Re: [Patch v3 03/13] x86/speculation: Add static key for Enhanced IBRS

From: Waiman Long <longman9394@gmail.com>
To: Tim Chen <tim.c.chen@linux.intel.com>,
	Jiri Kosina <jikos@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>
Cc: Tom Lendacky <thomas.lendacky@amd.com>,
	Ingo Molnar <mingo@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	Andrea Arcangeli <aarcange@redhat.com>,
	David Woodhouse <dwmw@amazon.co.uk>,
	Andi Kleen <ak@linux.intel.com>,
	Dave Hansen <dave.hansen@intel.com>,
	Casey Schaufler <casey.schaufler@intel.com>,
	Asit Mallick <asit.k.mallick@intel.com>,
	Arjan van de Ven <arjan@linux.intel.com>,
	Jon Masters <jcm@redhat.com>,
	linux-kernel@vger.kernel.org, x86@kernel.org
Subject: Re: [Patch v3 03/13] x86/speculation: Add static key for Enhanced IBRS
Date: Fri, 26 Oct 2018 12:58:16 -0400	[thread overview]
Message-ID: <b5b6f492-8bf0-7fcc-eb4f-3016b4d46c08@gmail.com> (raw)
In-Reply-To: <d62b81e9390c4e404ede02139f7a29327dfe1921.1539798901.git.tim.c.chen@linux.intel.com>

On 10/17/2018 01:59 PM, Tim Chen wrote:
> Add static key to indicate whether we are using Enhanced IBRS to mitigate
> Spectre v2.  This will be used in later patches to disengage STIBP code
> for Spectre v2 mitigation as STIBP is not needed when Enhanced IBRS is
> in use.
>
> Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
> ---
>  arch/x86/include/asm/nospec-branch.h | 3 +++
>  arch/x86/kernel/cpu/bugs.c           | 4 ++++
>  2 files changed, 7 insertions(+)
>
> diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
> index fd2a8c1..d57e84e 100644
> --- a/arch/x86/include/asm/nospec-branch.h
> +++ b/arch/x86/include/asm/nospec-branch.h
> @@ -3,6 +3,7 @@
>  #ifndef _ASM_X86_NOSPEC_BRANCH_H_
>  #define _ASM_X86_NOSPEC_BRANCH_H_
>  
> +#include <linux/static_key.h>
>  #include <asm/alternative.h>
>  #include <asm/alternative-asm.h>
>  #include <asm/cpufeatures.h>
> @@ -228,6 +229,8 @@ enum ssb_mitigation {
>  extern char __indirect_thunk_start[];
>  extern char __indirect_thunk_end[];
>  
> +DECLARE_STATIC_KEY_FALSE(spectre_v2_enhanced_ibrs);
> +
>  /*
>   * On VMEXIT we must ensure that no RSB predictions learned in the guest
>   * can be followed in the host, by overwriting the RSB completely. Both
> diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
> index b2f6b8b..2fc7b4e 100644
> --- a/arch/x86/kernel/cpu/bugs.c
> +++ b/arch/x86/kernel/cpu/bugs.c
> @@ -142,6 +142,9 @@ static const char *spectre_v2_strings[] = {
>  	[SPECTRE_V2_IBRS_ENHANCED]		= "Mitigation: Enhanced IBRS",
>  };
>  
> +DEFINE_STATIC_KEY_FALSE(spectre_v2_enhanced_ibrs);
> +EXPORT_SYMBOL(spectre_v2_enhanced_ibrs);
> +
>  #undef pr_fmt
>  #define pr_fmt(fmt)     "Spectre V2 : " fmt
>  
> @@ -386,6 +389,7 @@ static void __init spectre_v2_select_mitigation(void)
>  			/* Force it so VMEXIT will restore correctly */
>  			x86_spec_ctrl_base |= SPEC_CTRL_IBRS;
>  			wrmsrl(MSR_IA32_SPEC_CTRL, x86_spec_ctrl_base);
> +			static_branch_enable(&spectre_v2_enhanced_ibrs);
>  			goto specv2_set_mode;
>  		}
>  		if (IS_ENABLED(CONFIG_RETPOLINE))

Why you need a static key for enhanced IBRS? It is supposed to be set at
boot time and never get changed after that. It will be easier to use a
feature bit for that instead. We usually use static key when the value
can be changed at run time.

Cheers,
Longman