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, 175 insertions, 0 deletions

diff --git a/drivers/mailbox/pl320-ipc.c b/drivers/mailbox/pl320-ipc.c
new file mode 100644
index 000000000..606f26a2a
--- /dev/null
+++ b/drivers/mailbox/pl320-ipc.c
@@ -0,0 +1,175 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2012 Calxeda, Inc.
+ */
+#include <linux/types.h>
+#include <linux/err.h>
+#include <linux/delay.h>
+#include <linux/export.h>
+#include <linux/io.h>
+#include <linux/interrupt.h>
+#include <linux/completion.h>
+#include <linux/mutex.h>
+#include <linux/notifier.h>
+#include <linux/spinlock.h>
+#include <linux/device.h>
+#include <linux/amba/bus.h>
+
+#include <linux/pl320-ipc.h>
+
+#define IPCMxSOURCE(m)		((m) * 0x40)
+#define IPCMxDSET(m)		(((m) * 0x40) + 0x004)
+#define IPCMxDCLEAR(m)		(((m) * 0x40) + 0x008)
+#define IPCMxDSTATUS(m)		(((m) * 0x40) + 0x00C)
+#define IPCMxMODE(m)		(((m) * 0x40) + 0x010)
+#define IPCMxMSET(m)		(((m) * 0x40) + 0x014)
+#define IPCMxMCLEAR(m)		(((m) * 0x40) + 0x018)
+#define IPCMxMSTATUS(m)		(((m) * 0x40) + 0x01C)
+#define IPCMxSEND(m)		(((m) * 0x40) + 0x020)
+#define IPCMxDR(m, dr)		(((m) * 0x40) + ((dr) * 4) + 0x024)
+
+#define IPCMMIS(irq)		(((irq) * 8) + 0x800)
+#define IPCMRIS(irq)		(((irq) * 8) + 0x804)
+
+#define MBOX_MASK(n)		(1 << (n))
+#define IPC_TX_MBOX		1
+#define IPC_RX_MBOX		2
+
+#define CHAN_MASK(n)		(1 << (n))
+#define A9_SOURCE		1
+#define M3_SOURCE		0
+
+static void __iomem *ipc_base;
+static int ipc_irq;
+static DEFINE_MUTEX(ipc_m1_lock);
+static DECLARE_COMPLETION(ipc_completion);
+static ATOMIC_NOTIFIER_HEAD(ipc_notifier);
+
+static void __ipc_send(int mbox, u32 *data)
+{
+	int i;
+	for (i = 0; i < 7; i++)
+		writel_relaxed(data[i], ipc_base + IPCMxDR(mbox, i));
+	writel_relaxed(0x1, ipc_base + IPCMxSEND(mbox));
+}
+
+static u32 __ipc_rcv(int mbox, u32 *data)
+{
+	int i;
+	for (i = 0; i < 7; i++)
+		data[i] = readl_relaxed(ipc_base + IPCMxDR(mbox, i));
+	return data[1];
+}
+
+/* blocking implementation from the A9 side, not usable in interrupts! */
+int pl320_ipc_transmit(u32 *data)
+{
+	int ret;
+
+	mutex_lock(&ipc_m1_lock);
+
+	init_completion(&ipc_completion);
+	__ipc_send(IPC_TX_MBOX, data);
+	ret = wait_for_completion_timeout(&ipc_completion,
+					  msecs_to_jiffies(1000));
+	if (ret == 0) {
+		ret = -ETIMEDOUT;
+		goto out;
+	}
+
+	ret = __ipc_rcv(IPC_TX_MBOX, data);
+out:
+	mutex_unlock(&ipc_m1_lock);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(pl320_ipc_transmit);
+
+static irqreturn_t ipc_handler(int irq, void *dev)
+{
+	u32 irq_stat;
+	u32 data[7];
+
+	irq_stat = readl_relaxed(ipc_base + IPCMMIS(1));
+	if (irq_stat & MBOX_MASK(IPC_TX_MBOX)) {
+		writel_relaxed(0, ipc_base + IPCMxSEND(IPC_TX_MBOX));
+		complete(&ipc_completion);
+	}
+	if (irq_stat & MBOX_MASK(IPC_RX_MBOX)) {
+		__ipc_rcv(IPC_RX_MBOX, data);
+		atomic_notifier_call_chain(&ipc_notifier, data[0], data + 1);
+		writel_relaxed(2, ipc_base + IPCMxSEND(IPC_RX_MBOX));
+	}
+
+	return IRQ_HANDLED;
+}
+
+int pl320_ipc_register_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_register(&ipc_notifier, nb);
+}
+EXPORT_SYMBOL_GPL(pl320_ipc_register_notifier);
+
+int pl320_ipc_unregister_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&ipc_notifier, nb);
+}
+EXPORT_SYMBOL_GPL(pl320_ipc_unregister_notifier);
+
+static int pl320_probe(struct amba_device *adev, const struct amba_id *id)
+{
+	int ret;
+
+	ipc_base = ioremap(adev->res.start, resource_size(&adev->res));
+	if (ipc_base == NULL)
+		return -ENOMEM;
+
+	writel_relaxed(0, ipc_base + IPCMxSEND(IPC_TX_MBOX));
+
+	ipc_irq = adev->irq[0];
+	ret = request_irq(ipc_irq, ipc_handler, 0, dev_name(&adev->dev), NULL);
+	if (ret < 0)
+		goto err;
+
+	/* Init slow mailbox */
+	writel_relaxed(CHAN_MASK(A9_SOURCE),
+		       ipc_base + IPCMxSOURCE(IPC_TX_MBOX));
+	writel_relaxed(CHAN_MASK(M3_SOURCE),
+		       ipc_base + IPCMxDSET(IPC_TX_MBOX));
+	writel_relaxed(CHAN_MASK(M3_SOURCE) | CHAN_MASK(A9_SOURCE),
+		       ipc_base + IPCMxMSET(IPC_TX_MBOX));
+
+	/* Init receive mailbox */
+	writel_relaxed(CHAN_MASK(M3_SOURCE),
+		       ipc_base + IPCMxSOURCE(IPC_RX_MBOX));
+	writel_relaxed(CHAN_MASK(A9_SOURCE),
+		       ipc_base + IPCMxDSET(IPC_RX_MBOX));
+	writel_relaxed(CHAN_MASK(M3_SOURCE) | CHAN_MASK(A9_SOURCE),
+		       ipc_base + IPCMxMSET(IPC_RX_MBOX));
+
+	return 0;
+err:
+	iounmap(ipc_base);
+	return ret;
+}
+
+static const struct amba_id pl320_ids[] = {
+	{
+		.id	= 0x00041320,
+		.mask	= 0x000fffff,
+	},
+	{ 0, 0 },
+};
+
+static struct amba_driver pl320_driver = {
+	.drv = {
+		.name	= "pl320",
+	},
+	.id_table	= pl320_ids,
+	.probe		= pl320_probe,
+};
+
+static int __init ipc_init(void)
+{
+	return amba_driver_register(&pl320_driver);
+}
+subsys_initcall(ipc_init);