[llvm-objcopy] --weaken-symbol/--weaken: weaken STB_GNU_UNIQUE symbols

[MaskRay]

Summary

STB_GNU_UNIQUE is like STB_GLOBAL with extra semantics:

• gold and ld.lld: changed to STB_GLOBAL if --no-gnu-unique is specified
• glibc: unique even with dlopen RTLD_LOCAL, implies DF_1_NODELETE

Therefore, I think it makes sense for --weaken-symbol/--weaken-symbols/--weaken
to change STB_GNU_UNIQUE symbols.

binutils 2.39 will have the same behavior: https://sourceware.org/bugzilla/show_bug.cgi?id=28926

[jhenderson]

Looks good iff the GNU side are happy with the behaviour change.

[MaskRay]

In D120638#3348436, @jhenderson wrote:

Looks good iff the GNU side are happy with the behaviour change.

They haven't responded yet: https://sourceware.org/pipermail/binutils/2022-March/119961.html
I am inclined to land this instead of keep waiting.

[jhenderson]

In D120638#3375799, @MaskRay wrote:

In D120638#3348436, @jhenderson wrote:

Looks good iff the GNU side are happy with the behaviour change.

They haven't responded yet: https://sourceware.org/pipermail/binutils/2022-March/119961.html
I am inclined to land this instead of keep waiting.

I'm not. llvm-objcopy promises to match the behaviour of GNU objcopy except for where he GNU behaviour doesn't make sense. This doesn't fall into that exception, in my opinion (yes, what you're doing is not unreasonable, but the old behaviour is not nonsense).

[MaskRay]

binutils 2.39 will have the same behavior: https://sourceware.org/bugzilla/show_bug.cgi?id=28926

llvm/lib/ObjCopy/ELF/ELFObjcopy.cpp

if (!Config.SymbolsToKeepGlobal.empty() &&
         !Config.SymbolsToKeepGlobal.matches(Sym.Name) &&
         Sym.getShndx() != SHN_UNDEF)
       Sym.Binding = STB_LOCAL;
 
     if (Config.SymbolsToGlobalize.matches(Sym.Name) &&
         Sym.getShndx() != SHN_UNDEF)
       Sym.Binding = STB_GLOBAL;
 
-    if (Config.SymbolsToWeaken.matches(Sym.Name) && Sym.Binding == STB_GLOBAL)
+    // SymbolsToWeaken applies to both STB_GLOBAL and STB_GNU_UNIQUE.
+    if (Config.SymbolsToWeaken.matches(Sym.Name) && Sym.Binding != STB_LOCAL)
       Sym.Binding = STB_WEAK;
 
-    if (Config.Weaken && Sym.Binding == STB_GLOBAL &&
+    if (Config.Weaken && Sym.Binding != STB_LOCAL &&
         Sym.getShndx() != SHN_UNDEF)
       Sym.Binding = STB_WEAK;
 
     const auto I = Config.SymbolsToRename.find(Sym.Name);
     if (I != Config.SymbolsToRename.end())
       Sym.Name = std::string(I->getValue());
 
     if (!Config.SymbolsPrefix.empty() && Sym.Type != STT_SECTION)

llvm/test/tools/llvm-objcopy/ELF/weaken-all.test

- Name:     undefGlobal
     Size:     8
     Binding:  STB_GLOBAL
   - Name:     Weak
     Type:     STT_FUNC
     Size:     8
     Section:  .text
     Value:    0x1010
     Binding:  STB_WEAK
+  - Name:     GnuUnique
+    Section:  .text
+    Value:    0x1010
+    Binding:  STB_GNU_UNIQUE
 
 #CHECK: Symbols [
 #CHECK-NEXT:  Symbol {
 #CHECK-NEXT:    Name:
 #CHECK-NEXT:    Value: 0x0
 #CHECK-NEXT:    Size: 0
 #CHECK-NEXT:    Binding: Local
 #CHECK-NEXT:    Type: None

 #CHECK-NEXT:    Name: Weak
 #CHECK-NEXT:    Value: 0x1010
 #CHECK-NEXT:    Size: 8
 #CHECK-NEXT:    Binding: Weak
 #CHECK-NEXT:    Type: Function
 #CHECK-NEXT:    Other: 0
 #CHECK-NEXT:    Section: .text
 #CHECK-NEXT:  }
+#CHECK-NEXT:  Symbol {
+#CHECK-NEXT:    Name: GnuUnique
+#CHECK-NEXT:    Value: 0x1010
+#CHECK-NEXT:    Size: 0
+#CHECK-NEXT:    Binding: Weak
+#CHECK-NEXT:    Type: None
+#CHECK-NEXT:    Other: 0
+#CHECK-NEXT:    Section: .text
+#CHECK-NEXT:  }
 #CHECK-NEXT:]

llvm/test/tools/llvm-objcopy/ELF/weaken.test

 # RUN: yaml2obj %s -o %t
-# RUN: llvm-objcopy --weaken-symbol Global -W Local -W Weak -W Missing %t %t2
+# RUN: llvm-objcopy --weaken-symbol Global -W Local -W Weak -W GnuUnique -W Missing %t %t2
 # RUN: llvm-readobj --symbols %t2 | FileCheck %s
 # RUN: llvm-objcopy --regex --weaken-symbol='[LGW].*' %t %t3
 # RUN: cmp %t2 %t3
 # RUN: echo "  Global  # comment" > %t-list.txt
 # RUN: echo "Local" >> %t-list.txt
 # RUN: echo "Weak" >> %t-list.txt
+# RUN: echo "GnuUnique" >> %t-list.txt
 # RUN: echo "Missing" >> %t-list.txt
 # RUN: echo " # comment" >> %t-list.txt
 # RUN: llvm-objcopy --weaken-symbols %t-list.txt %t %t4
 # RUN: cmp %t2 %t4
 # RUN: echo ".*al # Global + Local == .*al" > %t-list2.txt
 # RUN: echo "Weak # weak" >> %t-list2.txt
+# RUN: echo " Gnu.*" >> %t-list2.txt
 # RUN: llvm-objcopy --regex --weaken-symbols %t-list2.txt %t %t5
 # RUN: cmp %t2 %t5
 
 !ELF
 FileHeader:
   Class:           ELFCLASS64
   Data:            ELFDATA2LSB
   Type:            ET_REL
- Name:     Global
     Value:    0x1010
     Binding:  STB_GLOBAL
   - Name:     Weak
     Type:     STT_FUNC
     Size:     8
     Section:  .text
     Value:    0x1008
     Binding:  STB_WEAK
+  - Name:     GnuUnique
+    Section:  .text
+    Binding:  STB_GNU_UNIQUE
   - Name:     Unspecified
     Section:  .text
     Binding:  STB_GLOBAL
 
 #CHECK: Symbols [
 #CHECK-NEXT:  Symbol {
 #CHECK-NEXT:    Name:
 #CHECK-NEXT:    Value: 0x0

 #CHECK-NEXT:    Value: 0x1008
 #CHECK-NEXT:    Size: 8
 #CHECK-NEXT:    Binding: Weak
 #CHECK-NEXT:    Type: Function
 #CHECK-NEXT:    Other: 0
 #CHECK-NEXT:    Section: .text
 #CHECK-NEXT:  }
 #CHECK-NEXT:  Symbol {
+#CHECK-NEXT:    Name: GnuUnique
+#CHECK-NEXT:    Value: 0x0
+#CHECK-NEXT:    Size: 0
+#CHECK-NEXT:    Binding: Weak
+#CHECK-NEXT:    Type: None
+#CHECK-NEXT:    Other: 0
+#CHECK-NEXT:    Section: .text
+#CHECK-NEXT:  }
+#CHECK-NEXT:  Symbol {
 #CHECK-NEXT:    Name: Unspecified
 #CHECK-NEXT:    Value: 0x0
 #CHECK-NEXT:    Size: 0
 #CHECK-NEXT:    Binding: Global
 #CHECK-NEXT:    Type: None
 #CHECK-NEXT:    Other: 0
 #CHECK-NEXT:    Section: .text
 #CHECK-NEXT:  }
 #CHECK-NEXT:]