Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 05 Dec 2019 19:15:03 +0000
To:     "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
From:   Krzysztof Piecuch <piecuch@protonmail.com>
Cc:     "tglx@linutronix.de" <tglx@linutronix.de>,
        "mingo@redhat.com" <mingo@redhat.com>,
        "bp@alien8.de" <bp@alien8.de>, "hpa@zytor.com" <hpa@zytor.com>,
        "x86@kernel.org" <x86@kernel.org>,
        "viresh.kumar@linaro.org" <viresh.kumar@linaro.org>,
        "rafael.j.wysocki@intel.com" <rafael.j.wysocki@intel.com>,
        "drake@endlessm.com" <drake@endlessm.com>,
        "piecuch@protonmail.com" <piecuch@protonmail.com>,
        "malat@debian.org" <malat@debian.org>,
        "peterz@infradead.org" <peterz@infradead.org>,
        "juri.lelli@redhat.com" <juri.lelli@redhat.com>
Reply-To: Krzysztof Piecuch <piecuch@protonmail.com>
Subject: [PATCH] x86/tsc: Don't use cpuid 0x16 leaf to determine cpu speed.
Message-ID: <SoluZg51N39Rx0tDCSJFbEvvgMrDnJ_g0RdRdN5mtCfag4GahIOPfok7UbkyeO5Qpl3wUHp8H8y73JtClcZvr1ARSIOBIFQAne0Z712el8M=@protonmail.com>
Feedback-ID: krphKiiPlx_XKIryTSpdJ_XtBwogkHXWA-Us-PsTeaBSrzOTAKWxwbFkseT4Z85b_7PMRvSnq3Ah7f9INXrOMw==:Ext:ProtonMail
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

This patch corrects tsc drift on systems with changed base clock frequency
(e.g. overclocking).

We can't use 0x16 cpu leaf as it's documented as "not reflecting actual
values" and is supposed to be used only as a mean to determine "processor
brand string and for determining the appropriate range to use when
displaying processor information e.g. frequency history graphs".

Signed-off-by: Krzysztof Piecuch <piecuch@protonmail.com>
---
 arch/x86/kernel/tsc.c | 21 ++++-----------------
 1 file changed, 4 insertions(+), 17 deletions(-)

diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index 7e322e2daaf5..fc9a000a814c 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -641,29 +641,16 @@ unsigned long native_calibrate_tsc(void)
 =09=09=09boot_cpu_data.x86_model =3D=3D INTEL_FAM6_ATOM_GOLDMONT_D)
 =09=09crystal_khz =3D 25000;

+=09if (crystal_khz =3D=3D 0)
+=09=09return 0;
+
 =09/*
 =09 * TSC frequency reported directly by CPUID is a "hardware reported"
 =09 * frequency and is the most accurate one so far we have. This
 =09 * is considered a known frequency.
 =09 */
-=09if (crystal_khz !=3D 0)
-=09=09setup_force_cpu_cap(X86_FEATURE_TSC_KNOWN_FREQ);
-
-=09/*
-=09 * Some Intel SoCs like Skylake and Kabylake don't report the crystal
-=09 * clock, but we can easily calculate it to a high degree of accuracy
-=09 * by considering the crystal ratio and the CPU speed.
-=09 */
-=09if (crystal_khz =3D=3D 0 && boot_cpu_data.cpuid_level >=3D 0x16) {
-=09=09unsigned int eax_base_mhz, ebx, ecx, edx;
-
-=09=09cpuid(0x16, &eax_base_mhz, &ebx, &ecx, &edx);
-=09=09crystal_khz =3D eax_base_mhz * 1000 *
-=09=09=09eax_denominator / ebx_numerator;
-=09}

-=09if (crystal_khz =3D=3D 0)
-=09=09return 0;
+=09setup_force_cpu_cap(X86_FEATURE_TSC_KNOWN_FREQ);

 =09/*
 =09 * For Atom SoCs TSC is the only reliable clocksource.
--
2.17.1