[llvm-objdump][BPF] --symbolize-operands: infer local labels for BPF (#100550)

Enable local labels computation for BPF disassembly when
`--symbolize-operands` option is specified.
This relies on `MCInstrAnalysis::evaluateBranch()` method, which is
already defined in `BPFMCInstrAnalysis::evaluateBranch`.

After this change the assembly code below:

        if r1 > 42 goto +1
        r1 -= 10
        ...

Would be printed as:

        if r1 > 42 goto +1 <L0>
        r1 -= 10
      <L0>:
        ...

(when `--symbolize-operands` option is set).

See https://reviews.llvm.org/D84191 for the main part of the
`--symbolize-operands` implementation logic.

1 files changed, 4 insertions, 2 deletions

diff --git a/llvm/tools/llvm-objdump/llvm-objdump.cpp b/llvm/tools/llvm-objdump/llvm-objdump.cpp
index d124002..768a976 100644
--- a/llvm/tools/llvm-objdump/llvm-objdump.cpp
+++ b/llvm/tools/llvm-objdump/llvm-objdump.cpp
@@ -1484,9 +1484,11 @@ collectLocalBranchTargets(ArrayRef<uint8_t> Bytes, MCInstrAnalysis *MIA,
                           const MCSubtargetInfo *STI, uint64_t SectionAddr,
                           uint64_t Start, uint64_t End,
                           std::unordered_map<uint64_t, std::string> &Labels) {
-  // So far only supports PowerPC and X86.
+  // Supported by certain targets.
   const bool isPPC = STI->getTargetTriple().isPPC();
-  if (!isPPC && !STI->getTargetTriple().isX86())
+  const bool isX86 = STI->getTargetTriple().isX86();
+  const bool isBPF = STI->getTargetTriple().isBPF();
+  if (!isPPC && !isX86 && !isBPF)
     return;
 
   if (MIA)