Update the fragments of symbols in compressed sections.

While unnamed relocations are already cached in side tables in
ELFObjectWriter::RecordRelocation, symbols still need their fragments
updated to refer to the newly compressed fragment (even if that fragment
isn't big enough to fit the offset). Even though we only create
temporary symbols in debug info sections this comes up in 32 bit builds
where even temporary symbols in mergeable sections (such as debug_str)
have to be emitted as named symbols.

I tried a few other ways to do this but they all didn't work for various
reasons:

1) Canonicalize the MCSymbolData in RecordRelocation, nulling out the
Fragment (so it didn't have to be updated by CompressDebugSection). This
doesn't work because some code relies on symbols having fragments to
indicate that they're defined, I think.

2) Canonicalize the MCSymbolData in RecordRelocation to be "first
fragment + absolute offset" so it would be cheaper to just test and
update the fragment in CompressDebugSections. This doesn't work because
the offset computed in RecordRelocation isn't that of the symbol's
fragment, it's the passed in fragment (I haven't figured out what that
fragment is - perhaps it's the location where the relocation is to be
written). And if the fragment offset has to be computed only for this
use we might as well just do it when we need to, in
CompressDebugSection.

I also added an assert to help catch this a bit more clearly, even
though it is UB. The test case improvements would either assert fail
and/or valgrind vail without the fix, even if they wouldn't necessarily
fail the FileCheck output.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@206653 91177308-0d34-0410-b5e6-96231b3b80d8

Author: dwblaikie

lib/MC/ELFObjectWriter.cpp

@@ -616,6 +616,10 @@ static const MCSymbol *getBaseSymbol(const MCAsmLayout &Layout,
 void ELFObjectWriter::WriteSymbol(SymbolTableWriter &Writer, ELFSymbolData &MSD,
                                   const MCAsmLayout &Layout) {
   MCSymbolData &OrigData = *MSD.SymbolData;
+  assert(!OrigData.getFragment() ||
+         (&OrigData.getFragment()->getParent()->getSection() ==
+          &OrigData.getSymbol().getSection()) &&
+             "The symbol's section doesn't match the fragment's symbol");
   const MCSymbol *Base = getBaseSymbol(Layout, OrigData.getSymbol());
 
   // This has to be in sync with when computeSymbolTable uses SHN_ABS or
@@ -1244,6 +1248,19 @@ getCompressedFragment(MCAsmLayout &Layout,
   return CompressedFragment;
 }
 
+static void UpdateSymbols(const MCAsmLayout &Layout, const MCSectionData &SD,
+                          MCAssembler::symbol_range Symbols,
+                          MCFragment *NewFragment) {
+  for (MCSymbolData &Data : Symbols) {
+    MCFragment *F = Data.getFragment();
+    if (F && F->getParent() == &SD) {
+      Data.setOffset(Data.getOffset() +
+                     Layout.getFragmentOffset(Data.Fragment));
+      Data.setFragment(NewFragment);
+    }
+  }
+}
+
 static void CompressDebugSection(MCAssembler &Asm, MCAsmLayout &Layout,
                                  const MCSectionELF &Section,
                                  MCSectionData &SD) {
@@ -1257,6 +1274,10 @@ static void CompressDebugSection(MCAssembler &Asm, MCAsmLayout &Layout,
   if (!CompressedFragment)
     return;
 
+  // Update the fragment+offsets of any symbols referring to fragments in this
+  // section to refer to the new fragment.
+  UpdateSymbols(Layout, SD, Asm.symbols(), CompressedFragment.get());
+
   // Invalidate the layout for the whole section since it will have new and
   // different fragments now.
   Layout.invalidateFragmentsFrom(&Fragments.front());

test/MC/ELF/compression.s

@@ -1,6 +1,8 @@
-// RUN: llvm-mc -filetype=obj -compress-debug-sections -triple x86_64-pc-linux-gnu %s -o %t
+// RUN: llvm-mc -filetype=obj -compress-debug-sections -triple x86_64-pc-linux-gnu < %s -o %t
 // RUN: llvm-objdump -s %t | FileCheck %s
 // RUN: llvm-dwarfdump -debug-dump=abbrev %t | FileCheck --check-prefix=ABBREV %s
+// RUN: llvm-mc -filetype=obj -compress-debug-sections -triple i386-pc-linux-gnu < %s \
+// RUN:     | llvm-readobj -symbols - | FileCheck --check-prefix=386-SYMBOLS %s
 
 // REQUIRES: zlib
 
@@ -22,12 +24,21 @@
 // ABBREV: Abbrev table for offset: 0x00000000
 // ABBREV: [1] DW_TAG_compile_unit DW_CHILDREN_no
 
+// In x86 32 bit named symbols are used for temporary symbols in merge
+// sections, so make sure we handle symbols inside compressed sections
+// 386-SYMBOLS: Name: .Linfo_string0
+// 386-SYMBOLS-NOT: }
+// 386-SYMBOLS: Section: .zdebug_str
+
 	.section	.debug_line,"",@progbits
 
 	.section	.debug_abbrev,"",@progbits
+.Lsection_abbrev:
 	.byte	1                       # Abbreviation Code
 	.byte	17                      # DW_TAG_compile_unit
 	.byte	0                       # DW_CHILDREN_no
+	.byte	27                      # DW_AT_comp_dir
+	.byte	14                      # DW_FORM_strp
 	.byte	0                       # EOM(1)
 	.byte	0                       # EOM(2)
 	.text
@@ -38,3 +49,15 @@ foo:
         nop
 	.cfi_endproc
 	.cfi_sections .debug_frame
+
+	.section        .debug_str,"MS",@progbits,1
+.Linfo_string0:
+        .asciz  "foo"
+
+	.section	.debug_info,"",@progbits
+	.long	40                      # Length of Unit
+	.short	4                       # DWARF version number
+	.long	.Lsection_abbrev        # Offset Into Abbrev. Section
+	.byte	4                       # Address Size (in bytes)
+	.byte	1                       # Abbrev [1] DW_TAG_compile_unit
+	.long	.Linfo_string0          # DW_AT_comp_dir