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, 130 insertions, 0 deletions

diff --git a/arch/nios2/mm/uaccess.c b/arch/nios2/mm/uaccess.c
new file mode 100644
index 000000000..34f10af8e
--- /dev/null
+++ b/arch/nios2/mm/uaccess.c
@@ -0,0 +1,130 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2009, Wind River Systems Inc
+ * Implemented by fredrik.markstrom@gmail.com and ivarholmqvist@gmail.com
+ */
+
+#include <linux/export.h>
+#include <linux/uaccess.h>
+
+asm(".global	raw_copy_from_user\n"
+	"   .type raw_copy_from_user, @function\n"
+	"raw_copy_from_user:\n"
+	"   movi  r2,7\n"
+	"   mov   r3,r4\n"
+	"   bge   r2,r6,1f\n"
+	"   xor   r2,r4,r5\n"
+	"   andi  r2,r2,3\n"
+	"   movi  r7,3\n"
+	"   beq   r2,zero,4f\n"
+	"1: addi  r6,r6,-1\n"
+	"   movi  r2,-1\n"
+	"   beq   r6,r2,3f\n"
+	"   mov   r7,r2\n"
+	"2: ldbu  r2,0(r5)\n"
+	"   addi  r6,r6,-1\n"
+	"   addi  r5,r5,1\n"
+	"   stb   r2,0(r3)\n"
+	"   addi  r3,r3,1\n"
+	"   bne   r6,r7,2b\n"
+	"3:\n"
+	"   addi  r2,r6,1\n"
+	"   ret\n"
+	"13:mov   r2,r6\n"
+	"   ret\n"
+	"4: andi  r2,r4,1\n"
+	"   cmpeq r2,r2,zero\n"
+	"   beq   r2,zero,7f\n"
+	"5: andi  r2,r3,2\n"
+	"   beq   r2,zero,6f\n"
+	"9: ldhu  r2,0(r5)\n"
+	"   addi  r6,r6,-2\n"
+	"   addi  r5,r5,2\n"
+	"   sth   r2,0(r3)\n"
+	"   addi  r3,r3,2\n"
+	"6: bge   r7,r6,1b\n"
+	"10:ldw   r2,0(r5)\n"
+	"   addi  r6,r6,-4\n"
+	"   addi  r5,r5,4\n"
+	"   stw   r2,0(r3)\n"
+	"   addi  r3,r3,4\n"
+	"   br    6b\n"
+	"7: ldbu  r2,0(r5)\n"
+	"   addi  r6,r6,-1\n"
+	"   addi  r5,r5,1\n"
+	"   addi  r3,r4,1\n"
+	"   stb   r2,0(r4)\n"
+	"   br    5b\n"
+	".section __ex_table,\"a\"\n"
+	".word 2b,3b\n"
+	".word 9b,13b\n"
+	".word 10b,13b\n"
+	".word 7b,13b\n"
+	".previous\n"
+	);
+EXPORT_SYMBOL(raw_copy_from_user);
+
+asm(
+	"   .global raw_copy_to_user\n"
+	"   .type raw_copy_to_user, @function\n"
+	"raw_copy_to_user:\n"
+	"   movi  r2,7\n"
+	"   mov   r3,r4\n"
+	"   bge   r2,r6,1f\n"
+	"   xor   r2,r4,r5\n"
+	"   andi  r2,r2,3\n"
+	"   movi  r7,3\n"
+	"   beq   r2,zero,4f\n"
+	/* Bail if we try to copy zero bytes  */
+	"1: addi  r6,r6,-1\n"
+	"   movi  r2,-1\n"
+	"   beq   r6,r2,3f\n"
+	/* Copy byte by byte for small copies and if src^dst != 0 */
+	"   mov   r7,r2\n"
+	"2: ldbu  r2,0(r5)\n"
+	"   addi  r5,r5,1\n"
+	"9: stb   r2,0(r3)\n"
+	"   addi  r6,r6,-1\n"
+	"   addi  r3,r3,1\n"
+	"   bne   r6,r7,2b\n"
+	"3: addi  r2,r6,1\n"
+	"   ret\n"
+	"13:mov   r2,r6\n"
+	"   ret\n"
+	/*  If 'to' is an odd address byte copy */
+	"4: andi  r2,r4,1\n"
+	"   cmpeq r2,r2,zero\n"
+	"   beq   r2,zero,7f\n"
+	/* If 'to' is not divideable by four copy halfwords */
+	"5: andi  r2,r3,2\n"
+	"   beq   r2,zero,6f\n"
+	"   ldhu  r2,0(r5)\n"
+	"   addi  r5,r5,2\n"
+	"10:sth   r2,0(r3)\n"
+	"   addi  r6,r6,-2\n"
+	"   addi  r3,r3,2\n"
+	/* Copy words */
+	"6: bge   r7,r6,1b\n"
+	"   ldw   r2,0(r5)\n"
+	"   addi  r5,r5,4\n"
+	"11:stw   r2,0(r3)\n"
+	"   addi  r6,r6,-4\n"
+	"   addi  r3,r3,4\n"
+	"   br    6b\n"
+	/* Copy remaining bytes */
+	"7: ldbu  r2,0(r5)\n"
+	"   addi  r5,r5,1\n"
+	"   addi  r3,r4,1\n"
+	"12: stb  r2,0(r4)\n"
+	"   addi  r6,r6,-1\n"
+	"   br    5b\n"
+	".section __ex_table,\"a\"\n"
+	".word 9b,3b\n"
+	".word 10b,13b\n"
+	".word 11b,13b\n"
+	".word 12b,13b\n"
+	".previous\n");
+EXPORT_SYMBOL(raw_copy_to_user);