linux-kernel - Re: [PATCH] static_call,x86: Robustify trampoline patching

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:   Tue, 2 Nov 2021 22:48:19 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     Ard Biesheuvel <ardb@...nel.org>
Cc:     Sami Tolvanen <samitolvanen@...gle.com>,
        Mark Rutland <mark.rutland@....com>, X86 ML <x86@...nel.org>,
        Kees Cook <keescook@...omium.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Nathan Chancellor <nathan@...nel.org>,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Sedat Dilek <sedat.dilek@...il.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        linux-hardening@...r.kernel.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        llvm@...ts.linux.dev, joao@...rdrivepizza.com
Subject: Re: [PATCH] static_call,x86: Robustify trampoline patching

On Tue, Nov 02, 2021 at 07:18:53PM +0100, Ard Biesheuvel wrote:

> > The range check isn't fundamental to CFI, having a check is the
> > important thing AFAIU.
> 
> Agreed. If the call site has a direct branch, it doesn't need the range check.

That, from the earlier email:

 | And have the actual indirect callsite look like:
 |
 |        # r11 - &foo
 |        ALTERNATIVE_2   "cs call __x86_indirect_thunk_r11",
 |                        "cs call __x86_indirect_cfi_deadbeef", X86_FEATURE_CFI
 |                        "cs call __x86_indirect_ibt_deadbeef", X86_FEATURE_IBT

So the callsite has a direct call to the hash-specific and cfi-type
specific thunk, which then does an (indirect) tail-call.

The CFI one does the hash check in the thunk and jumps to the function
proper, the IBT one on does it in the landing-pad.

The !CFI one ignore it all and simply does an indirect call (retpoline
aided or otherwise) to the function proper -- in which case we can free
all the thunks.