[NFC] Remove no-op code, duplicated, with multiple confusing comments, from x86AssemblyInspectionEngine::jmp_to_reg_p

[jasonmolenda]

Summary

Thanks to @fixathon for catching this code I added in https://reviews.llvm.org/D62764. This method clears the lower 3 bits of the second byte of an instruction, then compares it to a few values including 0x20 (fine) and 0x24, twice, with comments that seem to contradict the behavior. Of course, having masked off the low 3 bits, these duplicated entries are never matched.

[fixathon]

LGTM Thanks for fixing this!

lldb/source/Plugins/UnwindAssembly/x86/x86AssemblyInspectionEngine.cpp

 // ff 60 60                                jmpq   *0x60(%rax)
 bool x86AssemblyInspectionEngine::jmp_to_reg_p() {
   if (*m_cur_insn != 0xff)
     return false;
 
   // The second byte is a ModR/M /4 byte, strip off the registers
   uint8_t second_byte_sans_reg = *(m_cur_insn + 1) & ~7;
 
-  // Don't handle 0x24 disp32, because the target address is
-  // knowable statically - pc_rel_branch_or_jump_p() will
-  // return the target address.
-
   // [reg]
   if (second_byte_sans_reg == 0x20)
     return true;
 
   // [reg]+disp8
   if (second_byte_sans_reg == 0x60)
     return true;
 
   // [reg]+disp32
   if (second_byte_sans_reg == 0xa0)
     return true;
 
   // reg
   if (second_byte_sans_reg == 0xe0)
     return true;
 
-  // disp32
-  // jumps to an address stored in memory, the value can't be cached
-  // in an unwind plan.
-  if (second_byte_sans_reg == 0x24)
-    return true;
-
-  // use SIB byte
-  // ff 24 fe  jmpq   *(%rsi,%rdi,8)
-  if (second_byte_sans_reg == 0x24)
-    return true;
-
   return false;
 }
 
 // Detect branches to fixed pc-relative offsets.
 // Returns the offset from the address of the next instruction
 // that may be branch/jumped to.
 //
 // Cannot determine the offset of a JMP that jumps to the address in