[llvm] Fix undefined-behaviour in regex engine. (PR #73071)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-llvm-support

Author: Tanmay (tanmaysachan)

<details>
<summary>Changes</summary>

Running the `mlir-text-parser-fuzzer` discovers a path that causes application of offset to a null pointer (UB) in the regex engine.

This patch adds a check.

---
Full diff: https://github.com/llvm/llvm-project/pull/73071.diff


1 Files Affected:

- (modified) llvm/lib/Support/regengine.inc (+3-1) 


``````````diff
diff --git a/llvm/lib/Support/regengine.inc b/llvm/lib/Support/regengine.inc
index f23993abc6e7e71..54dd96ab9cfada5 100644
--- a/llvm/lib/Support/regengine.inc
+++ b/llvm/lib/Support/regengine.inc
@@ -146,7 +146,9 @@ matcher(struct re_guts *g, const char *string, size_t nmatch,
 	const char *stop;
 
 	/* simplify the situation where possible */
-	if (g->cflags&REG_NOSUB)
+        if (!string)
+		return(REG_INVARG);
+        if (g->cflags&REG_NOSUB)
 		nmatch = 0;
 	if (eflags&REG_STARTEND) {
 		start = string + pmatch[0].rm_so;

``````````

</details>


https://github.com/llvm/llvm-project/pull/73071