[llvm] r197526 - On APCS, only try to align aggregates to 32 bits instead of 64.

Rafael Espindola rafael.espindola at gmail.com
Tue Dec 17 13:36:55 PST 2013


Author: rafael
Date: Tue Dec 17 15:36:54 2013
New Revision: 197526

URL: http://llvm.org/viewvc/llvm-project?rev=197526&view=rev
Log:
On APCS, only try to align aggregates to 32 bits instead of 64.

This matches clang's behavior and since it is only a preference, it is not
an ABI issue.

Modified:
    llvm/trunk/lib/Target/ARM/ARMTargetMachine.cpp

Modified: llvm/trunk/lib/Target/ARM/ARMTargetMachine.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM/ARMTargetMachine.cpp?rev=197526&r1=197525&r2=197526&view=diff
==============================================================================
--- llvm/trunk/lib/Target/ARM/ARMTargetMachine.cpp (original)
+++ llvm/trunk/lib/Target/ARM/ARMTargetMachine.cpp Tue Dec 17 15:36:54 2013
@@ -92,9 +92,9 @@ static std::string computeDataLayout(ARM
   else
     Ret += "-v128:64:128";
 
-  // An aggregate of size 0 is ABI aligned to 0.
-  // FIXME: explain better what this means.
-  if (ST.isThumb())
+  // On thumb and APCS, only try to align aggregates to 32 bits (the default is
+  // 64 bits).
+  if (ST.isThumb() || ST.isAPCS_ABI())
     Ret += "-a:0:32";
 
   // Integer registers are 32 bits.





More information about the llvm-commits mailing list