From 7e52a03646af6fb45dd37c5375607858d65e9364 Mon Sep 17 00:00:00 2001 From: Thor Thayer Date: Fri, 17 Apr 2015 17:16:14 -0500 Subject: [PATCH 01/28] EDAC, altera: Do not build it as a module The SDRAM EDAC requires SDRAM configuration/initialization before SDRAM is accessed (in the preloader) and therefore before Linux is loaded. Having a module compile is not desired so force to be built into kernel. Signed-off-by: Thor Thayer Cc: mchehab@osg.samsung.com Cc: Takashi Iwai Link: http://lkml.kernel.org/r/1429308974-26380-3-git-send-email-tthayer@opensource.altera.com Signed-off-by: Borislav Petkov --- drivers/edac/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index cb59619df23f..2d2530cdf99d 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -377,8 +377,8 @@ config EDAC_OCTEON_PCI Cavium Octeon family of SOCs. config EDAC_ALTERA_MC - tristate "Altera SDRAM Memory Controller EDAC" - depends on EDAC_MM_EDAC && ARCH_SOCFPGA + bool "Altera SDRAM Memory Controller EDAC" + depends on EDAC_MM_EDAC=y && ARCH_SOCFPGA help Support for error detection and correction on the Altera SDRAM memory controller. Note that the From 30c7469ba6f5eccf33ac33ec7cb19414f50cf3bf Mon Sep 17 00:00:00 2001 From: Johannes Thumshirn Date: Thu, 7 May 2015 10:21:24 +0200 Subject: [PATCH 02/28] MAINTAINERS: Change Johannes Thumshirn's email address Change my email address to something that is not associated with my current or future employer. Suggested-by: Borislav Petkov Signed-off-by: Johannes Thumshirn Cc: linux-edac Cc: linux-watchdog@vger.kernel.org Cc: Mauro Carvalho Chehab Cc: Wim Van Sebroeck Cc: Greg Kroah-Hartman Link: http://lkml.kernel.org/r/1430986884-8760-1-git-send-email-jthumshirn@suse.de Signed-off-by: Borislav Petkov --- MAINTAINERS | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 2e5bbc0d68b2..be68a7156f13 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3708,7 +3708,7 @@ S: Maintained F: drivers/edac/ie31200_edac.c EDAC-MPC85XX -M: Johannes Thumshirn +M: Johannes Thumshirn L: linux-edac@vger.kernel.org W: bluesmoke.sourceforge.net S: Maintained @@ -6388,14 +6388,14 @@ F: include/linux/mtd/ F: include/uapi/mtd/ MEN A21 WATCHDOG DRIVER -M: Johannes Thumshirn +M: Johannes Thumshirn L: linux-watchdog@vger.kernel.org -S: Supported +S: Maintained F: drivers/watchdog/mena21_wdt.c MEN CHAMELEON BUS (mcb) -M: Johannes Thumshirn -S: Supported +M: Johannes Thumshirn +S: Maintained F: drivers/mcb/ F: include/linux/mcb.h From b01aec9b2c7d32f17a37553df63efa9f7c0fdaa0 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 21 May 2015 19:59:31 +0200 Subject: [PATCH 03/28] EDAC: Cleanup atomic_scrub mess So first of all, this atomic_scrub() function's naming is bad. It looks like an atomic_t helper. Change it to edac_atomic_scrub(). The bigger problem is that this function is arch-specific and every new arch which doesn't necessarily need that functionality still needs to define it, otherwise EDAC doesn't compile. So instead of doing that and including arch-specific headers, have each arch define an EDAC_ATOMIC_SCRUB symbol which can be used in edac_mc.c for ifdeffery. Much cleaner. And we already are doing this with another symbol - EDAC_SUPPORT. This is also much cleaner than having CONFIG_EDAC enumerate all the arches which need/have EDAC support and drivers. This way I can kill the useless edac.h header in tile too. Acked-by: Ralf Baechle Acked-by: Michael Ellerman Acked-by: Chris Metcalf Acked-by: Ingo Molnar Acked-by: Russell King Cc: Benjamin Herrenschmidt Cc: Doug Thompson Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac@vger.kernel.org Cc: linux-kernel@vger.kernel.org Cc: linux-mips@linux-mips.org Cc: linuxppc-dev@lists.ozlabs.org Cc: "Maciej W. Rozycki" Cc: Markos Chandras Cc: Mauro Carvalho Chehab Cc: Paul Mackerras Cc: "Steven J. Hill" Cc: x86@kernel.org Signed-off-by: Borislav Petkov --- arch/arm/Kconfig | 2 ++ arch/arm/include/asm/edac.h | 5 +++-- arch/mips/Kconfig | 1 + arch/mips/include/asm/edac.h | 4 ++-- arch/powerpc/Kconfig | 2 ++ arch/powerpc/include/asm/edac.h | 4 ++-- arch/tile/Kconfig | 1 + arch/tile/include/asm/edac.h | 29 ----------------------------- arch/x86/Kconfig | 2 ++ arch/x86/include/asm/edac.h | 2 +- drivers/edac/Kconfig | 7 ++++--- drivers/edac/edac_mc.c | 9 +++++++-- 12 files changed, 27 insertions(+), 41 deletions(-) delete mode 100644 arch/tile/include/asm/edac.h diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 45df48ba0b12..325d6f3a596a 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -15,6 +15,8 @@ config ARM select CLONE_BACKWARDS select CPU_PM if (SUSPEND || CPU_IDLE) select DCACHE_WORD_ACCESS if HAVE_EFFICIENT_UNALIGNED_ACCESS + select EDAC_SUPPORT + select EDAC_ATOMIC_SCRUB select GENERIC_ALLOCATOR select GENERIC_ATOMIC64 if (CPU_V7M || CPU_V6 || !CPU_32v6K || !AEABI) select GENERIC_CLOCKEVENTS_BROADCAST if SMP diff --git a/arch/arm/include/asm/edac.h b/arch/arm/include/asm/edac.h index 0df7a2c1fc3d..5189fa819b60 100644 --- a/arch/arm/include/asm/edac.h +++ b/arch/arm/include/asm/edac.h @@ -18,11 +18,12 @@ #define ASM_EDAC_H /* * ECC atomic, DMA, SMP and interrupt safe scrub function. - * Implements the per arch atomic_scrub() that EDAC use for software + * Implements the per arch edac_atomic_scrub() that EDAC use for software * ECC scrubbing. It reads memory and then writes back the original * value, allowing the hardware to detect and correct memory errors. */ -static inline void atomic_scrub(void *va, u32 size) + +static inline void edac_atomic_scrub(void *va, u32 size) { #if __LINUX_ARM_ARCH__ >= 6 unsigned int *virt_addr = va; diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig index f5016656494f..b65edf514b40 100644 --- a/arch/mips/Kconfig +++ b/arch/mips/Kconfig @@ -819,6 +819,7 @@ config CAVIUM_OCTEON_SOC select SYS_SUPPORTS_64BIT_KERNEL select SYS_SUPPORTS_BIG_ENDIAN select EDAC_SUPPORT + select EDAC_ATOMIC_SCRUB select SYS_SUPPORTS_LITTLE_ENDIAN select SYS_SUPPORTS_HOTPLUG_CPU if CPU_BIG_ENDIAN select SYS_HAS_EARLY_PRINTK diff --git a/arch/mips/include/asm/edac.h b/arch/mips/include/asm/edac.h index 94105d3f58f4..980b16527374 100644 --- a/arch/mips/include/asm/edac.h +++ b/arch/mips/include/asm/edac.h @@ -5,7 +5,7 @@ /* ECC atomic, DMA, SMP and interrupt safe scrub function */ -static inline void atomic_scrub(void *va, u32 size) +static inline void edac_atomic_scrub(void *va, u32 size) { unsigned long *virt_addr = va; unsigned long temp; @@ -21,7 +21,7 @@ static inline void atomic_scrub(void *va, u32 size) __asm__ __volatile__ ( " .set mips2 \n" - "1: ll %0, %1 # atomic_scrub \n" + "1: ll %0, %1 # edac_atomic_scrub \n" " addu %0, $0 \n" " sc %0, %1 \n" " beqz %0, 1b \n" diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 190cc48abc0c..5ef27113b898 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -153,6 +153,8 @@ config PPC select NO_BOOTMEM select HAVE_GENERIC_RCU_GUP select HAVE_PERF_EVENTS_NMI if PPC64 + select EDAC_SUPPORT + select EDAC_ATOMIC_SCRUB config GENERIC_CSUM def_bool CPU_LITTLE_ENDIAN diff --git a/arch/powerpc/include/asm/edac.h b/arch/powerpc/include/asm/edac.h index 6ead88bbfbb8..5571e23d253e 100644 --- a/arch/powerpc/include/asm/edac.h +++ b/arch/powerpc/include/asm/edac.h @@ -12,11 +12,11 @@ #define ASM_EDAC_H /* * ECC atomic, DMA, SMP and interrupt safe scrub function. - * Implements the per arch atomic_scrub() that EDAC use for software + * Implements the per arch edac_atomic_scrub() that EDAC use for software * ECC scrubbing. It reads memory and then writes back the original * value, allowing the hardware to detect and correct memory errors. */ -static __inline__ void atomic_scrub(void *va, u32 size) +static __inline__ void edac_atomic_scrub(void *va, u32 size) { unsigned int *virt_addr = va; unsigned int temp; diff --git a/arch/tile/Kconfig b/arch/tile/Kconfig index a07e31b50d3f..59cf0b911898 100644 --- a/arch/tile/Kconfig +++ b/arch/tile/Kconfig @@ -28,6 +28,7 @@ config TILE select HAVE_DEBUG_STACKOVERFLOW select ARCH_WANT_FRAME_POINTERS select HAVE_CONTEXT_TRACKING + select EDAC_SUPPORT # FIXME: investigate whether we need/want these options. # select HAVE_IOREMAP_PROT diff --git a/arch/tile/include/asm/edac.h b/arch/tile/include/asm/edac.h deleted file mode 100644 index 87fc83eeaffd..000000000000 --- a/arch/tile/include/asm/edac.h +++ /dev/null @@ -1,29 +0,0 @@ -/* - * Copyright 2011 Tilera Corporation. All Rights Reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation, version 2. - * - * This program is distributed in the hope that it will be useful, but - * WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or - * NON INFRINGEMENT. See the GNU General Public License for - * more details. - */ - -#ifndef _ASM_TILE_EDAC_H -#define _ASM_TILE_EDAC_H - -/* ECC atomic, DMA, SMP and interrupt safe scrub function */ - -static inline void atomic_scrub(void *va, u32 size) -{ - /* - * These is nothing to be done here because CE is - * corrected by the mshim. - */ - return; -} - -#endif /* _ASM_TILE_EDAC_H */ diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 226d5696e1d1..482c160a9fe9 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -143,6 +143,8 @@ config X86 select ACPI_LEGACY_TABLES_LOOKUP if ACPI select X86_FEATURE_NAMES if PROC_FS select SRCU + select EDAC_SUPPORT + select EDAC_ATOMIC_SCRUB config INSTRUCTION_DECODER def_bool y diff --git a/arch/x86/include/asm/edac.h b/arch/x86/include/asm/edac.h index e9b57ecc70c5..cf8fdf83b231 100644 --- a/arch/x86/include/asm/edac.h +++ b/arch/x86/include/asm/edac.h @@ -3,7 +3,7 @@ /* ECC atomic, DMA, SMP and interrupt safe scrub function */ -static inline void atomic_scrub(void *va, u32 size) +static inline void edac_atomic_scrub(void *va, u32 size) { u32 i, *virt_addr = va; diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index 2d2530cdf99d..a90e06ac1631 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -2,15 +2,16 @@ # EDAC Kconfig # Copyright (c) 2008 Doug Thompson www.softwarebitmaker.com # Licensed and distributed under the GPL -# + +config EDAC_ATOMIC_SCRUB + bool config EDAC_SUPPORT bool menuconfig EDAC bool "EDAC (Error Detection And Correction) reporting" - depends on HAS_IOMEM - depends on X86 || PPC || TILE || ARM || EDAC_SUPPORT + depends on HAS_IOMEM && EDAC_SUPPORT help EDAC is designed to report errors in the core system. These are low-level errors that are reported in the CPU or diff --git a/drivers/edac/edac_mc.c b/drivers/edac/edac_mc.c index af3be1914dbb..943ed8cf71b9 100644 --- a/drivers/edac/edac_mc.c +++ b/drivers/edac/edac_mc.c @@ -30,11 +30,16 @@ #include #include #include -#include #include "edac_core.h" #include "edac_module.h" #include +#ifdef CONFIG_EDAC_ATOMIC_SCRUB +#include +#else +#define edac_atomic_scrub(va, size) do { } while (0) +#endif + /* lock to memory controller's control array */ static DEFINE_MUTEX(mem_ctls_mutex); static LIST_HEAD(mc_devices); @@ -874,7 +879,7 @@ static void edac_mc_scrub_block(unsigned long page, unsigned long offset, virt_addr = kmap_atomic(pg); /* Perform architecture specific atomic scrub operation */ - atomic_scrub(virt_addr + offset, size); + edac_atomic_scrub(virt_addr + offset, size); /* Unmap and complete */ kunmap_atomic(virt_addr); From 6e681abc88376ae448071cbaa9445c20f2812368 Mon Sep 17 00:00:00 2001 From: Loc Ho Date: Fri, 22 May 2015 17:32:56 -0600 Subject: [PATCH 04/28] arm64: Enable EDAC on ARM64 Select EDAC_SUPPORT for arm64. Signed-off-by: Loc Ho Cc: devicetree@vger.kernel.org Cc: dougthompson@xmission.com Cc: ijc+devicetree@hellion.org.uk Cc: jcm@redhat.com Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: mchehab@osg.samsung.com Cc: patches@apm.com Cc: robh+dt@kernel.org Link: http://lkml.kernel.org/r/1432337580-3750-2-git-send-email-lho@apm.com Signed-off-by: Borislav Petkov --- arch/arm64/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 4269dba63cf1..577078faea49 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -22,6 +22,7 @@ config ARM64 select BUILDTIME_EXTABLE_SORT select CLONE_BACKWARDS select COMMON_CLK + select EDAC_SUPPORT select CPU_PM if (SUSPEND || CPU_IDLE) select DCACHE_WORD_ACCESS select GENERIC_ALLOCATOR From 995e1de45af8ed5c841583a29f94f338f0bd50ba Mon Sep 17 00:00:00 2001 From: Loc Ho Date: Fri, 22 May 2015 17:32:57 -0600 Subject: [PATCH 05/28] MAINTAINERS: Add entry for APM X-Gene SoC EDAC driver Add a MAINTAINERS entry for APM X-Gene SoC EDAC driver. Signed-off-by: Loc Ho Cc: devicetree@vger.kernel.org Cc: dougthompson@xmission.com Cc: ijc+devicetree@hellion.org.uk Cc: jcm@redhat.com Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: mchehab@osg.samsung.com Cc: patches@apm.com Cc: robh+dt@kernel.org Link: http://lkml.kernel.org/r/1432337580-3750-3-git-send-email-lho@apm.com Signed-off-by: Borislav Petkov --- MAINTAINERS | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index be68a7156f13..a51ec44f9711 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3735,6 +3735,13 @@ W: bluesmoke.sourceforge.net S: Maintained F: drivers/edac/sb_edac.c +EDAC-XGENE +APPLIED MICRO (APM) X-GENE SOC EDAC +M: Loc Ho +S: Supported +F: drivers/edac/xgene_edac.c +F: Documentation/devicetree/bindings/edac/apm-xgene-edac.txt + EDIROL UA-101/UA-1000 DRIVER M: Clemens Ladisch L: alsa-devel@alsa-project.org (moderated for non-subscribers) From ee66e6a24bd2f8a294f018ab90f660f23d8a7ff5 Mon Sep 17 00:00:00 2001 From: Loc Ho Date: Fri, 22 May 2015 17:32:58 -0600 Subject: [PATCH 06/28] Documentation: Add documentation for the APM X-Gene SoC EDAC DTS binding Add documentation for the APM X-Gene SoC EDAC DTS binding. Signed-off-by: Loc Ho Acked-by: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dougthompson@xmission.com Cc: ijc+devicetree@hellion.org.uk Cc: jcm@redhat.com Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: mchehab@osg.samsung.com Cc: patches@apm.com Cc: robh+dt@kernel.org Link: http://lkml.kernel.org/r/1432337580-3750-4-git-send-email-lho@apm.com Signed-off-by: Borislav Petkov --- .../bindings/edac/apm-xgene-edac.txt | 78 +++++++++++++++++++ 1 file changed, 78 insertions(+) create mode 100644 Documentation/devicetree/bindings/edac/apm-xgene-edac.txt diff --git a/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt b/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt new file mode 100644 index 000000000000..480911c38ff9 --- /dev/null +++ b/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt @@ -0,0 +1,78 @@ +* APM X-Gene SoC EDAC node + +EDAC node is defined to describe on-chip error detection and correction. +The follow error types are supported: + + memory controller - Memory controller + PMD (L1/L2) - Processor module unit (PMD) L1/L2 cache + +The following section describes the EDAC DT node binding. + +Required properties: +- compatible : Shall be "apm,xgene-edac". +- regmap-csw : Regmap of the CPU switch fabric (CSW) resource. +- regmap-mcba : Regmap of the MCB-A (memory bridge) resource. +- regmap-mcbb : Regmap of the MCB-B (memory bridge) resource. +- regmap-efuse : Regmap of the PMD efuse resource. +- reg : First resource shall be the CPU bus (PCP) resource. +- interrupts : Interrupt-specifier for MCU, PMD, L3, or SoC error + IRQ(s). + +Required properties for memory controller subnode: +- compatible : Shall be "apm,xgene-edac-mc". +- reg : First resource shall be the memory controller unit + (MCU) resource. +- memory-controller : Instance number of the memory controller. + +Required properties for PMD subnode: +- compatible : Shall be "apm,xgene-edac-pmd". +- reg : First resource shall be the PMD resource. +- pmd-controller : Instance number of the PMD controller. + +Example: + csw: csw@7e200000 { + compatible = "apm,xgene-csw", "syscon"; + reg = <0x0 0x7e200000 0x0 0x1000>; + }; + + mcba: mcba@7e700000 { + compatible = "apm,xgene-mcb", "syscon"; + reg = <0x0 0x7e700000 0x0 0x1000>; + }; + + mcbb: mcbb@7e720000 { + compatible = "apm,xgene-mcb", "syscon"; + reg = <0x0 0x7e720000 0x0 0x1000>; + }; + + efuse: efuse@1054a000 { + compatible = "apm,xgene-efuse", "syscon"; + reg = <0x0 0x1054a000 0x0 0x20>; + }; + + edac@78800000 { + compatible = "apm,xgene-edac"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + regmap-csw = <&csw>; + regmap-mcba = <&mcba>; + regmap-mcbb = <&mcbb>; + regmap-efuse = <&efuse>; + reg = <0x0 0x78800000 0x0 0x100>; + interrupts = <0x0 0x20 0x4>, + <0x0 0x21 0x4>, + <0x0 0x27 0x4>; + + edacmc@7e800000 { + compatible = "apm,xgene-edac-mc"; + reg = <0x0 0x7e800000 0x0 0x1000>; + memory-controller = <0>; + }; + + edacpmd@7c000000 { + compatible = "apm,xgene-edac-pmd"; + reg = <0x0 0x7c000000 0x0 0x200000>; + pmd-controller = <0>; + }; + }; From 0d4429301c4aa4028b0b96bf492ddb03e13206e3 Mon Sep 17 00:00:00 2001 From: Loc Ho Date: Fri, 22 May 2015 17:32:59 -0600 Subject: [PATCH 07/28] EDAC: Add APM X-Gene SoC EDAC driver Add support for the APM X-Gene SoC EDAC driver. Signed-off-by: Loc Ho Acked-by: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dougthompson@xmission.com Cc: ijc+devicetree@hellion.org.uk Cc: jcm@redhat.com Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: mchehab@osg.samsung.com Cc: patches@apm.com Cc: robh+dt@kernel.org Link: http://lkml.kernel.org/r/1432337580-3750-5-git-send-email-lho@apm.com Signed-off-by: Borislav Petkov --- drivers/edac/Kconfig | 7 + drivers/edac/Makefile | 1 + drivers/edac/xgene_edac.c | 1254 +++++++++++++++++++++++++++++++++++++ 3 files changed, 1262 insertions(+) create mode 100644 drivers/edac/xgene_edac.c diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index a90e06ac1631..9016ec8a0295 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -393,4 +393,11 @@ config EDAC_SYNOPSYS Support for error detection and correction on the Synopsys DDR memory controller. +config EDAC_XGENE + tristate "APM X-Gene SoC" + depends on EDAC_MM_EDAC && (ARM64 || COMPILE_TEST) + help + Support for error detection and correction on the + APM X-Gene family of SOCs. + endif # EDAC diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile index b255f362b1db..28ef2a519f65 100644 --- a/drivers/edac/Makefile +++ b/drivers/edac/Makefile @@ -68,3 +68,4 @@ obj-$(CONFIG_EDAC_OCTEON_PCI) += octeon_edac-pci.o obj-$(CONFIG_EDAC_ALTERA_MC) += altera_edac.o obj-$(CONFIG_EDAC_SYNOPSYS) += synopsys_edac.o +obj-$(CONFIG_EDAC_XGENE) += xgene_edac.o diff --git a/drivers/edac/xgene_edac.c b/drivers/edac/xgene_edac.c new file mode 100644 index 000000000000..b5158572f6ab --- /dev/null +++ b/drivers/edac/xgene_edac.c @@ -0,0 +1,1254 @@ +/* + * APM X-Gene SoC EDAC (error detection and correction) + * + * Copyright (c) 2015, Applied Micro Circuits Corporation + * Author: Feng Kan + * Loc Ho + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include "edac_core.h" + +#define EDAC_MOD_STR "xgene_edac" + +/* Global error configuration status registers (CSR) */ +#define PCPHPERRINTSTS 0x0000 +#define PCPHPERRINTMSK 0x0004 +#define MCU_CTL_ERR_MASK BIT(12) +#define IOB_PA_ERR_MASK BIT(11) +#define IOB_BA_ERR_MASK BIT(10) +#define IOB_XGIC_ERR_MASK BIT(9) +#define IOB_RB_ERR_MASK BIT(8) +#define L3C_UNCORR_ERR_MASK BIT(5) +#define MCU_UNCORR_ERR_MASK BIT(4) +#define PMD3_MERR_MASK BIT(3) +#define PMD2_MERR_MASK BIT(2) +#define PMD1_MERR_MASK BIT(1) +#define PMD0_MERR_MASK BIT(0) +#define PCPLPERRINTSTS 0x0008 +#define PCPLPERRINTMSK 0x000C +#define CSW_SWITCH_TRACE_ERR_MASK BIT(2) +#define L3C_CORR_ERR_MASK BIT(1) +#define MCU_CORR_ERR_MASK BIT(0) +#define MEMERRINTSTS 0x0010 +#define MEMERRINTMSK 0x0014 + +struct xgene_edac { + struct device *dev; + struct regmap *csw_map; + struct regmap *mcba_map; + struct regmap *mcbb_map; + struct regmap *efuse_map; + void __iomem *pcp_csr; + spinlock_t lock; + struct dentry *dfs; + + struct list_head mcus; + struct list_head pmds; + + struct mutex mc_lock; + int mc_active_mask; + int mc_registered_mask; +}; + +static void xgene_edac_pcp_rd(struct xgene_edac *edac, u32 reg, u32 *val) +{ + *val = readl(edac->pcp_csr + reg); +} + +static void xgene_edac_pcp_clrbits(struct xgene_edac *edac, u32 reg, + u32 bits_mask) +{ + u32 val; + + spin_lock(&edac->lock); + val = readl(edac->pcp_csr + reg); + val &= ~bits_mask; + writel(val, edac->pcp_csr + reg); + spin_unlock(&edac->lock); +} + +static void xgene_edac_pcp_setbits(struct xgene_edac *edac, u32 reg, + u32 bits_mask) +{ + u32 val; + + spin_lock(&edac->lock); + val = readl(edac->pcp_csr + reg); + val |= bits_mask; + writel(val, edac->pcp_csr + reg); + spin_unlock(&edac->lock); +} + +/* Memory controller error CSR */ +#define MCU_MAX_RANK 8 +#define MCU_RANK_STRIDE 0x40 + +#define MCUGECR 0x0110 +#define MCU_GECR_DEMANDUCINTREN_MASK BIT(0) +#define MCU_GECR_BACKUCINTREN_MASK BIT(1) +#define MCU_GECR_CINTREN_MASK BIT(2) +#define MUC_GECR_MCUADDRERREN_MASK BIT(9) +#define MCUGESR 0x0114 +#define MCU_GESR_ADDRNOMATCH_ERR_MASK BIT(7) +#define MCU_GESR_ADDRMULTIMATCH_ERR_MASK BIT(6) +#define MCU_GESR_PHYP_ERR_MASK BIT(3) +#define MCUESRR0 0x0314 +#define MCU_ESRR_MULTUCERR_MASK BIT(3) +#define MCU_ESRR_BACKUCERR_MASK BIT(2) +#define MCU_ESRR_DEMANDUCERR_MASK BIT(1) +#define MCU_ESRR_CERR_MASK BIT(0) +#define MCUESRRA0 0x0318 +#define MCUEBLRR0 0x031c +#define MCU_EBLRR_ERRBANK_RD(src) (((src) & 0x00000007) >> 0) +#define MCUERCRR0 0x0320 +#define MCU_ERCRR_ERRROW_RD(src) (((src) & 0xFFFF0000) >> 16) +#define MCU_ERCRR_ERRCOL_RD(src) ((src) & 0x00000FFF) +#define MCUSBECNT0 0x0324 +#define MCU_SBECNT_COUNT(src) ((src) & 0xFFFF) + +#define CSW_CSWCR 0x0000 +#define CSW_CSWCR_DUALMCB_MASK BIT(0) + +#define MCBADDRMR 0x0000 +#define MCBADDRMR_MCU_INTLV_MODE_MASK BIT(3) +#define MCBADDRMR_DUALMCU_MODE_MASK BIT(2) +#define MCBADDRMR_MCB_INTLV_MODE_MASK BIT(1) +#define MCBADDRMR_ADDRESS_MODE_MASK BIT(0) + +struct xgene_edac_mc_ctx { + struct list_head next; + char *name; + struct mem_ctl_info *mci; + struct xgene_edac *edac; + void __iomem *mcu_csr; + u32 mcu_id; +}; + +static ssize_t xgene_edac_mc_err_inject_write(struct file *file, + const char __user *data, + size_t count, loff_t *ppos) +{ + struct mem_ctl_info *mci = file->private_data; + struct xgene_edac_mc_ctx *ctx = mci->pvt_info; + int i; + + for (i = 0; i < MCU_MAX_RANK; i++) { + writel(MCU_ESRR_MULTUCERR_MASK | MCU_ESRR_BACKUCERR_MASK | + MCU_ESRR_DEMANDUCERR_MASK | MCU_ESRR_CERR_MASK, + ctx->mcu_csr + MCUESRRA0 + i * MCU_RANK_STRIDE); + } + return count; +} + +static const struct file_operations xgene_edac_mc_debug_inject_fops = { + .open = simple_open, + .write = xgene_edac_mc_err_inject_write, + .llseek = generic_file_llseek, +}; + +static void xgene_edac_mc_create_debugfs_node(struct mem_ctl_info *mci) +{ + if (!IS_ENABLED(CONFIG_EDAC_DEBUG)) + return; +#ifdef CONFIG_EDAC_DEBUG + if (!mci->debugfs) + return; + debugfs_create_file("inject_ctrl", S_IWUSR, mci->debugfs, mci, + &xgene_edac_mc_debug_inject_fops); +#endif +} + +static void xgene_edac_mc_check(struct mem_ctl_info *mci) +{ + struct xgene_edac_mc_ctx *ctx = mci->pvt_info; + unsigned int pcp_hp_stat; + unsigned int pcp_lp_stat; + u32 reg; + u32 rank; + u32 bank; + u32 count; + u32 col_row; + + xgene_edac_pcp_rd(ctx->edac, PCPHPERRINTSTS, &pcp_hp_stat); + xgene_edac_pcp_rd(ctx->edac, PCPLPERRINTSTS, &pcp_lp_stat); + if (!((MCU_UNCORR_ERR_MASK & pcp_hp_stat) || + (MCU_CTL_ERR_MASK & pcp_hp_stat) || + (MCU_CORR_ERR_MASK & pcp_lp_stat))) + return; + + for (rank = 0; rank < MCU_MAX_RANK; rank++) { + reg = readl(ctx->mcu_csr + MCUESRR0 + rank * MCU_RANK_STRIDE); + + /* Detect uncorrectable memory error */ + if (reg & (MCU_ESRR_DEMANDUCERR_MASK | + MCU_ESRR_BACKUCERR_MASK)) { + /* Detected uncorrectable memory error */ + edac_mc_chipset_printk(mci, KERN_ERR, "X-Gene", + "MCU uncorrectable error at rank %d\n", rank); + + edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci, + 1, 0, 0, 0, 0, 0, -1, mci->ctl_name, ""); + } + + /* Detect correctable memory error */ + if (reg & MCU_ESRR_CERR_MASK) { + bank = readl(ctx->mcu_csr + MCUEBLRR0 + + rank * MCU_RANK_STRIDE); + col_row = readl(ctx->mcu_csr + MCUERCRR0 + + rank * MCU_RANK_STRIDE); + count = readl(ctx->mcu_csr + MCUSBECNT0 + + rank * MCU_RANK_STRIDE); + edac_mc_chipset_printk(mci, KERN_WARNING, "X-Gene", + "MCU correctable error at rank %d bank %d column %d row %d count %d\n", + rank, MCU_EBLRR_ERRBANK_RD(bank), + MCU_ERCRR_ERRCOL_RD(col_row), + MCU_ERCRR_ERRROW_RD(col_row), + MCU_SBECNT_COUNT(count)); + + edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, + 1, 0, 0, 0, 0, 0, -1, mci->ctl_name, ""); + } + + /* Clear all error registers */ + writel(0x0, ctx->mcu_csr + MCUEBLRR0 + rank * MCU_RANK_STRIDE); + writel(0x0, ctx->mcu_csr + MCUERCRR0 + rank * MCU_RANK_STRIDE); + writel(0x0, ctx->mcu_csr + MCUSBECNT0 + + rank * MCU_RANK_STRIDE); + writel(reg, ctx->mcu_csr + MCUESRR0 + rank * MCU_RANK_STRIDE); + } + + /* Detect memory controller error */ + reg = readl(ctx->mcu_csr + MCUGESR); + if (reg) { + if (reg & MCU_GESR_ADDRNOMATCH_ERR_MASK) + edac_mc_chipset_printk(mci, KERN_WARNING, "X-Gene", + "MCU address miss-match error\n"); + if (reg & MCU_GESR_ADDRMULTIMATCH_ERR_MASK) + edac_mc_chipset_printk(mci, KERN_WARNING, "X-Gene", + "MCU address multi-match error\n"); + + writel(reg, ctx->mcu_csr + MCUGESR); + } +} + +static void xgene_edac_mc_irq_ctl(struct mem_ctl_info *mci, bool enable) +{ + struct xgene_edac_mc_ctx *ctx = mci->pvt_info; + unsigned int val; + + if (edac_op_state != EDAC_OPSTATE_INT) + return; + + mutex_lock(&ctx->edac->mc_lock); + + /* + * As there is only single bit for enable error and interrupt mask, + * we must only enable top level interrupt after all MCUs are + * registered. Otherwise, if there is an error and the corresponding + * MCU has not registered, the interrupt will never get cleared. To + * determine all MCU have registered, we will keep track of active + * MCUs and registered MCUs. + */ + if (enable) { + /* Set registered MCU bit */ + ctx->edac->mc_registered_mask |= 1 << ctx->mcu_id; + + /* Enable interrupt after all active MCU registered */ + if (ctx->edac->mc_registered_mask == + ctx->edac->mc_active_mask) { + /* Enable memory controller top level interrupt */ + xgene_edac_pcp_clrbits(ctx->edac, PCPHPERRINTMSK, + MCU_UNCORR_ERR_MASK | + MCU_CTL_ERR_MASK); + xgene_edac_pcp_clrbits(ctx->edac, PCPLPERRINTMSK, + MCU_CORR_ERR_MASK); + } + + /* Enable MCU interrupt and error reporting */ + val = readl(ctx->mcu_csr + MCUGECR); + val |= MCU_GECR_DEMANDUCINTREN_MASK | + MCU_GECR_BACKUCINTREN_MASK | + MCU_GECR_CINTREN_MASK | + MUC_GECR_MCUADDRERREN_MASK; + writel(val, ctx->mcu_csr + MCUGECR); + } else { + /* Disable MCU interrupt */ + val = readl(ctx->mcu_csr + MCUGECR); + val &= ~(MCU_GECR_DEMANDUCINTREN_MASK | + MCU_GECR_BACKUCINTREN_MASK | + MCU_GECR_CINTREN_MASK | + MUC_GECR_MCUADDRERREN_MASK); + writel(val, ctx->mcu_csr + MCUGECR); + + /* Disable memory controller top level interrupt */ + xgene_edac_pcp_setbits(ctx->edac, PCPHPERRINTMSK, + MCU_UNCORR_ERR_MASK | MCU_CTL_ERR_MASK); + xgene_edac_pcp_setbits(ctx->edac, PCPLPERRINTMSK, + MCU_CORR_ERR_MASK); + + /* Clear registered MCU bit */ + ctx->edac->mc_registered_mask &= ~(1 << ctx->mcu_id); + } + + mutex_unlock(&ctx->edac->mc_lock); +} + +static int xgene_edac_mc_is_active(struct xgene_edac_mc_ctx *ctx, int mc_idx) +{ + unsigned int reg; + u32 mcu_mask; + + if (regmap_read(ctx->edac->csw_map, CSW_CSWCR, ®)) + return 0; + + if (reg & CSW_CSWCR_DUALMCB_MASK) { + /* + * Dual MCB active - Determine if all 4 active or just MCU0 + * and MCU2 active + */ + if (regmap_read(ctx->edac->mcbb_map, MCBADDRMR, ®)) + return 0; + mcu_mask = (reg & MCBADDRMR_DUALMCU_MODE_MASK) ? 0xF : 0x5; + } else { + /* + * Single MCB active - Determine if MCU0/MCU1 or just MCU0 + * active + */ + if (regmap_read(ctx->edac->mcba_map, MCBADDRMR, ®)) + return 0; + mcu_mask = (reg & MCBADDRMR_DUALMCU_MODE_MASK) ? 0x3 : 0x1; + } + + /* Save active MC mask if hasn't set already */ + if (!ctx->edac->mc_active_mask) + ctx->edac->mc_active_mask = mcu_mask; + + return (mcu_mask & (1 << mc_idx)) ? 1 : 0; +} + +static int xgene_edac_mc_add(struct xgene_edac *edac, struct device_node *np) +{ + struct mem_ctl_info *mci; + struct edac_mc_layer layers[2]; + struct xgene_edac_mc_ctx tmp_ctx; + struct xgene_edac_mc_ctx *ctx; + struct resource res; + int rc; + + memset(&tmp_ctx, 0, sizeof(tmp_ctx)); + tmp_ctx.edac = edac; + + if (!devres_open_group(edac->dev, xgene_edac_mc_add, GFP_KERNEL)) + return -ENOMEM; + + rc = of_address_to_resource(np, 0, &res); + if (rc < 0) { + dev_err(edac->dev, "no MCU resource address\n"); + goto err_group; + } + tmp_ctx.mcu_csr = devm_ioremap_resource(edac->dev, &res); + if (IS_ERR(tmp_ctx.mcu_csr)) { + dev_err(edac->dev, "unable to map MCU resource\n"); + rc = PTR_ERR(tmp_ctx.mcu_csr); + goto err_group; + } + + /* Ignore non-active MCU */ + if (of_property_read_u32(np, "memory-controller", &tmp_ctx.mcu_id)) { + dev_err(edac->dev, "no memory-controller property\n"); + rc = -ENODEV; + goto err_group; + } + if (!xgene_edac_mc_is_active(&tmp_ctx, tmp_ctx.mcu_id)) { + rc = -ENODEV; + goto err_group; + } + + layers[0].type = EDAC_MC_LAYER_CHIP_SELECT; + layers[0].size = 4; + layers[0].is_virt_csrow = true; + layers[1].type = EDAC_MC_LAYER_CHANNEL; + layers[1].size = 2; + layers[1].is_virt_csrow = false; + mci = edac_mc_alloc(tmp_ctx.mcu_id, ARRAY_SIZE(layers), layers, + sizeof(*ctx)); + if (!mci) { + rc = -ENOMEM; + goto err_group; + } + + ctx = mci->pvt_info; + *ctx = tmp_ctx; /* Copy over resource value */ + ctx->name = "xgene_edac_mc_err"; + ctx->mci = mci; + mci->pdev = &mci->dev; + mci->ctl_name = ctx->name; + mci->dev_name = ctx->name; + + mci->mtype_cap = MEM_FLAG_RDDR | MEM_FLAG_RDDR2 | MEM_FLAG_RDDR3 | + MEM_FLAG_DDR | MEM_FLAG_DDR2 | MEM_FLAG_DDR3; + mci->edac_ctl_cap = EDAC_FLAG_SECDED; + mci->edac_cap = EDAC_FLAG_SECDED; + mci->mod_name = EDAC_MOD_STR; + mci->mod_ver = "0.1"; + mci->ctl_page_to_phys = NULL; + mci->scrub_cap = SCRUB_FLAG_HW_SRC; + mci->scrub_mode = SCRUB_HW_SRC; + + if (edac_op_state == EDAC_OPSTATE_POLL) + mci->edac_check = xgene_edac_mc_check; + + if (edac_mc_add_mc(mci)) { + dev_err(edac->dev, "edac_mc_add_mc failed\n"); + rc = -EINVAL; + goto err_free; + } + + xgene_edac_mc_create_debugfs_node(mci); + + list_add(&ctx->next, &edac->mcus); + + xgene_edac_mc_irq_ctl(mci, true); + + devres_remove_group(edac->dev, xgene_edac_mc_add); + + dev_info(edac->dev, "X-Gene EDAC MC registered\n"); + return 0; + +err_free: + edac_mc_free(mci); +err_group: + devres_release_group(edac->dev, xgene_edac_mc_add); + return rc; +} + +static int xgene_edac_mc_remove(struct xgene_edac_mc_ctx *mcu) +{ + xgene_edac_mc_irq_ctl(mcu->mci, false); + edac_mc_del_mc(&mcu->mci->dev); + edac_mc_free(mcu->mci); + return 0; +} + +/* CPU L1/L2 error CSR */ +#define MAX_CPU_PER_PMD 2 +#define CPU_CSR_STRIDE 0x00100000 +#define CPU_L2C_PAGE 0x000D0000 +#define CPU_MEMERR_L2C_PAGE 0x000E0000 +#define CPU_MEMERR_CPU_PAGE 0x000F0000 + +#define MEMERR_CPU_ICFECR_PAGE_OFFSET 0x0000 +#define MEMERR_CPU_ICFESR_PAGE_OFFSET 0x0004 +#define MEMERR_CPU_ICFESR_ERRWAY_RD(src) (((src) & 0xFF000000) >> 24) +#define MEMERR_CPU_ICFESR_ERRINDEX_RD(src) (((src) & 0x003F0000) >> 16) +#define MEMERR_CPU_ICFESR_ERRINFO_RD(src) (((src) & 0x0000FF00) >> 8) +#define MEMERR_CPU_ICFESR_ERRTYPE_RD(src) (((src) & 0x00000070) >> 4) +#define MEMERR_CPU_ICFESR_MULTCERR_MASK BIT(2) +#define MEMERR_CPU_ICFESR_CERR_MASK BIT(0) +#define MEMERR_CPU_LSUESR_PAGE_OFFSET 0x000c +#define MEMERR_CPU_LSUESR_ERRWAY_RD(src) (((src) & 0xFF000000) >> 24) +#define MEMERR_CPU_LSUESR_ERRINDEX_RD(src) (((src) & 0x003F0000) >> 16) +#define MEMERR_CPU_LSUESR_ERRINFO_RD(src) (((src) & 0x0000FF00) >> 8) +#define MEMERR_CPU_LSUESR_ERRTYPE_RD(src) (((src) & 0x00000070) >> 4) +#define MEMERR_CPU_LSUESR_MULTCERR_MASK BIT(2) +#define MEMERR_CPU_LSUESR_CERR_MASK BIT(0) +#define MEMERR_CPU_LSUECR_PAGE_OFFSET 0x0008 +#define MEMERR_CPU_MMUECR_PAGE_OFFSET 0x0010 +#define MEMERR_CPU_MMUESR_PAGE_OFFSET 0x0014 +#define MEMERR_CPU_MMUESR_ERRWAY_RD(src) (((src) & 0xFF000000) >> 24) +#define MEMERR_CPU_MMUESR_ERRINDEX_RD(src) (((src) & 0x007F0000) >> 16) +#define MEMERR_CPU_MMUESR_ERRINFO_RD(src) (((src) & 0x0000FF00) >> 8) +#define MEMERR_CPU_MMUESR_ERRREQSTR_LSU_MASK BIT(7) +#define MEMERR_CPU_MMUESR_ERRTYPE_RD(src) (((src) & 0x00000070) >> 4) +#define MEMERR_CPU_MMUESR_MULTCERR_MASK BIT(2) +#define MEMERR_CPU_MMUESR_CERR_MASK BIT(0) +#define MEMERR_CPU_ICFESRA_PAGE_OFFSET 0x0804 +#define MEMERR_CPU_LSUESRA_PAGE_OFFSET 0x080c +#define MEMERR_CPU_MMUESRA_PAGE_OFFSET 0x0814 + +#define MEMERR_L2C_L2ECR_PAGE_OFFSET 0x0000 +#define MEMERR_L2C_L2ESR_PAGE_OFFSET 0x0004 +#define MEMERR_L2C_L2ESR_ERRSYN_RD(src) (((src) & 0xFF000000) >> 24) +#define MEMERR_L2C_L2ESR_ERRWAY_RD(src) (((src) & 0x00FC0000) >> 18) +#define MEMERR_L2C_L2ESR_ERRCPU_RD(src) (((src) & 0x00020000) >> 17) +#define MEMERR_L2C_L2ESR_ERRGROUP_RD(src) (((src) & 0x0000E000) >> 13) +#define MEMERR_L2C_L2ESR_ERRACTION_RD(src) (((src) & 0x00001C00) >> 10) +#define MEMERR_L2C_L2ESR_ERRTYPE_RD(src) (((src) & 0x00000300) >> 8) +#define MEMERR_L2C_L2ESR_MULTUCERR_MASK BIT(3) +#define MEMERR_L2C_L2ESR_MULTICERR_MASK BIT(2) +#define MEMERR_L2C_L2ESR_UCERR_MASK BIT(1) +#define MEMERR_L2C_L2ESR_ERR_MASK BIT(0) +#define MEMERR_L2C_L2EALR_PAGE_OFFSET 0x0008 +#define CPUX_L2C_L2RTOCR_PAGE_OFFSET 0x0010 +#define MEMERR_L2C_L2EAHR_PAGE_OFFSET 0x000c +#define CPUX_L2C_L2RTOSR_PAGE_OFFSET 0x0014 +#define MEMERR_L2C_L2RTOSR_MULTERR_MASK BIT(1) +#define MEMERR_L2C_L2RTOSR_ERR_MASK BIT(0) +#define CPUX_L2C_L2RTOALR_PAGE_OFFSET 0x0018 +#define CPUX_L2C_L2RTOAHR_PAGE_OFFSET 0x001c +#define MEMERR_L2C_L2ESRA_PAGE_OFFSET 0x0804 + +/* + * Processor Module Domain (PMD) context - Context for a pair of processsors. + * Each PMD consists of 2 CPUs and a shared L2 cache. Each CPU consists of + * its own L1 cache. + */ +struct xgene_edac_pmd_ctx { + struct list_head next; + struct device ddev; + char *name; + struct xgene_edac *edac; + struct edac_device_ctl_info *edac_dev; + void __iomem *pmd_csr; + u32 pmd; +}; + +static void xgene_edac_pmd_l1_check(struct edac_device_ctl_info *edac_dev, + int cpu_idx) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + void __iomem *pg_f; + u32 val; + + pg_f = ctx->pmd_csr + cpu_idx * CPU_CSR_STRIDE + CPU_MEMERR_CPU_PAGE; + + val = readl(pg_f + MEMERR_CPU_ICFESR_PAGE_OFFSET); + if (val) { + dev_err(edac_dev->dev, + "CPU%d L1 memory error ICF 0x%08X Way 0x%02X Index 0x%02X Info 0x%02X\n", + ctx->pmd * MAX_CPU_PER_PMD + cpu_idx, val, + MEMERR_CPU_ICFESR_ERRWAY_RD(val), + MEMERR_CPU_ICFESR_ERRINDEX_RD(val), + MEMERR_CPU_ICFESR_ERRINFO_RD(val)); + if (val & MEMERR_CPU_ICFESR_CERR_MASK) + dev_err(edac_dev->dev, + "One or more correctable error\n"); + if (val & MEMERR_CPU_ICFESR_MULTCERR_MASK) + dev_err(edac_dev->dev, "Multiple correctable error\n"); + switch (MEMERR_CPU_ICFESR_ERRTYPE_RD(val)) { + case 1: + dev_err(edac_dev->dev, "L1 TLB multiple hit\n"); + break; + case 2: + dev_err(edac_dev->dev, "Way select multiple hit\n"); + break; + case 3: + dev_err(edac_dev->dev, "Physical tag parity error\n"); + break; + case 4: + case 5: + dev_err(edac_dev->dev, "L1 data parity error\n"); + break; + case 6: + dev_err(edac_dev->dev, "L1 pre-decode parity error\n"); + break; + } + + /* Clear any HW errors */ + writel(val, pg_f + MEMERR_CPU_ICFESR_PAGE_OFFSET); + + if (val & (MEMERR_CPU_ICFESR_CERR_MASK | + MEMERR_CPU_ICFESR_MULTCERR_MASK)) + edac_device_handle_ce(edac_dev, 0, 0, + edac_dev->ctl_name); + } + + val = readl(pg_f + MEMERR_CPU_LSUESR_PAGE_OFFSET); + if (val) { + dev_err(edac_dev->dev, + "CPU%d memory error LSU 0x%08X Way 0x%02X Index 0x%02X Info 0x%02X\n", + ctx->pmd * MAX_CPU_PER_PMD + cpu_idx, val, + MEMERR_CPU_LSUESR_ERRWAY_RD(val), + MEMERR_CPU_LSUESR_ERRINDEX_RD(val), + MEMERR_CPU_LSUESR_ERRINFO_RD(val)); + if (val & MEMERR_CPU_LSUESR_CERR_MASK) + dev_err(edac_dev->dev, + "One or more correctable error\n"); + if (val & MEMERR_CPU_LSUESR_MULTCERR_MASK) + dev_err(edac_dev->dev, "Multiple correctable error\n"); + switch (MEMERR_CPU_LSUESR_ERRTYPE_RD(val)) { + case 0: + dev_err(edac_dev->dev, "Load tag error\n"); + break; + case 1: + dev_err(edac_dev->dev, "Load data error\n"); + break; + case 2: + dev_err(edac_dev->dev, "WSL multihit error\n"); + break; + case 3: + dev_err(edac_dev->dev, "Store tag error\n"); + break; + case 4: + dev_err(edac_dev->dev, + "DTB multihit from load pipeline error\n"); + break; + case 5: + dev_err(edac_dev->dev, + "DTB multihit from store pipeline error\n"); + break; + } + + /* Clear any HW errors */ + writel(val, pg_f + MEMERR_CPU_LSUESR_PAGE_OFFSET); + + if (val & (MEMERR_CPU_LSUESR_CERR_MASK | + MEMERR_CPU_LSUESR_MULTCERR_MASK)) + edac_device_handle_ce(edac_dev, 0, 0, + edac_dev->ctl_name); + } + + val = readl(pg_f + MEMERR_CPU_MMUESR_PAGE_OFFSET); + if (val) { + dev_err(edac_dev->dev, + "CPU%d memory error MMU 0x%08X Way 0x%02X Index 0x%02X Info 0x%02X %s\n", + ctx->pmd * MAX_CPU_PER_PMD + cpu_idx, val, + MEMERR_CPU_MMUESR_ERRWAY_RD(val), + MEMERR_CPU_MMUESR_ERRINDEX_RD(val), + MEMERR_CPU_MMUESR_ERRINFO_RD(val), + val & MEMERR_CPU_MMUESR_ERRREQSTR_LSU_MASK ? "LSU" : + "ICF"); + if (val & MEMERR_CPU_MMUESR_CERR_MASK) + dev_err(edac_dev->dev, + "One or more correctable error\n"); + if (val & MEMERR_CPU_MMUESR_MULTCERR_MASK) + dev_err(edac_dev->dev, "Multiple correctable error\n"); + switch (MEMERR_CPU_MMUESR_ERRTYPE_RD(val)) { + case 0: + dev_err(edac_dev->dev, "Stage 1 UTB hit error\n"); + break; + case 1: + dev_err(edac_dev->dev, "Stage 1 UTB miss error\n"); + break; + case 2: + dev_err(edac_dev->dev, "Stage 1 UTB allocate error\n"); + break; + case 3: + dev_err(edac_dev->dev, + "TMO operation single bank error\n"); + break; + case 4: + dev_err(edac_dev->dev, "Stage 2 UTB error\n"); + break; + case 5: + dev_err(edac_dev->dev, "Stage 2 UTB miss error\n"); + break; + case 6: + dev_err(edac_dev->dev, "Stage 2 UTB allocate error\n"); + break; + case 7: + dev_err(edac_dev->dev, + "TMO operation multiple bank error\n"); + break; + } + + /* Clear any HW errors */ + writel(val, pg_f + MEMERR_CPU_MMUESR_PAGE_OFFSET); + + edac_device_handle_ce(edac_dev, 0, 0, edac_dev->ctl_name); + } +} + +static void xgene_edac_pmd_l2_check(struct edac_device_ctl_info *edac_dev) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + void __iomem *pg_d; + void __iomem *pg_e; + u32 val_hi; + u32 val_lo; + u32 val; + + /* Check L2 */ + pg_e = ctx->pmd_csr + CPU_MEMERR_L2C_PAGE; + val = readl(pg_e + MEMERR_L2C_L2ESR_PAGE_OFFSET); + if (val) { + val_lo = readl(pg_e + MEMERR_L2C_L2EALR_PAGE_OFFSET); + val_hi = readl(pg_e + MEMERR_L2C_L2EAHR_PAGE_OFFSET); + dev_err(edac_dev->dev, + "PMD%d memory error L2C L2ESR 0x%08X @ 0x%08X.%08X\n", + ctx->pmd, val, val_hi, val_lo); + dev_err(edac_dev->dev, + "ErrSyndrome 0x%02X ErrWay 0x%02X ErrCpu %d ErrGroup 0x%02X ErrAction 0x%02X\n", + MEMERR_L2C_L2ESR_ERRSYN_RD(val), + MEMERR_L2C_L2ESR_ERRWAY_RD(val), + MEMERR_L2C_L2ESR_ERRCPU_RD(val), + MEMERR_L2C_L2ESR_ERRGROUP_RD(val), + MEMERR_L2C_L2ESR_ERRACTION_RD(val)); + + if (val & MEMERR_L2C_L2ESR_ERR_MASK) + dev_err(edac_dev->dev, + "One or more correctable error\n"); + if (val & MEMERR_L2C_L2ESR_MULTICERR_MASK) + dev_err(edac_dev->dev, "Multiple correctable error\n"); + if (val & MEMERR_L2C_L2ESR_UCERR_MASK) + dev_err(edac_dev->dev, + "One or more uncorrectable error\n"); + if (val & MEMERR_L2C_L2ESR_MULTUCERR_MASK) + dev_err(edac_dev->dev, + "Multiple uncorrectable error\n"); + + switch (MEMERR_L2C_L2ESR_ERRTYPE_RD(val)) { + case 0: + dev_err(edac_dev->dev, "Outbound SDB parity error\n"); + break; + case 1: + dev_err(edac_dev->dev, "Inbound SDB parity error\n"); + break; + case 2: + dev_err(edac_dev->dev, "Tag ECC error\n"); + break; + case 3: + dev_err(edac_dev->dev, "Data ECC error\n"); + break; + } + + /* Clear any HW errors */ + writel(val, pg_e + MEMERR_L2C_L2ESR_PAGE_OFFSET); + + if (val & (MEMERR_L2C_L2ESR_ERR_MASK | + MEMERR_L2C_L2ESR_MULTICERR_MASK)) + edac_device_handle_ce(edac_dev, 0, 0, + edac_dev->ctl_name); + if (val & (MEMERR_L2C_L2ESR_UCERR_MASK | + MEMERR_L2C_L2ESR_MULTUCERR_MASK)) + edac_device_handle_ue(edac_dev, 0, 0, + edac_dev->ctl_name); + } + + /* Check if any memory request timed out on L2 cache */ + pg_d = ctx->pmd_csr + CPU_L2C_PAGE; + val = readl(pg_d + CPUX_L2C_L2RTOSR_PAGE_OFFSET); + if (val) { + val_lo = readl(pg_d + CPUX_L2C_L2RTOALR_PAGE_OFFSET); + val_hi = readl(pg_d + CPUX_L2C_L2RTOAHR_PAGE_OFFSET); + dev_err(edac_dev->dev, + "PMD%d L2C error L2C RTOSR 0x%08X @ 0x%08X.%08X\n", + ctx->pmd, val, val_hi, val_lo); + writel(val, pg_d + CPUX_L2C_L2RTOSR_PAGE_OFFSET); + } +} + +static void xgene_edac_pmd_check(struct edac_device_ctl_info *edac_dev) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + unsigned int pcp_hp_stat; + int i; + + xgene_edac_pcp_rd(ctx->edac, PCPHPERRINTSTS, &pcp_hp_stat); + if (!((PMD0_MERR_MASK << ctx->pmd) & pcp_hp_stat)) + return; + + /* Check CPU L1 error */ + for (i = 0; i < MAX_CPU_PER_PMD; i++) + xgene_edac_pmd_l1_check(edac_dev, i); + + /* Check CPU L2 error */ + xgene_edac_pmd_l2_check(edac_dev); +} + +static void xgene_edac_pmd_cpu_hw_cfg(struct edac_device_ctl_info *edac_dev, + int cpu) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + void __iomem *pg_f = ctx->pmd_csr + cpu * CPU_CSR_STRIDE + + CPU_MEMERR_CPU_PAGE; + + /* + * Enable CPU memory error: + * MEMERR_CPU_ICFESRA, MEMERR_CPU_LSUESRA, and MEMERR_CPU_MMUESRA + */ + writel(0x00000301, pg_f + MEMERR_CPU_ICFECR_PAGE_OFFSET); + writel(0x00000301, pg_f + MEMERR_CPU_LSUECR_PAGE_OFFSET); + writel(0x00000101, pg_f + MEMERR_CPU_MMUECR_PAGE_OFFSET); +} + +static bool xgene_edac_pmd_l2c_version1(void) +{ + /* Check all chips with PMD L2C version 1 HW */ + #define REVIDR_MINOR_REV(revidr) ((revidr) & 0x00000007) + + switch (MIDR_VARIANT(read_cpuid_id())) { + case 0: + switch (MIDR_REVISION(read_cpuid_id())) { + case 0: + + switch (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1))) { + case 1: + case 2: + return true; + }; + break; + case 1: + if (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1)) == 1) + return true; + break; + } + break; + case 1: + switch (MIDR_REVISION(read_cpuid_id())) { + case 0: + switch (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1))) { + case 1: + return true; + }; + break; + case 1: + switch (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1))) { + case 1: + case 0: + return true; + }; + break; + } + break; + } + + return false; +} + +static void xgene_edac_pmd_hw_cfg(struct edac_device_ctl_info *edac_dev) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + void __iomem *pg_d = ctx->pmd_csr + CPU_L2C_PAGE; + void __iomem *pg_e = ctx->pmd_csr + CPU_MEMERR_L2C_PAGE; + + /* Enable PMD memory error - MEMERR_L2C_L2ECR and L2C_L2RTOCR */ + writel(0x00000703, pg_e + MEMERR_L2C_L2ECR_PAGE_OFFSET); + /* Configure L2C HW request time out feature if supported */ + if (!xgene_edac_pmd_l2c_version1()) + writel(0x00000119, pg_d + CPUX_L2C_L2RTOCR_PAGE_OFFSET); +} + +static void xgene_edac_pmd_hw_ctl(struct edac_device_ctl_info *edac_dev, + bool enable) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + int i; + + /* Enable PMD error interrupt */ + if (edac_dev->op_state == OP_RUNNING_INTERRUPT) { + if (enable) + xgene_edac_pcp_clrbits(ctx->edac, PCPHPERRINTMSK, + PMD0_MERR_MASK << ctx->pmd); + else + xgene_edac_pcp_setbits(ctx->edac, PCPHPERRINTMSK, + PMD0_MERR_MASK << ctx->pmd); + } + + if (enable) { + xgene_edac_pmd_hw_cfg(edac_dev); + + /* Two CPUs per a PMD */ + for (i = 0; i < MAX_CPU_PER_PMD; i++) + xgene_edac_pmd_cpu_hw_cfg(edac_dev, i); + } +} + +static ssize_t xgene_edac_pmd_l1_inject_ctrl_write(struct file *file, + const char __user *data, + size_t count, loff_t *ppos) +{ + struct edac_device_ctl_info *edac_dev = file->private_data; + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + void __iomem *cpux_pg_f; + int i; + + for (i = 0; i < MAX_CPU_PER_PMD; i++) { + cpux_pg_f = ctx->pmd_csr + i * CPU_CSR_STRIDE + + CPU_MEMERR_CPU_PAGE; + + writel(MEMERR_CPU_ICFESR_MULTCERR_MASK | + MEMERR_CPU_ICFESR_CERR_MASK, + cpux_pg_f + MEMERR_CPU_ICFESRA_PAGE_OFFSET); + writel(MEMERR_CPU_LSUESR_MULTCERR_MASK | + MEMERR_CPU_LSUESR_CERR_MASK, + cpux_pg_f + MEMERR_CPU_LSUESRA_PAGE_OFFSET); + writel(MEMERR_CPU_MMUESR_MULTCERR_MASK | + MEMERR_CPU_MMUESR_CERR_MASK, + cpux_pg_f + MEMERR_CPU_MMUESRA_PAGE_OFFSET); + } + return count; +} + +static ssize_t xgene_edac_pmd_l2_inject_ctrl_write(struct file *file, + const char __user *data, + size_t count, loff_t *ppos) +{ + struct edac_device_ctl_info *edac_dev = file->private_data; + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + void __iomem *pg_e = ctx->pmd_csr + CPU_MEMERR_L2C_PAGE; + + writel(MEMERR_L2C_L2ESR_MULTUCERR_MASK | + MEMERR_L2C_L2ESR_MULTICERR_MASK | + MEMERR_L2C_L2ESR_UCERR_MASK | + MEMERR_L2C_L2ESR_ERR_MASK, + pg_e + MEMERR_L2C_L2ESRA_PAGE_OFFSET); + return count; +} + +static const struct file_operations xgene_edac_pmd_debug_inject_fops[] = { + { + .open = simple_open, + .write = xgene_edac_pmd_l1_inject_ctrl_write, + .llseek = generic_file_llseek, }, + { + .open = simple_open, + .write = xgene_edac_pmd_l2_inject_ctrl_write, + .llseek = generic_file_llseek, }, + { } +}; + +static void xgene_edac_pmd_create_debugfs_nodes( + struct edac_device_ctl_info *edac_dev) +{ + struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; + struct dentry *edac_debugfs; + char name[30]; + + if (!IS_ENABLED(CONFIG_EDAC_DEBUG)) + return; + + /* + * Todo: Switch to common EDAC debug file system for edac device + * when available. + */ + if (!ctx->edac->dfs) { + ctx->edac->dfs = debugfs_create_dir(edac_dev->dev->kobj.name, + NULL); + if (!ctx->edac->dfs) + return; + } + sprintf(name, "PMD%d", ctx->pmd); + edac_debugfs = debugfs_create_dir(name, ctx->edac->dfs); + if (!edac_debugfs) + return; + + debugfs_create_file("l1_inject_ctrl", S_IWUSR, edac_debugfs, edac_dev, + &xgene_edac_pmd_debug_inject_fops[0]); + debugfs_create_file("l2_inject_ctrl", S_IWUSR, edac_debugfs, edac_dev, + &xgene_edac_pmd_debug_inject_fops[1]); +} + +static int xgene_edac_pmd_available(u32 efuse, int pmd) +{ + return (efuse & (1 << pmd)) ? 0 : 1; +} + +static int xgene_edac_pmd_add(struct xgene_edac *edac, struct device_node *np) +{ + struct edac_device_ctl_info *edac_dev; + struct xgene_edac_pmd_ctx *ctx; + struct resource res; + char edac_name[10]; + u32 pmd; + int rc; + u32 val; + + if (!devres_open_group(edac->dev, xgene_edac_pmd_add, GFP_KERNEL)) + return -ENOMEM; + + /* Determine if this PMD is disabled */ + if (of_property_read_u32(np, "pmd-controller", &pmd)) { + dev_err(edac->dev, "no pmd-controller property\n"); + rc = -ENODEV; + goto err_group; + } + rc = regmap_read(edac->efuse_map, 0, &val); + if (rc) + goto err_group; + if (!xgene_edac_pmd_available(val, pmd)) { + rc = -ENODEV; + goto err_group; + } + + sprintf(edac_name, "l2c%d", pmd); + edac_dev = edac_device_alloc_ctl_info(sizeof(*ctx), + edac_name, 1, "l2c", 1, 2, NULL, + 0, edac_device_alloc_index()); + if (!edac_dev) { + rc = -ENOMEM; + goto err_group; + } + + ctx = edac_dev->pvt_info; + ctx->name = "xgene_pmd_err"; + ctx->pmd = pmd; + ctx->edac = edac; + ctx->edac_dev = edac_dev; + ctx->ddev = *edac->dev; + edac_dev->dev = &ctx->ddev; + edac_dev->ctl_name = ctx->name; + edac_dev->dev_name = ctx->name; + edac_dev->mod_name = EDAC_MOD_STR; + + rc = of_address_to_resource(np, 0, &res); + if (rc < 0) { + dev_err(edac->dev, "no PMD resource address\n"); + goto err_free; + } + ctx->pmd_csr = devm_ioremap_resource(edac->dev, &res); + if (IS_ERR(ctx->pmd_csr)) { + dev_err(edac->dev, + "devm_ioremap_resource failed for PMD resource address\n"); + rc = PTR_ERR(ctx->pmd_csr); + goto err_free; + } + + if (edac_op_state == EDAC_OPSTATE_POLL) + edac_dev->edac_check = xgene_edac_pmd_check; + + xgene_edac_pmd_create_debugfs_nodes(edac_dev); + + rc = edac_device_add_device(edac_dev); + if (rc > 0) { + dev_err(edac->dev, "edac_device_add_device failed\n"); + rc = -ENOMEM; + goto err_free; + } + + if (edac_op_state == EDAC_OPSTATE_INT) + edac_dev->op_state = OP_RUNNING_INTERRUPT; + + list_add(&ctx->next, &edac->pmds); + + xgene_edac_pmd_hw_ctl(edac_dev, 1); + + devres_remove_group(edac->dev, xgene_edac_pmd_add); + + dev_info(edac->dev, "X-Gene EDAC PMD%d registered\n", ctx->pmd); + return 0; + +err_free: + edac_device_free_ctl_info(edac_dev); +err_group: + devres_release_group(edac->dev, xgene_edac_pmd_add); + return rc; +} + +static int xgene_edac_pmd_remove(struct xgene_edac_pmd_ctx *pmd) +{ + struct edac_device_ctl_info *edac_dev = pmd->edac_dev; + + xgene_edac_pmd_hw_ctl(edac_dev, 0); + edac_device_del_device(edac_dev->dev); + edac_device_free_ctl_info(edac_dev); + return 0; +} + +static irqreturn_t xgene_edac_isr(int irq, void *dev_id) +{ + struct xgene_edac *ctx = dev_id; + struct xgene_edac_pmd_ctx *pmd; + unsigned int pcp_hp_stat; + unsigned int pcp_lp_stat; + + xgene_edac_pcp_rd(ctx, PCPHPERRINTSTS, &pcp_hp_stat); + xgene_edac_pcp_rd(ctx, PCPLPERRINTSTS, &pcp_lp_stat); + if ((MCU_UNCORR_ERR_MASK & pcp_hp_stat) || + (MCU_CTL_ERR_MASK & pcp_hp_stat) || + (MCU_CORR_ERR_MASK & pcp_lp_stat)) { + struct xgene_edac_mc_ctx *mcu; + + list_for_each_entry(mcu, &ctx->mcus, next) { + xgene_edac_mc_check(mcu->mci); + } + } + + list_for_each_entry(pmd, &ctx->pmds, next) { + if ((PMD0_MERR_MASK << pmd->pmd) & pcp_hp_stat) + xgene_edac_pmd_check(pmd->edac_dev); + } + + return IRQ_HANDLED; +} + +static int xgene_edac_probe(struct platform_device *pdev) +{ + struct xgene_edac *edac; + struct device_node *child; + struct resource *res; + int rc; + + edac = devm_kzalloc(&pdev->dev, sizeof(*edac), GFP_KERNEL); + if (!edac) + return -ENOMEM; + + edac->dev = &pdev->dev; + platform_set_drvdata(pdev, edac); + INIT_LIST_HEAD(&edac->mcus); + INIT_LIST_HEAD(&edac->pmds); + spin_lock_init(&edac->lock); + mutex_init(&edac->mc_lock); + + edac->csw_map = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "regmap-csw"); + if (IS_ERR(edac->csw_map)) { + dev_err(edac->dev, "unable to get syscon regmap csw\n"); + rc = PTR_ERR(edac->csw_map); + goto out_err; + } + + edac->mcba_map = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "regmap-mcba"); + if (IS_ERR(edac->mcba_map)) { + dev_err(edac->dev, "unable to get syscon regmap mcba\n"); + rc = PTR_ERR(edac->mcba_map); + goto out_err; + } + + edac->mcbb_map = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "regmap-mcbb"); + if (IS_ERR(edac->mcbb_map)) { + dev_err(edac->dev, "unable to get syscon regmap mcbb\n"); + rc = PTR_ERR(edac->mcbb_map); + goto out_err; + } + edac->efuse_map = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "regmap-efuse"); + if (IS_ERR(edac->efuse_map)) { + dev_err(edac->dev, "unable to get syscon regmap efuse\n"); + rc = PTR_ERR(edac->efuse_map); + goto out_err; + } + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + edac->pcp_csr = devm_ioremap_resource(&pdev->dev, res); + if (IS_ERR(edac->pcp_csr)) { + dev_err(&pdev->dev, "no PCP resource address\n"); + rc = PTR_ERR(edac->pcp_csr); + goto out_err; + } + + if (edac_op_state == EDAC_OPSTATE_INT) { + int irq; + int i; + + for (i = 0; i < 3; i++) { + irq = platform_get_irq(pdev, i); + if (irq < 0) { + dev_err(&pdev->dev, "No IRQ resource\n"); + rc = -EINVAL; + goto out_err; + } + rc = devm_request_irq(&pdev->dev, irq, + xgene_edac_isr, IRQF_SHARED, + dev_name(&pdev->dev), edac); + if (rc) { + dev_err(&pdev->dev, + "Could not request IRQ %d\n", irq); + goto out_err; + } + } + } + + for_each_child_of_node(pdev->dev.of_node, child) { + if (!of_device_is_available(child)) + continue; + if (of_device_is_compatible(child, "apm,xgene-edac-mc")) + xgene_edac_mc_add(edac, child); + if (of_device_is_compatible(child, "apm,xgene-edac-pmd")) + xgene_edac_pmd_add(edac, child); + } + + return 0; + +out_err: + return rc; +} + +static int xgene_edac_remove(struct platform_device *pdev) +{ + struct xgene_edac *edac = dev_get_drvdata(&pdev->dev); + struct xgene_edac_mc_ctx *mcu; + struct xgene_edac_mc_ctx *temp_mcu; + struct xgene_edac_pmd_ctx *pmd; + struct xgene_edac_pmd_ctx *temp_pmd; + + list_for_each_entry_safe(mcu, temp_mcu, &edac->mcus, next) { + xgene_edac_mc_remove(mcu); + } + + list_for_each_entry_safe(pmd, temp_pmd, &edac->pmds, next) { + xgene_edac_pmd_remove(pmd); + } + return 0; +} + +static const struct of_device_id xgene_edac_of_match[] = { + { .compatible = "apm,xgene-edac" }, + {}, +}; +MODULE_DEVICE_TABLE(of, xgene_edac_of_match); + +static struct platform_driver xgene_edac_driver = { + .probe = xgene_edac_probe, + .remove = xgene_edac_remove, + .driver = { + .name = "xgene-edac", + .owner = THIS_MODULE, + .of_match_table = xgene_edac_of_match, + }, +}; + +static int __init xgene_edac_init(void) +{ + int rc; + + /* Make sure error reporting method is sane */ + switch (edac_op_state) { + case EDAC_OPSTATE_POLL: + case EDAC_OPSTATE_INT: + break; + default: + edac_op_state = EDAC_OPSTATE_INT; + break; + } + + rc = platform_driver_register(&xgene_edac_driver); + if (rc) { + edac_printk(KERN_ERR, EDAC_MOD_STR, + "EDAC fails to register\n"); + goto reg_failed; + } + + return 0; + +reg_failed: + return rc; +} +module_init(xgene_edac_init); + +static void __exit xgene_edac_exit(void) +{ + platform_driver_unregister(&xgene_edac_driver); +} +module_exit(xgene_edac_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Feng Kan "); +MODULE_DESCRIPTION("APM X-Gene EDAC driver"); +module_param(edac_op_state, int, 0444); +MODULE_PARM_DESC(edac_op_state, + "EDAC error reporting state: 0=Poll, 2=Interrupt"); From 8f2ae6f30d5e935c9d7f18b869df66e307d41793 Mon Sep 17 00:00:00 2001 From: Loc Ho Date: Fri, 22 May 2015 17:33:00 -0600 Subject: [PATCH 08/28] arm64: Add APM X-Gene SoC EDAC DTS entries Add APM X-Gene SoC EDAC DTS entries. Signed-off-by: Loc Ho Acked-by: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dougthompson@xmission.com Cc: ijc+devicetree@hellion.org.uk Cc: jcm@redhat.com Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: mchehab@osg.samsung.com Cc: patches@apm.com Cc: robh+dt@kernel.org Link: http://lkml.kernel.org/r/1432337580-3750-6-git-send-email-lho@apm.com Signed-off-by: Borislav Petkov --- arch/arm64/boot/dts/apm/apm-storm.dtsi | 83 ++++++++++++++++++++++++++ 1 file changed, 83 insertions(+) diff --git a/arch/arm64/boot/dts/apm/apm-storm.dtsi b/arch/arm64/boot/dts/apm/apm-storm.dtsi index c8d3e0e86678..577799f0c5a3 100644 --- a/arch/arm64/boot/dts/apm/apm-storm.dtsi +++ b/arch/arm64/boot/dts/apm/apm-storm.dtsi @@ -374,6 +374,89 @@ dmaclk: dmaclk@1f27c000 { }; }; + csw: csw@7e200000 { + compatible = "apm,xgene-csw", "syscon"; + reg = <0x0 0x7e200000 0x0 0x1000>; + }; + + mcba: mcba@7e700000 { + compatible = "apm,xgene-mcb", "syscon"; + reg = <0x0 0x7e700000 0x0 0x1000>; + }; + + mcbb: mcbb@7e720000 { + compatible = "apm,xgene-mcb", "syscon"; + reg = <0x0 0x7e720000 0x0 0x1000>; + }; + + efuse: efuse@1054a000 { + compatible = "apm,xgene-efuse", "syscon"; + reg = <0x0 0x1054a000 0x0 0x20>; + }; + + edac@78800000 { + compatible = "apm,xgene-edac"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + regmap-csw = <&csw>; + regmap-mcba = <&mcba>; + regmap-mcbb = <&mcbb>; + regmap-efuse = <&efuse>; + reg = <0x0 0x78800000 0x0 0x100>; + interrupts = <0x0 0x20 0x4>, + <0x0 0x21 0x4>, + <0x0 0x27 0x4>; + + edacmc@7e800000 { + compatible = "apm,xgene-edac-mc"; + reg = <0x0 0x7e800000 0x0 0x1000>; + memory-controller = <0>; + }; + + edacmc@7e840000 { + compatible = "apm,xgene-edac-mc"; + reg = <0x0 0x7e840000 0x0 0x1000>; + memory-controller = <1>; + }; + + edacmc@7e880000 { + compatible = "apm,xgene-edac-mc"; + reg = <0x0 0x7e880000 0x0 0x1000>; + memory-controller = <2>; + }; + + edacmc@7e8c0000 { + compatible = "apm,xgene-edac-mc"; + reg = <0x0 0x7e8c0000 0x0 0x1000>; + memory-controller = <3>; + }; + + edacpmd@7c000000 { + compatible = "apm,xgene-edac-pmd"; + reg = <0x0 0x7c000000 0x0 0x200000>; + pmd-controller = <0>; + }; + + edacpmd@7c200000 { + compatible = "apm,xgene-edac-pmd"; + reg = <0x0 0x7c200000 0x0 0x200000>; + pmd-controller = <1>; + }; + + edacpmd@7c400000 { + compatible = "apm,xgene-edac-pmd"; + reg = <0x0 0x7c400000 0x0 0x200000>; + pmd-controller = <2>; + }; + + edacpmd@7c600000 { + compatible = "apm,xgene-edac-pmd"; + reg = <0x0 0x7c600000 0x0 0x200000>; + pmd-controller = <3>; + }; + }; + pcie0: pcie@1f2b0000 { status = "disabled"; device_type = "pci"; From cc14c5a80874cc5c5c7758c2f9f8c3153bbe645c Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 29 May 2015 22:01:00 +0200 Subject: [PATCH 09/28] EDAC, edac_stub: Drop arch-specific include contains only the arch-specific scrubbing function and is thus not needed in edac_stub.c. Kill it. Signed-off-by: Borislav Petkov --- drivers/edac/edac_stub.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/edac/edac_stub.c b/drivers/edac/edac_stub.c index 9d9e18aefaaa..ff07aae5b7fb 100644 --- a/drivers/edac/edac_stub.c +++ b/drivers/edac/edac_stub.c @@ -16,7 +16,6 @@ #include #include #include -#include int edac_op_state = EDAC_OPSTATE_INVAL; EXPORT_SYMBOL_GPL(edac_op_state); From 74210267a56eba10d6252215e72b095c257dcaf7 Mon Sep 17 00:00:00 2001 From: York Sun Date: Tue, 12 May 2015 18:03:41 +0800 Subject: [PATCH 10/28] EDAC, mpc8xxx: Adapt for FSL SoC Remove mpc83xx and mpc85xx as dependency. Signed-off-by: York Sun Acked-by: Johannes Thumshirn Cc: Mingkai.hu@freescale.com Link: http://lkml.kernel.org/r/1431425022-44766-1-git-send-email-Wenbin.Song@freescale.com Signed-off-by: songwenbin Signed-off-by: Borislav Petkov --- drivers/edac/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index 9016ec8a0295..8677ead2a8e1 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -263,10 +263,10 @@ config EDAC_SBRIDGE config EDAC_MPC85XX tristate "Freescale MPC83xx / MPC85xx" - depends on EDAC_MM_EDAC && FSL_SOC && (PPC_83xx || PPC_85xx) + depends on EDAC_MM_EDAC && FSL_SOC help Support for error detection and correction on the Freescale - MPC8349, MPC8560, MPC8540, MPC8548 + MPC8349, MPC8560, MPC8540, MPC8548, T4240 config EDAC_MV64X60 tristate "Marvell MV64x60" From 2ce39109a5062b06df440fbf2993ce9ab3bf6e08 Mon Sep 17 00:00:00 2001 From: York Sun Date: Tue, 12 May 2015 18:03:42 +0800 Subject: [PATCH 11/28] EDAC, mpc85xx: Extend error address to 64 bit Extend err_addr to cover 64 bits for DDR errors. Signed-off-by: York Sun Acked-by: Johannes Thumshirn Cc: Mingkai.hu@freescale.com Link: http://lkml.kernel.org/r/1431425022-44766-2-git-send-email-Wenbin.Song@freescale.com Signed-off-by: songwenbin Signed-off-by: Borislav Petkov --- drivers/edac/mpc85xx_edac.c | 10 +++++++--- drivers/edac/mpc85xx_edac.h | 1 + 2 files changed, 8 insertions(+), 3 deletions(-) diff --git a/drivers/edac/mpc85xx_edac.c b/drivers/edac/mpc85xx_edac.c index 68bf234bdfe6..23ef8e9f2c9a 100644 --- a/drivers/edac/mpc85xx_edac.c +++ b/drivers/edac/mpc85xx_edac.c @@ -811,6 +811,8 @@ static void sbe_ecc_decode(u32 cap_high, u32 cap_low, u32 cap_ecc, } } +#define make64(high, low) (((u64)(high) << 32) | (low)) + static void mpc85xx_mc_check(struct mem_ctl_info *mci) { struct mpc85xx_mc_pdata *pdata = mci->pvt_info; @@ -818,7 +820,7 @@ static void mpc85xx_mc_check(struct mem_ctl_info *mci) u32 bus_width; u32 err_detect; u32 syndrome; - u32 err_addr; + u64 err_addr; u32 pfn; int row_index; u32 cap_high; @@ -849,7 +851,9 @@ static void mpc85xx_mc_check(struct mem_ctl_info *mci) else syndrome &= 0xffff; - err_addr = in_be32(pdata->mc_vbase + MPC85XX_MC_CAPTURE_ADDRESS); + err_addr = make64( + in_be32(pdata->mc_vbase + MPC85XX_MC_CAPTURE_EXT_ADDRESS), + in_be32(pdata->mc_vbase + MPC85XX_MC_CAPTURE_ADDRESS)); pfn = err_addr >> PAGE_SHIFT; for (row_index = 0; row_index < mci->nr_csrows; row_index++) { @@ -886,7 +890,7 @@ static void mpc85xx_mc_check(struct mem_ctl_info *mci) mpc85xx_mc_printk(mci, KERN_ERR, "Captured Data / ECC:\t%#8.8x_%08x / %#2.2x\n", cap_high, cap_low, syndrome); - mpc85xx_mc_printk(mci, KERN_ERR, "Err addr: %#8.8x\n", err_addr); + mpc85xx_mc_printk(mci, KERN_ERR, "Err addr: %#8.8llx\n", err_addr); mpc85xx_mc_printk(mci, KERN_ERR, "PFN: %#8.8x\n", pfn); /* we are out of range */ diff --git a/drivers/edac/mpc85xx_edac.h b/drivers/edac/mpc85xx_edac.h index 4498baf9ce05..9352e88d53e5 100644 --- a/drivers/edac/mpc85xx_edac.h +++ b/drivers/edac/mpc85xx_edac.h @@ -43,6 +43,7 @@ #define MPC85XX_MC_ERR_INT_EN 0x0e48 #define MPC85XX_MC_CAPTURE_ATRIBUTES 0x0e4c #define MPC85XX_MC_CAPTURE_ADDRESS 0x0e50 +#define MPC85XX_MC_CAPTURE_EXT_ADDRESS 0x0e54 #define MPC85XX_MC_ERR_SBE 0x0e58 #define DSC_MEM_EN 0x80000000 From 451bb7fbccdc0f5d942227913afa5dea3afd12f7 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Mon, 1 Jun 2015 16:09:35 -0600 Subject: [PATCH 12/28] EDAC, xgene: Fix cpuid abuse The new x-gene EDAC driver incorrectly tried to figure out the version of one of its IP blocks by looking at the version of the CPU core, which is only vagely related. This removes the incorrect code and instead uses the version of the IP block in the compatible string where it belongs. Found using build testing on x86, which does not provide the arm64 cpuid interface. Signed-off-by: Arnd Bergmann [ Changed subnode to "apm,xgene-edac-pmd-v2", adjusted check. ] Signed-off-by: Loc Ho Cc: devicetree@vger.kernel.org Cc: dougthompson@xmission.com Cc: ijc+devicetree@hellion.org.uk Cc: jcm@redhat.com Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: mchehab@osg.samsung.com Cc: patches@apm.com Cc: robh+dt@kernel.org Link: http://lkml.kernel.org/r/3195065.IK73o60xya@wuerfel Signed-off-by: Borislav Petkov --- .../bindings/edac/apm-xgene-edac.txt | 3 +- drivers/edac/xgene_edac.c | 55 +++---------------- 2 files changed, 10 insertions(+), 48 deletions(-) diff --git a/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt b/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt index 480911c38ff9..78edb80002c8 100644 --- a/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt +++ b/Documentation/devicetree/bindings/edac/apm-xgene-edac.txt @@ -25,7 +25,8 @@ Required properties for memory controller subnode: - memory-controller : Instance number of the memory controller. Required properties for PMD subnode: -- compatible : Shall be "apm,xgene-edac-pmd". +- compatible : Shall be "apm,xgene-edac-pmd" or + "apm,xgene-edac-pmd-v2". - reg : First resource shall be the PMD resource. - pmd-controller : Instance number of the PMD controller. diff --git a/drivers/edac/xgene_edac.c b/drivers/edac/xgene_edac.c index b5158572f6ab..14636e4b6a08 100644 --- a/drivers/edac/xgene_edac.c +++ b/drivers/edac/xgene_edac.c @@ -523,6 +523,7 @@ struct xgene_edac_pmd_ctx { struct edac_device_ctl_info *edac_dev; void __iomem *pmd_csr; u32 pmd; + int version; }; static void xgene_edac_pmd_l1_check(struct edac_device_ctl_info *edac_dev, @@ -784,50 +785,6 @@ static void xgene_edac_pmd_cpu_hw_cfg(struct edac_device_ctl_info *edac_dev, writel(0x00000101, pg_f + MEMERR_CPU_MMUECR_PAGE_OFFSET); } -static bool xgene_edac_pmd_l2c_version1(void) -{ - /* Check all chips with PMD L2C version 1 HW */ - #define REVIDR_MINOR_REV(revidr) ((revidr) & 0x00000007) - - switch (MIDR_VARIANT(read_cpuid_id())) { - case 0: - switch (MIDR_REVISION(read_cpuid_id())) { - case 0: - - switch (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1))) { - case 1: - case 2: - return true; - }; - break; - case 1: - if (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1)) == 1) - return true; - break; - } - break; - case 1: - switch (MIDR_REVISION(read_cpuid_id())) { - case 0: - switch (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1))) { - case 1: - return true; - }; - break; - case 1: - switch (REVIDR_MINOR_REV(read_cpuid(REVIDR_EL1))) { - case 1: - case 0: - return true; - }; - break; - } - break; - } - - return false; -} - static void xgene_edac_pmd_hw_cfg(struct edac_device_ctl_info *edac_dev) { struct xgene_edac_pmd_ctx *ctx = edac_dev->pvt_info; @@ -837,7 +794,7 @@ static void xgene_edac_pmd_hw_cfg(struct edac_device_ctl_info *edac_dev) /* Enable PMD memory error - MEMERR_L2C_L2ECR and L2C_L2RTOCR */ writel(0x00000703, pg_e + MEMERR_L2C_L2ECR_PAGE_OFFSET); /* Configure L2C HW request time out feature if supported */ - if (!xgene_edac_pmd_l2c_version1()) + if (ctx->version > 1) writel(0x00000119, pg_d + CPUX_L2C_L2RTOCR_PAGE_OFFSET); } @@ -956,7 +913,8 @@ static int xgene_edac_pmd_available(u32 efuse, int pmd) return (efuse & (1 << pmd)) ? 0 : 1; } -static int xgene_edac_pmd_add(struct xgene_edac *edac, struct device_node *np) +static int xgene_edac_pmd_add(struct xgene_edac *edac, struct device_node *np, + int version) { struct edac_device_ctl_info *edac_dev; struct xgene_edac_pmd_ctx *ctx; @@ -998,6 +956,7 @@ static int xgene_edac_pmd_add(struct xgene_edac *edac, struct device_node *np) ctx->edac = edac; ctx->edac_dev = edac_dev; ctx->ddev = *edac->dev; + ctx->version = version; edac_dev->dev = &ctx->ddev; edac_dev->ctl_name = ctx->name; edac_dev->dev_name = ctx->name; @@ -1169,7 +1128,9 @@ static int xgene_edac_probe(struct platform_device *pdev) if (of_device_is_compatible(child, "apm,xgene-edac-mc")) xgene_edac_mc_add(edac, child); if (of_device_is_compatible(child, "apm,xgene-edac-pmd")) - xgene_edac_pmd_add(edac, child); + xgene_edac_pmd_add(edac, child, 1); + if (of_device_is_compatible(child, "apm,xgene-edac-pmd-v2")) + xgene_edac_pmd_add(edac, child, 2); } return 0; From e7f2ea1dbec30d607f0017706aa25b77477dd69c Mon Sep 17 00:00:00 2001 From: Aravind Gopalakrishnan Date: Tue, 2 Jun 2015 15:35:54 -0500 Subject: [PATCH 13/28] EDAC, mce_amd_inj: Use MCE_INJECT_GET macro for bank node too inj_bank_get() is generic enough that we can use the MCE_INJECT_GET macro instead. No functionality change. Signed-off-by: Aravind Gopalakrishnan Cc: linux-edac Cc: mchehab@osg.samsung.com Cc: x86-ml Link: http://lkml.kernel.org/r/1433277362-10911-2-git-send-email-Aravind.Gopalakrishnan@amd.com Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index f7681b553fd5..7f3a97a5ad17 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -187,13 +187,7 @@ static int inj_bank_set(void *data, u64 val) return 0; } -static int inj_bank_get(void *data, u64 *val) -{ - struct mce *m = (struct mce *)data; - - *val = m->bank; - return 0; -} +MCE_INJECT_GET(bank); DEFINE_SIMPLE_ATTRIBUTE(bank_fops, inj_bank_get, inj_bank_set, "%llu\n"); From 685d46d72ba456569d739cc3ff43577f4001d52f Mon Sep 17 00:00:00 2001 From: Aravind Gopalakrishnan Date: Wed, 27 May 2015 14:03:34 -0500 Subject: [PATCH 14/28] EDAC, mce_amd_inj: Read out number of MCE banks from the hardware The number of banks for a given processor is encoded in MSR_IA32_MCG_CAP[7:0]. So obtain the value from that MSR and use it for sanity checking in inj_bank_set() instead of doing a family/model check. Signed-off-by: Aravind Gopalakrishnan Cc: linux-edac Cc: mchehab@osg.samsung.com Link: http://lkml.kernel.org/r/1432753418-2985-3-git-send-email-Aravind.Gopalakrishnan@amd.com Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index 7f3a97a5ad17..15f6aa18b2f8 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -25,6 +25,8 @@ static struct mce i_mce; static struct dentry *dfs_inj; +static u8 n_banks; + #define MCE_INJECT_SET(reg) \ static int inj_##reg##_set(void *data, u64 val) \ { \ @@ -174,11 +176,9 @@ static int inj_bank_set(void *data, u64 val) { struct mce *m = (struct mce *)data; - if (val > 5) { - if (boot_cpu_data.x86 != 0x15 || val > 6) { - pr_err("Non-existent MCE bank: %llu\n", val); - return -EINVAL; - } + if (val >= n_banks) { + pr_err("Non-existent MCE bank: %llu\n", val); + return -EINVAL; } m->bank = val; @@ -207,6 +207,10 @@ static struct dfs_node { static int __init init_mce_inject(void) { int i; + u64 cap; + + rdmsrl(MSR_IA32_MCG_CAP, cap); + n_banks = cap & MCG_BANKCNT_MASK; dfs_inj = debugfs_create_dir("mce-inject", NULL); if (!dfs_inj) From 0451d14d05618e5fcbdc5017a30e3d609ddc8229 Mon Sep 17 00:00:00 2001 From: Aravind Gopalakrishnan Date: Tue, 2 Jun 2015 15:35:56 -0500 Subject: [PATCH 15/28] EDAC, mce_amd_inj: Modify flags attribute to use string arguments Use strings such as "hw" or "sw" to indicate the type of error injection to be performed. Current flags attribute derives the meanings of values that can be programmed into it from asm/mce.h. Moving to defined strings for the attribute allows this module to be self-sufficient and removes the dependency. Also, we can introduce new flags as and when needed without having to worry about conflicting with the flags already defined in asm/mce.h. Also, modify do_inject() to use the newly defined injection_type enum to figure out the injection mechanism we need to use Suggested-by: Borislav Petkov Signed-off-by: Aravind Gopalakrishnan Cc: linux-edac Cc: mchehab@osg.samsung.com Cc: x86-ml Link: http://lkml.kernel.org/r/1433277362-10911-4-git-send-email-Aravind.Gopalakrishnan@amd.com [ Use strstrip() return value. ] Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 83 +++++++++++++++++++++++++++++++++----- 1 file changed, 72 insertions(+), 11 deletions(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index 15f6aa18b2f8..2a0c829fef3d 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -15,6 +15,8 @@ #include #include #include +#include +#include #include #include "mce_amd.h" @@ -27,6 +29,23 @@ static struct dentry *dfs_inj; static u8 n_banks; +#define MAX_FLAG_OPT_SIZE 3 + +enum injection_type { + SW_INJ = 0, /* SW injection, simply decode the error */ + HW_INJ, /* Trigger a #MC */ + N_INJ_TYPES, +}; + +static const char * const flags_options[] = { + [SW_INJ] = "sw", + [HW_INJ] = "hw", + NULL +}; + +/* Set default injection to SW_INJ */ +enum injection_type inj_type = SW_INJ; + #define MCE_INJECT_SET(reg) \ static int inj_##reg##_set(void *data, u64 val) \ { \ @@ -81,24 +100,66 @@ static int toggle_hw_mce_inject(unsigned int cpu, bool enable) return err; } -static int flags_get(void *data, u64 *val) +static int __set_inj(const char *buf) { - struct mce *m = (struct mce *)data; + int i; - *val = m->inject_flags; - - return 0; + for (i = 0; i < N_INJ_TYPES; i++) { + if (!strncmp(flags_options[i], buf, strlen(flags_options[i]))) { + inj_type = i; + return 0; + } + } + return -EINVAL; } -static int flags_set(void *data, u64 val) +static ssize_t flags_read(struct file *filp, char __user *ubuf, + size_t cnt, loff_t *ppos) { - struct mce *m = (struct mce *)data; + char buf[MAX_FLAG_OPT_SIZE]; + int n; - m->inject_flags = (u8)val; - return 0; + n = sprintf(buf, "%s\n", flags_options[inj_type]); + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, n); } -DEFINE_SIMPLE_ATTRIBUTE(flags_fops, flags_get, flags_set, "%llu\n"); +static ssize_t flags_write(struct file *filp, const char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + char buf[MAX_FLAG_OPT_SIZE], *__buf; + int err; + size_t ret; + + if (cnt > MAX_FLAG_OPT_SIZE) + cnt = MAX_FLAG_OPT_SIZE; + + ret = cnt; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt - 1] = 0; + + /* strip whitespace */ + __buf = strstrip(buf); + + err = __set_inj(__buf); + if (err) { + pr_err("%s: Invalid flags value: %s\n", __func__, __buf); + return err; + } + + *ppos += ret; + + return ret; +} + +static const struct file_operations flags_fops = { + .read = flags_read, + .write = flags_write, + .llseek = generic_file_llseek, +}; /* * On which CPU to inject? @@ -130,7 +191,7 @@ static void do_inject(void) unsigned int cpu = i_mce.extcpu; u8 b = i_mce.bank; - if (!(i_mce.inject_flags & MCJ_EXCEPTION)) { + if (inj_type == SW_INJ) { amd_decode_mce(NULL, 0, &i_mce); return; } From 4c6034e8e1924b00bbe2c26e22c0b85c69709c99 Mon Sep 17 00:00:00 2001 From: Aravind Gopalakrishnan Date: Tue, 2 Jun 2015 15:35:58 -0500 Subject: [PATCH 16/28] EDAC, mce_amd_inj: Add individual permissions field to dfs_node Add per-file permissions to the dfs_fls[] array. In a later patch, we will add a README file that needs different permissions. Hence the move here to add a perm field. Signed-off-by: Aravind Gopalakrishnan Cc: linux-edac Cc: mchehab@osg.samsung.com Cc: x86-ml Link: http://lkml.kernel.org/r/1433277362-10911-6-git-send-email-Aravind.Gopalakrishnan@amd.com Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index 2a0c829fef3d..7501b5197a6a 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -256,13 +256,14 @@ static struct dfs_node { char *name; struct dentry *d; const struct file_operations *fops; + umode_t perm; } dfs_fls[] = { - { .name = "status", .fops = &status_fops }, - { .name = "misc", .fops = &misc_fops }, - { .name = "addr", .fops = &addr_fops }, - { .name = "bank", .fops = &bank_fops }, - { .name = "flags", .fops = &flags_fops }, - { .name = "cpu", .fops = &extcpu_fops }, + { .name = "status", .fops = &status_fops, .perm = S_IRUSR | S_IWUSR }, + { .name = "misc", .fops = &misc_fops, .perm = S_IRUSR | S_IWUSR }, + { .name = "addr", .fops = &addr_fops, .perm = S_IRUSR | S_IWUSR }, + { .name = "bank", .fops = &bank_fops, .perm = S_IRUSR | S_IWUSR }, + { .name = "flags", .fops = &flags_fops, .perm = S_IRUSR | S_IWUSR }, + { .name = "cpu", .fops = &extcpu_fops, .perm = S_IRUSR | S_IWUSR }, }; static int __init init_mce_inject(void) @@ -279,7 +280,7 @@ static int __init init_mce_inject(void) for (i = 0; i < ARRAY_SIZE(dfs_fls); i++) { dfs_fls[i].d = debugfs_create_file(dfs_fls[i].name, - S_IRUSR | S_IWUSR, + dfs_fls[i].perm, dfs_inj, &i_mce, dfs_fls[i].fops); From 99e21fea47f425f65f727c96ed8b64a7ac18e5cb Mon Sep 17 00:00:00 2001 From: Aravind Gopalakrishnan Date: Wed, 3 Jun 2015 17:13:58 +0200 Subject: [PATCH 17/28] EDAC, mce_amd_inj: Add README file Provide information about each injection file and its usage for ease of use and in-band documentation. This is a good idea adapted from ftrace. Signed-off-by: Aravind Gopalakrishnan Cc: linux-edac Cc: mchehab@osg.samsung.com Cc: Steven Rostedt Cc: x86-ml Link: http://lkml.kernel.org/r/1433277362-10911-7-git-send-email-Aravind.Gopalakrishnan@amd.com Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 46 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index 7501b5197a6a..b8385ccd1064 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -252,6 +252,51 @@ MCE_INJECT_GET(bank); DEFINE_SIMPLE_ATTRIBUTE(bank_fops, inj_bank_get, inj_bank_set, "%llu\n"); +static const char readme_msg[] = + "\nDescription of the files and their usages:\n\n" + "status: Set a value to be programmed into MCx_STATUS(bank)\n" + "\t The status bits provide insight into the type of\n" + "\t error that caused the MCE.\n\n" + "misc: Set value of MCx_MISC(bank)\n" + "\t misc register provides auxiliary info about the error. This\n" + "\t register is typically used for error thresholding purpose and\n" + "\t validity of the register is indicated by MCx_STATUS[MiscV]\n\n" + "addr: Error address value to be written to MCx_ADDR(bank)\n" + "\t This register is used to log address information associated\n" + "\t with the error.\n\n" + "Note: See respective BKDGs for the exact bit definitions of the\n" + "\t above registers as they mirror the MCi_[STATUS | MISC | ADDR]\n" + "\t hardware registers.\n\n" + "bank: Specify the bank you want to inject the error into.\n" + "\t The number of banks in a processor varies and is family/model\n" + "\t dependent. So, a sanity check performed while writing.\n" + "\t Writing to this file will trigger a #MC or APIC interrupts or\n" + "\t invoke the error decoder routines for AMD processors. The value\n" + "\t in 'flags' file decides which of above actions is triggered.\n\n" + "flags: Write to this file to speficy the error injection policy.\n" + "\t Allowed values:\n" + "\t\t\"sw\" - SW error injection, Only calls error decoder\n" + "\t\t\troutines to print error info in human readable format\n" + "\t\t\"hw\" - HW error injection, Forces a #MC,\n" + "\t\t\tcauses exception handler to handle the error\n" + "\t\t\tif UC or poll handler catches it if CE\n" + "\t\t\tWarning: Might cause system panic if MCx_STATUS[PCC]\n" + "\t\t\tis set. For debug purposes, consider setting\n" + "\t\t\t//mce/fake_panic\n" + "cpu: The cpu to inject the error on.\n\n"; + +static ssize_t +inj_readme_read(struct file *filp, char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + return simple_read_from_buffer(ubuf, cnt, ppos, + readme_msg, strlen(readme_msg)); +} + +static const struct file_operations readme_fops = { + .read = inj_readme_read, +}; + static struct dfs_node { char *name; struct dentry *d; @@ -264,6 +309,7 @@ static struct dfs_node { { .name = "bank", .fops = &bank_fops, .perm = S_IRUSR | S_IWUSR }, { .name = "flags", .fops = &flags_fops, .perm = S_IRUSR | S_IWUSR }, { .name = "cpu", .fops = &extcpu_fops, .perm = S_IRUSR | S_IWUSR }, + { .name = "README", .fops = &readme_fops, .perm = S_IRUSR | S_IRGRP | S_IROTH }, }; static int __init init_mce_inject(void) From f9ae487e04370e229a96c83a8c86510299712192 Mon Sep 17 00:00:00 2001 From: Thor Thayer Date: Thu, 4 Jun 2015 09:28:45 -0500 Subject: [PATCH 18/28] EDAC, altera: Generalize driver to use DT Memory size The Arria10 SOC uses a completely different SDRAM controller from the earlier CycloneV and ArriaV SoCs. The memory size is calculated in the bootloader and passed via the device tree. Using this device tree size is more generic than using the register fields to calculate the memory size for different SDRAM controllers. Signed-off-by: Thor Thayer Cc: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dinguyen@opensource.altera.com Cc: galak@codeaurora.org Cc: grant.likely@linaro.org Cc: ijc+devicetree@hellion.org.uk Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: m.chehab@samsung.com Cc: mark.rutland@arm.com Cc: pawel.moll@arm.com Cc: robh+dt@kernel.org Cc: tthayer.linux@gmail.com Link: http://lkml.kernel.org/r/1433428128-7292-2-git-send-email-tthayer@opensource.altera.com Signed-off-by: Borislav Petkov --- drivers/edac/altera_edac.c | 54 +++++++++++++++++--------------------- 1 file changed, 24 insertions(+), 30 deletions(-) diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c index 3c4929fda9d5..1d55c5b98024 100644 --- a/drivers/edac/altera_edac.c +++ b/drivers/edac/altera_edac.c @@ -219,36 +219,31 @@ static void altr_sdr_mc_create_debugfs_nodes(struct mem_ctl_info *mci) {} #endif -/* Get total memory size in bytes */ -static u32 altr_sdram_get_total_mem_size(struct regmap *mc_vbase) +/* Get total memory size from Open Firmware DTB */ +static unsigned long get_total_mem(void) { - u32 size, read_reg, row, bank, col, cs, width; + struct device_node *np = NULL; + const unsigned int *reg, *reg_end; + int len, sw, aw; + unsigned long start, size, total_mem = 0; - if (regmap_read(mc_vbase, DRAMADDRW_OFST, &read_reg) < 0) - return 0; + for_each_node_by_type(np, "memory") { + aw = of_n_addr_cells(np); + sw = of_n_size_cells(np); + reg = (const unsigned int *)of_get_property(np, "reg", &len); + reg_end = reg + (len / sizeof(u32)); - if (regmap_read(mc_vbase, DRAMIFWIDTH_OFST, &width) < 0) - return 0; - - col = (read_reg & DRAMADDRW_COLBIT_MASK) >> - DRAMADDRW_COLBIT_SHIFT; - row = (read_reg & DRAMADDRW_ROWBIT_MASK) >> - DRAMADDRW_ROWBIT_SHIFT; - bank = (read_reg & DRAMADDRW_BANKBIT_MASK) >> - DRAMADDRW_BANKBIT_SHIFT; - cs = (read_reg & DRAMADDRW_CSBIT_MASK) >> - DRAMADDRW_CSBIT_SHIFT; - - /* Correct for ECC as its not addressible */ - if (width == DRAMIFWIDTH_32B_ECC) - width = 32; - if (width == DRAMIFWIDTH_16B_ECC) - width = 16; - - /* calculate the SDRAM size base on this info */ - size = 1 << (row + bank + col); - size = size * cs * (width / 8); - return size; + total_mem = 0; + do { + start = of_read_number(reg, aw); + reg += aw; + size = of_read_number(reg, sw); + reg += sw; + total_mem += size; + } while (reg < reg_end); + } + edac_dbg(0, "total_mem 0x%lx\n", total_mem); + return total_mem; } static int altr_sdram_probe(struct platform_device *pdev) @@ -280,10 +275,9 @@ static int altr_sdram_probe(struct platform_device *pdev) } /* Grab memory size from device tree. */ - mem_size = altr_sdram_get_total_mem_size(mc_vbase); + mem_size = get_total_mem(); if (!mem_size) { - edac_printk(KERN_ERR, EDAC_MC, - "Unable to calculate memory size\n"); + edac_printk(KERN_ERR, EDAC_MC, "Unable to calculate memory size\n"); return -ENODEV; } From 143f4a5ac5af82a4055100c8f40b26187d5c20ba Mon Sep 17 00:00:00 2001 From: Thor Thayer Date: Thu, 4 Jun 2015 09:28:46 -0500 Subject: [PATCH 19/28] EDAC, altera: Refactor for Altera CycloneV SoC The Arria10 SoC uses a completely different SDRAM controller from the earlier CycloneV and ArriaV SoCs. This patch abstracts the SDRAM bits for the CycloneV/ArriaV SoCs in preparation for the Arria10 support. Signed-off-by: Thor Thayer Cc: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dinguyen@opensource.altera.com Cc: galak@codeaurora.org Cc: grant.likely@linaro.org Cc: ijc+devicetree@hellion.org.uk Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: m.chehab@samsung.com Cc: mark.rutland@arm.com Cc: pawel.moll@arm.com Cc: robh+dt@kernel.org Cc: tthayer.linux@gmail.com Link: http://lkml.kernel.org/r/1433428128-7292-3-git-send-email-tthayer@opensource.altera.com Signed-off-by: Borislav Petkov --- drivers/edac/altera_edac.c | 194 +++++++++++++++++-------------------- drivers/edac/altera_edac.h | 116 ++++++++++++++++++++++ 2 files changed, 206 insertions(+), 104 deletions(-) create mode 100644 drivers/edac/altera_edac.h diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c index 1d55c5b98024..4ac4e6c11ece 100644 --- a/drivers/edac/altera_edac.c +++ b/drivers/edac/altera_edac.c @@ -1,5 +1,5 @@ /* - * Copyright Altera Corporation (C) 2014. All rights reserved. + * Copyright Altera Corporation (C) 2014-2015. All rights reserved. * Copyright 2011-2012 Calxeda, Inc. * * This program is free software; you can redistribute it and/or modify it @@ -28,111 +28,64 @@ #include #include +#include "altera_edac.h" #include "edac_core.h" #include "edac_module.h" #define EDAC_MOD_STR "altera_edac" #define EDAC_VERSION "1" -/* SDRAM Controller CtrlCfg Register */ -#define CTLCFG_OFST 0x00 - -/* SDRAM Controller CtrlCfg Register Bit Masks */ -#define CTLCFG_ECC_EN 0x400 -#define CTLCFG_ECC_CORR_EN 0x800 -#define CTLCFG_GEN_SB_ERR 0x2000 -#define CTLCFG_GEN_DB_ERR 0x4000 - -#define CTLCFG_ECC_AUTO_EN (CTLCFG_ECC_EN | \ - CTLCFG_ECC_CORR_EN) - -/* SDRAM Controller Address Width Register */ -#define DRAMADDRW_OFST 0x2C - -/* SDRAM Controller Address Widths Field Register */ -#define DRAMADDRW_COLBIT_MASK 0x001F -#define DRAMADDRW_COLBIT_SHIFT 0 -#define DRAMADDRW_ROWBIT_MASK 0x03E0 -#define DRAMADDRW_ROWBIT_SHIFT 5 -#define DRAMADDRW_BANKBIT_MASK 0x1C00 -#define DRAMADDRW_BANKBIT_SHIFT 10 -#define DRAMADDRW_CSBIT_MASK 0xE000 -#define DRAMADDRW_CSBIT_SHIFT 13 - -/* SDRAM Controller Interface Data Width Register */ -#define DRAMIFWIDTH_OFST 0x30 - -/* SDRAM Controller Interface Data Width Defines */ -#define DRAMIFWIDTH_16B_ECC 24 -#define DRAMIFWIDTH_32B_ECC 40 - -/* SDRAM Controller DRAM Status Register */ -#define DRAMSTS_OFST 0x38 - -/* SDRAM Controller DRAM Status Register Bit Masks */ -#define DRAMSTS_SBEERR 0x04 -#define DRAMSTS_DBEERR 0x08 -#define DRAMSTS_CORR_DROP 0x10 - -/* SDRAM Controller DRAM IRQ Register */ -#define DRAMINTR_OFST 0x3C - -/* SDRAM Controller DRAM IRQ Register Bit Masks */ -#define DRAMINTR_INTREN 0x01 -#define DRAMINTR_SBEMASK 0x02 -#define DRAMINTR_DBEMASK 0x04 -#define DRAMINTR_CORRDROPMASK 0x08 -#define DRAMINTR_INTRCLR 0x10 - -/* SDRAM Controller Single Bit Error Count Register */ -#define SBECOUNT_OFST 0x40 - -/* SDRAM Controller Single Bit Error Count Register Bit Masks */ -#define SBECOUNT_MASK 0x0F - -/* SDRAM Controller Double Bit Error Count Register */ -#define DBECOUNT_OFST 0x44 - -/* SDRAM Controller Double Bit Error Count Register Bit Masks */ -#define DBECOUNT_MASK 0x0F - -/* SDRAM Controller ECC Error Address Register */ -#define ERRADDR_OFST 0x48 - -/* SDRAM Controller ECC Error Address Register Bit Masks */ -#define ERRADDR_MASK 0xFFFFFFFF - -/* Altera SDRAM Memory Controller data */ -struct altr_sdram_mc_data { - struct regmap *mc_vbase; +static const struct altr_sdram_prv_data c5_data = { + .ecc_ctrl_offset = CV_CTLCFG_OFST, + .ecc_ctl_en_mask = CV_CTLCFG_ECC_AUTO_EN, + .ecc_stat_offset = CV_DRAMSTS_OFST, + .ecc_stat_ce_mask = CV_DRAMSTS_SBEERR, + .ecc_stat_ue_mask = CV_DRAMSTS_DBEERR, + .ecc_saddr_offset = CV_ERRADDR_OFST, + .ecc_cecnt_offset = CV_SBECOUNT_OFST, + .ecc_uecnt_offset = CV_DBECOUNT_OFST, + .ecc_irq_en_offset = CV_DRAMINTR_OFST, + .ecc_irq_en_mask = CV_DRAMINTR_INTREN, + .ecc_irq_clr_offset = CV_DRAMINTR_OFST, + .ecc_irq_clr_mask = (CV_DRAMINTR_INTRCLR | CV_DRAMINTR_INTREN), + .ecc_cnt_rst_offset = CV_DRAMINTR_OFST, + .ecc_cnt_rst_mask = CV_DRAMINTR_INTRCLR, +#ifdef CONFIG_EDAC_DEBUG + .ce_ue_trgr_offset = CV_CTLCFG_OFST, + .ce_set_mask = CV_CTLCFG_GEN_SB_ERR, + .ue_set_mask = CV_CTLCFG_GEN_DB_ERR, +#endif }; static irqreturn_t altr_sdram_mc_err_handler(int irq, void *dev_id) { struct mem_ctl_info *mci = dev_id; struct altr_sdram_mc_data *drvdata = mci->pvt_info; + const struct altr_sdram_prv_data *priv = drvdata->data; u32 status, err_count, err_addr; /* Error Address is shared by both SBE & DBE */ - regmap_read(drvdata->mc_vbase, ERRADDR_OFST, &err_addr); + regmap_read(drvdata->mc_vbase, priv->ecc_saddr_offset, &err_addr); - regmap_read(drvdata->mc_vbase, DRAMSTS_OFST, &status); + regmap_read(drvdata->mc_vbase, priv->ecc_stat_offset, &status); - if (status & DRAMSTS_DBEERR) { - regmap_read(drvdata->mc_vbase, DBECOUNT_OFST, &err_count); + if (status & priv->ecc_stat_ue_mask) { + regmap_read(drvdata->mc_vbase, priv->ecc_uecnt_offset, + &err_count); panic("\nEDAC: [%d Uncorrectable errors @ 0x%08X]\n", err_count, err_addr); } - if (status & DRAMSTS_SBEERR) { - regmap_read(drvdata->mc_vbase, SBECOUNT_OFST, &err_count); + if (status & priv->ecc_stat_ce_mask) { + regmap_read(drvdata->mc_vbase, priv->ecc_cecnt_offset, + &err_count); edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, err_count, err_addr >> PAGE_SHIFT, err_addr & ~PAGE_MASK, 0, 0, 0, -1, mci->ctl_name, ""); } - regmap_write(drvdata->mc_vbase, DRAMINTR_OFST, - (DRAMINTR_INTRCLR | DRAMINTR_INTREN)); + regmap_write(drvdata->mc_vbase, priv->ecc_irq_clr_offset, + priv->ecc_irq_clr_mask); return IRQ_HANDLED; } @@ -144,6 +97,7 @@ static ssize_t altr_sdr_mc_err_inject_write(struct file *file, { struct mem_ctl_info *mci = file->private_data; struct altr_sdram_mc_data *drvdata = mci->pvt_info; + const struct altr_sdram_prv_data *priv = drvdata->data; u32 *ptemp; dma_addr_t dma_handle; u32 reg, read_reg; @@ -156,8 +110,9 @@ static ssize_t altr_sdr_mc_err_inject_write(struct file *file, return -ENOMEM; } - regmap_read(drvdata->mc_vbase, CTLCFG_OFST, &read_reg); - read_reg &= ~(CTLCFG_GEN_SB_ERR | CTLCFG_GEN_DB_ERR); + regmap_read(drvdata->mc_vbase, priv->ce_ue_trgr_offset, + &read_reg); + read_reg &= ~(priv->ce_set_mask | priv->ue_set_mask); /* Error are injected by writing a word while the SBE or DBE * bit in the CTLCFG register is set. Reading the word will @@ -166,20 +121,20 @@ static ssize_t altr_sdr_mc_err_inject_write(struct file *file, if (count == 3) { edac_printk(KERN_ALERT, EDAC_MC, "Inject Double bit error\n"); - regmap_write(drvdata->mc_vbase, CTLCFG_OFST, - (read_reg | CTLCFG_GEN_DB_ERR)); + regmap_write(drvdata->mc_vbase, priv->ce_ue_trgr_offset, + (read_reg | priv->ue_set_mask)); } else { edac_printk(KERN_ALERT, EDAC_MC, "Inject Single bit error\n"); - regmap_write(drvdata->mc_vbase, CTLCFG_OFST, - (read_reg | CTLCFG_GEN_SB_ERR)); + regmap_write(drvdata->mc_vbase, priv->ce_ue_trgr_offset, + (read_reg | priv->ce_set_mask)); } ptemp[0] = 0x5A5A5A5A; ptemp[1] = 0xA5A5A5A5; /* Clear the error injection bits */ - regmap_write(drvdata->mc_vbase, CTLCFG_OFST, read_reg); + regmap_write(drvdata->mc_vbase, priv->ce_ue_trgr_offset, read_reg); /* Ensure it has been written out */ wmb(); @@ -246,18 +201,29 @@ static unsigned long get_total_mem(void) return total_mem; } +static const struct of_device_id altr_sdram_ctrl_of_match[] = { + { .compatible = "altr,sdram-edac", .data = (void *)&c5_data}, + {}, +}; +MODULE_DEVICE_TABLE(of, altr_sdram_ctrl_of_match); + static int altr_sdram_probe(struct platform_device *pdev) { + const struct of_device_id *id; struct edac_mc_layer layers[2]; struct mem_ctl_info *mci; struct altr_sdram_mc_data *drvdata; + const struct altr_sdram_prv_data *priv; struct regmap *mc_vbase; struct dimm_info *dimm; - u32 read_reg, mem_size; - int irq; - int res = 0; + u32 read_reg; + int irq, res = 0; + unsigned long mem_size; + + id = of_match_device(altr_sdram_ctrl_of_match, &pdev->dev); + if (!id) + return -ENODEV; - /* Validate the SDRAM controller has ECC enabled */ /* Grab the register range from the sdr controller in device tree */ mc_vbase = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, "altr,sdr-syscon"); @@ -267,8 +233,13 @@ static int altr_sdram_probe(struct platform_device *pdev) return -ENODEV; } - if (regmap_read(mc_vbase, CTLCFG_OFST, &read_reg) || - ((read_reg & CTLCFG_ECC_AUTO_EN) != CTLCFG_ECC_AUTO_EN)) { + /* Check specific dependencies for the module */ + priv = of_match_node(altr_sdram_ctrl_of_match, + pdev->dev.of_node)->data; + + /* Validate the SDRAM controller has ECC enabled */ + if (regmap_read(mc_vbase, priv->ecc_ctrl_offset, &read_reg) || + ((read_reg & priv->ecc_ctl_en_mask) != priv->ecc_ctl_en_mask)) { edac_printk(KERN_ERR, EDAC_MC, "No ECC/ECC disabled [0x%08X]\n", read_reg); return -ENODEV; @@ -281,10 +252,27 @@ static int altr_sdram_probe(struct platform_device *pdev) return -ENODEV; } - /* Ensure the SDRAM Interrupt is disabled and cleared */ - if (regmap_write(mc_vbase, DRAMINTR_OFST, DRAMINTR_INTRCLR)) { + /* Ensure the SDRAM Interrupt is disabled */ + if (regmap_update_bits(mc_vbase, priv->ecc_irq_en_offset, + priv->ecc_irq_en_mask, 0)) { edac_printk(KERN_ERR, EDAC_MC, - "Error clearing SDRAM ECC IRQ\n"); + "Error disabling SDRAM ECC IRQ\n"); + return -ENODEV; + } + + /* Toggle to clear the SDRAM Error count */ + if (regmap_update_bits(mc_vbase, priv->ecc_cnt_rst_offset, + priv->ecc_cnt_rst_mask, + priv->ecc_cnt_rst_mask)) { + edac_printk(KERN_ERR, EDAC_MC, + "Error clearing SDRAM ECC count\n"); + return -ENODEV; + } + + if (regmap_update_bits(mc_vbase, priv->ecc_cnt_rst_offset, + priv->ecc_cnt_rst_mask, 0)) { + edac_printk(KERN_ERR, EDAC_MC, + "Error clearing SDRAM ECC count\n"); return -ENODEV; } @@ -309,9 +297,12 @@ static int altr_sdram_probe(struct platform_device *pdev) mci->pdev = &pdev->dev; drvdata = mci->pvt_info; drvdata->mc_vbase = mc_vbase; + drvdata->data = priv; platform_set_drvdata(pdev, mci); if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) { + edac_printk(KERN_ERR, EDAC_MC, + "Unable to get managed device resource\n"); res = -ENOMEM; goto free; } @@ -345,8 +336,9 @@ static int altr_sdram_probe(struct platform_device *pdev) goto err2; } - if (regmap_write(drvdata->mc_vbase, DRAMINTR_OFST, - (DRAMINTR_INTRCLR | DRAMINTR_INTREN))) { + /* Infrastructure ready - enable the IRQ */ + if (regmap_update_bits(drvdata->mc_vbase, priv->ecc_irq_en_offset, + priv->ecc_irq_en_mask, priv->ecc_irq_en_mask)) { edac_mc_printk(mci, KERN_ERR, "Error enabling SDRAM ECC IRQ\n"); res = -ENODEV; @@ -382,12 +374,6 @@ static int altr_sdram_remove(struct platform_device *pdev) return 0; } -static const struct of_device_id altr_sdram_ctrl_of_match[] = { - { .compatible = "altr,sdram-edac", }, - {}, -}; -MODULE_DEVICE_TABLE(of, altr_sdram_ctrl_of_match); - static struct platform_driver altr_sdram_edac_driver = { .probe = altr_sdram_probe, .remove = altr_sdram_remove, diff --git a/drivers/edac/altera_edac.h b/drivers/edac/altera_edac.h new file mode 100644 index 000000000000..b744d914d976 --- /dev/null +++ b/drivers/edac/altera_edac.h @@ -0,0 +1,116 @@ +/* + * + * Copyright (C) 2015 Altera Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see . + */ + +#ifndef _ALTERA_EDAC_H +#define _ALTERA_EDAC_H + +#include +#include + +/* SDRAM Controller CtrlCfg Register */ +#define CV_CTLCFG_OFST 0x00 + +/* SDRAM Controller CtrlCfg Register Bit Masks */ +#define CV_CTLCFG_ECC_EN 0x400 +#define CV_CTLCFG_ECC_CORR_EN 0x800 +#define CV_CTLCFG_GEN_SB_ERR 0x2000 +#define CV_CTLCFG_GEN_DB_ERR 0x4000 + +#define CV_CTLCFG_ECC_AUTO_EN (CV_CTLCFG_ECC_EN | \ + CV_CTLCFG_ECC_CORR_EN) + +/* SDRAM Controller Address Width Register */ +#define CV_DRAMADDRW_OFST 0x2C + +/* SDRAM Controller Address Widths Field Register */ +#define DRAMADDRW_COLBIT_MASK 0x001F +#define DRAMADDRW_COLBIT_SHIFT 0 +#define DRAMADDRW_ROWBIT_MASK 0x03E0 +#define DRAMADDRW_ROWBIT_SHIFT 5 +#define CV_DRAMADDRW_BANKBIT_MASK 0x1C00 +#define CV_DRAMADDRW_BANKBIT_SHIFT 10 +#define CV_DRAMADDRW_CSBIT_MASK 0xE000 +#define CV_DRAMADDRW_CSBIT_SHIFT 13 + +/* SDRAM Controller Interface Data Width Register */ +#define CV_DRAMIFWIDTH_OFST 0x30 + +/* SDRAM Controller Interface Data Width Defines */ +#define CV_DRAMIFWIDTH_16B_ECC 24 +#define CV_DRAMIFWIDTH_32B_ECC 40 + +/* SDRAM Controller DRAM Status Register */ +#define CV_DRAMSTS_OFST 0x38 + +/* SDRAM Controller DRAM Status Register Bit Masks */ +#define CV_DRAMSTS_SBEERR 0x04 +#define CV_DRAMSTS_DBEERR 0x08 +#define CV_DRAMSTS_CORR_DROP 0x10 + +/* SDRAM Controller DRAM IRQ Register */ +#define CV_DRAMINTR_OFST 0x3C + +/* SDRAM Controller DRAM IRQ Register Bit Masks */ +#define CV_DRAMINTR_INTREN 0x01 +#define CV_DRAMINTR_SBEMASK 0x02 +#define CV_DRAMINTR_DBEMASK 0x04 +#define CV_DRAMINTR_CORRDROPMASK 0x08 +#define CV_DRAMINTR_INTRCLR 0x10 + +/* SDRAM Controller Single Bit Error Count Register */ +#define CV_SBECOUNT_OFST 0x40 + +/* SDRAM Controller Double Bit Error Count Register */ +#define CV_DBECOUNT_OFST 0x44 + +/* SDRAM Controller ECC Error Address Register */ +#define CV_ERRADDR_OFST 0x48 + +struct altr_sdram_prv_data { + int ecc_ctrl_offset; + int ecc_ctl_en_mask; + int ecc_cecnt_offset; + int ecc_uecnt_offset; + int ecc_stat_offset; + int ecc_stat_ce_mask; + int ecc_stat_ue_mask; + int ecc_saddr_offset; + int ecc_daddr_offset; + int ecc_irq_en_offset; + int ecc_irq_en_mask; + int ecc_irq_clr_offset; + int ecc_irq_clr_mask; + int ecc_cnt_rst_offset; + int ecc_cnt_rst_mask; +#ifdef CONFIG_EDAC_DEBUG + struct edac_dev_sysfs_attribute *eccmgr_sysfs_attr; + int ecc_enable_mask; + int ce_set_mask; + int ue_set_mask; + int ce_ue_trgr_offset; +#endif +}; + +/* Altera SDRAM Memory Controller data */ +struct altr_sdram_mc_data { + struct regmap *mc_vbase; + int sb_irq; + int db_irq; + const struct altr_sdram_prv_data *data; +}; + +#endif /* #ifndef _ALTERA_EDAC_H */ From 73bcc942f4271fab2ea41e6a3992d3c2164faaa8 Mon Sep 17 00:00:00 2001 From: Thor Thayer Date: Thu, 4 Jun 2015 09:28:47 -0500 Subject: [PATCH 20/28] EDAC, altera: Add Arria10 EDAC support The Arria10 SDRAM and ECC system differs significantly from the Cyclone5 and Arria5 SoCs. This patch adds support for the Arria10 SoC. 1) IRQ handler needs to support SHARED IRQ 2) Support sberr and dberr address reporting. Signed-off-by: Thor Thayer Cc: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dinguyen@opensource.altera.com Cc: galak@codeaurora.org Cc: grant.likely@linaro.org Cc: ijc+devicetree@hellion.org.uk Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: m.chehab@samsung.com Cc: mark.rutland@arm.com Cc: pawel.moll@arm.com Cc: robh+dt@kernel.org Cc: tthayer.linux@gmail.com Link: http://lkml.kernel.org/r/1433428128-7292-4-git-send-email-tthayer@opensource.altera.com Signed-off-by: Borislav Petkov --- drivers/edac/altera_edac.c | 135 ++++++++++++++++++++++++++++++++----- drivers/edac/altera_edac.h | 85 +++++++++++++++++++++++ 2 files changed, 204 insertions(+), 16 deletions(-) diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c index 4ac4e6c11ece..182c741adf3e 100644 --- a/drivers/edac/altera_edac.c +++ b/drivers/edac/altera_edac.c @@ -42,6 +42,7 @@ static const struct altr_sdram_prv_data c5_data = { .ecc_stat_ce_mask = CV_DRAMSTS_SBEERR, .ecc_stat_ue_mask = CV_DRAMSTS_DBEERR, .ecc_saddr_offset = CV_ERRADDR_OFST, + .ecc_daddr_offset = CV_ERRADDR_OFST, .ecc_cecnt_offset = CV_SBECOUNT_OFST, .ecc_uecnt_offset = CV_DBECOUNT_OFST, .ecc_irq_en_offset = CV_DRAMINTR_OFST, @@ -57,37 +58,62 @@ static const struct altr_sdram_prv_data c5_data = { #endif }; +static const struct altr_sdram_prv_data a10_data = { + .ecc_ctrl_offset = A10_ECCCTRL1_OFST, + .ecc_ctl_en_mask = A10_ECCCTRL1_ECC_EN, + .ecc_stat_offset = A10_INTSTAT_OFST, + .ecc_stat_ce_mask = A10_INTSTAT_SBEERR, + .ecc_stat_ue_mask = A10_INTSTAT_DBEERR, + .ecc_saddr_offset = A10_SERRADDR_OFST, + .ecc_daddr_offset = A10_DERRADDR_OFST, + .ecc_irq_en_offset = A10_ERRINTEN_OFST, + .ecc_irq_en_mask = A10_ECC_IRQ_EN_MASK, + .ecc_irq_clr_offset = A10_INTSTAT_OFST, + .ecc_irq_clr_mask = (A10_INTSTAT_SBEERR | A10_INTSTAT_DBEERR), + .ecc_cnt_rst_offset = A10_ECCCTRL1_OFST, + .ecc_cnt_rst_mask = A10_ECC_CNT_RESET_MASK, +#ifdef CONFIG_EDAC_DEBUG + .ce_ue_trgr_offset = A10_DIAGINTTEST_OFST, + .ce_set_mask = A10_DIAGINT_TSERRA_MASK, + .ue_set_mask = A10_DIAGINT_TDERRA_MASK, +#endif +}; + static irqreturn_t altr_sdram_mc_err_handler(int irq, void *dev_id) { struct mem_ctl_info *mci = dev_id; struct altr_sdram_mc_data *drvdata = mci->pvt_info; const struct altr_sdram_prv_data *priv = drvdata->data; - u32 status, err_count, err_addr; - - /* Error Address is shared by both SBE & DBE */ - regmap_read(drvdata->mc_vbase, priv->ecc_saddr_offset, &err_addr); + u32 status, err_count = 1, err_addr; regmap_read(drvdata->mc_vbase, priv->ecc_stat_offset, &status); if (status & priv->ecc_stat_ue_mask) { - regmap_read(drvdata->mc_vbase, priv->ecc_uecnt_offset, - &err_count); + regmap_read(drvdata->mc_vbase, priv->ecc_daddr_offset, + &err_addr); + if (priv->ecc_uecnt_offset) + regmap_read(drvdata->mc_vbase, priv->ecc_uecnt_offset, + &err_count); panic("\nEDAC: [%d Uncorrectable errors @ 0x%08X]\n", err_count, err_addr); } if (status & priv->ecc_stat_ce_mask) { - regmap_read(drvdata->mc_vbase, priv->ecc_cecnt_offset, - &err_count); + regmap_read(drvdata->mc_vbase, priv->ecc_saddr_offset, + &err_addr); + if (priv->ecc_uecnt_offset) + regmap_read(drvdata->mc_vbase, priv->ecc_cecnt_offset, + &err_count); edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, err_count, err_addr >> PAGE_SHIFT, err_addr & ~PAGE_MASK, 0, 0, 0, -1, mci->ctl_name, ""); + /* Clear IRQ to resume */ + regmap_write(drvdata->mc_vbase, priv->ecc_irq_clr_offset, + priv->ecc_irq_clr_mask); + + return IRQ_HANDLED; } - - regmap_write(drvdata->mc_vbase, priv->ecc_irq_clr_offset, - priv->ecc_irq_clr_mask); - - return IRQ_HANDLED; + return IRQ_NONE; } #ifdef CONFIG_EDAC_DEBUG @@ -203,10 +229,60 @@ static unsigned long get_total_mem(void) static const struct of_device_id altr_sdram_ctrl_of_match[] = { { .compatible = "altr,sdram-edac", .data = (void *)&c5_data}, + { .compatible = "altr,sdram-edac-a10", .data = (void *)&a10_data}, {}, }; MODULE_DEVICE_TABLE(of, altr_sdram_ctrl_of_match); +static int a10_init(struct regmap *mc_vbase) +{ + if (regmap_update_bits(mc_vbase, A10_INTMODE_OFST, + A10_INTMODE_SB_INT, A10_INTMODE_SB_INT)) { + edac_printk(KERN_ERR, EDAC_MC, + "Error setting SB IRQ mode\n"); + return -ENODEV; + } + + if (regmap_write(mc_vbase, A10_SERRCNTREG_OFST, 1)) { + edac_printk(KERN_ERR, EDAC_MC, + "Error setting trigger count\n"); + return -ENODEV; + } + + return 0; +} + +static int a10_unmask_irq(struct platform_device *pdev, u32 mask) +{ + void __iomem *sm_base; + int ret = 0; + + if (!request_mem_region(A10_SYMAN_INTMASK_CLR, sizeof(u32), + dev_name(&pdev->dev))) { + edac_printk(KERN_ERR, EDAC_MC, + "Unable to request mem region\n"); + return -EBUSY; + } + + sm_base = ioremap(A10_SYMAN_INTMASK_CLR, sizeof(u32)); + if (!sm_base) { + edac_printk(KERN_ERR, EDAC_MC, + "Unable to ioremap device\n"); + + ret = -ENOMEM; + goto release; + } + + iowrite32(mask, sm_base); + + iounmap(sm_base); + +release: + release_mem_region(A10_SYMAN_INTMASK_CLR, sizeof(u32)); + + return ret; +} + static int altr_sdram_probe(struct platform_device *pdev) { const struct of_device_id *id; @@ -217,8 +293,8 @@ static int altr_sdram_probe(struct platform_device *pdev) struct regmap *mc_vbase; struct dimm_info *dimm; u32 read_reg; - int irq, res = 0; - unsigned long mem_size; + int irq, irq2, res = 0; + unsigned long mem_size, irqflags = 0; id = of_match_device(altr_sdram_ctrl_of_match, &pdev->dev); if (!id) @@ -283,6 +359,9 @@ static int altr_sdram_probe(struct platform_device *pdev) return -ENODEV; } + /* Arria10 has a 2nd IRQ */ + irq2 = platform_get_irq(pdev, 1); + layers[0].type = EDAC_MC_LAYER_CHIP_SELECT; layers[0].size = 1; layers[0].is_virt_csrow = true; @@ -327,8 +406,32 @@ static int altr_sdram_probe(struct platform_device *pdev) if (res < 0) goto err; + /* Only the Arria10 has separate IRQs */ + if (irq2 > 0) { + /* Arria10 specific initialization */ + res = a10_init(mc_vbase); + if (res < 0) + goto err2; + + res = devm_request_irq(&pdev->dev, irq2, + altr_sdram_mc_err_handler, + IRQF_SHARED, dev_name(&pdev->dev), mci); + if (res < 0) { + edac_mc_printk(mci, KERN_ERR, + "Unable to request irq %d\n", irq2); + res = -ENODEV; + goto err2; + } + + res = a10_unmask_irq(pdev, A10_DDR0_IRQ_MASK); + if (res < 0) + goto err2; + + irqflags = IRQF_SHARED; + } + res = devm_request_irq(&pdev->dev, irq, altr_sdram_mc_err_handler, - 0, dev_name(&pdev->dev), mci); + irqflags, dev_name(&pdev->dev), mci); if (res < 0) { edac_mc_printk(mci, KERN_ERR, "Unable to request irq %d\n", irq); diff --git a/drivers/edac/altera_edac.h b/drivers/edac/altera_edac.h index b744d914d976..7b64dc7c4eb7 100644 --- a/drivers/edac/altera_edac.h +++ b/drivers/edac/altera_edac.h @@ -80,6 +80,91 @@ /* SDRAM Controller ECC Error Address Register */ #define CV_ERRADDR_OFST 0x48 +/*-----------------------------------------*/ + +/* SDRAM Controller EccCtrl Register */ +#define A10_ECCCTRL1_OFST 0x00 + +/* SDRAM Controller EccCtrl Register Bit Masks */ +#define A10_ECCCTRL1_ECC_EN 0x001 +#define A10_ECCCTRL1_CNT_RST 0x010 +#define A10_ECCCTRL1_AWB_CNT_RST 0x100 +#define A10_ECC_CNT_RESET_MASK (A10_ECCCTRL1_CNT_RST | \ + A10_ECCCTRL1_AWB_CNT_RST) + +/* SDRAM Controller Address Width Register */ +#define CV_DRAMADDRW 0xFFC2502C +#define A10_DRAMADDRW 0xFFCFA0A8 + +/* SDRAM Controller Address Widths Field Register */ +#define DRAMADDRW_COLBIT_MASK 0x001F +#define DRAMADDRW_COLBIT_SHIFT 0 +#define DRAMADDRW_ROWBIT_MASK 0x03E0 +#define DRAMADDRW_ROWBIT_SHIFT 5 +#define CV_DRAMADDRW_BANKBIT_MASK 0x1C00 +#define CV_DRAMADDRW_BANKBIT_SHIFT 10 +#define CV_DRAMADDRW_CSBIT_MASK 0xE000 +#define CV_DRAMADDRW_CSBIT_SHIFT 13 + +#define A10_DRAMADDRW_BANKBIT_MASK 0x3C00 +#define A10_DRAMADDRW_BANKBIT_SHIFT 10 +#define A10_DRAMADDRW_GRPBIT_MASK 0xC000 +#define A10_DRAMADDRW_GRPBIT_SHIFT 14 +#define A10_DRAMADDRW_CSBIT_MASK 0x70000 +#define A10_DRAMADDRW_CSBIT_SHIFT 16 + +/* SDRAM Controller Interface Data Width Register */ +#define CV_DRAMIFWIDTH 0xFFC25030 +#define A10_DRAMIFWIDTH 0xFFCFB008 + +/* SDRAM Controller Interface Data Width Defines */ +#define CV_DRAMIFWIDTH_16B_ECC 24 +#define CV_DRAMIFWIDTH_32B_ECC 40 + +#define A10_DRAMIFWIDTH_16B 0x0 +#define A10_DRAMIFWIDTH_32B 0x1 +#define A10_DRAMIFWIDTH_64B 0x2 + +/* SDRAM Controller DRAM IRQ Register */ +#define A10_ERRINTEN_OFST 0x10 + +/* SDRAM Controller DRAM IRQ Register Bit Masks */ +#define A10_ERRINTEN_SERRINTEN 0x01 +#define A10_ERRINTEN_DERRINTEN 0x02 +#define A10_ECC_IRQ_EN_MASK (A10_ERRINTEN_SERRINTEN | \ + A10_ERRINTEN_DERRINTEN) + +/* SDRAM Interrupt Mode Register */ +#define A10_INTMODE_OFST 0x1C +#define A10_INTMODE_SB_INT 1 + +/* SDRAM Controller Error Status Register */ +#define A10_INTSTAT_OFST 0x20 + +/* SDRAM Controller Error Status Register Bit Masks */ +#define A10_INTSTAT_SBEERR 0x01 +#define A10_INTSTAT_DBEERR 0x02 + +/* SDRAM Controller ECC Error Address Register */ +#define A10_DERRADDR_OFST 0x2C +#define A10_SERRADDR_OFST 0x30 + +/* SDRAM Controller ECC Diagnostic Register */ +#define A10_DIAGINTTEST_OFST 0x24 + +#define A10_DIAGINT_TSERRA_MASK 0x0001 +#define A10_DIAGINT_TDERRA_MASK 0x0100 + +#define A10_SBERR_IRQ 34 +#define A10_DBERR_IRQ 32 + +/* SDRAM Single Bit Error Count Compare Set Register */ +#define A10_SERRCNTREG_OFST 0x3C + +#define A10_SYMAN_INTMASK_CLR 0xFFD06098 +#define A10_INTMASK_CLR_OFST 0x10 +#define A10_DDR0_IRQ_MASK BIT(17) + struct altr_sdram_prv_data { int ecc_ctrl_offset; int ecc_ctl_en_mask; From 54b4a8f57848bb08dcbdfba94b9b1ddef1c23358 Mon Sep 17 00:00:00 2001 From: Thor Thayer Date: Thu, 4 Jun 2015 09:28:48 -0500 Subject: [PATCH 21/28] arm: socfpga: dts: Add Arria10 SDRAM EDAC DTS support Add support for the Arria10 SDRAM EDAC. Update the bindings document for the new match string. Signed-off-by: Thor Thayer Cc: Arnd Bergmann Cc: devicetree@vger.kernel.org Cc: dinguyen@opensource.altera.com Cc: galak@codeaurora.org Cc: grant.likely@linaro.org Cc: ijc+devicetree@hellion.org.uk Cc: linux-arm-kernel@lists.infradead.org Cc: linux-edac Cc: m.chehab@samsung.com Cc: mark.rutland@arm.com Cc: pawel.moll@arm.com Cc: robh+dt@kernel.org Cc: tthayer.linux@gmail.com Link: http://lkml.kernel.org/r/1433428128-7292-5-git-send-email-tthayer@opensource.altera.com Signed-off-by: Borislav Petkov --- .../bindings/arm/altera/socfpga-sdram-edac.txt | 2 +- arch/arm/boot/dts/socfpga_arria10.dtsi | 11 +++++++++++ 2 files changed, 12 insertions(+), 1 deletion(-) diff --git a/Documentation/devicetree/bindings/arm/altera/socfpga-sdram-edac.txt b/Documentation/devicetree/bindings/arm/altera/socfpga-sdram-edac.txt index d0ce01da5c59..f5ad0ff69fae 100644 --- a/Documentation/devicetree/bindings/arm/altera/socfpga-sdram-edac.txt +++ b/Documentation/devicetree/bindings/arm/altera/socfpga-sdram-edac.txt @@ -2,7 +2,7 @@ Altera SOCFPGA SDRAM Error Detection & Correction [EDAC] The EDAC accesses a range of registers in the SDRAM controller. Required properties: -- compatible : should contain "altr,sdram-edac"; +- compatible : should contain "altr,sdram-edac" or "altr,sdram-edac-a10" - altr,sdr-syscon : phandle of the sdr module - interrupts : Should contain the SDRAM ECC IRQ in the appropriate format for the IRQ controller. diff --git a/arch/arm/boot/dts/socfpga_arria10.dtsi b/arch/arm/boot/dts/socfpga_arria10.dtsi index 8a05c47fd57f..4be75960a603 100644 --- a/arch/arm/boot/dts/socfpga_arria10.dtsi +++ b/arch/arm/boot/dts/socfpga_arria10.dtsi @@ -253,6 +253,17 @@ i2c4: i2c@ffc02600 { status = "disabled"; }; + sdr: sdr@ffc25000 { + compatible = "syscon"; + reg = <0xffcfb100 0x80>; + }; + + sdramedac { + compatible = "altr,sdram-edac-a10"; + altr,sdr-syscon = <&sdr>; + interrupts = <0 2 4>, <0 0 4>; + }; + L2: l2-cache@fffff000 { compatible = "arm,pl310-cache"; reg = <0xfffff000 0x1000>; From de2776787faa6859097a946e32f618f62fbce4c9 Mon Sep 17 00:00:00 2001 From: kbuild test robot Date: Fri, 5 Jun 2015 19:24:26 +0800 Subject: [PATCH 22/28] EDAC, mce_amd_inj: Make inj_type static It is used there only anyway. Signed-off-by: Fengguang Wu Cc: Aravind Gopalakrishnan Cc: Mauro Carvalho Chehab Cc: kbuild-all@01.org Cc: linux-edac Link: http://lkml.kernel.org/r/20150605112426.GA97073@lkp-sb04 Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index b8385ccd1064..eb8356e2df3e 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -44,7 +44,7 @@ static const char * const flags_options[] = { }; /* Set default injection to SW_INJ */ -enum injection_type inj_type = SW_INJ; +static enum injection_type inj_type = SW_INJ; #define MCE_INJECT_SET(reg) \ static int inj_##reg##_set(void *data, u64 val) \ From 6f2b6422d4787ffa72bd81abdc1b503dbf87a49a Mon Sep 17 00:00:00 2001 From: Alan Tull Date: Fri, 5 Jun 2015 08:49:15 -0500 Subject: [PATCH 23/28] EDAC, altera: Do not allow suspend when EDAC is enabled Suspend-to-RAM and EDAC support are mutually exclusive on SOCFPGA. If EDAC is enabled, it will prevent the platform from going into suspend. The reason is that the IRQ vectors for OCRAM reside on DDR and in Suspend-to-RAM mode we're executing out of OCRAM. If an ECC error occurs, we can't handle it so it was decided to make them mutually exclusive. Signed-off-by: Alan Tull Signed-off-by: Dinh Nguyen Cc: dinh.linux@gmail.com Cc: dougthompson@xmission.com Cc: linux-edac Cc: mchehab@osg.samsung.com Cc: tthayer@opensource.altera.com Link: http://lkml.kernel.org/r/1433512155-9906-1-git-send-email-dinguyen@opensource.altera.com Signed-off-by: Borislav Petkov --- drivers/edac/altera_edac.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c index 182c741adf3e..23ef0917483c 100644 --- a/drivers/edac/altera_edac.c +++ b/drivers/edac/altera_edac.c @@ -477,11 +477,31 @@ static int altr_sdram_remove(struct platform_device *pdev) return 0; } +/* + * If you want to suspend, need to disable EDAC by removing it + * from the device tree or defconfig. + */ +#ifdef CONFIG_PM +static int altr_sdram_prepare(struct device *dev) +{ + pr_err("Suspend not allowed when EDAC is enabled.\n"); + + return -EPERM; +} + +static const struct dev_pm_ops altr_sdram_pm_ops = { + .prepare = altr_sdram_prepare, +}; +#endif + static struct platform_driver altr_sdram_edac_driver = { .probe = altr_sdram_probe, .remove = altr_sdram_remove, .driver = { .name = "altr_sdram_edac", +#ifdef CONFIG_PM + .pm = &altr_sdram_pm_ops, +#endif .of_match_table = altr_sdram_ctrl_of_match, }, }; From f2f3dca1b7a66cbc91f45b59b8606bb325fa869f Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 10 Jun 2015 16:09:36 +0200 Subject: [PATCH 24/28] EDAC, mce_amd_inj: Cleanup and simplify README Save us an indentation level, widen to 80 cols, make the text more succinct and slender. Use i as the bank variable, same as what the documentation uses. Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 66 ++++++++++++++++++++------------------ 1 file changed, 35 insertions(+), 31 deletions(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index eb8356e2df3e..bfb0f350cb0e 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -253,37 +253,41 @@ MCE_INJECT_GET(bank); DEFINE_SIMPLE_ATTRIBUTE(bank_fops, inj_bank_get, inj_bank_set, "%llu\n"); static const char readme_msg[] = - "\nDescription of the files and their usages:\n\n" - "status: Set a value to be programmed into MCx_STATUS(bank)\n" - "\t The status bits provide insight into the type of\n" - "\t error that caused the MCE.\n\n" - "misc: Set value of MCx_MISC(bank)\n" - "\t misc register provides auxiliary info about the error. This\n" - "\t register is typically used for error thresholding purpose and\n" - "\t validity of the register is indicated by MCx_STATUS[MiscV]\n\n" - "addr: Error address value to be written to MCx_ADDR(bank)\n" - "\t This register is used to log address information associated\n" - "\t with the error.\n\n" - "Note: See respective BKDGs for the exact bit definitions of the\n" - "\t above registers as they mirror the MCi_[STATUS | MISC | ADDR]\n" - "\t hardware registers.\n\n" - "bank: Specify the bank you want to inject the error into.\n" - "\t The number of banks in a processor varies and is family/model\n" - "\t dependent. So, a sanity check performed while writing.\n" - "\t Writing to this file will trigger a #MC or APIC interrupts or\n" - "\t invoke the error decoder routines for AMD processors. The value\n" - "\t in 'flags' file decides which of above actions is triggered.\n\n" - "flags: Write to this file to speficy the error injection policy.\n" - "\t Allowed values:\n" - "\t\t\"sw\" - SW error injection, Only calls error decoder\n" - "\t\t\troutines to print error info in human readable format\n" - "\t\t\"hw\" - HW error injection, Forces a #MC,\n" - "\t\t\tcauses exception handler to handle the error\n" - "\t\t\tif UC or poll handler catches it if CE\n" - "\t\t\tWarning: Might cause system panic if MCx_STATUS[PCC]\n" - "\t\t\tis set. For debug purposes, consider setting\n" - "\t\t\t//mce/fake_panic\n" - "cpu: The cpu to inject the error on.\n\n"; +"Description of the files and their usages:\n" +"\n" +"Note1: i refers to the bank number below.\n" +"Note2: See respective BKDGs for the exact bit definitions of the files below\n" +"as they mirror the hardware registers.\n" +"\n" +"status:\t Set MCi_STATUS: the bits in that MSR control the error type and\n" +"\t attributes of the error which caused the MCE.\n" +"\n" +"misc:\t Set MCi_MISC: provide auxiliary info about the error. It is mostly\n" +"\t used for error thresholding purposes and its validity is indicated by\n" +"\t MCi_STATUS[MiscV].\n" +"\n" +"addr:\t Error address value to be written to MCi_ADDR. Log address information\n" +"\t associated with the error.\n" +"\n" +"cpu:\t The CPU to inject the error on.\n" +"\n" +"bank:\t Specify the bank you want to inject the error into: the number of\n" +"\t banks in a processor varies and is family/model-specific, therefore, the\n" +"\t supplied value is sanity-checked. Setting the bank value also triggers the\n" +"\t injection.\n" +"\n" +"flags:\t Injection type to be performed. Writing to this file will trigger a\n" +"\t real machine check, an APIC interrupt or invoke the error decoder routines\n" +"\t for AMD processors.\n" +"\n" +"\t Allowed error injection types:\n" +"\t - \"sw\": Software error injection. Decode error to a human-readable \n" +"\t format only. Safe to use.\n" +"\t - \"hw\": Hardware error injection. Causes the #MC exception handler to \n" +"\t handle the error. Be warned: might cause system panic if MCi_STATUS[PCC] \n" +"\t is set. Therefore, consider setting (debugfs_mountpoint)/mce/fake_panic \n" +"\t before injecting.\n" +"\n"; static ssize_t inj_readme_read(struct file *filp, char __user *ubuf, From 6d1e9bf5b04887a2220fbb44be074e9e8376e332 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 10 Jun 2015 16:17:13 +0200 Subject: [PATCH 25/28] EDAC, mce_amd_inj: Move bit preparations before the injection We do get_online_cpus() and then start noodling with the bits. Do that *before* we grab the hotplug lock. Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index bfb0f350cb0e..b92986b03f93 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -196,16 +196,16 @@ static void do_inject(void) return; } - get_online_cpus(); - if (!cpu_online(cpu)) - goto err; - /* prep MCE global settings for the injection */ mcg_status = MCG_STATUS_MCIP | MCG_STATUS_EIPV; if (!(i_mce.status & MCI_STATUS_PCC)) mcg_status |= MCG_STATUS_RIPV; + get_online_cpus(); + if (!cpu_online(cpu)) + goto err; + toggle_hw_mce_inject(cpu, true); wrmsr_on_cpu(cpu, MSR_IA32_MCG_STATUS, From cda9459da736481c6ff7e4368c994b98533a28cc Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Wed, 10 Jun 2015 16:20:56 +0200 Subject: [PATCH 26/28] EDAC, mce_amd_inj: Set MISCV on injection When during injection we populate MCi_MISC by writing into misc, we need to set the MiscV bit in the corresponding MCi_STATUS register which denotes that there's valid info in the MCi_MISC register. Signed-off-by: Borislav Petkov --- drivers/edac/mce_amd_inj.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/edac/mce_amd_inj.c b/drivers/edac/mce_amd_inj.c index b92986b03f93..4c73e4d03d46 100644 --- a/drivers/edac/mce_amd_inj.c +++ b/drivers/edac/mce_amd_inj.c @@ -191,6 +191,9 @@ static void do_inject(void) unsigned int cpu = i_mce.extcpu; u8 b = i_mce.bank; + if (i_mce.misc) + i_mce.status |= MCI_STATUS_MISCV; + if (inj_type == SW_INJ) { amd_decode_mce(NULL, 0, &i_mce); return; From 3aae9edd5a63e226baf3375bb8f7e8d05f5d9098 Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Fri, 19 Jun 2015 09:18:34 +0300 Subject: [PATCH 27/28] EDAC: Fix typos in Documentation/edac.txt Fix various typos in Documentation/edac.txt. Signed-off-by: Rami Rosen Link: http://lkml.kernel.org/r/1434694714-2924-1-git-send-email-ramirose@gmail.com Signed-off-by: Borislav Petkov --- Documentation/edac.txt | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/Documentation/edac.txt b/Documentation/edac.txt index 73fff13e848f..4df786e73e87 100644 --- a/Documentation/edac.txt +++ b/Documentation/edac.txt @@ -25,7 +25,7 @@ first time, it was renamed to 'EDAC'. The bluesmoke project at sourceforge.net is now utilized as a 'staging area' for EDAC development, before it is sent upstream to kernel.org -At the bluesmoke/EDAC project site is a series of quilt patches against +At the bluesmoke/EDAC project site, there is a series of quilt patches against recent kernels, stored in a SVN repository. For easier downloading, there is also a tarball snapshot available. @@ -235,7 +235,7 @@ In 'mcX' directories are EDAC control and attribute files for this 'X' instance of the memory controllers. For a description of the sysfs API, please see: - Documentation/ABI/testing/sysfs/devices-edac + Documentation/ABI/testing/sysfs-devices-edac ============================================================================ @@ -276,7 +276,7 @@ Total memory managed by this csrow attribute file: 'size_mb' - This attribute file displays, in count of megabytes, of memory + This attribute file displays, in count of megabytes, the memory that this csrow contains. @@ -516,7 +516,7 @@ Panic on PCI PARITY Error: 'panic_on_pci_parity' - This control files enables or disables panicking when a parity + This control file enables or disables panicking when a parity error has been detected. @@ -617,7 +617,7 @@ The 'test_device_edac' device adds 4 attributes and 1 control: reset all the above counters. -Use of the 'test_device_edac' driver should any others to create their own +Use of the 'test_device_edac' driver should enable any others to create their own unique drivers for their hardware systems. The 'test_device_edac' sample driver is located at the @@ -633,7 +633,7 @@ of the driver. Due to the way Nehalem exports Memory Controller data, some adjustments were done at i7core_edac driver. This chapter will cover those differences -1) On Nehalem, there are one Memory Controller per Quick Patch Interconnect +1) On Nehalem, there is one Memory Controller per Quick Patch Interconnect (QPI). At the driver, the term "socket" means one QPI. This is associated with a physical CPU socket. @@ -642,7 +642,7 @@ were done at i7core_edac driver. This chapter will cover those differences Each channel can have up to 3 DIMMs. The minimum known unity is DIMMs. There are no information about csrows. - As EDAC API maps the minimum unity is csrows, the driver sequencially + As EDAC API maps the minimum unity is csrows, the driver sequentially maps channel/dimm into different csrows. For example, supposing the following layout: @@ -664,7 +664,7 @@ exports one Each QPI is exported as a different memory controller. -2) Nehalem MC has the hability to generate errors. The driver implements this +2) Nehalem MC has the ability to generate errors. The driver implements this functionality via some error injection nodes: For injecting a memory error, there are some sysfs nodes, under @@ -771,5 +771,5 @@ exports one The standard error counters are generated when an mcelog error is received by the driver. Since, with udimm, this is counted by software, it is - possible that some errors could be lost. With rdimm's, they displays the + possible that some errors could be lost. With rdimm's, they display the contents of the registers From 043b43180efee8dcc41dde5ca710827b26d17510 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 19 Jun 2015 11:47:17 +0200 Subject: [PATCH 28/28] EDAC: Update Documentation/edac.txt Do some initial cleanup, more probably will come. - Move credits section to the end - Update maintainers - Drop sourceforge reference - project is long upstream now - Reformat sections - Reformat paragraphs - Clarify text - Bring it up-to-date - Drop useless "future hardware scanning" section Signed-off-by: Borislav Petkov --- Documentation/edac.txt | 275 ++++++++++++++++++++--------------------- 1 file changed, 131 insertions(+), 144 deletions(-) diff --git a/Documentation/edac.txt b/Documentation/edac.txt index 4df786e73e87..0cf27a3544a5 100644 --- a/Documentation/edac.txt +++ b/Documentation/edac.txt @@ -1,53 +1,34 @@ - - EDAC - Error Detection And Correction - -Written by Doug Thompson -7 Dec 2005 -17 Jul 2007 Updated - -(c) Mauro Carvalho Chehab -05 Aug 2009 Nehalem interface - -EDAC is maintained and written by: - - Doug Thompson, Dave Jiang, Dave Peterson et al, - original author: Thayne Harbaugh, - -Contact: - website: bluesmoke.sourceforge.net - mailing list: bluesmoke-devel@lists.sourceforge.net +===================================== "bluesmoke" was the name for this device driver when it was "out-of-tree" and maintained at sourceforge.net. When it was pushed into 2.6.16 for the first time, it was renamed to 'EDAC'. -The bluesmoke project at sourceforge.net is now utilized as a 'staging area' -for EDAC development, before it is sent upstream to kernel.org +PURPOSE +------- -At the bluesmoke/EDAC project site, there is a series of quilt patches against -recent kernels, stored in a SVN repository. For easier downloading, there -is also a tarball snapshot available. - -============================================================================ -EDAC PURPOSE - -The 'edac' kernel module goal is to detect and report errors that occur -within the computer system running under linux. +The 'edac' kernel module's goal is to detect and report hardware errors +that occur within the computer system running under linux. MEMORY +------ -In the initial release, memory Correctable Errors (CE) and Uncorrectable -Errors (UE) are the primary errors being harvested. These types of errors -are harvested by the 'edac_mc' class of device. +Memory Correctable Errors (CE) and Uncorrectable Errors (UE) are the +primary errors being harvested. These types of errors are harvested by +the 'edac_mc' device. Detecting CE events, then harvesting those events and reporting them, -CAN be a predictor of future UE events. With CE events, the system can -continue to operate, but with less safety. Preventive maintenance and -proactive part replacement of memory DIMMs exhibiting CEs can reduce -the likelihood of the dreaded UE events and system 'panics'. +*can* but must not necessarily be a predictor of future UE events. With +CE events only, the system can and will continue to operate as no data +has been damaged yet. -NON-MEMORY +However, preventive maintenance and proactive part replacement of memory +DIMMs exhibiting CEs can reduce the likelihood of the dreaded UE events +and system panics. + +OTHER HARDWARE ELEMENTS +----------------------- A new feature for EDAC, the edac_device class of device, was added in the 2.6.23 version of the kernel. @@ -56,70 +37,57 @@ This new device type allows for non-memory type of ECC hardware detectors to have their states harvested and presented to userspace via the sysfs interface. -Some architectures have ECC detectors for L1, L2 and L3 caches, along with DMA -engines, fabric switches, main data path switches, interconnections, -and various other hardware data paths. If the hardware reports it, then -a edac_device device probably can be constructed to harvest and present -that to userspace. +Some architectures have ECC detectors for L1, L2 and L3 caches, +along with DMA engines, fabric switches, main data path switches, +interconnections, and various other hardware data paths. If the hardware +reports it, then a edac_device device probably can be constructed to +harvest and present that to userspace. PCI BUS SCANNING +---------------- -In addition, PCI Bus Parity and SERR Errors are scanned for on PCI devices -in order to determine if errors are occurring on data transfers. +In addition, PCI devices are scanned for PCI Bus Parity and SERR Errors +in order to determine if errors are occurring during data transfers. The presence of PCI Parity errors must be examined with a grain of salt. -There are several add-in adapters that do NOT follow the PCI specification +There are several add-in adapters that do *not* follow the PCI specification with regards to Parity generation and reporting. The specification says the vendor should tie the parity status bits to 0 if they do not intend to generate parity. Some vendors do not do this, and thus the parity bit can "float" giving false positives. -In the kernel there is a PCI device attribute located in sysfs that is -checked by the EDAC PCI scanning code. If that attribute is set, -PCI parity/error scanning is skipped for that device. The attribute -is: +There is a PCI device attribute located in sysfs that is checked by +the EDAC PCI scanning code. If that attribute is set, PCI parity/error +scanning is skipped for that device. The attribute is: broken_parity_status -as is located in /sys/devices/pci/0000:XX:YY.Z directories for +and is located in /sys/devices/pci/0000:XX:YY.Z directories for PCI devices. -FUTURE HARDWARE SCANNING -EDAC will have future error detectors that will be integrated with -EDAC or added to it, in the following list: - - MCE Machine Check Exception - MCA Machine Check Architecture - NMI NMI notification of ECC errors - MSRs Machine Specific Register error cases - and other mechanisms. - -These errors are usually bus errors, ECC errors, thermal throttling -and the like. - - -============================================================================ -EDAC VERSIONING +VERSIONING +---------- EDAC is composed of a "core" module (edac_core.ko) and several Memory -Controller (MC) driver modules. On a given system, the CORE -is loaded and one MC driver will be loaded. Both the CORE and -the MC driver (or edac_device driver) have individual versions that reflect -current release level of their respective modules. +Controller (MC) driver modules. On a given system, the CORE is loaded +and one MC driver will be loaded. Both the CORE and the MC driver (or +edac_device driver) have individual versions that reflect current +release level of their respective modules. -Thus, to "report" on what version a system is running, one must report both -the CORE's and the MC driver's versions. +Thus, to "report" on what version a system is running, one must report +both the CORE's and the MC driver's versions. LOADING +------- -If 'edac' was statically linked with the kernel then no loading is -necessary. If 'edac' was built as modules then simply modprobe the -'edac' pieces that you need. You should be able to modprobe -hardware-specific modules and have the dependencies load the necessary core -modules. +If 'edac' was statically linked with the kernel then no loading +is necessary. If 'edac' was built as modules then simply modprobe +the 'edac' pieces that you need. You should be able to modprobe +hardware-specific modules and have the dependencies load the necessary +core modules. Example: @@ -129,35 +97,33 @@ loads both the amd76x_edac.ko memory controller module and the edac_mc.ko core module. -============================================================================ -EDAC sysfs INTERFACE +SYSFS INTERFACE +--------------- -EDAC presents a 'sysfs' interface for control, reporting and attribute -reporting purposes. +EDAC presents a 'sysfs' interface for control and reporting purposes. It +lives in the /sys/devices/system/edac directory. -EDAC lives in the /sys/devices/system/edac directory. - -Within this directory there currently reside 2 'edac' components: +Within this directory there currently reside 2 components: mc memory controller(s) system pci PCI control and status system -============================================================================ + Memory Controller (mc) Model +---------------------------- -First a background on the memory controller's model abstracted in EDAC. -Each 'mc' device controls a set of DIMM memory modules. These modules are -laid out in a Chip-Select Row (csrowX) and Channel table (chX). There can -be multiple csrows and multiple channels. +Each 'mc' device controls a set of DIMM memory modules. These modules +are laid out in a Chip-Select Row (csrowX) and Channel table (chX). +There can be multiple csrows and multiple channels. -Memory controllers allow for several csrows, with 8 csrows being a typical value. -Yet, the actual number of csrows depends on the electrical "loading" -of a given motherboard, memory controller and DIMM characteristics. +Memory controllers allow for several csrows, with 8 csrows being a +typical value. Yet, the actual number of csrows depends on the layout of +a given motherboard, memory controller and DIMM characteristics. -Dual channels allows for 128 bit data transfers to the CPU from memory. -Some newer chipsets allow for more than 2 channels, like Fully Buffered DIMMs -(FB-DIMMs). The following example will assume 2 channels: +Dual channels allows for 128 bit data transfers to/from the CPU from/to +memory. Some newer chipsets allow for more than 2 channels, like Fully +Buffered DIMMs (FB-DIMMs). The following example will assume 2 channels: Channel 0 Channel 1 @@ -179,12 +145,12 @@ for memory DIMMs: DIMM_A1 DIMM_B1 -Labels for these slots are usually silk screened on the motherboard. Slots -labeled 'A' are channel 0 in this example. Slots labeled 'B' -are channel 1. Notice that there are two csrows possible on a -physical DIMM. These csrows are allocated their csrow assignment -based on the slot into which the memory DIMM is placed. Thus, when 1 DIMM -is placed in each Channel, the csrows cross both DIMMs. +Labels for these slots are usually silk-screened on the motherboard. +Slots labeled 'A' are channel 0 in this example. Slots labeled 'B' are +channel 1. Notice that there are two csrows possible on a physical DIMM. +These csrows are allocated their csrow assignment based on the slot into +which the memory DIMM is placed. Thus, when 1 DIMM is placed in each +Channel, the csrows cross both DIMMs. Memory DIMMs come single or dual "ranked". A rank is a populated csrow. Thus, 2 single ranked DIMMs, placed in slots DIMM_A0 and DIMM_B0 above @@ -193,8 +159,8 @@ when 2 dual ranked DIMMs are similarly placed, then both csrow0 and csrow1 will be populated. The pattern repeats itself for csrow2 and csrow3. -The representation of the above is reflected in the directory tree -in EDAC's sysfs interface. Starting in directory +The representation of the above is reflected in the directory +tree in EDAC's sysfs interface. Starting in directory /sys/devices/system/edac/mc each memory controller will be represented by its own 'mcX' directory, where 'X' is the index of the MC. @@ -217,19 +183,19 @@ Under each 'mcX' directory each 'csrowX' is again represented by a |->csrow3 .... -Notice that there is no csrow1, which indicates that csrow0 is -composed of a single ranked DIMMs. This should also apply in both -Channels, in order to have dual-channel mode be operational. Since -both csrow2 and csrow3 are populated, this indicates a dual ranked -set of DIMMs for channels 0 and 1. +Notice that there is no csrow1, which indicates that csrow0 is composed +of a single ranked DIMMs. This should also apply in both Channels, in +order to have dual-channel mode be operational. Since both csrow2 and +csrow3 are populated, this indicates a dual ranked set of DIMMs for +channels 0 and 1. -Within each of the 'mcX' and 'csrowX' directories are several -EDAC control and attribute files. +Within each of the 'mcX' and 'csrowX' directories are several EDAC +control and attribute files. -============================================================================ -'mcX' DIRECTORIES +'mcX' directories +----------------- In 'mcX' directories are EDAC control and attribute files for this 'X' instance of the memory controllers. @@ -238,13 +204,14 @@ For a description of the sysfs API, please see: Documentation/ABI/testing/sysfs-devices-edac -============================================================================ -'csrowX' DIRECTORIES -When CONFIG_EDAC_LEGACY_SYSFS is enabled, the sysfs will contain the -csrowX directories. As this API doesn't work properly for Rambus, FB-DIMMs -and modern Intel Memory Controllers, this is being deprecated in favor -of dimmX directories. +'csrowX' directories +-------------------- + +When CONFIG_EDAC_LEGACY_SYSFS is enabled, sysfs will contain the csrowX +directories. As this API doesn't work properly for Rambus, FB-DIMMs and +modern Intel Memory Controllers, this is being deprecated in favor of +dimmX directories. In the 'csrowX' directories are EDAC control and attribute files for this 'X' instance of csrow: @@ -265,11 +232,11 @@ Total Correctable Errors count attribute file: 'ce_count' This attribute file displays the total count of correctable - errors that have occurred on this csrow. This - count is very important to examine. CEs provide early - indications that a DIMM is beginning to fail. This count - field should be monitored for non-zero values and report - such information to the system administrator. + errors that have occurred on this csrow. This count is very + important to examine. CEs provide early indications that a + DIMM is beginning to fail. This count field should be + monitored for non-zero values and report such information + to the system administrator. Total memory managed by this csrow attribute file: @@ -377,11 +344,13 @@ Channel 1 DIMM Label control file: motherboard specific and determination of this information must occur in userland at this time. -============================================================================ -SYSTEM LOGGING -If logging for UEs and CEs are enabled then system logs will have -error notices indicating errors that have been detected: + +SYSTEM LOGGING +-------------- + +If logging for UEs and CEs is enabled, then system logs will contain +information indicating that errors have been detected: EDAC MC0: CE page 0x283, offset 0xce0, grain 8, syndrome 0x6ec3, row 0, channel 1 "DIMM_B1": amd76x_edac @@ -404,24 +373,23 @@ The structure of the message is: and then an optional, driver-specific message that may have additional information. -Both UEs and CEs with no info will lack all but memory controller, -error type, a notice of "no info" and then an optional, -driver-specific error message. +Both UEs and CEs with no info will lack all but memory controller, error +type, a notice of "no info" and then an optional, driver-specific error +message. -============================================================================ PCI Bus Parity Detection +------------------------ - -On Header Type 00 devices the primary status is looked at -for any parity error regardless of whether Parity is enabled on the -device. (The spec indicates parity is generated in some cases). -On Header Type 01 bridges, the secondary status register is also -looked at to see if parity occurred on the bus on the other side of -the bridge. +On Header Type 00 devices, the primary status is looked at for any +parity error regardless of whether parity is enabled on the device or +not. (The spec indicates parity is generated in some cases). On Header +Type 01 bridges, the secondary status register is also looked at to see +if parity occurred on the bus on the other side of the bridge. SYSFS CONFIGURATION +------------------- Under /sys/devices/system/edac/pci are control and attribute files as follows: @@ -450,8 +418,9 @@ Parity Count: have been detected. -============================================================================ + MODULE PARAMETERS +----------------- Panic on UE control file: @@ -530,10 +499,8 @@ Panic on PCI PARITY Error: -======================================================================= - - -EDAC_DEVICE type of device +EDAC device type +---------------- In the header file, edac_core.h, there is a series of edac_device structures and APIs for the EDAC_DEVICE. @@ -573,6 +540,7 @@ The test_device_edac device adds at least one of its own custom control: The symlink points to the 'struct dev' that is registered for this edac_device. INSTANCES +--------- One or more instance directories are present. For the 'test_device_edac' case: @@ -586,6 +554,7 @@ counter in deeper subdirectories. ue_count total of UE events of subdirectories BLOCKS +------ At the lowest directory level is the 'block' directory. There can be 0, 1 or more blocks specified in each instance. @@ -623,8 +592,9 @@ unique drivers for their hardware systems. The 'test_device_edac' sample driver is located at the bluesmoke.sourceforge.net project site for EDAC. -======================================================================= + NEHALEM USAGE OF EDAC APIs +-------------------------- This chapter documents some EXPERIMENTAL mappings for EDAC API to handle Nehalem EDAC driver. They will likely be changed on future versions @@ -773,3 +743,20 @@ exports one by the driver. Since, with udimm, this is counted by software, it is possible that some errors could be lost. With rdimm's, they display the contents of the registers + +CREDITS: +======== + +Written by Doug Thompson +7 Dec 2005 +17 Jul 2007 Updated + +(c) Mauro Carvalho Chehab +05 Aug 2009 Nehalem interface + +EDAC authors/maintainers: + + Doug Thompson, Dave Jiang, Dave Peterson et al, + Mauro Carvalho Chehab + Borislav Petkov + original author: Thayne Harbaugh