x86 codegen source file skeletons
Add source files for x86 code generation. These are basically just
the MIPS versions with all bodies ifdef'd out and replaced with
UNIMPLEMENTED(WARNING).
There is also one real change in this CL related to the way the
assembler computes instruction size. We now set the initial
instruction size into the LIR during flag generation. For x86, this
size may be changed during assembly when we detect the ability to
substitute a general instruction for an equivalent one with a shorter
encoding.
NOTE: the new files are not yet ready for compilation. That will be
the next CL.
Change-Id: Id94fbeeb7c3e36ba4a8d93c3807351ea20c74133
diff --git a/src/compiler/codegen/x86/Assemble.cc b/src/compiler/codegen/x86/Assemble.cc
new file mode 100644
index 0000000..3614dce
--- /dev/null
+++ b/src/compiler/codegen/x86/Assemble.cc
@@ -0,0 +1,324 @@
+/*
+ * Copyright (C) 2012 The Android Open Source Project
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include "../../Dalvik.h"
+#include "../../CompilerInternals.h"
+#include "X86LIR.h"
+#include "Codegen.h"
+#include <sys/mman.h> /* for protection change */
+
+namespace art {
+
+#define MAX_ASSEMBLER_RETRIES 50
+
+/*
+ * opcode: MipsOpCode enum
+ * skeleton: pre-designated bit-pattern for this opcode
+ * k0: key to applying ds/de
+ * ds: dest start bit position
+ * de: dest end bit position
+ * k1: key to applying s1s/s1e
+ * s1s: src1 start bit position
+ * s1e: src1 end bit position
+ * k2: key to applying s2s/s2e
+ * s2s: src2 start bit position
+ * s2e: src2 end bit position
+ * operands: number of operands (for sanity check purposes)
+ * name: mnemonic name
+ * fmt: for pretty-printing
+ */
+#define ENCODING_MAP(opcode, skeleton, k0, ds, de, k1, s1s, s1e, k2, s2s, s2e, \
+ k3, k3s, k3e, flags, name, fmt, size) \
+ {skeleton, {{k0, ds, de}, {k1, s1s, s1e}, {k2, s2s, s2e}, \
+ {k3, k3s, k3e}}, opcode, flags, name, fmt, size}
+
+/* Instruction dump string format keys: !pf, where "!" is the start
+ * of the key, "p" is which numeric operand to use and "f" is the
+ * print format.
+ *
+ * [p]ositions:
+ * 0 -> operands[0] (dest)
+ * 1 -> operands[1] (src1)
+ * 2 -> operands[2] (src2)
+ * 3 -> operands[3] (extra)
+ *
+ * [f]ormats:
+ * h -> 4-digit hex
+ * d -> decimal
+ * E -> decimal*4
+ * F -> decimal*2
+ * c -> branch condition (beq, bne, etc.)
+ * t -> pc-relative target
+ * T -> pc-region target
+ * u -> 1st half of bl[x] target
+ * v -> 2nd half ob bl[x] target
+ * R -> register list
+ * s -> single precision floating point register
+ * S -> double precision floating point register
+ * m -> Thumb2 modified immediate
+ * n -> complimented Thumb2 modified immediate
+ * M -> Thumb2 16-bit zero-extended immediate
+ * b -> 4-digit binary
+ * N -> append a NOP
+ *
+ * [!] escape. To insert "!", use "!!"
+ */
+/* NOTE: must be kept in sync with enum MipsOpcode from LIR.h */
+/*
+ * TUNING: We're currently punting on the branch delay slots. All branch
+ * instructions in this map are given a size of 8, which during assembly
+ * is expanded to include a nop. This scheme should be replaced with
+ * an assembler pass to fill those slots when possible.
+ */
+MipsEncodingMap EncodingMap[kX86Last] = {
+};
+
+
+/*
+ * Assemble the LIR into binary instruction format. Note that we may
+ * discover that pc-relative displacements may not fit the selected
+ * instruction. In those cases we will try to substitute a new code
+ * sequence or request that the trace be shortened and retried.
+ */
+AssemblerStatus oatAssembleInstructions(CompilationUnit *cUnit,
+ intptr_t startAddr)
+{
+ UNIMPLEMENTED(WARNING) << "oatAssembleInstructions";
+ return kSuccess;
+#if 0
+ LIR *lir;
+ AssemblerStatus res = kSuccess; // Assume success
+
+ for (lir = (LIR *) cUnit->firstLIRInsn; lir; lir = NEXT_LIR(lir)) {
+ if (lir->opcode < 0) {
+ continue;
+ }
+
+
+ if (lir->flags.isNop) {
+ continue;
+ }
+
+ if (lir->flags.pcRelFixup) {
+ if (lir->opcode == kMipsDelta) {
+ /*
+ * The "Delta" pseudo-ops load the difference between
+ * two pc-relative locations into a the target register
+ * found in operands[0]. The delta is determined by
+ * (label2 - label1), where label1 is a standard
+ * kPseudoTargetLabel and is stored in operands[2].
+ * If operands[3] is null, then label2 is a kPseudoTargetLabel
+ * and is found in lir->target. If operands[3] is non-NULL,
+ * then it is a Switch/Data table.
+ */
+ int offset1 = ((LIR*)lir->operands[2])->offset;
+ SwitchTable *tabRec = (SwitchTable*)lir->operands[3];
+ int offset2 = tabRec ? tabRec->offset : lir->target->offset;
+ int delta = offset2 - offset1;
+ if ((delta & 0xffff) == delta) {
+ // Fits
+ lir->operands[1] = delta;
+ } else {
+ // Doesn't fit - must expand to kMipsDelta[Hi|Lo] pair
+ LIR *newDeltaHi =
+ rawLIR(cUnit, lir->dalvikOffset, kMipsDeltaHi,
+ lir->operands[0], 0, lir->operands[2],
+ lir->operands[3], lir->target);
+ oatInsertLIRBefore((LIR*)lir, (LIR*)newDeltaHi);
+ LIR *newDeltaLo =
+ rawLIR(cUnit, lir->dalvikOffset, kMipsDeltaLo,
+ lir->operands[0], 0, lir->operands[2],
+ lir->operands[3], lir->target);
+ oatInsertLIRBefore((LIR*)lir, (LIR*)newDeltaLo);
+ lir->flags.isNop = true;
+ res = kRetryAll;
+ }
+ } else if (lir->opcode == kMipsDeltaLo) {
+ int offset1 = ((LIR*)lir->operands[2])->offset;
+ SwitchTable *tabRec = (SwitchTable*)lir->operands[3];
+ int offset2 = tabRec ? tabRec->offset : lir->target->offset;
+ int delta = offset2 - offset1;
+ lir->operands[1] = delta & 0xffff;
+ } else if (lir->opcode == kMipsDeltaHi) {
+ int offset1 = ((LIR*)lir->operands[2])->offset;
+ SwitchTable *tabRec = (SwitchTable*)lir->operands[3];
+ int offset2 = tabRec ? tabRec->offset : lir->target->offset;
+ int delta = offset2 - offset1;
+ lir->operands[1] = (delta >> 16) & 0xffff;
+ } else if (lir->opcode == kMipsB || lir->opcode == kMipsBal) {
+ LIR *targetLIR = (LIR *) lir->target;
+ intptr_t pc = lir->offset + 4;
+ intptr_t target = targetLIR->offset;
+ int delta = target - pc;
+ if (delta & 0x3) {
+ LOG(FATAL) << "PC-rel offset not multiple of 4: " << delta;
+ }
+ if (delta > 131068 || delta < -131069) {
+ res = kRetryAll;
+ convertShortToLongBranch(cUnit, lir);
+ } else {
+ lir->operands[0] = delta >> 2;
+ }
+ } else if (lir->opcode >= kMipsBeqz && lir->opcode <= kMipsBnez) {
+ LIR *targetLIR = (LIR *) lir->target;
+ intptr_t pc = lir->offset + 4;
+ intptr_t target = targetLIR->offset;
+ int delta = target - pc;
+ if (delta & 0x3) {
+ LOG(FATAL) << "PC-rel offset not multiple of 4: " << delta;
+ }
+ if (delta > 131068 || delta < -131069) {
+ res = kRetryAll;
+ convertShortToLongBranch(cUnit, lir);
+ } else {
+ lir->operands[1] = delta >> 2;
+ }
+ } else if (lir->opcode == kMipsBeq || lir->opcode == kMipsBne) {
+ LIR *targetLIR = (LIR *) lir->target;
+ intptr_t pc = lir->offset + 4;
+ intptr_t target = targetLIR->offset;
+ int delta = target - pc;
+ if (delta & 0x3) {
+ LOG(FATAL) << "PC-rel offset not multiple of 4: " << delta;
+ }
+ if (delta > 131068 || delta < -131069) {
+ res = kRetryAll;
+ convertShortToLongBranch(cUnit, lir);
+ } else {
+ lir->operands[2] = delta >> 2;
+ }
+ } else if (lir->opcode == kMipsJal) {
+ intptr_t curPC = (startAddr + lir->offset + 4) & ~3;
+ intptr_t target = lir->operands[0];
+ /* ensure PC-region branch can be used */
+ DCHECK_EQ((curPC & 0xF0000000), (target & 0xF0000000));
+ if (target & 0x3) {
+ LOG(FATAL) << "Jump target not multiple of 4: " << target;
+ }
+ lir->operands[0] = target >> 2;
+ } else if (lir->opcode == kMipsLahi) { /* ld address hi (via lui) */
+ LIR *targetLIR = (LIR *) lir->target;
+ intptr_t target = startAddr + targetLIR->offset;
+ lir->operands[1] = target >> 16;
+ } else if (lir->opcode == kMipsLalo) { /* ld address lo (via ori) */
+ LIR *targetLIR = (LIR *) lir->target;
+ intptr_t target = startAddr + targetLIR->offset;
+ lir->operands[2] = lir->operands[2] + target;
+ }
+ }
+
+ /*
+ * If one of the pc-relative instructions expanded we'll have
+ * to make another pass. Don't bother to fully assemble the
+ * instruction.
+ */
+ if (res != kSuccess) {
+ continue;
+ }
+ const MipsEncodingMap *encoder = &EncodingMap[lir->opcode];
+ u4 bits = encoder->skeleton;
+ int i;
+ for (i = 0; i < 4; i++) {
+ u4 operand;
+ u4 value;
+ operand = lir->operands[i];
+ switch(encoder->fieldLoc[i].kind) {
+ case kFmtUnused:
+ break;
+ case kFmtBitBlt:
+ if (encoder->fieldLoc[i].start == 0 && encoder->fieldLoc[i].end == 31) {
+ value = operand;
+ } else {
+ value = (operand << encoder->fieldLoc[i].start) &
+ ((1 << (encoder->fieldLoc[i].end + 1)) - 1);
+ }
+ bits |= value;
+ break;
+ case kFmtBlt5_2:
+ value = (operand & 0x1f);
+ bits |= (value << encoder->fieldLoc[i].start);
+ bits |= (value << encoder->fieldLoc[i].end);
+ break;
+ case kFmtDfp: {
+ DCHECK(DOUBLEREG(operand));
+ DCHECK((operand & 0x1) == 0);
+ value = ((operand & FP_REG_MASK) << encoder->fieldLoc[i].start) &
+ ((1 << (encoder->fieldLoc[i].end + 1)) - 1);
+ bits |= value;
+ break;
+ }
+ case kFmtSfp:
+ DCHECK(SINGLEREG(operand));
+ value = ((operand & FP_REG_MASK) << encoder->fieldLoc[i].start) &
+ ((1 << (encoder->fieldLoc[i].end + 1)) - 1);
+ bits |= value;
+ break;
+ default:
+ LOG(FATAL) << "Bad encoder format: "
+ << (int)encoder->fieldLoc[i].kind;
+ }
+ }
+ // FIXME: need multi-endian handling here
+ cUnit->codeBuffer.push_back((bits >> 16) & 0xffff);
+ cUnit->codeBuffer.push_back(bits & 0xffff);
+ // TUNING: replace with proper delay slot handling
+ if (encoder->size == 8) {
+ const MipsEncodingMap *encoder = &EncodingMap[kMipsNop];
+ u4 bits = encoder->skeleton;
+ cUnit->codeBuffer.push_back((bits >> 16) & 0xffff);
+ cUnit->codeBuffer.push_back(bits & 0xffff);
+ }
+ }
+ return res;
+#endif
+}
+
+int oatGetInsnSize(LIR* lir)
+{
+ return EncodingMap[lir->opcode].size;
+}
+/*
+ * Target-dependent offset assignment.
+ * independent.
+ */
+int oatAssignInsnOffsets(CompilationUnit* cUnit)
+{
+ LIR* x86LIR;
+ int offset = 0;
+
+ for (x86LIR = (LIR *) cUnit->firstLIRInsn;
+ x86LIR;
+ x86LIR = NEXT_LIR(x86LIR)) {
+ x86LIR->offset = offset;
+ if (x86LIR->opcode >= 0) {
+ if (!x86LIR->flags.isNop) {
+ offset += x86LIR->flags.size;
+ }
+ } else if (x86LIR->opcode == kPseudoPseudoAlign4) {
+ if (offset & 0x2) {
+ offset += 2;
+ x86LIR->operands[0] = 1;
+ } else {
+ x86LIR->operands[0] = 0;
+ }
+ }
+ /* Pseudo opcodes don't consume space */
+ }
+
+ return offset;
+}
+
+} // namespace art