Package ghidra.program.model.correlate

Class HashedFunctionAddressCorrelation

java.lang.Object
ghidra.program.model.correlate.HashedFunctionAddressCorrelation
All Implemented Interfaces:
ListingAddressCorrelation
public class HashedFunctionAddressCorrelation extends Object implements ListingAddressCorrelation
Correlator to construct a 1-1 map between the Instructions of two similar Functions. Matching is performed via a greedy algorithm that looks for sequences (n-grams) of Instructions that are similar between the two Functions. Similarity of two sequences is determined by comparing hashes generated by the HashCalculator object. 1) Potential sequences and their hashes are generated for both functions (see HashStore). 2) Sequences are pulled from the HashStore based on the uniqueness of a potential match and on the size of the sequence. 3) If a unique match is found between sequences, it is extended in either direction as far as possible, as constrained by HashCalculator and the containing basic-blocks. 4) The matching Instruction pairs are put in the final map and removed from further sequence lists to allow other potential matches to be considered. 5) Sequences with no corresponding match are also removed from consideration. 6) Sequences are limited to a single basic-block, and the algorithm is basic-block aware. Once a match establishes a correspondence between a pair of basic blocks, the algorithm uses 7) If a particular sequence has matches that are not unique, the algorithm tries to disambiguate the potential matches by looking at parent/child relationships of the containing basic-blocks. (see DisambiguateStrategy) 8) Multiple passes are attempted, each time the set of potential sequences is completely regenerated, varying the range of sequence sizes for which a match is attempted and other hash parameters. This allows matches discovered by earlier passes to disambiguate sequences in later passes.

  • Constructor Details

    • HashedFunctionAddressCorrelation

      public HashedFunctionAddressCorrelation(Function leftFunction, Function rightFunction, TaskMonitor monitor) throws CancelledException, MemoryAccessException
      Correlates addresses between the two specified functions.
      Parameters:
      leftFunction - the first function
      rightFunction - the second function
      monitor - the task monitor that indicates progress and allows the user to cancel.
      Throws:
      CancelledException - if the user cancels
      MemoryAccessException - if either functions memory can't be accessed.

  • Method Details

    • getProgram

      public Program getProgram(Duo.Side side)
      Description copied from interface: ListingAddressCorrelation
      Gets the program for the given side.
      Specified by:
      getProgram in interface ListingAddressCorrelation
      Parameters:
      side - LEFT or RIGHT
      Returns:
      the program for the given side

    • getAddresses

      public AddressSetView getAddresses(Duo.Side side)
      Description copied from interface: ListingAddressCorrelation
      Gets the addresses that are part of the correlator for the given side
      Specified by:
      getAddresses in interface ListingAddressCorrelation
      Parameters:
      side - LEFT or RIGHT
      Returns:
      the addresses that are part of the correlator for the given side

    • getTotalInstructionsInFirst

      public int getTotalInstructionsInFirst()
      Gets the total number of instructions that are in the first function.
      Returns:
      the first function's instruction count.

    • getTotalInstructionsInSecond

      public int getTotalInstructionsInSecond()
      Gets the total number of instructions that are in the second function.
      Returns:
      the second function's instruction count.

    • numMatchedInstructionsInFirst

      public int numMatchedInstructionsInFirst()
      Determines the number of instructions from the first function that match an instruction in the second function.
      Returns:
      the number of instructions in the first function that have matches.

    • numMatchedInstructionsInSecond

      public int numMatchedInstructionsInSecond()
      Determines the number of instructions from the second function that match an instruction in the first function.
      Returns:
      the number of instructions in the second function that have matches.

    • getUnmatchedInstructionsInFirst

      public List<Instruction> getUnmatchedInstructionsInFirst()
      Determines the number of instructions from the first function that do not match an instruction in the second function.
      Returns:
      the number of instructions in the first function without matches.

    • getUnmatchedInstructionsInSecond

      public List<Instruction> getUnmatchedInstructionsInSecond()
      Determines the number of instructions from the second function that do not match an instruction in the first function.
      Returns:
      the number of instructions in the second function without matches.

    • getFirstToSecondIterator

      public Iterator<Map.Entry<Address,Address>> getFirstToSecondIterator()
      Gets an iterator of the matching addresses from the first function to the second.
      Returns:
      the iterator

    • getAddress

      public Address getAddress(Duo.Side side, Address otherSideAddress)
      Description copied from interface: ListingAddressCorrelation
      Gets the address for the given side that matches the given address from the other side.
      Specified by:
      getAddress in interface ListingAddressCorrelation
      Parameters:
      side - the side to get an address for
      otherSideAddress - the address from the other side to find a match for
      Returns:
      the address for the given side that matches the given address from the other side.

    • getFunction

      public Function getFunction(Duo.Side side)
      Description copied from interface: ListingAddressCorrelation
      Gets the function for the given side. This will be null if the addresses are not function based.
      Specified by:
      getFunction in interface ListingAddressCorrelation
      Parameters:
      side - LEFT or RIGHT
      Returns:
      the function for the given side or null if not function based