[ELF] Add OVERWRITE_SECTIONS command

[MaskRay]

Summary

This implements https://sourceware.org/bugzilla/show_bug.cgi?id=26404

An OVERWRITE_SECTIONS command is a SECTIONS variant which contains several
output section descriptions. The output sections do not have specify an order.
Similar to INSERT [BEFORE|AFTER], LinkerScript::hasSectionsCommand is not
set, so the built-in rules (see docs/ELF/linker_script.rst) still apply.
OVERWRITE_SECTIONS can be more convenient than INSERT because it does not
need an anchor section.

The initial syntax is intentionally narrow to facilitate backward compatible
extensions in the future. Symbol assignments cannot be used.

This feature is versatile. To list a few usage:

• Use section : { KEEP(...) } to retain input sections under GC
• Define encapsulation symbols (start/end) for an output section
• Use section : ALIGN(...) : { ... } to overalign an output section (similar to ld64 -sectalign)

When an output section is specified by both OVERWRITE_SECTIONS and
INSERT, INSERT is processed after overwrite sections. To make this work,
this patch changes InsertCommand to use name based matching instead of pointer
based matching. (This may cause a difference when INSERT moves one output
section more than once. Such duplicate commands should not be used in practice
(seems that in GNU ld the output sections may just disappear).)

A linker script can be used without -T/--script. The traditional SECTIONS
commands are concatenated, so a wrong rule can be more noticeable from the
section order. This feature if misused can be less noticeable, just like
INSERT.

[peter.smith]

Will take a look this week. Apologies, bit behind due to UK public holiday on Monday.

[peter.smith]

Sorry for the delay in responding. I'm in favour of adding this as it should combine well with LLD when no default linker script is used.

I'm thinking overwrite-section-script may be a better name than --overwrite-script. It would reinforce the narrow nature of the script, to overwrite specific SECTIONS commands, and also help people not seeing it as overwriting their whole linker script.

With respect to Roland's comment in https://sourceware.org/bugzilla/show_bug.cgi?id=26404#c10 one way of having this work by adding another script fragment would be to add a keyword at the top of the linker script for example:

/* This linker script fragment is an overwrite script */
OVERWRITE_SCRIPT()

Will be good to have a test to show how --overwrite-script combines with INSERT AFTER and INSERT BEFORE

lld/ELF/Driver.cpp

222 ↗

(On Diff #348461)

Could we put /* overwrite */ before the false.

lld/ELF/Options.td

369 ↗

(On Diff #348461)

I'd recommend being more specific here
"Read a script containing SECTION commands that replace SECTIONS in the main or default linker script."

Will be good to update docs/ld.lld.1 and perhaps the release notes?

lld/ELF/ScriptParser.cpp

284

expect("SECTIONS") will do a good job of preventing other linker script commands, but it may leave people wondering why? Would it be possible to do something like:

expect("SECTIONS")
if (!errorCount())
  // Diagnostic "overwrite script can only contain SECTIONS commands."

lld/test/ELF/linkerscript/overwrite-script.test

106 ↗

(On Diff #348461)

Will be good to have a diagnostic for an overwrite script that doesn't contain SECTIONS.

[MaskRay]

• Rename to --overwrite-section-script
• Test INSERT
• Test A = 1; outside SECTIONS

[MaskRay]

In D103303#2796607, @peter.smith wrote:

Sorry for the delay in responding. I'm in favour of adding this as it should combine well with LLD when no default linker script is used.

I'm thinking overwrite-section-script may be a better name than --overwrite-script. It would reinforce the narrow nature of the script, to overwrite specific SECTIONS commands, and also help people not seeing it as overwriting their whole linker script.

Much appreciate your review. I agree that --overwrite-section-script is better.

With respect to Roland's comment in https://sourceware.org/bugzilla/show_bug.cgi?id=26404#c10 one way of having this work by adding another script fragment would be to add a keyword at the top of the linker script for example:

/* This linker script fragment is an overwrite script */
OVERWRITE_SCRIPT()

This would add some complexity to readLinkerScript as we will need to check the first token. Do we really need it?
This is just a matter between $file and -Wl,--overwrite-script=$file in the build system.

[mcgrathr]

In D103303#2797099, @MaskRay wrote:

This would add some complexity to readLinkerScript as we will need to check the first token. Do we really need it?
This is just a matter between $file and -Wl,--overwrite-script=$file in the build system.

That is no small matter in many build systems. Switches that name input files are especially difficult in general.

Even if you have no sympathy for build systems you don't know about, requiring a switch precludes using this functionality as a "transparent" part of a library implementation. Supporting "just add -lfoo to your compile" as the end-user API for using a library is highly desireable, and often a non-negotiable requirement. Having a libfoo.a input linker script that wraps an INPUT ( libfoo-impl.a ) along with other magic is a well-worn and very successful technique. The use cases you cite as motivations for adding this feature in the first place are exactly the kinds of things one would like to encapsulate as "hidden magic" in a library that's "just linked".

[peter.smith]

Thanks for the updates.

One idea that might work better for the linker script fragment is to use a new keyword. For example instead of using --overwrite-section-script and SECTIONS. Use a new keyword like OVERWRITE_SECTIONS this could essentially be mixed in with any other linker script without needing --overwrite-section-script.

[mcgrathr]

In D103303#2798557, @peter.smith wrote:

Thanks for the updates.

One idea that might work better for the linker script fragment is to use a new keyword. For example instead of using --overwrite-section-script and SECTIONS. Use a new keyword like OVERWRITE_SECTIONS this could essentially be mixed in with any other linker script without needing --overwrite-section-script.

I think that's an excellent idea. The new command really has different semantics from the old one, so making it a different keyword feels right.
It also makes it much simpler to have clear rules about what linker constructs are allowed in what kinds of contexts. Another syntax that would have equivalent properties from the user's point of view would be a trailing OVERWRITE keyword. That frankly seems less clear than OVERWRITE_SECTIONS, but it's more consistent with the existing SECTIONS { ... } INSERT ... syntax.

[MaskRay]

Use OVERWRITE_SECTIONS
Delete --overwrite-section-script

[MaskRay]

Introducing new syntax was my initial thought. Now I agree that having overwrite in the name is useful, OVERWRITE_SECTIONS looks great to me.
It now feels to me that --overwrite-section-script is less essential so I just removed the option from the patch.

[peter.smith]

Thanks very much for the update. I've made a few suggestions for the documentation.

One case I wanted to check was something like the following:

OVERWRITE_SECTIONS {
  orphan_section_name_matching_nothing_seen_before : { ... }
}

SECTIONS {
  another_orphan_section_name_matching_nothing_seen_before_including_overwrite : { ... }
} INSERT AFTER orphan_section_name_matching_nothing_seen_before

Would that place "orphan_section_name_matching_nothing_seen_before" using orphan rules then "orphan_section_name_matching_nothing_seen_before" immediately after it?

lld/docs/ELF/linker_script.rst

57

Suggest adding something like the insertion occurs after input sections have been mapped to output sections but before orphan sections have been processed.

59

A suggestion to avoid the "absence of a non-insert" which is difficult to parse.

In the case where no linker script has been provided or every `SECTIONS` command is followed by INSERT, LLD applies built-in rules which are similar to GNU ld's internal linker scripts.

147

"also appears in a `SECTIONS` command," might read a bit better. Only a weak suggestion though.

151

"also appears in"?

156

I got a little confused about this at first. Maybe worth being more specific:
"will be provided by the description in the `OVERWRITE_SECTIONS` command."

[MaskRay]

Thanks for the comments. Addressed and add another INSERT test.

[peter.smith]

LGTM thanks. One tiny typo spotted in a comment in the test.

lld/test/ELF/linkerscript/overwrite-sections.test

81

typo "than -> then"

[MaskRay]

fix a typo

[MaskRay]

I'll push this next week.

lld/ELF/LinkerScript.h

struct ByteCommand : BaseCommand {
   // This is just an offset of this assignment command in the output section.
   unsigned offset;
 
   // Size of this data command.
   unsigned size;
 };
 
 struct InsertCommand {
-  OutputSection *os;
+  StringRef name;
   bool isAfter;
   StringRef where;
 };
 
 struct PhdrsCommand {
   StringRef name;
   unsigned type = llvm::ELF::PT_NULL;
   bool hasFilehdr = false;
public:
 
   // A list of symbols referenced by the script.
   std::vector<llvm::StringRef> referencedSymbols;
 
   // Used to implement INSERT [AFTER|BEFORE]. Contains output sections that need
   // to be reordered.
   std::vector<InsertCommand> insertCommands;
 
+  // OutputSections specified by OVERWRITE_SECTIONS.
+  std::vector<OutputSection *> overwriteSections;
+
   // Sections that will be warned/errored by --orphan-handling.
   std::vector<const InputSectionBase *> orphanSections;
 };
 
 extern LinkerScript *script;
 
 } // end namespace elf
 } // end namespace lld
 
 #endif // LLD_ELF_LINKER_SCRIPT_H

lld/ELF/LinkerScript.cpp

 }
 
 // Process INSERT [AFTER|BEFORE] commands. For each command, we move the
 // specified output section to the designated place.
 void LinkerScript::processInsertCommands() {
   for (const InsertCommand &cmd : insertCommands) {
     // If cmd.os is empty, it may have been discarded by
     // adjustSectionsBeforeSorting(). We do not handle such output sections.
-    auto from = llvm::find(sectionCommands, cmd.os);
+    auto from = llvm::find_if(sectionCommands, [&](BaseCommand *base) {
+      return isa<OutputSection>(base) &&
+             cast<OutputSection>(base)->name == cmd.name;
+    });
     if (from == sectionCommands.end())
       continue;
+    OutputSection *osec = cast<OutputSection>(*from);
     sectionCommands.erase(from);
 
     auto insertPos = llvm::find_if(sectionCommands, [&cmd](BaseCommand *base) {
       auto *to = dyn_cast<OutputSection>(base);
       return to != nullptr && to->name == cmd.where;
     });
     if (insertPos == sectionCommands.end()) {
-      error("unable to insert " + cmd.os->name +
+      error("unable to insert " + osec->name +
             (cmd.isAfter ? " after " : " before ") + cmd.where);
     } else {
       if (cmd.isAfter)
         ++insertPos;
-      sectionCommands.insert(insertPos, cmd.os);
+      sectionCommands.insert(insertPos, osec);
     }
   }
 }
 
 // Symbols defined in script should not be inlined by LTO. At the same time
 // we don't know their final values until late stages of link. Here we scan
 // over symbol assignment commands and create placeholder symbols if needed.
 void LinkerScript::declareSymbols() {
if (auto *cmd = dyn_cast<InputSectionDescription>(base)) {
       ret.insert(ret.end(), cmd->sectionBases.begin(), cmd->sectionBases.end());
     }
   }
   return ret;
 }
 
 // Create output sections described by SECTIONS commands.
 void LinkerScript::processSectionCommands() {
-  size_t i = 0;
-  for (BaseCommand *base : sectionCommands) {
-    if (auto *sec = dyn_cast<OutputSection>(base)) {
-      std::vector<InputSectionBase *> v = createInputSectionList(*sec);
+  auto process = [this](OutputSection *osec) {
+    std::vector<InputSectionBase *> v = createInputSectionList(*osec);
 
-      // The output section name `/DISCARD/' is special.
-      // Any input section assigned to it is discarded.
-      if (sec->name == "/DISCARD/") {
-        for (InputSectionBase *s : v)
-          discard(s);
-        discardSynthetic(*sec);
-        sec->sectionCommands.clear();
-        continue;
-      }
+    if (osec->name == "/DISCARD/") {
+      discardSynthetic(*osec);
+      osec->sectionCommands.clear();
+      return false;
 
-      // This is for ONLY_IF_RO and ONLY_IF_RW. An output section directive
-      // ".foo : ONLY_IF_R[OW] { ... }" is handled only if all member input
-      // sections satisfy a given constraint. If not, a directive is handled
-      // as if it wasn't present from the beginning.
-      //
-      // Because we'll iterate over SectionCommands many more times, the easy
-      // way to "make it as if it wasn't present" is to make it empty.
-      if (!matchConstraints(v, sec->constraint)) {
-        for (InputSectionBase *s : v)
-          s->parent = nullptr;
-        sec->sectionCommands.clear();
-        continue;
-      }
+    if (!matchConstraints(v, osec->constraint)) {
+      osec->sectionCommands.clear();
+      return false;
 
-      // Handle subalign (e.g. ".foo : SUBALIGN(32) { ... }"). If subalign
-      // is given, input sections are aligned to that value, whether the
-      // given value is larger or smaller than the original section alignment.
-      if (sec->subalignExpr) {
-        uint32_t subalign = sec->subalignExpr().getValue();
-        for (InputSectionBase *s : v)
-          s->alignment = subalign;
-      }
+    if (osec->subalignExpr) {
+      uint32_t subalign = osec->subalignExpr().getValue();
 
-      // Set the partition field the same way OutputSection::recordSection()
-      // does. Partitions cannot be used with the SECTIONS command, so this is
-      // always 1.
-      sec->partition = 1;
+    osec->partition = 1;
+    return true;
+  };
 
-      sec->sectionIndex = i++;
+  // Process OVERWRITE_SECTIONS first so that it can overwrite the main script
+  // or orphans.
+  DenseMap<StringRef, OutputSection *> map;
+  size_t i = 0;
+  for (OutputSection *osec : overwriteSections)
+    if (process(osec) && !map.try_emplace(osec->name, osec).second)
+      warn("OVERWRITE_SECTIONS specifies duplicate " + osec->name);
+  for (BaseCommand *&base : sectionCommands)
+    if (auto *osec = dyn_cast<OutputSection>(base)) {
+      if (OutputSection *overwrite = map.lookup(osec->name)) {
+        log(overwrite->location + " overwrites " + osec->name);
+        overwrite->sectionIndex = i++;
+        base = overwrite;
+      } else if (process(osec)) {
+        osec->sectionIndex = i++;
       }
     }
+
+  // If an OVERWRITE_SECTIONS specified output section is not in
+  // sectionCommands, append it to the end. The section will be inserted by
+  // orphan placement.
+  for (OutputSection *osec : overwriteSections)
+    if (osec->partition == 1 && osec->sectionIndex == UINT32_MAX)
+      sectionCommands.push_back(osec);
 }
 
 void LinkerScript::processSymbolAssignments() {
   // Dot outside an output section still represents a relative address, whose
   // sh_shndx should not be SHN_UNDEF or SHN_ABS. Create a dummy aether section
   // that fills the void outside a section. It has an index of one, which is
   // indistinguishable from any other regular section index.
   aether = make<OutputSection>("", 0, SHF_ALLOC);

lld/ELF/ScriptParser.cpp

private:
   void readExtern();
   void readGroup();
   void readInclude();
   void readInput();
   void readMemory();
   void readOutput();
   void readOutputArch();
   void readOutputFormat();
+  void readOverwriteSections();
   void readPhdrs();
   void readRegionAlias();
   void readSearchDir();
   void readSections();
   void readTarget();
   void readVersion();
   void readVersionScriptCommand();
 
while (!atEOF()) {
     } else if (tok == "MEMORY") {
       readMemory();
     } else if (tok == "OUTPUT") {
       readOutput();
     } else if (tok == "OUTPUT_ARCH") {
       readOutputArch();
     } else if (tok == "OUTPUT_FORMAT") {
       readOutputFormat();
+    } else if (tok == "OVERWRITE_SECTIONS") {
+      readOverwriteSections();
     } else if (tok == "PHDRS") {
       readPhdrs();
     } else if (tok == "REGION_ALIAS") {
       readRegionAlias();
     } else if (tok == "SEARCH_DIR") {
       readSearchDir();
     } else if (tok == "SECTIONS") {
       readSections();

 
 void ScriptParser::readDefsym(StringRef name) {
   if (errorCount())
     return;
   Expr e = readExpr();
   if (!atEOF())
     setError("EOF expected, but got " + next());
   SymbolAssignment *cmd = make<SymbolAssignment>(name, e, getCurrentLocation());
   script->sectionCommands.push_back(cmd);
 }
 
 void ScriptParser::addFile(StringRef s) {
   if (isUnderSysroot && s.startswith("/")) {
     SmallString<128> pathData;
     StringRef path = (config->sysroot + s).toStringRef(pathData);
     if (sys::fs::exists(path)) {
       driver->addFile(saver.save(path), /*withLOption=*/false);
Expr moveDot = [=] {
     for (BaseCommand *cmd : v)
       max = std::max(max, cast<OutputSection>(cmd)->size);
     return addrExpr().getValue() + max;
   };
   v.push_back(make<SymbolAssignment>(".", moveDot, getCurrentLocation()));
   return v;
 }
 
+void ScriptParser::readOverwriteSections() {
+  expect("{");
+  while (!errorCount() && !consume("}"))
+    script->overwriteSections.push_back(readOutputSectionDescription(next()));
+}
+
 void ScriptParser::readSections() {
   expect("{");
   std::vector<BaseCommand *> v;
   while (!errorCount() && !consume("}")) {
     StringRef tok = next();
     if (tok == "OVERLAY") {
       for (BaseCommand *cmd : readOverlay())
         v.push_back(cmd);
void ScriptParser::readSections() {
   bool isAfter = false;
   if (consume("AFTER"))
     isAfter = true;
   else if (!consume("BEFORE"))
     setError("expected AFTER/BEFORE, but got '" + next() + "'");
   StringRef where = next();
   for (BaseCommand *cmd : v)
     if (auto *os = dyn_cast<OutputSection>(cmd))
-      script->insertCommands.push_back({os, isAfter, where});
+      script->insertCommands.push_back({os->name, isAfter, where});
 }
 
 void ScriptParser::readTarget() {
   // TARGET(foo) is an alias for "--format foo". Unlike GNU linkers,
   // we accept only a limited set of BFD names (i.e. "elf" or "binary")
   // for --format. We recognize only /^elf/ and "binary" in the linker
   // script as well.
   expect("(");

lld/docs/ELF/linker_script.rst

 overridden. The other form requires ``symbol`` to be already defined.
 
 For a simple assignment like ``alias = aliasee;``, the ``st_type`` field is
 copied from the original symbol. Any arithmetic operation (e.g. ``+ 0`` will
 reset ``st_type`` to ``STT_NOTYPE``.
 
 The ``st_size`` field is set to 0.
 
+SECTIONS command
+~~~~~~~~~~~~~~~~
+
+A ``SECTIONS`` command looks like:
+
+::
+
+  SECTIONS {
+    section-command
+    section-command
+    ...
+  } [INSERT [AFTER|BEFORE] anchor_section;]
+
+Each section-command can be a symbol assignment, an output section description,
+or an overlay description.
+
+When the ``INSERT`` keyword is present, the ``SECTIONS`` command describes some
+output sections which should be inserted after or before the specified anchor
+section. The insertion occurs after input sections have been mapped to output
+sections but before orphan sections have been processed.
+
+In the case where no linker script has been provided or every ``SECTIONS``
+command is followed by ``INSERT``, LLD applies built-in rules which are similar
+to GNU ld's internal linker scripts.
+
+- Align the first section in a ``PT_LOAD`` segment according to ``-z noseparate-code``,
+  ``-z separate-code``, or ``-z separate-loadable-segments``
+- Define ``__bss_start``, ``end``, ``_end``, ``etext``, ``_etext``, ``edata``, ``_edata``
+- Sort ``.ctors.*``/``.dtors.*``/``.init_array.*``/``.fini_array.*`` and PowerPC64 specific ``.toc``
+- Place input ``.text.*`` into output ``.text``, and handle certain variants
+  (``.text.hot.``, ``.text.unknown.``, ``.text.unlikely.``, etc) in the precense of
+  ``-z keep-text-section-prefix``.
+
 Output section description
 ~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 The description of an output section looks like:
 
 ::
 
   section [address] [(type)] : [AT(lma)] [ALIGN(section_align)] [SUBALIGN](subsection_align)] {

 The two keywords cannot be specified at the same time.
 
 If neither ``AT(lma)`` nor ``AT>lma_region`` is specified:
 
 - If the previous section is also in the default LMA region, and the two
   section have the same memory regions, the difference between the LMA and the
   VMA is computed to be the same as the previous difference.
 - Otherwise, the LMA is set to the VMA.
+
+Overwrite sections
+~~~~~~~~~~~~~~~~~~
+
+An ``OVERWRITE_SECTIONS`` command looks like:
+
+::
+
+  OVERWRITE_SECTIONS {
+    output-section-description
+    output-section-description
+    ...
+  }
+
+Unlike a ``SECTIONS`` command, ``OVERWRITE_SECTIONS``  does not specify a
+section order or suppress the built-in rules.
+
+If a described output section description also appears in a ``SECTIONS``
+command, the ``OVERWRITE_SECTIONS`` command wins; otherwise, the output section
+will be added somewhere following the usual orphan section placement rules.
+
+If a described output section description also appears in an ``INSERT
+[AFTER|BEFORE]`` command, the description will be provided by the
+description in the ``OVERWRITE_SECTIONS`` command while the insert command
+still applies (possibly after orphan section placement). It is recommended to
+leave the brace empty (i.e. ``section : {}``) for the insert command, because
+its description will be ignored anyway.

lld/docs/ReleaseNotes.rst

 
 ELF Improvements
 ----------------
 
 * ``-Bsymbolic -Bsymbolic-functions`` has been changed to behave the same as ``-Bsymbolic-functions``. This matches GNU ld.
   (`D102461 <https://reviews.llvm.org/D102461>`_)
 * ``-Bno-symbolic`` has been added.
   (`D102461 <https://reviews.llvm.org/D102461>`_)
+* A new linker script command ``OVERWRITE_SECTIONS`` has been added.
+  (`D103303 <https://reviews.llvm.org/D103303>`_)
 
 Breaking changes
 ----------------
 
 * ``--shuffle-sections=<seed>`` has been changed to ``--shuffle-sections=<section-glob>=<seed>``.
   Specify ``*`` as ``<section-glob>`` to get the previous behavior.
 
 COFF Improvements

lld/test/ELF/linkerscript/insert-duplicate.test

 # REQUIRES: x86
-## Test that we can handle cases where an output section is specified by multiple
-## INSERT commands. Each output section description creates a new instance.
-## A redundant description matches no input sections and thus is a no-op.
+## Document the behavior when an output section is specified by multiple
+## INSERT commands. It is discouraged in the real world.
 
 # RUN: llvm-mc -filetype=obj -triple=x86_64 %p/Inputs/insert-after.s -o %t.o
 # RUN: ld.lld -T %s %t.o -o %t
 # RUN: llvm-readelf -S -l %t | FileCheck %s
 
 # CHECK:      Name      Type     Address          Off
 # CHECK-NEXT:           NULL     0000000000000000 000000
-# CHECK-NEXT: .text     PROGBITS 00000000002011c8 0001c8
-# CHECK-NEXT: .foo.data PROGBITS 00000000002021d0 0001d0
-# CHECK-NEXT: .foo.text PROGBITS 00000000002031d8 0001d8
+# CHECK-NEXT: .text     PROGBITS 0000000000201158 000158
+# CHECK-NEXT: .foo.text PROGBITS 0000000000201160 000160
+# CHECK-NEXT: .foo.data PROGBITS 0000000000202168 000168
 # CHECK:      Type
 # CHECK-NEXT: PHDR {{.*}} R
 # CHECK-NEXT: LOAD {{.*}} R
 # CHECK-NEXT: LOAD {{.*}} R E
 # CHECK-NEXT: LOAD {{.*}} RW
-# CHECK-NEXT: LOAD {{.*}} R E
-# CHECK-NEXT: LOAD {{.*}} RW
 # CHECK-NEXT: GNU_STACK {{.*}} RW
 
 ## First, move .foo.data after .foo.text
 SECTIONS { .foo.data : { *(.foo.data) } } INSERT AFTER .foo.text;
 
 ## Next, move .foo.text after .foo.data
 SECTIONS { .foo.text : { *(.foo.text) } } INSERT AFTER .foo.data;
 
-## No-op. The .foo.data output section is a different instance and matches no
-## input sections.
+## Then, move .foo.data after .foo.text again.
 SECTIONS { .foo.data : { *(.foo.data) } } INSERT AFTER .foo.text;

lld/test/ELF/linkerscript/overwrite-sections-discard.test

+# REQUIRES: x86
+# RUN: split-file %s %t
+# RUN: llvm-mc -filetype=obj -triple=x86_64 %t/a.s -o %t/a.o
+
+## /DISCARD/ specified by OVERWRITE_SECTIONS can discard sections as well.
+
+# RUN: ld.lld %t/overwrite.lds -T %t/main.lds %t/a.o -o %t1 2>&1 | count 0
+# RUN: llvm-readelf -S -l %t1 | FileCheck %s
+# RUN: ld.lld %t/overwrite.lds %t/overwrite.lds -T %t/main.lds %t/a.o -o %t1 2>&1 | count 0
+# RUN: llvm-readelf -S -l %t1 | FileCheck %s
+
+# CHECK:     Name
+# CHECK-NOT: .data
+
+#--- a.s
+.globl _start
+_start:
+
+.section .data.1,"aw"; .byte 1
+.section .data.2,"aw"; .byte 2
+
+#--- main.lds
+SECTIONS {
+  /DISCARD/ : { *(.data.1) }
+}
+
+#--- overwrite.lds
+OVERWRITE_SECTIONS {
+  /DISCARD/ : { *(.data.2) }
+}

lld/test/ELF/linkerscript/overwrite-sections.test

+# REQUIRES: x86
+# RUN: rm -rf %t && split-file %s %t
+# RUN: llvm-mc -filetype=obj -triple=x86_64 %t/a.s -o %t/a.o
+
+## There is no main linker script. OVERWRITE_SECTIONS defines output section
+## descriptions and follows the usual orphan placement rules.
+
+# RUN: ld.lld %t/overwrite1.lds %t/a.o -o %t1
+# RUN: llvm-readelf -S %t1 | FileCheck %s --check-prefix=CHECK1
+
+# CHECK1:      Name      Type     Address  Off      Size   ES Flg Lk Inf Al
+# CHECK1-NEXT:           NULL     [[#%x,]] [[#%x,]] 000000
+# CHECK1-NEXT: .text     PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  4
+# CHECK1-NEXT: .foo.text PROGBITS [[#%x,]] [[#%x,]] 000002 00 WAX  0   0  8
+# CHECK1-NEXT: .data     PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK1-NEXT: .comment  PROGBITS {{.*}}
+
+# RUN: ld.lld %t/overwrite1.lds %t/overwrite1.lds %t/a.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=WARN1
+
+# WARN1: warning: OVERWRITE_SECTIONS specifies duplicate .foo.text
+
+## The output section description order (.foo.data .foo.text) does not affect
+## the output order.
+
+# RUN: ld.lld %t/overwrite2.lds %t/a.o -o %t2
+# RUN: llvm-readelf -S -s %t2 | FileCheck %s --check-prefix=CHECK2
+
+# CHECK2:      [Nr] Name      Type     Address  Off      Size   ES Flg Lk Inf Al
+# CHECK2-NEXT: [ 0]           NULL     [[#%x,]] [[#%x,]] 000000
+# CHECK2-NEXT: [ 1] .text     PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  4
+# CHECK2-NEXT: [ 2] .foo.text PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  1
+# CHECK2-NEXT: [ 3] .data     PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK2-NEXT: [ 4] .foo.data PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK2-NEXT: [ 5] .comment  PROGBITS {{.*}}
+
+# CHECK2:        Num: Value            Size Type   Bind   Vis     Ndx Name
+# CHECK2:      [[#]]: [[#%x,ADDR:]]       0  NOTYPE GLOBAL DEFAULT   4 FOO_DATA_START
+# CHECK2-NEXT: [[#]]: {{0*}}[[#ADDR+1]]   0  NOTYPE GLOBAL DEFAULT   4 FOO_DATA_END
+# CHECK2-NEXT: [[#]]: [[#%x,ADDR:]]       0  NOTYPE GLOBAL DEFAULT   2 FOO_TEXT_START
+# CHECK2-NEXT: [[#]]: {{0*}}[[#ADDR+1]]   0  NOTYPE GLOBAL DEFAULT   2 FOO_TEXT_END
+
+## OVERWRITE_SECTIONS is processed before the main script. The main script
+## specifies the output order. The alignment of .foo.text is specified by
+## OVERWRITE_SECTIONS insteaad of the main script.
+
+# RUN: ld.lld -T %t/main.lds %t/overwrite1.lds %t/a.o -o %t3 2>&1 | count 0
+# RUN: llvm-readelf -S %t3 | FileCheck %s --check-prefix=CHECK3
+# RUN: ld.lld %t/overwrite1.lds -T %t/main.lds %t/a.o -o %t3 2>&1 | count 0
+# RUN: llvm-readelf -S %t3 | FileCheck %s --check-prefix=CHECK3
+
+# CHECK3:      Name      Type     Address  Off      Size   ES Flg Lk Inf Al
+# CHECK3-NEXT:           NULL     [[#%x,]] [[#%x,]] 000000
+# CHECK3-NEXT: .data     PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK3-NEXT: .text     PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  4
+# CHECK3-NEXT: .foo.text PROGBITS [[#%x,]] [[#%x,]] 000002 00 WAX  0   0  8
+# CHECK3-NEXT: .comment  PROGBITS {{.*}}
+
+# RUN: ld.lld -T %t/main.lds %t/overwrite1.lds %t/a.o -o %t3 --verbose 2>&1 | FileCheck %s --check-prefix=VERBOSE3
+
+# VERBOSE3: {{.*}}overwrite1.lds:[[#]] overwrites .foo.text
+
+## If INSERT commands are specified, INSERT commands are processed after overwrite sections.
+# RUN: ld.lld %t/insert.lds %t/overwrite2.lds %t/a.o -o %t4
+# RUN: llvm-readelf -S -s %t4 | FileCheck %s --check-prefix=CHECK4
+
+# CHECK4:      [Nr] Name      Type     Address  Off      Size   ES Flg Lk Inf Al
+# CHECK4-NEXT: [ 0]           NULL     [[#%x,]] [[#%x,]] 000000
+# CHECK4-NEXT: [ 1] .text     PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  4
+# CHECK4-NEXT: [ 2] .data     PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK4-NEXT: [ 3] .foo.data PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK4-NEXT: [ 4] .foo.text PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  1
+# CHECK4-NEXT: [ 5] .comment  PROGBITS {{.*}}
+
+# CHECK4:        Num: Value            Size Type   Bind   Vis     Ndx Name
+# CHECK4:      [[#]]: [[#%x,ADDR:]]       0  NOTYPE GLOBAL DEFAULT   4 FOO_TEXT_START
+# CHECK4-NEXT: [[#]]: {{0*}}[[#ADDR+1]]   0  NOTYPE GLOBAL DEFAULT   4 FOO_TEXT_END
+# CHECK4-NEXT: [[#]]: [[#%x,ADDR:]]       0  NOTYPE GLOBAL DEFAULT   3 FOO_DATA_START
+# CHECK4-NEXT: [[#]]: {{0*}}[[#ADDR+1]]   0  NOTYPE GLOBAL DEFAULT   3 FOO_DATA_END
+
+## OVERWRITE_SECTIONS describes a section used as an anchor of an insert command.
+## Place .foo.data as an orphan than insert .foo.text after .foo.data.
+# RUN: ld.lld %t/insert.lds %t/overwrite3.lds %t/a.o -o %t5
+# RUN: llvm-readelf -S -s %t4 | FileCheck %s --check-prefix=CHECK5
+
+# CHECK5:      [Nr] Name      Type     Address  Off      Size   ES Flg Lk Inf Al
+# CHECK5-NEXT: [ 0]           NULL     [[#%x,]] [[#%x,]] 000000
+# CHECK5-NEXT: [ 1] .text     PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  4
+# CHECK5-NEXT: [ 2] .data     PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK5-NEXT: [ 3] .foo.data PROGBITS [[#%x,]] [[#%x,]] 000001 00  WA  0   0  1
+# CHECK5-NEXT: [ 4] .foo.text PROGBITS [[#%x,]] [[#%x,]] 000001 00  AX  0   0  1
+# CHECK5-NEXT: [ 5] .comment  PROGBITS {{.*}}
+
+
+#--- a.s
+.globl _start
+_start:
+
+.section .foo.text,"ax"; .byte 1
+.section .foo.data,"aw"; .byte 2
+.section .text.1,"ax"; .byte 3
+.section .data.1,"aw"; .byte 4
+
+#--- main.lds
+SECTIONS {
+  .data : { *(.data*) }
+  .foo.data : { *(.foo.data) }
+  .text : { *(.text*) }
+  .foo.text : ALIGN(16) { *(.foo.text) }
+}
+
+#--- overwrite1.lds
+OVERWRITE_SECTIONS {
+  .foo.text : ALIGN(8) { *(.foo.data .foo.text) }
+}
+
+#--- overwrite2.lds
+OVERWRITE_SECTIONS {
+  .foo.data : { FOO_DATA_START = .; *(.foo.data) FOO_DATA_END = .;  }
+  .foo.text : { FOO_TEXT_START = .; *(.foo.text) FOO_TEXT_END = .;  }
+}
+
+#--- overwrite3.lds
+OVERWRITE_SECTIONS {
+  .foo.data : {}
+}
+
+#--- insert.lds
+SECTIONS { .foo.text : {} } INSERT AFTER .foo.data;
+
+#--- err1.lds
+## TODO Fix the diagnostic 'malformed number'
+# RUN: not ld.lld %t/err1.lds %t/a.o -o /dev/null 2>&1 | FileCheck %s -DFILE=%t/err1.lds --check-prefix=ERR1
+# ERR1: error: [[FILE]]:[[#@LINE+2]]: malformed number: =
+OVERWRITE_SECTIONS {
+  A = 1;
+}