From 6565a0d4b2c98722eb8fee9093cdde4f37928986 Mon Sep 17 00:00:00 2001
From: "Duncan P. N. Exon Smith" <dexonsmith@apple.com>
Date: Sun, 27 Mar 2016 23:17:54 +0000
Subject: Reapply ~"Bitcode: Collect all MDString records into a single blob"

Spiritually reapply commit r264409 (reverted in r264410), albeit with a
bit of a redesign.

Firstly, avoid splitting the big blob into multiple chunks of strings.

r264409 imposed an arbitrary limit to avoid a massive allocation on the
shared 'Record' SmallVector.  The bug with that commit only reproduced
when there were more than "chunk-size" strings.  A test for this would
have been useless long-term, since we're liable to adjust the chunk-size
in the future.

Thus, eliminate the motivation for chunk-ing by storing the string sizes
in the blob.  Here's the layout:

    vbr6: # of strings
    vbr6: offset-to-blob
    blob:
       [vbr6]: string lengths
       [char]: concatenated strings

Secondly, make the output of llvm-bcanalyzer readable.

I noticed when debugging r264409 that llvm-bcanalyzer was outputting a
massive blob all in one line.  Past a small number, the strings were
impossible to split in my head, and the lines were way too long.  This
version adds support in llvm-bcanalyzer for pretty-printing.

    <STRINGS abbrevid=4 op0=3 op1=9/> num-strings = 3 {
      'abc'
      'def'
      'ghi'
    }

From the original commit:

Inspired by Mehdi's similar patch, http://reviews.llvm.org/D18342, this
should (a) slightly reduce bitcode size, since there is less record
overhead, and (b) greatly improve reading speed, since blobs are super
cheap to deserialize.

llvm-svn: 264551
---
 llvm/lib/Bitcode/Writer/ValueEnumerator.cpp | 23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)

(limited to 'llvm/lib/Bitcode/Writer/ValueEnumerator.cpp')

diff --git a/llvm/lib/Bitcode/Writer/ValueEnumerator.cpp b/llvm/lib/Bitcode/Writer/ValueEnumerator.cpp
index 08b5e45703a..69cafb7eeee 100644
--- a/llvm/lib/Bitcode/Writer/ValueEnumerator.cpp
+++ b/llvm/lib/Bitcode/Writer/ValueEnumerator.cpp
@@ -280,8 +280,7 @@ static bool isIntOrIntVectorValue(const std::pair<const Value*, unsigned> &V) {
 
 ValueEnumerator::ValueEnumerator(const Module &M,
                                  bool ShouldPreserveUseListOrder)
-    : HasMDString(false),
-      ShouldPreserveUseListOrder(ShouldPreserveUseListOrder) {
+    : ShouldPreserveUseListOrder(ShouldPreserveUseListOrder) {
   if (ShouldPreserveUseListOrder)
     UseListOrders = predictUseListOrder(M);
 
@@ -375,6 +374,9 @@ ValueEnumerator::ValueEnumerator(const Module &M,
 
   // Optimize constant ordering.
   OptimizeConstants(FirstConstant, Values.size());
+
+  // Organize metadata ordering.
+  organizeMetadata();
 }
 
 unsigned ValueEnumerator::getInstructionID(const Instruction *Inst) const {
@@ -530,8 +532,8 @@ void ValueEnumerator::EnumerateMetadata(const Metadata *MD) {
     EnumerateMDNodeOperands(N);
   else if (auto *C = dyn_cast<ConstantAsMetadata>(MD))
     EnumerateValue(C->getValue());
-
-  HasMDString |= isa<MDString>(MD);
+  else
+    ++NumMDStrings;
 
   // Replace the dummy ID inserted above with the correct one.  MetadataMap may
   // have changed by inserting operands, so we need a fresh lookup here.
@@ -557,6 +559,19 @@ void ValueEnumerator::EnumerateFunctionLocalMetadata(
   FunctionLocalMDs.push_back(Local);
 }
 
+void ValueEnumerator::organizeMetadata() {
+  if (!NumMDStrings)
+    return;
+
+  // Put the strings first.
+  std::stable_partition(MDs.begin(), MDs.end(),
+                        [](const Metadata *MD) { return isa<MDString>(MD); });
+
+  // Renumber.
+  for (unsigned I = 0, E = MDs.size(); I != E; ++I)
+    MetadataMap[MDs[I]] = I + 1;
+}
+
 void ValueEnumerator::EnumerateValue(const Value *V) {
   assert(!V->getType()->isVoidTy() && "Can't insert void values!");
   assert(!isa<MetadataAsValue>(V) && "EnumerateValue doesn't handle Metadata!");
-- 
cgit v1.2.3