linux-kernel - Re: [PATCH v8 05/14] x86: conditionally place regular ASM functions into separate sections

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]

Date:   Fri, 10 Dec 2021 12:01:02 +0100
From:   Alexander Lobakin <alexandr.lobakin@...el.com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Alexander Lobakin <alexandr.lobakin@...el.com>,
        linux-hardening@...r.kernel.org, x86@...nel.org,
        Jesse Brandeburg <jesse.brandeburg@...el.com>,
        Kristen Carlson Accardi <kristen@...ux.intel.com>,
        Kees Cook <keescook@...omium.org>,
        Miklos Szeredi <miklos@...redi.hu>,
        Ard Biesheuvel <ardb@...nel.org>,
        Tony Luck <tony.luck@...el.com>,
        Bruce Schlobohm <bruce.schlobohm@...el.com>,
        Jessica Yu <jeyu@...nel.org>,
        kernel test robot <lkp@...el.com>,
        Miroslav Benes <mbenes@...e.cz>,
        Evgenii Shatokhin <eshatokhin@...tuozzo.com>,
        Jonathan Corbet <corbet@....net>,
        Masahiro Yamada <masahiroy@...nel.org>,
        Michal Marek <michal.lkml@...kovi.net>,
        "H . J . Lu" <hjl.tools@...il.com>,
        Nicolas Pitre <nico@...xnic.net>,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Herbert Xu <herbert@...dor.apana.org.au>,
        "David S. Miller" <davem@...emloft.net>,
        Thomas Gleixner <tglx@...utronix.de>,
        Will Deacon <will@...nel.org>, Ingo Molnar <mingo@...hat.com>,
        Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        "H. Peter Anvin" <hpa@...or.com>,
        Andy Lutomirski <luto@...nel.org>,
        Arnd Bergmann <arnd@...db.de>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Nathan Chancellor <nathan@...nel.org>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Marios Pomonis <pomonis@...gle.com>,
        Sami Tolvanen <samitolvanen@...gle.com>,
        linux-kernel@...r.kernel.org, linux-kbuild@...r.kernel.org,
        linux-arch@...r.kernel.org, live-patching@...r.kernel.org,
        llvm@...ts.linux.dev
Subject: Re: [PATCH v8 05/14] x86: conditionally place regular ASM functions into separate sections

From: Peter Zijlstra <peterz@...radead.org>
Date: Fri, 3 Dec 2021 10:44:10 +0100

> On Thu, Dec 02, 2021 at 11:32:05PM +0100, Alexander Lobakin wrote:
> > Use the newly introduces macros to create unique separate sections
> > for (almost) every "regular" ASM function (i.e. for those which
> > aren't explicitly put into a specific one).
> > There should be no leftovers as input .text will be size-asserted
> > in the LD script generated for FG-KASLR.
> 
> *groan*...
> 
> Please, can't we do something like:
> 
> #define SYM_PUSH_SECTION(name)	\
> .if section == .text		\
> .push_section .text.##name	\
> .else				\
> .push_section .text		\
> .endif

This condition

.pushsection .text
.if section == .text
# do something
.endif
.popsection

doesn't really works. `do something` doesn't happen.
This works only when

.pushsection .text
.equ section, .text

but it's not really okayish I'd say to find all .{,push}section
occurences and replace them with a macro (which would also do .equ).

I don't really know how %S with --sectname-subst should help me as

.if %S == .text
# do something
.endif

doesn't work at all (syntax error) -- and it shouldn't, %S is
supposed to work only inside .{,push}section directives.

I could do unconditional

.pushsection %S.##name
                ^^^^^^ function name

but this would involve changing LDS scripts (and vmlinux.lds.h) to
let's say replace *(.noinstr.text) with *(.noinstr.text*).

So I hope there is a way to get current section name? If not, then
the last option is the least harmful I suppose.
At least not as harmful as current approach with alternative macros,
far from it lol.

> 
> #define SYM_POP_SECTION()	\
> .pop_section
> 
> and wrap that inside the existing SYM_FUNC_START*() SYM_FUNC_END()
> macros.

Thanks,
Al