MFC 261991:

Upgrade our copy of llvm/clang to 3.4 release.  This version supports
all of the features in the current working draft of the upcoming C++
standard, provisionally named C++1y.

The code generator's performance is greatly increased, and the loop
auto-vectorizer is now enabled at -Os and -O2 in addition to -O3.  The
PowerPC backend has made several major improvements to code generation
quality and compile time, and the X86, SPARC, ARM32, Aarch64 and SystemZ
backends have all seen major feature work.

Release notes for llvm and clang can be found here:
<http://llvm.org/releases/3.4/docs/ReleaseNotes.html>
<http://llvm.org/releases/3.4/tools/clang/docs/ReleaseNotes.html>

MFC 262121 (by emaste):

Update lldb for clang/llvm 3.4 import

This commit largely restores the lldb source to the upstream r196259
snapshot with the addition of threaded inferior support and a few bug
fixes.

Specific upstream lldb revisions restored include:
   SVN      git
  181387  779e6ac
  181703  7bef4e2
  182099  b31044e
  182650  f2dcf35
  182683  0d91b80
  183862  15c1774
  183929  99447a6
  184177  0b2934b
  184948  4dc3761
  184954  007e7bc
  186990  eebd175

Sponsored by:	DARPA, AFRL

MFC 262186 (by emaste):

Fix mismerge in r262121

A break statement was lost in the merge.  The error had no functional
impact, but restore it to reduce the diff against upstream.

MFC 262303:

Pull in r197521 from upstream clang trunk (by rdivacky):

  Use the integrated assembler by default on FreeBSD/ppc and ppc64.

Requested by:	jhibbits

MFC 262611:

Pull in r196874 from upstream llvm trunk:

  Fix a crash that occurs when PWD is invalid.

  MCJIT needs to be able to run in hostile environments, even when PWD
  is invalid. There's no need to crash MCJIT in this case.

  The obvious fix is to simply leave MCContext's CompilationDir empty
  when PWD can't be determined. This way, MCJIT clients,
  and other clients that link with LLVM don't need a valid working directory.

  If we do want to guarantee valid CompilationDir, that should be done
  only for clients of getCompilationDir(). This is as simple as checking
  for an empty string.

  The only current use of getCompilationDir is EmitGenDwarfInfo, which
  won't conceivably run with an invalid working dir. However, in the
  purely hypothetically and untestable case that this happens, the
  AT_comp_dir will be omitted from the compilation_unit DIE.

This should help fix assertions occurring with ports-mgmt/tinderbox,
when it is using jails, and sometimes invalidates clang's current
working directory.

Reported by:	decke

MFC 262809:

Pull in r203007 from upstream clang trunk:

  Don't produce an alias between destructors with different calling conventions.

  Fixes pr19007.

(Please note that is an LLVM PR identifier, not a FreeBSD one.)

This should fix Firefox and/or libxul crashes (due to problems with
regparm/stdcall calling conventions) on i386.

Reported by:	multiple users on freebsd-current
PR:		bin/187103

MFC 263048:

Repair recognition of "CC" as an alias for the C++ compiler, since it
was silently broken by upstream for a Windows-specific use-case.

Apparently some versions of CMake still rely on this archaic feature...

Reported by:	rakuco

MFC 263049:

Garbage collect the old way of adding the libstdc++ include directories
in clang's InitHeaderSearch.cpp.  This has been superseded by David
Chisnall's commit in r255321.

Moreover, if libc++ is used, the libstdc++ include directories should
not be in the search path at all.  These directories are now only used
if you pass -stdlib=libstdc++.

1 files changed, 25 insertions, 48 deletions

diff --git a/contrib/llvm/lib/CodeGen/TargetSchedule.cpp b/contrib/llvm/lib/CodeGen/TargetSchedule.cpp
index 1bf14db..b0f2ca6 100644
--- a/contrib/llvm/lib/CodeGen/TargetSchedule.cpp
+++ b/contrib/llvm/lib/CodeGen/TargetSchedule.cpp
@@ -93,33 +93,10 @@ unsigned TargetSchedModel::getNumMicroOps(const MachineInstr *MI,
 // effectively means infinite latency. Since users of the TargetSchedule API
 // don't know how to handle this, we convert it to a very large latency that is
 // easy to distinguish when debugging the DAG but won't induce overflow.
-static unsigned convertLatency(int Cycles) {
+static unsigned capLatency(int Cycles) {
   return Cycles >= 0 ? Cycles : 1000;
 }
 
-/// If we can determine the operand latency from the def only, without machine
-/// model or itinerary lookup, do so. Otherwise return -1.
-int TargetSchedModel::getDefLatency(const MachineInstr *DefMI,
-                                    bool FindMin) const {
-
-  // Return a latency based on the itinerary properties and defining instruction
-  // if possible. Some common subtargets don't require per-operand latency,
-  // especially for minimum latencies.
-  if (FindMin) {
-    // If MinLatency is invalid, then use the itinerary for MinLatency. If no
-    // itinerary exists either, then use single cycle latency.
-    if (SchedModel.MinLatency < 0 && !hasInstrItineraries()) {
-      return 1;
-    }
-    return SchedModel.MinLatency;
-  }
-  else if (!hasInstrSchedModel() && !hasInstrItineraries()) {
-    return TII->defaultDefLatency(&SchedModel, DefMI);
-  }
-  // ...operand lookup required
-  return -1;
-}
-
 /// Return the MCSchedClassDesc for this instruction. Some SchedClasses require
 /// evaluation of predicates that depend on instruction operands or flags.
 const MCSchedClassDesc *TargetSchedModel::
@@ -177,18 +154,16 @@ static unsigned findUseIdx(const MachineInstr *MI, unsigned UseOperIdx) {
 // Top-level API for clients that know the operand indices.
 unsigned TargetSchedModel::computeOperandLatency(
   const MachineInstr *DefMI, unsigned DefOperIdx,
-  const MachineInstr *UseMI, unsigned UseOperIdx,
-  bool FindMin) const {
+  const MachineInstr *UseMI, unsigned UseOperIdx) const {
 
-  int DefLatency = getDefLatency(DefMI, FindMin);
-  if (DefLatency >= 0)
-    return DefLatency;
+  if (!hasInstrSchedModel() && !hasInstrItineraries())
+    return TII->defaultDefLatency(&SchedModel, DefMI);
 
   if (hasInstrItineraries()) {
     int OperLatency = 0;
     if (UseMI) {
-      OperLatency =
-        TII->getOperandLatency(&InstrItins, DefMI, DefOperIdx, UseMI, UseOperIdx);
+      OperLatency = TII->getOperandLatency(&InstrItins, DefMI, DefOperIdx,
+                                           UseMI, UseOperIdx);
     }
     else {
       unsigned DefClass = DefMI->getDesc().getSchedClass();
@@ -205,13 +180,11 @@ unsigned TargetSchedModel::computeOperandLatency(
     // hook to allow subtargets to specialize latency. This hook is only
     // applicable to the InstrItins model. InstrSchedModel should model all
     // special cases without TII hooks.
-    if (!FindMin)
-      InstrLatency = std::max(InstrLatency,
-                              TII->defaultDefLatency(&SchedModel, DefMI));
+    InstrLatency = std::max(InstrLatency,
+                            TII->defaultDefLatency(&SchedModel, DefMI));
     return InstrLatency;
   }
-  assert(!FindMin && hasInstrSchedModel() &&
-         "Expected a SchedModel for this cpu");
+  // hasInstrSchedModel()
   const MCSchedClassDesc *SCDesc = resolveSchedClass(DefMI);
   unsigned DefIdx = findDefIdx(DefMI, DefOperIdx);
   if (DefIdx < SCDesc->NumWriteLatencyEntries) {
@@ -219,7 +192,7 @@ unsigned TargetSchedModel::computeOperandLatency(
     const MCWriteLatencyEntry *WLEntry =
       STI->getWriteLatencyEntry(SCDesc, DefIdx);
     unsigned WriteID = WLEntry->WriteResourceID;
-    unsigned Latency = convertLatency(WLEntry->Cycles);
+    unsigned Latency = capLatency(WLEntry->Cycles);
     if (!UseMI)
       return Latency;
 
@@ -228,13 +201,17 @@ unsigned TargetSchedModel::computeOperandLatency(
     if (UseDesc->NumReadAdvanceEntries == 0)
       return Latency;
     unsigned UseIdx = findUseIdx(UseMI, UseOperIdx);
-    return Latency - STI->getReadAdvanceCycles(UseDesc, UseIdx, WriteID);
+    int Advance = STI->getReadAdvanceCycles(UseDesc, UseIdx, WriteID);
+    if (Advance > 0 && (unsigned)Advance > Latency) // unsigned wrap
+      return 0;
+    return Latency - Advance;
   }
   // If DefIdx does not exist in the model (e.g. implicit defs), then return
   // unit latency (defaultDefLatency may be too conservative).
 #ifndef NDEBUG
   if (SCDesc->isValid() && !DefMI->getOperand(DefOperIdx).isImplicit()
-      && !DefMI->getDesc().OpInfo[DefOperIdx].isOptionalDef()) {
+      && !DefMI->getDesc().OpInfo[DefOperIdx].isOptionalDef()
+      && SchedModel.isComplete()) {
     std::string Err;
     raw_string_ostream ss(Err);
     ss << "DefIdx " << DefIdx << " exceeds machine model writes for "
@@ -248,10 +225,13 @@ unsigned TargetSchedModel::computeOperandLatency(
   return DefMI->isTransient() ? 0 : TII->defaultDefLatency(&SchedModel, DefMI);
 }
 
-unsigned TargetSchedModel::computeInstrLatency(const MachineInstr *MI) const {
+unsigned
+TargetSchedModel::computeInstrLatency(const MachineInstr *MI,
+                                      bool UseDefaultDefLatency) const {
   // For the itinerary model, fall back to the old subtarget hook.
   // Allow subtargets to compute Bundle latencies outside the machine model.
-  if (hasInstrItineraries() || MI->isBundle())
+  if (hasInstrItineraries() || MI->isBundle() ||
+      (!hasInstrSchedModel() && !UseDefaultDefLatency))
     return TII->getInstrLatency(&InstrItins, MI);
 
   if (hasInstrSchedModel()) {
@@ -263,7 +243,7 @@ unsigned TargetSchedModel::computeInstrLatency(const MachineInstr *MI) const {
         // Lookup the definition's write latency in SubtargetInfo.
         const MCWriteLatencyEntry *WLEntry =
           STI->getWriteLatencyEntry(SCDesc, DefIdx);
-        Latency = std::max(Latency, convertLatency(WLEntry->Cycles));
+        Latency = std::max(Latency, capLatency(WLEntry->Cycles));
       }
       return Latency;
     }
@@ -274,13 +254,10 @@ unsigned TargetSchedModel::computeInstrLatency(const MachineInstr *MI) const {
 unsigned TargetSchedModel::
 computeOutputLatency(const MachineInstr *DefMI, unsigned DefOperIdx,
                      const MachineInstr *DepMI) const {
-  // MinLatency == -1 is for in-order processors that always have unit
-  // MinLatency. MinLatency > 0 is for in-order processors with varying min
-  // latencies, but since this is not a RAW dep, we always use unit latency.
-  if (SchedModel.MinLatency != 0)
+  if (SchedModel.MicroOpBufferSize <= 1)
     return 1;
 
-  // MinLatency == 0 indicates an out-of-order processor that can dispatch
+  // MicroOpBufferSize > 1 indicates an out-of-order processor that can dispatch
   // WAW dependencies in the same cycle.
 
   // Treat predication as a data dependency for out-of-order cpus. In-order
@@ -302,7 +279,7 @@ computeOutputLatency(const MachineInstr *DefMI, unsigned DefOperIdx,
     if (SCDesc->isValid()) {
       for (const MCWriteProcResEntry *PRI = STI->getWriteProcResBegin(SCDesc),
              *PRE = STI->getWriteProcResEnd(SCDesc); PRI != PRE; ++PRI) {
-        if (!SchedModel.getProcResource(PRI->ProcResourceIdx)->IsBuffered)
+        if (!SchedModel.getProcResource(PRI->ProcResourceIdx)->BufferSize)
           return 1;
       }
     }