From:   Joshua Scott <joshua.scott@alliedtelesis.co.nz>
To:     Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        linux-kernel@vger.kernel.org
Cc:     x86@kernel.org, Joshua Scott <joshua.scott@alliedtelesis.co.nz>
Subject: [PATCH] x86/cpu/amd: Work-around for family 17h models 00h-0fh erratum 1019
Date:   Wed,  8 Feb 2023 14:14:52 +1300
Message-Id: <20230208011452.5000-1-joshua.scott@alliedtelesis.co.nz>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Errata 1019 can result in a false-positive micro-op queue parity error,
which could result in a system reset.

Set the appropriate mask bit to disable detection of the affected error.

Signed-off-by: Joshua Scott <joshua.scott@alliedtelesis.co.nz>
---
 arch/x86/include/asm/msr-index.h |  2 ++
 arch/x86/kernel/cpu/amd.c        | 12 ++++++++++++
 2 files changed, 14 insertions(+)

diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-=
index.h
index d3fe82c5d6b6..3c8b95829838 100644
--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -527,6 +527,8 @@
 #define MSR_AMD64_OSVW_STATUS		0xc0010141
 #define MSR_AMD_PPIN_CTL		0xc00102f0
 #define MSR_AMD_PPIN			0xc00102f1
+#define MSR_AMD64_DE_MC_CTRL_MASK	0xc0010403
+#define MSR_AMD64_DE_MC_CTRL_MASK_UOPQ_BIT	 3
 #define MSR_AMD64_CPUID_FN_1		0xc0011004
 #define MSR_AMD64_LS_CFG		0xc0011020
 #define MSR_AMD64_DC_CFG		0xc0011022
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index f769d6d08b43..c329be80930f 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -29,6 +29,7 @@
=20
 static const int amd_erratum_383[];
 static const int amd_erratum_400[];
+static const int amd_erratum_1019[];
 static const int amd_erratum_1054[];
 static bool cpu_has_amd_erratum(struct cpuinfo_x86 *cpu, const int *erra=
tum);
=20
@@ -995,6 +996,13 @@ static void init_amd(struct cpuinfo_x86 *c)
 	    !cpu_has_amd_erratum(c, amd_erratum_1054))
 		msr_set_bit(MSR_K7_HWCR, MSR_K7_HWCR_IRPERF_EN_BIT);
=20
+	/*
+	 *  Turn off detection of micro-op queue parity errors on
+	 *  machines susceptible to erratum #1019.
+	 */
+	if (cpu_has_amd_erratum(c, amd_erratum_1019))
+		msr_set_bit(MSR_AMD64_DE_MC_CTRL_MASK, MSR_AMD64_DE_MC_CTRL_MASK_UOPQ_=
BIT);
+
 	check_null_seg_clears_base(c);
 }
=20
@@ -1123,6 +1131,10 @@ static const int amd_erratum_400[] =3D
 static const int amd_erratum_383[] =3D
 	AMD_OSVW_ERRATUM(3, AMD_MODEL_RANGE(0x10, 0, 0, 0xff, 0xf));
=20
+/* #1019: A Thread May Spuriously Report a Micro-Op Queue Parity Error *=
/
+static const int amd_erratum_1019[] =3D
+	AMD_LEGACY_ERRATUM(AMD_MODEL_RANGE(0x17, 0, 0, 0x0f, 0xf));
+
 /* #1054: Instructions Retired Performance Counter May Be Inaccurate */
 static const int amd_erratum_1054[] =3D
 	AMD_LEGACY_ERRATUM(AMD_MODEL_RANGE(0x17, 0, 0, 0x2f, 0xf));
--=20
2.39.1