GP-3793 escape non-ASCII symbols during ELF Import

Ghidra/Features/Base/src/main/java/ghidra/app/util/opinion/ElfProgramBuilder.java

@@ -649,8 +649,7 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 
 		long entry = elf.e_entry(); // already adjusted for pre-link
 		if (entry != 0 && (elf.isExecutable() || elf.isSharedObject())) {
-			Address entryAddr =
+			Address entryAddr = createEntryFunction(ElfLoader.ELF_ENTRY_FUNCTION_NAME, entry);
-				createEntryFunction(ElfLoader.ELF_ENTRY_FUNCTION_NAME, entry);
 			if (entryAddr != null) {
 				addElfHeaderReferenceMarkup(elf.getEntryComponentOrdinal(), entryAddr);
 				Function entryFunc = program.getFunctionManager().getFunctionAt(entryAddr);
@@ -996,6 +995,9 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 
 			int symbolIndex = reloc.getSymbolIndex();
 			String symbolName = symbolTable != null ? symbolTable.getSymbolName(symbolIndex) : "";
+			if (symbolName != null && isUnsupportedASCIISymbolName(symbolName)) {
+				symbolName = getEscapedSymbolName(symbolName);
+			}
 
 			Address baseAddress = relocationSpace.getTruncatedAddress(baseWordOffset, true);
 
@@ -1116,9 +1118,8 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 				hasConflict = nextRelocAddr != null && nextRelocAddr.compareTo(maxAddr) <= 0;
 			}
 			if (hasConflict) {
-				Msg.warn(this,
+				Msg.warn(this, "Artificial relocation for " + address +
-					"Artificial relocation for " + address +
+					" conflicts with a previous relocation");
-						" conflicts with a previous relocation");
 			}
 			relocationTable.add(address, Status.APPLIED_OTHER, 0, null, length, null);
 			return true;
@@ -2093,6 +2094,12 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 					isPrimary = (existingSym == null);
 				}
 
+				if (name != null && isUnsupportedASCIISymbolName(name)) {
+					String escapedName = getEscapedSymbolName(name);
+					log("Unsupported ASCII symbol name has been escaped: \"" + escapedName + "\"");
+					name = escapedName;
+				}
+
 				createSymbol(address, name, isPrimary, elfSymbol.isAbsolute(), null);
 
 				// NOTE: treat weak symbols as global so that other programs may link to them.
@@ -2127,6 +2134,39 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 		}
 	}
 
+	private boolean isUnsupportedASCIISymbolName(String name) {
+		for (char c : name.toCharArray()) {
+			if (!StringUtilities.isAsciiChar(c) || c == ' ') {
+				return true;
+			}
+		}
+		return false;
+	}
+
+	private String getEscapedSymbolName(String name) {
+		StringBuilder escapedBuf = new StringBuilder();
+		for (char c : name.toCharArray()) {
+			if (c < 0 || c > 0x7f) {
+				// format non-ASCII as escaped numeric hex value \xNN
+				escapedBuf.append("\\x%02x".formatted((int) c));
+			}
+			else if (c < ' ') {
+				// format as ^Control character for consistency with readelf
+				c += 64; // get ASCII control character, starts with ^@
+				escapedBuf.append('^');
+				escapedBuf.append(c);
+			}
+			else if (c == 0x7f) {
+				// format as ^? character for consistency with readelf
+				escapedBuf.append("^?");
+			}
+			else {
+				escapedBuf.append(c);
+			}
+		}
+		return escapedBuf.toString();
+	}
+
 	@Override
 	public void setElfSymbolAddress(ElfSymbol elfSymbol, Address address) {
 		symbolMap.put(elfSymbol, address);
@@ -3627,14 +3667,13 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 //		Msg.debug(this,
 //			"Loading block " + name + " at " + start + " from file offset " + fileOffset);
 
-		long compSectOrigSize = loadable instanceof ElfSectionHeader header && header.isCompressed()
+		long compSectOrigSize =
-				? header.getSize()
+			loadable instanceof ElfSectionHeader header && header.isCompressed() ? header.getSize()
-				: -1;
+					: -1;
 
 		String blockComment = comment;
 		if (compSectOrigSize >= 0) {
-			blockComment +=
+			blockComment += " (decompressed, original length: 0x%x)".formatted(compSectOrigSize);
-				" (decompressed, original length: 0x%x)".formatted(compSectOrigSize);
 		}
 		else if ((fileOffset + revisedLength - 1) >= fileBytes.getSize()) {
 			// ensure valid length for non-compressed items
@@ -3650,15 +3689,13 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 		try {
 			if (loadable != null && loadable.hasFilteredLoadInputStream(this, start)) {
 				// block is unable to map directly to file bytes - load from input stream
-				try (InputStream is =
+				try (InputStream is = loadable.getFilteredLoadInputStream(this, start,
-					loadable.getFilteredLoadInputStream(this, start, revisedLength,
+					revisedLength, (errorMsg, th) -> {
-						(errorMsg, th) -> {
+						String loadableTypeStr = compSectOrigSize >= 0 ? "compressed section " : "";
-							String loadableTypeStr =
+						log.appendMsg("Error when reading %s[%s]: %s".formatted(loadableTypeStr,
-								compSectOrigSize >= 0 ? "compressed section " : "";
+							name, errorMsg));
-							log.appendMsg("Error when reading %s[%s]: %s".formatted(loadableTypeStr,
+						Msg.error(this, errorMsg, th);
-								name, errorMsg));
+					})) {
-							Msg.error(this, errorMsg, th);
-						})) {
 					block = MemoryBlockUtils.createInitializedBlock(program, isOverlay, name, start,
 						is, revisedLength, blockComment, BLOCK_SOURCE_NAME, r, w, x, log, monitor);
 				}
@@ -3673,8 +3710,8 @@ class ElfProgramBuilder extends MemorySectionResolver implements ElfLoadHelper {
 		finally {
 			if (block == null) {
 				Address end = start.addNoWrap(revisedLength - 1);
-				log("Unexpected ELF memory block load conflict when creating '" + name +
+				log("Unexpected ELF memory block load conflict when creating '" + name + "' at " +
-					"' at " + start.toString(true) + "-" + end.toString(true));
+					start.toString(true) + "-" + end.toString(true));
 			}
 		}
 		return block;