Re: [Patch v7 10/18] x86/speculation: Turn on or off STIBP according to a task's TIF_STIBP

From: Thomas Gleixner <tglx@linutronix.de>
To: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Jiri Kosina <jikos@kernel.org>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	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>, Waiman Long <longman9394@gmail.com>,
	Greg KH <gregkh@linuxfoundation.org>,
	Dave Stewart <david.c.stewart@intel.com>,
	linux-kernel@vger.kernel.org, x86@kernel.org,
	stable@vger.kernel.org
Subject: Re: [Patch v7 10/18] x86/speculation: Turn on or off STIBP according to a task's TIF_STIBP
Date: Wed, 21 Nov 2018 14:33:59 +0100 (CET)	[thread overview]
Message-ID: <alpine.DEB.2.21.1811211426470.1665@nanos.tec.linutronix.de> (raw)
In-Reply-To: <b6a7deb9f976680d5d96a2c1affc032604bb1879.1542758656.git.tim.c.chen@linux.intel.com>

On Tue, 20 Nov 2018, Tim Chen wrote:
> diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
> index 74bef48..48fcd46 100644
> --- a/arch/x86/kernel/process.c
> +++ b/arch/x86/kernel/process.c
> @@ -406,6 +406,8 @@ static __always_inline void spec_ctrl_update_msr(unsigned long tifn)
>  	if (static_cpu_has(X86_FEATURE_SSBD))
>  		msr |= ssbd_tif_to_spec_ctrl(tifn);
>  
> +	msr |= stibp_tif_to_spec_ctrl(tifn);

This is wrong. If STIBP is disabled, but the TIF flag is set to control
IBPB and this path is entered to handle a SSBD change, then this will set
STIBP as well.

> +
>  	wrmsrl(MSR_IA32_SPEC_CTRL, msr);
>  }
>  
> @@ -418,7 +420,17 @@ static __always_inline void spec_ctrl_update_msr(unsigned long tifn)
>  static __always_inline void __speculation_ctrl_update(unsigned long tifp,
>  						      unsigned long tifn)
>  {
> -	bool updmsr = false;
> +	bool updmsr;
> +
> +	/*
> +	 * Need STIBP defense against Spectre v2 attack
> +	 * if SMT is in use and enhanced IBRS is unsupported.
> +	 */
> +	if (static_cpu_has(X86_FEATURE_STIBP) && cpu_use_smt_and_hotplug &&
> +	    !static_cpu_has(X86_FEATURE_USE_IBRS_ENHANCED))
> +		updmsr = !!((tifp ^ tifn) & _TIF_SPEC_INDIR_BRANCH);

At the end of the series this has then two alternative patch bits and two
static keys. This can be done with a single static key which is controlled
from the bug code.

I'll send out a cleaned up version of that stuff later today.

Thanks,

	tglx