linux-kernel - Re: [PATCH v3 01/10] x86/cfi: Add warn option

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: <202502190941.CA60135@keescook>
Date: Wed, 19 Feb 2025 09:50:54 -0800
From: Kees Cook <kees@...nel.org>
To: Peter Zijlstra <peterz@...radead.org>
Cc: x86@...nel.org, linux-kernel@...r.kernel.org, alyssa.milburn@...el.com,
	scott.d.constable@...el.com, joao@...rdrivepizza.com,
	andrew.cooper3@...rix.com, jpoimboe@...nel.org,
	jose.marchesi@...cle.com, hjl.tools@...il.com,
	ndesaulniers@...gle.com, samitolvanen@...gle.com, nathan@...nel.org,
	ojeda@...nel.org, alexei.starovoitov@...il.com, mhiramat@...nel.org,
	jmill@....edu
Subject: Re: [PATCH v3 01/10] x86/cfi: Add warn option

On Wed, Feb 19, 2025 at 05:21:08PM +0100, Peter Zijlstra wrote:
> Rebuilding with CFI_PERMISSIVE toggled is such a pain, esp. since
> clang is so slow.

This seems too complex; report_cfi_failure() already has the fail/warn
logic test. I would have expected cfi_warn to take CONFIG_CFI_PERMISSIVE
as a default instead, like:

+bool cfi_warn __ro_after_init = IS_ENABLED(CONFIG_CFI_PERMISSIVE);

and then just replace report_cfi_failure()'s check of
CONFIG_CFI_PERMISSIVE with cfi_warn:

-        if (IS_ENABLED(CONFIG_CFI_PERMISSIVE)) {
+        if (cfi_warn) {

-Kees

(I do worry about data-only attacks going after page tables and flipping
pages to r/w and changing cfi_warn to 1, but that's probably on the same
order of difficulty as targeting the cfi handler function itself. Hmpf.)

-- 
Kees Cook