	struct cpufreq_frequency_table *pos;
+	const u32 *max_freqp;
+	u32 max_freq;
+	int cur_astate, idx;
+	struct resource res;
+	struct device_node *cpu, *dn;
+	int err = -ENODEV;
+
+	cpu = of_get_cpu_node(policy->cpu, NULL);
+	if (!cpu)
+		goto out;
+
+	max_freqp = of_get_property(cpu, "clock-frequency", NULL);
+	of_node_put(cpu);
+	if (!max_freqp) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* we need the freq in kHz */
+	max_freq = *max_freqp / 1000;
+
+	dn = of_find_compatible_node(NULL, NULL, "1682m-sdc");
+	if (!dn)
+		dn = of_find_compatible_node(NULL, NULL,
+					     "pasemi,pwrficient-sdc");
+	if (!dn)
+		goto out;
+	err = of_address_to_resource(dn, 0, &res);
+	of_node_put(dn);
+	if (err)
+		goto out;
+	sdcasr_mapbase = ioremap(res.start + SDCASR_OFFSET, 0x2000);
+	if (!sdcasr_mapbase) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	dn = of_find_compatible_node(NULL, NULL, "1682m-gizmo");
+	if (!dn)
+		dn = of_find_compatible_node(NULL, NULL,
+					     "pasemi,pwrficient-gizmo");
+	if (!dn) {
+		err = -ENODEV;
+		goto out_unmap_sdcasr;
+	}
+	err = of_address_to_resource(dn, 0, &res);
+	of_node_put(dn);
+	if (err)
+		goto out_unmap_sdcasr;
+	sdcpwr_mapbase = ioremap(res.start, 0x1000);
+	if (!sdcpwr_mapbase) {
+		err = -EINVAL;
+		goto out_unmap_sdcasr;
+	}
+
+	pr_debug("init cpufreq on CPU %d\n", policy->cpu);
+	pr_debug("max clock-frequency is at %u kHz\n", max_freq);
+	pr_debug("initializing frequency table\n");
+
+	/* initialize frequency table */
+	cpufreq_for_each_entry_idx(pos, pas_freqs, idx) {
+		pos->frequency = get_astate_freq(pos->driver_data) * 100000;
+		pr_debug("%d: %d\n", idx, pos->frequency);
+	}
+
+	cur_astate = get_cur_astate(policy->cpu);
+	pr_debug("current astate is at %d\n",cur_astate);
+
+	policy->cur = pas_freqs[cur_astate].frequency;
+	ppc_proc_freq = policy->cur * 1000ul;
+
+	cpufreq_generic_init(policy, pas_freqs, get_gizmo_latency());
+	return 0;
+
+out_unmap_sdcasr:
+	iounmap(sdcasr_mapbase);
+out:
+	return err;
+}
+
+static void pas_cpufreq_cpu_exit(struct cpufreq_policy *policy)
+{
+	/*
+	 * We don't support CPU hotplug. Don't unmap after the system
+	 * has already made it to a running state.
+	 */
+	if (system_state >= SYSTEM_RUNNING)
+		return;
+
+	if (sdcasr_mapbase)
+		iounmap(sdcasr_mapbase);
+	if (sdcpwr_mapbase)
+		iounmap(sdcpwr_mapbase);
+}
+
+static int pas_cpufreq_target(struct cpufreq_policy *policy,
+			      unsigned int pas_astate_new)
+{
+	int i;
+
+	pr_debug("setting frequency for cpu %d to %d kHz, 1/%d of max frequency\n",
+		 policy->cpu,
+		 pas_freqs[pas_astate_new].frequency,
+		 pas_freqs[pas_astate_new].driver_data);
+
+	current_astate = pas_astate_new;
+
+	for_each_online_cpu(i)
+		set_astate(i, pas_astate_new);
+
+	ppc_proc_freq = pas_freqs[pas_astate_new].frequency * 1000ul;
+	return 0;
+}
+
+static struct cpufreq_driver pas_cpufreq_driver = {
+	.name		= "pas-cpufreq",
+	.flags		= CPUFREQ_CONST_LOOPS,
+	.init		= pas_cpufreq_cpu_init,
+	.exit		= pas_cpufreq_cpu_exit,
+	.verify		= cpufreq_generic_frequency_table_verify,
+	.target_index	= pas_cpufreq_target,
+};
+
+/*
+ * module init and destoy
+ */
+
+static int __init pas_cpufreq_init(void)
+{
+	if (!of_machine_is_compatible("PA6T-1682M") &&
+	    !of_machine_is_compatible("pasemi,pwrficient"))
+		return -ENODEV;
+
+	return cpufreq_register_driver(&pas_cpufreq_driver);
+}
+
+static void __exit pas_cpufreq_exit(void)
+{
+	cpufreq_unregister_driver(&pas_cpufreq_driver);
+}
+
+module_init(pas_cpufreq_init);
+module_exit(pas_cpufreq_exit);
+
+MODULE_DESCRIPTION("cpufreq driver for PA Semi PWRficient");
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Egor Martovetsky <egor@pasemi.com>, Olof Johansson <olof@lixom.net>");