linux-kernel - Re: [PATCH 06/10] x86/fpu: Log XSAVE disablement consistently

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: <d67648a0-925c-433a-a33d-1cdba02cacef@intel.com>
Date: Wed, 16 Apr 2025 09:56:14 -0700
From: Sohil Mehta <sohil.mehta@...el.com>
To: "Chang S. Bae" <chang.seok.bae@...el.com>, <mingo@...hat.com>
CC: <linux-kernel@...r.kernel.org>, <x86@...nel.org>, <tglx@...utronix.de>,
	<bp@...en8.de>, <dave.hansen@...ux.intel.com>
Subject: Re: [PATCH 06/10] x86/fpu: Log XSAVE disablement consistently

On 4/15/2025 7:16 PM, Chang S. Bae wrote:
> Not all paths that lead to fpu__init_disable_system_xstate() currently
> emit a message indicating that XSAVE has been disabled. Move the print
> statement into the function to ensure the message in all cases.
> 
> Suggested-by: Dave Hansen <dave.hansen@...ux.intel.com>
> Signed-off-by: Chang S. Bae <chang.seok.bae@...el.com>
> Link: https://lore.kernel.org/lkml/d6d19e39-2749-4d45-aeab-a209a0ecba17@intel.com
> ---
> New patch for following up patch 3.
> ---
>  arch/x86/kernel/fpu/xstate.c | 8 +++++---
>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
> index 2ac1fc182273..8b14c9d3a1df 100644
> --- a/arch/x86/kernel/fpu/xstate.c
> +++ b/arch/x86/kernel/fpu/xstate.c
> @@ -751,6 +751,8 @@ static int __init init_xstate_size(void)
>   */
>  static void __init fpu__init_disable_system_xstate(unsigned int legacy_size)
>  {
> +	pr_info("x86/fpu: XSAVE disabled\n");
> +

There is a mix of pr_info(), pr_warn() and pr_err() to log these related
messages. Would it be useful to make the log level consistent in this
patch or a follow-up?

I think new the "XSAVE disabled" print should be a pr_warn() at least.

>  	fpu_kernel_cfg.max_features = 0;
>  	cr4_clear_bits(X86_CR4_OSXSAVE);
>  	setup_clear_cpu_cap(X86_FEATURE_XSAVE);
> @@ -821,7 +823,7 @@ void __init fpu__init_system_xstate(unsigned int legacy_size)
>  		 * This is a problematic CPU configuration where two
>  		 * conflicting state components are both enumerated.
>  		 */
> -		pr_err("x86/fpu: Both APX/MPX present in the CPU's xstate features: 0x%llx, disabling XSAVE.\n",
> +		pr_err("x86/fpu: Both APX/MPX present in the CPU's xstate features: 0x%llx.\n",
>  		       fpu_kernel_cfg.max_features);
>  		goto out_disable;
>  	}
> @@ -900,7 +902,7 @@ void __init fpu__init_system_xstate(unsigned int legacy_size)
>  	init_fpstate.xfeatures		= fpu_kernel_cfg.default_features;
>  
>  	if (init_fpstate.size > sizeof(init_fpstate.regs)) {
> -		pr_warn("x86/fpu: init_fpstate buffer too small (%zu < %d), disabling XSAVE\n",
> +		pr_warn("x86/fpu: init_fpstate buffer too small (%zu < %d)\n",
>  			sizeof(init_fpstate.regs), init_fpstate.size);
>  		goto out_disable;
>  	}
> @@ -912,7 +914,7 @@ void __init fpu__init_system_xstate(unsigned int legacy_size)
>  	 * xfeatures mask.
>  	 */
>  	if (xfeatures != fpu_kernel_cfg.max_features) {
> -		pr_err("x86/fpu: xfeatures modified from 0x%016llx to 0x%016llx during init, disabling XSAVE\n",
> +		pr_err("x86/fpu: xfeatures modified from 0x%016llx to 0x%016llx during init\n",
>  		       xfeatures, fpu_kernel_cfg.max_features);
>  		goto out_disable;
>  	}