Skip to content
GitLab
Commit df437344 authored by Matthias Braun's avatar Matthias Braun
Browse files

fix Mulh, Div, IDiv handling

parent 4b56e387
Hide whitespace changes
Inline Side-by-side
ir/be/ia32/ia32_emitter.c
View file @ df437344
......@@ -343,7 +343,7 @@ void ia32_emit_x87_name(ia32_emit_env_t *env, const ir_node *node, int pos)
{
ia32_attr_t *attr = get_ia32_attr(node);
assert(pos < 7);
assert(pos < 3);
ia32_emit_char(env, '%');
ia32_emit_string(env, attr->x87[pos]->name);
}
......@@ -515,11 +515,10 @@ void ia32_emit_unop(ia32_emit_env_t *env, const ir_node *node) {
ia32_emit_char(env, '$');
ia32_emit_immediate(env, node);
} else {
if (is_ia32_IMul(node) || is_ia32_Mulh(node)) {
/* MulS and Mulh implicitly multiply by EAX */
if (is_ia32_Mul(node) || is_ia32_IMul1OP(node)) {
ia32_emit_source_register(env, node, 3);
} else if(is_ia32_IDiv(node)) {
ia32_emit_source_register(env, node, 1);
} else if(is_ia32_IDiv(node) || is_ia32_Div(node)) {
ia32_emit_source_register(env, node, 4);
} else if(is_ia32_Push(node)) {
ia32_emit_source_register(env, node, 2);
} else if(is_ia32_Pop(node)) {
......
ir/be/ia32/ia32_spec.pl
View file @ df437344
......@@ -114,7 +114,7 @@ $arch = "ia32";
{ "name" => "ebp", "type" => 2 },
{ "name" => "esp", "type" => 4 },
{ "name" => "gp_NOREG", "type" => 4 | 8 | 16 }, # we need a dummy register for NoReg nodes
{ "name" => "gp_UKNWN", "type" => 4 | 8 | 16}, # we need a dummy register for Unknown nodes
{ "name" => "gp_UKNWN", "type" => 4 | 8 | 16 }, # we need a dummy register for Unknown nodes
{ "mode" => "mode_Iu" }
],
"xmm" => [
......@@ -127,7 +127,7 @@ $arch = "ia32";
{ "name" => "xmm6", "type" => 1 },
{ "name" => "xmm7", "type" => 1 },
{ "name" => "xmm_NOREG", "type" => 4 | 8 | 16 }, # we need a dummy register for NoReg nodes
{ "name" => "xmm_UKNWN", "type" => 4 | 8 | 16}, # we need a dummy register for Unknown nodes
{ "name" => "xmm_UKNWN", "type" => 4 | 8 | 16 }, # we need a dummy register for Unknown nodes
{ "mode" => "mode_E" }
],
"vfp" => [
......@@ -140,7 +140,7 @@ $arch = "ia32";
{ "name" => "vf6", "type" => 1 | 16 },
{ "name" => "vf7", "type" => 1 | 16 },
{ "name" => "vfp_NOREG", "type" => 4 | 8 | 16 }, # we need a dummy register for NoReg nodes
{ "name" => "vfp_UKNWN", "type" => 4 | 8 | 16}, # we need a dummy register for Unknown nodes
{ "name" => "vfp_UKNWN", "type" => 4 | 8 | 16 }, # we need a dummy register for Unknown nodes
{ "mode" => "mode_E" }
],
"st" => [
......@@ -323,6 +323,16 @@ $default_cmp_attr = "return ia32_compare_attr(attr_a, attr_b);";
"mode" => "mode_Iu",
},
"IMul1OP" => {
"irn_flags" => "R",
"comment" => "construct Mul (1 operand format): Mul(a, b) = Mul(b, a) = a * b",
"reg_req" => { "in" => [ "gp", "gp", "eax", "gp", "none" ], "out" => [ "eax", "edx", "none" ] },
"emit" => '. imull %unop',
"outs" => [ "EAX", "EDX", "M" ],
"latency" => 5,
"units" => [ "GP" ],
},
"l_IMul" => {
"op_flags" => "C",
"cmp_attr" => "return 1;",
......@@ -330,20 +340,6 @@ $default_cmp_attr = "return ia32_compare_attr(attr_a, attr_b);";
"arity" => 2
},
# Mulh is an exception from the 4 INs with AM because the target is always EAX:EDX
# Matze: It's not clear to me why we have a separate Mulh node and not just use
# the IMul node...
"Mulh" => {
# we should not rematrialize this node. It produces 2 results and has
# very strict constrains
"comment" => "construct Mul: Mul(a, b) = Mul(b, a) = a * b",
"reg_req" => { "in" => [ "gp", "gp", "eax", "gp", "none" ], "out" => [ "eax", "edx" ] },
"emit" => '. imull %unop',
"outs" => [ "EAX", "EDX", "M" ],
"latency" => 5,
"units" => [ "GP" ],
},
"And" => {
"irn_flags" => "R",
"comment" => "construct And: And(a, b) = And(b, a) = a AND b",
......@@ -428,7 +424,7 @@ $default_cmp_attr = "return ia32_compare_attr(attr_a, attr_b);";
"IDiv" => {
"op_flags" => "F|L",
"state" => "exc_pinned",
"reg_req" => { "in" => [ "eax", "gp", "edx", "none" ], "out" => [ "eax", "edx", "none" ] },
"reg_req" => { "in" => [ "gp", "gp", "eax", "edx", "gp", "none" ], "out" => [ "eax", "edx", "none" ] },
"attr" => "ia32_op_flavour_t dm_flav",
"init_attr" => "attr->data.op_flav = dm_flav;",
"emit" => ". idivl %unop",
......@@ -440,7 +436,7 @@ $default_cmp_attr = "return ia32_compare_attr(attr_a, attr_b);";
"Div" => {
"op_flags" => "F|L",
"state" => "exc_pinned",
"reg_req" => { "in" => [ "eax", "gp", "edx", "none" ], "out" => [ "eax", "edx", "none" ] },
"reg_req" => { "in" => [ "gp", "gp", "eax", "edx", "gp", "none" ], "out" => [ "eax", "edx", "none" ] },
"attr" => "ia32_op_flavour_t dm_flav",
"init_attr" => "attr->data.op_flav = dm_flav;",
"emit" => ". divl %unop",
......@@ -1125,7 +1121,7 @@ if (get_ia32_immop_type(node) == ia32_ImmNone) {
"state" => "exc_pinned",
"comment" => "store ST0 onto stack",
"reg_req" => { "in" => [ "gp", "gp", "none" ] },
"emit" => '. fstp%M %AM',
"emit" => '. fstp%XM %AM',
"latency" => 4,
"units" => [ "SSE" ],
"mode" => "mode_M",
......@@ -1726,7 +1722,7 @@ if (get_ia32_immop_type(node) == ia32_ImmNone) {
"state" => "exc_pinned",
"comment" => "x87 fp Load: Load(ptr, mem) = LD ptr -> reg",
"reg_req" => { },
"emit" => '. fld%M %AM',
"emit" => '. fld%XM %AM',
},
"fst" => {
......@@ -1735,7 +1731,7 @@ if (get_ia32_immop_type(node) == ia32_ImmNone) {
"state" => "exc_pinned",
"comment" => "x87 fp Store: Store(ptr, val, mem) = ST ptr,val",
"reg_req" => { },
"emit" => '. fst%M %AM',
"emit" => '. fst%XM %AM',
"mode" => "mode_M",
},
......@@ -1745,7 +1741,7 @@ if (get_ia32_immop_type(node) == ia32_ImmNone) {
"state" => "exc_pinned",
"comment" => "x87 fp Store: Store(ptr, val, mem) = ST ptr,val",
"reg_req" => { },
"emit" => '. fstp%M %AM',
"emit" => '. fstp%XM %AM',
"mode" => "mode_M",
},
......@@ -1756,7 +1752,7 @@ if (get_ia32_immop_type(node) == ia32_ImmNone) {
"rd_constructor" => "NONE",
"comment" => "x87 fp integer Load: Load(ptr, mem) = iLD ptr -> reg",
"reg_req" => { },
"emit" => '. fild%M %AM',
"emit" => '. fild%XM %AM',
},
"fist" => {
......
ir/be/ia32/ia32_transform.c
View file @ df437344
......@@ -521,7 +521,7 @@ static ir_node *gen_binop(ia32_transform_env_t *env, ir_node *node,
ir_node *new_op2 = transform_node(env, op2);
new_node = func(dbg, irg, block, noreg_gp, noreg_gp, new_op1, new_op2, nomem);
if(func == new_rd_ia32_Mul) {
if(func == new_rd_ia32_IMul) {
set_ia32_am_support(new_node, ia32_am_Source);
} else {
set_ia32_am_support(new_node, ia32_am_Full);
......@@ -884,18 +884,24 @@ static ir_node *gen_Mulh(ia32_transform_env_t *env, ir_node *node) {
ir_node *new_op1 = transform_node(env, op1);
ir_node *new_op2 = transform_node(env, op2);
ir_node *noreg = ia32_new_NoReg_gp(env->cg);
ir_node *proj_EAX, *proj_EDX, *mulh;
ir_node *proj_EAX, *proj_EDX, *res;
ir_mode *mode = get_irn_mode(node);
ir_node *in[1];
assert(!mode_is_float(mode) && "Mulh with float not supported");
mulh = new_rd_ia32_Mulh(dbg, irg, block, noreg, noreg, new_op1, new_op2, new_NoMem());
set_ia32_commutative(mulh);
set_ia32_am_support(mulh, ia32_am_Source);
if(mode_is_signed(mode)) {
res = new_rd_ia32_IMul1OP(dbg, irg, block, noreg, noreg, new_op1, new_op2, new_NoMem());
} else {
res = new_rd_ia32_Mul(dbg, irg, block, noreg, noreg, new_op1, new_op2, new_NoMem());
}
/* imediates are not supported, so no fold_immediate */
proj_EAX = new_rd_Proj(dbg, irg, block, mulh, mode_Iu, pn_EAX);
proj_EDX = new_rd_Proj(dbg, irg, block, mulh, mode_Iu, pn_EDX);
set_ia32_commutative(res);
set_ia32_am_support(res, ia32_am_Source);
set_ia32_am_support(res, ia32_am_Source);
proj_EAX = new_rd_Proj(dbg, irg, block, res, mode_Iu, pn_EAX);
proj_EDX = new_rd_Proj(dbg, irg, block, res, mode_Iu, pn_EDX);
/* keep EAX */
in[0] = proj_EAX;
......@@ -1196,6 +1202,7 @@ static ir_node *generate_DivMod(ia32_transform_env_t *env, ir_node *node,
ir_node *in_keep[1];
ir_node *mem, *new_mem;
ir_node *projs[pn_DivMod_max];
ir_node *noreg = ia32_new_NoReg_gp(env->cg);
ir_node *new_dividend = transform_node(env, dividend);
ir_node *new_divisor = transform_node(env, divisor);
......@@ -1233,10 +1240,17 @@ static ir_node *generate_DivMod(ia32_transform_env_t *env, ir_node *node,
}
if(mode_is_signed(mode)) {
res = new_rd_ia32_IDiv(dbg, irg, block, new_dividend, new_divisor, edx_node, new_mem, dm_flav);
res = new_rd_ia32_IDiv(dbg, irg, block, noreg, noreg, new_dividend, edx_node, new_divisor, new_mem, dm_flav);
} else {
res = new_rd_ia32_Div(dbg, irg, block, new_dividend, new_divisor, edx_node, new_mem, dm_flav);
res = new_rd_ia32_Div(dbg, irg, block, noreg, noreg, new_dividend, edx_node, new_divisor, new_mem, dm_flav);
}
/* Matze: code can't handle this at the moment... */
#if 0
/* set AM support */
set_ia32_am_support(res, ia32_am_Source);
#endif
set_ia32_n_res(res, 2);
/* Only one proj is used -> We must add a second proj and */
......@@ -3387,7 +3401,7 @@ static ir_node *gen_Proj_be_AddSP(ia32_transform_env_t *env, ir_node *node) {
dbg_info *dbg = get_irn_dbg_info(node);
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
if(proj == pn_be_AddSP_res) {
ir_node *res = new_rd_Proj(dbg, irg, block, new_pred, mode_Iu, pn_ia32_AddSP_stack);
......@@ -3407,7 +3421,7 @@ static ir_node *gen_Proj_be_SubSP(ia32_transform_env_t *env, ir_node *node) {
dbg_info *dbg = get_irn_dbg_info(node);
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
if(proj == pn_be_SubSP_res) {
ir_node *res = new_rd_Proj(dbg, irg, block, new_pred, mode_Iu, pn_ia32_AddSP_stack);
......@@ -3427,7 +3441,7 @@ static ir_node *gen_Proj_Load(ia32_transform_env_t *env, ir_node *node) {
dbg_info *dbg = get_irn_dbg_info(node);
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
/* renumber the proj */
if(is_ia32_Load(new_pred)) {
......@@ -3462,7 +3476,7 @@ static ir_node *gen_Proj_DivMod(ia32_transform_env_t *env, ir_node *node) {
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
assert(is_ia32_Div(new_pred) || is_ia32_IDiv(new_pred));
......@@ -3516,7 +3530,7 @@ static ir_node *gen_Proj_CopyB(ia32_transform_env_t *env, ir_node *node)
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
switch(proj) {
case pn_CopyB_M_regular:
......@@ -3545,7 +3559,7 @@ static ir_node *gen_Proj_l_vfdiv(ia32_transform_env_t *env, ir_node *node)
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
switch(proj) {
case pn_ia32_l_vfdiv_M:
......@@ -3568,7 +3582,7 @@ static ir_node *gen_Proj_Quot(ia32_transform_env_t *env, ir_node *node)
ir_node *pred = get_Proj_pred(node);
ir_node *new_pred = transform_node(env, pred);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
switch(proj) {
case pn_Quot_M:
......@@ -3607,11 +3621,30 @@ static ir_node *gen_Proj_tls(ia32_transform_env_t *env, ir_node *node) {
return res;
}
static ir_node *gen_Proj_be_Call(ia32_transform_env_t *env, ir_node *node) {
ir_graph *irg = env->irg;
dbg_info *dbg = get_irn_dbg_info(node);
ir_node *pred = get_Proj_pred(node);
long proj = get_Proj_proj(node);
ir_mode *mode = get_irn_mode(node);
ir_node *block = transform_node(env, get_nodes_block(node));
ir_node *new_pred = transform_node(env, pred);
/* transform call modes to the mode_Iu or mode_E */
if(mode_is_float(mode)) {
mode = mode_E;
} else if(mode != mode_M) {
mode = mode_Iu;
}
return new_rd_Proj(dbg, irg, block, new_pred, mode, proj);
}
static ir_node *gen_Proj(ia32_transform_env_t *env, ir_node *node) {
ir_graph *irg = env->irg;
dbg_info *dbg = get_irn_dbg_info(node);
ir_node *pred = get_Proj_pred(node);
int proj = get_Proj_proj(node);
long proj = get_Proj_proj(node);
if(is_Store(pred) || be_is_FrameStore(pred)) {
if(proj == pn_Store_M) {
......@@ -3634,6 +3667,8 @@ static ir_node *gen_Proj(ia32_transform_env_t *env, ir_node *node) {
return gen_Proj_be_SubSP(env, node);
} else if(be_is_AddSP(pred)) {
return gen_Proj_be_AddSP(env, node);
} else if(be_is_Call(pred)) {
return gen_Proj_be_Call(env, node);
} else if(get_irn_op(pred) == op_Start) {
if(proj == pn_Start_X_initial_exec) {
ir_node *block = get_nodes_block(pred);
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment

Imprint