[libcxx-commits] [libcxx] [ASan][libc++] Initialize `__r_` variable with lambda (PR #77394)

[via libcxx-commits]

https://github.com/AdvenamTacet updated https://github.com/llvm/llvm-project/pull/77394

>From 4ea954bb63d4d956242f4e1e4366b9046823c304 Mon Sep 17 00:00:00 2001
From: Advenam Tacet <advenam.tacet at trailofbits.com>
Date: Tue, 9 Jan 2024 00:51:04 +0100
Subject: [PATCH 1/2] [ASan][libc++] Initialize a variable with lambda

This commit is a refactor (increases readability) and optimization fix.

This is a fixed commit of https://github.com/llvm/llvm-project/pull/76200
First reverthed here: https://github.com/llvm/llvm-project/commit/1ea7a56057492d9da1124787a9855cc2edca7df9

The difference is a return type of the lambda.

Original description:

This commit addresses optimization and instrumentation challenges encountered within comma constructors.
  1) _LIBCPP_STRING_INTERNAL_MEMORY_ACCESS does not work in comma constructors.
  2) Code inside comma constructors is not always correctly optimized. Problematic code examples:
        - `: __r_(((__str.__is_long() ? 0 : (__str.__annotate_delete(), 0)), std::move(__str.__r_))) {`
        - `: __r_(__r_([&](){ if(!__s.__is_long()) __s.__annotate_delete(); return std::move(__s.__r_);}())) {`

However, lambda with argument seems to be correctly optimized. This patch uses that fact.

Use of lambda based on idea from @ldionne.
---
 libcxx/include/string | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libcxx/include/string b/libcxx/include/string
index c676182fba8bac..c3cbb560efa911 100644
--- a/libcxx/include/string
+++ b/libcxx/include/string
@@ -922,7 +922,7 @@ public:
       // Turning off ASan instrumentation for variable initialization with _LIBCPP_STRING_INTERNAL_MEMORY_ACCESS
       // does not work consistently during initialization of __r_, so we instead unpoison __str's memory manually first.
       // __str's memory needs to be unpoisoned only in the case where it's a short string.
-      : __r_(((__str.__is_long() ? 0 : (__str.__annotate_delete(), 0)), std::move(__str.__r_))) {
+      : __r_([](basic_string &__s) -> auto&& { if(!__s.__is_long()) __s.__annotate_delete(); return std::move(__s.__r_); }(__str)) {
     __str.__r_.first() = __rep();
     __str.__annotate_new(0);
     if (!__is_long())

>From 59849af6b4813c0fb55dfa7ed82110bc7667e800 Mon Sep 17 00:00:00 2001
From: Advenam Tacet <advenam.tacet at trailofbits.com>
Date: Tue, 9 Jan 2024 11:20:58 +0100
Subject: [PATCH 2/2] Replace auto with decltype

---
 libcxx/include/string | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libcxx/include/string b/libcxx/include/string
index c3cbb560efa911..bdc9a6e5616fd1 100644
--- a/libcxx/include/string
+++ b/libcxx/include/string
@@ -922,7 +922,7 @@ public:
       // Turning off ASan instrumentation for variable initialization with _LIBCPP_STRING_INTERNAL_MEMORY_ACCESS
       // does not work consistently during initialization of __r_, so we instead unpoison __str's memory manually first.
       // __str's memory needs to be unpoisoned only in the case where it's a short string.
-      : __r_([](basic_string &__s) -> auto&& { if(!__s.__is_long()) __s.__annotate_delete(); return std::move(__s.__r_); }(__str)) {
+      : __r_([](basic_string &__s) -> decltype(__s.__r_)&& { if(!__s.__is_long()) __s.__annotate_delete(); return std::move(__s.__r_); }(__str)) {
     __str.__r_.first() = __rep();
     __str.__annotate_new(0);
     if (!__is_long())