False positive msan error with strtol and -std=c2x or _GNU_SOURCE

[tavianator]

A recent glibc update added a new strtol() implementation to support the 0b prefix from C23. This new symbol seems to be missing an interceptor:

tavianator@graphene $ cat foo.c
#include <assert.h>
#include <stdlib.h>

int main(void) {
        char *end;
        long l = strtol("42", &end, 0);
        assert(l == 42);
        assert(*end == '\0');
        return 0;
}
tavianator@graphene $ clang -fsanitize=memory foo.c -o foo && ./foo
tavianator@graphene $ clang -std=c2x -fsanitize=memory foo.c -o foo && ./foo
==58541==WARNING: MemorySanitizer: use-of-uninitialized-value
    #0 0x55994bbe00da in main (/home/tavianator/code/bfs/foo+0xb50da) (BuildId: 43c18b79db0838f533698c3e83e9322a882086d9)
    #1 0x7f1a12c27ccf  (/usr/lib/libc.so.6+0x27ccf) (BuildId: 316d0d3666387f0e8fb98773f51aa1801027c5ab)
    #2 0x7f1a12c27d89 in __libc_start_main (/usr/lib/libc.so.6+0x27d89) (BuildId: 316d0d3666387f0e8fb98773f51aa1801027c5ab)
    #3 0x55994bb4d134 in _start (/home/tavianator/code/bfs/foo+0x22134) (BuildId: 43c18b79db0838f533698c3e83e9322a882086d9)

SUMMARY: MemorySanitizer: use-of-uninitialized-value (/home/tavianator/code/bfs/foo+0xb50da) (BuildId: 43c18b79db0838f533698c3e83e9322a882086d9) in main
Exiting
tavianator@graphene $ clang -D_GNU_SOURCE -fsanitize=memory foo.c -o foo && ./foo
==58556==WARNING: MemorySanitizer: use-of-uninitialized-value
    #0 0x5643ee5720da in main (/home/tavianator/code/bfs/foo+0xb50da) (BuildId: 43c18b79db0838f533698c3e83e9322a882086d9)
    #1 0x7f302fc27ccf  (/usr/lib/libc.so.6+0x27ccf) (BuildId: 316d0d3666387f0e8fb98773f51aa1801027c5ab)
    #2 0x7f302fc27d89 in __libc_start_main (/usr/lib/libc.so.6+0x27d89) (BuildId: 316d0d3666387f0e8fb98773f51aa1801027c5ab)
    #3 0x5643ee4df134 in _start (/home/tavianator/code/bfs/foo+0x22134) (BuildId: 43c18b79db0838f533698c3e83e9322a882086d9)

SUMMARY: MemorySanitizer: use-of-uninitialized-value (/home/tavianator/code/bfs/foo+0xb50da) (BuildId: 43c18b79db0838f533698c3e83e9322a882086d9) in main
Exiting
tavianator@graphene $ nm -u foo | grep strtol
                 U __isoc23_strtol@GLIBC_2.38
tavianator@graphene $ clang --version
clang version 15.0.7
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin
tavianator@graphene $ /usr/lib/libc.so.6
GNU C Library (GNU libc) stable release version 2.38.
Copyright (C) 2023 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 13.2.1 20230801.
libc ABIs: UNIQUE IFUNC ABSOLUTE
Minimum supported kernel: 4.4.0
For bug reporting instructions, please see:
<https://bugs.archlinux.org/>.

[MaskRay]

Patch: https://reviews.llvm.org/D158943 [sanitizer] Intercept glibc 2.38 __isoc23_* functions

@tavianator It'd be nice if you can verify whether the patch helps.

curl -s https://raw.githubusercontent.com/chromium/chromium/main/tools/clang/scripts/update.py | python3 - --output-dir=~/Stable

# Use https://github.com/MaskRay/llvm-project/tree/san-c2x
cmake -GNinja -Sllvm -Bout/release -DCMAKE_CXX_COMPILER=$HOME/Stable/bin/clang++ -DCMAKE_C_COMPILER=$HOME/Stable/bin/clang  -DLLVM_ENABLE_PROJECTS='clang;lld;compiler-rt' -DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD=host -DCLANG_ENABLE_ARCMT=off -DCLANG_ENABLE_STATIC_ANALYZER=off
ninja -C out/release clang asan msan compiler-rt-headers
# Use out/release/bin/clang

[tavianator]

@MaskRay I can confirm that MaskRay@9e305a0 fixes it for me, thanks!

[EugeneZelenko]

@MaskRay: Should it be backported to 17.x?

[MaskRay]

Filed release/17.x backport request on #64388

@MaskRay: Should it be backported to 17.x?

Yes! My commit message mentioned

The strtol family functions are spreaded, which is not great, but the
patch (intended for release/17.x) does not attempt to address the issue.