Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 06 Jan 2023 13:25:07 -0800
From:   "H. Peter Anvin" <hpa@zytor.com>
To:     Xin Li <xin3.li@intel.com>, linux-kernel@vger.kernel.org,
        x86@kernel.org, kvm@vger.kernel.org
CC:     tglx@linutronix.de, mingo@redhat.com, bp@alien8.de,
        dave.hansen@linux.intel.com, peterz@infradead.org,
        andrew.cooper3@citrix.com, seanjc@google.com, pbonzini@redhat.com,
        ravi.v.shankar@intel.com
Subject: =?US-ASCII?Q?Re=3A_=5BRFC_PATCH_v2_32/32=5D_x86/fred=3A_dis?= =?US-ASCII?Q?able_FRED_by_default_in_its_early_stage?=
User-Agent: K-9 Mail for Android
In-Reply-To: <20230106085617.17248-33-xin3.li@intel.com>
References: <20230106085617.17248-1-xin3.li@intel.com> <20230106085617.17248-33-xin3.li@intel.com>
Message-ID: <46772FCB-7313-47E9-ADCD-8BD556E8F265@zytor.com>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On January 6, 2023 12:56:17 AM PST, Xin Li <xin3=2Eli@intel=2Ecom> wrote:
>Disable FRED by default in its early stage=2E
>
>To enable FRED, a new kernel command line option "fred" needs to be added=
=2E
>
>Signed-off-by: Xin Li <xin3=2Eli@intel=2Ecom>
>---
> Documentation/admin-guide/kernel-parameters=2Etxt | 4 ++++
> arch/x86/kernel/cpu/common=2Ec                    | 3 +++
> 2 files changed, 7 insertions(+)
>
>diff --git a/Documentation/admin-guide/kernel-parameters=2Etxt b/Document=
ation/admin-guide/kernel-parameters=2Etxt
>index 42af9ca0127e=2E=2E0bc76d926dd4 100644
>--- a/Documentation/admin-guide/kernel-parameters=2Etxt
>+++ b/Documentation/admin-guide/kernel-parameters=2Etxt
>@@ -1506,6 +1506,10 @@
> 			Warning: use of this parameter will taint the kernel
> 			and may cause unknown problems=2E
>=20
>+	fred
>+			Forcefully enable flexible return and event delivery,
>+			which is otherwise disabled by default=2E
>+
> 	ftrace=3D[tracer]
> 			[FTRACE] will set and start the specified tracer
> 			as early as possible in order to facilitate early
>diff --git a/arch/x86/kernel/cpu/common=2Ec b/arch/x86/kernel/cpu/common=
=2Ec
>index 5de68356fe62=2E=2E1a160337ad41 100644
>--- a/arch/x86/kernel/cpu/common=2Ec
>+++ b/arch/x86/kernel/cpu/common=2Ec
>@@ -1434,6 +1434,9 @@ static void __init cpu_parse_early_param(void)
> 	char *argptr =3D arg, *opt;
> 	int arglen, taint =3D 0;
>=20
>+	if (!cmdline_find_option_bool(boot_command_line, "fred"))
>+		setup_clear_cpu_cap(X86_FEATURE_FRED);
>+
> #ifdef CONFIG_X86_32
> 	if (cmdline_find_option_bool(boot_command_line, "no387"))
> #ifdef CONFIG_MATH_EMULATION

Specifically, the intent is to enable it by default once hardware is widel=
y available=2E This allows e=2Eg=2E distros to compile in the support for o=
pt-in field testing before throwing the switch=2E

The motivation for this comes from the SMAP experience where the code was =
functional and correct when merged, but a change elsewhere in the kernel co=
de base broke it during the intervening period=2E