Date: Fri, 07 Jun 2024 18:35:25 +0000 From: bugzilla-noreply@freebsd.org To: standards@FreeBSD.org Subject: [Bug 279560] FreeBSD 14.1 ships libc++ which passes wrong size to sized operator delete Message-ID: <bug-279560-99-vDvQrXEqHP@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-279560-99@https.bugs.freebsd.org/bugzilla/> References: <bug-279560-99@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D279560 --- Comment #3 from Dimitry Andric <dim@FreeBSD.org> --- This is a bit of a false positive due to internal details of libc++ std::string. Normally the size of a std::string is 24 bytes on amd64, but only if it contains a 'short' string. When the contained string is too long to fit in = the object itself, it's reallocated on the heap instead. In this particular case some TCMalloc test string comes out as 25 bytes, and that is what libc++ eventually calls malloc() with. However due to internal administration issues, it calculates the wrong size for the corresponding delete. This was also noticed by AddressSanitizer after <https://github.com/llvm/llvm-project/pull/83774>; ("Enable sized deallocati= on by default in C++14 onwards"), and fixed by Vitaly Buka in <https://github.com/llvm/llvm-project/commit/d129ea8d2fa3>. I will merge that fix into -CURRENT, with one prerequisite fix added, so we= can see if there's any fallout (but it's not likely). Then I will MFC it, and we can think about a 14.1 erratum. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-279560-99-vDvQrXEqHP>