[lld][ELF] Add --shuffle-sections=seed to shuffle input sections

Summary:
This option causes lld to shuffle sections by assigning different
priorities in each run.

The use case for this is to introduce randomization in benchmarks. The
idea is inspired by the paper "Producing Wrong Data Without Doing
Anything Obviously Wrong!"
(https://www.inf.usi.ch/faculty/hauswirth/publications/asplos09.pdf). Unlike
the paper, we shuffle individual sections, not just input files.

Doing this in lld is particularly convenient as the --reproduce option
makes it easy to collect all the necessary bits for relinking the
program being benchmarked. Once that it is done, all that is needed is
to add --shuffle-sections=0 to the response file and relink before each
run of the benchmark.

Differential Revision: https://reviews.llvm.org/D74791

Author: espindola

lld/ELF/Config.h

@@ -182,6 +182,7 @@ struct Configuration {
   bool relocatable;
   bool relrPackDynRelocs;
   bool saveTemps;
+  llvm::Optional<uint32_t> shuffleSectionSeed;
   bool singleRoRx;
   bool shared;
   bool isStatic = false;

lld/ELF/Driver.cpp

@@ -956,6 +956,8 @@ static void readConfigs(opt::InputArgList &args) {
   config->rpath = getRpath(args);
   config->relocatable = args.hasArg(OPT_relocatable);
   config->saveTemps = args.hasArg(OPT_save_temps);
+  if (args.hasArg(OPT_shuffle_sections))
+    config->shuffleSectionSeed = args::getInteger(args, OPT_shuffle_sections, 0);
   config->searchPaths = args::getStrings(args, OPT_library_path);
   config->sectionStartMap = getSectionStartMap(args);
   config->shared = args.hasArg(OPT_shared);

lld/ELF/Options.td

@@ -499,6 +499,8 @@ def opt_remarks_format: Separate<["--"], "opt-remarks-format">,
   HelpText<"The format used for serializing remarks (default: YAML)">;
 defm plugin_opt: Eq<"plugin-opt", "specifies LTO options for compatibility with GNU linkers">;
 def save_temps: F<"save-temps">;
+def shuffle_sections: J<"shuffle-sections=">, MetaVarName<"<seed>">,
+  HelpText<"Shuffle input sections using the given seed. If 0, use a random seed">;
 def thinlto_cache_dir: J<"thinlto-cache-dir=">,
   HelpText<"Path to ThinLTO cached object file directory">;
 defm thinlto_cache_policy: Eq<"thinlto-cache-policy", "Pruning policy for the ThinLTO cache">;

lld/ELF/Writer.cpp

@@ -1205,6 +1205,27 @@ findOrphanPos(std::vector<BaseCommand *>::iterator b,
   return i;
 }
 
+// Adds random priorities to sections not already in the map.
+static void maybeShuffle(DenseMap<const InputSectionBase *, int> &order) {
+  if (!config->shuffleSectionSeed)
+    return;
+
+  std::vector<int> priorities(inputSections.size() - order.size());
+  // Existing priorities are < 0, so use priorities >= 0 for the missing
+  // sections.
+  int curPrio = 0;
+  for (int &prio : priorities)
+    prio = curPrio++;
+  uint32_t seed = *config->shuffleSectionSeed;
+  std::mt19937 g(seed ? seed : std::random_device()());
+  std::shuffle(priorities.begin(), priorities.end(), g);
+  int prioIndex = 0;
+  for (InputSectionBase *sec : inputSections) {
+    if (order.try_emplace(sec, priorities[prioIndex]).second)
+      ++prioIndex;
+  }
+}
+
 // Builds section order for handling --symbol-ordering-file.
 static DenseMap<const InputSectionBase *, int> buildSectionOrder() {
   DenseMap<const InputSectionBase *, int> sectionOrder;
@@ -1384,6 +1405,7 @@ static void sortSection(OutputSection *sec,
 template <class ELFT> void Writer<ELFT>::sortInputSections() {
   // Build the order once since it is expensive.
   DenseMap<const InputSectionBase *, int> order = buildSectionOrder();
+  maybeShuffle(order);
   for (BaseCommand *base : script->sectionCommands)
     if (auto *sec = dyn_cast<OutputSection>(base))
       sortSection(sec, order);

lld/docs/ld.lld.1

@@ -456,6 +456,8 @@ were concatenated in the order they appeared on the command line.
 Set address of section.
 .It Fl -shared , Fl -Bsharable
 Build a shared object.
+.It Fl -shuffle-sections Ns = Ns Ar seed
+Shuffle input sections using the given seed. If 0, use a random seed.
 .It Fl -soname Ns = Ns Ar value , Fl h Ar value
 Set
 .Dv DT_SONAME

lld/test/ELF/shuffle-sections.s

@@ -0,0 +1,47 @@
+# REQUIRES: x86
+# RUN: llvm-mc -filetype=obj -triple=x86_64 %s -o %t.o
+
+# RUN: ld.lld %t.o -o %t.out
+# RUN: llvm-readelf -x .text %t.out | FileCheck %s
+# CHECK: Hex dump of section '.text':
+# CHECK-NEXT: 01020304
+
+## --shuffle-sections= shuffles input sections.
+# RUN: ld.lld --shuffle-sections=1 %t.o -o %t1.out
+# RUN: llvm-readelf -x .text %t1.out | FileCheck %s --check-prefix=SHUFFLE1
+# SHUFFLE1: Hex dump of section '.text':
+# SHUFFLE1-NEXT: 01020403
+
+## Test that --shuffle-sections= can be used with --symbol-ordering-file
+# RUN: echo "foo" > %t_order.txt
+# RUN: echo "_start " >> %t_order.txt
+
+# RUN: ld.lld --symbol-ordering-file %t_order.txt --shuffle-sections=2 %t.o -o %t2.out
+# RUN: llvm-readelf -x .text %t2.out | FileCheck %s --check-prefix=SHUFFLE2
+# SHUFFLE2: Hex dump of section '.text':
+# SHUFFLE2-NEXT: 02cccccc 010403
+
+# RUN: ld.lld --symbol-ordering-file %t_order.txt --shuffle-sections=3 %t.o -o %t3.out
+# RUN: llvm-readelf -x .text %t3.out | FileCheck %s --check-prefix=SHUFFLE3
+# SHUFFLE3: Hex dump of section '.text':
+# SHUFFLE3-NEXT: 02cccccc 010304
+
+## .text has an alignment of 4.
+.global _start
+_start:
+  .byte 1
+
+.section .text.foo,"ax"
+.global foo
+foo:
+  .byte 2
+
+.section .text.bar,"ax"
+.global bar
+bar:
+  .byte 3
+
+.section .text.zed,"ax"
+.global zed
+zed:
+  .byte 4