linux-kernel - Re: [PATCH] lib/string.c: implement stpcpy

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <562003af51ca0b08f2108147b8d6e75cec18f3fd.camel@perches.com>
Date:   Fri, 14 Aug 2020 17:52:00 -0700
From:   Joe Perches <joe@...ches.com>
To:     Nick Desaulniers <ndesaulniers@...gle.com>,
        Andrew Morton <akpm@...ux-foundation.org>
Cc:     Dávid Bolvanský <david.bolvansky@...il.com>,
        Eli Friedman <efriedma@...cinc.com>, stable@...r.kernel.org,
        Sami Tolvanen <samitolvanen@...gle.com>,
        Tony Luck <tony.luck@...el.com>,
        Yury Norov <yury.norov@...il.com>,
        Daniel Axtens <dja@...ens.net>,
        Arvind Sankar <nivedita@...m.mit.edu>,
        Dan Williams <dan.j.williams@...el.com>,
        "Joel Fernandes (Google)" <joel@...lfernandes.org>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Kees Cook <keescook@...omium.org>,
        Alexandru Ardelean <alexandru.ardelean@...log.com>,
        linux-kernel@...r.kernel.org, clang-built-linux@...glegroups.com
Subject: Re: [PATCH] lib/string.c: implement stpcpy

On Fri, 2020-08-14 at 17:24 -0700, Nick Desaulniers wrote:
> LLVM implemented a recent "libcall optimization" that lowers calls to
> `sprintf(dest, "%s", str)` where the return value is used to
> `stpcpy(dest, str) - dest`. This generally avoids the machinery involved
> in parsing format strings.
[]
> diff --git a/include/linux/string.h b/include/linux/string.h
[]
> @@ -31,6 +31,9 @@ size_t strlcpy(char *, const char *, size_t);
>  #ifndef __HAVE_ARCH_STRSCPY
>  ssize_t strscpy(char *, const char *, size_t);
>  #endif
> +#ifndef __HAVE_ARCH_STPCPY
> +extern char *stpcpy(char *__restrict, const char *__restrict__);

Why use two different forms for __restrict and __restrict__?
Any real reason to use __restrict__ at all?

It's used nowhere else in the kernel.

$ git grep -w -P '__restrict_{0,2}'
scripts/genksyms/keywords.c:    // According to rth, c99 defines "_Bool", __restrict", __restrict__", "restrict".  KAO
scripts/genksyms/keywords.c:    { "__restrict__", RESTRICT_KEYW },