peixin added a comment.
Thanks for this patch. This is one big progress to use `flang-new` for real
workloads. Mostly looks good to me. I have several minor comments. If supporing
`-Ofast` requires more changes, I prefer to push forward with the current patch.
summary: fronted -> frontend
Confirmed with the following case:
program m
integer :: x = 1, y = 2, z = 0
call add(x, y, z)
print *, z
contains
subroutine add(a, b, c)
integer :: a, b, c
integer :: i
do i = 1, 10
c = c + a + b
enddo
end
end
$ flang-new -fc1 -emit-llvm test.f90 -O3
$ cat test.ll
; ModuleID = 'FIRModule'
source_filename = "FIRModule"
target triple = "aarch64-unknown-linux-gnu"
@_QFEz = internal unnamed_addr global i32 0
@_QQcl.2E2F746573742E66393000 = linkonce constant [11 x i8] c"./test.f90\00"
define void @_QQmain() local_unnamed_addr !dbg !3 {
%1 = load i32, ptr @_QFEz, align 4, !dbg !7
%2 = add i32 %1, 30, !dbg !7
store i32 %2, ptr @_QFEz, align 4, !dbg !7
%3 = tail call ptr @_FortranAioBeginExternalListOutput(i32 -1, ptr nonnull
@_QQcl.2E2F746573742E66393000, i32 5), !dbg !12
%4 = tail call i1 @_FortranAioOutputInteger32(ptr %3, i32 %2), !dbg !13
%5 = tail call i32 @_FortranAioEndIoStatement(ptr %3), !dbg !12
ret void, !dbg !14
}
; Function Attrs: argmemonly nofree norecurse nosync nounwind
define void @_QFPadd(ptr nocapture readonly %0, ptr nocapture readonly %1,
ptr nocapture %2) local_unnamed_addr #0 !dbg !9 {
...
================
Comment at: clang/include/clang/Driver/Options.td:732
+def O_flag : Flag<["-"], "O">, Flags<[CC1Option,FC1Option]>, Alias<O>,
AliasArgs<["1"]>;
def Ofast : Joined<["-"], "Ofast">, Group<O_Group>, Flags<[CC1Option]>;
def P : Flag<["-"], "P">, Flags<[CC1Option,FlangOption,FC1Option]>,
Group<Preprocessor_Group>,
----------------
Will enabling Ofast require more changes in the flang frontend? If yes, it is
OK to support it in another patch later.
================
Comment at: flang/include/flang/Frontend/CodeGenOptions.def:30
+
+VALUE_CODEGENOPT(OptimizationLevel, 2, 0) ///< The -O[0-3] option specified.
+
----------------
I saw `clang/include/clang/Basic/CodeGenOptions.def` has the following:
```
VALUE_CODEGENOPT(OptimizationLevel, 2, 0) ///< The -O[0-3] option specified.
VALUE_CODEGENOPT(OptimizeSize, 2, 0) ///< If -Os (==1) or -Oz (==2) is
specified.
```
Do `-Os` and `-Oz` need extra processing in flang drivers? If yes, it is OK to
support it in another patch later.
================
Comment at: flang/lib/Frontend/CodeGenOptions.cpp:8
+//===----------------------------------------------------------------------===//
+
+#include "flang/Frontend/CodeGenOptions.h"
----------------
Miss the following?
```
//
// Coding style: https://mlir.llvm.org/getting_started/DeveloperGuide/
//
//===----------------------------------------------------------------------===//
```
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D128043/new/
https://reviews.llvm.org/D128043
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
