[lld] 62a3f0c - [lld-macho] Support absolute symbols
Jez Ng via llvm-commits
llvm-commits at lists.llvm.org
Fri Sep 25 11:28:54 PDT 2020
Author: Jez Ng
Date: 2020-09-25T11:28:35-07:00
New Revision: 62a3f0c9844bb89a48173440145b26212be60f83
URL: https://github.com/llvm/llvm-project/commit/62a3f0c9844bb89a48173440145b26212be60f83
DIFF: https://github.com/llvm/llvm-project/commit/62a3f0c9844bb89a48173440145b26212be60f83.diff
LOG: [lld-macho] Support absolute symbols
They operate like Defined symbols but with no associated InputSection.
Note that `ld64` seems to treat the weak definition flag like a no-op for
absolute symbols, so I have replicated that behavior.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D87909
Added:
lld/test/MachO/abs-symbols.s
lld/test/MachO/invalid/abs-duplicate.s
Modified:
lld/MachO/ExportTrie.cpp
lld/MachO/InputFiles.cpp
lld/MachO/InputFiles.h
lld/MachO/SymbolTable.cpp
lld/MachO/Symbols.cpp
lld/MachO/Symbols.h
lld/MachO/SyntheticSections.cpp
lld/test/MachO/invalid/dso-handle-duplicate.s
Removed:
################################################################################
diff --git a/lld/MachO/ExportTrie.cpp b/lld/MachO/ExportTrie.cpp
index 0093907ea959..bd0c0004309c 100644
--- a/lld/MachO/ExportTrie.cpp
+++ b/lld/MachO/ExportTrie.cpp
@@ -62,11 +62,18 @@ struct ExportInfo {
uint8_t flags = 0;
ExportInfo(const Symbol &sym, uint64_t imageBase)
: address(sym.getVA() - imageBase) {
+ // Set the symbol type.
if (sym.isWeakDef())
flags |= EXPORT_SYMBOL_FLAGS_WEAK_DEFINITION;
- if (sym.isTlv())
- flags |= EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL;
// TODO: Add proper support for re-exports & stub-and-resolver flags.
+
+ // Set the symbol kind.
+ if (sym.isTlv()) {
+ flags |= EXPORT_SYMBOL_FLAGS_KIND_THREAD_LOCAL;
+ } else if (auto *defined = dyn_cast<Defined>(&sym)) {
+ if (defined->isAbsolute())
+ flags |= EXPORT_SYMBOL_FLAGS_KIND_ABSOLUTE;
+ }
}
};
diff --git a/lld/MachO/InputFiles.cpp b/lld/MachO/InputFiles.cpp
index ad2d6a25849a..35211ea25e73 100644
--- a/lld/MachO/InputFiles.cpp
+++ b/lld/MachO/InputFiles.cpp
@@ -222,6 +222,50 @@ void InputFile::parseRelocations(const section_64 &sec,
}
}
+static macho::Symbol *createDefined(const structs::nlist_64 &sym,
+ StringRef name, InputSection *isec,
+ uint32_t value) {
+ if (sym.n_type & N_EXT)
+ // Global defined symbol
+ return symtab->addDefined(name, isec, value, sym.n_desc & N_WEAK_DEF);
+ // Local defined symbol
+ return make<Defined>(name, isec, value, sym.n_desc & N_WEAK_DEF,
+ /*isExternal=*/false);
+}
+
+// Absolute symbols are defined symbols that do not have an associated
+// InputSection. They cannot be weak.
+static macho::Symbol *createAbsolute(const structs::nlist_64 &sym,
+ StringRef name) {
+ if (sym.n_type & N_EXT)
+ return symtab->addDefined(name, nullptr, sym.n_value, /*isWeakDef=*/false);
+ return make<Defined>(name, nullptr, sym.n_value, /*isWeakDef=*/false,
+ /*isExternal=*/false);
+}
+
+macho::Symbol *InputFile::parseNonSectionSymbol(const structs::nlist_64 &sym,
+ StringRef name) {
+ uint8_t type = sym.n_type & N_TYPE;
+ switch (type) {
+ case N_UNDF:
+ return sym.n_value == 0
+ ? symtab->addUndefined(name)
+ : symtab->addCommon(name, this, sym.n_value,
+ 1 << GET_COMM_ALIGN(sym.n_desc));
+ case N_ABS:
+ return createAbsolute(sym, name);
+ case N_PBUD:
+ case N_INDR:
+ error("TODO: support symbols of type " + std::to_string(type));
+ return nullptr;
+ case N_SECT:
+ llvm_unreachable(
+ "N_SECT symbols should not be passed to parseNonSectionSymbol");
+ default:
+ llvm_unreachable("invalid symbol type");
+ }
+}
+
void InputFile::parseSymbols(ArrayRef<structs::nlist_64> nList,
const char *strtab, bool subsectionsViaSymbols) {
// resize(), not reserve(), because we are going to create N_ALT_ENTRY symbols
@@ -229,26 +273,12 @@ void InputFile::parseSymbols(ArrayRef<structs::nlist_64> nList,
symbols.resize(nList.size());
std::vector<size_t> altEntrySymIdxs;
- auto createDefined = [&](const structs::nlist_64 &sym, InputSection *isec,
- uint32_t value) -> Symbol * {
- StringRef name = strtab + sym.n_strx;
- if (sym.n_type & N_EXT)
- // Global defined symbol
- return symtab->addDefined(name, isec, value, sym.n_desc & N_WEAK_DEF);
- // Local defined symbol
- return make<Defined>(name, isec, value, sym.n_desc & N_WEAK_DEF,
- /*isExternal=*/false);
- };
-
for (size_t i = 0, n = nList.size(); i < n; ++i) {
const structs::nlist_64 &sym = nList[i];
+ StringRef name = strtab + sym.n_strx;
- if ((sym.n_type & N_TYPE) == N_UNDF) {
- StringRef name = strtab + sym.n_strx;
- symbols[i] = sym.n_value == 0
- ? symtab->addUndefined(name)
- : symtab->addCommon(name, this, sym.n_value,
- 1 << GET_COMM_ALIGN(sym.n_desc));
+ if ((sym.n_type & N_TYPE) != N_SECT) {
+ symbols[i] = parseNonSectionSymbol(sym, name);
continue;
}
@@ -260,7 +290,7 @@ void InputFile::parseSymbols(ArrayRef<structs::nlist_64> nList,
// use the same subsection. Otherwise, we must split the sections along
// symbol boundaries.
if (!subsectionsViaSymbols) {
- symbols[i] = createDefined(sym, subsecMap[0], offset);
+ symbols[i] = createDefined(sym, name, subsecMap[0], offset);
continue;
}
@@ -282,7 +312,7 @@ void InputFile::parseSymbols(ArrayRef<structs::nlist_64> nList,
if (firstSize == 0) {
// Alias of an existing symbol, or the first symbol in the section. These
// are handled by reusing the existing section.
- symbols[i] = createDefined(sym, firstIsec, 0);
+ symbols[i] = createDefined(sym, name, firstIsec, 0);
continue;
}
@@ -298,15 +328,16 @@ void InputFile::parseSymbols(ArrayRef<structs::nlist_64> nList,
subsecMap[offset] = secondIsec;
// By construction, the symbol will be at offset zero in the new section.
- symbols[i] = createDefined(sym, secondIsec, 0);
+ symbols[i] = createDefined(sym, name, secondIsec, 0);
}
for (size_t idx : altEntrySymIdxs) {
const structs::nlist_64 &sym = nList[idx];
+ StringRef name = strtab + sym.n_strx;
SubsectionMap &subsecMap = subsections[sym.n_sect - 1];
uint32_t off = sym.n_value - sectionHeaders[sym.n_sect - 1].addr;
InputSection *subsec = findContainingSubsection(subsecMap, &off);
- symbols[idx] = createDefined(sym, subsec, off);
+ symbols[idx] = createDefined(sym, name, subsec, off);
}
}
diff --git a/lld/MachO/InputFiles.h b/lld/MachO/InputFiles.h
index bb83c2091178..59b0e41d5de3 100644
--- a/lld/MachO/InputFiles.h
+++ b/lld/MachO/InputFiles.h
@@ -65,6 +65,8 @@ class InputFile {
void parseSymbols(ArrayRef<lld::structs::nlist_64> nList, const char *strtab,
bool subsectionsViaSymbols);
+ Symbol *parseNonSectionSymbol(const structs::nlist_64 &sym, StringRef name);
+
void parseRelocations(const llvm::MachO::section_64 &, SubsectionMap &);
private:
diff --git a/lld/MachO/SymbolTable.cpp b/lld/MachO/SymbolTable.cpp
index 0aa8828d1718..6a3ee9230f41 100644
--- a/lld/MachO/SymbolTable.cpp
+++ b/lld/MachO/SymbolTable.cpp
@@ -131,9 +131,11 @@ Symbol *SymbolTable::addDSOHandle(const MachHeaderSection *header) {
bool wasInserted;
std::tie(s, wasInserted) = insert(DSOHandle::name);
if (!wasInserted) {
+ // FIXME: Make every symbol (including absolute symbols) contain a
+ // reference to their originating file, then add that file name to this
+ // error message.
if (auto *defined = dyn_cast<Defined>(s))
- error("found defined symbol from " + defined->isec->file->getName() +
- " with illegal name " + DSOHandle::name);
+ error("found defined symbol with illegal name " + DSOHandle::name);
}
replaceSymbol<DSOHandle>(s, header);
return s;
diff --git a/lld/MachO/Symbols.cpp b/lld/MachO/Symbols.cpp
index af5d9d25029f..75c699781a61 100644
--- a/lld/MachO/Symbols.cpp
+++ b/lld/MachO/Symbols.cpp
@@ -14,6 +14,21 @@ using namespace llvm;
using namespace lld;
using namespace lld::macho;
+uint64_t Defined::getVA() const {
+ if (isAbsolute())
+ return value;
+ return isec->getVA() + value;
+}
+
+uint64_t Defined::getFileOffset() const {
+ if (isAbsolute()) {
+ error("absolute symbol " + toString(*this) +
+ " does not have a file offset");
+ return 0;
+ }
+ return isec->getFileOffset() + value;
+}
+
void LazySymbol::fetchArchiveMember() { file->fetch(sym); }
// Returns a symbol for an error message.
diff --git a/lld/MachO/Symbols.h b/lld/MachO/Symbols.h
index 14c1ee813420..f0d77d5ce0fe 100644
--- a/lld/MachO/Symbols.h
+++ b/lld/MachO/Symbols.h
@@ -89,18 +89,17 @@ class Defined : public Symbol {
overridesWeakDef(false), weakDef(isWeakDef), external(isExternal) {}
bool isWeakDef() const override { return weakDef; }
-
- bool isTlv() const override { return isThreadLocalVariables(isec->flags); }
+ bool isTlv() const override {
+ return !isAbsolute() && isThreadLocalVariables(isec->flags);
+ }
bool isExternal() const { return external; }
+ bool isAbsolute() const { return isec == nullptr; }
- static bool classof(const Symbol *s) { return s->kind() == DefinedKind; }
-
- uint64_t getVA() const override { return isec->getVA() + value; }
+ uint64_t getVA() const override;
+ uint64_t getFileOffset() const override;
- uint64_t getFileOffset() const override {
- return isec->getFileOffset() + value;
- }
+ static bool classof(const Symbol *s) { return s->kind() == DefinedKind; }
InputSection *isec;
uint32_t value;
diff --git a/lld/MachO/SyntheticSections.cpp b/lld/MachO/SyntheticSections.cpp
index e40b832195b7..5d603d6f90d3 100644
--- a/lld/MachO/SyntheticSections.cpp
+++ b/lld/MachO/SyntheticSections.cpp
@@ -594,11 +594,17 @@ void SymtabSection::writeTo(uint8_t *buf) const {
// TODO support other symbol types
// TODO populate n_desc with more flags
if (auto *defined = dyn_cast<Defined>(entry.sym)) {
- nList->n_type = MachO::N_EXT | MachO::N_SECT;
- nList->n_sect = defined->isec->parent->index;
+ if (defined->isAbsolute()) {
+ nList->n_type = MachO::N_EXT | MachO::N_ABS;
+ nList->n_sect = MachO::NO_SECT;
+ nList->n_value = defined->value;
+ } else {
+ nList->n_type = MachO::N_EXT | MachO::N_SECT;
+ nList->n_sect = defined->isec->parent->index;
+ // For the N_SECT symbol type, n_value is the address of the symbol
+ nList->n_value = defined->value + defined->isec->getVA();
+ }
nList->n_desc |= defined->isWeakDef() ? MachO::N_WEAK_DEF : 0;
- // For the N_SECT symbol type, n_value is the address of the symbol
- nList->n_value = defined->value + defined->isec->getVA();
}
++nList;
}
diff --git a/lld/test/MachO/abs-symbols.s b/lld/test/MachO/abs-symbols.s
new file mode 100644
index 000000000000..25f6c87ca41e
--- /dev/null
+++ b/lld/test/MachO/abs-symbols.s
@@ -0,0 +1,24 @@
+# REQUIRES: x86
+# RUN: llvm-mc -filetype=obj -triple=x86_64-apple-darwin %s -o %t.o
+# RUN: lld -flavor darwinnew -syslibroot %S/Inputs/MacOSX.sdk -lSystem %t.o -o %t
+# RUN: llvm-objdump --macho --syms --exports-trie %t | FileCheck %s
+
+# CHECK-LABEL: SYMBOL TABLE:
+# CHECK-DAG: 000000000000dead g *ABS* _foo
+# CHECK-DAG: 000000000000beef g *ABS* _weakfoo
+
+# CHECK-LABEL: Exports trie:
+# CHECK-DAG: 0x0000DEAD _foo [absolute]
+# CHECK-DAG: 0x0000BEEF _weakfoo [absolute]
+
+.globl _foo, _weakfoo, _main
+.weak_definition _weakfoo
+_foo = 0xdead
+_weakfoo = 0xbeef
+
+.text
+_main:
+ ret
+
+## TODO: once we support emitting local symbols in the symtab, test local
+## absolute symbols too
diff --git a/lld/test/MachO/invalid/abs-duplicate.s b/lld/test/MachO/invalid/abs-duplicate.s
new file mode 100644
index 000000000000..e46c78351385
--- /dev/null
+++ b/lld/test/MachO/invalid/abs-duplicate.s
@@ -0,0 +1,23 @@
+# REQUIRES: x86
+# RUN: split-file %s %t
+# RUN: llvm-mc -filetype=obj -triple=x86_64-apple-darwin %t/test.s -o %t/test.o
+# RUN: llvm-mc -filetype=obj -triple=x86_64-apple-darwin %t/weakfoo.s -o %t/weakfoo.o
+# RUN: not lld -flavor darwinnew -syslibroot %S/../Inputs/MacOSX.sdk -lSystem %t/test.o %t/weakfoo.o -o %t/test 2>&1 | FileCheck %s
+
+# CHECK: lld: error: duplicate symbol: _weakfoo
+
+#--- weakfoo.s
+.globl _weakfoo
+## The weak attribute is ignored for absolute symbols, so we will have a
+## duplicate symbol error for _weakfoo.
+.weak_definition _weakfoo
+_weakfoo = 0x1234
+
+#--- test.s
+.globl _main, _weakfoo
+.weak_definition _weakfoo
+_weakfoo = 0x5678
+
+.text
+_main:
+ ret
diff --git a/lld/test/MachO/invalid/dso-handle-duplicate.s b/lld/test/MachO/invalid/dso-handle-duplicate.s
index 5991c6faff88..4390da62c33e 100644
--- a/lld/test/MachO/invalid/dso-handle-duplicate.s
+++ b/lld/test/MachO/invalid/dso-handle-duplicate.s
@@ -7,8 +7,8 @@
## far-out edge case that should be safe to ignore.
# RUN: llvm-mc -filetype=obj -triple=x86_64-apple-darwin %s -o %t.o
-# RUN: not lld -flavor darwinnew -dylib %t.o -o %t.dylib 2>&1 | FileCheck %s -DFILE=%t.o
-# CHECK: error: found defined symbol from [[FILE]] with illegal name ___dso_handle
+# RUN: not lld -flavor darwinnew -dylib %t.o -o %t.dylib 2>&1 | FileCheck %s
+# CHECK: error: found defined symbol with illegal name ___dso_handle
.globl _main, ___dso_handle
.text
More information about the llvm-commits
mailing list