j*cxz/pusha/popa alternate size forms

these all existed since forever but the library did not distinguish them
and did not provide prefix information for users to tell which had been
decoded.

1 files changed, 37 insertions, 3 deletions

diff --git a/src/real_mode/mod.rs b/src/real_mode/mod.rs
index f538fa5..8dd29b1 100644
--- a/src/real_mode/mod.rs
+++ b/src/real_mode/mod.rs
@@ -2508,6 +2508,13 @@ pub enum Opcode {
     PVALIDATE,
     RMPADJUST,
     RMPUPDATE,
+
+    // these were part of the base (i386) ISA but were reported as distinct opcodes relatively late.
+    PUSHAD,
+    POPAD,
+
+    // likewise with alternate-size jecxz/jcxz..
+    JECXZ,
 }
 
 impl PartialEq for Instruction {
@@ -3992,6 +3999,7 @@ enum OperandCase {
     ModRM_0xf30f38fa,
     ModRM_0xf30f38fb,
     ModRM_0xf30f3af0,
+    CXZ, // catch-all for "j*cxz and prefix handling"
 }
 
 #[allow(non_camel_case_types)]
@@ -4319,6 +4327,7 @@ enum OperandCode {
     ModRM_0xc4 = OperandCodeBuilder::new().operand_case(OperandCase::ModRM_0xc4).bits(),
     ModRM_0xc5 = OperandCodeBuilder::new().operand_case(OperandCase::ModRM_0xc5).bits(),
     MASKMOVDQU = OperandCodeBuilder::new().read_E().reg_mem().operand_case(OperandCase::MASKMOVDQU).bits(),
+    CXZ = OperandCodeBuilder::new().operand_case(OperandCase::CXZ).bits(),
 }
 
 fn base_opcode_map(v: u8) -> Opcode {
@@ -4622,7 +4631,7 @@ const OPCODES: [OpcodeRecord; 256] = [
     OpcodeRecord::new(Interpretation::Instruction(Opcode::LOOPNZ), OperandCode::Ibs),
     OpcodeRecord::new(Interpretation::Instruction(Opcode::LOOPZ), OperandCode::Ibs),
     OpcodeRecord::new(Interpretation::Instruction(Opcode::LOOP), OperandCode::Ibs),
-    OpcodeRecord::new(Interpretation::Instruction(Opcode::JCXZ), OperandCode::Ibs),
+    OpcodeRecord::new(Interpretation::Instruction(Opcode::JCXZ), OperandCode::CXZ),
     OpcodeRecord::new(Interpretation::Instruction(Opcode::IN), OperandCode::AL_Ib),
     OpcodeRecord::new(Interpretation::Instruction(Opcode::IN), OperandCode::AX_Ib),
     OpcodeRecord::new(Interpretation::Instruction(Opcode::OUT), OperandCode::Ib_AL),
@@ -6510,9 +6519,19 @@ fn read_operands<
             } else if instruction.opcode == Opcode::XLAT {
                 instruction.mem_size = 1;
             } else if instruction.opcode == Opcode::PUSHA {
-                instruction.mem_size = 2 * 8;
+                if !instruction.prefixes.operand_size() {
+                    instruction.mem_size = 2 * 8;
+                } else {
+                    instruction.opcode = Opcode::PUSHAD;
+                    instruction.mem_size = 4 * 8;
+                }
             } else if instruction.opcode == Opcode::POPA {
-                instruction.mem_size = 2 * 8;
+                if !instruction.prefixes.operand_size() {
+                    instruction.mem_size = 2 * 8;
+                } else {
+                    instruction.opcode = Opcode::POPAD;
+                    instruction.mem_size = 4 * 8;
+                }
             }
             instruction.operands[0] = OperandSpec::Nothing;
             instruction.operand_count = 0;
@@ -9050,6 +9069,21 @@ fn read_operands<
             instruction.mem_size = 2;
             instruction.operand_count = 2;
         },
+        OperandCase::CXZ => {
+            if instruction.prefixes.address_size() {
+                // address-size overridden from 16-bit to 32-bit
+                instruction.opcode = Opcode::JECXZ;
+            }
+            instruction.imm =
+                read_imm_signed(words, 1)? as u32;
+            sink.record(
+                words.offset() as u32 * 8 - 8,
+                words.offset() as u32 * 8 - 1,
+                InnerDescription::Number("1-byte immediate", instruction.imm as i64)
+                    .with_id(words.offset() as u32 * 8),
+            );
+            instruction.operands[0] = OperandSpec::ImmI8;
+        },
     };
     Ok(())
 }