[llvm-readobj/readelf] - Don't fail dumping when unable to read the name of the SHT_DYNSYM section.

[grimar]

Summary

We have an issue currently: we are trying to read the name of the SHT_DYNSYM section
very early and using unwrapOrError call for that.

The name is needed only for the GNU output. Because of the current logic, the tool
fails to dump the whole object when something is wrong with the name of the .dynsym section.

This patch delays reading the name and also allows it to be broken.

[jhenderson]

llvm/test/tools/llvm-readobj/ELF/dyn-symbols.test

93

I'm not convinced we want this warning in the LLVM case - it doesn't actually need the name for anything, so emitting the warning doesn't make much sense.

Looking at the code, it looks like the SHT_SYMTAB name lookup has a similar problem? Could the name not be looked up inside the printSymtabMessage function?

141–142

This doesn't check that the warning is emitted.

[grimar]

• Addressed review comments.

llvm/test/tools/llvm-readobj/ELF/dyn-symbols.test

93

Looking at the code, it looks like the SHT_SYMTAB name lookup has a similar problem?

Yes.

Could the name not be looked up inside the printSymtabMessage function?

Yes. It can be done in GNUStyle<ELFT>::printSymtabMessage instead of printSymbolsHelper.
But it should be fixed in a follow-up, I think, because the possible change relies on the DotDynsymSec
introduced in this patch. I will post the fix soon.

141–142

Fixed.

[grimar]

llvm/test/tools/llvm-readobj/ELF/dyn-symbols.test

93

Follow-up: D84231

[jhenderson]

LGTM.

llvm/test/tools/llvm-readobj/ELF/dyn-symbols.test

 ## b) Check --dyn-syms equals --dyn-symbols, --dt for llvm-readobj.
 # RUN: llvm-readobj --dyn-syms %t1.so > %t.readobj-dyn-syms
 # RUN: cmp %t.readobj-dt-alias %t.readobj-dyn-syms
 ## c) Check --dyn-syms equals --dyn-symbols for llvm-readelf.
 # RUN: llvm-readelf --dyn-syms %t1.so > %t.readelf-dyn-syms
 # RUN: llvm-readelf --dyn-symbols %t1.so > %t.readelf-dyn-symbols
 # RUN: cmp %t.readelf-dyn-symbols %t.readelf-dyn-syms
 
-## Case 3: Check that we are able to dump the dynamic symbol table even when we have no program headers.
+## Case 3.1: Check that we are able to dump the dynamic symbol table even when we have no program headers.
 ## In this case we find the table by it's type (SHT_DYNSYM) and ignore the DT_SYMTAB value.
 # RUN: yaml2obj --docnum=2 %s -o %t2.so
 # RUN: llvm-readobj %t2.so --dyn-symbols | FileCheck %s --check-prefix=NOPHDRS-LLVM
-# RUN: llvm-readelf %t2.so --dyn-symbols | FileCheck %s --check-prefix=NOPHDRS-GNU
+# RUN: llvm-readelf %t2.so --dyn-symbols | FileCheck %s -DNAME=.dynsym --check-prefix=NOPHDRS-GNU
 
-# NOPHDRS-LLVM: Name: foo
-# NOPHDRS-GNU:  1: 0000000000000000 0 NOTYPE LOCAL DEFAULT UND foo
+# NOPHDRS-LLVM:      DynamicSymbols [
+# NOPHDRS-NAME-LLVM: warning: '[[FILE]]': unable to get the name of the SHT_DYNSYM section: a section [index 2] has an invalid sh_name (0xffffffff) offset which goes past the end of the section name string table
+# NOPHDRS-LLVM-NEXT:   Symbol {
+# NOPHDRS-LLVM-NEXT:     Name:  (0)
+# NOPHDRS-LLVM-NEXT:     Value: 0x0
+# NOPHDRS-LLVM-NEXT:     Size: 0
+# NOPHDRS-LLVM-NEXT:     Binding: Local (0x0)
+# NOPHDRS-LLVM-NEXT:     Type: None (0x0)
+# NOPHDRS-LLVM-NEXT:     Other: 0
+# NOPHDRS-LLVM-NEXT:     Section: Undefined (0x0)
+# NOPHDRS-LLVM-NEXT:   }
+# NOPHDRS-LLVM-NEXT:   Symbol {
+# NOPHDRS-LLVM-NEXT:     Name: foo (1)
+# NOPHDRS-LLVM-NEXT:     Value: 0x0
+# NOPHDRS-LLVM-NEXT:     Size: 0
+# NOPHDRS-LLVM-NEXT:     Binding: Local (0x0)
+# NOPHDRS-LLVM-NEXT:     Type: None (0x0)
+# NOPHDRS-LLVM-NEXT:     Other: 0
+# NOPHDRS-LLVM-NEXT:     Section: Undefined (0x0)
+# NOPHDRS-LLVM-NEXT:   }
+# NOPHDRS-LLVM-NEXT: ]
+
+# NOPHDRS-GNU:      Symbol table '[[NAME]]' contains 2 entries:
+# NOPHDRS-GNU-NEXT:  Num:    Value          Size Type    Bind   Vis       Ndx Name
+# NOPHDRS-GNU-NEXT:    0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT   UND
+# NOPHDRS-GNU-NEXT:    1: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT   UND foo
 
 --- !ELF
 FileHeader:
   Class:   ELFCLASS64
   Data:    ELFDATA2LSB
   Type:    ET_DYN
   Machine: EM_X86_64
 Sections:
   - Name:    .dynamic
     Type:    SHT_DYNAMIC
     Entries:
       - Tag:   DT_SYMTAB
         Value: 0xffff1234
       - Tag:   DT_NULL
         Value: 0
 DynamicSymbols:
   - Name: foo
 
+## Case 3.2: the same as 3.1, but the sh_name field of the SHT_DYNSYM section is invalid.
+##           Check we are still able to dump symbols.
+# RUN: yaml2obj --docnum=3 %s -o %t2.broken.name
+# RUN: llvm-readobj %t2.broken.name --dyn-symbols 2>&1 | \
+# RUN:   FileCheck %s -DFILE=%t2.broken.name --check-prefixes=NOPHDRS-LLVM,NOPHDRS-NAME-LLVM
+# RUN: llvm-readelf %t2.broken.name --dyn-symbols 2>&1 | \
+# RUN:   FileCheck %s -DFILE=%t2.broken.name -DNAME="<?>" --check-prefix=NOPHDRS-GNU
+
+--- !ELF
+FileHeader:
+  Class:   ELFCLASS64
+  Data:    ELFDATA2LSB
+  Type:    ET_DYN
+  Machine: EM_X86_64
+Sections:
+  - Name:    .dynamic
+    Type:    SHT_DYNAMIC
+    Entries:
+      - Tag:   DT_SYMTAB
+        Value: 0xffff1234
+      - Tag:   DT_NULL
+        Value: 0
+  - Name:   .dynsym
+    Type:   SHT_DYNSYM
+    ShName: 0xffffffff
+DynamicSymbols:
+  - Name: foo
+
 ## Case 4: Check we report a warning when there is no SHT_DYNSYM section and we can't map the DT_SYMTAB value
 ## to an address because of the absence of a corresponding PT_LOAD program header.
-# RUN: yaml2obj --docnum=3 %s -o %t3.so
+# RUN: yaml2obj --docnum=4 %s -o %t3.so
 # RUN: llvm-readobj %t3.so --dyn-symbols 2>&1 | FileCheck %s -DFILE=%t3.so --check-prefixes=NOSHT-DYNSYM,NOSHT-DYNSYM-LLVM
 # RUN: llvm-readelf %t3.so --dyn-symbols 2>&1 | FileCheck %s -DFILE=%t3.so --check-prefix=NOSHT-DYNSYM
 
 # NOSHT-DYNSYM:           warning: '[[FILE]]': Unable to parse DT_SYMTAB: virtual address is not in any segment: 0x0
 # NOSHT-DYNSYM-LLVM:      DynamicSymbols [
 # NOSHT-DYNSYM-LLVM-NEXT: ]
 
 --- !ELF
Entries:
         Value: 0
       - Tag:   DT_NULL
         Value: 0
 DynamicSymbols:
   - Name: foo
 
 ## Case 5: Check that when we can't map the value of the DT_SYMTAB tag to an address, we report a warning and
 ## use the information in the section header table to locate the dynamic symbol table.
-# RUN: yaml2obj --docnum=4 %s -o %t4.so
+# RUN: yaml2obj --docnum=5 %s -o %t4.so
 # RUN: llvm-readobj %t4.so --dyn-symbols 2>&1 | FileCheck -DFILE=%t4.so %s --check-prefixes=BROKEN-DTSYMTAB,BROKEN-DTSYMTAB-LLVM
 # RUN: llvm-readelf %t4.so --dyn-symbols 2>&1 | FileCheck -DFILE=%t4.so %s --check-prefixes=BROKEN-DTSYMTAB,BROKEN-DTSYMTAB-GNU
 
 # BROKEN-DTSYMTAB:      warning: '[[FILE]]': Unable to parse DT_SYMTAB: virtual address is not in any segment: 0xffff1234
 # BROKEN-DTSYMTAB-LLVM: Name: foo
 # BROKEN-DTSYMTAB-GNU:  1: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT   UND foo
 
 --- !ELF

 ProgramHeaders:
   - Type: PT_LOAD
     VAddr: 0x0000
     Sections:
       - Section: .dynsym
 
 ## Case 6: Check that if we can get the location of the dynamic symbol table using both the DT_SYMTAB value
 ## and the section headers table then we prefer the former and report a warning.
-# RUN: yaml2obj --docnum=5 %s -o %t5.so
+# RUN: yaml2obj --docnum=6 %s -o %t5.so
 # RUN: llvm-readobj %t5.so --dyn-symbols 2>&1 | FileCheck -DFILE=%t5.so %s --check-prefixes=PREFER-DTSYMTAB,PREFER-DTSYMTAB-LLVM
 # RUN: llvm-readelf %t5.so --dyn-symbols 2>&1 | FileCheck -DFILE=%t5.so %s --check-prefixes=PREFER-DTSYMTAB,PREFER-DTSYMTAB-GNU
 
 # PREFER-DTSYMTAB:      warning: '[[FILE]]': SHT_DYNSYM section header and DT_SYMTAB disagree about the location of the dynamic symbol table
 # PREFER-DTSYMTAB-LLVM: Name: o
 # PREFER-DTSYMTAB-GNU:  1: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT   UND o
 
 --- !ELF

 ProgramHeaders:
   - Type: PT_LOAD
     VAddr: 0x0000
     Sections:
       - Section: .mydynsym
 
 ## Case 7: Check how we dump versioned symbols. Use both -V and --dyn-symbols
 ## to check that printed version is consistent.
-# RUN: yaml2obj %s --docnum=6 -o %t6
+# RUN: yaml2obj %s --docnum=7 -o %t6
 # RUN: llvm-readobj -V --dyn-symbols %t6 | FileCheck %s --check-prefix=VERSIONED-LLVM
 # RUN: llvm-readelf -V --dyn-symbols %t6 | FileCheck %s --check-prefix=VERSIONED-GNU
 
 # VERSIONED-LLVM:      Symbol {
 # VERSIONED-LLVM:        Name: foo (16)
 # VERSIONED-LLVM-NEXT:   Value: 0x0
 # VERSIONED-LLVM-NEXT:   Size: 0
 # VERSIONED-LLVM-NEXT:   Binding: Local (0x0)
Entries:
           - NonDefault
 DynamicSymbols:
   - Name: foo
   - Name: bar
   - Name: zed
 
 ## Case 8: Check what we print when:
 ## a) The dynamic symbol table does not exist.
-# RUN: yaml2obj %s --docnum=7 -o %t7
+# RUN: yaml2obj %s --docnum=8 -o %t7
 # RUN: llvm-readobj --dyn-symbols %t7 | FileCheck %s --check-prefix=NO-DYNSYM-LLVM
 # RUN: llvm-readelf --dyn-symbols %t7 | count 0
 ## b) The dynamic symbol table has a size of 0.
-# RUN: yaml2obj %s --docnum=8 -o %t8
+# RUN: yaml2obj %s --docnum=9 -o %t8
 # RUN: llvm-readobj --dyn-symbols %t8 | FileCheck %s --check-prefix=NO-DYNSYM-LLVM
 # RUN: llvm-readelf --dyn-symbols %t8 | count 0
 ## c) The dynamic symbol table only contains the null symbol.
-# RUN: yaml2obj %s --docnum=9 -o %t9
+# RUN: yaml2obj %s --docnum=10 -o %t9
 # RUN: llvm-readobj --dyn-symbols %t9 | FileCheck %s --check-prefix=DYNSYM-EMPTY-LLVM
 # RUN: llvm-readelf --dyn-symbols %t9 | FileCheck %s --check-prefix=DYNSYM-EMPTY-GNU
 
 # NO-DYNSYM-LLVM:      DynamicSymbols [
 # NO-DYNSYM-LLVM-NEXT: ]
 
 # DYNSYM-EMPTY-LLVM:      DynamicSymbols [
 # DYNSYM-EMPTY-LLVM-NEXT:  Symbol {
FileHeader:
   Class:   ELFCLASS32
   Data:    ELFDATA2LSB
   Type:    ET_DYN
   Machine: EM_386
 DynamicSymbols: []
 
 ## Case 9: Check what we print when:
 ## a) The size of the dynamic symbol table is not a multiple of its entry size.
-# RUN: yaml2obj %s --docnum=10 -o %t10
+# RUN: yaml2obj %s --docnum=11 -o %t10
 # RUN: llvm-readobj --dyn-symbols %t10 2>&1 | FileCheck %s -DFILE=%t10 --check-prefix=DYNSYM-SIZE-INVALID1
 # RUN: llvm-readelf --dyn-symbols %t10 2>&1 | FileCheck %s -DFILE=%t10 --check-prefix=DYNSYM-SIZE-INVALID1
 
 # DYNSYM-SIZE-INVALID1: warning: '[[FILE]]': section with index 1 has invalid size (0x1) or entry size (0x10)
 
 ## b) The same, but the DT_SYMTAB tag is present. In this case the dynamic tag has priority over the
 ##    information about a location and an entity size of the dynamic symbol table from the section header.
 ##    The code uses sizeof(Elf_Sym) for an entity size, so it can't be incorrect and
 ##    the message printed is a bit shorter.
-# RUN: yaml2obj %s --docnum=11 -o %t11
+# RUN: yaml2obj %s --docnum=12 -o %t11
 # RUN: llvm-readobj --dyn-symbols %t11 2>&1 | FileCheck %s -DFILE=%t11 --check-prefix=DYNSYM-SIZE-INVALID2
 # RUN: llvm-readelf --dyn-symbols %t11 2>&1 | FileCheck %s -DFILE=%t11 --check-prefix=DYNSYM-SIZE-INVALID2
 
 # DYNSYM-SIZE-INVALID2: warning: '[[FILE]]': section with index 2 has invalid size (0x1){{$}}
 
 ## c) In the case when the DT_SYMENT tag is present, we report when it's value does not match the
 #     value of the symbol size for the platform.
-# RUN: yaml2obj %s -D BITS=32 --docnum=12 -o %t12
+# RUN: yaml2obj %s -D BITS=32 --docnum=13 -o %t12
 # RUN: llvm-readobj --dyn-symbols %t12 2>&1 | FileCheck %s -DFILE=%t12 --check-prefix=DYNSYM-SIZE-INVALID3
 # RUN: llvm-readelf --dyn-symbols %t12 2>&1 | FileCheck %s -DFILE=%t12 --check-prefix=DYNSYM-SIZE-INVALID3
-# RUN: yaml2obj %s -D BITS=64 --docnum=12 -o %t13
+# RUN: yaml2obj %s -D BITS=64 --docnum=13 -o %t13
 # RUN: llvm-readobj --dyn-symbols %t13 2>&1 | FileCheck %s -DFILE=%t13 --check-prefix=DYNSYM-SIZE-INVALID4
 # RUN: llvm-readelf --dyn-symbols %t13 2>&1 | FileCheck %s -DFILE=%t13 --check-prefix=DYNSYM-SIZE-INVALID4
 
 # DYNSYM-SIZE-INVALID3: warning: '[[FILE]]': DT_SYMENT value of 0x123 is not the size of a symbol (0x10){{$}}
 # DYNSYM-SIZE-INVALID4: warning: '[[FILE]]': DT_SYMENT value of 0x123 is not the size of a symbol (0x18){{$}}
 
 --- !ELF
 FileHeader:
Entries:
       - Tag:   DT_SYMENT
         Value: 0x123
       - Tag:   DT_NULL
         Value: 0
 
 ## Check we report a warning when the DT_STRSZ value is broken so that the dynamic string
 ## table goes past the end of the file. Document we stop dumping symbols and report an error.
 
-# RUN: yaml2obj %s --docnum=13 -o %t14
+# RUN: yaml2obj %s --docnum=14 -o %t14
 # RUN: llvm-readobj --dyn-symbols %t14 2>&1 | \
 # RUN:   FileCheck %s -DFILE=%t14 --check-prefix=DYNSTR-INVALID-LLVM
 # RUN: llvm-readelf --dyn-symbols %t14 2>&1 | \
 # RUN:   FileCheck %s -DFILE=%t14 --check-prefix=DYNSTR-INVALID-GNU
 
 # DYNSTR-INVALID-LLVM: warning: '[[FILE]]': the dynamic string table at 0x78 goes past the end of the file (0x2a8) with DT_STRSZ = 0xffffffff
 # DYNSTR-INVALID-LLVM:      DynamicSymbols [
 # DYNSTR-INVALID-LLVM-NEXT:   Symbol {

 ProgramHeaders:
   - Type:  PT_LOAD
     Flags: [ PF_R ]
     Sections:
       - Section: .dynstr
       - Section: .dynamic
 
 ## Check we report a warning when the entry size of the dynamic symbol table is zero.
-# RUN: yaml2obj %s --docnum=14 -o %t15
+# RUN: yaml2obj %s --docnum=15 -o %t15
 # RUN: llvm-readobj --dyn-symbols %t15 2>&1 | FileCheck %s -DFILE=%t15 --check-prefix=DYNSYM-ZERO-ENTSIZE-LLVM
 # RUN: llvm-readelf --dyn-symbols %t15 2>&1 | \
 # RUN:   FileCheck %s -DFILE=%t15 --check-prefix=DYNSYM-ZERO-ENTSIZE-GNU --implicit-check-not="Symbol table"
 
 # DYNSYM-ZERO-ENTSIZE-LLVM:      DynamicSymbols [
 # DYNSYM-ZERO-ENTSIZE-LLVM-NEXT:  warning: '[[FILE]]': section with index 1 has invalid size (0x10) or entry size (0x0)
 # DYNSYM-ZERO-ENTSIZE-LLVM-NEXT: ]
 

llvm/tools/llvm-readobj/ELFDumper.cpp

private:
   DynRegionInfo DynPLTRelRegion;
   Optional<DynRegionInfo> DynSymRegion;
   DynRegionInfo DynamicTable;
   StringRef DynamicStringTable;
   StringRef SOName = "<Not found>";
   const Elf_Hash *HashTable = nullptr;
   const Elf_GnuHash *GnuHashTable = nullptr;
   const Elf_Shdr *DotSymtabSec = nullptr;
+  const Elf_Shdr *DotDynsymSec = nullptr;
   const Elf_Shdr *DotCGProfileSec = nullptr;
   const Elf_Shdr *DotAddrsigSec = nullptr;
-  StringRef DynSymtabName;
   ArrayRef<Elf_Word> ShndxTable;
 
   const Elf_Shdr *SymbolVersionSection = nullptr;   // .gnu.version
   const Elf_Shdr *SymbolVersionNeedSection = nullptr; // .gnu.version_r
   const Elf_Shdr *SymbolVersionDefSection = nullptr; // .gnu.version_d
 
   struct VersionEntry {
     std::string Name;
void ELFDumper<ELFT>::printSymbolsHelper(bool IsDynamic) const {
   Optional<StringRef> StrTable;
   StringRef SymtabName;
   size_t Entries = 0;
   Elf_Sym_Range Syms(nullptr, nullptr);
   const ELFFile<ELFT> *Obj = ObjF->getELFFile();
   if (IsDynamic) {
     StrTable = DynamicStringTable;
     Syms = dynamic_symbols();
-    SymtabName = DynSymtabName;
+
+    if (DotDynsymSec) {
+      if (Expected<StringRef> NameOrErr = Obj->getSectionName(DotDynsymSec)) {
+        SymtabName = *NameOrErr;
+      } else {
+        reportUniqueWarning(
+            createError("unable to get the name of the SHT_DYNSYM section: " +
+                        toString(NameOrErr.takeError())));
+        SymtabName = "<?>";
+      }
+    }
+
     Entries = Syms.size();
   } else {
     if (!DotSymtabSec)
       return;
 
     if (Expected<StringRef> StrTableOrErr =
             Obj->getStringTableForSymtab(*DotSymtabSec))
       StrTable = *StrTableOrErr;
ELFDumper<ELFT>::ELFDumper(const object::ELFObjectFile<ELFT> *ObjF,
   typename ELFT::ShdrRange Sections = cantFail(Obj->sections());
   for (const Elf_Shdr &Sec : Sections) {
     switch (Sec.sh_type) {
     case ELF::SHT_SYMTAB:
       if (!DotSymtabSec)
         DotSymtabSec = &Sec;
       break;
     case ELF::SHT_DYNSYM:
+      if (!DotDynsymSec)
+        DotDynsymSec = &Sec;
+
       if (!DynSymRegion) {
         DynSymRegion = createDRIFrom(&Sec);
         DynSymRegion->Context =
             ("section with index " + Twine(&Sec - &Sections.front())).str();
-        // This is only used (if Elf_Shdr present)for naming section in GNU
-        // style
-        DynSymtabName =
-            unwrapOrError(ObjF->getFileName(), Obj->getSectionName(&Sec));
 
         if (Expected<StringRef> E = Obj->getStringTableForSymtab(Sec))
           DynamicStringTable = *E;
         else
           reportWarning(E.takeError(), ObjF->getFileName());
       }
       break;
     case ELF::SHT_SYMTAB_SHNDX: