linux-hardening - Re: [RFC PATCH 00/11] Kernel FineIBT Support

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: <e57e01bb682da9443d0ad6e29eeb69dac22e2ffa.camel@intel.com>
Date:   Wed, 20 Apr 2022 23:34:20 +0000
From:   "Edgecombe, Rick P" <rick.p.edgecombe@...el.com>
To:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-hardening@...r.kernel.org" <linux-hardening@...r.kernel.org>,
        "joao@...rdrivepizza.com" <joao@...rdrivepizza.com>
CC:     "peterz@...radead.org" <peterz@...radead.org>,
        "keescook@...omium.org" <keescook@...omium.org>,
        "hjl.tools@...il.com" <hjl.tools@...il.com>,
        "Cooper, Andrew" <andrew.cooper3@...rix.com>,
        "Poimboe, Josh" <jpoimboe@...hat.com>,
        "samitolvanen@...gle.com" <samitolvanen@...gle.com>,
        "mark.rutland@....com" <mark.rutland@....com>,
        "alyssa.milburn@...ux.intel.com" <alyssa.milburn@...ux.intel.com>,
        "gabriel.gomes@...ux.intel.com" <gabriel.gomes@...ux.intel.com>,
        "ndesaulniers@...gle.com" <ndesaulniers@...gle.com>
Subject: Re: [RFC PATCH 00/11] Kernel FineIBT Support

On Tue, 2022-04-19 at 17:42 -0700, joao@...rdrivepizza.com wrote:
> A debatable point is the fact that on FineIBT the checks are made on
> the callee
> side. On a quick look, this seems to be cool because it allows strict
> reachability refinement of more security-critical functions (like
> hardware
> feature disabling ones) while still allowing other less critical
> functions to be
> relaxed/coarse-grained; under caller-side checks, if one single
> function is
> required to be relaxed, this leads into an indirect call instruction
> being
> relaxed, then becoming a branch capable of reaching all the functions
> in the
> executable address space, including those considered security-
> critical. Inputs
> and opinions on this are very welcome, as there are other
> perspectives about
> this I might be missing.

One minor point: In the course IBT implementation there are places in
the kernel where IBT is toggled because of missing endbranches (calling
into BIOS). So for caller checked solutions, these calls would probably
need to be annotated or something such that caller checks were not
generated.

I haven't been following kCFI, so apologies if this is already handled
somehow.