Chromium Code Reviews Chromium Code Reviews
Issue 1694833002:
MIPS: Support r6 max, min floating point instructions. (Closed)
Base URL: https://chromium.googlesource.com/v8/v8.git@master| Index: src/mips/builtins-mips.cc |
| diff --git a/src/mips/builtins-mips.cc b/src/mips/builtins-mips.cc |
| index 99226e7da19bc996552abf71aedccd42cc6fd814..6935e90756d8af08e1d5a3839c34eb9df623ce02 100644 |
| --- a/src/mips/builtins-mips.cc |
| +++ b/src/mips/builtins-mips.cc |
| @@ -149,17 +149,15 @@ void Builtins::Generate_MathMaxMin(MacroAssembler* masm, MathMaxMinKind kind) { |
| // -- sp[(argc - n) * 8] : arg[n] (zero-based) |
| // -- sp[(argc + 1) * 8] : receiver |
| // ----------------------------------- |
| - Condition const cc = (kind == MathMaxMinKind::kMin) ? ge : le; |
| Heap::RootListIndex const root_index = |
| (kind == MathMaxMinKind::kMin) ? Heap::kInfinityValueRootIndex |
| : Heap::kMinusInfinityValueRootIndex; |
| - DoubleRegister const reg = (kind == MathMaxMinKind::kMin) ? f2 : f0; |
| // Load the accumulator with the default return value (either -Infinity or |
| // +Infinity), with the tagged value in a1 and the double value in f0. |
| __ LoadRoot(a1, root_index); |
| __ ldc1(f0, FieldMemOperand(a1, HeapNumber::kValueOffset)); |
| - __ mov(a3, a0); |
| + __ Addu(a3, a0, Operand(1)); |
| Label done_loop, loop; |
| __ bind(&loop); |
| @@ -182,8 +180,6 @@ void Builtins::Generate_MathMaxMin(MacroAssembler* masm, MathMaxMinKind kind) { |
| { |
| // Parameter is not a Number, use the ToNumberStub to convert it. |
| FrameScope scope(masm, StackFrame::INTERNAL); |
| - __ SmiTag(a0); |
|
Michael Lippautz
2016/03/29 15:20:58
This is not GC save. There exists a path (at least
|
| - __ SmiTag(a3); |
| __ Push(a0, a1, a3); |
| __ mov(a0, a2); |
| ToNumberStub stub(masm->isolate()); |
| @@ -200,8 +196,6 @@ void Builtins::Generate_MathMaxMin(MacroAssembler* masm, MathMaxMinKind kind) { |
| __ SmiToDoubleFPURegister(a1, f0, t0); |
| __ bind(&done_restore); |
| } |
| - __ SmiUntag(a3); |
| - __ SmiUntag(a0); |
| } |
| __ jmp(&convert); |
| __ bind(&convert_number); |
| @@ -211,21 +205,24 @@ void Builtins::Generate_MathMaxMin(MacroAssembler* masm, MathMaxMinKind kind) { |
| __ SmiToDoubleFPURegister(a2, f2, t0); |
| __ bind(&done_convert); |
| - // Perform the actual comparison with the accumulator value on the left hand |
| - // side (f0) and the next parameter value on the right hand side (f2). |
| - Label compare_equal, compare_nan, compare_swap; |
| - __ BranchF(&compare_equal, &compare_nan, eq, f0, f2); |
| - __ BranchF(&compare_swap, nullptr, cc, f0, f2); |
| - __ Branch(&loop); |
| - |
| - // Left and right hand side are equal, check for -0 vs. +0. |
| - __ bind(&compare_equal); |
| - __ FmoveHigh(t0, reg); |
| - __ Branch(&loop, ne, t0, Operand(0x80000000)); |
| - |
| - // Result is on the right hand side. |
| - __ bind(&compare_swap); |
| - __ mov_d(f0, f2); |
| + // Perform the actual comparison with using Min/Max macro instructions the |
| + // accumulator value on the left hand side (f0) and the next parameter value |
| + // on the right hand side (f2). |
| + // We need to work out which HeapNumber (or smi) the result came from. |
| + Label compare_nan, set_value; |
| + __ BranchF(nullptr, &compare_nan, eq, f0, f2); |
| + __ Move(t0, t1, f0); |
| + if (kind == MathMaxMinKind::kMin) { |
| + __ MinNaNCheck_d(f0, f0, f2); |
| + } else { |
| + DCHECK(kind == MathMaxMinKind::kMax); |
| + __ MaxNaNCheck_d(f0, f0, f2); |
| + } |
| + __ Move(at, t8, f0); |
| + __ Branch(&set_value, ne, t0, Operand(at)); |
| + __ Branch(&set_value, ne, t1, Operand(t8)); |
| + __ jmp(&loop); |
| + __ bind(&set_value); |
| __ mov(a1, a2); |
| __ jmp(&loop); |
| @@ -238,8 +235,8 @@ void Builtins::Generate_MathMaxMin(MacroAssembler* masm, MathMaxMinKind kind) { |
| __ bind(&done_loop); |
| __ Lsa(sp, sp, a3, kPointerSizeLog2); |
| - __ mov(v0, a1); |
| - __ DropAndRet(1); |
| + __ Ret(USE_DELAY_SLOT); |
| + __ mov(v0, a1); // In delay slot. |
| } |
| // static |
