linux-kernel - Re: [PATCH v3 16/16] objtool: Validate kCFI calls

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: <20250714112122.GL1613633@noisy.programming.kicks-ass.net>
Date: Mon, 14 Jul 2025 13:21:22 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: x86@...nel.org
Cc: kys@...rosoft.com, haiyangz@...rosoft.com, wei.liu@...nel.org,
	decui@...rosoft.com, tglx@...utronix.de, mingo@...hat.com,
	bp@...en8.de, dave.hansen@...ux.intel.com, hpa@...or.com,
	seanjc@...gle.com, pbonzini@...hat.com, ardb@...nel.org,
	kees@...nel.org, Arnd Bergmann <arnd@...db.de>,
	gregkh@...uxfoundation.org, jpoimboe@...nel.org,
	linux-hyperv@...r.kernel.org, linux-kernel@...r.kernel.org,
	kvm@...r.kernel.org, linux-efi@...r.kernel.org,
	samitolvanen@...gle.com, ojeda@...nel.org
Subject: Re: [PATCH v3 16/16] objtool: Validate kCFI calls

On Mon, Jul 14, 2025 at 12:49:19PM +0200, Peter Zijlstra wrote:
> On Mon, Jul 14, 2025 at 12:20:27PM +0200, Peter Zijlstra wrote:
> 
> > --- a/arch/x86/platform/efi/efi_stub_64.S
> > +++ b/arch/x86/platform/efi/efi_stub_64.S
> > @@ -11,6 +11,10 @@
> >  #include <asm/nospec-branch.h>
> >  
> >  SYM_FUNC_START(__efi_call)
> > +	/*
> > +	 * The EFI code doesn't have any CFI, annotate away the CFI violation.
> > +	 */
> > +	ANNOTATE_NOCFI_SYM
> >  	pushq %rbp
> >  	movq %rsp, %rbp
> >  	and $~0xf, %rsp
> 
> FWIW, we should probably do something like this as well.
> 
> ---
> 
> --- a/arch/x86/kernel/cpu/common.c
> +++ b/arch/x86/kernel/cpu/common.c
> @@ -562,6 +562,13 @@ __noendbr u64 ibt_save(bool disable)
>  {
>  	u64 msr = 0;
>  
> +	/*
> +	 * Firmware code will not provide the same level of
> +	 * control-flow-integriry. Taint the kernel to let the user know.
> +	 */
> +	if (disable || (IS_ENABLED(CONFIG_CFI_CLANG) && cfi_mode != CFI_OFF))
> +		add_taint(TAINT_CFI, LOCKDEP_STILL_OK);

Or perhaps:

	WARN_TAINT_ONCE(disable || IS_ENABLED(CONFIG_CFI_CLANG) && cfi_mode != CFI_OFF),
			TAINT_CFI, "Firmware has weaker CFI");

> +
>  	if (cpu_feature_enabled(X86_FEATURE_IBT)) {
>  		rdmsrq(MSR_IA32_S_CET, msr);
>  		if (disable)
> --- a/include/linux/panic.h
> +++ b/include/linux/panic.h
> @@ -73,7 +73,8 @@ static inline void set_arch_panic_timeou
>  #define TAINT_RANDSTRUCT		17
>  #define TAINT_TEST			18
>  #define TAINT_FWCTL			19
> -#define TAINT_FLAGS_COUNT		20
> +#define TAINT_CFI			20
> +#define TAINT_FLAGS_COUNT		21
>  #define TAINT_FLAGS_MAX			((1UL << TAINT_FLAGS_COUNT) - 1)
>  
>  struct taint_flag {