linux-kernel - [PATCH 1/2] x86/tsc: use CPUID.0x16 to calculate missing crystal frequency

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20190422101526.28415-1-drake@endlessm.com>
Date:   Mon, 22 Apr 2019 18:15:25 +0800
From:   Daniel Drake <drake@...lessm.com>
To:     tglx@...utronix.de, mingo@...hat.com, bp@...en8.de
Cc:     hpa@...or.com, x86@...nel.org, linux-kernel@...r.kernel.org,
        len.brown@...el.com, rafael.j.wysocki@...el.com, linux@...lessm.com
Subject: [PATCH 1/2] x86/tsc: use CPUID.0x16 to calculate missing crystal frequency

native_calibrate_tsc() had a hardcoded table of Intel CPU families
and crystal clock, but we have found that it is possible to
calculate the crystal clock speed, and this is preferred to a hardcoded
table.

Where crystal clock frequency was not reported by CPUID.0x15,
use CPUID.0x16 data to calculate the crystal clock.

Using CPUID dump data from http://instlatx64.atw.hu/, the calculation
results can be seen to be sufficiently close to the previously hardcoded
values:
SKYLAKE_MOBILE: 24074074 Hz
SKYLAKE_DESKTOP: 23913043 Hz
KABYLAKE_MOBILE: 23893805 Hz
KABYLAKE_DESKTOP: 24050632 Hz
GOLDMONT: 19.2MHz crystal clock correctly reported by CPUID.0x15

Additionally, crystal clock frequency for platforms that were missing
from the list (e.g. SKYLAKE_X) will now be provided.

GOLDMONT_X was left as a hardcoded value, as the CPUID data on that site
indicates that the hardware does not report crystal clock nor CPUID.0x16
data.

Going forward, Intel have hopefully now started providing crystal
frequency in CPUID.0x15. At least ApolloLake, GeminiLake and CannonLake
CPUs all provide the relevant data directly.

Link: https://lkml.kernel.org/r/20190419083533.32388-1-drake@endlessm.com
Suggested-by: Thomas Gleixner <tglx@...utronix.de>
Signed-off-by: Daniel Drake <drake@...lessm.com>
---
 arch/x86/kernel/tsc.c | 33 ++++++++++++++++++---------------
 1 file changed, 18 insertions(+), 15 deletions(-)

diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index 3fae23834069..3971c837584a 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -629,23 +629,26 @@ unsigned long native_calibrate_tsc(void)
 
 	crystal_khz = ecx_hz / 1000;
 
-	if (crystal_khz == 0) {
-		switch (boot_cpu_data.x86_model) {
-		case INTEL_FAM6_SKYLAKE_MOBILE:
-		case INTEL_FAM6_SKYLAKE_DESKTOP:
-		case INTEL_FAM6_KABYLAKE_MOBILE:
-		case INTEL_FAM6_KABYLAKE_DESKTOP:
-			crystal_khz = 24000;	/* 24.0 MHz */
-			break;
-		case INTEL_FAM6_ATOM_GOLDMONT_X:
-			crystal_khz = 25000;	/* 25.0 MHz */
-			break;
-		case INTEL_FAM6_ATOM_GOLDMONT:
-			crystal_khz = 19200;	/* 19.2 MHz */
-			break;
-		}
+	/*
+	 * Some Intel SoCs like Skylake and Kabylake don't report the crystal
+	 * clock, but we can easily calculate it by considering the crystal
+	 * ratio and the CPU speed.
+	 */
+	if (crystal_khz == 0 && boot_cpu_data.cpuid_level >= 0x16) {
+		unsigned int eax_base_mhz, ebx, ecx, edx;
+		cpuid(0x16, &eax_base_mhz, &ebx, &ecx, &edx);
+		crystal_khz = eax_base_mhz * 1000 * \
+			eax_denominator / ebx_numerator;
 	}
 
+	/*
+	 * Denverton SoCs don't report crystal clock, and also don't support
+	 * CPUID.0x16, so hardcode the 25MHz crystal clock.
+	 */
+	if (crystal_khz == 0 &&
+			boot_cpu_data.x86_model == INTEL_FAM6_ATOM_GOLDMONT_X)
+		crystal_khz = 25000;
+
 	if (crystal_khz == 0)
 		return 0;
 	/*
-- 
2.19.1