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, 325 insertions, 0 deletions

diff --git a/drivers/crypto/ccp/sp-dev.c b/drivers/crypto/ccp/sp-dev.c
new file mode 100644
index 000000000..f204aa5df
--- /dev/null
+++ b/drivers/crypto/ccp/sp-dev.c
@@ -0,0 +1,325 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * AMD Secure Processor driver
+ *
+ * Copyright (C) 2017-2018 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ * Author: Gary R Hook <gary.hook@amd.com>
+ * Author: Brijesh Singh <brijesh.singh@amd.com>
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/kthread.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+#include <linux/spinlock_types.h>
+#include <linux/types.h>
+#include <linux/ccp.h>
+
+#include "sev-dev.h"
+#include "ccp-dev.h"
+#include "sp-dev.h"
+
+MODULE_AUTHOR("Tom Lendacky <thomas.lendacky@amd.com>");
+MODULE_AUTHOR("Gary R Hook <gary.hook@amd.com>");
+MODULE_LICENSE("GPL");
+MODULE_VERSION("1.1.0");
+MODULE_DESCRIPTION("AMD Secure Processor driver");
+
+/* List of SPs, SP count, read-write access lock, and access functions
+ *
+ * Lock structure: get sp_unit_lock for reading whenever we need to
+ * examine the SP list.
+ */
+static DEFINE_RWLOCK(sp_unit_lock);
+static LIST_HEAD(sp_units);
+
+/* Ever-increasing value to produce unique unit numbers */
+static atomic_t sp_ordinal;
+
+static void sp_add_device(struct sp_device *sp)
+{
+	unsigned long flags;
+
+	write_lock_irqsave(&sp_unit_lock, flags);
+
+	list_add_tail(&sp->entry, &sp_units);
+
+	write_unlock_irqrestore(&sp_unit_lock, flags);
+}
+
+static void sp_del_device(struct sp_device *sp)
+{
+	unsigned long flags;
+
+	write_lock_irqsave(&sp_unit_lock, flags);
+
+	list_del(&sp->entry);
+
+	write_unlock_irqrestore(&sp_unit_lock, flags);
+}
+
+static irqreturn_t sp_irq_handler(int irq, void *data)
+{
+	struct sp_device *sp = data;
+
+	if (sp->ccp_irq_handler)
+		sp->ccp_irq_handler(irq, sp->ccp_irq_data);
+
+	if (sp->psp_irq_handler)
+		sp->psp_irq_handler(irq, sp->psp_irq_data);
+
+	return IRQ_HANDLED;
+}
+
+int sp_request_ccp_irq(struct sp_device *sp, irq_handler_t handler,
+		       const char *name, void *data)
+{
+	int ret;
+
+	if ((sp->psp_irq == sp->ccp_irq) && sp->dev_vdata->psp_vdata) {
+		/* Need a common routine to manage all interrupts */
+		sp->ccp_irq_data = data;
+		sp->ccp_irq_handler = handler;
+
+		if (!sp->irq_registered) {
+			ret = request_irq(sp->ccp_irq, sp_irq_handler, 0,
+					  sp->name, sp);
+			if (ret)
+				return ret;
+
+			sp->irq_registered = true;
+		}
+	} else {
+		/* Each sub-device can manage it's own interrupt */
+		ret = request_irq(sp->ccp_irq, handler, 0, name, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+int sp_request_psp_irq(struct sp_device *sp, irq_handler_t handler,
+		       const char *name, void *data)
+{
+	int ret;
+
+	if ((sp->psp_irq == sp->ccp_irq) && sp->dev_vdata->ccp_vdata) {
+		/* Need a common routine to manage all interrupts */
+		sp->psp_irq_data = data;
+		sp->psp_irq_handler = handler;
+
+		if (!sp->irq_registered) {
+			ret = request_irq(sp->psp_irq, sp_irq_handler, 0,
+					  sp->name, sp);
+			if (ret)
+				return ret;
+
+			sp->irq_registered = true;
+		}
+	} else {
+		/* Each sub-device can manage it's own interrupt */
+		ret = request_irq(sp->psp_irq, handler, 0, name, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+void sp_free_ccp_irq(struct sp_device *sp, void *data)
+{
+	if ((sp->psp_irq == sp->ccp_irq) && sp->dev_vdata->psp_vdata) {
+		/* Using common routine to manage all interrupts */
+		if (!sp->psp_irq_handler) {
+			/* Nothing else using it, so free it */
+			free_irq(sp->ccp_irq, sp);
+
+			sp->irq_registered = false;
+		}
+
+		sp->ccp_irq_handler = NULL;
+		sp->ccp_irq_data = NULL;
+	} else {
+		/* Each sub-device can manage it's own interrupt */
+		free_irq(sp->ccp_irq, data);
+	}
+}
+
+void sp_free_psp_irq(struct sp_device *sp, void *data)
+{
+	if ((sp->psp_irq == sp->ccp_irq) && sp->dev_vdata->ccp_vdata) {
+		/* Using common routine to manage all interrupts */
+		if (!sp->ccp_irq_handler) {
+			/* Nothing else using it, so free it */
+			free_irq(sp->psp_irq, sp);
+
+			sp->irq_registered = false;
+		}
+
+		sp->psp_irq_handler = NULL;
+		sp->psp_irq_data = NULL;
+	} else {
+		/* Each sub-device can manage it's own interrupt */
+		free_irq(sp->psp_irq, data);
+	}
+}
+
+/**
+ * sp_alloc_struct - allocate and initialize the sp_device struct
+ *
+ * @dev: device struct of the SP
+ */
+struct sp_device *sp_alloc_struct(struct device *dev)
+{
+	struct sp_device *sp;
+
+	sp = devm_kzalloc(dev, sizeof(*sp), GFP_KERNEL);
+	if (!sp)
+		return NULL;
+
+	sp->dev = dev;
+	sp->ord = atomic_inc_return(&sp_ordinal);
+	snprintf(sp->name, SP_MAX_NAME_LEN, "sp-%u", sp->ord);
+
+	return sp;
+}
+
+int sp_init(struct sp_device *sp)
+{
+	sp_add_device(sp);
+
+	if (sp->dev_vdata->ccp_vdata)
+		ccp_dev_init(sp);
+
+	if (sp->dev_vdata->psp_vdata)
+		psp_dev_init(sp);
+	return 0;
+}
+
+void sp_destroy(struct sp_device *sp)
+{
+	if (sp->dev_vdata->ccp_vdata)
+		ccp_dev_destroy(sp);
+
+	if (sp->dev_vdata->psp_vdata)
+		psp_dev_destroy(sp);
+
+	sp_del_device(sp);
+}
+
+int sp_suspend(struct sp_device *sp)
+{
+	if (sp->dev_vdata->ccp_vdata) {
+		ccp_dev_suspend(sp);
+	}
+
+	return 0;
+}
+
+int sp_resume(struct sp_device *sp)
+{
+	if (sp->dev_vdata->ccp_vdata) {
+		ccp_dev_resume(sp);
+	}
+
+	return 0;
+}
+
+int sp_restore(struct sp_device *sp)
+{
+	if (sp->psp_data) {
+		int ret = psp_restore(sp);
+
+		if (ret)
+			return ret;
+	}
+
+	return sp_resume(sp);
+}
+
+struct sp_device *sp_get_psp_master_device(void)
+{
+	struct sp_device *i, *ret = NULL;
+	unsigned long flags;
+
+	write_lock_irqsave(&sp_unit_lock, flags);
+	if (list_empty(&sp_units))
+		goto unlock;
+
+	list_for_each_entry(i, &sp_units, entry) {
+		if (i->psp_data && i->get_psp_master_device) {
+			ret = i->get_psp_master_device();
+			break;
+		}
+	}
+
+unlock:
+	write_unlock_irqrestore(&sp_unit_lock, flags);
+	return ret;
+}
+
+static int __init sp_mod_init(void)
+{
+#ifdef CONFIG_X86
+	static bool initialized;
+	int ret;
+
+	if (initialized)
+		return 0;
+
+	ret = sp_pci_init();
+	if (ret)
+		return ret;
+
+#ifdef CONFIG_CRYPTO_DEV_SP_PSP
+	psp_pci_init();
+#endif
+
+	initialized = true;
+
+	return 0;
+#endif
+
+#ifdef CONFIG_ARM64
+	int ret;
+
+	ret = sp_platform_init();
+	if (ret)
+		return ret;
+
+	return 0;
+#endif
+
+	return -ENODEV;
+}
+
+#if IS_BUILTIN(CONFIG_KVM_AMD) && IS_ENABLED(CONFIG_KVM_AMD_SEV)
+int __init sev_module_init(void)
+{
+	return sp_mod_init();
+}
+#endif
+
+static void __exit sp_mod_exit(void)
+{
+#ifdef CONFIG_X86
+
+#ifdef CONFIG_CRYPTO_DEV_SP_PSP
+	psp_pci_exit();
+#endif
+
+	sp_pci_exit();
+#endif
+
+#ifdef CONFIG_ARM64
+	sp_platform_exit();
+#endif
+}
+
+module_init(sp_mod_init);
+module_exit(sp_mod_exit);