[llvm-objcopy] Add elf32-sparc and elf32-sparcel target

[seiya]

Summary

The "sparc"/"sparcel" architectures appears in ArchMap (used by -B option) but not in OutputFormatMap (used by -I/-O option). Add their targets into OutputFormatMap for consistency.

Note that AFAIK there're no targets for 32-bit little-endian SPARC ("elf32-sparcel") in GNU binutils.

[jhenderson]

Looks good, I think, but does GNU objcopy have sparc as -I/-O type? Is the syntax the same as you have used?

[seiya]

In D63238#1541201, @jhenderson wrote:

Looks good, I think, but does GNU objcopy have sparc as -I/-O type? Is the syntax the same as you have used?

Yes, it has the same target name. I'm not familiar with SPARC at all, but now I noticed that 32-bit SPARC is big endian and gnu-objcopy -Oelf32-sparc certainly emits a big endian ELF file. I think I should fix the SPARC entry in ArchMap before committing this patch. What do you think on this?

llvm/tools/llvm-objcopy/CopyConfig.cpp

266–267

I mean this 32-bit SPARC entry should be big endian.

[seiya]

In LLVM, there's a target named SPARCV8el, the 32-bit little-endian SPARC target so adding sparc/ sparcel into ArchMap and elf32-sparc/elf32-sparcel into OutputFormatMap would be better I think.

[jhenderson]

In D63238#1541295, @seiya wrote:

In D63238#1541201, @jhenderson wrote:

Looks good, I think, but does GNU objcopy have sparc as -I/-O type? Is the syntax the same as you have used?

Yes, it has the same target name. I'm not familiar with SPARC at all, but now I noticed that 32-bit SPARC is big endian and gnu-objcopy -Oelf32-sparc certainly emits a big endian ELF file. I think I should fix the SPARC entry in ArchMap before committing this patch. What do you think on this?

I'm not familiar with SPARC either. Are you confident about it being big endian? If so, fixing the map seems fine and relevant. I have no problem with you adding the sparcel target if you want.

[seiya]

In D63238#1541370, @jhenderson wrote:

In D63238#1541295, @seiya wrote:

In D63238#1541201, @jhenderson wrote:

Looks good, I think, but does GNU objcopy have sparc as -I/-O type? Is the syntax the same as you have used?

Yes, it has the same target name. I'm not familiar with SPARC at all, but now I noticed that 32-bit SPARC is big endian and gnu-objcopy -Oelf32-sparc certainly emits a big endian ELF file. I think I should fix the SPARC entry in ArchMap before committing this patch. What do you think on this?

I'm not familiar with SPARC either. Are you confident about it being big endian? If so, fixing the map seems fine and relevant. I have no problem with you adding the sparcel target if you want.

According to bfd/elf32-sparc.c in GNU binutils and clang/lib/Basic/Targets/Sparc.h, it seems to be big endian. I'll submit another patch for this.

[seiya]

• FIx endianess: see https://reviews.llvm.org/rL363336
• Add elf32-sparcel for 32-bit little endian SPARC.

[jhenderson]

LGTM.

llvm/test/tools/llvm-objcopy/ELF/cross-arch-headers.test

 # RUN: llvm-objcopy %t.o -O elf64-tradbigmips %t.elf64_tradbigmips.o --split-dwo=%t.elf64_tradbigmips.dwo
 # RUN: llvm-readobj --file-headers %t.elf64_tradbigmips.o | FileCheck %s --check-prefixes=CHECK,BE,MIPS,MIPS64BE,64,SYSV
 # RUN: llvm-readobj --file-headers %t.elf64_tradbigmips.dwo | FileCheck %s --check-prefixes=CHECK,BE,MIPS,MIPS64BE,64,SYSV
 
 # RUN: llvm-objcopy %t.o -O elf64-tradlittlemips %t.elf64_tradlittlemips.o --split-dwo=%t.elf64_tradlittlemips.dwo
 # RUN: llvm-readobj --file-headers %t.elf64_tradlittlemips.o | FileCheck %s --check-prefixes=CHECK,LE,MIPS,MIPS64LE,64,SYSV
 # RUN: llvm-readobj --file-headers %t.elf64_tradlittlemips.dwo | FileCheck %s --check-prefixes=CHECK,LE,MIPS,MIPS64LE,64,SYSV
 
+# RUN: llvm-objcopy %t.o -O elf32-sparc %t.elf32_sparc.o --split-dwo=%t.elf32_sparc.dwo
+# RUN: llvm-readobj --file-headers %t.elf32_sparc.o | FileCheck %s --check-prefixes=CHECK,BE,SPARC,32,SYSV
+# RUN: llvm-readobj --file-headers %t.elf32_sparc.dwo | FileCheck %s --check-prefixes=CHECK,BE,SPARC,32,SYSV
+
+# RUN: llvm-objcopy %t.o -O elf32-sparcel %t.elf32_sparcel.o --split-dwo=%t.elf32_sparcel.dwo
+# RUN: llvm-readobj --file-headers %t.elf32_sparcel.o | FileCheck %s --check-prefixes=CHECK,LE,SPARCEL,32,SYSV
+# RUN: llvm-readobj --file-headers %t.elf32_sparcel.dwo | FileCheck %s --check-prefixes=CHECK,LE,SPARCEL,32,SYSV
+
 !ELF
 FileHeader:
   Class:           ELFCLASS32
   Data:            ELFDATA2LSB
   Type:            ET_EXEC
   # Arbitrary values that do not match any value we convert to via --output-format.
   Machine:         EM_NONE
   OSABI:           ELFOSABI_STANDALONE

 # AARCH-SAME:   aarch64
 # ARM-SAME:     arm-little
 # MIPS-SAME:    mips
 # PPC-SAME:     ppc{{$}}
 # PPC64BE-SAME: ppc64{{$}}
 # PPC64LE-SAME: ppc64{{$}}
 # RISCV32-SAME: riscv{{$}}
 # RISCV64-SAME: riscv{{$}}
+# SPARC-SAME:   sparc
 # X86-64-SAME:  x86-64
 # DEFAULT-SAME: unknown
 
 # I386-NEXT:     Arch: i386
 # IAMCU-NEXT:    Arch: i386
 # AARCH-NEXT:    Arch: aarch64
 # ARM-NEXT:      Arch: arm
 # MIPSBE-NEXT:   Arch: mips{{$}}
 # MIPSLE-NEXT:   Arch: mipsel{{$}}
 # MIPS64BE-NEXT: Arch: mips64{{$}}
 # MIPS64LE-NEXT: Arch: mips64el{{$}}
 # PPC-NEXT:      Arch: powerpc{{$}}
 # PPC64BE-NEXT:  Arch: powerpc64{{$}}
 # PPC64LE-NEXT:  Arch: powerpc64le
 # RISCV32-NEXT:  Arch: riscv32
 # RISCV64-NEXT:  Arch: riscv64
+# SPARC-NEXT:    Arch: sparc{{$}}
+# SPARCEL-NEXT:  Arch: sparcel
 # X86-64-NEXT:   Arch: x86_64
 # DEFAULT-NEXT:  Arch: unknown
 
 # 32-NEXT:      AddressSize: 32bit
 # 64-NEXT:      AddressSize: 64bit
 
 # 32: Class: 32-bit
 # 64: Class: 64-bit
 # LE: DataEncoding: LittleEndian
 # BE: DataEncoding: BigEndian
 
 # SYSV:    OS/ABI: SystemV (0x0)
 # FREEBSD: OS/ABI: FreeBSD (0x9)
 # DEFAULT: OS/ABI: Standalone (0xFF)
 
 # AARCH:   Machine: EM_AARCH64 (0xB7)
 # ARM:     Machine: EM_ARM (0x28)
 # I386:    Machine: EM_386 (0x3)
 # IAMCU:   Machine: EM_IAMCU (0x6)
 # MIPS:    Machine: EM_MIPS (0x8)
 # PPC:     Machine: EM_PPC (0x14)
 # PPC64:   Machine: EM_PPC64 (0x15)
 # RISCV32: Machine: EM_RISCV (0xF3)
 # RISCV64: Machine: EM_RISCV (0xF3)
+# SPARC:   Machine: EM_SPARC (0x2)
 # X86-64:  Machine: EM_X86_64 (0x3E)
 
 # 32: HeaderSize: 52
 # 64: HeaderSize: 64
 
 # 32: SectionHeaderEntrySize: 40
 # 64: SectionHeaderEntrySize: 64

llvm/tools/llvm-objcopy/CopyConfig.cpp

static const StringMap<MachineInfo> ArchMap{
     {"aarch64", {ELF::EM_AARCH64, true, true}},
     {"arm", {ELF::EM_ARM, false, true}},
     {"i386", {ELF::EM_386, false, true}},
     {"i386:x86-64", {ELF::EM_X86_64, true, true}},
     {"mips", {ELF::EM_MIPS, false, false}},
     {"powerpc:common64", {ELF::EM_PPC64, true, true}},
     {"riscv:rv32", {ELF::EM_RISCV, false, true}},
     {"riscv:rv64", {ELF::EM_RISCV, true, true}},
     {"sparc", {ELF::EM_SPARC, false, false}},
     {"sparcel", {ELF::EM_SPARC, false, true}},
     {"x86-64", {ELF::EM_X86_64, true, true}},
 };
 
 static Expected<const MachineInfo &> getMachineInfo(StringRef Arch) {
   auto Iter = ArchMap.find(Arch);
   if (Iter == std::end(ArchMap))
     return createStringError(errc::invalid_argument,
                              "Invalid architecture: '%s'", Arch.str().c_str());
static const StringMap<MachineInfo> OutputFormatMap{
     // MIPS
     {"elf32-bigmips", {ELF::EM_MIPS, false, false}},
     {"elf32-ntradbigmips", {ELF::EM_MIPS, false, false}},
     {"elf32-ntradlittlemips", {ELF::EM_MIPS, false, true}},
     {"elf32-tradbigmips", {ELF::EM_MIPS, false, false}},
     {"elf32-tradlittlemips", {ELF::EM_MIPS, false, true}},
     {"elf64-tradbigmips", {ELF::EM_MIPS, true, false}},
     {"elf64-tradlittlemips", {ELF::EM_MIPS, true, true}},
+    // SPARC
+    {"elf32-sparc", {ELF::EM_SPARC, false, false}},
+    {"elf32-sparcel", {ELF::EM_SPARC, false, true}},
 };
 
 static Expected<MachineInfo> getOutputFormatMachineInfo(StringRef Format) {
   StringRef OriginalFormat = Format;
   bool IsFreeBSD = Format.consume_back("-freebsd");
   auto Iter = OutputFormatMap.find(Format);
   if (Iter == std::end(OutputFormatMap))
     return createStringError(errc::invalid_argument,