x86/32: Remove schedule_tail_wrapper()

The unwinder expects a return address at the very top of the kernel
stack just below pt_regs and before any stack frame is created.  Instead
of calling a wrapper, set up a return address as if ret_from_fork()
was called from the syscall entry code.
Signed-off-by: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Sami Tolvanen <samitolvanen@google.com>
Link: https://lkml.kernel.org/r/20230623225529.34590-2-brgerst@gmail.com

arch/x86/entry/entry_32.S

@@ -719,26 +719,6 @@ SYM_CODE_START(__switch_to_asm)
 SYM_CODE_END(__switch_to_asm)
 .popsection
 
-/*
- * The unwinder expects the last frame on the stack to always be at the same
- * offset from the end of the page, which allows it to validate the stack.
- * Calling schedule_tail() directly would break that convention because its an
- * asmlinkage function so its argument has to be pushed on the stack.  This
- * wrapper creates a proper "end of stack" frame header before the call.
- */
-.pushsection .text, "ax"
-SYM_FUNC_START(schedule_tail_wrapper)
-	FRAME_BEGIN
-
-	pushl	%eax
-	call	schedule_tail
-	popl	%eax
-
-	FRAME_END
-	RET
-SYM_FUNC_END(schedule_tail_wrapper)
-.popsection
-
 /*
  * A newly forked process directly context switches into this address.
  *
@@ -748,16 +728,23 @@ SYM_FUNC_END(schedule_tail_wrapper)
  */
 .pushsection .text, "ax"
 SYM_CODE_START(ret_from_fork)
-	call	schedule_tail_wrapper
+	/* return address for the stack unwinder */
+	pushl	$.Lsyscall_32_done
+
+	FRAME_BEGIN
+	pushl	%eax
+	call	schedule_tail
+	addl	$4, %esp
+	FRAME_END
 
 	testl	%ebx, %ebx
 	jnz	1f		/* kernel threads are uncommon */
 
 2:
 	/* When we fork, we trace the syscall return in the child, too. */
-	movl    %esp, %eax
+	leal    4(%esp), %eax
 	call    syscall_exit_to_user_mode
-	jmp     .Lsyscall_32_done
+	RET
 
 	/* kernel thread */
 1:	movl	%edi, %eax