Refactor data-in-code annotations.
Use a dedicated MachO load command to annotate data-in-code regions.
This is the same format the linker produces for final executable images,
allowing consistency of representation and use of introspection tools
for both object and executable files.
Data-in-code regions are annotated via ".data_region"/".end_data_region"
directive pairs, with an optional region type.
data_region_directive := ".data_region" { region_type }
region_type := "jt8" | "jt16" | "jt32" | "jta32"
end_data_region_directive := ".end_data_region"
The previous handling of ARM-style "$d.*" labels was broken and has
been removed. Specifically, it didn't handle ARM vs. Thumb mode when
marking the end of the section.
rdar://11459456
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157062 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/MC/MachObjectWriter.cpp b/lib/MC/MachObjectWriter.cpp
index 8e4066c..5820a22 100644
--- a/lib/MC/MachObjectWriter.cpp
+++ b/lib/MC/MachObjectWriter.cpp
@@ -21,6 +21,7 @@
#include "llvm/MC/MCMachOSymbolFlags.h"
#include "llvm/MC/MCValue.h"
#include "llvm/Object/MachOFormat.h"
+#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include <vector>
@@ -351,6 +352,21 @@
Write32(Address);
}
+void MachObjectWriter::WriteLinkeditLoadCommand(uint32_t Type,
+ uint32_t DataOffset,
+ uint32_t DataSize) {
+ uint64_t Start = OS.tell();
+ (void) Start;
+
+ Write32(Type);
+ Write32(macho::LinkeditLoadCommandSize);
+ Write32(DataOffset);
+ Write32(DataSize);
+
+ assert(OS.tell() - Start == macho::LinkeditLoadCommandSize);
+}
+
+
void MachObjectWriter::RecordRelocation(const MCAssembler &Asm,
const MCAsmLayout &Layout,
const MCFragment *Fragment,
@@ -654,6 +670,13 @@
macho::DysymtabLoadCommandSize);
}
+ // Add the data-in-code load command size, if used.
+ unsigned NumDataRegions = Asm.getDataRegions().size();
+ if (NumDataRegions) {
+ ++NumLoadCommands;
+ LoadCommandsSize += macho::LinkeditLoadCommandSize;
+ }
+
// Compute the total size of the section data, as well as its file size and vm
// size.
uint64_t SectionDataStart = (is64Bit() ? macho::Header64Size :
@@ -701,6 +724,15 @@
RelocTableEnd += NumRelocs * macho::RelocationInfoSize;
}
+ // Write the data-in-code load command, if used.
+ uint64_t DataInCodeTableEnd = RelocTableEnd + NumDataRegions * 8;
+ if (NumDataRegions) {
+ uint64_t DataRegionsOffset = RelocTableEnd;
+ uint64_t DataRegionsSize = NumDataRegions * 8;
+ WriteLinkeditLoadCommand(macho::LCT_DataInCode, DataRegionsOffset,
+ DataRegionsSize);
+ }
+
// Write the symbol table load command, if used.
if (NumSymbols) {
unsigned FirstLocalSymbol = 0;
@@ -717,10 +749,10 @@
// If used, the indirect symbols are written after the section data.
if (NumIndirectSymbols)
- IndirectSymbolOffset = RelocTableEnd;
+ IndirectSymbolOffset = DataInCodeTableEnd;
// The symbol table is written after the indirect symbol data.
- uint64_t SymbolTableOffset = RelocTableEnd + IndirectSymbolSize;
+ uint64_t SymbolTableOffset = DataInCodeTableEnd + IndirectSymbolSize;
// The string table is written after symbol table.
uint64_t StringTableOffset =
@@ -760,6 +792,23 @@
}
}
+ // Write out the data-in-code region payload, if there is one.
+ for (MCAssembler::const_data_region_iterator
+ it = Asm.data_region_begin(), ie = Asm.data_region_end();
+ it != ie; ++it) {
+ const DataRegionData *Data = &(*it);
+ uint64_t Start = getSymbolAddress(&Layout.getAssembler().getSymbolData(*Data->Start), Layout);
+ uint64_t End = getSymbolAddress(&Layout.getAssembler().getSymbolData(*Data->End), Layout);
+ DEBUG(dbgs() << "data in code region-- kind: " << Data->Kind
+ << " start: " << Start << "(" << Data->Start->getName() << ")"
+ << " end: " << End << "(" << Data->End->getName() << ")"
+ << " size: " << End - Start
+ << "\n");
+ Write32(Start);
+ Write16(End - Start);
+ Write16(Data->Kind);
+ }
+
// Write the symbol table data, if used.
if (NumSymbols) {
// Write the indirect symbol entries.