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, 252 insertions, 0 deletions

diff --git a/drivers/pmdomain/bcm/raspberrypi-power.c b/drivers/pmdomain/bcm/raspberrypi-power.c
new file mode 100644
index 000000000..b87ea7adb
--- /dev/null
+++ b/drivers/pmdomain/bcm/raspberrypi-power.c
@@ -0,0 +1,252 @@
+// SPDX-License-Identifier: GPL-2.0
+/* (C) 2015 Pengutronix, Alexander Aring <aar@pengutronix.de>
+ *
+ * Authors:
+ * Alexander Aring <aar@pengutronix.de>
+ * Eric Anholt <eric@anholt.net>
+ */
+
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_domain.h>
+#include <dt-bindings/power/raspberrypi-power.h>
+#include <soc/bcm2835/raspberrypi-firmware.h>
+
+/*
+ * Firmware indices for the old power domains interface.  Only a few
+ * of them were actually implemented.
+ */
+#define RPI_OLD_POWER_DOMAIN_USB		3
+#define RPI_OLD_POWER_DOMAIN_V3D		10
+
+struct rpi_power_domain {
+	u32 domain;
+	bool enabled;
+	bool old_interface;
+	struct generic_pm_domain base;
+	struct rpi_firmware *fw;
+};
+
+struct rpi_power_domains {
+	bool has_new_interface;
+	struct genpd_onecell_data xlate;
+	struct rpi_firmware *fw;
+	struct rpi_power_domain domains[RPI_POWER_DOMAIN_COUNT];
+};
+
+/*
+ * Packet definition used by RPI_FIRMWARE_SET_POWER_STATE and
+ * RPI_FIRMWARE_SET_DOMAIN_STATE
+ */
+struct rpi_power_domain_packet {
+	u32 domain;
+	u32 state;
+};
+
+/*
+ * Asks the firmware to enable or disable power on a specific power
+ * domain.
+ */
+static int rpi_firmware_set_power(struct generic_pm_domain *domain, bool on)
+{
+	struct rpi_power_domain *rpi_domain =
+		container_of(domain, struct rpi_power_domain, base);
+	bool old_interface = rpi_domain->old_interface;
+	struct rpi_power_domain_packet packet;
+	int ret;
+
+	packet.domain = rpi_domain->domain;
+	packet.state = on;
+
+	ret = rpi_firmware_property(rpi_domain->fw, old_interface ?
+				    RPI_FIRMWARE_SET_POWER_STATE :
+				    RPI_FIRMWARE_SET_DOMAIN_STATE,
+				    &packet, sizeof(packet));
+	if (ret)
+		dev_err(&domain->dev, "Failed to set %s to %u (%d)\n",
+			old_interface ? "power" : "domain", on, ret);
+	else
+		dev_dbg(&domain->dev, "Set %s to %u\n",
+			old_interface ? "power" : "domain", on);
+
+	return ret;
+}
+
+static int rpi_domain_off(struct generic_pm_domain *domain)
+{
+	return rpi_firmware_set_power(domain, false);
+}
+
+static int rpi_domain_on(struct generic_pm_domain *domain)
+{
+	return rpi_firmware_set_power(domain, true);
+}
+
+static void rpi_common_init_power_domain(struct rpi_power_domains *rpi_domains,
+					 int xlate_index, const char *name)
+{
+	struct rpi_power_domain *dom = &rpi_domains->domains[xlate_index];
+
+	dom->fw = rpi_domains->fw;
+
+	dom->base.name = name;
+	dom->base.flags = GENPD_FLAG_ACTIVE_WAKEUP;
+	dom->base.power_on = rpi_domain_on;
+	dom->base.power_off = rpi_domain_off;
+
+	/*
+	 * Treat all power domains as off at boot.
+	 *
+	 * The firmware itself may be keeping some domains on, but
+	 * from Linux's perspective all we control is the refcounts
+	 * that we give to the firmware, and we can't ask the firmware
+	 * to turn off something that we haven't ourselves turned on.
+	 */
+	pm_genpd_init(&dom->base, NULL, true);
+
+	rpi_domains->xlate.domains[xlate_index] = &dom->base;
+}
+
+static void rpi_init_power_domain(struct rpi_power_domains *rpi_domains,
+				  int xlate_index, const char *name)
+{
+	struct rpi_power_domain *dom = &rpi_domains->domains[xlate_index];
+
+	if (!rpi_domains->has_new_interface)
+		return;
+
+	/* The DT binding index is the firmware's domain index minus one. */
+	dom->domain = xlate_index + 1;
+
+	rpi_common_init_power_domain(rpi_domains, xlate_index, name);
+}
+
+static void rpi_init_old_power_domain(struct rpi_power_domains *rpi_domains,
+				      int xlate_index, int domain,
+				      const char *name)
+{
+	struct rpi_power_domain *dom = &rpi_domains->domains[xlate_index];
+
+	dom->old_interface = true;
+	dom->domain = domain;
+
+	rpi_common_init_power_domain(rpi_domains, xlate_index, name);
+}
+
+/*
+ * Detects whether the firmware supports the new power domains interface.
+ *
+ * The firmware doesn't actually return an error on an unknown tag,
+ * and just skips over it, so we do the detection by putting an
+ * unexpected value in the return field and checking if it was
+ * unchanged.
+ */
+static bool
+rpi_has_new_domain_support(struct rpi_power_domains *rpi_domains)
+{
+	struct rpi_power_domain_packet packet;
+	int ret;
+
+	packet.domain = RPI_POWER_DOMAIN_ARM;
+	packet.state = ~0;
+
+	ret = rpi_firmware_property(rpi_domains->fw,
+				    RPI_FIRMWARE_GET_DOMAIN_STATE,
+				    &packet, sizeof(packet));
+
+	return ret == 0 && packet.state != ~0;
+}
+
+static int rpi_power_probe(struct platform_device *pdev)
+{
+	struct device_node *fw_np;
+	struct device *dev = &pdev->dev;
+	struct rpi_power_domains *rpi_domains;
+
+	rpi_domains = devm_kzalloc(dev, sizeof(*rpi_domains), GFP_KERNEL);
+	if (!rpi_domains)
+		return -ENOMEM;
+
+	rpi_domains->xlate.domains =
+		devm_kcalloc(dev,
+			     RPI_POWER_DOMAIN_COUNT,
+			     sizeof(*rpi_domains->xlate.domains),
+			     GFP_KERNEL);
+	if (!rpi_domains->xlate.domains)
+		return -ENOMEM;
+
+	rpi_domains->xlate.num_domains = RPI_POWER_DOMAIN_COUNT;
+
+	fw_np = of_parse_phandle(pdev->dev.of_node, "firmware", 0);
+	if (!fw_np) {
+		dev_err(&pdev->dev, "no firmware node\n");
+		return -ENODEV;
+	}
+
+	rpi_domains->fw = devm_rpi_firmware_get(&pdev->dev, fw_np);
+	of_node_put(fw_np);
+	if (!rpi_domains->fw)
+		return -EPROBE_DEFER;
+
+	rpi_domains->has_new_interface =
+		rpi_has_new_domain_support(rpi_domains);
+
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_I2C0, "I2C0");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_I2C1, "I2C1");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_I2C2, "I2C2");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_VIDEO_SCALER,
+			      "VIDEO_SCALER");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_VPU1, "VPU1");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_HDMI, "HDMI");
+
+	/*
+	 * Use the old firmware interface for USB power, so that we
+	 * can turn it on even if the firmware hasn't been updated.
+	 */
+	rpi_init_old_power_domain(rpi_domains, RPI_POWER_DOMAIN_USB,
+				  RPI_OLD_POWER_DOMAIN_USB, "USB");
+
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_VEC, "VEC");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_JPEG, "JPEG");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_H264, "H264");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_V3D, "V3D");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_ISP, "ISP");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_UNICAM0, "UNICAM0");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_UNICAM1, "UNICAM1");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_CCP2RX, "CCP2RX");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_CSI2, "CSI2");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_CPI, "CPI");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_DSI0, "DSI0");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_DSI1, "DSI1");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_TRANSPOSER,
+			      "TRANSPOSER");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_CCP2TX, "CCP2TX");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_CDP, "CDP");
+	rpi_init_power_domain(rpi_domains, RPI_POWER_DOMAIN_ARM, "ARM");
+
+	of_genpd_add_provider_onecell(dev->of_node, &rpi_domains->xlate);
+
+	platform_set_drvdata(pdev, rpi_domains);
+
+	return 0;
+}
+
+static const struct of_device_id rpi_power_of_match[] = {
+	{ .compatible = "raspberrypi,bcm2835-power", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, rpi_power_of_match);
+
+static struct platform_driver rpi_power_driver = {
+	.driver = {
+		.name = "raspberrypi-power",
+		.of_match_table = rpi_power_of_match,
+	},
+	.probe		= rpi_power_probe,
+};
+builtin_platform_driver(rpi_power_driver);
+
+MODULE_AUTHOR("Alexander Aring <aar@pengutronix.de>");
+MODULE_AUTHOR("Eric Anholt <eric@anholt.net>");
+MODULE_DESCRIPTION("Raspberry Pi power domain driver");