llvm-project/llvm/tools/llvm-size/llvm-size.cpp

762 lines
29 KiB
C++
Raw Normal View History

//===-- llvm-size.cpp - Print the size of each object section ---*- C++ -*-===//
2011-09-29 04:57:46 +08:00
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This program is a utility that works like traditional Unix "size",
// that is, it prints out the size of each section, and the total size of all
// sections.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/APInt.h"
#include "llvm/Object/Archive.h"
#include "llvm/Object/ELFObjectFile.h"
#include "llvm/Object/MachO.h"
#include "llvm/Object/MachOUniversal.h"
#include "llvm/Object/ObjectFile.h"
2011-09-29 04:57:46 +08:00
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/FileSystem.h"
#include "llvm/Support/Format.h"
#include "llvm/Support/ManagedStatic.h"
#include "llvm/Support/MemoryBuffer.h"
#include "llvm/Support/PrettyStackTrace.h"
#include "llvm/Support/Signals.h"
#include "llvm/Support/raw_ostream.h"
2011-09-29 04:57:46 +08:00
#include <algorithm>
#include <string>
#include <system_error>
2011-09-29 04:57:46 +08:00
using namespace llvm;
using namespace object;
enum OutputFormatTy { berkeley, sysv, darwin };
static cl::opt<OutputFormatTy>
OutputFormat("format", cl::desc("Specify output format"),
cl::values(clEnumVal(sysv, "System V format"),
clEnumVal(berkeley, "Berkeley format"),
clEnumVal(darwin, "Darwin -m format"), clEnumValEnd),
cl::init(berkeley));
static cl::opt<OutputFormatTy> OutputFormatShort(
cl::desc("Specify output format"),
cl::values(clEnumValN(sysv, "A", "System V format"),
clEnumValN(berkeley, "B", "Berkeley format"),
clEnumValN(darwin, "m", "Darwin -m format"), clEnumValEnd),
cl::init(berkeley));
static bool BerkeleyHeaderPrinted = false;
static bool MoreThanOneFile = false;
cl::opt<bool>
DarwinLongFormat("l", cl::desc("When format is darwin, use long format "
"to include addresses and offsets."));
cl::opt<bool>
ELFCommons("common",
cl::desc("Print common symbols in the ELF file. When using "
"Berkely format, this is added to bss."),
cl::init(false));
static cl::list<std::string>
ArchFlags("arch", cl::desc("architecture(s) from a Mach-O file to dump"),
cl::ZeroOrMore);
bool ArchAll = false;
enum RadixTy { octal = 8, decimal = 10, hexadecimal = 16 };
static cl::opt<unsigned int>
Radix("-radix", cl::desc("Print size in radix. Only 8, 10, and 16 are valid"),
cl::init(decimal));
static cl::opt<RadixTy>
RadixShort(cl::desc("Print size in radix:"),
cl::values(clEnumValN(octal, "o", "Print size in octal"),
clEnumValN(decimal, "d", "Print size in decimal"),
clEnumValN(hexadecimal, "x", "Print size in hexadecimal"),
clEnumValEnd),
cl::init(decimal));
static cl::list<std::string>
InputFilenames(cl::Positional, cl::desc("<input files>"), cl::ZeroOrMore);
static std::string ToolName;
2016-02-10 05:32:56 +08:00
/// If ec is not success, print the error and return true.
static bool error(std::error_code ec) {
if (!ec)
return false;
2011-09-29 04:57:46 +08:00
errs() << ToolName << ": error reading file: " << ec.message() << ".\n";
errs().flush();
2011-09-29 04:57:46 +08:00
return true;
}
2016-02-10 05:32:56 +08:00
/// Get the length of the string that represents @p num in Radix including the
/// leading 0x or 0 for hexadecimal and octal respectively.
static size_t getNumLengthAsString(uint64_t num) {
2011-09-29 04:57:46 +08:00
APInt conv(64, num);
SmallString<32> result;
conv.toString(result, Radix, false, true);
2011-09-29 04:57:46 +08:00
return result.size();
}
2016-02-10 05:32:56 +08:00
/// Return the printing format for the Radix.
static const char *getRadixFmt() {
switch (Radix) {
case octal:
return PRIo64;
case decimal:
return PRIu64;
case hexadecimal:
return PRIx64;
}
return nullptr;
}
/// Remove unneeded ELF sections from calculation
static bool considerForSize(ObjectFile *Obj, SectionRef Section) {
if (!Obj->isELF())
return true;
switch (static_cast<ELFSectionRef>(Section).getType()) {
case ELF::SHT_NULL:
case ELF::SHT_SYMTAB:
case ELF::SHT_STRTAB:
case ELF::SHT_REL:
case ELF::SHT_RELA:
return false;
}
return true;
}
/// Total size of all ELF common symbols
static uint64_t getCommonSize(ObjectFile *Obj) {
uint64_t TotalCommons = 0;
for (auto &Sym : Obj->symbols())
if (Obj->getSymbolFlags(Sym.getRawDataRefImpl()) & SymbolRef::SF_Common)
TotalCommons += Obj->getCommonSymbolSize(Sym.getRawDataRefImpl());
return TotalCommons;
}
2016-02-10 05:32:56 +08:00
/// Print the size of each Mach-O segment and section in @p MachO.
///
/// This is when used when @c OutputFormat is darwin and produces the same
/// output as darwin's size(1) -m output.
static void printDarwinSectionSizes(MachOObjectFile *MachO) {
std::string fmtbuf;
raw_string_ostream fmt(fmtbuf);
const char *radix_fmt = getRadixFmt();
if (Radix == hexadecimal)
fmt << "0x";
fmt << "%" << radix_fmt;
uint32_t Filetype = MachO->getHeader().filetype;
uint64_t total = 0;
for (const auto &Load : MachO->load_commands()) {
if (Load.C.cmd == MachO::LC_SEGMENT_64) {
MachO::segment_command_64 Seg = MachO->getSegment64LoadCommand(Load);
outs() << "Segment " << Seg.segname << ": "
<< format(fmt.str().c_str(), Seg.vmsize);
if (DarwinLongFormat)
outs() << " (vmaddr 0x" << format("%" PRIx64, Seg.vmaddr) << " fileoff "
<< Seg.fileoff << ")";
outs() << "\n";
total += Seg.vmsize;
uint64_t sec_total = 0;
for (unsigned J = 0; J < Seg.nsects; ++J) {
MachO::section_64 Sec = MachO->getSection64(Load, J);
if (Filetype == MachO::MH_OBJECT)
outs() << "\tSection (" << format("%.16s", &Sec.segname) << ", "
<< format("%.16s", &Sec.sectname) << "): ";
else
outs() << "\tSection " << format("%.16s", &Sec.sectname) << ": ";
outs() << format(fmt.str().c_str(), Sec.size);
if (DarwinLongFormat)
outs() << " (addr 0x" << format("%" PRIx64, Sec.addr) << " offset "
<< Sec.offset << ")";
outs() << "\n";
sec_total += Sec.size;
}
if (Seg.nsects != 0)
outs() << "\ttotal " << format(fmt.str().c_str(), sec_total) << "\n";
} else if (Load.C.cmd == MachO::LC_SEGMENT) {
MachO::segment_command Seg = MachO->getSegmentLoadCommand(Load);
uint64_t Seg_vmsize = Seg.vmsize;
outs() << "Segment " << Seg.segname << ": "
<< format(fmt.str().c_str(), Seg_vmsize);
if (DarwinLongFormat)
outs() << " (vmaddr 0x" << format("%" PRIx32, Seg.vmaddr) << " fileoff "
<< Seg.fileoff << ")";
outs() << "\n";
total += Seg.vmsize;
uint64_t sec_total = 0;
for (unsigned J = 0; J < Seg.nsects; ++J) {
MachO::section Sec = MachO->getSection(Load, J);
if (Filetype == MachO::MH_OBJECT)
outs() << "\tSection (" << format("%.16s", &Sec.segname) << ", "
<< format("%.16s", &Sec.sectname) << "): ";
else
outs() << "\tSection " << format("%.16s", &Sec.sectname) << ": ";
uint64_t Sec_size = Sec.size;
outs() << format(fmt.str().c_str(), Sec_size);
if (DarwinLongFormat)
outs() << " (addr 0x" << format("%" PRIx32, Sec.addr) << " offset "
<< Sec.offset << ")";
outs() << "\n";
sec_total += Sec.size;
}
if (Seg.nsects != 0)
outs() << "\ttotal " << format(fmt.str().c_str(), sec_total) << "\n";
}
}
outs() << "total " << format(fmt.str().c_str(), total) << "\n";
}
2016-02-10 05:32:56 +08:00
/// Print the summary sizes of the standard Mach-O segments in @p MachO.
///
/// This is when used when @c OutputFormat is berkeley with a Mach-O file and
/// produces the same output as darwin's size(1) default output.
static void printDarwinSegmentSizes(MachOObjectFile *MachO) {
uint64_t total_text = 0;
uint64_t total_data = 0;
uint64_t total_objc = 0;
uint64_t total_others = 0;
for (const auto &Load : MachO->load_commands()) {
if (Load.C.cmd == MachO::LC_SEGMENT_64) {
MachO::segment_command_64 Seg = MachO->getSegment64LoadCommand(Load);
if (MachO->getHeader().filetype == MachO::MH_OBJECT) {
for (unsigned J = 0; J < Seg.nsects; ++J) {
MachO::section_64 Sec = MachO->getSection64(Load, J);
StringRef SegmentName = StringRef(Sec.segname);
if (SegmentName == "__TEXT")
total_text += Sec.size;
else if (SegmentName == "__DATA")
total_data += Sec.size;
else if (SegmentName == "__OBJC")
total_objc += Sec.size;
else
total_others += Sec.size;
}
} else {
StringRef SegmentName = StringRef(Seg.segname);
if (SegmentName == "__TEXT")
total_text += Seg.vmsize;
else if (SegmentName == "__DATA")
total_data += Seg.vmsize;
else if (SegmentName == "__OBJC")
total_objc += Seg.vmsize;
else
total_others += Seg.vmsize;
}
} else if (Load.C.cmd == MachO::LC_SEGMENT) {
MachO::segment_command Seg = MachO->getSegmentLoadCommand(Load);
if (MachO->getHeader().filetype == MachO::MH_OBJECT) {
for (unsigned J = 0; J < Seg.nsects; ++J) {
MachO::section Sec = MachO->getSection(Load, J);
StringRef SegmentName = StringRef(Sec.segname);
if (SegmentName == "__TEXT")
total_text += Sec.size;
else if (SegmentName == "__DATA")
total_data += Sec.size;
else if (SegmentName == "__OBJC")
total_objc += Sec.size;
else
total_others += Sec.size;
}
} else {
StringRef SegmentName = StringRef(Seg.segname);
if (SegmentName == "__TEXT")
total_text += Seg.vmsize;
else if (SegmentName == "__DATA")
total_data += Seg.vmsize;
else if (SegmentName == "__OBJC")
total_objc += Seg.vmsize;
else
total_others += Seg.vmsize;
}
}
}
uint64_t total = total_text + total_data + total_objc + total_others;
if (!BerkeleyHeaderPrinted) {
outs() << "__TEXT\t__DATA\t__OBJC\tothers\tdec\thex\n";
BerkeleyHeaderPrinted = true;
}
outs() << total_text << "\t" << total_data << "\t" << total_objc << "\t"
<< total_others << "\t" << total << "\t" << format("%" PRIx64, total)
<< "\t";
}
2016-02-10 05:32:56 +08:00
/// Print the size of each section in @p Obj.
///
/// The format used is determined by @c OutputFormat and @c Radix.
static void printObjectSectionSizes(ObjectFile *Obj) {
2011-09-29 04:57:46 +08:00
uint64_t total = 0;
std::string fmtbuf;
raw_string_ostream fmt(fmtbuf);
const char *radix_fmt = getRadixFmt();
// If OutputFormat is darwin and we have a MachOObjectFile print as darwin's
// size(1) -m output, else if OutputFormat is darwin and not a Mach-O object
// let it fall through to OutputFormat berkeley.
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(Obj);
if (OutputFormat == darwin && MachO)
printDarwinSectionSizes(MachO);
// If we have a MachOObjectFile and the OutputFormat is berkeley print as
// darwin's default berkeley format for Mach-O files.
else if (MachO && OutputFormat == berkeley)
printDarwinSegmentSizes(MachO);
else if (OutputFormat == sysv) {
2011-09-29 04:57:46 +08:00
// Run two passes over all sections. The first gets the lengths needed for
// formatting the output. The second actually does the output.
std::size_t max_name_len = strlen("section");
std::size_t max_size_len = strlen("size");
std::size_t max_addr_len = strlen("addr");
for (const SectionRef &Section : Obj->sections()) {
if (!considerForSize(Obj, Section))
continue;
uint64_t size = Section.getSize();
2011-09-29 04:57:46 +08:00
total += size;
StringRef name;
if (error(Section.getName(name)))
return;
uint64_t addr = Section.getAddress();
2011-09-29 04:57:46 +08:00
max_name_len = std::max(max_name_len, name.size());
max_size_len = std::max(max_size_len, getNumLengthAsString(size));
max_addr_len = std::max(max_addr_len, getNumLengthAsString(addr));
2011-09-29 04:57:46 +08:00
}
// Add extra padding.
2011-09-29 04:57:46 +08:00
max_name_len += 2;
max_size_len += 2;
max_addr_len += 2;
// Setup header format.
2011-09-29 04:57:46 +08:00
fmt << "%-" << max_name_len << "s "
<< "%" << max_size_len << "s "
<< "%" << max_addr_len << "s\n";
// Print header
outs() << format(fmt.str().c_str(), static_cast<const char *>("section"),
static_cast<const char *>("size"),
static_cast<const char *>("addr"));
2011-09-29 04:57:46 +08:00
fmtbuf.clear();
// Setup per section format.
fmt << "%-" << max_name_len << "s "
<< "%#" << max_size_len << radix_fmt << " "
<< "%#" << max_addr_len << radix_fmt << "\n";
// Print each section.
for (const SectionRef &Section : Obj->sections()) {
if (!considerForSize(Obj, Section))
continue;
2011-09-29 04:57:46 +08:00
StringRef name;
if (error(Section.getName(name)))
return;
uint64_t size = Section.getSize();
uint64_t addr = Section.getAddress();
2011-09-29 04:57:46 +08:00
std::string namestr = name;
outs() << format(fmt.str().c_str(), namestr.c_str(), size, addr);
2011-09-29 04:57:46 +08:00
}
if (ELFCommons) {
uint64_t CommonSize = getCommonSize(Obj);
total += CommonSize;
outs() << format(fmt.str().c_str(), std::string("*COM*").c_str(),
CommonSize, static_cast<uint64_t>(0));
}
2011-09-29 04:57:46 +08:00
// Print total.
fmtbuf.clear();
fmt << "%-" << max_name_len << "s "
<< "%#" << max_size_len << radix_fmt << "\n";
outs() << format(fmt.str().c_str(), static_cast<const char *>("Total"),
2011-09-29 04:57:46 +08:00
total);
} else {
// The Berkeley format does not display individual section sizes. It
// displays the cumulative size for each section type.
2011-09-29 04:57:46 +08:00
uint64_t total_text = 0;
uint64_t total_data = 0;
uint64_t total_bss = 0;
// Make one pass over the section table to calculate sizes.
for (const SectionRef &Section : Obj->sections()) {
uint64_t size = Section.getSize();
bool isText = Section.isText();
bool isData = Section.isData();
bool isBSS = Section.isBSS();
2011-09-29 04:57:46 +08:00
if (isText)
total_text += size;
else if (isData)
total_data += size;
else if (isBSS)
total_bss += size;
}
if (ELFCommons)
total_bss += getCommonSize(Obj);
2011-09-29 04:57:46 +08:00
total = total_text + total_data + total_bss;
if (!BerkeleyHeaderPrinted) {
outs() << " text data bss "
<< (Radix == octal ? "oct" : "dec") << " hex filename\n";
BerkeleyHeaderPrinted = true;
}
2011-09-29 04:57:46 +08:00
// Print result.
fmt << "%#7" << radix_fmt << " "
<< "%#7" << radix_fmt << " "
<< "%#7" << radix_fmt << " ";
outs() << format(fmt.str().c_str(), total_text, total_data, total_bss);
2011-09-29 04:57:46 +08:00
fmtbuf.clear();
fmt << "%7" << (Radix == octal ? PRIo64 : PRIu64) << " "
<< "%7" PRIx64 " ";
outs() << format(fmt.str().c_str(), total, total);
2011-09-29 04:57:46 +08:00
}
}
2016-02-10 05:32:56 +08:00
/// Checks to see if the @p o ObjectFile is a Mach-O file and if it is and there
/// is a list of architecture flags specified then check to make sure this
/// Mach-O file is one of those architectures or all architectures was
/// specificed. If not then an error is generated and this routine returns
/// false. Else it returns true.
static bool checkMachOAndArchFlags(ObjectFile *o, StringRef file) {
if (isa<MachOObjectFile>(o) && !ArchAll && ArchFlags.size() != 0) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
bool ArchFound = false;
MachO::mach_header H;
MachO::mach_header_64 H_64;
Triple T;
if (MachO->is64Bit()) {
H_64 = MachO->MachOObjectFile::getHeader64();
T = MachOObjectFile::getArchTriple(H_64.cputype, H_64.cpusubtype);
} else {
H = MachO->MachOObjectFile::getHeader();
T = MachOObjectFile::getArchTriple(H.cputype, H.cpusubtype);
}
unsigned i;
for (i = 0; i < ArchFlags.size(); ++i) {
if (ArchFlags[i] == T.getArchName())
ArchFound = true;
break;
}
if (!ArchFound) {
errs() << ToolName << ": file: " << file
<< " does not contain architecture: " << ArchFlags[i] << ".\n";
return false;
}
}
return true;
}
2016-02-10 05:32:56 +08:00
/// Print the section sizes for @p file. If @p file is an archive, print the
/// section sizes for each archive member.
static void printFileSectionSizes(StringRef file) {
2011-09-29 04:57:46 +08:00
// Attempt to open the binary.
Thread Expected<...> up from createMachOObjectFile() to allow llvm-objdump to produce a real error message Produce the first specific error message for a malformed Mach-O file describing the problem instead of the generic message for object_error::parse_failed of "Invalid data was encountered while parsing the file”.  Many more good error messages will follow after this first one. This is built on Lang Hames’ great work of adding the ’Error' class for structured error handling and threading Error through MachOObjectFile construction. And making createMachOObjectFile return Expected<...> . So to to get the error to the llvm-obdump tool, I changed the stack of these methods to also return Expected<...> : object::ObjectFile::createObjectFile() object::SymbolicFile::createSymbolicFile() object::createBinary() Then finally in ParseInputMachO() in MachODump.cpp the error can be reported and the specific error message can be printed in llvm-objdump and can be seen in the existing test case for the existing malformed binary but with the updated error message. Converting these interfaces to Expected<> from ErrorOr<> does involve touching a number of places. To contain the changes for now use of errorToErrorCode() and errorOrToExpected() are used where the callers are yet to be converted. Also there some were bugs in the existing code that did not deal with the old ErrorOr<> return values. So now with Expected<> since they must be checked and the error handled, I added a TODO and a comment: “// TODO: Actually report errors helpfully” and a call something like consumeError(ObjOrErr.takeError()) so the buggy code will not crash since needed to deal with the Error. Note there is one fix also needed to lld/COFF/InputFiles.cpp that goes along with this that I will commit right after this. So expect lld not to built after this commit and before the next one. llvm-svn: 265606
2016-04-07 06:14:09 +08:00
Expected<OwningBinary<Binary>> BinaryOrErr = createBinary(file);
if (!BinaryOrErr) {
error(errorToErrorCode(BinaryOrErr.takeError()));
2011-09-29 04:57:46 +08:00
return;
Thread Expected<...> up from createMachOObjectFile() to allow llvm-objdump to produce a real error message Produce the first specific error message for a malformed Mach-O file describing the problem instead of the generic message for object_error::parse_failed of "Invalid data was encountered while parsing the file”.  Many more good error messages will follow after this first one. This is built on Lang Hames’ great work of adding the ’Error' class for structured error handling and threading Error through MachOObjectFile construction. And making createMachOObjectFile return Expected<...> . So to to get the error to the llvm-obdump tool, I changed the stack of these methods to also return Expected<...> : object::ObjectFile::createObjectFile() object::SymbolicFile::createSymbolicFile() object::createBinary() Then finally in ParseInputMachO() in MachODump.cpp the error can be reported and the specific error message can be printed in llvm-objdump and can be seen in the existing test case for the existing malformed binary but with the updated error message. Converting these interfaces to Expected<> from ErrorOr<> does involve touching a number of places. To contain the changes for now use of errorToErrorCode() and errorOrToExpected() are used where the callers are yet to be converted. Also there some were bugs in the existing code that did not deal with the old ErrorOr<> return values. So now with Expected<> since they must be checked and the error handled, I added a TODO and a comment: “// TODO: Actually report errors helpfully” and a call something like consumeError(ObjOrErr.takeError()) so the buggy code will not crash since needed to deal with the Error. Note there is one fix also needed to lld/COFF/InputFiles.cpp that goes along with this that I will commit right after this. So expect lld not to built after this commit and before the next one. llvm-svn: 265606
2016-04-07 06:14:09 +08:00
}
Binary &Bin = *BinaryOrErr.get().getBinary();
2011-09-29 04:57:46 +08:00
if (Archive *a = dyn_cast<Archive>(&Bin)) {
// This is an archive. Iterate over each member and display its sizes.
for (object::Archive::child_iterator i = a->child_begin(),
e = a->child_end();
i != e; ++i) {
if (error(i->getError()))
Reapply r250906 with many suggested updates from Rafael Espindola. The needed lld matching changes to be submitted immediately next, but this revision will cause lld failures with this alone which is expected. This removes the eating of the error in Archive::Child::getSize() when the characters in the size field in the archive header for the member is not a number. To do this we have all of the needed methods return ErrorOr to push them up until we get out of lib. Then the tools and can handle the error in whatever way is appropriate for that tool. So the solution is to plumb all the ErrorOr stuff through everything that touches archives. This include its iterators as one can create an Archive object but the first or any other Child object may fail to be created due to a bad size field in its header. Thanks to Lang Hames on the changes making child_iterator contain an ErrorOr<Child> instead of a Child and the needed changes to ErrorOr.h to add operator overloading for * and -> . We don’t want to use llvm_unreachable() as it calls abort() and is produces a “crash” and using report_fatal_error() to move the error checking will cause the program to stop, neither of which are really correct in library code. There are still some uses of these that should be cleaned up in this library code for other than the size field. The test cases use archives with text files so one can see the non-digit character, in this case a ‘%’, in the size field. These changes will require corresponding changes to the lld project. That will be committed immediately after this change. But this revision will cause lld failures with this alone which is expected. llvm-svn: 252192
2015-11-06 03:24:56 +08:00
exit(1);
ErrorOr<std::unique_ptr<Binary>> ChildOrErr = i->get().getAsBinary();
if (error(ChildOrErr.getError()))
2011-09-29 04:57:46 +08:00
continue;
if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (!checkMachOAndArchFlags(o, file))
return;
2011-09-29 04:57:46 +08:00
if (OutputFormat == sysv)
outs() << o->getFileName() << " (ex " << a->getFileName() << "):\n";
else if (MachO && OutputFormat == darwin)
outs() << a->getFileName() << "(" << o->getFileName() << "):\n";
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (MachO)
outs() << a->getFileName() << "(" << o->getFileName() << ")\n";
else
outs() << o->getFileName() << " (ex " << a->getFileName() << ")\n";
}
2011-09-29 04:57:46 +08:00
}
}
} else if (MachOUniversalBinary *UB =
dyn_cast<MachOUniversalBinary>(&Bin)) {
// If we have a list of architecture flags specified dump only those.
if (!ArchAll && ArchFlags.size() != 0) {
// Look for a slice in the universal binary that matches each ArchFlag.
bool ArchFound;
for (unsigned i = 0; i < ArchFlags.size(); ++i) {
ArchFound = false;
for (MachOUniversalBinary::object_iterator I = UB->begin_objects(),
E = UB->end_objects();
I != E; ++I) {
if (ArchFlags[i] == I->getArchTypeName()) {
ArchFound = true;
ErrorOr<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile();
if (UO) {
if (ObjectFile *o = dyn_cast<ObjectFile>(&*UO.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (OutputFormat == sysv)
outs() << o->getFileName() << " :\n";
else if (MachO && OutputFormat == darwin) {
if (MoreThanOneFile || ArchFlags.size() > 1)
outs() << o->getFileName() << " (for architecture "
<< I->getArchTypeName() << "): \n";
}
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (!MachO || MoreThanOneFile || ArchFlags.size() > 1)
outs() << o->getFileName() << " (for architecture "
<< I->getArchTypeName() << ")";
outs() << "\n";
}
}
} else if (ErrorOr<std::unique_ptr<Archive>> AOrErr =
I->getAsArchive()) {
std::unique_ptr<Archive> &UA = *AOrErr;
// This is an archive. Iterate over each member and display its
// sizes.
for (object::Archive::child_iterator i = UA->child_begin(),
e = UA->child_end();
i != e; ++i) {
if (error(i->getError()))
Reapply r250906 with many suggested updates from Rafael Espindola. The needed lld matching changes to be submitted immediately next, but this revision will cause lld failures with this alone which is expected. This removes the eating of the error in Archive::Child::getSize() when the characters in the size field in the archive header for the member is not a number. To do this we have all of the needed methods return ErrorOr to push them up until we get out of lib. Then the tools and can handle the error in whatever way is appropriate for that tool. So the solution is to plumb all the ErrorOr stuff through everything that touches archives. This include its iterators as one can create an Archive object but the first or any other Child object may fail to be created due to a bad size field in its header. Thanks to Lang Hames on the changes making child_iterator contain an ErrorOr<Child> instead of a Child and the needed changes to ErrorOr.h to add operator overloading for * and -> . We don’t want to use llvm_unreachable() as it calls abort() and is produces a “crash” and using report_fatal_error() to move the error checking will cause the program to stop, neither of which are really correct in library code. There are still some uses of these that should be cleaned up in this library code for other than the size field. The test cases use archives with text files so one can see the non-digit character, in this case a ‘%’, in the size field. These changes will require corresponding changes to the lld project. That will be committed immediately after this change. But this revision will cause lld failures with this alone which is expected. llvm-svn: 252192
2015-11-06 03:24:56 +08:00
exit(1);
ErrorOr<std::unique_ptr<Binary>> ChildOrErr = i->get().getAsBinary();
if (error(ChildOrErr.getError()))
continue;
if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (OutputFormat == sysv)
outs() << o->getFileName() << " (ex " << UA->getFileName()
<< "):\n";
else if (MachO && OutputFormat == darwin)
outs() << UA->getFileName() << "(" << o->getFileName()
<< ")"
<< " (for architecture " << I->getArchTypeName()
<< "):\n";
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (MachO) {
outs() << UA->getFileName() << "(" << o->getFileName()
<< ")";
if (ArchFlags.size() > 1)
outs() << " (for architecture " << I->getArchTypeName()
<< ")";
outs() << "\n";
} else
outs() << o->getFileName() << " (ex " << UA->getFileName()
<< ")\n";
}
}
}
}
}
}
if (!ArchFound) {
errs() << ToolName << ": file: " << file
<< " does not contain architecture" << ArchFlags[i] << ".\n";
return;
}
}
return;
}
// No architecture flags were specified so if this contains a slice that
// matches the host architecture dump only that.
if (!ArchAll) {
StringRef HostArchName = MachOObjectFile::getHostArch().getArchName();
for (MachOUniversalBinary::object_iterator I = UB->begin_objects(),
E = UB->end_objects();
I != E; ++I) {
if (HostArchName == I->getArchTypeName()) {
ErrorOr<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile();
if (UO) {
if (ObjectFile *o = dyn_cast<ObjectFile>(&*UO.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (OutputFormat == sysv)
outs() << o->getFileName() << " :\n";
else if (MachO && OutputFormat == darwin) {
if (MoreThanOneFile)
outs() << o->getFileName() << " (for architecture "
<< I->getArchTypeName() << "):\n";
}
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (!MachO || MoreThanOneFile)
outs() << o->getFileName() << " (for architecture "
<< I->getArchTypeName() << ")";
outs() << "\n";
}
}
} else if (ErrorOr<std::unique_ptr<Archive>> AOrErr =
I->getAsArchive()) {
std::unique_ptr<Archive> &UA = *AOrErr;
// This is an archive. Iterate over each member and display its
// sizes.
for (object::Archive::child_iterator i = UA->child_begin(),
e = UA->child_end();
i != e; ++i) {
if (error(i->getError()))
Reapply r250906 with many suggested updates from Rafael Espindola. The needed lld matching changes to be submitted immediately next, but this revision will cause lld failures with this alone which is expected. This removes the eating of the error in Archive::Child::getSize() when the characters in the size field in the archive header for the member is not a number. To do this we have all of the needed methods return ErrorOr to push them up until we get out of lib. Then the tools and can handle the error in whatever way is appropriate for that tool. So the solution is to plumb all the ErrorOr stuff through everything that touches archives. This include its iterators as one can create an Archive object but the first or any other Child object may fail to be created due to a bad size field in its header. Thanks to Lang Hames on the changes making child_iterator contain an ErrorOr<Child> instead of a Child and the needed changes to ErrorOr.h to add operator overloading for * and -> . We don’t want to use llvm_unreachable() as it calls abort() and is produces a “crash” and using report_fatal_error() to move the error checking will cause the program to stop, neither of which are really correct in library code. There are still some uses of these that should be cleaned up in this library code for other than the size field. The test cases use archives with text files so one can see the non-digit character, in this case a ‘%’, in the size field. These changes will require corresponding changes to the lld project. That will be committed immediately after this change. But this revision will cause lld failures with this alone which is expected. llvm-svn: 252192
2015-11-06 03:24:56 +08:00
exit(1);
ErrorOr<std::unique_ptr<Binary>> ChildOrErr = i->get().getAsBinary();
if (error(ChildOrErr.getError()))
continue;
if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (OutputFormat == sysv)
outs() << o->getFileName() << " (ex " << UA->getFileName()
<< "):\n";
else if (MachO && OutputFormat == darwin)
outs() << UA->getFileName() << "(" << o->getFileName() << ")"
<< " (for architecture " << I->getArchTypeName()
<< "):\n";
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (MachO)
outs() << UA->getFileName() << "(" << o->getFileName()
<< ")\n";
else
outs() << o->getFileName() << " (ex " << UA->getFileName()
<< ")\n";
}
}
}
}
return;
}
}
}
// Either all architectures have been specified or none have been specified
// and this does not contain the host architecture so dump all the slices.
bool MoreThanOneArch = UB->getNumberOfObjects() > 1;
for (MachOUniversalBinary::object_iterator I = UB->begin_objects(),
E = UB->end_objects();
I != E; ++I) {
ErrorOr<std::unique_ptr<ObjectFile>> UO = I->getAsObjectFile();
if (UO) {
if (ObjectFile *o = dyn_cast<ObjectFile>(&*UO.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (OutputFormat == sysv)
outs() << o->getFileName() << " :\n";
else if (MachO && OutputFormat == darwin) {
if (MoreThanOneFile || MoreThanOneArch)
outs() << o->getFileName() << " (for architecture "
<< I->getArchTypeName() << "):";
outs() << "\n";
}
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (!MachO || MoreThanOneFile || MoreThanOneArch)
outs() << o->getFileName() << " (for architecture "
<< I->getArchTypeName() << ")";
outs() << "\n";
}
}
} else if (ErrorOr<std::unique_ptr<Archive>> AOrErr =
I->getAsArchive()) {
std::unique_ptr<Archive> &UA = *AOrErr;
// This is an archive. Iterate over each member and display its sizes.
for (object::Archive::child_iterator i = UA->child_begin(),
e = UA->child_end();
i != e; ++i) {
if (error(i->getError()))
Reapply r250906 with many suggested updates from Rafael Espindola. The needed lld matching changes to be submitted immediately next, but this revision will cause lld failures with this alone which is expected. This removes the eating of the error in Archive::Child::getSize() when the characters in the size field in the archive header for the member is not a number. To do this we have all of the needed methods return ErrorOr to push them up until we get out of lib. Then the tools and can handle the error in whatever way is appropriate for that tool. So the solution is to plumb all the ErrorOr stuff through everything that touches archives. This include its iterators as one can create an Archive object but the first or any other Child object may fail to be created due to a bad size field in its header. Thanks to Lang Hames on the changes making child_iterator contain an ErrorOr<Child> instead of a Child and the needed changes to ErrorOr.h to add operator overloading for * and -> . We don’t want to use llvm_unreachable() as it calls abort() and is produces a “crash” and using report_fatal_error() to move the error checking will cause the program to stop, neither of which are really correct in library code. There are still some uses of these that should be cleaned up in this library code for other than the size field. The test cases use archives with text files so one can see the non-digit character, in this case a ‘%’, in the size field. These changes will require corresponding changes to the lld project. That will be committed immediately after this change. But this revision will cause lld failures with this alone which is expected. llvm-svn: 252192
2015-11-06 03:24:56 +08:00
exit(1);
ErrorOr<std::unique_ptr<Binary>> ChildOrErr = i->get().getAsBinary();
if (error(ChildOrErr.getError()))
continue;
if (ObjectFile *o = dyn_cast<ObjectFile>(&*ChildOrErr.get())) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (OutputFormat == sysv)
outs() << o->getFileName() << " (ex " << UA->getFileName()
<< "):\n";
else if (MachO && OutputFormat == darwin)
outs() << UA->getFileName() << "(" << o->getFileName() << ")"
<< " (for architecture " << I->getArchTypeName() << "):\n";
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
if (MachO)
outs() << UA->getFileName() << "(" << o->getFileName() << ")"
<< " (for architecture " << I->getArchTypeName()
<< ")\n";
else
outs() << o->getFileName() << " (ex " << UA->getFileName()
<< ")\n";
}
}
}
}
}
} else if (ObjectFile *o = dyn_cast<ObjectFile>(&Bin)) {
if (!checkMachOAndArchFlags(o, file))
return;
2011-09-29 04:57:46 +08:00
if (OutputFormat == sysv)
outs() << o->getFileName() << " :\n";
printObjectSectionSizes(o);
if (OutputFormat == berkeley) {
MachOObjectFile *MachO = dyn_cast<MachOObjectFile>(o);
if (!MachO || MoreThanOneFile)
outs() << o->getFileName();
outs() << "\n";
}
2011-09-29 04:57:46 +08:00
} else {
errs() << ToolName << ": " << file << ": "
<< "Unrecognized file type.\n";
2011-09-29 04:57:46 +08:00
}
// System V adds an extra newline at the end of each file.
2011-09-29 04:57:46 +08:00
if (OutputFormat == sysv)
outs() << "\n";
}
int main(int argc, char **argv) {
// Print a stack trace if we signal out.
sys::PrintStackTraceOnErrorSignal();
PrettyStackTraceProgram X(argc, argv);
llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
2011-09-29 04:57:46 +08:00
cl::ParseCommandLineOptions(argc, argv, "llvm object size dumper\n");
ToolName = argv[0];
if (OutputFormatShort.getNumOccurrences())
OutputFormat = static_cast<OutputFormatTy>(OutputFormatShort);
2011-09-29 04:57:46 +08:00
if (RadixShort.getNumOccurrences())
Radix = RadixShort;
2011-09-29 04:57:46 +08:00
for (unsigned i = 0; i < ArchFlags.size(); ++i) {
if (ArchFlags[i] == "all") {
ArchAll = true;
} else {
if (!MachOObjectFile::isValidArch(ArchFlags[i])) {
outs() << ToolName << ": for the -arch option: Unknown architecture "
<< "named '" << ArchFlags[i] << "'";
return 1;
}
}
}
2011-09-29 04:57:46 +08:00
if (InputFilenames.size() == 0)
InputFilenames.push_back("a.out");
MoreThanOneFile = InputFilenames.size() > 1;
2011-09-29 04:57:46 +08:00
std::for_each(InputFilenames.begin(), InputFilenames.end(),
printFileSectionSizes);
2011-09-29 04:57:46 +08:00
return 0;
}