[lld] ae1ba61 - [ELF] Replace uncompressed InputSectionBase::data() with rawData. NFC

Fangrui Song via llvm-commits llvm-commits at lists.llvm.org
Mon Feb 21 00:39:32 PST 2022


Author: Fangrui Song
Date: 2022-02-21T00:39:26-08:00
New Revision: ae1ba6194f09b7e310fd49cf18a2829dcbeb7f6b

URL: https://github.com/llvm/llvm-project/commit/ae1ba6194f09b7e310fd49cf18a2829dcbeb7f6b
DIFF: https://github.com/llvm/llvm-project/commit/ae1ba6194f09b7e310fd49cf18a2829dcbeb7f6b.diff

LOG: [ELF] Replace uncompressed InputSectionBase::data() with rawData. NFC

In many call sites we know uncompression cannot happen (non-SHF_ALLOC, or the
data (even if compressed) must have been uncompressed by a previous pass).
Prefer rawData in these cases. data() increases code size and prevents
optimization on rawData.

Added: 
    

Modified: 
    lld/ELF/AArch64ErrataFix.cpp
    lld/ELF/ARMErrataFix.cpp
    lld/ELF/Arch/X86_64.cpp
    lld/ELF/Driver.cpp
    lld/ELF/EhFrame.cpp
    lld/ELF/ICF.cpp
    lld/ELF/InputFiles.cpp
    lld/ELF/InputSection.cpp
    lld/ELF/InputSection.h
    lld/ELF/MarkLive.cpp
    lld/ELF/Relocations.cpp
    lld/ELF/SyntheticSections.cpp
    lld/ELF/Writer.cpp

Removed: 
    


################################################################################
diff  --git a/lld/ELF/AArch64ErrataFix.cpp b/lld/ELF/AArch64ErrataFix.cpp
index 9c9bd41d6e6c4..c0774be3fd2c9 100644
--- a/lld/ELF/AArch64ErrataFix.cpp
+++ b/lld/ELF/AArch64ErrataFix.cpp
@@ -349,7 +349,7 @@ static uint64_t scanCortexA53Errata843419(InputSection *isec, uint64_t &off,
   }
 
   uint64_t patchOff = 0;
-  const uint8_t *buf = isec->data().begin();
+  const uint8_t *buf = isec->rawData.begin();
   const ulittle32_t *instBuf = reinterpret_cast<const ulittle32_t *>(buf + off);
   uint32_t instr1 = *instBuf++;
   uint32_t instr2 = *instBuf++;
@@ -408,7 +408,7 @@ uint64_t Patch843419Section::getLDSTAddr() const {
 void Patch843419Section::writeTo(uint8_t *buf) {
   // Copy the instruction that we will be replacing with a branch in the
   // patchee Section.
-  write32le(buf, read32le(patchee->data().begin() + patcheeOffset));
+  write32le(buf, read32le(patchee->rawData.begin() + patcheeOffset));
 
   // Apply any relocation transferred from the original patchee section.
   relocateAlloc(buf, buf + getSize());
@@ -591,7 +591,7 @@ AArch64Err843419Patcher::patchInputSectionDescription(
       auto dataSym = std::next(codeSym);
       uint64_t off = (*codeSym)->value;
       uint64_t limit =
-          (dataSym == mapSyms.end()) ? isec->data().size() : (*dataSym)->value;
+          (dataSym == mapSyms.end()) ? isec->rawData.size() : (*dataSym)->value;
 
       while (off < limit) {
         uint64_t startAddr = isec->getVA(off);

diff  --git a/lld/ELF/ARMErrataFix.cpp b/lld/ELF/ARMErrataFix.cpp
index bf277b298587d..bfeeb5db9f33b 100644
--- a/lld/ELF/ARMErrataFix.cpp
+++ b/lld/ELF/ARMErrataFix.cpp
@@ -265,7 +265,7 @@ static ScanResult scanCortexA8Errata657417(InputSection *isec, uint64_t &off,
   }
 
   ScanResult scanRes = {0, 0, nullptr};
-  const uint8_t *buf = isec->data().begin();
+  const uint8_t *buf = isec->rawData.begin();
   // ARMv7-A Thumb 32-bit instructions are encoded 2 consecutive
   // little-endian halfwords.
   const ulittle16_t *instBuf = reinterpret_cast<const ulittle16_t *>(buf + off);
@@ -497,7 +497,7 @@ ARMErr657417Patcher::patchInputSectionDescription(
     while (thumbSym != mapSyms.end()) {
       auto nonThumbSym = std::next(thumbSym);
       uint64_t off = (*thumbSym)->value;
-      uint64_t limit = (nonThumbSym == mapSyms.end()) ? isec->data().size()
+      uint64_t limit = (nonThumbSym == mapSyms.end()) ? isec->rawData.size()
                                                       : (*nonThumbSym)->value;
 
       while (off < limit) {

diff  --git a/lld/ELF/Arch/X86_64.cpp b/lld/ELF/Arch/X86_64.cpp
index ebf0a479b62a1..8c2333cd9bd8c 100644
--- a/lld/ELF/Arch/X86_64.cpp
+++ b/lld/ELF/Arch/X86_64.cpp
@@ -263,7 +263,7 @@ bool X86_64::deleteFallThruJmpInsn(InputSection &is, InputFile *file,
   Relocation &r = is.relocations[rIndex];
 
   // Check if the relocation corresponds to a direct jmp.
-  const uint8_t *secContents = is.data().data();
+  const uint8_t *secContents = is.rawData.data();
   // If it is not a direct jmp instruction, there is nothing to do here.
   if (*(secContents + r.offset - 1) != 0xe9)
     return false;

diff  --git a/lld/ELF/Driver.cpp b/lld/ELF/Driver.cpp
index 4f5f58dadbd06..4910a7d5a1633 100644
--- a/lld/ELF/Driver.cpp
+++ b/lld/ELF/Driver.cpp
@@ -1923,7 +1923,7 @@ static void readSymbolPartitionSection(InputSectionBase *s) {
   if (!isa<Defined>(sym) || !sym->includeInDynsym())
     return;
 
-  StringRef partName = reinterpret_cast<const char *>(s->data().data());
+  StringRef partName = reinterpret_cast<const char *>(s->rawData.data());
   for (Partition &part : partitions) {
     if (part.name == partName) {
       sym->partition = part.getNumber();

diff  --git a/lld/ELF/EhFrame.cpp b/lld/ELF/EhFrame.cpp
index 9ac2ed772073a..794fe04346a99 100644
--- a/lld/ELF/EhFrame.cpp
+++ b/lld/ELF/EhFrame.cpp
@@ -42,7 +42,7 @@ class EhReader {
 private:
   template <class P> void failOn(const P *loc, const Twine &msg) {
     fatal("corrupted .eh_frame: " + msg + "\n>>> defined in " +
-          isec->getObjMsg((const uint8_t *)loc - isec->data().data()));
+          isec->getObjMsg((const uint8_t *)loc - isec->rawData.data()));
   }
 
   uint8_t readByte();

diff  --git a/lld/ELF/ICF.cpp b/lld/ELF/ICF.cpp
index 4c2c0a76f01db..76beff491f52e 100644
--- a/lld/ELF/ICF.cpp
+++ b/lld/ELF/ICF.cpp
@@ -312,7 +312,7 @@ bool ICF<ELFT>::constantEq(const InputSection *secA, ArrayRef<RelTy> ra,
 template <class ELFT>
 bool ICF<ELFT>::equalsConstant(const InputSection *a, const InputSection *b) {
   if (a->flags != b->flags || a->getSize() != b->getSize() ||
-      a->data() != b->data())
+      a->rawData != b->rawData)
     return false;
 
   // If two sections have 
diff erent output sections, we cannot merge them.
@@ -491,7 +491,7 @@ template <class ELFT> void ICF<ELFT>::run() {
   // Initially, we use hash values to partition sections.
   parallelForEach(sections, [&](InputSection *s) {
     // Set MSB to 1 to avoid collisions with unique IDs.
-    s->eqClass[0] = xxHash64(s->data()) | (1U << 31);
+    s->eqClass[0] = xxHash64(s->rawData) | (1U << 31);
   });
 
   // Perform 2 rounds of relocation hash propagation. 2 is an empirical value to

diff  --git a/lld/ELF/InputFiles.cpp b/lld/ELF/InputFiles.cpp
index 4f5c69f4e08ac..e4509c1f78880 100644
--- a/lld/ELF/InputFiles.cpp
+++ b/lld/ELF/InputFiles.cpp
@@ -789,10 +789,10 @@ template <class ELFT> static uint32_t readAndFeatures(const InputSection &sec) {
   using Elf_Note = typename ELFT::Note;
 
   uint32_t featuresSet = 0;
-  ArrayRef<uint8_t> data = sec.data();
+  ArrayRef<uint8_t> data = sec.rawData;
   auto reportFatal = [&](const uint8_t *place, const char *msg) {
     fatal(toString(sec.file) + ":(" + sec.name + "+0x" +
-          Twine::utohexstr(place - sec.data().data()) + "): " + msg);
+          Twine::utohexstr(place - sec.rawData.data()) + "): " + msg);
   };
   while (!data.empty()) {
     // Read one NOTE record.

diff  --git a/lld/ELF/InputSection.cpp b/lld/ELF/InputSection.cpp
index 01a10a797d5eb..7f6d275e327d4 100644
--- a/lld/ELF/InputSection.cpp
+++ b/lld/ELF/InputSection.cpp
@@ -367,6 +367,7 @@ template <class ELFT, class RelTy>
 void InputSection::copyRelocations(uint8_t *buf, ArrayRef<RelTy> rels) {
   const TargetInfo &target = *elf::target;
   InputSectionBase *sec = getRelocatedSection();
+  (void)sec->data(); // uncompress if needed
 
   for (const RelTy &rel : rels) {
     RelType type = rel.getType(config->isMips64EL);
@@ -419,7 +420,7 @@ void InputSection::copyRelocations(uint8_t *buf, ArrayRef<RelTy> rels) {
       }
 
       int64_t addend = getAddend<ELFT>(rel);
-      const uint8_t *bufLoc = sec->data().begin() + rel.r_offset;
+      const uint8_t *bufLoc = sec->rawData.begin() + rel.r_offset;
       if (!RelTy::IsRela)
         addend = target.getImplicitAddend(bufLoc, type);
 
@@ -1423,7 +1424,7 @@ void MergeInputSection::splitIntoPieces() {
 }
 
 SectionPiece *MergeInputSection::getSectionPiece(uint64_t offset) {
-  if (this->data().size() <= offset)
+  if (this->rawData.size() <= offset)
     fatal(toString(this) + ": offset is outside the section");
 
   // If Offset is not at beginning of a section piece, it is not in the map.

diff  --git a/lld/ELF/InputSection.h b/lld/ELF/InputSection.h
index e4fb354592a09..a6413ff2e11ba 100644
--- a/lld/ELF/InputSection.h
+++ b/lld/ELF/InputSection.h
@@ -215,18 +215,18 @@ class InputSectionBase : public SectionBase {
 
 
   template <typename T> llvm::ArrayRef<T> getDataAs() const {
-    size_t s = data().size();
+    size_t s = rawData.size();
     assert(s % sizeof(T) == 0);
-    return llvm::makeArrayRef<T>((const T *)data().data(), s / sizeof(T));
+    return llvm::makeArrayRef<T>((const T *)rawData.data(), s / sizeof(T));
   }
 
+  mutable ArrayRef<uint8_t> rawData;
+
 protected:
   template <typename ELFT>
   void parseCompressedHeader();
   void uncompress() const;
 
-  mutable ArrayRef<uint8_t> rawData;
-
   // This field stores the uncompressed size of the compressed data in rawData,
   // or -1 if rawData is not compressed (either because the section wasn't
   // compressed in the first place, or because we ended up uncompressing it).
@@ -277,8 +277,8 @@ class MergeInputSection : public InputSectionBase {
   llvm::CachedHashStringRef getData(size_t i) const {
     size_t begin = pieces[i].inputOff;
     size_t end =
-        (pieces.size() - 1 == i) ? data().size() : pieces[i + 1].inputOff;
-    return {toStringRef(data().slice(begin, end - begin)), pieces[i].hash};
+        (pieces.size() - 1 == i) ? rawData.size() : pieces[i + 1].inputOff;
+    return {toStringRef(rawData.slice(begin, end - begin)), pieces[i].hash};
   }
 
   // Returns the SectionPiece at a given input section offset.
@@ -300,7 +300,7 @@ struct EhSectionPiece {
       : inputOff(off), sec(sec), size(size), firstRelocation(firstRelocation) {}
 
   ArrayRef<uint8_t> data() const {
-    return {sec->data().data() + this->inputOff, size};
+    return {sec->rawData.data() + this->inputOff, size};
   }
 
   size_t inputOff;

diff  --git a/lld/ELF/MarkLive.cpp b/lld/ELF/MarkLive.cpp
index a55758b20f99c..b197dd45d765b 100644
--- a/lld/ELF/MarkLive.cpp
+++ b/lld/ELF/MarkLive.cpp
@@ -74,7 +74,7 @@ template <class ELFT> class MarkLive {
 template <class ELFT>
 static uint64_t getAddend(InputSectionBase &sec,
                           const typename ELFT::Rel &rel) {
-  return target->getImplicitAddend(sec.data().begin() + rel.r_offset,
+  return target->getImplicitAddend(sec.rawData.begin() + rel.r_offset,
                                    rel.getType(config->isMips64EL));
 }
 

diff  --git a/lld/ELF/Relocations.cpp b/lld/ELF/Relocations.cpp
index 20433b4552edd..8bc52ed2e3771 100644
--- a/lld/ELF/Relocations.cpp
+++ b/lld/ELF/Relocations.cpp
@@ -497,7 +497,7 @@ int64_t RelocationScanner::computeMipsAddend(const RelTy &rel, RelExpr expr,
   if (pairTy == R_MIPS_NONE)
     return 0;
 
-  const uint8_t *buf = sec.data().data();
+  const uint8_t *buf = sec.rawData.data();
   uint32_t symIndex = rel.getSymbol(config->isMips64EL);
 
   // To make things worse, paired relocations might not be contiguous in
@@ -524,7 +524,7 @@ int64_t RelocationScanner::computeAddend(const RelTy &rel, RelExpr expr,
   if (RelTy::IsRela) {
     addend = getAddend<ELFT>(rel);
   } else {
-    const uint8_t *buf = sec.data().data();
+    const uint8_t *buf = sec.rawData.data();
     addend = target.getImplicitAddend(buf + rel.r_offset, type);
   }
 
@@ -1326,7 +1326,7 @@ template <class ELFT, class RelTy> void RelocationScanner::scanOne(RelTy *&i) {
       maybeReportUndefined(cast<Undefined>(sym), sec, offset))
     return;
 
-  const uint8_t *relocatedAddr = sec.data().begin() + offset;
+  const uint8_t *relocatedAddr = sec.rawData.begin() + offset;
   RelExpr expr = target.getRelExpr(type, sym, relocatedAddr);
 
   // Ignore R_*_NONE and other marker relocations.

diff  --git a/lld/ELF/SyntheticSections.cpp b/lld/ELF/SyntheticSections.cpp
index 37b6877d699d1..2a93e66e7c029 100644
--- a/lld/ELF/SyntheticSections.cpp
+++ b/lld/ELF/SyntheticSections.cpp
@@ -112,7 +112,7 @@ std::unique_ptr<MipsAbiFlagsSection<ELFT>> MipsAbiFlagsSection<ELFT>::create() {
     create = true;
 
     std::string filename = toString(sec->file);
-    const size_t size = sec->data().size();
+    const size_t size = sec->rawData.size();
     // Older version of BFD (such as the default FreeBSD linker) concatenate
     // .MIPS.abiflags instead of merging. To allow for this case (or potential
     // zero padding) we ignore everything after the first Elf_Mips_ABIFlags
@@ -121,7 +121,7 @@ std::unique_ptr<MipsAbiFlagsSection<ELFT>> MipsAbiFlagsSection<ELFT>::create() {
             Twine(size) + " instead of " + Twine(sizeof(Elf_Mips_ABIFlags)));
       return nullptr;
     }
-    auto *s = reinterpret_cast<const Elf_Mips_ABIFlags *>(sec->data().data());
+    auto *s = reinterpret_cast<const Elf_Mips_ABIFlags *>(sec->rawData.data());
     if (s->version != 0) {
       error(filename + ": unexpected .MIPS.abiflags version " +
             Twine(s->version));
@@ -184,7 +184,7 @@ std::unique_ptr<MipsOptionsSection<ELFT>> MipsOptionsSection<ELFT>::create() {
     sec->markDead();
 
     std::string filename = toString(sec->file);
-    ArrayRef<uint8_t> d = sec->data();
+    ArrayRef<uint8_t> d = sec->rawData;
 
     while (!d.empty()) {
       if (d.size() < sizeof(Elf_Mips_Options)) {
@@ -240,12 +240,12 @@ std::unique_ptr<MipsReginfoSection<ELFT>> MipsReginfoSection<ELFT>::create() {
   for (InputSectionBase *sec : sections) {
     sec->markDead();
 
-    if (sec->data().size() != sizeof(Elf_Mips_RegInfo)) {
+    if (sec->rawData.size() != sizeof(Elf_Mips_RegInfo)) {
       error(toString(sec->file) + ": invalid size of .reginfo section");
       return nullptr;
     }
 
-    auto *r = reinterpret_cast<const Elf_Mips_RegInfo *>(sec->data().data());
+    auto *r = reinterpret_cast<const Elf_Mips_RegInfo *>(sec->rawData.data());
     reginfo.ri_gprmask |= r->ri_gprmask;
     sec->getFile<ELFT>()->mipsGp0 = r->ri_gp_value;
   };
@@ -3535,7 +3535,7 @@ void ARMExidxSyntheticSection::writeTo(uint8_t *buf) {
   for (InputSection *isec : executableSections) {
     assert(isec->getParent() != nullptr);
     if (InputSection *d = findExidxSection(isec)) {
-      memcpy(buf + offset, d->data().data(), d->data().size());
+      memcpy(buf + offset, d->rawData.data(), d->rawData.size());
       d->relocateAlloc(buf + d->outSecOff, buf + d->outSecOff + d->getSize());
       offset += d->getSize();
     } else {

diff  --git a/lld/ELF/Writer.cpp b/lld/ELF/Writer.cpp
index 0282d7d6b5a78..cd43e79b82760 100644
--- a/lld/ELF/Writer.cpp
+++ b/lld/ELF/Writer.cpp
@@ -1695,7 +1695,7 @@ static void fixSymbolsAfterShrinking() {
       if (!inputSec || !inputSec->bytesDropped)
         return;
 
-      const size_t OldSize = inputSec->data().size();
+      const size_t OldSize = inputSec->rawData.size();
       const size_t NewSize = OldSize - inputSec->bytesDropped;
 
       if (def->value > NewSize && def->value <= OldSize) {


        


More information about the llvm-commits mailing list