summaryrefslogtreecommitdiffhomepage
path: root/ir/be/ia32/ia32_fpu.c
blob: 84d207be3defcea9c523ea46666c526758528170 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
/*
 * This file is part of libFirm.
 * Copyright (C) 2012 University of Karlsruhe.
 */

/**
 * @file
 * @brief   Handles fpu rounding modes
 * @author  Matthias Braun
 *
 * The problem we deal with here is that the x86 ABI says the user can control
 * the fpu rounding mode, which means that when we do some operations like float
 * to int conversion which are specified as truncation in the C standard we have
 * to spill, change and restore the fpu rounding mode between spills.
 */
#include "array.h"
#include "bearch.h"
#include "benode.h"
#include "besched.h"
#include "bessaconstr.h"
#include "bestate.h"
#include "gen_ia32_regalloc_if.h"
#include "ia32_architecture.h"
#include "ia32_bearch_t.h"
#include "ia32_new_nodes.h"
#include "ia32_transform.h"
#include "ircons.h"
#include "irgwalk.h"
#include "tv.h"

static ir_entity *fpcw_round    = NULL;
static ir_entity *fpcw_truncate = NULL;

static ir_entity *create_ent(ir_entity **const dst, int value, const char *name)
{
	if (!*dst) {
		ir_mode   *const mode = mode_Hu;
		ir_type   *const type = get_type_for_mode(mode);
		ir_type   *const glob = get_glob_type();
		ident     *const id   = new_id_from_str(name);
		ir_entity *const ent
			= new_global_entity(glob, id, type, ir_visibility_local,
			                    IR_LINKAGE_CONSTANT | IR_LINKAGE_NO_IDENTITY);

		ir_tarval        *const cnst = new_tarval_from_long(value, mode);
		ir_initializer_t *const init = create_initializer_tarval(cnst);
		set_entity_initializer(ent, init);
		*dst = ent;
	}
	return *dst;
}

static ir_node *create_fnstcw(ir_node *const block, ir_node *const frame, ir_node *const noreg, ir_node *const nomem, ir_node *const state)
{
	ir_node *const fnstcw = new_bd_ia32_FnstCW(NULL, block, frame, noreg, nomem,
	                                           state);
	ia32_attr_t *const attr = get_ia32_attr(fnstcw);
	attr->addr.variant = X86_ADDR_BASE;
	set_ia32_op_type(fnstcw, ia32_AddrModeD);
	set_ia32_frame_use(fnstcw, IA32_FRAME_USE_32BIT);
	return fnstcw;
}

static ir_node *create_fpu_mode_spill(void *const env, ir_node *const state, bool const force, ir_node *const after)
{
	(void)env;

	if (!force && is_ia32_ChangeCW(state))
		return NULL;

	ir_node       *spill;
	ir_node *const block = get_nodes_block(state);
	/* Don't spill the fpcw in unsafe mode. */
	if (ia32_cg_config.use_unsafe_floatconv) {
		spill = new_bd_ia32_FnstCWNOP(NULL, block, state);
	} else {
		ir_graph *const irg   = get_irn_irg(state);
		ir_node  *const noreg = ia32_new_NoReg_gp(irg);
		ir_node  *const nomem = get_irg_no_mem(irg);
		ir_node  *const frame = get_irg_frame(irg);
		spill = create_fnstcw(block, frame, noreg, nomem, state);
	}
	sched_add_after(skip_Proj(after), spill);
	return spill;
}

static ir_node *create_fpu_mode_reload(void *const env, ir_node *const state, ir_node *const spill, ir_node *const before, ir_node *const last_state)
{
	(void)env;
	(void)state;

	ir_node        *reload;
	ir_node  *const block = get_nodes_block(before);
	ir_graph *const irg   = get_irn_irg(block);
	ir_node  *const noreg = ia32_new_NoReg_gp(irg);
	ir_node  *const nomem = get_irg_no_mem(irg);
	if (ia32_cg_config.use_unsafe_floatconv) {
		reload = new_bd_ia32_FldCW(NULL, block, noreg, noreg, nomem);
		ir_entity *const rounding_mode = spill ?
			create_ent(&fpcw_round,    0xC7F, "_fpcw_round") :
			create_ent(&fpcw_truncate, 0x37F, "_fpcw_truncate");
		ia32_attr_t *const attr = get_ia32_attr(reload);
		attr->addr.immediate.entity = rounding_mode;
		attr->addr.immediate.kind   = X86_IMM_ADDR;
		attr->addr.variant          = X86_ADDR_JUST_IMM;
	} else {
		ir_node       *mem;
		ir_node *const frame = get_irg_frame(irg);
		if (spill) {
			mem = spill;
		} else {
			assert(last_state);
			ir_node *const cwstore = create_fnstcw(block, frame, noreg, nomem, last_state);
			sched_add_before(before, cwstore);

			ir_node *const load = new_bd_ia32_Load(NULL, block, frame, noreg, cwstore, X86_SIZE_16, false);
			ia32_attr_t *const load_attr = get_ia32_attr(load);
			load_attr->addr.variant = X86_ADDR_BASE;
			set_ia32_op_type(load, ia32_AddrModeS);
			set_ia32_frame_use(load, IA32_FRAME_USE_32BIT);
			sched_add_before(before, load);

			ir_node *const load_res = be_new_Proj(load, pn_ia32_Load_res);

			/* TODO: Make the actual mode configurable in ChangeCW. */
			ir_node *const or_const = ia32_create_Immediate(irg, 0xC00);
			ir_node *const orn      = new_bd_ia32_Or(NULL, block, noreg, noreg,
			                                         nomem, load_res, or_const,
			                                         X86_SIZE_32);
			sched_add_before(before, orn);

			ir_node *const store = new_bd_ia32_Store(NULL, block, frame, noreg,
			                                         nomem, orn, X86_SIZE_32);
			ia32_attr_t *const store_attr = get_ia32_attr(store);
			/* Use ia32_mode_gp, as movl has a shorter opcode than movw. */
			store_attr->addr.variant = X86_ADDR_BASE;
			set_ia32_op_type(store, ia32_AddrModeD);
			set_ia32_frame_use(store, IA32_FRAME_USE_32BIT);
			sched_add_before(before, store);
			mem = be_new_Proj(store, pn_ia32_Store_M);
		}

		reload = new_bd_ia32_FldCW(NULL, block, frame, noreg, mem);
		ia32_attr_t *const attr = get_ia32_attr(reload);
		attr->addr.variant = X86_ADDR_BASE;
		set_ia32_frame_use(reload, IA32_FRAME_USE_32BIT);
	}

	get_ia32_attr(reload)->size = X86_SIZE_16;
	set_ia32_op_type(reload, ia32_AddrModeS);
	arch_set_irn_register(reload, &ia32_registers[REG_FPCW]);
	sched_add_before(before, reload);
	return reload;
}

typedef struct collect_fpu_mode_nodes_env_t {
	ir_node **state_nodes;
} collect_fpu_mode_nodes_env_t;

static void collect_fpu_mode_nodes_walker(ir_node *node, void *data)
{
	if (is_Proj(node) || is_ia32_ChangeCW(node))
		return;

	collect_fpu_mode_nodes_env_t *env = (collect_fpu_mode_nodes_env_t*)data;
	be_foreach_out(node, o) {
		const arch_register_t *reg = arch_get_irn_register_out(node, o);
		if (reg != &ia32_registers[REG_FPCW])
			continue;
		ir_node *value = node;
		if (get_irn_mode(value) == mode_T)
			value = be_get_or_make_Proj_for_pn(node, o);
		ARR_APP1(ir_node*, env->state_nodes, value);
	}
}

static void rewire_fpu_mode_nodes(ir_graph *irg)
{
	/* do ssa construction for the fpu modes */
	collect_fpu_mode_nodes_env_t env;
	env.state_nodes = NEW_ARR_F(ir_node*, 0);
	irg_walk_graph(irg, collect_fpu_mode_nodes_walker, NULL, &env);

	/* nothing needs to be done, in fact we must not continue as for endless
	 * loops noone is using the initial_value and it will point to a bad node
	 * now */
	if (ARR_LEN(env.state_nodes) == 0) {
		DEL_ARR_F(env.state_nodes);
		return;
	}

	be_ssa_construction_env_t senv;
	be_ssa_construction_init(&senv, irg);
	be_ssa_construction_add_copies(&senv, env.state_nodes,
	                               ARR_LEN(env.state_nodes));
	arch_register_t const *const reg           = &ia32_registers[REG_FPCW];
	ir_node               *const initial_value = be_get_Start_proj(irg, reg);
	be_ssa_construction_fix_users(&senv, initial_value);

	/* set registers for the phis */
	ir_node **phis = be_ssa_construction_get_new_phis(&senv);
	for (size_t i = 0, len = ARR_LEN(phis); i < len; ++i) {
		ir_node *phi = phis[i];
		arch_set_irn_register(phi, reg);
	}
	be_ssa_construction_destroy(&senv);
	DEL_ARR_F(env.state_nodes);

	be_invalidate_live_sets(irg);
}

void ia32_setup_fpu_mode(ir_graph *irg)
{
	/* do ssa construction for the fpu modes */
	rewire_fpu_mode_nodes(irg);

	/* ensure correct fpu mode for operations */
	be_assure_state(irg, &ia32_registers[REG_FPCW],
	                NULL, create_fpu_mode_spill, create_fpu_mode_reload);
}