From c7056b31ebbf5a176f49e9609aded8a4a4b89fb0 Mon Sep 17 00:00:00 2001
From: Jacob Panov <84049465+jacobpanov@users.noreply.github.com>
Date: Wed, 30 Jul 2025 11:00:09 -0400
Subject: [PATCH] Fix llvm-mca tool preprocessing of debug directives (#7964)

This PR fixes the "unassigned file number" error in llvm-mca when
processing GCC-generated assembly with debug information #6795 .

**Problem**: GCC with `-g` flag generates `.loc` and `.file` debug
directives that cause llvm-mca to fail with "unassigned file number"
errors.

**Solution**: Filter out `.loc` and `.file` directives during assembly
preprocessing, similar to how other problematic directives are already
handled.

**Changes**:
- Add regex filtering for `.loc` and `.file` directives in
`rewriteAsm()`
- Make `rewriteAsm()` method public for testability
- Add comprehensive test coverage for debug directive removal

The fix is minimal, targeted, and follows the existing pattern of
directive filtering.
---
 lib/tooling/llvm-mca-tool.ts |  5 +++--
 test/analysis-tests.ts       | 16 ++++++++++++++++
 2 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/lib/tooling/llvm-mca-tool.ts b/lib/tooling/llvm-mca-tool.ts
index 3d24c1596..9528f7bc0 100644
--- a/lib/tooling/llvm-mca-tool.ts
+++ b/lib/tooling/llvm-mca-tool.ts
@@ -36,12 +36,13 @@ export class LLVMMcaTool extends BaseTool {
         return 'llvm-mca-tool';
     }
 
-    rewriteAsm(asm: string) {
+    public rewriteAsm(asm: string) {
         return asm
             .replaceAll(/.hword\s/gim, '.short ')
             .replaceAll(/offset flat:/gim, '')
             .replaceAll(/ptr\s%fs/gim, 'PTR fs')
-            .replaceAll(/^\s*\.(fnstart|eabi_attribute|fpu).*/gim, ''); // https://github.com/compiler-explorer/compiler-explorer/issues/1270
+            .replaceAll(/^\s*\.(fnstart|eabi_attribute|fpu).*/gim, '') // https://github.com/compiler-explorer/compiler-explorer/issues/1270
+            .replaceAll(/^\s*\.(loc|file).*/gim, ''); // Remove debug directives that can cause "unassigned file number" errors
     }
 
     writeAsmFile(data: string, destination: string) {
diff --git a/test/analysis-tests.ts b/test/analysis-tests.ts
index 9dc22bd8b..42f56cdae 100644
--- a/test/analysis-tests.ts
+++ b/test/analysis-tests.ts
@@ -26,6 +26,9 @@ import {beforeAll, describe, expect, it} from 'vitest';
 
 import {CompilationEnvironment} from '../lib/compilation-env.js';
 import {AnalysisTool, LLVMmcaTool} from '../lib/compilers/index.js';
+import {ToolEnv} from '../lib/tooling/base-tool.interface.js';
+import {LLVMMcaTool as LLVMMcaTooling} from '../lib/tooling/llvm-mca-tool.js';
+import {ToolInfo} from '../types/tool.interfaces.js';
 
 import {
     makeCompilationEnvironment,
@@ -100,4 +103,17 @@ describe('LLVM-mca tool definition', () => {
         });
         expect(new AnalysisTool(info, ce).getInfo().disabledFilters).toEqual(['labels', 'directives', 'debugCalls']);
     });
+
+    it('should remove .loc and .file directives from assembly', () => {
+        const mcaTool = new LLVMMcaTooling({} as ToolInfo, {} as ToolEnv);
+
+        const asmWithDebugInfo = '.file "test.c"\n.loc 1 1 12\nmovl $2, %eax\n.file 0 "/path" "test.c"\nret\n';
+        const result = mcaTool.rewriteAsm(asmWithDebugInfo);
+
+        // Verify that .loc and .file directives are removed
+        expect(result).not.toContain('.loc');
+        expect(result).not.toContain('.file');
+        expect(result).toContain('movl $2, %eax');
+        expect(result).toContain('ret');
+    });
 });