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, 191 insertions, 0 deletions

diff --git a/drivers/mtd/ubi/nvmem.c b/drivers/mtd/ubi/nvmem.c
new file mode 100644
index 000000000..2dade9727
--- /dev/null
+++ b/drivers/mtd/ubi/nvmem.c
@@ -0,0 +1,191 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2023 Daniel Golle <daniel@makrotopia.org>
+ */
+
+/* UBI NVMEM provider */
+#include "ubi.h"
+#include <linux/nvmem-provider.h>
+
+/* List of all NVMEM devices */
+static LIST_HEAD(nvmem_devices);
+static DEFINE_MUTEX(devices_mutex);
+
+struct ubi_nvmem {
+	struct nvmem_device *nvmem;
+	int ubi_num;
+	int vol_id;
+	int usable_leb_size;
+	struct list_head list;
+};
+
+static int ubi_nvmem_reg_read(void *priv, unsigned int from,
+			      void *val, size_t bytes)
+{
+	size_t to_read, bytes_left = bytes;
+	struct ubi_nvmem *unv = priv;
+	struct ubi_volume_desc *desc;
+	uint32_t offs;
+	uint32_t lnum;
+	int err = 0;
+
+	desc = ubi_open_volume(unv->ubi_num, unv->vol_id, UBI_READONLY);
+	if (IS_ERR(desc))
+		return PTR_ERR(desc);
+
+	offs = from % unv->usable_leb_size;
+	lnum = from / unv->usable_leb_size;
+	while (bytes_left) {
+		to_read = unv->usable_leb_size - offs;
+
+		if (to_read > bytes_left)
+			to_read = bytes_left;
+
+		err = ubi_read(desc, lnum, val, offs, to_read);
+		if (err)
+			break;
+
+		lnum += 1;
+		offs = 0;
+		bytes_left -= to_read;
+		val += to_read;
+	}
+	ubi_close_volume(desc);
+
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int ubi_nvmem_add(struct ubi_volume_info *vi)
+{
+	struct device_node *np = dev_of_node(vi->dev);
+	struct nvmem_config config = {};
+	struct ubi_nvmem *unv;
+	int ret;
+
+	if (!np)
+		return 0;
+
+	if (!of_get_child_by_name(np, "nvmem-layout"))
+		return 0;
+
+	if (WARN_ON_ONCE(vi->usable_leb_size <= 0) ||
+	    WARN_ON_ONCE(vi->size <= 0))
+		return -EINVAL;
+
+	unv = kzalloc_obj(struct ubi_nvmem);
+	if (!unv)
+		return -ENOMEM;
+
+	config.id = NVMEM_DEVID_NONE;
+	config.dev = vi->dev;
+	config.name = dev_name(vi->dev);
+	config.owner = THIS_MODULE;
+	config.priv = unv;
+	config.reg_read = ubi_nvmem_reg_read;
+	config.size = vi->usable_leb_size * vi->size;
+	config.word_size = 1;
+	config.stride = 1;
+	config.read_only = true;
+	config.root_only = true;
+	config.ignore_wp = true;
+	config.of_node = np;
+
+	unv->ubi_num = vi->ubi_num;
+	unv->vol_id = vi->vol_id;
+	unv->usable_leb_size = vi->usable_leb_size;
+	unv->nvmem = nvmem_register(&config);
+	if (IS_ERR(unv->nvmem)) {
+		ret = dev_err_probe(vi->dev, PTR_ERR(unv->nvmem),
+				    "Failed to register NVMEM device\n");
+		kfree(unv);
+		return ret;
+	}
+
+	mutex_lock(&devices_mutex);
+	list_add_tail(&unv->list, &nvmem_devices);
+	mutex_unlock(&devices_mutex);
+
+	return 0;
+}
+
+static void ubi_nvmem_remove(struct ubi_volume_info *vi)
+{
+	struct ubi_nvmem *unv_c, *unv = NULL;
+
+	mutex_lock(&devices_mutex);
+	list_for_each_entry(unv_c, &nvmem_devices, list)
+		if (unv_c->ubi_num == vi->ubi_num && unv_c->vol_id == vi->vol_id) {
+			unv = unv_c;
+			break;
+		}
+
+	if (!unv) {
+		mutex_unlock(&devices_mutex);
+		return;
+	}
+
+	list_del(&unv->list);
+	mutex_unlock(&devices_mutex);
+	nvmem_unregister(unv->nvmem);
+	kfree(unv);
+}
+
+/**
+ * nvmem_notify - UBI notification handler.
+ * @nb: registered notifier block
+ * @l: notification type
+ * @ns_ptr: pointer to the &struct ubi_notification object
+ */
+static int nvmem_notify(struct notifier_block *nb, unsigned long l,
+			 void *ns_ptr)
+{
+	struct ubi_notification *nt = ns_ptr;
+
+	switch (l) {
+	case UBI_VOLUME_RESIZED:
+		ubi_nvmem_remove(&nt->vi);
+		fallthrough;
+	case UBI_VOLUME_ADDED:
+		ubi_nvmem_add(&nt->vi);
+		break;
+	case UBI_VOLUME_SHUTDOWN:
+		ubi_nvmem_remove(&nt->vi);
+		break;
+	default:
+		break;
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block nvmem_notifier = {
+	.notifier_call = nvmem_notify,
+};
+
+static int __init ubi_nvmem_init(void)
+{
+	return ubi_register_volume_notifier(&nvmem_notifier, 0);
+}
+
+static void __exit ubi_nvmem_exit(void)
+{
+	struct ubi_nvmem *unv, *tmp;
+
+	mutex_lock(&devices_mutex);
+	list_for_each_entry_safe(unv, tmp, &nvmem_devices, list) {
+		nvmem_unregister(unv->nvmem);
+		list_del(&unv->list);
+		kfree(unv);
+	}
+	mutex_unlock(&devices_mutex);
+
+	ubi_unregister_volume_notifier(&nvmem_notifier);
+}
+
+module_init(ubi_nvmem_init);
+module_exit(ubi_nvmem_exit);
+MODULE_DESCRIPTION("NVMEM layer over UBI volumes");
+MODULE_AUTHOR("Daniel Golle");
+MODULE_LICENSE("GPL");