diff options
Diffstat (limited to 'drivers/edac')
-rw-r--r-- | drivers/edac/Kconfig | 23 | ||||
-rw-r--r-- | drivers/edac/Makefile | 4 | ||||
-rw-r--r-- | drivers/edac/amd64_edac.c | 17 | ||||
-rw-r--r-- | drivers/edac/amd64_edac.h | 5 | ||||
-rw-r--r-- | drivers/edac/amd64_edac_dbg.c | 207 | ||||
-rw-r--r-- | drivers/edac/edac_device_sysfs.c | 18 | ||||
-rw-r--r-- | drivers/edac/edac_mc_sysfs.c | 11 | ||||
-rw-r--r-- | drivers/edac/edac_mce_amd.c | 452 | ||||
-rw-r--r-- | drivers/edac/edac_module.c | 79 | ||||
-rw-r--r-- | drivers/edac/edac_module.h | 1 | ||||
-rw-r--r-- | drivers/edac/edac_pci_sysfs.c | 10 | ||||
-rw-r--r-- | drivers/edac/edac_stub.c | 51 | ||||
-rw-r--r-- | drivers/edac/i7300_edac.c | 1247 | ||||
-rw-r--r-- | drivers/edac/i7core_edac.c | 1 | ||||
-rw-r--r-- | drivers/edac/i82443bxgx_edac.c | 2 | ||||
-rw-r--r-- | drivers/edac/mce_amd.c | 680 | ||||
-rw-r--r-- | drivers/edac/mce_amd.h (renamed from drivers/edac/edac_mce_amd.h) | 59 | ||||
-rw-r--r-- | drivers/edac/mce_amd_inj.c | 171 |
18 files changed, 2275 insertions, 763 deletions
diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig index 70bb350de99..f436a2fa9f3 100644 --- a/drivers/edac/Kconfig +++ b/drivers/edac/Kconfig @@ -39,7 +39,7 @@ config EDAC_DEBUG there're four debug levels (x=0,1,2,3 from low to high). Usually you should select 'N'. - config EDAC_DECODE_MCE +config EDAC_DECODE_MCE tristate "Decode MCEs in human-readable form (only on AMD for now)" depends on CPU_SUP_AMD && X86_MCE default y @@ -51,6 +51,16 @@ config EDAC_DEBUG which occur really early upon boot, before the module infrastructure has been initialized. +config EDAC_MCE_INJ + tristate "Simple MCE injection interface over /sysfs" + depends on EDAC_DECODE_MCE + default n + help + This is a simple interface to inject MCEs over /sysfs and test + the MCE decoding code in EDAC. + + This is currently AMD-only. + config EDAC_MM_EDAC tristate "Main Memory EDAC (Error Detection And Correction) reporting" help @@ -66,13 +76,13 @@ config EDAC_MCE config EDAC_AMD64 tristate "AMD64 (Opteron, Athlon64) K8, F10h, F11h" - depends on EDAC_MM_EDAC && K8_NB && X86_64 && PCI && EDAC_DECODE_MCE + depends on EDAC_MM_EDAC && AMD_NB && X86_64 && PCI && EDAC_DECODE_MCE help Support for error detection and correction on the AMD 64 Families of Memory Controllers (K8, F10h and F11h) config EDAC_AMD64_ERROR_INJECTION - bool "Sysfs Error Injection facilities" + bool "Sysfs HW Error injection facilities" depends on EDAC_AMD64 help Recent Opterons (Family 10h and later) provide for Memory Error @@ -199,6 +209,13 @@ config EDAC_I5100 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 + help + Support for error detection and correction the Intel + Clarksboro MCH (Intel 7300 chipset). + config EDAC_MPC85XX tristate "Freescale MPC83xx / MPC85xx" depends on EDAC_MM_EDAC && FSL_SOC && (PPC_83xx || PPC_85xx) diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile index ca6b1bb24cc..b3781399b38 100644 --- a/drivers/edac/Makefile +++ b/drivers/edac/Makefile @@ -17,6 +17,9 @@ ifdef CONFIG_PCI edac_core-objs += edac_pci.o edac_pci_sysfs.o endif +obj-$(CONFIG_EDAC_MCE_INJ) += mce_amd_inj.o + +edac_mce_amd-objs := mce_amd.o obj-$(CONFIG_EDAC_DECODE_MCE) += edac_mce_amd.o obj-$(CONFIG_EDAC_AMD76X) += amd76x_edac.o @@ -24,6 +27,7 @@ obj-$(CONFIG_EDAC_CPC925) += cpc925_edac.o obj-$(CONFIG_EDAC_I5000) += i5000_edac.o obj-$(CONFIG_EDAC_I5100) += i5100_edac.o obj-$(CONFIG_EDAC_I5400) += i5400_edac.o +obj-$(CONFIG_EDAC_I7300) += i7300_edac.o obj-$(CONFIG_EDAC_I7CORE) += i7core_edac.o obj-$(CONFIG_EDAC_E7XXX) += e7xxx_edac.o obj-$(CONFIG_EDAC_E752X) += e752x_edac.o diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c index e7d5d6b5dcf..8521401bbd7 100644 --- a/drivers/edac/amd64_edac.c +++ b/drivers/edac/amd64_edac.c @@ -1,5 +1,5 @@ #include "amd64_edac.h" -#include <asm/k8.h> +#include <asm/amd_nb.h> static struct edac_pci_ctl_info *amd64_ctl_pci; @@ -2073,11 +2073,18 @@ static inline void __amd64_decode_bus_error(struct mem_ctl_info *mci, amd64_handle_ue(mci, info); } -void amd64_decode_bus_error(int node_id, struct err_regs *regs) +void amd64_decode_bus_error(int node_id, struct mce *m, u32 nbcfg) { struct mem_ctl_info *mci = mci_lookup[node_id]; + struct err_regs regs; - __amd64_decode_bus_error(mci, regs); + regs.nbsl = (u32) m->status; + regs.nbsh = (u32)(m->status >> 32); + regs.nbeal = (u32) m->addr; + regs.nbeah = (u32)(m->addr >> 32); + regs.nbcfg = nbcfg; + + __amd64_decode_bus_error(mci, ®s); /* * Check the UE bit of the NB status high register, if set generate some @@ -2086,7 +2093,7 @@ void amd64_decode_bus_error(int node_id, struct err_regs *regs) * * FIXME: this should go somewhere else, if at all. */ - if (regs->nbsh & K8_NBSH_UC_ERR && !report_gart_errors) + if (regs.nbsh & K8_NBSH_UC_ERR && !report_gart_errors) edac_mc_handle_ue_no_info(mci, "UE bit is set"); } @@ -2927,7 +2934,7 @@ static int __init amd64_edac_init(void) * to finish initialization of the MC instances. */ err = -ENODEV; - for (nb = 0; nb < num_k8_northbridges; nb++) { + for (nb = 0; nb < k8_northbridges.num; nb++) { if (!pvt_lookup[nb]) continue; diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h index 613b9381e71..044aee4f944 100644 --- a/drivers/edac/amd64_edac.h +++ b/drivers/edac/amd64_edac.h @@ -72,7 +72,7 @@ #include <linux/edac.h> #include <asm/msr.h> #include "edac_core.h" -#include "edac_mce_amd.h" +#include "mce_amd.h" #define amd64_printk(level, fmt, arg...) \ edac_printk(level, "amd64", fmt, ##arg) @@ -482,11 +482,10 @@ extern const char *rrrr_msgs[16]; extern const char *to_msgs[2]; extern const char *pp_msgs[4]; extern const char *ii_msgs[4]; -extern const char *ext_msgs[32]; extern const char *htlink_msgs[8]; #ifdef CONFIG_EDAC_DEBUG -#define NUM_DBG_ATTRS 9 +#define NUM_DBG_ATTRS 5 #else #define NUM_DBG_ATTRS 0 #endif diff --git a/drivers/edac/amd64_edac_dbg.c b/drivers/edac/amd64_edac_dbg.c index 59cf2cf6e11..e3562288f4c 100644 --- a/drivers/edac/amd64_edac_dbg.c +++ b/drivers/edac/amd64_edac_dbg.c @@ -1,167 +1,16 @@ #include "amd64_edac.h" -/* - * accept a hex value and store it into the virtual error register file, field: - * nbeal and nbeah. Assume virtual error values have already been set for: NBSL, - * NBSH and NBCFG. Then proceed to map the error values to a MC, CSROW and - * CHANNEL - */ -static ssize_t amd64_nbea_store(struct mem_ctl_info *mci, const char *data, - size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long long value; - int ret = 0; - - ret = strict_strtoull(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBEA= 0x%llx\n", value); - - /* place the value into the virtual error packet */ - pvt->ctl_error_info.nbeal = (u32) value; - value >>= 32; - pvt->ctl_error_info.nbeah = (u32) value; - - /* Process the Mapping request */ - /* TODO: Add race prevention */ - amd_decode_nb_mce(pvt->mc_node_id, &pvt->ctl_error_info, 1); - - return count; - } - return ret; -} - -/* display back what the last NBEA (MCA NB Address (MC4_ADDR)) was written */ -static ssize_t amd64_nbea_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u64 value; - - value = pvt->ctl_error_info.nbeah; - value <<= 32; - value |= pvt->ctl_error_info.nbeal; - - return sprintf(data, "%llx\n", value); -} - -/* store the NBSL (MCA NB Status Low (MC4_STATUS)) value user desires */ -static ssize_t amd64_nbsl_store(struct mem_ctl_info *mci, const char *data, - size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long value; - int ret = 0; - - ret = strict_strtoul(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBSL= 0x%lx\n", value); - - pvt->ctl_error_info.nbsl = (u32) value; - - return count; - } - return ret; -} - -/* display back what the last NBSL value written */ -static ssize_t amd64_nbsl_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u32 value; - - value = pvt->ctl_error_info.nbsl; - - return sprintf(data, "%x\n", value); -} - -/* store the NBSH (MCA NB Status High) value user desires */ -static ssize_t amd64_nbsh_store(struct mem_ctl_info *mci, const char *data, - size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long value; - int ret = 0; - - ret = strict_strtoul(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBSH= 0x%lx\n", value); - - pvt->ctl_error_info.nbsh = (u32) value; - - return count; - } - return ret; -} - -/* display back what the last NBSH value written */ -static ssize_t amd64_nbsh_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - u32 value; - - value = pvt->ctl_error_info.nbsh; - - return sprintf(data, "%x\n", value); +#define EDAC_DCT_ATTR_SHOW(reg) \ +static ssize_t amd64_##reg##_show(struct mem_ctl_info *mci, char *data) \ +{ \ + struct amd64_pvt *pvt = mci->pvt_info; \ + return sprintf(data, "0x%016llx\n", (u64)pvt->reg); \ } -/* accept and store the NBCFG (MCA NB Configuration) value user desires */ -static ssize_t amd64_nbcfg_store(struct mem_ctl_info *mci, - const char *data, size_t count) -{ - struct amd64_pvt *pvt = mci->pvt_info; - unsigned long value; - int ret = 0; - - ret = strict_strtoul(data, 16, &value); - if (ret != -EINVAL) { - debugf0("received NBCFG= 0x%lx\n", value); - - pvt->ctl_error_info.nbcfg = (u32) value; - - return count; - } - return ret; -} - -/* various show routines for the controls of a MCI */ -static ssize_t amd64_nbcfg_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%x\n", pvt->ctl_error_info.nbcfg); -} - - -static ssize_t amd64_dhar_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%x\n", pvt->dhar); -} - - -static ssize_t amd64_dbam_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%x\n", pvt->dbam0); -} - - -static ssize_t amd64_topmem_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%llx\n", pvt->top_mem); -} - - -static ssize_t amd64_topmem2_show(struct mem_ctl_info *mci, char *data) -{ - struct amd64_pvt *pvt = mci->pvt_info; - - return sprintf(data, "%llx\n", pvt->top_mem2); -} +EDAC_DCT_ATTR_SHOW(dhar); +EDAC_DCT_ATTR_SHOW(dbam0); +EDAC_DCT_ATTR_SHOW(top_mem); +EDAC_DCT_ATTR_SHOW(top_mem2); static ssize_t amd64_hole_show(struct mem_ctl_info *mci, char *data) { @@ -182,38 +31,6 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { { .attr = { - .name = "nbea_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbea_show, - .store = amd64_nbea_store, - }, - { - .attr = { - .name = "nbsl_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbsl_show, - .store = amd64_nbsl_store, - }, - { - .attr = { - .name = "nbsh_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbsh_show, - .store = amd64_nbsh_store, - }, - { - .attr = { - .name = "nbcfg_ctl", - .mode = (S_IRUGO | S_IWUSR) - }, - .show = amd64_nbcfg_show, - .store = amd64_nbcfg_store, - }, - { - .attr = { .name = "dhar", .mode = (S_IRUGO) }, @@ -225,7 +42,7 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { .name = "dbam", .mode = (S_IRUGO) }, - .show = amd64_dbam_show, + .show = amd64_dbam0_show, .store = NULL, }, { @@ -233,7 +50,7 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { .name = "topmem", .mode = (S_IRUGO) }, - .show = amd64_topmem_show, + .show = amd64_top_mem_show, .store = NULL, }, { @@ -241,7 +58,7 @@ struct mcidev_sysfs_attribute amd64_dbg_attrs[] = { .name = "topmem2", .mode = (S_IRUGO) }, - .show = amd64_topmem2_show, + .show = amd64_top_mem2_show, .store = NULL, }, { diff --git a/drivers/edac/edac_device_sysfs.c b/drivers/edac/edac_device_sysfs.c index 070968178a2..400de071cab 100644 --- a/drivers/edac/edac_device_sysfs.c +++ b/drivers/edac/edac_device_sysfs.c @@ -1,7 +1,7 @@ /* * file for managing the edac_device class of devices for EDAC * - * (C) 2007 SoftwareBitMaker (http://www.softwarebitmaker.com) + * (C) 2007 SoftwareBitMaker * * This file may be distributed under the terms of the * GNU General Public License. @@ -13,6 +13,7 @@ #include <linux/ctype.h> #include <linux/module.h> #include <linux/slab.h> +#include <linux/edac.h> #include "edac_core.h" #include "edac_module.h" @@ -235,7 +236,7 @@ int edac_device_register_sysfs_main_kobj(struct edac_device_ctl_info *edac_dev) debugf1("%s()\n", __func__); /* get the /sys/devices/system/edac reference */ - edac_class = edac_get_edac_class(); + edac_class = edac_get_sysfs_class(); if (edac_class == NULL) { debugf1("%s() no edac_class error\n", __func__); err = -ENODEV; @@ -255,7 +256,7 @@ int edac_device_register_sysfs_main_kobj(struct edac_device_ctl_info *edac_dev) if (!try_module_get(edac_dev->owner)) { err = -ENODEV; - goto err_out; + goto err_mod_get; } /* register */ @@ -282,6 +283,9 @@ int edac_device_register_sysfs_main_kobj(struct edac_device_ctl_info *edac_dev) err_kobj_reg: module_put(edac_dev->owner); +err_mod_get: + edac_put_sysfs_class(); + err_out: return err; } @@ -290,12 +294,11 @@ err_out: * edac_device_unregister_sysfs_main_kobj: * the '..../edac/<name>' kobject */ -void edac_device_unregister_sysfs_main_kobj( - struct edac_device_ctl_info *edac_dev) +void edac_device_unregister_sysfs_main_kobj(struct edac_device_ctl_info *dev) { debugf0("%s()\n", __func__); debugf4("%s() name of kobject is: %s\n", - __func__, kobject_name(&edac_dev->kobj)); + __func__, kobject_name(&dev->kobj)); /* * Unregister the edac device's kobject and @@ -304,7 +307,8 @@ void edac_device_unregister_sysfs_main_kobj( * a) module_put() this module * b) 'kfree' the memory */ - kobject_put(&edac_dev->kobj); + kobject_put(&dev->kobj); + edac_put_sysfs_class(); } /* edac_dev -> instance information */ diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c index 8aad94d10c0..a4135860149 100644 --- a/drivers/edac/edac_mc_sysfs.c +++ b/drivers/edac/edac_mc_sysfs.c @@ -11,6 +11,7 @@ #include <linux/ctype.h> #include <linux/slab.h> +#include <linux/edac.h> #include <linux/bug.h> #include "edac_core.h" @@ -1011,13 +1012,13 @@ void edac_remove_sysfs_mci_device(struct mem_ctl_info *mci) */ int edac_sysfs_setup_mc_kset(void) { - int err = 0; + int err = -EINVAL; struct sysdev_class *edac_class; debugf1("%s()\n", __func__); /* get the /sys/devices/system/edac class reference */ - edac_class = edac_get_edac_class(); + edac_class = edac_get_sysfs_class(); if (edac_class == NULL) { debugf1("%s() no edac_class error=%d\n", __func__, err); goto fail_out; @@ -1028,15 +1029,16 @@ int edac_sysfs_setup_mc_kset(void) if (!mc_kset) { err = -ENOMEM; debugf1("%s() Failed to register '.../edac/mc'\n", __func__); - goto fail_out; + goto fail_kset; } debugf1("%s() Registered '.../edac/mc' kobject\n", __func__); return 0; +fail_kset: + edac_put_sysfs_class(); - /* error unwind stack */ fail_out: return err; } @@ -1049,5 +1051,6 @@ fail_out: void edac_sysfs_teardown_mc_kset(void) { kset_unregister(mc_kset); + edac_put_sysfs_class(); } diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c deleted file mode 100644 index 9014df6f605..00000000000 --- a/drivers/edac/edac_mce_amd.c +++ /dev/null @@ -1,452 +0,0 @@ -#include <linux/module.h> -#include "edac_mce_amd.h" - -static bool report_gart_errors; -static void (*nb_bus_decoder)(int node_id, struct err_regs *regs); - -void amd_report_gart_errors(bool v) -{ - report_gart_errors = v; -} -EXPORT_SYMBOL_GPL(amd_report_gart_errors); - -void amd_register_ecc_decoder(void (*f)(int, struct err_regs *)) -{ - nb_bus_decoder = f; -} -EXPORT_SYMBOL_GPL(amd_register_ecc_decoder); - -void amd_unregister_ecc_decoder(void (*f)(int, struct err_regs *)) -{ - if (nb_bus_decoder) { - WARN_ON(nb_bus_decoder != f); - - nb_bus_decoder = NULL; - } -} -EXPORT_SYMBOL_GPL(amd_unregister_ecc_decoder); - -/* - * string representation for the different MCA reported error types, see F3x48 - * or MSR0000_0411. - */ -const char *tt_msgs[] = { /* transaction type */ - "instruction", - "data", - "generic", - "reserved" -}; -EXPORT_SYMBOL_GPL(tt_msgs); - -const char *ll_msgs[] = { /* cache level */ - "L0", - "L1", - "L2", - "L3/generic" -}; -EXPORT_SYMBOL_GPL(ll_msgs); - -const char *rrrr_msgs[] = { - "generic", - "generic read", - "generic write", - "data read", - "data write", - "inst fetch", - "prefetch", - "evict", - "snoop", - "reserved RRRR= 9", - "reserved RRRR= 10", - "reserved RRRR= 11", - "reserved RRRR= 12", - "reserved RRRR= 13", - "reserved RRRR= 14", - "reserved RRRR= 15" -}; -EXPORT_SYMBOL_GPL(rrrr_msgs); - -const char *pp_msgs[] = { /* participating processor */ - "local node originated (SRC)", - "local node responded to request (RES)", - "local node observed as 3rd party (OBS)", - "generic" -}; -EXPORT_SYMBOL_GPL(pp_msgs); - -const char *to_msgs[] = { - "no timeout", - "timed out" -}; -EXPORT_SYMBOL_GPL(to_msgs); - -const char *ii_msgs[] = { /* memory or i/o */ - "mem access", - "reserved", - "i/o access", - "generic" -}; -EXPORT_SYMBOL_GPL(ii_msgs); - -/* - * Map the 4 or 5 (family-specific) bits of Extended Error code to the - * string table. - */ -const char *ext_msgs[] = { - "K8 ECC error", /* 0_0000b */ - "CRC error on link", /* 0_0001b */ - "Sync error packets on link", /* 0_0010b */ - "Master Abort during link operation", /* 0_0011b */ - "Target Abort during link operation", /* 0_0100b */ - "Invalid GART PTE entry during table walk", /* 0_0101b */ - "Unsupported atomic RMW command received", /* 0_0110b */ - "WDT error: NB transaction timeout", /* 0_0111b */ - "ECC/ChipKill ECC error", /* 0_1000b */ - "SVM DEV Error", /* 0_1001b */ - "Link Data error", /* 0_1010b */ - "Link/L3/Probe Filter Protocol error", /* 0_1011b */ - "NB Internal Arrays Parity error", /* 0_1100b */ - "DRAM Address/Control Parity error", /* 0_1101b */ - "Link Transmission error", /* 0_1110b */ - "GART/DEV Table Walk Data error" /* 0_1111b */ - "Res 0x100 error", /* 1_0000b */ - "Res 0x101 error", /* 1_0001b */ - "Res 0x102 error", /* 1_0010b */ - "Res 0x103 error", /* 1_0011b */ - "Res 0x104 error", /* 1_0100b */ - "Res 0x105 error", /* 1_0101b */ - "Res 0x106 error", /* 1_0110b */ - "Res 0x107 error", /* 1_0111b */ - "Res 0x108 error", /* 1_1000b */ - "Res 0x109 error", /* 1_1001b */ - "Res 0x10A error", /* 1_1010b */ - "Res 0x10B error", /* 1_1011b */ - "ECC error in L3 Cache Data", /* 1_1100b */ - "L3 Cache Tag error", /* 1_1101b */ - "L3 Cache LRU Parity error", /* 1_1110b */ - "Probe Filter error" /* 1_1111b */ -}; -EXPORT_SYMBOL_GPL(ext_msgs); - -static void amd_decode_dc_mce(u64 mc0_status) -{ - u32 ec = mc0_status & 0xffff; - u32 xec = (mc0_status >> 16) & 0xf; - - pr_emerg("Data Cache Error"); - - if (xec == 1 && TLB_ERROR(ec)) - pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); - else if (xec == 0) { - if (mc0_status & (1ULL << 40)) - pr_cont(" during Data Scrub.\n"); - else if (TLB_ERROR(ec)) - pr_cont(": %s TLB parity error.\n", LL_MSG(ec)); - else if (MEM_ERROR(ec)) { - u8 ll = ec & 0x3; - u8 tt = (ec >> 2) & 0x3; - u8 rrrr = (ec >> 4) & 0xf; - - /* see F10h BKDG (31116), Table 92. */ - if (ll == 0x1) { - if (tt != 0x1) - goto wrong_dc_mce; - - pr_cont(": Data/Tag %s error.\n", RRRR_MSG(ec)); - - } else if (ll == 0x2 && rrrr == 0x3) - pr_cont(" during L1 linefill from L2.\n"); - else - goto wrong_dc_mce; - } else if (BUS_ERROR(ec) && boot_cpu_data.x86 == 0xf) - pr_cont(" during system linefill.\n"); - else - goto wrong_dc_mce; - } else - goto wrong_dc_mce; - - return; - -wrong_dc_mce: - pr_warning("Corrupted DC MCE info?\n"); -} - -static void amd_decode_ic_mce(u64 mc1_status) -{ - u32 ec = mc1_status & 0xffff; - u32 xec = (mc1_status >> 16) & 0xf; - - pr_emerg("Instruction Cache Error"); - - if (xec == 1 && TLB_ERROR(ec)) - pr_cont(": %s TLB multimatch.\n", LL_MSG(ec)); - else if (xec == 0) { - if (TLB_ERROR(ec)) - pr_cont(": %s TLB Parity error.\n", LL_MSG(ec)); - else if (BUS_ERROR(ec)) { - if (boot_cpu_data.x86 == 0xf && - (mc1_status & (1ULL << 58))) - pr_cont(" during system linefill.\n"); - else - pr_cont(" during attempted NB data read.\n"); - } else if (MEM_ERROR(ec)) { - u8 ll = ec & 0x3; - u8 rrrr = (ec >> 4) & 0xf; - - if (ll == 0x2) - pr_cont(" during a linefill from L2.\n"); - else if (ll == 0x1) { - - switch (rrrr) { - case 0x5: - pr_cont(": Parity error during " - "data load.\n"); - break; - - case 0x7: - pr_cont(": Copyback Parity/Victim" - " error.\n"); - break; - - case 0x8: - pr_cont(": Tag Snoop error.\n"); - break; - - default: - goto wrong_ic_mce; - break; - } - } - } else - goto wrong_ic_mce; - } else - goto wrong_ic_mce; - - return; - -wrong_ic_mce: - pr_warning("Corrupted IC MCE info?\n"); -} - -static void amd_decode_bu_mce(u64 mc2_status) -{ - u32 ec = mc2_status & 0xffff; - u32 xec = (mc2_status >> 16) & 0xf; - - pr_emerg("Bus Unit Error"); - - if (xec == 0x1) - pr_cont(" in the write data buffers.\n"); - else if (xec == 0x3) - pr_cont(" in the victim data buffers.\n"); - else if (xec == 0x2 && MEM_ERROR(ec)) - pr_cont(": %s error in the L2 cache tags.\n", RRRR_MSG(ec)); - else if (xec == 0x0) { - if (TLB_ERROR(ec)) - pr_cont(": %s error in a Page Descriptor Cache or " - "Guest TLB.\n", TT_MSG(ec)); - else if (BUS_ERROR(ec)) - pr_cont(": %s/ECC error in data read from NB: %s.\n", - RRRR_MSG(ec), PP_MSG(ec)); - else if (MEM_ERROR(ec)) { - u8 rrrr = (ec >> 4) & 0xf; - - if (rrrr >= 0x7) - pr_cont(": %s error during data copyback.\n", - RRRR_MSG(ec)); - else if (rrrr <= 0x1) - pr_cont(": %s parity/ECC error during data " - "access from L2.\n", RRRR_MSG(ec)); - else - goto wrong_bu_mce; - } else - goto wrong_bu_mce; - } else - goto wrong_bu_mce; - - return; - -wrong_bu_mce: - pr_warning("Corrupted BU MCE info?\n"); -} - -static void amd_decode_ls_mce(u64 mc3_status) -{ - u32 ec = mc3_status & 0xffff; - u32 xec = (mc3_status >> 16) & 0xf; - - pr_emerg("Load Store Error"); - - if (xec == 0x0) { - u8 rrrr = (ec >> 4) & 0xf; - - if (!BUS_ERROR(ec) || (rrrr != 0x3 && rrrr != 0x4)) - goto wrong_ls_mce; - - pr_cont(" during %s.\n", RRRR_MSG(ec)); - } - return; - -wrong_ls_mce: - pr_warning("Corrupted LS MCE info?\n"); -} - -void amd_decode_nb_mce(int node_id, struct err_regs *regs, int handle_errors) -{ - u32 ec = ERROR_CODE(regs->nbsl); - - if (!handle_errors) - return; - - /* - * GART TLB error reporting is disabled by default. Bail out early. - */ - if (TLB_ERROR(ec) && !report_gart_errors) - return; - - pr_emerg("Northbridge Error, node %d", node_id); - - /* - * F10h, revD can disable ErrCpu[3:0] so check that first and also the - * value encoding has changed so interpret those differently - */ - if ((boot_cpu_data.x86 == 0x10) && - (boot_cpu_data.x86_model > 7)) { - if (regs->nbsh & K8_NBSH_ERR_CPU_VAL) - pr_cont(", core: %u\n", (u8)(regs->nbsh & 0xf)); - } else { - u8 assoc_cpus = regs->nbsh & 0xf; - - if (assoc_cpus > 0) - pr_cont(", core: %d", fls(assoc_cpus) - 1); - - pr_cont("\n"); - } - - pr_emerg("%s.\n", EXT_ERR_MSG(regs->nbsl)); - - if (BUS_ERROR(ec) && nb_bus_decoder) - nb_bus_decoder(node_id, regs); -} -EXPORT_SYMBOL_GPL(amd_decode_nb_mce); - -static void amd_decode_fr_mce(u64 mc5_status) -{ - /* we have only one error signature so match all fields at once. */ - if ((mc5_status & 0xffff) == 0x0f0f) - pr_emerg(" FR Error: CPU Watchdog timer expire.\n"); - else - pr_warning("Corrupted FR MCE info?\n"); -} - -static inline void amd_decode_err_code(unsigned int ec) -{ - if (TLB_ERROR(ec)) { - pr_emerg("Transaction: %s, Cache Level %s\n", - TT_MSG(ec), LL_MSG(ec)); - } else if (MEM_ERROR(ec)) { - pr_emerg("Transaction: %s, Type: %s, Cache Level: %s", - RRRR_MSG(ec), TT_MSG(ec), LL_MSG(ec)); - } else if (BUS_ERROR(ec)) { - pr_emerg("Transaction type: %s(%s), %s, Cache Level: %s, " - "Participating Processor: %s\n", - RRRR_MSG(ec), II_MSG(ec), TO_MSG(ec), LL_MSG(ec), - PP_MSG(ec)); - } else - pr_warning("Huh? Unknown MCE error 0x%x\n", ec); -} - -static int amd_decode_mce(struct notifier_block *nb, unsigned long val, - void *data) -{ - struct mce *m = (struct mce *)data; - struct err_regs regs; - int node, ecc; - - pr_emerg("MC%d_STATUS: ", m->bank); - - pr_cont("%sorrected error, other errors lost: %s, " - "CPU context corrupt: %s", - ((m->status & MCI_STATUS_UC) ? "Unc" : "C"), - ((m->status & MCI_STATUS_OVER) ? "yes" : "no"), - ((m->status & MCI_STATUS_PCC) ? "yes" : "no")); - - /* do the two bits[14:13] together */ - ecc = (m->status >> 45) & 0x3; - if (ecc) - pr_cont(", %sECC Error", ((ecc == 2) ? "C" : "U")); - - pr_cont("\n"); - - switch (m->bank) { - case 0: - amd_decode_dc_mce(m->status); - break; - - case 1: - amd_decode_ic_mce(m->status); - break; - - case 2: - amd_decode_bu_mce(m->status); - break; - - case 3: - amd_decode_ls_mce(m->status); - break; - - case 4: - regs.nbsl = (u32) m->status; - regs.nbsh = (u32)(m->status >> 32); - regs.nbeal = (u32) m->addr; - regs.nbeah = (u32)(m->addr >> 32); - node = amd_get_nb_id(m->extcpu); - - amd_decode_nb_mce(node, ®s, 1); - break; - - case 5: - amd_decode_fr_mce(m->status); - break; - - default: - break; - } - - amd_decode_err_code(m->status & 0xffff); - - return NOTIFY_STOP; -} - -static struct notifier_block amd_mce_dec_nb = { - .notifier_call = amd_decode_mce, -}; - -static int __init mce_amd_init(void) -{ - /* - * We can decode MCEs for K8, F10h and F11h CPUs: - */ - if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD) - return 0; - - if (boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x11) - return 0; - - atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb); - - return 0; -} -early_initcall(mce_amd_init); - -#ifdef MODULE -static void __exit mce_amd_exit(void) -{ - atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb); -} - -MODULE_DESCRIPTION("AMD MCE decoder"); -MODULE_ALIAS("edac-mce-amd"); -MODULE_LICENSE("GPL"); -module_exit(mce_amd_exit); -#endif diff --git a/drivers/edac/edac_module.c b/drivers/edac/edac_module.c index 7e1374afd96..be4b075c309 100644 --- a/drivers/edac/edac_module.c +++ b/drivers/edac/edac_module.c @@ -27,15 +27,6 @@ EXPORT_SYMBOL_GPL(edac_debug_level); struct workqueue_struct *edac_workqueue; /* - * sysfs object: /sys/devices/system/edac - * need to export to other files in this modules - */ -static struct sysdev_class edac_class = { - .name = "edac", -}; -static int edac_class_valid; - -/* * edac_op_state_to_string() */ char *edac_op_state_to_string(int opstate) @@ -55,60 +46,6 @@ char *edac_op_state_to_string(int opstate) } /* - * edac_get_edac_class() - * - * return pointer to the edac class of 'edac' - */ -struct sysdev_class *edac_get_edac_class(void) -{ - struct sysdev_class *classptr = NULL; - - if (edac_class_valid) - classptr = &edac_class; - - return classptr; -} - -/* - * edac_register_sysfs_edac_name() - * - * register the 'edac' into /sys/devices/system - * - * return: - * 0 success - * !0 error - */ -static int edac_register_sysfs_edac_name(void) -{ - int err; - - /* create the /sys/devices/system/edac directory */ - err = sysdev_class_register(&edac_class); - - if (err) { - debugf1("%s() error=%d\n", __func__, err); - return err; - } - - edac_class_valid = 1; - return 0; -} - -/* - * sysdev_class_unregister() - * - * unregister the 'edac' from /sys/devices/system - */ -static void edac_unregister_sysfs_edac_name(void) -{ - /* only if currently registered, then unregister it */ - if (edac_class_valid) - sysdev_class_unregister(&edac_class); - - edac_class_valid = 0; -} - -/* * edac_workqueue_setup < |