lld: Target-specific section-folding logic for ICF

[pranavk]

This is needed for AArch64 when we try to fold two sections with unpaired ADRP/LDR relocations in them.

[github-actions]

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command:

git-clang-format --diff HEAD~1 HEAD --extensions h,cpp -- lld/ELF/Arch/AArch64.cpp lld/ELF/ICF.cpp lld/ELF/Target.h

View the diff from clang-format here.

diff --git a/lld/ELF/Arch/AArch64.cpp b/lld/ELF/Arch/AArch64.cpp
index b42711a81..23a670c9d 100644
--- a/lld/ELF/Arch/AArch64.cpp
+++ b/lld/ELF/Arch/AArch64.cpp
@@ -982,11 +982,12 @@ void AArch64::relocateAlloc(InputSectionBase &sec, uint8_t *buf) const {
 bool AArch64::canFoldSection(const SmallVector<Relocation> &relocs) const {
   // Section cannot be folded as part of ICF if it contains unpaired relocations
   // eg: ADR_GOT_PAGE and LD64_GOT_LO12_NC don't point to same symbol.
-  SmallSet<Symbol*, 4> syms;
+  SmallSet<Symbol *, 4> syms;
   for (const Relocation &reloc : relocs) {
     if (reloc.type == R_AARCH64_ADR_GOT_PAGE)
       syms.insert(reloc.sym);
-    else if (reloc.type == R_AARCH64_LD64_GOT_LO12_NC && !syms.contains(reloc.sym))
+    else if (reloc.type == R_AARCH64_LD64_GOT_LO12_NC &&
+             !syms.contains(reloc.sym))
       return false;
   }
 
diff --git a/lld/ELF/ICF.cpp b/lld/ELF/ICF.cpp
index 4493ba4a0..846aed50d 100644
--- a/lld/ELF/ICF.cpp
+++ b/lld/ELF/ICF.cpp
@@ -322,7 +322,8 @@ bool ICF<ELFT>::constantEq(const InputSection *secA, Relocs<RelTy> ra,
   }
 
   // Target-specific section-folding logic based on section's relocation list
-  if (!ctx.target->canFoldSection(relocsA) || !ctx.target->canFoldSection(relocsB))
+  if (!ctx.target->canFoldSection(relocsA) ||
+      !ctx.target->canFoldSection(relocsB))
     return false;
 
   return true;
diff --git a/lld/ELF/Target.h b/lld/ELF/Target.h
index 169bd17ec..76c3d4a39 100644
--- a/lld/ELF/Target.h
+++ b/lld/ELF/Target.h
@@ -102,8 +102,11 @@ public:
   virtual void applyJumpInstrMod(uint8_t *loc, JumpModType type,
                                  JumpModType val) const {}
 
-  // Used by ICF to determine if section full of @relocs can be folded safely with another
-  virtual bool canFoldSection(const SmallVector<Relocation> &relocs) const { return true; }
+  // Used by ICF to determine if section full of @relocs can be folded safely
+  // with another
+  virtual bool canFoldSection(const SmallVector<Relocation> &relocs) const {
+    return true;
+  }
 
   virtual ~TargetInfo();
 

[pranavk]

happy to either only pass rel.type and Symbol to TargetInfo::canFoldSection but passing Relocation sounded better to me.

[smithp35]

Thanks for making the patch!

It sounds like #131630 (comment) might end up with a different approach. I'll wait to see if that succeeds before commenting on the details here. There are some more relocation types we'd need to add to canFoldSection for example.