Subzero: Randomly insert nops.

src/IceTargetLoweringX8632.cpp

Index: src/IceTargetLoweringX8632.cpp
diff --git a/src/IceTargetLoweringX8632.cpp b/src/IceTargetLoweringX8632.cpp
index ebc0bae319fc44d9d509e67999f088e905586076..94682a2fcb709c06adb4c595b587bf13325c20ff 100644
--- a/src/IceTargetLoweringX8632.cpp
+++ b/src/IceTargetLoweringX8632.cpp
@@ -128,6 +128,9 @@ const uint32_t X86_CHAR_BIT = 8;
 const uint32_t X86_STACK_ALIGNMENT_BYTES = 16;
 // Size of the return address on the stack
 const uint32_t X86_RET_IP_SIZE_BYTES = 4;
+// The maximum supported length of a NOP instruction (all smaller
+// lengths are also supported)
+const uint32_t X86_MAX_NOP_LEN = 9;

[jvoung (off chromium), 2014/08/14 15:17:17]

Does it actually help to add nops of up to 9 bytes?

The UCI CGO paper only has a table of 7 nop candidates (of which 5 are not awful
for performance). The choices are just the 1-byte nop, and several 2-byte nop
variants. By default only one of the 5 may be randomly inserted. It sounds like
the main point is that the nop can turn non-bundle-aligned streams of
instructions into things that are less useful when executed -- e.g., some of the
2 byte nops, if jumped into the middle-of, are interpreted as privileged
instructions that cannot be run:
(https://www.ics.uci.edu/~ahomescu/multicompiler_cgo13.pdf)

[wala, 2014/08/14 23:29:50]

Thanks for the link.

That paper makes it clear that length is not the most important attribute of the
nop.

To this end, I've changed the NOP instruction to take a "nop variant" parameter
instead of a "length" parameter.

Variant is currently an integer, but eventually it should be expanded into an
enum once it is decided which nop sequences to use.

 
 // Value is a size in bytes.  Return Value adjusted to the next highest
 // multiple of the stack alignment.
@@ -392,6 +395,11 @@ void TargetX8632::translateO2() {
     return;
   T_genFrame.printElapsedUs(Context, "genFrame()");
   Func->dump("After stack frame mapping");
+
+  // Nop insertion
+  if (shouldDoNopInsertion()) {
+    Func->doNopInsertion();
+  }
 }
 
 void TargetX8632::translateOm1() {
@@ -430,6 +438,11 @@ void TargetX8632::translateOm1() {
     return;
   T_genFrame.printElapsedUs(Context, "genFrame()");
   Func->dump("After stack frame mapping");
+
+  // Nop insertion
+  if (shouldDoNopInsertion()) {
+    Func->doNopInsertion();
+  }
 }
 
 IceString TargetX8632::RegNames[] = {
@@ -3597,6 +3610,13 @@ void TargetX8632::doAddressOptLoad() {
   }
 }
 
+void TargetX8632::randomlyInsertNop(float Probability) {
+  RandomNumberGeneratorWrapper RNG = Ctx->getRNG();
+  if (RNG.getTrueWithProbability(Probability)) {
+    _nop(1 + RNG.next(X86_MAX_NOP_LEN));
+  }
+}
+
 void TargetX8632::lowerPhi(const InstPhi * /*Inst*/) {
   Func->setError("Phi found in regular instruction list");
 }