feat(PRO-1733): allow attaching to matching analyses by hash only

Enforce boundaries checks at function level

Author: robertmarsal

src/main/java/ai/reveng/toolkit/ghidra/binarysimilarity/ui/autounstrip/AutoUnstripDialog.java

@@ -4,6 +4,7 @@
 import ai.reveng.toolkit.ghidra.core.services.api.GhidraRevengService;
 import ai.reveng.toolkit.ghidra.core.services.api.types.AnalysisID;
 import ai.reveng.toolkit.ghidra.core.services.api.types.AutoUnstripResponse;
+import ai.reveng.toolkit.ghidra.core.services.api.types.FunctionID;
 import ai.reveng.toolkit.ghidra.core.types.ProgramWithBinaryID;
 import ai.reveng.toolkit.ghidra.plugins.ReaiPluginPackage;
 import ghidra.framework.plugintool.PluginTool;
@@ -107,6 +108,9 @@ private void importFunctionNames(AutoUnstripResponse autoUnstripResponse) {
         // Retrieve the mangled names map once outside the transaction
         var mangledNameMapOpt = revengService.getFunctionMangledNamesMap(program);
 
+        // Retrieve the function ID map once outside the transaction
+        var functionMap = revengService.getFunctionMap(program);
+
         program.withTransaction("Apply Auto-Unstrip Function Names", () -> {
                     try {
                         var revengMatchNamespace = program.getSymbolTable().getOrCreateNameSpace(
@@ -121,6 +125,7 @@ private void importFunctionNames(AutoUnstripResponse autoUnstripResponse) {
 
                             var revEngMangledName = match.suggested_name();
                             var revEngDemangledName = match.suggested_demangled_name();
+                            var functionID = functionMap.get(new FunctionID(match.function_id().value()));
 
                             if (
                                     func != null &&
@@ -132,6 +137,8 @@ private void importFunctionNames(AutoUnstripResponse autoUnstripResponse) {
                                     // Only accept valid names (no spaces)
                                     !revEngMangledName.contains(" ") &&
                                     !revEngDemangledName.contains(" ")
+                                    // Only rename if the function ID is known (boundaries matched)
+                                    && functionID != null
                             ) {
                                 try {
                                     // Capture original name before renaming

src/main/java/ai/reveng/toolkit/ghidra/binarysimilarity/ui/functionmatching/AbstractFunctionMatchingDialog.java

@@ -135,7 +135,7 @@ protected void processFunctionMatchingResults(FunctionMatchingBatchResponse resp
                 Function localFunction = functionMap.get(new FunctionID(matchResult.getFunctionId()));
 
                 if (localFunction == null) {
-                    // If we can't find the local function, skip this match
+                    // If we can't find the local function, skip this match (boundaries do not match the remote ones)
                     return;
                 }
 

src/main/java/ai/reveng/toolkit/ghidra/binarysimilarity/ui/recentanalyses/RecentAnalysesTableModel.java

@@ -48,16 +48,6 @@ protected void doLoad(Accumulator<LegacyAnalysisResult> accumulator, TaskMonitor
                         return;
                     }
 
-                    // Filter out analyses where the function boundaries hash does not match our program
-                    var functionBoundariesHash = functionBoundariesService.getFunctionBoundariesHash();
-                    if (!result.function_boundaries_hash().equals(functionBoundariesHash)) {
-                        loggingService.info(
-                            "[RevEng] Skipping analysis for " + result.binary_id() + " as function boundaries hash does" +
-                            " not match. Expected " + functionBoundariesHash + " but got " +
-                            result.function_boundaries_hash());
-                        return;
-                    }
-
                     accumulator.add(result);
                 }
         );

src/main/java/ai/reveng/toolkit/ghidra/core/services/api/GhidraRevengService.java

@@ -188,6 +188,7 @@ private void loadFunctionInfo(Program program, BinaryID binID) throws ApiExcepti
         StringPropertyMap finalMangledNameMap = mangledNameMap;
 
         AtomicInteger ghidraRenamedFunctions = new AtomicInteger();
+        AtomicInteger ghidraBoundariesMatchedFunction = new AtomicInteger();
         functionInfo.forEach(
                 info -> {
                     var oFunc = getFunctionFor(info, program);
@@ -215,9 +216,11 @@ private void loadFunctionInfo(Program program, BinaryID binID) throws ApiExcepti
                     }
 
                     var funcSize = func.getBody().getNumAddresses();
+
                     // For unclear reasons the func size is off by one
                     if (funcSize - 1 != info.functionSize() && funcSize != info.functionSize()){
                         Msg.warn(this, "Function size mismatch for function %s: %d vs %d".formatted(ghidraMangledName, funcSize, info.functionSize()));
+                        return;
                     }
 
                     // Source types:
@@ -238,6 +241,8 @@ private void loadFunctionInfo(Program program, BinaryID binID) throws ApiExcepti
                     }
                     finalFunctionIDMap.add(func.getEntryPoint(), info.functionID().value());
                     finalMangledNameMap.add(func.getEntryPoint(), revEngMangledName);
+
+                    ghidraBoundariesMatchedFunction.getAndIncrement();
                 }
         );
 
@@ -256,14 +261,14 @@ private void loadFunctionInfo(Program program, BinaryID binID) throws ApiExcepti
         program.endTransaction(transactionID, true);
 
         // Print summary
-        Msg.debug(
-    this,
-            "Loaded %d functions from RevEng.AI, renamed %d, Ghidra has %d functions".formatted(
+        Msg.showInfo(this, null, ReaiPluginPackage.WINDOW_PREFIX + "Function loading summary",
+            ("Found %d functions from RevEng.AI. Renamed %d. Your local Ghidra instance has %d/%d matching function " +
+                "boundaries. For better results, please start a new analysis from this plugin.").formatted(
                 functionInfo.size(),
                 ghidraRenamedFunctions.get(),
+                ghidraBoundariesMatchedFunction.get(),
                 ghidraFunctionCount.get()
-            )
-        );
+        ));
     }
 
     /**

src/main/java/ai/reveng/toolkit/ghidra/core/services/function/export/ExportFunctionBoundariesService.java

@@ -29,11 +29,4 @@ public interface ExportFunctionBoundariesService {
 	 * @return
 	 */
 	public JSONArray getFunctionsArray();
-
-    /**
-     * Return a hash of the function boundaries for change detection
-     * Note that this algorithm must match that used on the API server side!
-     * @return
-     */
-    public String getFunctionBoundariesHash();
 }

src/main/java/ai/reveng/toolkit/ghidra/core/services/function/export/ExportFunctionBoundariesServiceImpl.java

@@ -80,58 +80,4 @@ public JSONArray getFunctionsArray() {
 		}
 		return fArray;
 	}
-
-    @Override
-    public String getFunctionBoundariesHash() {
-        if (!isReady)
-            init();
-
-        // Collect all function boundaries into a list
-        List<JSONObject> boundaries = new ArrayList<>();
-        for (Function f : fm.getFunctions(true)) {
-            boundaries.add(getFunctionAt(f.getEntryPoint()));
-        }
-
-        // Sort the boundaries by start address (convert hex string to long for proper sorting)
-        boundaries.sort(Comparator.comparingLong(b -> Long.parseUnsignedLong(
-            b.getString("start_addr").substring(2), 16)));
-
-        // Create a formatted string representation of the boundaries
-        StringBuilder boundariesStr = new StringBuilder();
-        for (int i = 0; i < boundaries.size(); i++) {
-            JSONObject b = boundaries.get(i);
-            if (i > 0) {
-                boundariesStr.append(",");
-            }
-
-            // Convert hex addresses to integer representation
-            String startAddrHex = b.getString("start_addr");
-            String endAddrHex = b.getString("end_addr");
-            long startAddrInt = Long.parseUnsignedLong(startAddrHex.substring(2), 16);
-            long endAddrInt = Long.parseUnsignedLong(endAddrHex.substring(2), 16);
-
-            boundariesStr.append(startAddrInt)
-                         .append("-")
-                         .append(endAddrInt);
-        }
-
-        // Generate SHA-256 hash of the boundaries string
-        try {
-            MessageDigest digest = MessageDigest.getInstance("SHA-256");
-            byte[] hashBytes = digest.digest(boundariesStr.toString().getBytes());
-
-            // Convert to hexadecimal string
-            StringBuilder hexString = new StringBuilder();
-            for (byte b : hashBytes) {
-                String hex = Integer.toHexString(0xff & b);
-                if (hex.length() == 1) {
-                    hexString.append('0');
-                }
-                hexString.append(hex);
-            }
-            return hexString.toString();
-        } catch (NoSuchAlgorithmException e) {
-            throw new RuntimeException("SHA-256 algorithm not available", e);
-        }
-    }
 }