Eugene Zelenko via llvm-dev
2016-Dec-09 20:10 UTC
[llvm-dev] Weird code in MipsGenSubtargetInfo.inc
Hi! Clang-tidy modernize-use-bool-literals pointed to some weird code in: MipsGenSubtargetInfo::resolveSchedClass(). For example: case 3: // II_ADD if (SchedModel->getProcessorID() == 2) { // MipsP5600Model if ((1)) return 779; // P5600WriteALU if ((0)) return 780; // P5600WriteAL2 } break; I think will be good idea if somebody who knows MIPS and table generation specific will look onto this issue. Eugene.
Simon Atanasyan via llvm-dev
2016-Dec-10 06:56 UTC
[llvm-dev] Weird code in MipsGenSubtargetInfo.inc
Hi, On Fri, Dec 9, 2016 at 11:10 PM, Eugene Zelenko via llvm-dev <llvm-dev at lists.llvm.org> wrote:> Clang-tidy modernize-use-bool-literals pointed to some weird code in: > MipsGenSubtargetInfo::resolveSchedClass(). For example: > > case 3: // II_ADD > if (SchedModel->getProcessorID() == 2) { // MipsP5600Model > if ((1)) > return 779; // P5600WriteALU > if ((0)) > return 780; // P5600WriteAL2 > } > break; > > I think will be good idea if somebody who knows MIPS and table > generation specific will look onto this issue.This code corresponds "FIXME: Implement selection predicate" comments in the MipsScheduleP5600.td. The code was added in the r248725. -- Simon Atanasyan