[flang-commits] [flang] 219c14a - [Flang] Remove dead -mvscale-{min, max} logic from getVScaleRange. NFCI (#78133)

[via flang-commits]

Author: Luke Lau
Date: 2024-01-16T00:20:35+07:00
New Revision: 219c14a2609aaced2823f5a1b0a22584559bddad

URL: https://github.com/llvm/llvm-project/commit/219c14a2609aaced2823f5a1b0a22584559bddad
DIFF: https://github.com/llvm/llvm-project/commit/219c14a2609aaced2823f5a1b0a22584559bddad.diff

LOG: [Flang] Remove dead -mvscale-{min,max} logic from getVScaleRange. NFCI (#78133)

After #77905, setting -mvscale-min or -mvscale-max on targets other than
AArch64 and RISC-V should be an error now, so we no longer need this
target-agnostic code in getVScaleRange.

Added: 
    

Modified: 
    flang/lib/Frontend/FrontendActions.cpp

Removed: 
    


################################################################################
diff  --git a/flang/lib/Frontend/FrontendActions.cpp b/flang/lib/Frontend/FrontendActions.cpp
index 8fb1998f01703a..74e3992d5ab62b 100644
--- a/flang/lib/Frontend/FrontendActions.cpp
+++ b/flang/lib/Frontend/FrontendActions.cpp
@@ -769,7 +769,6 @@ getRISCVVScaleRange(CompilerInstance &ci) {
 // too much of clang, so for now, replicate the functionality.
 static std::optional<std::pair<unsigned, unsigned>>
 getVScaleRange(CompilerInstance &ci) {
-  const auto &langOpts = ci.getInvocation().getLangOpts();
   const llvm::Triple triple(ci.getInvocation().getTargetOpts().triple);
 
   if (triple.isAArch64())
@@ -777,10 +776,8 @@ getVScaleRange(CompilerInstance &ci) {
   if (triple.isRISCV())
     return getRISCVVScaleRange(ci);
 
-  if (langOpts.VScaleMin || langOpts.VScaleMax)
-    return std::pair<unsigned, unsigned>(
-        langOpts.VScaleMin ? langOpts.VScaleMin : 1, langOpts.VScaleMax);
-
+  // All other architectures that don't support scalable vectors (i.e. don't
+  // need vscale)
   return std::nullopt;
 }