diff options
26 files changed, 2343 insertions, 236 deletions
diff --git a/MAINTAINERS b/MAINTAINERS index 1bb06c5f7716..6e8897594291 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4715,6 +4715,7 @@ L: linux-edac@vger.kernel.org L: linux-mips@linux-mips.org S: Supported F: drivers/edac/octeon_edac* +F: drivers/edac/thunderx_edac* EDAC-E752X M: Mark Gross <mark.gross@intel.com> diff --git a/arch/arm/configs/multi_v7_defconfig b/arch/arm/configs/multi_v7_defconfig index a94126fb02c2..6aa7be191f1a 100644 --- a/arch/arm/configs/multi_v7_defconfig +++ b/arch/arm/configs/multi_v7_defconfig @@ -748,7 +748,6 @@ CONFIG_LEDS_TRIGGER_DEFAULT_ON=y CONFIG_LEDS_TRIGGER_TRANSIENT=y CONFIG_LEDS_TRIGGER_CAMERA=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_HIGHBANK_MC=y CONFIG_EDAC_HIGHBANK_L2=y CONFIG_RTC_CLASS=y diff --git a/arch/arm/configs/pxa_defconfig b/arch/arm/configs/pxa_defconfig index 2aac99fd1c41..1318f61589dc 100644 --- a/arch/arm/configs/pxa_defconfig +++ b/arch/arm/configs/pxa_defconfig @@ -635,8 +635,7 @@ CONFIG_LEDS_TRIGGER_GPIO=m CONFIG_LEDS_TRIGGER_DEFAULT_ON=m CONFIG_LEDS_TRIGGER_TRANSIENT=m CONFIG_LEDS_TRIGGER_CAMERA=m -CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=m +CONFIG_EDAC=m CONFIG_RTC_CLASS=y CONFIG_RTC_DEBUG=y CONFIG_RTC_DRV_DS1307=m diff --git a/arch/powerpc/configs/85xx-hw.config b/arch/powerpc/configs/85xx-hw.config index 528ff0e714e6..c03d0fb16665 100644 --- a/arch/powerpc/configs/85xx-hw.config +++ b/arch/powerpc/configs/85xx-hw.config @@ -16,9 +16,8 @@ CONFIG_DAVICOM_PHY=y CONFIG_DMADEVICES=y CONFIG_E1000E=y CONFIG_E1000=y -CONFIG_EDAC_MM_EDAC=y -CONFIG_EDAC_MPC85XX=y CONFIG_EDAC=y +CONFIG_EDAC_MPC85XX=y CONFIG_EEPROM_AT24=y CONFIG_EEPROM_LEGACY=y CONFIG_FB_FSL_DIU=y diff --git a/arch/powerpc/configs/85xx/ge_imp3a_defconfig b/arch/powerpc/configs/85xx/ge_imp3a_defconfig index c79283be5680..a917f7afb4f9 100644 --- a/arch/powerpc/configs/85xx/ge_imp3a_defconfig +++ b/arch/powerpc/configs/85xx/ge_imp3a_defconfig @@ -155,7 +155,6 @@ CONFIG_USB_OHCI_HCD_PPC_OF_BE=y CONFIG_USB_OHCI_HCD_PPC_OF_LE=y CONFIG_USB_STORAGE=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_MPC85XX=y CONFIG_RTC_CLASS=y # CONFIG_RTC_INTF_PROC is not set diff --git a/arch/powerpc/configs/85xx/xes_mpc85xx_defconfig b/arch/powerpc/configs/85xx/xes_mpc85xx_defconfig index dbd961de251e..72900b84d3e0 100644 --- a/arch/powerpc/configs/85xx/xes_mpc85xx_defconfig +++ b/arch/powerpc/configs/85xx/xes_mpc85xx_defconfig @@ -116,7 +116,6 @@ CONFIG_LEDS_TRIGGERS=y CONFIG_LEDS_TRIGGER_TIMER=y CONFIG_LEDS_TRIGGER_HEARTBEAT=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y CONFIG_RTC_DRV_CMOS=y diff --git a/arch/powerpc/configs/cell_defconfig b/arch/powerpc/configs/cell_defconfig index 2d7fcbe047ac..aa564599e368 100644 --- a/arch/powerpc/configs/cell_defconfig +++ b/arch/powerpc/configs/cell_defconfig @@ -179,7 +179,6 @@ CONFIG_INFINIBAND_MTHCA=m CONFIG_INFINIBAND_IPOIB=m CONFIG_INFINIBAND_IPOIB_DEBUG_DATA=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_CELL=y CONFIG_UIO=m CONFIG_EXT2_FS=y diff --git a/arch/powerpc/configs/pasemi_defconfig b/arch/powerpc/configs/pasemi_defconfig index 5553c5ce4274..fe43ff47bd2f 100644 --- a/arch/powerpc/configs/pasemi_defconfig +++ b/arch/powerpc/configs/pasemi_defconfig @@ -142,7 +142,6 @@ CONFIG_USB_UHCI_HCD=y CONFIG_USB_SL811_HCD=y CONFIG_USB_STORAGE=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_PASEMI=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y diff --git a/arch/powerpc/configs/ppc64_defconfig b/arch/powerpc/configs/ppc64_defconfig index 4f1288b04303..f2e03f032041 100644 --- a/arch/powerpc/configs/ppc64_defconfig +++ b/arch/powerpc/configs/ppc64_defconfig @@ -262,7 +262,6 @@ CONFIG_INFINIBAND_IPOIB_CM=y CONFIG_INFINIBAND_SRP=m CONFIG_INFINIBAND_ISER=m CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_EDAC_PASEMI=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y diff --git a/arch/powerpc/configs/ppc64e_defconfig b/arch/powerpc/configs/ppc64e_defconfig index 11a3473f9e2e..6340e6c53c54 100644 --- a/arch/powerpc/configs/ppc64e_defconfig +++ b/arch/powerpc/configs/ppc64e_defconfig @@ -173,7 +173,6 @@ CONFIG_INFINIBAND_MTHCA=m CONFIG_INFINIBAND_IPOIB=m CONFIG_INFINIBAND_ISER=m CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_DS1307=y CONFIG_FS_DAX=y diff --git a/arch/powerpc/configs/ppc6xx_defconfig b/arch/powerpc/configs/ppc6xx_defconfig index 1d2d69dd6409..18d0d60dadbf 100644 --- a/arch/powerpc/configs/ppc6xx_defconfig +++ b/arch/powerpc/configs/ppc6xx_defconfig @@ -988,8 +988,7 @@ CONFIG_LEDS_TRIGGER_BACKLIGHT=m CONFIG_LEDS_TRIGGER_DEFAULT_ON=m CONFIG_ACCESSIBILITY=y CONFIG_A11Y_BRAILLE_CONSOLE=y -CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=m +CONFIG_EDAC=m CONFIG_RTC_CLASS=y # CONFIG_RTC_HCTOSYS is not set CONFIG_RTC_DRV_DS1307=m diff --git a/arch/tile/configs/tilegx_defconfig b/arch/tile/configs/tilegx_defconfig index fd122ef45b00..0d925fa0f0c1 100644 --- a/arch/tile/configs/tilegx_defconfig +++ b/arch/tile/configs/tilegx_defconfig @@ -249,7 +249,6 @@ CONFIG_USB_EHCI_HCD=y CONFIG_USB_OHCI_HCD=y CONFIG_USB_STORAGE=y CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_TILE=y CONFIG_EXT2_FS=y diff --git a/arch/tile/configs/tilepro_defconfig b/arch/tile/configs/tilepro_defconfig index eb6a55944191..149d8e8eacb8 100644 --- a/arch/tile/configs/tilepro_defconfig +++ b/arch/tile/configs/tilepro_defconfig @@ -358,7 +358,6 @@ CONFIG_WATCHDOG_NOWAYOUT=y # CONFIG_VGA_ARB is not set # CONFIG_USB_SUPPORT is not set CONFIG_EDAC=y -CONFIG_EDAC_MM_EDAC=y CONFIG_RTC_CLASS=y CONFIG_RTC_DRV_TILE=y CONFIG_EXT2_FS=y diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c index a723ae9440ab..446c8aa09b9b 100644 --- a/arch/x86/kernel/nmi.c +++ b/arch/x86/kernel/nmi.c @@ -222,17 +222,6 @@ pci_serr_error(unsigned char reason, struct pt_regs *regs) pr_emerg("NMI: PCI system error (SERR) for reason %02x on CPU %d.\n", reason, smp_processor_id()); - /* - * On some machines, PCI SERR line is used to report memory - * errors. EDAC makes use of it. - */ -#if defined(CONFIG_EDAC) - if (edac_handler_set()) { - edac_atomic_assert_error(); - return; - } -#endif - if (panic_on_unrecovered_nmi) nmi_panic(regs, "NMI: Not continuing"); diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index 83e5f7e1a20d..a20cfcbee694 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -469,9 +469,8 @@ config ACPI_WATCHDOG config ACPI_EXTLOG tristate "Extended Error Log support" - depends on X86_MCE && X86_LOCAL_APIC + depends on X86_MCE && X86_LOCAL_APIC && EDAC select UEFI_CPER - select RAS default n help Certain usages such as Predictive Failure Analysis (PFA) require diff --git a/drivers/acpi/acpi_extlog.c b/drivers/acpi/acpi_extlog.c index a15270a806fc..502ea4dc2080 100644 --- a/drivers/acpi/acpi_extlog.c +++ b/drivers/acpi/acpi_extlog.c @@ -229,7 +229,7 @@ static int __init extlog_init(void) if (!(cap & MCG_ELOG_P) || !extlog_get_l1addr()) return -ENODEV; - if (get_edac_report_status() == EDAC_REPORTING_FORCE) { + if (edac_get_report_status() == EDAC_REPORTING_FORCE) { pr_warn("Not loading eMCA, error reporting force-enabled through EDAC.\n"); return -EPERM; } @@ -285,8 +285,8 @@ static int __init extlog_init(void) * eMCA event report method has higher priority than EDAC method, * unless EDAC event report method is mandatory. */ - old_edac_report_status = get_edac_report_status(); - set_edac_report_status(EDAC_REPORTING_DISABLED); + old_edac_report_status = edac_get_report_status(); + edac_set_report_status(EDAC_REPORTING_DISABLED); mce_register_decode_chain(&extlog_mce_dec); /* enable OS to be involved to take over management from BIOS */ ((struct extlog_l1_head *)extlog_l1_addr)->flags |= FLAG_OS_OPTIN; @@ -308,7 +308,7 @@ err: static void __exit extlog_exit(void) { - set_edac_report_status(old_edac_report_status); + edac_set_report_status(old_edac_report_status); mce_unregister_decode_chain(&extlog_mce_dec); ((struct extlog_l1_head *)extlog_l1_addr)->flags &= ~FLAG_OS_OPTIN; if (extlog_l1_addr) diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index 4773f2867234..96afb2aeed18 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -10,26 +10,16 @@ config EDAC_SUPPORT bool menuconfig EDAC - bool "EDAC (Error Detection And Correction) reporting" - depends on HAS_IOMEM && EDAC_SUPPORT + tristate "EDAC (Error Detection And Correction) reporting" + depends on HAS_IOMEM && EDAC_SUPPORT && RAS help - EDAC is designed to report errors in the core system. - These are low-level errors that are reported in the CPU or - supporting chipset or other subsystems: + EDAC is a subsystem along with hardware-specific drivers designed to + report hardware errors. These are low-level errors that are reported + in the CPU or supporting chipset or other subsystems: memory errors, cache errors, PCI errors, thermal throttling, etc.. If unsure, select 'Y'. - If this code is reporting problems on your system, please - see the EDAC project web pages for more information at: - - <http://bluesmoke.sourceforge.net/> - - and: - - <http://buttersideup.com/edacwiki> - - There is also a mailing list for the EDAC project, which can - be found via the sourceforge page. + The mailing list for the EDAC project is linux-edac@vger.kernel.org. if EDAC @@ -62,21 +52,9 @@ config EDAC_DECODE_MCE which occur really early upon boot, before the module infrastructure has been initialized. -config EDAC_MM_EDAC - tristate "Main Memory EDAC (Error Detection And Correction) reporting" - select RAS - help - Some systems are able to detect and correct errors in main - memory. EDAC can report statistics on memory error - detection and correction (EDAC - or commonly referred to ECC - errors). EDAC will also try to decode where these errors - occurred so that a particular failing memory module can be - replaced. If unsure, select 'Y'. - config EDAC_GHES bool "Output ACPI APEI/GHES BIOS detected errors via EDAC" - depends on ACPI_APEI_GHES && (EDAC_MM_EDAC=y) - default y + depends on ACPI_APEI_GHES && (EDAC=y) help Not all machines support hardware-driven error report. Some of those provide a BIOS-driven error report mechanism via ACPI, using the @@ -98,7 +76,7 @@ config EDAC_GHES config EDAC_AMD64 tristate "AMD64 (Opteron, Athlon64)" - depends on EDAC_MM_EDAC && AMD_NB && EDAC_DECODE_MCE + depends on AMD_NB && EDAC_DECODE_MCE help Support for error detection and correction of DRAM ECC errors on the AMD64 families (>= K8) of memory controllers. @@ -124,28 +102,28 @@ config EDAC_AMD64_ERROR_INJECTION config EDAC_AMD76X tristate "AMD 76x (760, 762, 768)" - depends on EDAC_MM_EDAC && PCI && X86_32 + depends on PCI && X86_32 help Support for error detection and correction on the AMD 76x series of chipsets used with the Athlon processor. config EDAC_E7XXX tristate "Intel e7xxx (e7205, e7500, e7501, e7505)" - depends on EDAC_MM_EDAC && PCI && X86_32 + depends on PCI && X86_32 help Support for error detection and correction on the Intel E7205, E7500, E7501 and E7505 server chipsets. config EDAC_E752X tristate "Intel e752x (e7520, e7525, e7320) and 3100" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction on the Intel E7520, E7525, E7320 server chipsets. config EDAC_I82443BXGX tristate "Intel 82443BX/GX (440BX/GX)" - depends on EDAC_MM_EDAC && PCI && X86_32 + depends on PCI && X86_32 depends on BROKEN help Support for error detection and correction on the Intel @@ -153,56 +131,56 @@ config EDAC_I82443BXGX config EDAC_I82875P tristate "Intel 82875p (D82875P, E7210)" - depends on EDAC_MM_EDAC && PCI && X86_32 + depends on PCI && X86_32 help Support for error detection and correction on the Intel DP82785P and E7210 server chipsets. config EDAC_I82975X tristate "Intel 82975x (D82975x)" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction on the Intel DP82975x server chipsets. config EDAC_I3000 tristate "Intel 3000/3010" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction on the Intel 3000 and 3010 server chipsets. config EDAC_I3200 tristate "Intel 3200" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction on the Intel 3200 and 3210 server chipsets. config EDAC_IE31200 tristate "Intel e312xx" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction on the Intel E3-1200 based DRAM controllers. config EDAC_X38 tristate "Intel X38" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction on the Intel X38 server chipsets. config EDAC_I5400 tristate "Intel 5400 (Seaburg) chipsets" - depends on EDAC_MM_EDAC && PCI && X86 + depends on PCI && X86 help Support for error detection and correction the Intel i5400 MCH chipset (Seaburg). config EDAC_I7CORE tristate "Intel i7 Core (Nehalem) processors" - depends on EDAC_MM_EDAC && PCI && X86 && X86_MCE_INTEL + depends on PCI && X86 && X86_MCE_INTEL help Support for error detection and correction the Intel i7 Core (Nehalem) Integrated Memory Controller that exists on @@ -211,58 +189,56 @@ config EDAC_I7CORE config EDAC_I82860 tristate "Intel 82860" - depends on EDAC_MM_EDAC && PCI && X86_32 + depends on PCI && X86_32 help Support for error detection and correction on the Intel 82860 chipset. config EDAC_R82600 tristate "Radisys 82600 embedded chipset" - depends on EDAC_MM_EDAC && PCI && X86_32 + depends on PCI && X86_32 help Support for error detection and correction on the Radisys 82600 embedded chipset. config EDAC_I5000 tristate "Intel Greencreek/Blackford chipset" - depends on EDAC_MM_EDAC && X86 && PCI + depends on X86 && PCI help Support for error detection and correction the Intel Greekcreek/Blackford chipsets. config EDAC_I5100 tristate "Intel San Clemente MCH" - depends on EDAC_MM_EDAC && X86 && PCI + depends on X86 && PCI help Support for error detection and correction the Intel San Clemente MCH. config EDAC_I7300 tristate "Intel Clarksboro MCH" - depends on EDAC_MM_EDAC && X86 && PCI + depends on X86 && PCI help Support for error detection and correction the Intel Clarksboro MCH (Intel 7300 chipset). config EDAC_SBRIDGE tristate "Intel Sandy-Bridge/Ivy-Bridge/Haswell Integrated MC" - depends on EDAC_MM_EDAC && PCI && X86_64 && X86_MCE_INTEL - depends on PCI_MMCONFIG + depends on PCI && X86_64 && X86_MCE_INTEL && PCI_MMCONFIG help Support for error detection and correction the Intel Sandy Bridge, Ivy Bridge and Haswell Integrated Memory Controllers. config EDAC_SKX tristate "Intel Skylake server Integrated MC" - depends on EDAC_MM_EDAC && PCI && X86_64 && X86_MCE_INTEL - depends on PCI_MMCONFIG + depends on PCI && X86_64 && X86_MCE_INTEL && PCI_MMCONFIG help Support for error detection and correction the Intel Skylake server Integrated Memory Controllers. config EDAC_PND2 tristate "Intel Pondicherry2" - depends on EDAC_MM_EDAC && PCI && X86_64 && X86_MCE_INTEL + depends on PCI && X86_64 && X86_MCE_INTEL help Support for error detection and correction on the Intel Pondicherry2 Integrated Memory Controller. This SoC IP is @@ -271,36 +247,35 @@ config EDAC_PND2 config EDAC_MPC85XX tristate "Freescale MPC83xx / MPC85xx" - depends on EDAC_MM_EDAC && FSL_SOC + depends on FSL_SOC help Support for error detection and correction on the Freescale MPC8349, MPC8560, MPC8540, MPC8548, T4240 config EDAC_LAYERSCAPE tristate "Freescale Layerscape DDR" - depends on EDAC_MM_EDAC && ARCH_LAYERSCAPE + depends on ARCH_LAYERSCAPE help Support for error detection and correction on Freescale memory controllers on Layerscape SoCs. config EDAC_MV64X60 tristate "Marvell MV64x60" - depends on EDAC_MM_EDAC && MV64X60 + depends on MV64X60 help Support for error detection and correction on the Marvell MV64360 and MV64460 chipsets. config EDAC_PASEMI tristate "PA Semi PWRficient" - depends on EDAC_MM_EDAC && PCI - depends on PPC_PASEMI + depends on PPC_PASEMI && PCI help Support for error detection and correction on PA Semi PWRficient. config EDAC_CELL tristate "Cell Broadband Engine memory controller" - depends on EDAC_MM_EDAC && PPC_CELL_COMMON + depends on PPC_CELL_COMMON help Support for error detection and correction on the Cell Broadband Engine internal memory controller @@ -308,7 +283,7 @@ config EDAC_CELL config EDAC_PPC4XX tristate "PPC4xx IBM DDR2 Memory Controller" - depends on EDAC_MM_EDAC && 4xx + depends on 4xx help This enables support for EDAC on the ECC memory used with the IBM DDR2 memory controller found in various @@ -317,7 +292,7 @@ config EDAC_PPC4XX config EDAC_AMD8131 tristate "AMD8131 HyperTransport PCI-X Tunnel" - depends on EDAC_MM_EDAC && PCI && PPC_MAPLE + depends on PCI && PPC_MAPLE help Support for error detection and correction on the AMD8131 HyperTransport PCI-X Tunnel chip. @@ -326,7 +301,7 @@ config EDAC_AMD8131 config EDAC_AMD8111 tristate "AMD8111 HyperTransport I/O Hub" - depends on EDAC_MM_EDAC && PCI && PPC_MAPLE + depends on PCI && PPC_MAPLE help Support for error detection and correction on the AMD8111 HyperTransport I/O Hub chip. @@ -335,7 +310,7 @@ config EDAC_AMD8111 config EDAC_CPC925 tristate "IBM CPC925 Memory Controller (PPC970FX)" - depends on EDAC_MM_EDAC && PPC64 + depends on PPC64 help Support for error detection and correction on the IBM CPC925 Bridge and Memory Controller, which is @@ -344,7 +319,7 @@ config EDAC_CPC925 config EDAC_TILE tristate "Tilera Memory Controller" - depends on EDAC_MM_EDAC && TILE + depends on TILE default y help Support for error detection and correction on the @@ -352,49 +327,59 @@ config EDAC_TILE config EDAC_HIGHBANK_MC tristate "Highbank Memory Controller" - depends on EDAC_MM_EDAC && ARCH_HIGHBANK + depends on ARCH_HIGHBANK help Support for error detection and correction on the Calxeda Highbank memory controller. config EDAC_HIGHBANK_L2 tristate "Highbank L2 Cache" - depends on EDAC_MM_EDAC && ARCH_HIGHBANK + depends on ARCH_HIGHBANK help Support for error detection and correction on the Calxeda Highbank memory controller. config EDAC_OCTEON_PC tristate "Cavium Octeon Primary Caches" - depends on EDAC_MM_EDAC && CPU_CAVIUM_OCTEON + depends on CPU_CAVIUM_OCTEON help Support for error detection and correction on the primary caches of the cnMIPS cores of Cavium Octeon family SOCs. config EDAC_OCTEON_L2C tristate "Cavium Octeon Secondary Caches (L2C)" - depends on EDAC_MM_EDAC && CAVIUM_OCTEON_SOC + depends on CAVIUM_OCTEON_SOC help Support for error detection and correction on the Cavium Octeon family of SOCs. config EDAC_OCTEON_LMC tristate "Cavium Octeon DRAM Memory Controller (LMC)" - depends on EDAC_MM_EDAC && CAVIUM_OCTEON_SOC + depends on CAVIUM_OCTEON_SOC help Support for error detection and correction on the Cavium Octeon family of SOCs. config EDAC_OCTEON_PCI tristate "Cavium Octeon PCI Controller" - depends on EDAC_MM_EDAC && PCI && CAVIUM_OCTEON_SOC + depends on PCI && CAVIUM_OCTEON_SOC help Support for error detection and correction on the Cavium Octeon family of SOCs. +config EDAC_THUNDERX + tristate "Cavium ThunderX EDAC" + depends on ARM64 + depends on PCI + help + Support for error detection and correction on the + Cavium ThunderX memory controllers (LMC), Cache + Coherent Processor Interconnect (CCPI) and L2 cache + blocks (TAD, CBC, MCI). + config EDAC_ALTERA bool "Altera SOCFPGA ECC" - depends on EDAC_MM_EDAC=y && ARCH_SOCFPGA + depends on EDAC=y && ARCH_SOCFPGA help Support for error detection and correction on the Altera SOCs. This must be selected for SDRAM ECC. @@ -460,14 +445,14 @@ config EDAC_ALTERA_SDMMC config EDAC_SYNOPSYS tristate "Synopsys DDR Memory Controller" - depends on EDAC_MM_EDAC && ARCH_ZYNQ + depends on ARCH_ZYNQ help 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) + depends on (ARM64 || COMPILE_TEST) help Support for error detection and correction on the APM X-Gene family of SOCs. diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile index 587107e90996..0fd9ffa63299 100644 --- a/drivers/edac/Makefile +++ b/drivers/edac/Makefile @@ -6,8 +6,7 @@ # GNU General Public License. # -obj-$(CONFIG_EDAC) := edac_stub.o -obj-$(CONFIG_EDAC_MM_EDAC) += edac_core.o +obj-$(CONFIG_EDAC) := edac_core.o edac_core-y := edac_mc.o edac_device.o edac_mc_sysfs.o edac_core-y += edac_module.o edac_device_sysfs.o wq.o @@ -67,13 +66,14 @@ obj-$(CONFIG_EDAC_AMD8131) += amd8131_edac.o obj-$(CONFIG_EDAC_TILE) += tile_edac.o -obj-$(CONFIG_EDAC_HIGHBANK_MC) += highbank_mc_edac.o -obj-$(CONFIG_EDAC_HIGHBANK_L2) += highbank_l2_edac.o +obj-$(CONFIG_EDAC_HIGHBANK_MC) += highbank_mc_edac.o +obj-$(CONFIG_EDAC_HIGHBANK_L2) += highbank_l2_edac.o obj-$(CONFIG_EDAC_OCTEON_PC) += octeon_edac-pc.o obj-$(CONFIG_EDAC_OCTEON_L2C) += octeon_edac-l2c.o obj-$(CONFIG_EDAC_OCTEON_LMC) += octeon_edac-lmc.o obj-$(CONFIG_EDAC_OCTEON_PCI) += octeon_edac-pci.o +obj-$(CONFIG_EDAC_THUNDERX) += thunderx_edac.o obj-$(CONFIG_EDAC_ALTERA) += altera_edac.o obj-$(CONFIG_EDAC_SYNOPSYS) += synopsys_edac.o diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c index c5a5b91f37f0..7717b094fabb 100644 --- a/drivers/edac/altera_edac.c +++ b/drivers/edac/altera_edac.c @@ -1023,13 +1023,23 @@ out: return ret; } +static int socfpga_is_a10(void) +{ + return of_machine_is_compatible("altr,socfpga-arria10"); +} + static int validate_parent_available(struct device_node *np); static const struct of_device_id altr_edac_a10_device_of_match[]; static int __init __maybe_unused altr_init_a10_ecc_device_type(char *compat) { int irq; - struct device_node *child, *np = of_find_compatible_node(NULL, NULL, - "altr,socfpga-a10-ecc-manager"); + struct device_node *child, *np; + + if (!socfpga_is_a10()) + return -ENODEV; + + np = of_find_compatible_node(NULL, NULL, + "altr,socfpga-a10-ecc-manager"); if (!np) { edac_printk(KERN_ERR, EDAC_DEVICE, "ECC Manager not found\n"); return -ENODEV; @@ -1545,8 +1555,12 @@ static const struct edac_device_prv_data a10_sdmmceccb_data = { static int __init socfpga_init_sdmmc_ecc(void) { int rc = -ENODEV; - struct device_node *child = of_find_compatible_node(NULL, NULL, - "altr,socfpga-sdmmc-ecc"); + struct device_node *child; + + if (!socfpga_is_a10()) + return -ENODEV; + + child = of_find_compatible_node(NULL, NULL, "altr,socfpga-sdmmc-ecc"); if (!child) { edac_printk(KERN_WARNING, EDAC_DEVICE, "SDMMC node not found\n"); return -ENODEV; diff --git a/drivers/edac/edac_mc.c b/drivers/edac/edac_mc.c index e5573c56b15e..480072139b7a 100644 --- a/drivers/edac/edac_mc.c +++ b/drivers/edac/edac_mc.c @@ -40,6 +40,11 @@ #define edac_atomic_scrub(va, size) do { } while (0) #endif +int edac_op_state = EDAC_OPSTATE_INVAL; +EXPORT_SYMBOL_GPL(edac_op_state); + +static int edac_report = EDAC_REPORTING_ENABLED; + /* lock to memory controller's control array */ static DEFINE_MUTEX(mem_ctls_mutex); static LIST_HEAD(mc_devices); @@ -52,6 +57,65 @@ static void const *edac_mc_owner; static struct bus_type mc_bus[EDAC_MAX_MCS]; +int edac_get_report_status(void) +{ + return edac_report; +} +EXPORT_SYMBOL_GPL(edac_get_report_status); + +void edac_set_report_status(int new) +{ + if (new == EDAC_REPORTING_ENABLED || + new == EDAC_REPORTING_DISABLED || + new == EDAC_REPORTING_FORCE) + edac_report = new; +} +EXPORT_SYMBOL_GPL(edac_set_report_status); + +static int edac_report_set(const char *str, const struct kernel_param *kp) +{ + if (!str) + return -EINVAL; + + if (!strncmp(str, "on", 2)) + edac_report = EDAC_REPORTING_ENABLED; + else if (!strncmp(str, "off", 3)) + edac_report = EDAC_REPORTING_DISABLED; + else if (!strncmp(str, "force", 5)) + edac_report = EDAC_REPORTING_FORCE; + + return 0; +} + +static int edac_report_get(char *buffer, const struct kernel_param *kp) +{ + int ret = 0; + + switch (edac_report) { + case EDAC_REPORTING_ENABLED: + ret = sprintf(buffer, "on"); + break; + case EDAC_REPORTING_DISABLED: + ret = sprintf(buffer, "off"); + break; + case EDAC_REPORTING_FORCE: + ret = sprintf(buffer, "force"); + break; + default: + ret = -EINVAL; + break; + } + + return ret; +} + +static const struct kernel_param_ops edac_report_ops = { + .set = edac_report_set, + .get = edac_report_get, +}; + +module_param_cb(edac_report, &edac_report_ops, &edac_report, 0644); + unsigned edac_dimm_info_location(struct dimm_info *dimm, char *buf, unsigned len) { @@ -505,22 +569,6 @@ struct mem_ctl_info *find_mci_by_dev(struct device *dev) EXPORT_SYMBOL_GPL(find_mci_by_dev); /* - * handler for EDAC to check if NMI type handler has asserted interrupt - */ -static int edac_mc_assert_error_check_and_clear(void) -{ - int old_state; - - if (edac_op_state == EDAC_OPSTATE_POLL) - return 1; - - old_state = edac_err_assert; - edac_err_assert = 0; - - return old_state; -} - -/* * edac_mc_workq_function * performs the operation scheduled by a workq request */ @@ -536,7 +584,7 @@ static void edac_mc_workq_function(struct work_struct *work_req) return; } - if (edac_mc_assert_error_check_and_clear()) + if (edac_op_state == EDAC_OPSTATE_POLL) mci->edac_check(mci); mutex_unlock(&mem_ctls_mutex); @@ -601,7 +649,6 @@ static int add_mc_to_global_list(struct mem_ctl_info *mci) } list_add_tail_rcu(&mci->link, insert_before); - atomic_inc(&edac_handlers); return 0; fail0: @@ -619,7 +666,6 @@ fail1: static int del_mc_from_global_list(struct mem_ctl_info *mci) { - int handlers = atomic_dec_return(&edac_handlers); list_del_rcu(&mci->link); /* these are for safe removal of devices from global list while @@ -628,7 +674,7 @@ static int del_mc_from_global_list(struct mem_ctl_info *mci) synchronize_rcu(); INIT_LIST_HEAD(&mci->link); - return handlers; + return list_empty(&mc_devices); } struct mem_ctl_info *edac_mc_find(int idx) @@ -763,7 +809,7 @@ struct mem_ctl_info *edac_mc_del_mc(struct device *dev) /* mark MCI offline: */ mci->op_state = OP_OFFLINE; - if (!del_mc_from_global_list(mci)) + if (del_mc_from_global_list(mci)) edac_mc_owner = NULL; mutex_unlock(&mem_ctls_mutex); @@ -1195,10 +1241,13 @@ void edac_mc_handle_error(const enum hw_event_mc_err_type type, /* Report the error via the trace interface */ grain_bits = fls_long(e->grain) + 1; - trace_mc_event(type, e->msg, e->label, e->error_count, - mci->mc_idx, e->top_layer, e->mid_layer, e->low_layer, - (e->page_frame_number << PAGE_SHIFT) | e->offset_in_page, - grain_bits, e->syndrome, e->other_detail); + + if (IS_ENABLED(CONFIG_RAS)) + trace_mc_event(type, e->msg, e->label, e->error_count, + mci->mc_idx, e->top_layer, e->mid_layer, + e->low_layer, + (e->page_frame_number << PAGE_SHIFT) | e->offset_in_page, + grain_bits, e->syndrome, e->other_detail); edac_raw_mc_handle_error(type, mci, e); } diff --git a/drivers/edac/edac_stub.c b/drivers/edac/edac_stub.c deleted file mode 100644 index 952e411f01f2..000000000000 --- a/drivers/edac/edac_stub.c +++ /dev/null @@ -1,68 +0,0 @@ -/* - * common EDAC components that must be in kernel - * - * Author: Dave Jiang <djiang@mvista.com> - * - * 2007 (c) MontaVista Software, Inc. - * 2010 (c) Advanced Micro Devices Inc. - * Borislav Petkov <bp@alien8.de> - * - * This file is licensed under the terms of the GNU General Public - * License version 2. This program is licensed "as is" without any - * warranty of any kind, whether express or implied. - * - */ -#include <linux/module.h> -#include <linux/edac.h> -#include <linux/atomic.h> -#include <linux/device.h> - -int edac_op_state = EDAC_OPSTATE_INVAL; -EXPORT_SYMBOL_GPL(edac_op_state); - -atomic_t edac_handlers = ATOMIC_INIT(0); -EXPORT_SYMBOL_GPL(edac_handlers); - -int edac_err_assert = 0; -EXPORT_SYMBOL_GPL(edac_err_assert); - -int edac_report_status = EDAC_REPORTING_ENABLED; -EXPORT_SYMBOL_GPL(edac_report_status); - -static int __init edac_report_setup(char *str) -{ - if (!str) - return -EINVAL; - - if (!strncmp(str, "on", 2)) - set_edac_report_status(EDAC_REPORTING_ENABLED); - else if (!strncmp(str, "off", 3)) - set_edac_report_status(EDAC_REPORTING_DISABLED); - else if (!strncmp(str, "force", 5)) - set_edac_report_status(EDAC_REPORTING_FORCE); - - return 0; -} -__setup("edac_report=", edac_report_setup); - -/* - * called to determine if there is an EDAC driver interested in - * knowing an event (such as NMI) occurred - */ -int edac_handler_set(void) -{ - if (edac_op_state == EDAC_OPSTATE_POLL) - return 0; - - return atomic_read(&edac_handlers); -} -EXPORT_SYMBOL_GPL(edac_handler_set); - -/* - * handler for NMI type of interrupts to assert error - */ -void edac_atomic_assert_error(void) -{ - edac_err_assert++; -} -EXPORT_SYMBOL_GPL(edac_atomic_assert_error); diff --git a/drivers/edac/pnd2_edac.c b/drivers/edac/pnd2_edac.c index 928e0dba41fc..1cad5a9af8d0 100644 --- a/drivers/edac/pnd2_edac.c +++ b/drivers/edac/pnd2_edac.c @@ -1349,7 +1349,7 @@ static int pnd2_mce_check_error(struct notifier_block *nb, unsigned long val, vo struct dram_addr daddr; char *type; - if (get_edac_report_status() == EDAC_REPORTING_DISABLED) + if (edac_get_report_status() == EDAC_REPORTING_DISABLED) return NOTIFY_DONE; mci = pnd2_mci; diff --git a/drivers/edac/sb_edac.c b/drivers/edac/sb_edac.c index a65ea44e3b0b..ea21cb651b3c 100644 --- a/drivers/edac/sb_edac.c +++ b/drivers/edac/sb_edac.c @@ -3075,7 +3075,7 @@ static int sbridge_mce_check_error(struct notifier_block *nb, unsigned long val, struct sbridge_pvt *pvt; char *type; - if (get_edac_report_status() == EDAC_REPORTING_DISABLED) + if (edac_get_report_status() == EDAC_REPORTING_DISABLED) return NOTIFY_DONE; mci = get_mci_for_node_id(mce->socketid); @@ -3441,7 +3441,7 @@ static int __init sbridge_init(void) if (rc >= 0) { mce_register_decode_chain(&sbridge_mce_dec); - if (get_edac_report_status() == EDAC_REPORTING_DISABLED) + if (edac_get_report_status() == EDAC_REPORTING_DISABLED) sbridge_printk(KERN_WARNING, "Loading driver, error reporting disabled.\n"); return 0; } diff --git a/drivers/edac/skx_edac.c b/drivers/edac/skx_edac.c index 1159dba4671f..64bef6c9cfb4 100644 --- a/drivers/edac/skx_edac.c +++ b/drivers/edac/skx_edac.c @@ -971,7 +971,7 @@ static int skx_mce_check_error(struct notifier_block *nb, unsigned long val, struct mem_ctl_info *mci; char *type; - if (get_edac_report_status() == EDAC_REPORTING_DISABLED) + if (edac_get_report_status() == EDAC_REPORTING_DISABLED) return NOTIFY_DONE; /* ignore unless this is memory related with an address */ diff --git a/drivers/edac/thunderx_edac.c b/drivers/edac/thunderx_edac.c new file mode 100644 index 000000000000..86d585cb6d32 --- /dev/null +++ b/drivers/edac/thunderx_edac.c @@ -0,0 +1,2174 @@ +/* + * Cavium ThunderX memory controller kernel module + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + * + * Copyright Cavium, Inc. (C) 2015-2017. All rights reserved. + * + */ + +#include <linux/module.h> +#include <linux/pci.h> +#include <linux/edac.h> +#include <linux/interrupt.h> +#include <linux/string.h> +#include <linux/stop_machine.h> +#include <linux/delay.h> +#include <linux/sizes.h> +#include <linux/atomic.h> +#include <linux/bitfield.h> +#include <linux/circ_buf.h> + +#include <asm/page.h> + +#include "edac_module.h" + +#define phys_to_pfn(phys) (PFN_DOWN(phys)) + +#define THUNDERX_NODE GENMASK(45, 44) + +enum { + ERR_CORRECTED = 1, + ERR_UNCORRECTED = 2, + ERR_UNKNOWN = 3, +}; + +#define MAX_SYNDROME_REGS 4 + +struct error_syndrome { + u64 reg[MAX_SYNDROME_REGS]; +}; + +struct error_descr { + int type; + u64 mask; + char *descr; +}; + +static void decode_register(char *str, size_t size, + const struct error_descr *descr, + const uint64_t reg) +{ + int ret = 0; + + while (descr->type && descr->mask && descr->descr) { + if (reg & descr->mask) { + ret = snprintf(str, size, "\n\t%s, %s", + descr->type == ERR_CORRECTED ? + "Corrected" : "Uncorrected", + descr->descr); + str += ret; + size -= ret; + } + descr++; + } +} + +static unsigned long get_bits(unsigned long data, int pos, int width) +{ + return (data >> pos) & ((1 << width) - 1); +} + +#define L2C_CTL 0x87E080800000 +#define L2C_CTL_DISIDXALIAS BIT(0) + +#define PCI_DEVICE_ID_THUNDER_LMC 0xa022 + +#define LMC_FADR 0x20 +#define LMC_FADR_FDIMM(x) ((x >> 37) & 0x1) +#define LMC_FADR_FBUNK(x) ((x >> 36) & 0x1) +#define LMC_FADR_FBANK(x) ((x >> 32) & 0xf) +#define LMC_FADR_FROW(x) ((x >> 14) & 0xffff) +#define LMC_FADR_FCOL(x) ((x >> 0) & 0x1fff) + +#define LMC_NXM_FADR 0x28 +#define LMC_ECC_SYND 0x38 + +#define LMC_ECC_PARITY_TEST 0x108 + +#define LMC_INT_W1S 0x150 + +#define LMC_INT_ENA_W1C 0x158 +#define LMC_INT_ENA_W1S 0x160 + +#define LMC_CONFIG 0x188 + +#define LMC_CONFIG_BG2 BIT(62) +#define LMC_CONFIG_RANK_ENA BIT(42) +#define LMC_CONFIG_PBANK_LSB(x) (((x) >> 5) & 0xF) +#define LMC_CONFIG_ROW_LSB(x) (((x) >> 2) & 0x7) + +#define LMC_CONTROL 0x190 +#define LMC_CONTROL_XOR_BANK BIT(16) + +#define LMC_INT 0x1F0 + +#define LMC_INT_DDR_ERR BIT(11) +#define LMC_INT_DED_ERR (0xFUL << 5) +#define LMC_INT_SEC_ERR (0xFUL << 1) +#define LMC_INT_NXM_WR_MASK BIT(0) + +#define LMC_DDR_PLL_CTL 0x258 +#define LMC_DDR_PLL_CTL_DDR4 BIT(29) + +#define LMC_FADR_SCRAMBLED 0x330 + +#define LMC_INT_UE (LMC_INT_DDR_ERR | LMC_INT_DED_ERR | \ + LMC_INT_NXM_WR_MASK) + +#define LMC_INT_CE (LMC_INT_SEC_ERR) + +static const struct error_descr lmc_errors[] = { + { + .type = ERR_CORRECTED, + .mask = LMC_INT_SEC_ERR, + .descr = "Single-bit ECC error", + }, + { + .type = ERR_UNCORRECTED, + .mask = LMC_INT_DDR_ERR, + .descr = "DDR chip error", + }, + { + .type = ERR_UNCORRECTED, + .mask = LMC_INT_DED_ERR, + .descr = "Double-bit ECC error", + }, + { + .type = ERR_UNCORRECTED, + .mask = LMC_INT_NXM_WR_MASK, + .descr = "Non-existent memory write", + }, + {0, 0, NULL}, +}; + +#define LMC_INT_EN_DDR_ERROR_ALERT_ENA BIT(5) +#define LMC_INT_EN_DLCRAM_DED_ERR BIT(4) +#define LMC_INT_EN_DLCRAM_SEC_ERR BIT(3) +#define LMC_INT_INTR_DED_ENA BIT(2) +#define LMC_INT_INTR_SEC_ENA BIT(1) +#define LMC_INT_INTR_NXM_WR_ENA BIT(0) + +#define LMC_INT_ENA_ALL GENMASK(5, 0) + +#define LMC_DDR_PLL_CTL 0x258 +#define LMC_DDR_PLL_CTL_DDR4 BIT(29) + +#define LMC_CONTROL 0x190 +#define LMC_CONTROL_RDIMM BIT(0) + +#define LMC_SCRAM_FADR 0x330 + +#define LMC_CHAR_MASK0 0x228 +#define LMC_CHAR_MASK2 0x238 + +#define RING_ENTRIES 8 + +struct debugfs_entry { + const char *name; + umode_t mode; + const struct file_operations fops; +}; + +struct lmc_err_ctx { + u64 reg_int; + u64 reg_fadr; + u64 reg_nxm_fadr; + u64 reg_scram_fadr; + u64 reg_ecc_synd; +}; + +struct thunderx_lmc { + void __iomem *regs; + struct pci_dev *pdev; + struct msix_entry msix_ent; + + atomic_t ecc_int; + + u64 mask0; + u64 mask2; + u64 parity_test; + u64 node; + + int xbits; + int bank_width; + int pbank_lsb; + int dimm_lsb; + int rank_lsb; + int bank_lsb; + int row_lsb; + int col_hi_lsb; + + int xor_bank; + int l2c_alias; + + struct page *mem; + + struct lmc_err_ctx err_ctx[RING_ENTRIES]; + unsigned long ring_head; + unsigned long ring_tail; +}; + +#define ring_pos(pos, size) ((pos) & (size - 1)) + +#define DEBUGFS_STRUCT(_name, _mode, _write, _read) \ +static struct debugfs_entry debugfs_##_name = { \ + .name = __stringify(_name), \ + .mode = VERIFY_OCTAL_PERMISSIONS(_mode), \ + .fops = { \ + .open = simple_open, \ + .write = _write, \ + .read = _read, \ + .llseek = generic_file_llseek, \ + }, \ +} + +#define DEBUGFS_FIELD_ATTR(_type, _field) \ +static ssize_t thunderx_##_type##_##_field##_read(struct file *file, \ + char __user *data, \ + size_t count, loff_t *ppos) \ +{ \ + struct thunderx_##_type *pdata = file->private_data; \ + char buf[20]; \ + \ + snprintf(buf, count, "0x%016llx", pdata->_field); \ + return simple_read_from_buffer(data, count, ppos, \ + buf, sizeof(buf)); \ +} \ + \ +static ssize_t thunderx_##_type##_##_field##_write(struct file *file, \ + const char __user *data, \ + size_t count, loff_t *ppos) \ +{ \ + struct thunderx_##_type *pdata = file->private_data; \ + int res; \ + \ + res = kstrtoull_from_user(data, count, 0, &pdata->_field); \ + \ + return res ? res : count; \ +} \ + \ +DEBUGFS_STRUCT(_field, 0600, \ + thunderx_##_type##_##_field##_write, \ + thunderx_##_type##_##_field##_read) \ + +#define DEBUGFS_REG_ATTR(_type, _name, _reg) \ +static ssize_t thunderx_##_type##_##_name##_read(struct file *file, \ + char __user *data, \ + size_t count, loff_t *ppos) \ +{ \ + struct thunderx_##_type *pdata = file->private_data; \ + char buf[20]; \ + \ + sprintf(buf, "0x%016llx", readq(pdata->regs + _reg)); \ + return simple_read_from_buffer(data, count, ppos, \ + buf, sizeof(buf)); \ +} \ + \ +static ssize_t thunderx_##_type##_##_name##_write(struct file *file, \ + const char __user *data, \ + size_t count, loff_t *ppos) \ +{ \ + struct thunderx_##_type *pdata = file->private_data; \ + u64 val; \ + int res; \ + \ + res = kstrtoull_from_user(data, count, 0, &val); \ + \ + if (!res) { \ + writeq(val, pdata->regs + _reg); \ + res = count; \ + } \ + \ + return res; \ +} \ + \ +DEBUGFS_STRUCT(_name, 0600, \ + thunderx_##_type##_##_name##_write, \ + thunderx_##_type##_##_name##_read) + +#define LMC_DEBUGFS_ENT(_field) DEBUGFS_FIELD_ATTR(lmc, _field) + +/* + * To get an ECC error injected, the following steps are needed: + * - Setup the ECC injection by writing the appropriate parameters: + * echo <bit mask value> > /sys/kernel/debug/<device number>/ecc_mask0 + * echo <bit mask value> > /sys/kernel/debug/<device number>/ecc_mask2 + * echo 0x802 > /sys/kernel/debug/<device number>/ecc_parity_test + * - Do the actual injection: + * echo 1 > /sys/kernel/debug/<device number>/inject_ecc + */ +static ssize_t thunderx_lmc_inject_int_write(struct file *file, + const char __user *data, + size_t count, loff_t *ppos) +{ + struct thunderx_lmc *lmc = file->private_data; + u64 val; + int res; + + res = kstrtoull_from_user(data, count, 0, &val); + + if (!res) { + /* Trigger the interrupt */ + writeq(val, lmc->regs + LMC_INT_W1S); + res = count; + } + + return res; +} + +static ssize_t thunderx_lmc_int_read(struct file *file, + char __user *data, + size_t count, loff_t *ppos) +{ + struct thunderx_lmc *lmc = file->private_data; + char buf[20]; + u64 lmc_int = readq(lmc->regs + LMC_INT); + + snprintf(buf, sizeof(buf), "0x%016llx", lmc_int); + return simple_read_from_buffer(data, count, ppos, buf, sizeof(buf)); +} + +#define TEST_PATTERN 0xa5 + +static int inject_ecc_fn(void *arg) +{ + struct thunderx_lmc *lmc = arg; + uintptr_t addr, phys; + unsigned int cline_size = cache_line_size(); + const unsigned int lines = PAGE_SIZE / cline_size; + unsigned int i, cl_idx; + + addr = (uintptr_t)page_address(lmc->mem); + phys = (uintptr_t)page_to_phys(lmc->mem); + + cl_idx = (phys & 0x7f) >> 4; + lmc->parity_test &= ~(7ULL << 8); + lmc->parity_test |= (cl_idx << 8); + + writeq(lmc->mask0, lmc->regs + LMC_CHAR_MASK0); + writeq(lmc->mask2, lmc->regs + LMC_CHAR_MASK2); + writeq(lmc->parity_test, lmc->regs + LMC_ECC_PARITY_TEST); + + readq(lmc->regs + LMC_CHAR_MASK0); + readq(lmc->regs + LMC_CHAR_MASK2); + readq(lmc->regs + LMC_ECC_PARITY_TEST); + + for (i = 0; i < lines; i++) { + memset((void *)addr, TEST_PATTERN, cline_size); + barrier(); + + /* + * Flush L1 cachelines to the PoC (L2). + * This will cause cacheline eviction to the L2. + */ + asm volatile("dc civac, %0\n" + "dsb sy\n" + : : "r"(addr + i * cline_size)); + } + + for (i = 0; i < lines; i++) { + /* + * Flush L2 cachelines to the DRAM. + * This will cause cacheline eviction to the DRAM + * and ECC corruption according to the masks set. + */ + __asm__ volatile("sys #0,c11,C1,#2, %0\n" + : : "r"(phys + i * cline_size)); + } + + for (i = 0; i < lines; i++) { + /* + * Invalidate L2 cachelines. + * The subsequent load will cause cacheline fetch + * from the DRAM and an error interrupt + */ + __asm__ volatile("sys #0,c11,C1,#1, %0" + : : "r"(phys + i * cline_size)); + } + + for (i = 0; i < lines; i++) { + /* + * Invalidate L1 cachelines. + * The subsequent load will cause cacheline fetch + * from the L2 and/or DRAM + */ + asm volatile("dc ivac, %0\n" + "dsb sy\n" + : : "r"(addr + i * cline_size)); + } + + return 0; +} + +static ssize_t thunderx_lmc_inject_ecc_write(struct file *file, + const char __user *data, + size_t count, loff_t *ppos) +{ + struct thunderx_lmc *lmc = file->private_data; + + unsigned int cline_size = cache_line_size(); + + u8 tmp[cline_size]; + void __iomem *addr; + unsigned int offs, timeout = 100000; + + atomic_set(&lmc->ecc_int, 0); + + lmc->mem = alloc_pages_node(lmc->node, GFP_KERNEL, 0); + + if (!lmc->mem) + return -ENOMEM; + + addr = page_address(lmc->mem); + + while (!atomic_read(&lmc->ecc_int) && timeout--) { + stop_machine(inject_ecc_fn, lmc, NULL); + + for (offs = 0; offs < PAGE_SIZE; offs += sizeof(tmp)) { + /* + * Do a load from the previously rigged location + * This should generate an error interrupt. + */ + memcpy(tmp, addr + offs, cline_size); + asm volatile("dsb ld\n"); + } + } + + __free_pages(lmc->mem, 0); + + return count; +} + +LMC_DEBUGFS_ENT(mask0); +LMC_DEBUGFS_ENT(mask2); +LMC_DEBUGFS_ENT(parity_test); + +DEBUGFS_STRUCT(inject_int, 0200, thunderx_lmc_inject_int_write, NULL); +DEBUGFS_STRUCT(inject_ecc, 0200, thunderx_lmc_inject_ecc_write, NULL); +DEBUGFS_STRUCT(int_w1c, 0400, NULL, thunderx_lmc_int_read); + +struct debugfs_entry *lmc_dfs_ents[] = { + &debugfs_mask0, + &debugfs_mask2, + &debugfs_parity_test, + &debugfs_inject_ecc, + &debugfs_inject_int, + &debugfs_int_w1c, +}; + +static int thunderx_create_debugfs_nodes(struct dentry *parent, + struct debugfs_entry *attrs[], + void *data, + size_t num) +{ + int i; + struct dentry *ent; + + if (!IS_ENABLED(CONFIG_EDAC_DEBUG)) + return 0; + + if (!parent) + return -ENOENT; + + for (i = 0; i < num; i++) { + ent = edac_debugfs_create_file(attrs[i]->name, attrs[i]->mode, + parent, data, &attrs[i]->fops); + + if (!ent) + break; + } + + return i; +} + +static phys_addr_t thunderx_faddr_to_phys(u64 faddr, struct thunderx_lmc *lmc) +{ + phys_addr_t addr = 0; + int bank, xbits; + + addr |= lmc->node << 40; + addr |= LMC_FADR_FDIMM(faddr) << lmc->dimm_lsb; + addr |= LMC_FADR_FBUNK(faddr) << lmc->rank_lsb; + addr |= LMC_FADR_FROW(faddr) << lmc->row_lsb; + addr |= (LMC_FADR_FCOL(faddr) >> 4) << lmc->col_hi_lsb; + + bank = LMC_FADR_FBANK(faddr) << lmc->bank_lsb; + + if (lmc->xor_bank) + bank ^= get_bits(addr, 12 + lmc->xbits, lmc->bank_width); + + addr |= bank << lmc->bank_lsb; + + xbits = PCI_FUNC(lmc->pdev->devfn); + + if (lmc->l2c_alias) + xbits ^= get_bits(addr, 20, lmc->xbits) ^ + get_bits(addr, 12, lmc->xbits); + + addr |= xbits << 7; + + return addr; +} + +static unsigned int thunderx_get_num_lmcs(unsigned int node) +{ + unsigned int number = 0; + struct pci_dev *pdev = NULL; + + do { + pdev = pci_get_device(PCI_VENDOR_ID_CAVIUM, + PCI_DEVICE_ID_THUNDER_LMC, + pdev); + if (pdev) { +#ifdef CONFIG_NUMA + if (pdev->dev.numa_node == node) + number++; +#else + number++; +#endif + } + } while (pdev); + + return number; +} + +#define LMC_MESSAGE_SIZE 120 +#define LMC_OTHER_SIZE (50 * ARRAY_SIZE(lmc_errors)) + +static irqreturn_t thunderx_lmc_err_isr(int irq, void *dev_id) +{ + struct mem_ctl_info *mci = dev_id; + struct thunderx_lmc *lmc = mci->pvt_info; + + unsigned long head = ring_pos(lmc->ring_head, ARRAY_SIZE(lmc->err_ctx)); + struct lmc_err_ctx *ctx = &lmc->err_ctx[head]; + + writeq(0, lmc->regs + LMC_CHAR_MASK0); + writeq(0, lmc->regs + LMC_CHAR_MASK2); + writeq(0x2, lmc->regs + LMC_ECC_PARITY_TEST); + + ctx->reg_int = readq(lmc->regs + LMC_INT); + ctx->reg_fadr = readq(lmc->regs + LMC_FADR); + ctx->reg_nxm_fadr = readq(lmc->regs + LMC_NXM_FADR); + ctx->reg_scram_fadr = readq(lmc->regs + LMC_SCRAM_FADR); + ctx->reg_ecc_synd = readq(lmc->regs + LMC_ECC_SYND); + + lmc->ring_head++; + + atomic_set(&lmc->ecc_int, 1); + + /* Clear the interrupt */ + writeq(ctx->reg_int, lmc->regs + LMC_INT); + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t thunderx_lmc_threaded_isr(int irq, void *dev_id) +{ + struct mem_ctl_info *mci = dev_id; + struct thunderx_lmc *lmc = mci->pvt_info; + phys_addr_t phys_addr; + + unsigned long tail; + struct lmc_err_ctx *ctx; + + irqreturn_t ret = IRQ_NONE; + + char *msg; + char *other; + + msg = kmalloc(LMC_MESSAGE_SIZE, GFP_KERNEL); + other = kmalloc(LMC_OTHER_SIZE, GFP_KERNEL); + + if (!msg || !other) + goto err_free; + + while (CIRC_CNT(lmc->ring_head, lmc->ring_tail, + ARRAY_SIZE(lmc->err_ctx))) { + tail = ring_pos(lmc->ring_tail, ARRAY_SIZE(lmc->err_ctx)); + + ctx = &lmc->err_ctx[tail]; + + dev_dbg(&lmc->pdev->dev, "LMC_INT: %016llx\n", + ctx->reg_int); + dev_dbg(&lmc->pdev->dev, "LMC_FADR: %016llx\n", + ctx->reg_fadr); + dev_dbg(&lmc->pdev->dev, "LMC_NXM_FADR: %016llx\n", + ctx->reg_nxm_fadr); + dev_dbg(&lmc->pdev->dev, "LMC_SCRAM_FADR: %016llx\n", + ctx->reg_scram_fadr); + dev_dbg(&lmc->pdev->dev, "LMC_ECC_SYND: %016llx\n", + ctx->reg_ecc_synd); + + snprintf(msg, LMC_MESSAGE_SIZE, + "DIMM %lld rank %lld bank %lld row %lld col %lld", + LMC_FADR_FDIMM(ctx->reg_scram_fadr), + LMC_FADR_FBUNK(ctx->reg_scram_fadr), + LMC_FADR_FBANK(ctx->reg_scram_fadr), + LMC_FADR_FROW(ctx->reg_scram_fadr), + LMC_FADR_FCOL(ctx->reg_scram_fadr)); + + decode_register(other, LMC_OTHER_SIZE, lmc_errors, + ctx->reg_int); + + phys_addr = thunderx_faddr_to_phys(ctx->reg_fadr, lmc); + + if (ctx->reg_int & LMC_INT_UE) + edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci, 1, + phys_to_pfn(phys_addr), + offset_in_page(phys_addr), + 0, -1, -1, -1, msg, other); + else if (ctx->reg_int & LMC_INT_CE) + edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, 1, + phys_to_pfn(phys_addr), + offset_in_page(phys_addr), + 0, -1, -1, -1, msg, other); + + lmc->ring_tail++; + } + + ret = IRQ_HANDLED; + +err_free: + kfree(msg); + kfree(other); + + return ret; +} + +#ifdef CONFIG_PM +static int thunderx_lmc_suspend(struct pci_dev *pdev, pm_message_t state) +{ + pci_save_state(pdev); + pci_disable_device(pdev); + + pci_set_power_state(pdev, pci_choose_state(pdev, state)); + + return 0; +} + +static int thunderx_lmc_resume(struct pci_dev *pdev) +{ + pci_set_power_state(pdev, PCI_D0); + pci_enable_wake(pdev, PCI_D0, 0); + pci_restore_state(pdev); + + return 0; +} +#endif + +static const struct pci_device_id thunderx_lmc_pci_tbl[] = { + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_LMC) }, + { 0, }, +}; + +static inline int pci_dev_to_mc_idx(struct pci_dev *pdev) +{ + int node = dev_to_node(&pdev->dev); + int ret = PCI_FUNC(pdev->devfn); + + ret += max(node, 0) << 3; + + return ret; +} + +static int thunderx_lmc_probe(struct pci_dev *pdev, + const struct pci_device_id *id) +{ + struct thunderx_lmc *lmc; + struct edac_mc_layer layer; + struct mem_ctl_info *mci; + u64 lmc_control, lmc_ddr_pll_ctl, lmc_config; + int ret; + u64 lmc_int; + void *l2c_ioaddr; + + layer.type = EDAC_MC_LAYER_SLOT; + layer.size = 2; + layer.is_virt_csrow = false; + + ret = pcim_enable_device(pdev); + if (ret) { + dev_err(&pdev->dev, "Cannot enable PCI device: %d\n", ret); + return ret; + } + + ret = pcim_iomap_regions(pdev, BIT(0), "thunderx_lmc"); + if (ret) { + dev_err(&pdev->dev, "Cannot map PCI resources: %d\n", ret); + return ret; + } + + mci = edac_mc_alloc(pci_dev_to_mc_idx(pdev), 1, &layer, + sizeof(struct thunderx_lmc)); + if (!mci) + return -ENOMEM; + + mci->pdev = &pdev->dev; + lmc = mci->pvt_info; + + pci_set_drvdata(pdev, mci); + + lmc->regs = pcim_iomap_table(pdev)[0]; + + lmc_control = readq(lmc->regs + LMC_CONTROL); + lmc_ddr_pll_ctl = readq(lmc->regs + LMC_DDR_PLL_CTL); + lmc_config = readq(lmc->regs + LMC_CONFIG); + + if (lmc_control & LMC_CONTROL_RDIMM) { + mci->mtype_cap = FIELD_GET(LMC_DDR_PLL_CTL_DDR4, + lmc_ddr_pll_ctl) ? + MEM_RDDR4 : MEM_RDDR3; + } else { + mci->mtype_cap = FIELD_GET(LMC_DDR_PLL_CTL_DDR4, + lmc_ddr_pll_ctl) ? + MEM_DDR4 : MEM_DDR3; + } + + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED; + mci->edac_cap = EDAC_FLAG_SECDED; + + mci->mod_name = "thunderx-lmc"; + mci->mod_ver = "1"; + mci->ctl_name = "thunderx-lmc"; + mci->dev_name = dev_name(&pdev->dev); + mci->scrub_mode = SCRUB_NONE; + + lmc->pdev = pdev; + lmc->msix_ent.entry = 0; + + lmc->ring_head = 0; + lmc->ring_tail = 0; + + ret = pci_enable_msix_exact(pdev, &lmc->msix_ent, 1); + if (ret) { + dev_err(&pdev->dev, "Cannot enable interrupt: %d\n", ret); + goto err_free; + } + + ret = devm_request_threaded_irq(&pdev->dev, lmc->msix_ent.vector, + thunderx_lmc_err_isr, + thunderx_lmc_threaded_isr, 0, + "[EDAC] ThunderX LMC", mci); + if (ret) { + dev_err(&pdev->dev, "Cannot set ISR: %d\n", ret); + goto err_free; + } + + lmc->node = FIELD_GET(THUNDERX_NODE, pci_resource_start(pdev, 0)); + + lmc->xbits = thunderx_get_num_lmcs(lmc->node) >> 1; + lmc->bank_width = (FIELD_GET(LMC_DDR_PLL_CTL_DDR4, lmc_ddr_pll_ctl) && + FIELD_GET(LMC_CONFIG_BG2, lmc_config)) ? 4 : 3; + + lmc->pbank_lsb = (lmc_config >> 5) & 0xf; + lmc->dimm_lsb = 28 + lmc->pbank_lsb + lmc->xbits; + lmc->rank_lsb = lmc->dimm_lsb; + lmc->rank_lsb -= FIELD_GET(LMC_CONFIG_RANK_ENA, lmc_config) ? 1 : 0; + lmc->bank_lsb = 7 + lmc->xbits; + lmc->row_lsb = 14 + LMC_CONFIG_ROW_LSB(lmc_config) + lmc->xbits; + + lmc->col_hi_lsb = lmc->bank_lsb + lmc->bank_width; + + lmc->xor_bank = lmc_control & LMC_CONTROL_XOR_BANK; + + l2c_ioaddr = ioremap(L2C_CTL | FIELD_PREP(THUNDERX_NODE, lmc->node), + PAGE_SIZE); + + if (!l2c_ioaddr) { + dev_err(&pdev->dev, "Cannot map L2C_CTL\n"); + goto err_free; + } + + lmc->l2c_alias = !(readq(l2c_ioaddr) & L2C_CTL_DISIDXALIAS); + + iounmap(l2c_ioaddr); + + ret = edac_mc_add_mc(mci); + if (ret) { + dev_err(&pdev->dev, "Cannot add the MC: %d\n", ret); + goto err_free; + } + + lmc_int = readq(lmc->regs + LMC_INT); + writeq(lmc_int, lmc->regs + LMC_INT); + + writeq(LMC_INT_ENA_ALL, lmc->regs + LMC_INT_ENA_W1S); + + if (IS_ENABLED(CONFIG_EDAC_DEBUG)) { + ret = thunderx_create_debugfs_nodes(mci->debugfs, + lmc_dfs_ents, + lmc, + ARRAY_SIZE(lmc_dfs_ents)); + + if (ret != ARRAY_SIZE(lmc_dfs_ents)) { + dev_warn(&pdev->dev, "Error creating debugfs entries: %d%s\n", + ret, ret >= 0 ? " created" : ""); + } + } + + return 0; + +err_free: + pci_set_drvdata(pdev, NULL); + edac_mc_free(mci); + + return ret; +} + +static void thunderx_lmc_remove(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci = pci_get_drvdata(pdev); + struct thunderx_lmc *lmc = mci->pvt_info; + + writeq(LMC_INT_ENA_ALL, lmc->regs + LMC_INT_ENA_W1C); + + edac_mc_del_mc(&pdev->dev); + edac_mc_free(mci); +} + +MODULE_DEVICE_TABLE(pci, thunderx_lmc_pci_tbl); + +static struct pci_driver thunderx_lmc_driver = { + .name = "thunderx_lmc_edac", + .probe = thunderx_lmc_probe, + .remove = thunderx_lmc_remove, +#ifdef CONFIG_PM + .suspend = thunderx_lmc_suspend, + .resume = thunderx_lmc_resume, +#endif + .id_table = thunderx_lmc_pci_tbl, +}; + +/*---------------------- OCX driver ---------------------------------*/ + +#define PCI_DEVICE_ID_THUNDER_OCX 0xa013 + +#define OCX_LINK_INTS 3 +#define OCX_INTS (OCX_LINK_INTS + 1) +#define OCX_RX_LANES 24 +#define OCX_RX_LANE_STATS 15 + +#define OCX_COM_INT 0x100 +#define OCX_COM_INT_W1S 0x108 +#define OCX_COM_INT_ENA_W1S 0x110 +#define OCX_COM_INT_ENA_W1C 0x118 + +#define OCX_COM_IO_BADID BIT(54) +#define OCX_COM_MEM_BADID BIT(53) +#define OCX_COM_COPR_BADID BIT(52) +#define OCX_COM_WIN_REQ_BADID BIT(51) +#define OCX_COM_WIN_REQ_TOUT BIT(50) +#define OCX_COM_RX_LANE GENMASK(23, 0) + +#define OCX_COM_INT_CE (OCX_COM_IO_BADID | \ + OCX_COM_MEM_BADID | \ + OCX_COM_COPR_BADID | \ + OCX_COM_WIN_REQ_BADID | \ + OCX_COM_WIN_REQ_TOUT) + +static const struct error_descr ocx_com_errors[] = { + { + .type = ERR_CORRECTED, + .mask = OCX_COM_IO_BADID, + .descr = "Invalid IO transaction node ID", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_MEM_BADID, + .descr = "Invalid memory transaction node ID", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_COPR_BADID, + .descr = "Invalid coprocessor transaction node ID", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_WIN_REQ_BADID, + .descr = "Invalid SLI transaction node ID", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_WIN_REQ_TOUT, + .descr = "Window/core request timeout", + }, + {0, 0, NULL}, +}; + +#define OCX_COM_LINKX_INT(x) (0x120 + (x) * 8) +#define OCX_COM_LINKX_INT_W1S(x) (0x140 + (x) * 8) +#define OCX_COM_LINKX_INT_ENA_W1S(x) (0x160 + (x) * 8) +#define OCX_COM_LINKX_INT_ENA_W1C(x) (0x180 + (x) * 8) + +#define OCX_COM_LINK_BAD_WORD BIT(13) +#define OCX_COM_LINK_ALIGN_FAIL BIT(12) +#define OCX_COM_LINK_ALIGN_DONE BIT(11) +#define OCX_COM_LINK_UP BIT(10) +#define OCX_COM_LINK_STOP BIT(9) +#define OCX_COM_LINK_BLK_ERR BIT(8) +#define OCX_COM_LINK_REINIT BIT(7) +#define OCX_COM_LINK_LNK_DATA BIT(6) +#define OCX_COM_LINK_RXFIFO_DBE BIT(5) +#define OCX_COM_LINK_RXFIFO_SBE BIT(4) +#define OCX_COM_LINK_TXFIFO_DBE BIT(3) +#define OCX_COM_LINK_TXFIFO_SBE BIT(2) +#define OCX_COM_LINK_REPLAY_DBE BIT(1) +#define OCX_COM_LINK_REPLAY_SBE BIT(0) + +static const struct error_descr ocx_com_link_errors[] = { + { + .type = ERR_CORRECTED, + .mask = OCX_COM_LINK_REPLAY_SBE, + .descr = "Replay buffer single-bit error", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_LINK_TXFIFO_SBE, + .descr = "TX FIFO single-bit error", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_LINK_RXFIFO_SBE, + .descr = "RX FIFO single-bit error", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_LINK_BLK_ERR, + .descr = "Block code error", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_LINK_ALIGN_FAIL, + .descr = "Link alignment failure", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_COM_LINK_BAD_WORD, + .descr = "Bad code word", + }, + { + .type = ERR_UNCORRECTED, + .mask = OCX_COM_LINK_REPLAY_DBE, + .descr = "Replay buffer double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = OCX_COM_LINK_TXFIFO_DBE, + .descr = "TX FIFO double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = OCX_COM_LINK_RXFIFO_DBE, + .descr = "RX FIFO double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = OCX_COM_LINK_STOP, + .descr = "Link stopped", + }, + {0, 0, NULL}, +}; + +#define OCX_COM_LINK_INT_UE (OCX_COM_LINK_REPLAY_DBE | \ + OCX_COM_LINK_TXFIFO_DBE | \ + OCX_COM_LINK_RXFIFO_DBE | \ + OCX_COM_LINK_STOP) + +#define OCX_COM_LINK_INT_CE (OCX_COM_LINK_REPLAY_SBE | \ + OCX_COM_LINK_TXFIFO_SBE | \ + OCX_COM_LINK_RXFIFO_SBE | \ + OCX_COM_LINK_BLK_ERR | \ + OCX_COM_LINK_ALIGN_FAIL | \ + OCX_COM_LINK_BAD_WORD) + +#define OCX_LNE_INT(x) (0x8018 + (x) * 0x100) +#define OCX_LNE_INT_EN(x) (0x8020 + (x) * 0x100) +#define OCX_LNE_BAD_CNT(x) (0x8028 + (x) * 0x100) +#define OCX_LNE_CFG(x) (0x8000 + (x) * 0x100) +#define OCX_LNE_STAT(x, y) (0x8040 + (x) * 0x100 + (y) * 8) + +#define OCX_LNE_CFG_RX_BDRY_LOCK_DIS BIT(8) +#define OCX_LNE_CFG_RX_STAT_WRAP_DIS BIT(2) +#define OCX_LNE_CFG_RX_STAT_RDCLR BIT(1) +#define OCX_LNE_CFG_RX_STAT_ENA BIT(0) + + +#define OCX_LANE_BAD_64B67B BIT(8) +#define OCX_LANE_DSKEW_FIFO_OVFL BIT(5) +#define OCX_LANE_SCRM_SYNC_LOSS BIT(4) +#define OCX_LANE_UKWN_CNTL_WORD BIT(3) +#define OCX_LANE_CRC32_ERR BIT(2) +#define OCX_LANE_BDRY_SYNC_LOSS BIT(1) +#define OCX_LANE_SERDES_LOCK_LOSS BIT(0) + +#define OCX_COM_LANE_INT_UE (0) +#define OCX_COM_LANE_INT_CE (OCX_LANE_SERDES_LOCK_LOSS | \ + OCX_LANE_BDRY_SYNC_LOSS | \ + OCX_LANE_CRC32_ERR | \ + OCX_LANE_UKWN_CNTL_WORD | \ + OCX_LANE_SCRM_SYNC_LOSS | \ + OCX_LANE_DSKEW_FIFO_OVFL | \ + OCX_LANE_BAD_64B67B) + +static const struct error_descr ocx_lane_errors[] = { + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_SERDES_LOCK_LOSS, + .descr = "RX SerDes lock lost", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_BDRY_SYNC_LOSS, + .descr = "RX word boundary lost", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_CRC32_ERR, + .descr = "CRC32 error", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_UKWN_CNTL_WORD, + .descr = "Unknown control word", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_SCRM_SYNC_LOSS, + .descr = "Scrambler synchronization lost", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_DSKEW_FIFO_OVFL, + .descr = "RX deskew FIFO overflow", + }, + { + .type = ERR_CORRECTED, + .mask = OCX_LANE_BAD_64B67B, + .descr = "Bad 64B/67B codeword", + }, + {0, 0, NULL}, +}; + +#define OCX_LNE_INT_ENA_ALL (GENMASK(9, 8) | GENMASK(6, 0)) +#define OCX_COM_INT_ENA_ALL (GENMASK(54, 50) | GENMASK(23, 0)) +#define OCX_COM_LINKX_INT_ENA_ALL (GENMASK(13, 12) | \ + GENMASK(9, 7) | GENMASK(5, 0)) + +#define OCX_TLKX_ECC_CTL(x) (0x10018 + (x) * 0x2000) +#define OCX_RLKX_ECC_CTL(x) (0x18018 + (x) * 0x2000) + +struct ocx_com_err_ctx { + u64 reg_com_int; + u64 reg_lane_int[OCX_RX_LANES]; + u64 reg_lane_stat11[OCX_RX_LANES]; +}; + +struct ocx_link_err_ctx { + u64 reg_com_link_int; + int link; +}; + +struct thunderx_ocx { + void __iomem *regs; + int com_link; + struct pci_dev *pdev; + struct edac_device_ctl_info *edac_dev; + + struct dentry *debugfs; + struct msix_entry msix_ent[OCX_INTS]; + + struct ocx_com_err_ctx com_err_ctx[RING_ENTRIES]; + struct ocx_link_err_ctx link_err_ctx[RING_ENTRIES]; + + unsigned long com_ring_head; + unsigned long com_ring_tail; + + unsigned long link_ring_head; + unsigned long link_ring_tail; +}; + +#define OCX_MESSAGE_SIZE SZ_1K +#define OCX_OTHER_SIZE (50 * ARRAY_SIZE(ocx_com_link_errors)) + +/* This handler is threaded */ +static irqreturn_t thunderx_ocx_com_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_ocx *ocx = container_of(msix, struct thunderx_ocx, + msix_ent[msix->entry]); + + int lane; + unsigned long head = ring_pos(ocx->com_ring_head, + ARRAY_SIZE(ocx->com_err_ctx)); + struct ocx_com_err_ctx *ctx = &ocx->com_err_ctx[head]; + + ctx->reg_com_int = readq(ocx->regs + OCX_COM_INT); + + for (lane = 0; lane < OCX_RX_LANES; lane++) { + ctx->reg_lane_int[lane] = + readq(ocx->regs + OCX_LNE_INT(lane)); + ctx->reg_lane_stat11[lane] = + readq(ocx->regs + OCX_LNE_STAT(lane, 11)); + + writeq(ctx->reg_lane_int[lane], ocx->regs + OCX_LNE_INT(lane)); + } + + writeq(ctx->reg_com_int, ocx->regs + OCX_COM_INT); + + ocx->com_ring_head++; + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t thunderx_ocx_com_threaded_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_ocx *ocx = container_of(msix, struct thunderx_ocx, + msix_ent[msix->entry]); + + irqreturn_t ret = IRQ_NONE; + + unsigned long tail; + struct ocx_com_err_ctx *ctx; + int lane; + char *msg; + char *other; + + msg = kmalloc(OCX_MESSAGE_SIZE, GFP_KERNEL); + other = kmalloc(OCX_OTHER_SIZE, GFP_KERNEL); + + if (!msg || !other) + goto err_free; + + while (CIRC_CNT(ocx->com_ring_head, ocx->com_ring_tail, + ARRAY_SIZE(ocx->com_err_ctx))) { + tail = ring_pos(ocx->com_ring_tail, + ARRAY_SIZE(ocx->com_err_ctx)); + ctx = &ocx->com_err_ctx[tail]; + + snprintf(msg, OCX_MESSAGE_SIZE, "%s: OCX_COM_INT: %016llx", + ocx->edac_dev->ctl_name, ctx->reg_com_int); + + decode_register(other, OCX_OTHER_SIZE, + ocx_com_errors, ctx->reg_com_int); + + strncat(msg, other, OCX_MESSAGE_SIZE); + + for (lane = 0; lane < OCX_RX_LANES; lane++) + if (ctx->reg_com_int & BIT(lane)) { + snprintf(other, OCX_OTHER_SIZE, + "\n\tOCX_LNE_INT[%02d]: %016llx OCX_LNE_STAT11[%02d]: %016llx", + lane, ctx->reg_lane_int[lane], + lane, ctx->reg_lane_stat11[lane]); + + strncat(msg, other, OCX_MESSAGE_SIZE); + + decode_register(other, OCX_OTHER_SIZE, + ocx_lane_errors, + ctx->reg_lane_int[lane]); + strncat(msg, other, OCX_MESSAGE_SIZE); + } + + if (ctx->reg_com_int & OCX_COM_INT_CE) + edac_device_handle_ce(ocx->edac_dev, 0, 0, msg); + + ocx->com_ring_tail++; + } + + ret = IRQ_HANDLED; + +err_free: + kfree(other); + kfree(msg); + + return ret; +} + +static irqreturn_t thunderx_ocx_lnk_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_ocx *ocx = container_of(msix, struct thunderx_ocx, + msix_ent[msix->entry]); + unsigned long head = ring_pos(ocx->link_ring_head, + ARRAY_SIZE(ocx->link_err_ctx)); + struct ocx_link_err_ctx *ctx = &ocx->link_err_ctx[head]; + + ctx->link = msix->entry; + ctx->reg_com_link_int = readq(ocx->regs + OCX_COM_LINKX_INT(ctx->link)); + + writeq(ctx->reg_com_link_int, ocx->regs + OCX_COM_LINKX_INT(ctx->link)); + + ocx->link_ring_head++; + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t thunderx_ocx_lnk_threaded_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_ocx *ocx = container_of(msix, struct thunderx_ocx, + msix_ent[msix->entry]); + irqreturn_t ret = IRQ_NONE; + unsigned long tail; + struct ocx_link_err_ctx *ctx; + + char *msg; + char *other; + + msg = kmalloc(OCX_MESSAGE_SIZE, GFP_KERNEL); + other = kmalloc(OCX_OTHER_SIZE, GFP_KERNEL); + + if (!msg || !other) + goto err_free; + + while (CIRC_CNT(ocx->link_ring_head, ocx->link_ring_tail, + ARRAY_SIZE(ocx->link_err_ctx))) { + tail = ring_pos(ocx->link_ring_head, + ARRAY_SIZE(ocx->link_err_ctx)); + + ctx = &ocx->link_err_ctx[tail]; + + snprintf(msg, OCX_MESSAGE_SIZE, + "%s: OCX_COM_LINK_INT[%d]: %016llx", + ocx->edac_dev->ctl_name, + ctx->link, ctx->reg_com_link_int); + + decode_register(other, OCX_OTHER_SIZE, + ocx_com_link_errors, ctx->reg_com_link_int); + + strncat(msg, other, OCX_MESSAGE_SIZE); + + if (ctx->reg_com_link_int & OCX_COM_LINK_INT_UE) + edac_device_handle_ue(ocx->edac_dev, 0, 0, msg); + else if (ctx->reg_com_link_int & OCX_COM_LINK_INT_CE) + edac_device_handle_ce(ocx->edac_dev, 0, 0, msg); + + ocx->link_ring_tail++; + } + + ret = IRQ_HANDLED; +err_free: + kfree(other); + kfree(msg); + + return ret; +} + +#define OCX_DEBUGFS_ATTR(_name, _reg) DEBUGFS_REG_ATTR(ocx, _name, _reg) + +OCX_DEBUGFS_ATTR(tlk0_ecc_ctl, OCX_TLKX_ECC_CTL(0)); +OCX_DEBUGFS_ATTR(tlk1_ecc_ctl, OCX_TLKX_ECC_CTL(1)); +OCX_DEBUGFS_ATTR(tlk2_ecc_ctl, OCX_TLKX_ECC_CTL(2)); + +OCX_DEBUGFS_ATTR(rlk0_ecc_ctl, OCX_RLKX_ECC_CTL(0)); +OCX_DEBUGFS_ATTR(rlk1_ecc_ctl, OCX_RLKX_ECC_CTL(1)); +OCX_DEBUGFS_ATTR(rlk2_ecc_ctl, OCX_RLKX_ECC_CTL(2)); + +OCX_DEBUGFS_ATTR(com_link0_int, OCX_COM_LINKX_INT_W1S(0)); +OCX_DEBUGFS_ATTR(com_link1_int, OCX_COM_LINKX_INT_W1S(1)); +OCX_DEBUGFS_ATTR(com_link2_int, OCX_COM_LINKX_INT_W1S(2)); + +OCX_DEBUGFS_ATTR(lne00_badcnt, OCX_LNE_BAD_CNT(0)); +OCX_DEBUGFS_ATTR(lne01_badcnt, OCX_LNE_BAD_CNT(1)); +OCX_DEBUGFS_ATTR(lne02_badcnt, OCX_LNE_BAD_CNT(2)); +OCX_DEBUGFS_ATTR(lne03_badcnt, OCX_LNE_BAD_CNT(3)); +OCX_DEBUGFS_ATTR(lne04_badcnt, OCX_LNE_BAD_CNT(4)); +OCX_DEBUGFS_ATTR(lne05_badcnt, OCX_LNE_BAD_CNT(5)); +OCX_DEBUGFS_ATTR(lne06_badcnt, OCX_LNE_BAD_CNT(6)); +OCX_DEBUGFS_ATTR(lne07_badcnt, OCX_LNE_BAD_CNT(7)); + +OCX_DEBUGFS_ATTR(lne08_badcnt, OCX_LNE_BAD_CNT(8)); +OCX_DEBUGFS_ATTR(lne09_badcnt, OCX_LNE_BAD_CNT(9)); +OCX_DEBUGFS_ATTR(lne10_badcnt, OCX_LNE_BAD_CNT(10)); +OCX_DEBUGFS_ATTR(lne11_badcnt, OCX_LNE_BAD_CNT(11)); +OCX_DEBUGFS_ATTR(lne12_badcnt, OCX_LNE_BAD_CNT(12)); +OCX_DEBUGFS_ATTR(lne13_badcnt, OCX_LNE_BAD_CNT(13)); +OCX_DEBUGFS_ATTR(lne14_badcnt, OCX_LNE_BAD_CNT(14)); +OCX_DEBUGFS_ATTR(lne15_badcnt, OCX_LNE_BAD_CNT(15)); + +OCX_DEBUGFS_ATTR(lne16_badcnt, OCX_LNE_BAD_CNT(16)); +OCX_DEBUGFS_ATTR(lne17_badcnt, OCX_LNE_BAD_CNT(17)); +OCX_DEBUGFS_ATTR(lne18_badcnt, OCX_LNE_BAD_CNT(18)); +OCX_DEBUGFS_ATTR(lne19_badcnt, OCX_LNE_BAD_CNT(19)); +OCX_DEBUGFS_ATTR(lne20_badcnt, OCX_LNE_BAD_CNT(20)); +OCX_DEBUGFS_ATTR(lne21_badcnt, OCX_LNE_BAD_CNT(21)); +OCX_DEBUGFS_ATTR(lne22_badcnt, OCX_LNE_BAD_CNT(22)); +OCX_DEBUGFS_ATTR(lne23_badcnt, OCX_LNE_BAD_CNT(23)); + +OCX_DEBUGFS_ATTR(com_int, OCX_COM_INT_W1S); + +struct debugfs_entry *ocx_dfs_ents[] = { + &debugfs_tlk0_ecc_ctl, + &debugfs_tlk1_ecc_ctl, + &debugfs_tlk2_ecc_ctl, + + &debugfs_rlk0_ecc_ctl, + &debugfs_rlk1_ecc_ctl, + &debugfs_rlk2_ecc_ctl, + + &debugfs_com_link0_int, + &debugfs_com_link1_int, + &debugfs_com_link2_int, + + &debugfs_lne00_badcnt, + &debugfs_lne01_badcnt, + &debugfs_lne02_badcnt, + &debugfs_lne03_badcnt, + &debugfs_lne04_badcnt, + &debugfs_lne05_badcnt, + &debugfs_lne06_badcnt, + &debugfs_lne07_badcnt, + &debugfs_lne08_badcnt, + &debugfs_lne09_badcnt, + &debugfs_lne10_badcnt, + &debugfs_lne11_badcnt, + &debugfs_lne12_badcnt, + &debugfs_lne13_badcnt, + &debugfs_lne14_badcnt, + &debugfs_lne15_badcnt, + &debugfs_lne16_badcnt, + &debugfs_lne17_badcnt, + &debugfs_lne18_badcnt, + &debugfs_lne19_badcnt, + &debugfs_lne20_badcnt, + &debugfs_lne21_badcnt, + &debugfs_lne22_badcnt, + &debugfs_lne23_badcnt, + + &debugfs_com_int, +}; + +static const struct pci_device_id thunderx_ocx_pci_tbl[] = { + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_OCX) }, + { 0, }, +}; + +static void thunderx_ocx_clearstats(struct thunderx_ocx *ocx) +{ + int lane, stat, cfg; + + for (lane = 0; lane < OCX_RX_LANES; lane++) { + cfg = readq(ocx->regs + OCX_LNE_CFG(lane)); + cfg |= OCX_LNE_CFG_RX_STAT_RDCLR; + cfg &= ~OCX_LNE_CFG_RX_STAT_ENA; + writeq(cfg, ocx->regs + OCX_LNE_CFG(lane)); + + for (stat = 0; stat < OCX_RX_LANE_STATS; stat++) + readq(ocx->regs + OCX_LNE_STAT(lane, stat)); + } +} + +static int thunderx_ocx_probe(struct pci_dev *pdev, + const struct pci_device_id *id) +{ + struct thunderx_ocx *ocx; + struct edac_device_ctl_info *edac_dev; + char name[32]; + int idx; + int i; + int ret; + u64 reg; + + ret = pcim_enable_device(pdev); + if (ret) { + dev_err(&pdev->dev, "Cannot enable PCI device: %d\n", ret); + return ret; + } + + ret = pcim_iomap_regions(pdev, BIT(0), "thunderx_ocx"); + if (ret) { + dev_err(&pdev->dev, "Cannot map PCI resources: %d\n", ret); + return ret; + } + + idx = edac_device_alloc_index(); + snprintf(name, sizeof(name), "OCX%d", idx); + edac_dev = edac_device_alloc_ctl_info(sizeof(struct thunderx_ocx), + name, 1, "CCPI", 1, + 0, NULL, 0, idx); + if (!edac_dev) { + dev_err(&pdev->dev, "Cannot allocate EDAC device: %d\n", ret); + return -ENOMEM; + } + ocx = edac_dev->pvt_info; + ocx->edac_dev = edac_dev; + ocx->com_ring_head = 0; + ocx->com_ring_tail = 0; + ocx->link_ring_head = 0; + ocx->link_ring_tail = 0; + + ocx->regs = pcim_iomap_table(pdev)[0]; + if (!ocx->regs) { + dev_err(&pdev->dev, "Cannot map PCI resources: %d\n", ret); + ret = -ENODEV; + goto err_free; + } + + ocx->pdev = pdev; + + for (i = 0; i < OCX_INTS; i++) { + ocx->msix_ent[i].entry = i; + ocx->msix_ent[i].vector = 0; + } + + ret = pci_enable_msix_exact(pdev, ocx->msix_ent, OCX_INTS); + if (ret) { + dev_err(&pdev->dev, "Cannot enable interrupt: %d\n", ret); + goto err_free; + } + + for (i = 0; i < OCX_INTS; i++) { + ret = devm_request_threaded_irq(&pdev->dev, + ocx->msix_ent[i].vector, + (i == 3) ? + thunderx_ocx_com_isr : + thunderx_ocx_lnk_isr, + (i == 3) ? + thunderx_ocx_com_threaded_isr : + thunderx_ocx_lnk_threaded_isr, + 0, "[EDAC] ThunderX OCX", + &ocx->msix_ent[i]); + if (ret) + goto err_free; + } + + edac_dev->dev = &pdev->dev; + edac_dev->dev_name = dev_name(&pdev->dev); + edac_dev->mod_name = "thunderx-ocx"; + edac_dev->ctl_name = "thunderx-ocx"; + + ret = edac_device_add_device(edac_dev); + if (ret) { + dev_err(&pdev->dev, "Cannot add EDAC device: %d\n", ret); + goto err_free; + } + + if (IS_ENABLED(CONFIG_EDAC_DEBUG)) { + ocx->debugfs = edac_debugfs_create_dir(pdev->dev.kobj.name); + + ret = thunderx_create_debugfs_nodes(ocx->debugfs, + ocx_dfs_ents, + ocx, + ARRAY_SIZE(ocx_dfs_ents)); + if (ret != ARRAY_SIZE(ocx_dfs_ents)) { + dev_warn(&pdev->dev, "Error creating debugfs entries: %d%s\n", + ret, ret >= 0 ? " created" : ""); + } + } + + pci_set_drvdata(pdev, edac_dev); + + thunderx_ocx_clearstats(ocx); + + for (i = 0; i < OCX_RX_LANES; i++) { + writeq(OCX_LNE_INT_ENA_ALL, + ocx->regs + OCX_LNE_INT_EN(i)); + + reg = readq(ocx->regs + OCX_LNE_INT(i)); + writeq(reg, ocx->regs + OCX_LNE_INT(i)); + + } + + for (i = 0; i < OCX_LINK_INTS; i++) { + reg = readq(ocx->regs + OCX_COM_LINKX_INT(i)); + writeq(reg, ocx->regs + OCX_COM_LINKX_INT(i)); + + writeq(OCX_COM_LINKX_INT_ENA_ALL, + ocx->regs + OCX_COM_LINKX_INT_ENA_W1S(i)); + } + + reg = readq(ocx->regs + OCX_COM_INT); + writeq(reg, ocx->regs + OCX_COM_INT); + + writeq(OCX_COM_INT_ENA_ALL, ocx->regs + OCX_COM_INT_ENA_W1S); + + return 0; +err_free: + edac_device_free_ctl_info(edac_dev); + + return ret; +} + +static void thunderx_ocx_remove(struct pci_dev *pdev) +{ + struct edac_device_ctl_info *edac_dev = pci_get_drvdata(pdev); + struct thunderx_ocx *ocx = edac_dev->pvt_info; + int i; + + writeq(OCX_COM_INT_ENA_ALL, ocx->regs + OCX_COM_INT_ENA_W1C); + + for (i = 0; i < OCX_INTS; i++) { + writeq(OCX_COM_LINKX_INT_ENA_ALL, + ocx->regs + OCX_COM_LINKX_INT_ENA_W1C(i)); + } + + edac_debugfs_remove_recursive(ocx->debugfs); + + edac_device_del_device(&pdev->dev); + edac_device_free_ctl_info(edac_dev); +} + +MODULE_DEVICE_TABLE(pci, thunderx_ocx_pci_tbl); + +static struct pci_driver thunderx_ocx_driver = { + .name = "thunderx_ocx_edac", + .probe = thunderx_ocx_probe, + .remove = thunderx_ocx_remove, + .id_table = thunderx_ocx_pci_tbl, +}; + +/*---------------------- L2C driver ---------------------------------*/ + +#define PCI_DEVICE_ID_THUNDER_L2C_TAD 0xa02e +#define PCI_DEVICE_ID_THUNDER_L2C_CBC 0xa02f +#define PCI_DEVICE_ID_THUNDER_L2C_MCI 0xa030 + +#define L2C_TAD_INT_W1C 0x40000 +#define L2C_TAD_INT_W1S 0x40008 + +#define L2C_TAD_INT_ENA_W1C 0x40020 +#define L2C_TAD_INT_ENA_W1S 0x40028 + + +#define L2C_TAD_INT_L2DDBE BIT(1) +#define L2C_TAD_INT_SBFSBE BIT(2) +#define L2C_TAD_INT_SBFDBE BIT(3) +#define L2C_TAD_INT_FBFSBE BIT(4) +#define L2C_TAD_INT_FBFDBE BIT(5) +#define L2C_TAD_INT_TAGDBE BIT(9) +#define L2C_TAD_INT_RDDISLMC BIT(15) +#define L2C_TAD_INT_WRDISLMC BIT(16) +#define L2C_TAD_INT_LFBTO BIT(17) +#define L2C_TAD_INT_GSYNCTO BIT(18) +#define L2C_TAD_INT_RTGSBE BIT(32) +#define L2C_TAD_INT_RTGDBE BIT(33) +#define L2C_TAD_INT_RDDISOCI BIT(34) +#define L2C_TAD_INT_WRDISOCI BIT(35) + +#define L2C_TAD_INT_ECC (L2C_TAD_INT_L2DDBE | \ + L2C_TAD_INT_SBFSBE | L2C_TAD_INT_SBFDBE | \ + L2C_TAD_INT_FBFSBE | L2C_TAD_INT_FBFDBE) + +#define L2C_TAD_INT_CE (L2C_TAD_INT_SBFSBE | \ + L2C_TAD_INT_FBFSBE) + +#define L2C_TAD_INT_UE (L2C_TAD_INT_L2DDBE | \ + L2C_TAD_INT_SBFDBE | \ + L2C_TAD_INT_FBFDBE | \ + L2C_TAD_INT_TAGDBE | \ + L2C_TAD_INT_RTGDBE | \ + L2C_TAD_INT_WRDISOCI | \ + L2C_TAD_INT_RDDISOCI | \ + L2C_TAD_INT_WRDISLMC | \ + L2C_TAD_INT_RDDISLMC | \ + L2C_TAD_INT_LFBTO | \ + L2C_TAD_INT_GSYNCTO) + +static const struct error_descr l2_tad_errors[] = { + { + .type = ERR_CORRECTED, + .mask = L2C_TAD_INT_SBFSBE, + .descr = "SBF single-bit error", + }, + { + .type = ERR_CORRECTED, + .mask = L2C_TAD_INT_FBFSBE, + .descr = "FBF single-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_L2DDBE, + .descr = "L2D double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_SBFDBE, + .descr = "SBF double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_FBFDBE, + .descr = "FBF double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_TAGDBE, + .descr = "TAG double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_RTGDBE, + .descr = "RTG double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_WRDISOCI, + .descr = "Write to a disabled CCPI", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_RDDISOCI, + .descr = "Read from a disabled CCPI", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_WRDISLMC, + .descr = "Write to a disabled LMC", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_RDDISLMC, + .descr = "Read from a disabled LMC", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_LFBTO, + .descr = "LFB entry timeout", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_TAD_INT_GSYNCTO, + .descr = "Global sync CCPI timeout", + }, + {0, 0, NULL}, +}; + +#define L2C_TAD_INT_TAG (L2C_TAD_INT_TAGDBE) + +#define L2C_TAD_INT_RTG (L2C_TAD_INT_RTGDBE) + +#define L2C_TAD_INT_DISLMC (L2C_TAD_INT_WRDISLMC | L2C_TAD_INT_RDDISLMC) + +#define L2C_TAD_INT_DISOCI (L2C_TAD_INT_WRDISOCI | L2C_TAD_INT_RDDISOCI) + +#define L2C_TAD_INT_ENA_ALL (L2C_TAD_INT_ECC | L2C_TAD_INT_TAG | \ + L2C_TAD_INT_RTG | \ + L2C_TAD_INT_DISLMC | L2C_TAD_INT_DISOCI | \ + L2C_TAD_INT_LFBTO) + +#define L2C_TAD_TIMETWO 0x50000 +#define L2C_TAD_TIMEOUT 0x50100 +#define L2C_TAD_ERR 0x60000 +#define L2C_TAD_TQD_ERR 0x60100 +#define L2C_TAD_TTG_ERR 0x60200 + + +#define L2C_CBC_INT_W1C 0x60000 + +#define L2C_CBC_INT_RSDSBE BIT(0) +#define L2C_CBC_INT_RSDDBE BIT(1) + +#define L2C_CBC_INT_RSD (L2C_CBC_INT_RSDSBE | L2C_CBC_INT_RSDDBE) + +#define L2C_CBC_INT_MIBSBE BIT(4) +#define L2C_CBC_INT_MIBDBE BIT(5) + +#define L2C_CBC_INT_MIB (L2C_CBC_INT_MIBSBE | L2C_CBC_INT_MIBDBE) + +#define L2C_CBC_INT_IORDDISOCI BIT(6) +#define L2C_CBC_INT_IOWRDISOCI BIT(7) + +#define L2C_CBC_INT_IODISOCI (L2C_CBC_INT_IORDDISOCI | \ + L2C_CBC_INT_IOWRDISOCI) + +#define L2C_CBC_INT_CE (L2C_CBC_INT_RSDSBE | L2C_CBC_INT_MIBSBE) +#define L2C_CBC_INT_UE (L2C_CBC_INT_RSDDBE | L2C_CBC_INT_MIBDBE) + + +static const struct error_descr l2_cbc_errors[] = { + { + .type = ERR_CORRECTED, + .mask = L2C_CBC_INT_RSDSBE, + .descr = "RSD single-bit error", + }, + { + .type = ERR_CORRECTED, + .mask = L2C_CBC_INT_MIBSBE, + .descr = "MIB single-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_CBC_INT_RSDDBE, + .descr = "RSD double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_CBC_INT_MIBDBE, + .descr = "MIB double-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_CBC_INT_IORDDISOCI, + .descr = "Read from a disabled CCPI", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_CBC_INT_IOWRDISOCI, + .descr = "Write to a disabled CCPI", + }, + {0, 0, NULL}, +}; + +#define L2C_CBC_INT_W1S 0x60008 +#define L2C_CBC_INT_ENA_W1C 0x60020 + +#define L2C_CBC_INT_ENA_ALL (L2C_CBC_INT_RSD | L2C_CBC_INT_MIB | \ + L2C_CBC_INT_IODISOCI) + +#define L2C_CBC_INT_ENA_W1S 0x60028 + +#define L2C_CBC_IODISOCIERR 0x80008 +#define L2C_CBC_IOCERR 0x80010 +#define L2C_CBC_RSDERR 0x80018 +#define L2C_CBC_MIBERR 0x80020 + + +#define L2C_MCI_INT_W1C 0x0 + +#define L2C_MCI_INT_VBFSBE BIT(0) +#define L2C_MCI_INT_VBFDBE BIT(1) + +static const struct error_descr l2_mci_errors[] = { + { + .type = ERR_CORRECTED, + .mask = L2C_MCI_INT_VBFSBE, + .descr = "VBF single-bit error", + }, + { + .type = ERR_UNCORRECTED, + .mask = L2C_MCI_INT_VBFDBE, + .descr = "VBF double-bit error", + }, + {0, 0, NULL}, +}; + +#define L2C_MCI_INT_W1S 0x8 +#define L2C_MCI_INT_ENA_W1C 0x20 + +#define L2C_MCI_INT_ENA_ALL (L2C_MCI_INT_VBFSBE | L2C_MCI_INT_VBFDBE) + +#define L2C_MCI_INT_ENA_W1S 0x28 + +#define L2C_MCI_ERR 0x10000 + +#define L2C_MESSAGE_SIZE SZ_1K +#define L2C_OTHER_SIZE (50 * ARRAY_SIZE(l2_tad_errors)) + +struct l2c_err_ctx { + char *reg_ext_name; + u64 reg_int; + u64 reg_ext; +}; + +struct thunderx_l2c { + void __iomem *regs; + struct pci_dev *pdev; + struct edac_device_ctl_info *edac_dev; + + struct dentry *debugfs; + + int index; + + struct msix_entry msix_ent; + + struct l2c_err_ctx err_ctx[RING_ENTRIES]; + unsigned long ring_head; + unsigned long ring_tail; +}; + +static irqreturn_t thunderx_l2c_tad_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_l2c *tad = container_of(msix, struct thunderx_l2c, + msix_ent); + + unsigned long head = ring_pos(tad->ring_head, ARRAY_SIZE(tad->err_ctx)); + struct l2c_err_ctx *ctx = &tad->err_ctx[head]; + + ctx->reg_int = readq(tad->regs + L2C_TAD_INT_W1C); + + if (ctx->reg_int & L2C_TAD_INT_ECC) { + ctx->reg_ext_name = "TQD_ERR"; + ctx->reg_ext = readq(tad->regs + L2C_TAD_TQD_ERR); + } else if (ctx->reg_int & L2C_TAD_INT_TAG) { + ctx->reg_ext_name = "TTG_ERR"; + ctx->reg_ext = readq(tad->regs + L2C_TAD_TTG_ERR); + } else if (ctx->reg_int & L2C_TAD_INT_LFBTO) { + ctx->reg_ext_name = "TIMEOUT"; + ctx->reg_ext = readq(tad->regs + L2C_TAD_TIMEOUT); + } else if (ctx->reg_int & L2C_TAD_INT_DISOCI) { + ctx->reg_ext_name = "ERR"; + ctx->reg_ext = readq(tad->regs + L2C_TAD_ERR); + } + + writeq(ctx->reg_int, tad->regs + L2C_TAD_INT_W1C); + + tad->ring_head++; + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t thunderx_l2c_cbc_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_l2c *cbc = container_of(msix, struct thunderx_l2c, + msix_ent); + + unsigned long head = ring_pos(cbc->ring_head, ARRAY_SIZE(cbc->err_ctx)); + struct l2c_err_ctx *ctx = &cbc->err_ctx[head]; + + ctx->reg_int = readq(cbc->regs + L2C_CBC_INT_W1C); + + if (ctx->reg_int & L2C_CBC_INT_RSD) { + ctx->reg_ext_name = "RSDERR"; + ctx->reg_ext = readq(cbc->regs + L2C_CBC_RSDERR); + } else if (ctx->reg_int & L2C_CBC_INT_MIB) { + ctx->reg_ext_name = "MIBERR"; + ctx->reg_ext = readq(cbc->regs + L2C_CBC_MIBERR); + } else if (ctx->reg_int & L2C_CBC_INT_IODISOCI) { + ctx->reg_ext_name = "IODISOCIERR"; + ctx->reg_ext = readq(cbc->regs + L2C_CBC_IODISOCIERR); + } + + writeq(ctx->reg_int, cbc->regs + L2C_CBC_INT_W1C); + + cbc->ring_head++; + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t thunderx_l2c_mci_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_l2c *mci = container_of(msix, struct thunderx_l2c, + msix_ent); + + unsigned long head = ring_pos(mci->ring_head, ARRAY_SIZE(mci->err_ctx)); + struct l2c_err_ctx *ctx = &mci->err_ctx[head]; + + ctx->reg_int = readq(mci->regs + L2C_MCI_INT_W1C); + ctx->reg_ext = readq(mci->regs + L2C_MCI_ERR); + + writeq(ctx->reg_int, mci->regs + L2C_MCI_INT_W1C); + + ctx->reg_ext_name = "ERR"; + + mci->ring_head++; + + return IRQ_WAKE_THREAD; +} + +static irqreturn_t thunderx_l2c_threaded_isr(int irq, void *irq_id) +{ + struct msix_entry *msix = irq_id; + struct thunderx_l2c *l2c = container_of(msix, struct thunderx_l2c, + msix_ent); + + unsigned long tail = ring_pos(l2c->ring_tail, ARRAY_SIZE(l2c->err_ctx)); + struct l2c_err_ctx *ctx = &l2c->err_ctx[tail]; + irqreturn_t ret = IRQ_NONE; + + u64 mask_ue, mask_ce; + const struct error_descr *l2_errors; + char *reg_int_name; + + char *msg; + char *other; + + msg = kmalloc(OCX_MESSAGE_SIZE, GFP_KERNEL); + other = kmalloc(OCX_OTHER_SIZE, GFP_KERNEL); + + if (!msg || !other) + goto err_free; + + switch (l2c->pdev->device) { + case PCI_DEVICE_ID_THUNDER_L2C_TAD: + reg_int_name = "L2C_TAD_INT"; + mask_ue = L2C_TAD_INT_UE; + mask_ce = L2C_TAD_INT_CE; + l2_errors = l2_tad_errors; + break; + case PCI_DEVICE_ID_THUNDER_L2C_CBC: + reg_int_name = "L2C_CBC_INT"; + mask_ue = L2C_CBC_INT_UE; + mask_ce = L2C_CBC_INT_CE; + l2_errors = l2_cbc_errors; + break; + case PCI_DEVICE_ID_THUNDER_L2C_MCI: + reg_int_name = "L2C_MCI_INT"; + mask_ue = L2C_MCI_INT_VBFDBE; + mask_ce = L2C_MCI_INT_VBFSBE; + l2_errors = l2_mci_errors; + break; + default: + dev_err(&l2c->pdev->dev, "Unsupported device: %04x\n", + l2c->pdev->device); + return IRQ_NONE; + } + + while (CIRC_CNT(l2c->ring_head, l2c->ring_tail, + ARRAY_SIZE(l2c->err_ctx))) { + snprintf(msg, L2C_MESSAGE_SIZE, + "%s: %s: %016llx, %s: %016llx", + l2c->edac_dev->ctl_name, reg_int_name, ctx->reg_int, + ctx->reg_ext_name, ctx->reg_ext); + + decode_register(other, L2C_OTHER_SIZE, l2_errors, ctx->reg_int); + + strncat(msg, other, L2C_MESSAGE_SIZE); + + if (ctx->reg_int & mask_ue) + edac_device_handle_ue(l2c->edac_dev, 0, 0, msg); + else if (ctx->reg_int & mask_ce) + edac_device_handle_ce(l2c->edac_dev, 0, 0, msg); + + l2c->ring_tail++; + } + + return IRQ_HANDLED; + +err_free: + kfree(other); + kfree(msg); + + return ret; +} + +#define L2C_DEBUGFS_ATTR(_name, _reg) DEBUGFS_REG_ATTR(l2c, _name, _reg) + +L2C_DEBUGFS_ATTR(tad_int, L2C_TAD_INT_W1S); + +struct debugfs_entry *l2c_tad_dfs_ents[] = { + &debugfs_tad_int, +}; + +L2C_DEBUGFS_ATTR(cbc_int, L2C_CBC_INT_W1S); + +struct debugfs_entry *l2c_cbc_dfs_ents[] = { + &debugfs_cbc_int, +}; + +L2C_DEBUGFS_ATTR(mci_int, L2C_MCI_INT_W1S); + +struct debugfs_entry *l2c_mci_dfs_ents[] = { + &debugfs_mci_int, +}; + +static const struct pci_device_id thunderx_l2c_pci_tbl[] = { + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_L2C_TAD), }, + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_L2C_CBC), }, + { PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVICE_ID_THUNDER_L2C_MCI), }, + { 0, }, +}; + +static int thunderx_l2c_probe(struct pci_dev *pdev, + const struct pci_device_id *id) +{ + struct thunderx_l2c *l2c; + struct edac_device_ctl_info *edac_dev; + struct debugfs_entry **l2c_devattr; + size_t dfs_entries; + irqreturn_t (*thunderx_l2c_isr)(int, void *) = NULL; + char name[32]; + const char *fmt; + u64 reg_en_offs, reg_en_mask; + int idx; + int ret; + + ret = pcim_enable_device(pdev); + if (ret) { + dev_err(&pdev->dev, "Cannot enable PCI device: %d\n", ret); + return ret; + } + + ret = pcim_iomap_regions(pdev, BIT(0), "thunderx_l2c"); + if (ret) { + dev_err(&pdev->dev, "Cannot map PCI resources: %d\n", ret); + return ret; + } + + switch (pdev->device) { + case PCI_DEVICE_ID_THUNDER_L2C_TAD: + thunderx_l2c_isr = thunderx_l2c_tad_isr; + l2c_devattr = l2c_tad_dfs_ents; + dfs_entries = ARRAY_SIZE(l2c_tad_dfs_ents); + fmt = "L2C-TAD%d"; + reg_en_offs = L2C_TAD_INT_ENA_W1S; + reg_en_mask = L2C_TAD_INT_ENA_ALL; + break; + case PCI_DEVICE_ID_THUNDER_L2C_CBC: + thunderx_l2c_isr = thunderx_l2c_cbc_isr; + l2c_devattr = l2c_cbc_dfs_ents; + dfs_entries = ARRAY_SIZE(l2c_cbc_dfs_ents); + fmt = "L2C-CBC%d"; + reg_en_offs = L2C_CBC_INT_ENA_W1S; + reg_en_mask = L2C_CBC_INT_ENA_ALL; + break; + case PCI_DEVICE_ID_THUNDER_L2C_MCI: + thunderx_l2c_isr = thunderx_l2c_mci_isr; + l2c_devattr = l2c_mci_dfs_ents; + dfs_entries = ARRAY_SIZE(l2c_mci_dfs_ents); + fmt = "L2C-MCI%d"; + reg_en_offs = L2C_MCI_INT_ENA_W1S; + reg_en_mask = L2C_MCI_INT_ENA_ALL; + break; + default: + //Should never ever get here + dev_err(&pdev->dev, "Unsupported PCI device: %04x\n", + pdev->device); + return -EINVAL; + } + + idx = edac_device_alloc_index(); + snprintf(name, sizeof(name), fmt, idx); + + edac_dev = edac_device_alloc_ctl_info(sizeof(struct thunderx_l2c), + name, 1, "L2C", 1, 0, + NULL, 0, idx); + if (!edac_dev) { + dev_err(&pdev->dev, "Cannot allocate EDAC device\n"); + return -ENOMEM; + } + + l2c = edac_dev->pvt_info; + l2c->edac_dev = edac_dev; + + l2c->regs = pcim_iomap_table(pdev)[0]; + if (!l2c->regs) { + dev_err(&pdev->dev, "Cannot map PCI resources\n"); + ret = -ENODEV; + goto err_free; + } + + l2c->pdev = pdev; + + l2c->ring_head = 0; + l2c->ring_tail = 0; + + l2c->msix_ent.entry = 0; + l2c->msix_ent.vector = 0; + + ret = pci_enable_msix_exact(pdev, &l2c->msix_ent, 1); + if (ret) { + dev_err(&pdev->dev, "Cannot enable interrupt: %d\n", ret); + goto err_free; + } + + ret = devm_request_threaded_irq(&pdev->dev, l2c->msix_ent.vector, + thunderx_l2c_isr, + thunderx_l2c_threaded_isr, + 0, "[EDAC] ThunderX L2C", + &l2c->msix_ent); + if (ret) + goto err_free; + + edac_dev->dev = &pdev->dev; + edac_dev->dev_name = dev_name(&pdev->dev); + edac_dev->mod_name = "thunderx-l2c"; + edac_dev->ctl_name = "thunderx-l2c"; + + ret = edac_device_add_device(edac_dev); + if (ret) { + dev_err(&pdev->dev, "Cannot add EDAC device: %d\n", ret); + goto err_free; + } + + if (IS_ENABLED(CONFIG_EDAC_DEBUG)) { + l2c->debugfs = edac_debugfs_create_dir(pdev->dev.kobj.name); + + thunderx_create_debugfs_nodes(l2c->debugfs, l2c_devattr, + l2c, dfs_entries); + + if (ret != dfs_entries) { + dev_warn(&pdev->dev, "Error creating debugfs entries: %d%s\n", + ret, ret >= 0 ? " created" : ""); + } + } + + pci_set_drvdata(pdev, edac_dev); + + writeq(reg_en_mask, l2c->regs + reg_en_offs); + + return 0; + +err_free: + edac_device_free_ctl_info(edac_dev); + + return ret; +} + +static void thunderx_l2c_remove(struct pci_dev *pdev) +{ + struct edac_device_ctl_info *edac_dev = pci_get_drvdata(pdev); + struct thunderx_l2c *l2c = edac_dev->pvt_info; + + switch (pdev->device) { + case PCI_DEVICE_ID_THUNDER_L2C_TAD: + writeq(L2C_TAD_INT_ENA_ALL, l2c->regs + L2C_TAD_INT_ENA_W1C); + break; + case PCI_DEVICE_ID_THUNDER_L2C_CBC: + writeq(L2C_CBC_INT_ENA_ALL, l2c->regs + L2C_CBC_INT_ENA_W1C); + break; + case PCI_DEVICE_ID_THUNDER_L2C_MCI: + writeq(L2C_MCI_INT_ENA_ALL, l2c->regs + L2C_MCI_INT_ENA_W1C); + break; + } + + edac_debugfs_remove_recursive(l2c->debugfs); + + edac_device_del_device(&pdev->dev); + edac_device_free_ctl_info(edac_dev); +} + +MODULE_DEVICE_TABLE(pci, thunderx_l2c_pci_tbl); + +static struct pci_driver thunderx_l2c_driver = { + .name = "thunderx_l2c_edac", + .probe = thunderx_l2c_probe, + .remove = thunderx_l2c_remove, + .id_table = thunderx_l2c_pci_tbl, +}; + +static int __init thunderx_edac_init(void) +{ + int rc = 0; + + rc = pci_register_driver(&thunderx_lmc_driver); + if (rc) + return rc; + + rc = pci_register_driver(&thunderx_ocx_driver); + if (rc) + goto err_lmc; + + rc = pci_register_driver(&thunderx_l2c_driver); + if (rc) + goto err_ocx; + + return rc; +err_ocx: + pci_unregister_driver(&thunderx_ocx_driver); +err_lmc: + pci_unregister_driver(&thunderx_lmc_driver); + + return rc; +} + +static void __exit thunderx_edac_exit(void) +{ + pci_unregister_driver(&thunderx_l2c_driver); + pci_unregister_driver(&thunderx_ocx_driver); + pci_unregister_driver(&thunderx_lmc_driver); + +} + +module_init(thunderx_edac_init); +module_exit(thunderx_edac_exit); + +MODULE_LICENSE("GPL v2"); +MODULE_AUTHOR("Cavium, Inc."); +MODULE_DESCRIPTION("EDAC Driver for Cavium ThunderX"); diff --git a/include/linux/edac.h b/include/linux/edac.h index 5b6adf964248..8ae0f45fafd6 100644 --- a/include/linux/edac.h +++ b/include/linux/edac.h @@ -28,12 +28,10 @@ struct device; #define EDAC_OPSTATE_INT 2 extern int edac_op_state; -extern int edac_err_assert; -extern atomic_t edac_handlers; -extern int edac_handler_set(void); -extern void edac_atomic_assert_error(void); -extern struct bus_type *edac_get_sysfs_subsys(void); +struct bus_type *edac_get_sysfs_subsys(void); +int edac_get_report_status(void); +void edac_set_report_status(int new); enum { EDAC_REPORTING_ENABLED, @@ -41,28 +39,6 @@ enum { EDAC_REPORTING_FORCE }; -extern int edac_report_status; -#ifdef CONFIG_EDAC -static inline int get_edac_report_status(void) -{ - return edac_report_status; -} - -static inline void set_edac_report_status(int new) -{ - edac_report_status = new; -} -#else -static inline int get_edac_report_status(void) -{ - return EDAC_REPORTING_DISABLED; -} - -static inline void set_edac_report_status(int new) -{ -} -#endif - static inline void opstate_init(void) { switch (edac_op_state) { |