From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1751590AbdITDSc (ORCPT <rfc822;w@1wt.eu>);
        Tue, 19 Sep 2017 23:18:32 -0400
Received: from mail-pg0-f67.google.com ([74.125.83.67]:32780 "EHLO
        mail-pg0-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751457AbdITDSb (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Tue, 19 Sep 2017 23:18:31 -0400
X-Google-Smtp-Source: AOwi7QCYse5i2pW3IMMt3Rk3RJ3M6YrJlBKSCOig0SuAdvwty+NmGId0Ldquxyh8W8Tn9d3CvRFY3w==
Date: Tue, 19 Sep 2017 20:18:28 -0700
From: Eric Biggers <ebiggers3@gmail.com>
To: x86@kernel.org
Cc: linux-kernel@vger.kernel.org, kernel-hardening@lists.openwall.com,
        Andy Lutomirski <luto@kernel.org>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Dmitry Vyukov <dvyukov@google.com>, Fenghua Yu <fenghua.yu@intel.com>,
        Ingo Molnar <mingo@kernel.org>, Kevin Hao <haokexin@gmail.com>,
        Oleg Nesterov <oleg@redhat.com>, Wanpeng Li <wanpeng.li@hotmail.com>,
        Yu-cheng Yu <yu-cheng.yu@intel.com>,
        Michael Halcrow <mhalcrow@google.com>,
        Eric Biggers <ebiggers@google.com>
Subject: Re: [PATCH v2 3/3] x86/fpu: reinitialize FPU registers if restoring
 FPU state fails
Message-ID: <20170920031828.GA537@zzz.localdomain>
References: <20170920004434.35308-1-ebiggers3@gmail.com>
 <20170920004434.35308-4-ebiggers3@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170920004434.35308-4-ebiggers3@gmail.com>
User-Agent: Mutt/1.9.0 (2017-09-02)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, Sep 19, 2017 at 05:44:34PM -0700, Eric Biggers wrote:
> +/*
> + * We should never get here because the fpregs_state stored in 'struct fpu'
> + * should always be readable and contain a valid FPU state.  However, past bugs
> + * have allowed userspace to set reserved bits in the XSAVE area using
> + * PTRACE_SETREGSET or sys_rt_sigreturn().  These caused XRSTOR to fail when
> + * switching to the task, leaking the FPU registers of the task previously
> + * executing on the CPU.  Mitigate this class of vulnerability by restoring from
> + * the initial state (essentially, zeroing out all the FPU registers) if we
> + * can't restore from the task's FPU state.
> + */
> +void __handle_bad_fpstate(union fpregs_state *fpstate, u64 mask)
> +{
> +	WARN_ONCE(1, "Bad FPU state detected, reinitializing FPU registers");
> +	____copy_kernel_to_fpregs(&init_fpstate, mask);
> +}

Sorry, a small oversight here --- this needs to be exported to modules, since
kvm can be built as a module, and kvm uses __copy_kernel_to_fpregs().

Eric