[LLVMbugs] [Bug 1023] NEW: folding load of non-constant initializer

bugzilla-daemon at cs.uiuc.edu bugzilla-daemon at cs.uiuc.edu
Tue Nov 28 13:57:23 PST 2006


           Summary: folding load of non-constant initializer
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: djg at cray.com

The logic that folds loads of static initializers doesn't take into account
whether the variable being initialized may be modified at runtime.

Here's a patch that fixes it:

Index: SelectionDAGISel.cpp
RCS file: /var/cvs/llvm/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp,v
retrieving revision 1.319
diff -u -r1.319 SelectionDAGISel.cpp
--- SelectionDAGISel.cpp
+++ SelectionDAGISel.cpp
@@ -3298,7 +3298,7 @@
         if (G) {
           GlobalVariable *GV = dyn_cast<GlobalVariable>(G->getGlobal());
-          if (GV) {
+          if (GV && GV->isConstant()) {
             Str = GV->getStringValue(false);
             if (!Str.empty()) {
               CopyFromStr = true;

And here's a testcase that reproduces it:

target datalayout = "e-p:32:32"
target endian = little
target pointersize = 32
target triple = "i686-pc-linux-gnu"
%fmt = global [4 x sbyte] c"%x\0A\00"
%bytes = global [4 x sbyte] c"\AA\BB\CC\DD"


void %foo() {
        %y = alloca uint
        %c = cast uint* %y to sbyte*
        %z = getelementptr [4 x sbyte]* %bytes, int 0, int 0
        call void %llvm.memcpy.i32( sbyte* %c, sbyte* %z, uint 4, uint 1 )
        %r = load uint* %y
        %t = cast [4 x sbyte]* %fmt to sbyte*
        %tmp = call int (sbyte*, ...)* %printf( sbyte* %t, uint %r )
        ret void

declare void %llvm.memcpy.i32(sbyte*, sbyte*, uint, uint)

declare int %printf(sbyte*, ...)

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

More information about the llvm-bugs mailing list