/*	$NetBSD: dtrace_asm.S,v 1.7 2018/05/28 21:05:03 chs Exp $	*/

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License, Version 1.0 only
 * (the "License").  You may not use this file except in compliance
 * with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 *
 * $FreeBSD: head/sys/cddl/dev/dtrace/i386/dtrace_asm.S 298171 2016-04-17 23:08:47Z markj $
 */
/*
 * Copyright 2004 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 */

#define _ASM

#include "assym.h"

#include <sys/cpuvar_defs.h>
#include <sys/dtrace.h>
#include <machine/asm.h>
#include <machine/frameasm.h>
#include <machine/trap.h>

#define INTR_POP		\
	addl	$16, %esp;	\
	popl	%edi;		\
	popl	%esi;		\
	popl	%ebp;		\
	popl	%ebx;		\
	popl	%edx;		\
	popl	%ecx;		\
	popl	%eax;		\
	addl	$8, %esp

	ENTRY(dtrace_invop_start)
	
	/* Store a trapframe for dtrace. */
	pushl	$0
	pushl	$T_PRIVINFLT
	pushl	%eax
	pushl	%ecx
	pushl	%edx
	pushl	%ebx
	pushl	%ebp
	pushl	%esi
	pushl	%edi
	subl	$16,%esp		/* dummy for segment regs */
	cld

	/* Store the args to dtrace_invop(). */
	pushl	%eax			/* push %eax -- may be return value */
	pushl	%esp			/* push stack pointer */
	addl	$4, (%esp)		/* skip first arg and segment regs */
	pushl	TF_EIP+8(%esp)		/* push calling EIP */

	/*
	 * Call dtrace_invop to let it check if the exception was
	 * a fbt one. The return value in %eax will tell us what
	 * dtrace_invop wants us to do.
	 */
	call	dtrace_invop
	ALTENTRY(dtrace_invop_callsite)
	addl	$12, %esp
	cmpl	$DTRACE_INVOP_PUSHL_EBP, %eax
	je	invop_push
	cmpl	$DTRACE_INVOP_POPL_EBP, %eax
	je	invop_pop
	cmpl	$DTRACE_INVOP_LEAVE, %eax
	je	invop_leave
	cmpl	$DTRACE_INVOP_NOP, %eax
	je	invop_nop

	/* When all else fails handle the trap in the usual way. */
	jmpl	*dtrace_invop_calltrap_addr

invop_push:
	/*
	 * We must emulate a "pushl %ebp".  To do this, we pull the stack
	 * down 4 bytes, and then store the base pointer.
	 */
	INTR_POP
	subl	$4, %esp		/* make room for %ebp */
	pushl	%eax			/* push temp */
	movl	8(%esp), %eax		/* load calling EIP */
	incl	%eax			/* increment over LOCK prefix */
	movl	%eax, 4(%esp)		/* store calling EIP */
	movl	12(%esp), %eax		/* load calling CS */
	movl	%eax, 8(%esp)		/* store calling CS */
	movl	16(%esp), %eax		/* load calling EFLAGS */
	movl	%eax, 12(%esp)		/* store calling EFLAGS */
	movl	%ebp, 16(%esp)		/* push %ebp */
	popl	%eax			/* pop off temp */
	iret				/* Return from interrupt. */
invop_pop:
	/*
	 * We must emulate a "popl %ebp".  To do this, we do the opposite of
	 * the above:  we remove the %ebp from the stack, and squeeze up the
	 * saved state from the trap.
	 */
	INTR_POP
	pushl	%eax			/* push temp */
	movl	16(%esp), %ebp		/* pop %ebp */
	movl	12(%esp), %eax		/* load calling EFLAGS */
	movl	%eax, 16(%esp)		/* store calling EFLAGS */
	movl	8(%esp), %eax		/* load calling CS */
	movl	%eax, 12(%esp)		/* store calling CS */
	movl	4(%esp), %eax		/* load calling EIP */
	incl	%eax			/* increment over LOCK prefix */
	movl	%eax, 8(%esp)		/* store calling EIP */
	popl	%eax			/* pop off temp */
	addl	$4, %esp		/* adjust stack pointer */
	iret				/* Return from interrupt. */
invop_leave:
	/*
	 * We must emulate a "leave", which is the same as a "movl %ebp, %esp"
	 * followed by a "popl %ebp".  This looks similar to the above, but
	 * requires two temporaries:  one for the new base pointer, and one
	 * for the staging register.
	 */
	INTR_POP
	pushl	%eax			/* push temp */
	pushl	%ebx			/* push temp */
	movl	%ebp, %ebx		/* set temp to old %ebp */
	movl	(%ebx), %ebp		/* pop %ebp */
	movl	16(%esp), %eax		/* load calling EFLAGS */
	movl	%eax, (%ebx)		/* store calling EFLAGS */
	movl	12(%esp), %eax		/* load calling CS */
	movl	%eax, -4(%ebx)		/* store calling CS */
	movl	8(%esp), %eax		/* load calling EIP */
	incl	%eax			/* increment over LOCK prefix */
	movl	%eax, -8(%ebx)		/* store calling EIP */
	subl	$8, %ebx		/* adjust for three pushes, one pop */
	movl	%ebx, 8(%esp)		/* temporarily store new %esp */
	popl	%ebx			/* pop off temp */
	popl	%eax			/* pop off temp */
	movl	(%esp), %esp		/* set stack pointer */
	iret				/* return from interrupt */
invop_nop:
	/*
	 * We must emulate a "nop".  This is obviously not hard:  we need only
	 * advance the %eip by one.
	 */
	INTR_POP
	incl	(%esp)
	iret				/* return from interrupt */

	END(dtrace_invop_start)

/*
void dtrace_invop_init(void)
*/
	ENTRY(dtrace_invop_init)
	movl	$dtrace_invop_start, dtrace_invop_jump_addr
	ret
	END(dtrace_invop_init)

/*
void dtrace_invop_uninit(void)
*/
	ENTRY(dtrace_invop_uninit)
	movl	$0, dtrace_invop_jump_addr
	ret
	END(dtrace_invop_uninit)

/*
greg_t dtrace_getfp(void)
*/

	ENTRY(dtrace_getfp)
	movl	%ebp, %eax
	ret
	END(dtrace_getfp)

/*
uint32_t dtrace_cas32(uint32_t *target, uint32_t cmp, uint32_t new)
*/

	ENTRY(dtrace_cas32)
	ALTENTRY(dtrace_casptr)
	movl	4(%esp), %edx
	movl	8(%esp), %eax
	movl	12(%esp), %ecx
	lock
	cmpxchgl %ecx, (%edx)
	ret
	END(dtrace_casptr)
	END(dtrace_cas32)

/*
uintptr_t dtrace_caller(int aframes)
*/

	ENTRY(dtrace_caller)
	movl	$-1, %eax
	ret
	END(dtrace_caller)

/*
void dtrace_copy(uintptr_t src, uintptr_t dest, size_t size)
*/

	ENTRY(dtrace_copy)
	pushl	%ebp
	movl	%esp, %ebp
	pushl	%esi
	pushl	%edi

	movl	8(%ebp), %esi		/* Load source address */
	movl	12(%ebp), %edi		/* Load destination address */
	movl	16(%ebp), %ecx		/* Load count */
	repz				/* Repeat for count... */
	smovb				/*   move from %ds:si to %es:di */

	popl	%edi
	popl	%esi
	movl	%ebp, %esp
	popl	%ebp
	ret
	END(dtrace_copy)

/*
void dtrace_copystr(uintptr_t uaddr, uintptr_t kaddr, size_t size)
*/

	ENTRY(dtrace_copystr)

	pushl	%ebp			/* Setup stack frame */
	movl	%esp, %ebp
	pushl	%ebx			/* Save registers */

	movl	8(%ebp), %ebx		/* Load source address */
	movl	12(%ebp), %edx		/* Load destination address */
	movl	16(%ebp), %ecx		/* Load count */

0:
	movb	(%ebx), %al		/* Load from source */
	movb	%al, (%edx)		/* Store to destination */
	incl	%ebx			/* Increment source pointer */
	incl	%edx			/* Increment destination pointer */
	decl	%ecx			/* Decrement remaining count */
	cmpb	$0, %al
	je	1f
	cmpl	$0, %ecx
	jne	0b

1:
	popl	%ebx
	movl	%ebp, %esp
	popl	%ebp
	ret

	END(dtrace_copystr)

/*
uintptr_t dtrace_fulword(void *addr)
*/

	ENTRY(dtrace_fulword)
	movl	4(%esp), %ecx
	xorl	%eax, %eax
	movl	(%ecx), %eax
	ret
	END(dtrace_fulword)

/*
uint8_t dtrace_fuword8_nocheck(void *addr)
*/

	ENTRY(dtrace_fuword8_nocheck)
	movl	4(%esp), %ecx
	xorl	%eax, %eax
	movzbl	(%ecx), %eax
	ret
	END(dtrace_fuword8_nocheck)

/*
uint16_t dtrace_fuword16_nocheck(void *addr)
*/

	ENTRY(dtrace_fuword16_nocheck)
	movl	4(%esp), %ecx
	xorl	%eax, %eax
	movzwl	(%ecx), %eax
	ret
	END(dtrace_fuword16_nocheck)

/*
uint32_t dtrace_fuword32_nocheck(void *addr)
*/

	ENTRY(dtrace_fuword32_nocheck)
	movl	4(%esp), %ecx
	xorl	%eax, %eax
	movl	(%ecx), %eax
	ret
	END(dtrace_fuword32_nocheck)

/*
uint64_t dtrace_fuword64_nocheck(void *addr)
*/

	ENTRY(dtrace_fuword64_nocheck)
	movl	4(%esp), %ecx
	xorl	%eax, %eax
	xorl	%edx, %edx
	movl	(%ecx), %eax
	movl	4(%ecx), %edx
	ret
	END(dtrace_fuword64_nocheck)

/*
void dtrace_probe_error(dtrace_state_t *state, dtrace_epid_t epid, int which, int fault, int fltoffs, uintptr_t illval)
*/

	ENTRY(dtrace_probe_error)
	pushl	%ebp
	movl	%esp, %ebp
	pushl	0x1c(%ebp)
	pushl	0x18(%ebp)
	pushl	0x14(%ebp)
	pushl	0x10(%ebp)
	pushl	0xc(%ebp)
	pushl	0x8(%ebp)
	pushl	dtrace_probeid_error
	call	dtrace_probe
	movl	%ebp, %esp
	popl	%ebp
	ret
	END(dtrace_probe_error)

/*
void dtrace_membar_producer(void)
*/

	ENTRY(dtrace_membar_producer)
	rep;	ret	/* use 2 byte return instruction when branch target */
			/* AMD Software Optimization Guide - Section 6.2 */
	END(dtrace_membar_producer)

/*
void dtrace_membar_consumer(void)
*/

	ENTRY(dtrace_membar_consumer)
	rep;	ret	/* use 2 byte return instruction when branch target */
			/* AMD Software Optimization Guide - Section 6.2 */
	END(dtrace_membar_consumer)

/*
dtrace_icookie_t dtrace_interrupt_disable(void)
*/
	ENTRY(dtrace_interrupt_disable)
	pushfl
	popl	%eax
	cli
	ret
	END(dtrace_interrupt_disable)

/*
void dtrace_interrupt_enable(dtrace_icookie_t cookie)
*/
	ENTRY(dtrace_interrupt_enable)
	movl	4(%esp), %eax
	pushl	%eax
	popfl
	ret
	END(dtrace_interrupt_enable)