/**
* Try to place a Shl into an address mode.
*
* @param addr the address mode data so far
* @param node the node to place
* @return true on success
*/
static bool eat_shl(x86_address_t *addr, ir_node *node)
{
/* we can only eat a shl if we don't have a scale or index set yet */
if (addr->scale != 0 || addr->index != NULL)
return false;
ir_node *shifted_val;
long val;
if (is_Shl(node)) {
/* we can use shl with 0,1,2 or 3 shift */
ir_node *right = get_Shl_right(node);
if (!is_Const(right))
return false;
ir_tarval *tv = get_Const_tarval(right);
if (!tarval_is_long(tv))
return false;
val = get_tarval_long(tv);
if (val < 0 || val > 3)
return false;
if (val == 0)
be_warningf(node, "found unoptimized Shl x,0");
shifted_val = get_Shl_left(node);
} else if (is_Add(node)) {
/* might be an add x, x */
ir_node *left = get_Add_left(node);
ir_node *right = get_Add_right(node);
if (left != right)
return false;
if (is_Const(left))
return false;
val = 1;
shifted_val = left;
} else {
return false;
}
if (x86_is_non_address_mode_node(node))
return false;
addr->scale = val;
addr->index = shifted_val;
return true;
}
static bool eat_imm(x86_address_t *const addr, ir_node const *const node)
{
switch (get_irn_opcode(node)) {
case iro_Add:
/* Add is supported as long as both operands are immediates. */
return
!x86_is_non_address_mode_node(node) &&
eat_imm(addr, get_Add_left(node)) &&
eat_imm(addr, get_Add_right(node));
case iro_Address:
/* The first Address of a DAG can be folded into an immediate. */
if (addr->imm.entity)
return false;
addr->imm.entity = get_Address_entity(node);
addr->imm.kind = X86_IMM_ADDR;
if (is_tls_entity(addr->imm.entity))
addr->tls_segment = true;
return true;
case iro_Const: {
/* Add the value to the offset. */
ir_tarval *const tv = get_Const_tarval(node);
if (!tarval_possible(tv))
return false;
addr->imm.offset += get_tarval_long(tv);
return true;
}
case iro_Unknown:
/* Use '0' for Unknowns. */
return true;
default:
if (be_is_Relocation(node)) {
if (addr->imm.entity)
return false;
addr->imm.entity = be_get_Relocation_entity(node);
addr->imm.kind = (x86_immediate_kind_t)be_get_Relocation_kind(node);
return true;
}
/* All other nodes are no immediates. */
return false;
}
}
/**
* Process a call node.
*
* @param call A ir_node to be checked.
* @param callee The entity of the callee
* @param hmap The quadruple-set containing the calls with constant parameters
*/
static void process_call(ir_node *call, ir_entity *callee, q_set *hmap)
{
/* TODO
* Beware: We cannot clone variadic parameters as well as the
* last non-variadic one, which might be needed for the va_start()
* magic. */
/* In this for loop we collect the calls, that have
a constant parameter. */
size_t const n_params = get_Call_n_params(call);
for (size_t i = n_params; i-- > 0;) {
ir_node *const call_param = get_Call_param(call, i);
if (is_Const(call_param)) {
/* we have found a Call to collect and we save the information
* we need.*/
if (!hmap->map)
hmap->map = new_pset(entry_cmp, 8);
entry_t *const key = OALLOC(&hmap->obst, entry_t);
key->q.ent = callee;
key->q.pos = i;
key->q.tv = get_Const_tarval(call_param);
key->q.calls = NULL;
key->weight = 0.0F;
key->next = NULL;
/* Insert entry or get existing equivalent entry */
entry_t *const entry = (entry_t*)pset_insert(hmap->map, key, hash_entry(key));
/* Free memory if entry already is in set */
if (entry != key)
obstack_free(&hmap->obst, key);
/* add the call to the list */
if (!entry->q.calls) {
entry->q.calls = NEW_ARR_F(ir_node*, 1);
entry->q.calls[0] = call;
} else {
ARR_APP1(ir_node *, entry->q.calls, call);
}
}
Worklist::Worklist(ir_graph* functionGraph, GraphHandler& handler): functionGraph(functionGraph), handler(handler)
{
typedef void (*ir_func)(ir_node*, void*);
struct envMembers
{
std::queue<ir_node*>* pQueue;
std::unordered_map<ir_node*, bool>* pIsQueued;
};
envMembers envInstance;
envInstance.pQueue = &this->worklist;
envInstance.pIsQueued = &this->isQueued;
ir_func addPhis = [](ir_node * node, void* env)
{
if (is_Phi(node))
{
auto envInstance = (envMembers*)env;
set_irn_link(node, (void*)tarval_unknown);
envInstance->pQueue->push(node);
(*envInstance->pIsQueued)[node] = true;
}
};
ir_func addToWorklist = [](ir_node * node, void* env)
{
if (is_Phi(node))
return;
auto envInstance = (envMembers*)env;
ir_tarval* tarval;
auto isBadNode = [&] (Node node) -> bool
{
if (is_Call(node) || is_Load(node) || is_Start(node))
return true;
/*else if (Node(node).getChildCount() > 0)
{
for (Node child : Node(node).getChildren())
if (child.getTarval() == tarval_bad)
return true;
}*/
return false;
};
// TODO: Support other modes such as Bu, Lu
if (is_Const(node) && Node(node).getTarval().isNumeric())
tarval = get_Const_tarval(node);
else if (isBadNode(Node(node)))
tarval = tarval_bad;
else
tarval = tarval_unknown;
set_irn_link(node, (void*)tarval);
envInstance->pQueue->push(node);
(*envInstance->pIsQueued)[node] = true;
};
walk_topological(functionGraph, addPhis, (void*)&envInstance);
walk_topological(functionGraph, addToWorklist, (void*)&envInstance);
}
ir_node *copy_const_value(dbg_info *dbg, ir_node *n, ir_node *block)
{
ir_graph *irg = get_irn_irg(block);
/* @@@ GL I think we should implement this using the routines from irgopt
* for dead node elimination/inlineing. */
ir_mode *m = get_irn_mode(n);
ir_node *nn;
switch (get_irn_opcode(n)) {
case iro_Const:
nn = new_rd_Const(dbg, irg, get_Const_tarval(n));
break;
case iro_SymConst:
nn = new_rd_SymConst(dbg, irg, get_irn_mode(n), get_SymConst_symbol(n), get_SymConst_kind(n));
break;
case iro_Add:
nn = new_rd_Add(dbg, block,
copy_const_value(dbg, get_Add_left(n), block),
copy_const_value(dbg, get_Add_right(n), block), m);
break;
case iro_Sub:
nn = new_rd_Sub(dbg, block,
copy_const_value(dbg, get_Sub_left(n), block),
copy_const_value(dbg, get_Sub_right(n), block), m);
break;
case iro_Mul:
nn = new_rd_Mul(dbg, block,
copy_const_value(dbg, get_Mul_left(n), block),
copy_const_value(dbg, get_Mul_right(n), block), m);
break;
case iro_And:
nn = new_rd_And(dbg, block,
copy_const_value(dbg, get_And_left(n), block),
copy_const_value(dbg, get_And_right(n), block), m);
break;
case iro_Or:
nn = new_rd_Or(dbg, block,
copy_const_value(dbg, get_Or_left(n), block),
copy_const_value(dbg, get_Or_right(n), block), m);
break;
case iro_Eor:
nn = new_rd_Eor(dbg, block,
copy_const_value(dbg, get_Eor_left(n), block),
copy_const_value(dbg, get_Eor_right(n), block), m);
break;
case iro_Conv:
nn = new_rd_Conv(dbg, block,
copy_const_value(dbg, get_Conv_op(n), block), m);
break;
case iro_Minus:
nn = new_rd_Minus(dbg, block,
copy_const_value(dbg, get_Minus_op(n), block), m);
break;
case iro_Not:
nn = new_rd_Not(dbg, block,
copy_const_value(dbg, get_Not_op(n), block), m);
break;
case iro_Unknown:
nn = new_r_Unknown(irg, m);
break;
default:
panic("opcode invalid or not implemented %+F", n);
}
return nn;
}
void x86_create_address_mode(x86_address_t *addr, ir_node *node,
x86_create_am_flags_t flags)
{
addr->imm.kind = X86_IMM_VALUE;
if (eat_immediate(addr, node, true)) {
addr->variant = addr->ip_base ? X86_ADDR_RIP : X86_ADDR_JUST_IMM;
return;
}
assert(!addr->ip_base);
if (!(flags & x86_create_am_force) && x86_is_non_address_mode_node(node)
&& (!(flags & x86_create_am_double_use) || get_irn_n_edges(node) > 2)) {
addr->variant = X86_ADDR_BASE;
addr->base = node;
return;
}
ir_node *eat_imms = eat_immediates(addr, node, flags, false);
if (eat_imms != node) {
if (flags & x86_create_am_force)
eat_imms = be_skip_downconv(eat_imms, true);
node = eat_imms;
if (x86_is_non_address_mode_node(node)) {
addr->variant = X86_ADDR_BASE;
addr->base = node;
return;
}
}
/* starting point Add, Sub or Shl, FrameAddr */
if (is_Shl(node)) {
/* We don't want to eat add x, x as shl here, so only test for real Shl
* instructions, because we want the former as Lea x, x, not Shl x, 1 */
if (eat_shl(addr, node)) {
addr->variant = X86_ADDR_INDEX;
return;
}
} else if (eat_immediate(addr, node, true)) {
/* we can hit this case in x86_create_am_force mode */
addr->variant = addr->ip_base ? X86_ADDR_RIP : X86_ADDR_JUST_IMM;
return;
} else if (is_Add(node)) {
ir_node *left = get_Add_left(node);
ir_node *right = get_Add_right(node);
if (flags & x86_create_am_force) {
left = be_skip_downconv(left, true);
right = be_skip_downconv(right, true);
}
left = eat_immediates(addr, left, flags, false);
right = eat_immediates(addr, right, flags, false);
if (eat_shl(addr, left)) {
left = NULL;
} else if (eat_shl(addr, right)) {
right = NULL;
}
/* (x & 0xFFFFFFFC) + (x >> 2) -> lea(x >> 2, x >> 2, 4) */
if (left != NULL && right != NULL) {
ir_node *and;
ir_node *shr;
if (is_And(left) && (is_Shr(right) || is_Shrs(right))) {
and = left;
shr = right;
goto tryit;
}
if (is_And(right) && (is_Shr(left) || is_Shrs(left))) {
and = right;
shr = left;
tryit:
if (get_And_left(and) == get_binop_left(shr)) {
ir_node *and_right = get_And_right(and);
ir_node *shr_right = get_binop_right(shr);
if (is_Const(and_right) && is_Const(shr_right)) {
ir_tarval *and_mask = get_Const_tarval(and_right);
ir_tarval *shift_amount = get_Const_tarval(shr_right);
ir_mode *mode = get_irn_mode(and);
ir_tarval *all_one = get_mode_all_one(mode);
ir_tarval *shift_mask = tarval_shl(tarval_shr(all_one, shift_amount), shift_amount);
long val = get_tarval_long(shift_amount);
if (and_mask == shift_mask && val >= 0 && val <= 3) {
addr->variant = X86_ADDR_BASE_INDEX;
addr->base = shr;
addr->index = shr;
addr->scale = val;
return;
}
}
}
}
}
if (left != NULL) {
ir_node *base = addr->base;
if (base == NULL) {
addr->variant = addr->index != NULL ? X86_ADDR_BASE_INDEX
: X86_ADDR_BASE;
addr->base = left;
} else {
addr->variant = X86_ADDR_BASE_INDEX;
assert(addr->index == NULL && addr->scale == 0);
assert(right == NULL);
/* esp must be used as base */
if (is_Proj(left) && is_Start(get_Proj_pred(left))) {
addr->index = base;
addr->base = left;
} else {
addr->index = left;
}
}
}
if (right != NULL) {
ir_node *base = addr->base;
if (base == NULL) {
addr->variant = addr->index != NULL ? X86_ADDR_BASE_INDEX
: X86_ADDR_BASE;
addr->base = right;
} else {
addr->variant = X86_ADDR_BASE_INDEX;
assert(addr->index == NULL && addr->scale == 0);
/* esp must be used as base */
if (is_Proj(right) && is_Start(get_Proj_pred(right))) {
addr->index = base;
addr->base = right;
} else {
addr->index = right;
}
}
}
return;
}
addr->variant = X86_ADDR_BASE;
addr->base = node;
}