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, 231 insertions, 0 deletions

diff --git a/drivers/base/regmap/regmap-fsi.c b/drivers/base/regmap/regmap-fsi.c
new file mode 100644
index 000000000..3d2f3cb31
--- /dev/null
+++ b/drivers/base/regmap/regmap-fsi.c
@@ -0,0 +1,231 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Register map access API - FSI support
+//
+// Copyright 2022 IBM Corp
+//
+// Author: Eddie James <eajames@linux.ibm.com>
+
+#include <linux/fsi.h>
+#include <linux/module.h>
+#include <linux/regmap.h>
+
+#include "internal.h"
+
+static int regmap_fsi32_reg_read(void *context, unsigned int reg, unsigned int *val)
+{
+	u32 v;
+	int ret;
+
+	ret = fsi_slave_read(context, reg, &v, sizeof(v));
+	if (ret)
+		return ret;
+
+	*val = v;
+	return 0;
+}
+
+static int regmap_fsi32_reg_write(void *context, unsigned int reg, unsigned int val)
+{
+	u32 v = val;
+
+	return fsi_slave_write(context, reg, &v, sizeof(v));
+}
+
+static const struct regmap_bus regmap_fsi32 = {
+	.reg_write = regmap_fsi32_reg_write,
+	.reg_read = regmap_fsi32_reg_read,
+};
+
+static int regmap_fsi32le_reg_read(void *context, unsigned int reg, unsigned int *val)
+{
+	__be32 v;
+	int ret;
+
+	ret = fsi_slave_read(context, reg, &v, sizeof(v));
+	if (ret)
+		return ret;
+
+	*val = be32_to_cpu(v);
+	return 0;
+}
+
+static int regmap_fsi32le_reg_write(void *context, unsigned int reg, unsigned int val)
+{
+	__be32 v = cpu_to_be32(val);
+
+	return fsi_slave_write(context, reg, &v, sizeof(v));
+}
+
+static const struct regmap_bus regmap_fsi32le = {
+	.reg_write = regmap_fsi32le_reg_write,
+	.reg_read = regmap_fsi32le_reg_read,
+};
+
+static int regmap_fsi16_reg_read(void *context, unsigned int reg, unsigned int *val)
+{
+	u16 v;
+	int ret;
+
+	ret = fsi_slave_read(context, reg, &v, sizeof(v));
+	if (ret)
+		return ret;
+
+	*val = v;
+	return 0;
+}
+
+static int regmap_fsi16_reg_write(void *context, unsigned int reg, unsigned int val)
+{
+	u16 v;
+
+	if (val > 0xffff)
+		return -EINVAL;
+
+	v = val;
+	return fsi_slave_write(context, reg, &v, sizeof(v));
+}
+
+static const struct regmap_bus regmap_fsi16 = {
+	.reg_write = regmap_fsi16_reg_write,
+	.reg_read = regmap_fsi16_reg_read,
+};
+
+static int regmap_fsi16le_reg_read(void *context, unsigned int reg, unsigned int *val)
+{
+	__be16 v;
+	int ret;
+
+	ret = fsi_slave_read(context, reg, &v, sizeof(v));
+	if (ret)
+		return ret;
+
+	*val = be16_to_cpu(v);
+	return 0;
+}
+
+static int regmap_fsi16le_reg_write(void *context, unsigned int reg, unsigned int val)
+{
+	__be16 v;
+
+	if (val > 0xffff)
+		return -EINVAL;
+
+	v = cpu_to_be16(val);
+	return fsi_slave_write(context, reg, &v, sizeof(v));
+}
+
+static const struct regmap_bus regmap_fsi16le = {
+	.reg_write = regmap_fsi16le_reg_write,
+	.reg_read = regmap_fsi16le_reg_read,
+};
+
+static int regmap_fsi8_reg_read(void *context, unsigned int reg, unsigned int *val)
+{
+	u8 v;
+	int ret;
+
+	ret = fsi_slave_read(context, reg, &v, sizeof(v));
+	if (ret)
+		return ret;
+
+	*val = v;
+	return 0;
+}
+
+static int regmap_fsi8_reg_write(void *context, unsigned int reg, unsigned int val)
+{
+	u8 v;
+
+	if (val > 0xff)
+		return -EINVAL;
+
+	v = val;
+	return fsi_slave_write(context, reg, &v, sizeof(v));
+}
+
+static const struct regmap_bus regmap_fsi8 = {
+	.reg_write = regmap_fsi8_reg_write,
+	.reg_read = regmap_fsi8_reg_read,
+};
+
+static const struct regmap_bus *regmap_get_fsi_bus(struct fsi_device *fsi_dev,
+						   const struct regmap_config *config)
+{
+	const struct regmap_bus *bus = NULL;
+
+	if (config->reg_bits == 8 || config->reg_bits == 16 || config->reg_bits == 32) {
+		switch (config->val_bits) {
+		case 8:
+			bus = &regmap_fsi8;
+			break;
+		case 16:
+			switch (regmap_get_val_endian(&fsi_dev->dev, NULL, config)) {
+			case REGMAP_ENDIAN_LITTLE:
+#ifdef __LITTLE_ENDIAN
+			case REGMAP_ENDIAN_NATIVE:
+#endif
+				bus = &regmap_fsi16le;
+				break;
+			case REGMAP_ENDIAN_DEFAULT:
+			case REGMAP_ENDIAN_BIG:
+#ifdef __BIG_ENDIAN
+			case REGMAP_ENDIAN_NATIVE:
+#endif
+				bus = &regmap_fsi16;
+				break;
+			default:
+				break;
+			}
+			break;
+		case 32:
+			switch (regmap_get_val_endian(&fsi_dev->dev, NULL, config)) {
+			case REGMAP_ENDIAN_LITTLE:
+#ifdef __LITTLE_ENDIAN
+			case REGMAP_ENDIAN_NATIVE:
+#endif
+				bus = &regmap_fsi32le;
+				break;
+			case REGMAP_ENDIAN_DEFAULT:
+			case REGMAP_ENDIAN_BIG:
+#ifdef __BIG_ENDIAN
+			case REGMAP_ENDIAN_NATIVE:
+#endif
+				bus = &regmap_fsi32;
+				break;
+			default:
+				break;
+			}
+			break;
+		}
+	}
+
+	return bus ?: ERR_PTR(-EOPNOTSUPP);
+}
+
+struct regmap *__regmap_init_fsi(struct fsi_device *fsi_dev, const struct regmap_config *config,
+				 struct lock_class_key *lock_key, const char *lock_name)
+{
+	const struct regmap_bus *bus = regmap_get_fsi_bus(fsi_dev, config);
+
+	if (IS_ERR(bus))
+		return ERR_CAST(bus);
+
+	return __regmap_init(&fsi_dev->dev, bus, fsi_dev->slave, config, lock_key, lock_name);
+}
+EXPORT_SYMBOL_GPL(__regmap_init_fsi);
+
+struct regmap *__devm_regmap_init_fsi(struct fsi_device *fsi_dev,
+				      const struct regmap_config *config,
+				      struct lock_class_key *lock_key, const char *lock_name)
+{
+	const struct regmap_bus *bus = regmap_get_fsi_bus(fsi_dev, config);
+
+	if (IS_ERR(bus))
+		return ERR_CAST(bus);
+
+	return __devm_regmap_init(&fsi_dev->dev, bus, fsi_dev->slave, config, lock_key, lock_name);
+}
+EXPORT_SYMBOL_GPL(__devm_regmap_init_fsi);
+
+MODULE_LICENSE("GPL");