diff options
Diffstat (limited to 'drivers/edac')
34 files changed, 1208 insertions, 547 deletions
diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c index 3c9e4e98c65..f8bf00010d4 100644 --- a/drivers/edac/amd64_edac.c +++ b/drivers/edac/amd64_edac.c @@ -1,7 +1,7 @@ #include "amd64_edac.h" #include <asm/amd_nb.h> -static struct edac_pci_ctl_info *amd64_ctl_pci; +static struct edac_pci_ctl_info *pci_ctl; static int report_gart_errors; module_param(report_gart_errors, int, 0644); @@ -162,7 +162,7 @@ static int f15_read_dct_pci_cfg(struct amd64_pvt *pvt, int addr, u32 *val, * scan the scrub rate mapping table for a close or matching bandwidth value to * issue. If requested is too big, then use last maximum value found. */ -static int __amd64_set_scrub_rate(struct pci_dev *ctl, u32 new_bw, u32 min_rate) +static int __set_scrub_rate(struct pci_dev *ctl, u32 new_bw, u32 min_rate) { u32 scrubval; int i; @@ -198,7 +198,7 @@ static int __amd64_set_scrub_rate(struct pci_dev *ctl, u32 new_bw, u32 min_rate) return 0; } -static int amd64_set_scrub_rate(struct mem_ctl_info *mci, u32 bw) +static int set_scrub_rate(struct mem_ctl_info *mci, u32 bw) { struct amd64_pvt *pvt = mci->pvt_info; u32 min_scrubrate = 0x5; @@ -210,10 +210,10 @@ static int amd64_set_scrub_rate(struct mem_ctl_info *mci, u32 bw) if (pvt->fam == 0x15 && pvt->model < 0x10) f15h_select_dct(pvt, 0); - return __amd64_set_scrub_rate(pvt->F3, bw, min_scrubrate); + return __set_scrub_rate(pvt->F3, bw, min_scrubrate); } -static int amd64_get_scrub_rate(struct mem_ctl_info *mci) +static int get_scrub_rate(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; u32 scrubval = 0; @@ -240,8 +240,7 @@ static int amd64_get_scrub_rate(struct mem_ctl_info *mci) * returns true if the SysAddr given by sys_addr matches the * DRAM base/limit associated with node_id */ -static bool amd64_base_limit_match(struct amd64_pvt *pvt, u64 sys_addr, - u8 nid) +static bool base_limit_match(struct amd64_pvt *pvt, u64 sys_addr, u8 nid) { u64 addr; @@ -285,7 +284,7 @@ static struct mem_ctl_info *find_mc_by_sys_addr(struct mem_ctl_info *mci, if (intlv_en == 0) { for (node_id = 0; node_id < DRAM_RANGES; node_id++) { - if (amd64_base_limit_match(pvt, sys_addr, node_id)) + if (base_limit_match(pvt, sys_addr, node_id)) goto found; } goto err_no_match; @@ -309,7 +308,7 @@ static struct mem_ctl_info *find_mc_by_sys_addr(struct mem_ctl_info *mci, } /* sanity test for sys_addr */ - if (unlikely(!amd64_base_limit_match(pvt, sys_addr, node_id))) { + if (unlikely(!base_limit_match(pvt, sys_addr, node_id))) { amd64_warn("%s: sys_addr 0x%llx falls outside base/limit address" "range for node %d with node interleaving enabled.\n", __func__, sys_addr, node_id); @@ -339,8 +338,8 @@ static void get_cs_base_and_mask(struct amd64_pvt *pvt, int csrow, u8 dct, if (pvt->fam == 0xf && pvt->ext_model < K8_REV_F) { csbase = pvt->csels[dct].csbases[csrow]; csmask = pvt->csels[dct].csmasks[csrow]; - base_bits = GENMASK(21, 31) | GENMASK(9, 15); - mask_bits = GENMASK(21, 29) | GENMASK(9, 15); + base_bits = GENMASK_ULL(31, 21) | GENMASK_ULL(15, 9); + mask_bits = GENMASK_ULL(29, 21) | GENMASK_ULL(15, 9); addr_shift = 4; /* @@ -352,16 +351,16 @@ static void get_cs_base_and_mask(struct amd64_pvt *pvt, int csrow, u8 dct, csbase = pvt->csels[dct].csbases[csrow]; csmask = pvt->csels[dct].csmasks[csrow >> 1]; - *base = (csbase & GENMASK(5, 15)) << 6; - *base |= (csbase & GENMASK(19, 30)) << 8; + *base = (csbase & GENMASK_ULL(15, 5)) << 6; + *base |= (csbase & GENMASK_ULL(30, 19)) << 8; *mask = ~0ULL; /* poke holes for the csmask */ - *mask &= ~((GENMASK(5, 15) << 6) | - (GENMASK(19, 30) << 8)); + *mask &= ~((GENMASK_ULL(15, 5) << 6) | + (GENMASK_ULL(30, 19) << 8)); - *mask |= (csmask & GENMASK(5, 15)) << 6; - *mask |= (csmask & GENMASK(19, 30)) << 8; + *mask |= (csmask & GENMASK_ULL(15, 5)) << 6; + *mask |= (csmask & GENMASK_ULL(30, 19)) << 8; return; } else { @@ -370,9 +369,11 @@ static void get_cs_base_and_mask(struct amd64_pvt *pvt, int csrow, u8 dct, addr_shift = 8; if (pvt->fam == 0x15) - base_bits = mask_bits = GENMASK(19,30) | GENMASK(5,13); + base_bits = mask_bits = + GENMASK_ULL(30,19) | GENMASK_ULL(13,5); else - base_bits = mask_bits = GENMASK(19,28) | GENMASK(5,13); + base_bits = mask_bits = + GENMASK_ULL(28,19) | GENMASK_ULL(13,5); } *base = (csbase & base_bits) << addr_shift; @@ -561,7 +562,7 @@ static u64 sys_addr_to_dram_addr(struct mem_ctl_info *mci, u64 sys_addr) * section 3.4.2 of AMD publication 24592: AMD x86-64 Architecture * Programmer's Manual Volume 1 Application Programming. */ - dram_addr = (sys_addr & GENMASK(0, 39)) - dram_base; + dram_addr = (sys_addr & GENMASK_ULL(39, 0)) - dram_base; edac_dbg(2, "using DRAM Base register to translate SysAddr 0x%lx to DramAddr 0x%lx\n", (unsigned long)sys_addr, (unsigned long)dram_addr); @@ -597,7 +598,7 @@ static u64 dram_addr_to_input_addr(struct mem_ctl_info *mci, u64 dram_addr) * concerning translating a DramAddr to an InputAddr. */ intlv_shift = num_node_interleave_bits(dram_intlv_en(pvt, 0)); - input_addr = ((dram_addr >> intlv_shift) & GENMASK(12, 35)) + + input_addr = ((dram_addr >> intlv_shift) & GENMASK_ULL(35, 12)) + (dram_addr & 0xfff); edac_dbg(2, " Intlv Shift=%d DramAddr=0x%lx maps to InputAddr=0x%lx\n", @@ -658,7 +659,7 @@ static int get_channel_from_ecc_syndrome(struct mem_ctl_info *, u16); * Determine if the DIMMs have ECC enabled. ECC is enabled ONLY if all the DIMMs * are ECC capable. */ -static unsigned long amd64_determine_edac_cap(struct amd64_pvt *pvt) +static unsigned long determine_edac_cap(struct amd64_pvt *pvt) { u8 bit; unsigned long edac_cap = EDAC_FLAG_NONE; @@ -673,9 +674,9 @@ static unsigned long amd64_determine_edac_cap(struct amd64_pvt *pvt) return edac_cap; } -static void amd64_debug_display_dimm_sizes(struct amd64_pvt *, u8); +static void debug_display_dimm_sizes(struct amd64_pvt *, u8); -static void amd64_dump_dramcfg_low(struct amd64_pvt *pvt, u32 dclr, int chan) +static void debug_dump_dramcfg_low(struct amd64_pvt *pvt, u32 dclr, int chan) { edac_dbg(1, "F2x%d90 (DRAM Cfg Low): 0x%08x\n", chan, dclr); @@ -709,7 +710,7 @@ static void dump_misc_regs(struct amd64_pvt *pvt) (pvt->nbcap & NBCAP_SECDED) ? "yes" : "no", (pvt->nbcap & NBCAP_CHIPKILL) ? "yes" : "no"); - amd64_dump_dramcfg_low(pvt, pvt->dclr0, 0); + debug_dump_dramcfg_low(pvt, pvt->dclr0, 0); edac_dbg(1, "F3xB0 (Online Spare): 0x%08x\n", pvt->online_spare); @@ -720,19 +721,19 @@ static void dump_misc_regs(struct amd64_pvt *pvt) edac_dbg(1, " DramHoleValid: %s\n", dhar_valid(pvt) ? "yes" : "no"); - amd64_debug_display_dimm_sizes(pvt, 0); + debug_display_dimm_sizes(pvt, 0); /* everything below this point is Fam10h and above */ if (pvt->fam == 0xf) return; - amd64_debug_display_dimm_sizes(pvt, 1); + debug_display_dimm_sizes(pvt, 1); amd64_info("using %s syndromes.\n", ((pvt->ecc_sym_sz == 8) ? "x8" : "x4")); /* Only if NOT ganged does dclr1 have valid info */ if (!dct_ganging_enabled(pvt)) - amd64_dump_dramcfg_low(pvt, pvt->dclr1, 1); + debug_dump_dramcfg_low(pvt, pvt->dclr1, 1); } /* @@ -798,7 +799,7 @@ static void read_dct_base_mask(struct amd64_pvt *pvt) } } -static enum mem_type amd64_determine_memory_type(struct amd64_pvt *pvt, int cs) +static enum mem_type determine_memory_type(struct amd64_pvt *pvt, int cs) { enum mem_type type; @@ -849,7 +850,7 @@ static u64 get_error_address(struct amd64_pvt *pvt, struct mce *m) end_bit = 39; } - addr = m->addr & GENMASK(start_bit, end_bit); + addr = m->addr & GENMASK_ULL(end_bit, start_bit); /* * Erratum 637 workaround @@ -861,7 +862,7 @@ static u64 get_error_address(struct amd64_pvt *pvt, struct mce *m) u16 mce_nid; u8 intlv_en; - if ((addr & GENMASK(24, 47)) >> 24 != 0x00fdf7) + if ((addr & GENMASK_ULL(47, 24)) >> 24 != 0x00fdf7) return addr; mce_nid = amd_get_nb_id(m->extcpu); @@ -871,7 +872,7 @@ static u64 get_error_address(struct amd64_pvt *pvt, struct mce *m) intlv_en = tmp >> 21 & 0x7; /* add [47:27] + 3 trailing bits */ - cc6_base = (tmp & GENMASK(0, 20)) << 3; + cc6_base = (tmp & GENMASK_ULL(20, 0)) << 3; /* reverse and add DramIntlvEn */ cc6_base |= intlv_en ^ 0x7; @@ -880,18 +881,18 @@ static u64 get_error_address(struct amd64_pvt *pvt, struct mce *m) cc6_base <<= 24; if (!intlv_en) - return cc6_base | (addr & GENMASK(0, 23)); + return cc6_base | (addr & GENMASK_ULL(23, 0)); amd64_read_pci_cfg(pvt->F1, DRAM_LOCAL_NODE_BASE, &tmp); /* faster log2 */ - tmp_addr = (addr & GENMASK(12, 23)) << __fls(intlv_en + 1); + tmp_addr = (addr & GENMASK_ULL(23, 12)) << __fls(intlv_en + 1); /* OR DramIntlvSel into bits [14:12] */ - tmp_addr |= (tmp & GENMASK(21, 23)) >> 9; + tmp_addr |= (tmp & GENMASK_ULL(23, 21)) >> 9; /* add remaining [11:0] bits from original MC4_ADDR */ - tmp_addr |= addr & GENMASK(0, 11); + tmp_addr |= addr & GENMASK_ULL(11, 0); return cc6_base | tmp_addr; } @@ -952,12 +953,12 @@ static void read_dram_base_limit_regs(struct amd64_pvt *pvt, unsigned range) amd64_read_pci_cfg(f1, DRAM_LOCAL_NODE_LIM, &llim); - pvt->ranges[range].lim.lo &= GENMASK(0, 15); + pvt->ranges[range].lim.lo &= GENMASK_ULL(15, 0); /* {[39:27],111b} */ pvt->ranges[range].lim.lo |= ((llim & 0x1fff) << 3 | 0x7) << 16; - pvt->ranges[range].lim.hi &= GENMASK(0, 7); + pvt->ranges[range].lim.hi &= GENMASK_ULL(7, 0); /* [47:40] */ pvt->ranges[range].lim.hi |= llim >> 13; @@ -1238,9 +1239,17 @@ static u8 f15_m30h_determine_channel(struct amd64_pvt *pvt, u64 sys_addr, if (num_dcts_intlv == 2) { select = (sys_addr >> 8) & 0x3; channel = select ? 0x3 : 0; - } else if (num_dcts_intlv == 4) - channel = (sys_addr >> 8) & 0x7; - + } else if (num_dcts_intlv == 4) { + u8 intlv_addr = dct_sel_interleave_addr(pvt); + switch (intlv_addr) { + case 0x4: + channel = (sys_addr >> 8) & 0x3; + break; + case 0x5: + channel = (sys_addr >> 9) & 0x3; + break; + } + } return channel; } @@ -1330,7 +1339,7 @@ static u64 f1x_get_norm_dct_addr(struct amd64_pvt *pvt, u8 range, chan_off = dram_base; } - return (sys_addr & GENMASK(6,47)) - (chan_off & GENMASK(23,47)); + return (sys_addr & GENMASK_ULL(47,6)) - (chan_off & GENMASK_ULL(47,23)); } /* @@ -1576,7 +1585,7 @@ static int f15_m30h_match_to_this_node(struct amd64_pvt *pvt, unsigned range, num_dcts_intlv, dct_sel); /* Verify we stay within the MAX number of channels allowed */ - if (channel > 4 || channel < 0) + if (channel > 3) return -EINVAL; leg_mmio_hole = (u8) (dct_cont_base_reg >> 1 & BIT(0)); @@ -1700,7 +1709,7 @@ static void f1x_map_sysaddr_to_csrow(struct mem_ctl_info *mci, u64 sys_addr, * debug routine to display the memory sizes of all logical DIMMs and its * CSROWs */ -static void amd64_debug_display_dimm_sizes(struct amd64_pvt *pvt, u8 ctrl) +static void debug_display_dimm_sizes(struct amd64_pvt *pvt, u8 ctrl) { int dimm, size0, size1; u32 *dcsb = ctrl ? pvt->csels[1].csbases : pvt->csels[0].csbases; @@ -1742,7 +1751,7 @@ static void amd64_debug_display_dimm_sizes(struct amd64_pvt *pvt, u8 ctrl) } } -static struct amd64_family_type amd64_family_types[] = { +static struct amd64_family_type family_types[] = { [K8_CPUS] = { .ctl_name = "K8", .f1_id = PCI_DEVICE_ID_AMD_K8_NB_ADDRMAP, @@ -1798,6 +1807,17 @@ static struct amd64_family_type amd64_family_types[] = { .read_dct_pci_cfg = f10_read_dct_pci_cfg, } }, + [F16_M30H_CPUS] = { + .ctl_name = "F16h_M30h", + .f1_id = PCI_DEVICE_ID_AMD_16H_M30H_NB_F1, + .f3_id = PCI_DEVICE_ID_AMD_16H_M30H_NB_F3, + .ops = { + .early_channel_count = f1x_early_channel_count, + .map_sysaddr_to_csrow = f1x_map_sysaddr_to_csrow, + .dbam_to_cs = f16_dbam_to_chip_select, + .read_dct_pci_cfg = f10_read_dct_pci_cfg, + } + }, }; /* @@ -2003,9 +2023,9 @@ static void __log_bus_error(struct mem_ctl_info *mci, struct err_info *err, string, ""); } -static inline void __amd64_decode_bus_error(struct mem_ctl_info *mci, - struct mce *m) +static inline void decode_bus_error(int node_id, struct mce *m) { + struct mem_ctl_info *mci = mcis[node_id]; struct amd64_pvt *pvt = mci->pvt_info; u8 ecc_type = (m->status >> 45) & 0x3; u8 xec = XEC(m->status, 0x1f); @@ -2033,11 +2053,6 @@ static inline void __amd64_decode_bus_error(struct mem_ctl_info *mci, __log_bus_error(mci, &err, ecc_type); } -void amd64_decode_bus_error(int node_id, struct mce *m) -{ - __amd64_decode_bus_error(mcis[node_id], m); -} - /* * Use pvt->F2 which contains the F2 CPU PCI device to get the related * F1 (AddrMap) and F3 (Misc) devices. Return negative value on error. @@ -2194,7 +2209,7 @@ static void read_mc_regs(struct amd64_pvt *pvt) * encompasses * */ -static u32 amd64_csrow_nr_pages(struct amd64_pvt *pvt, u8 dct, int csrow_nr) +static u32 get_csrow_nr_pages(struct amd64_pvt *pvt, u8 dct, int csrow_nr) { u32 cs_mode, nr_pages; u32 dbam = dct ? pvt->dbam1 : pvt->dbam0; @@ -2261,19 +2276,19 @@ static int init_csrows(struct mem_ctl_info *mci) pvt->mc_node_id, i); if (row_dct0) { - nr_pages = amd64_csrow_nr_pages(pvt, 0, i); + nr_pages = get_csrow_nr_pages(pvt, 0, i); csrow->channels[0]->dimm->nr_pages = nr_pages; } /* K8 has only one DCT */ if (pvt->fam != 0xf && row_dct1) { - int row_dct1_pages = amd64_csrow_nr_pages(pvt, 1, i); + int row_dct1_pages = get_csrow_nr_pages(pvt, 1, i); csrow->channels[1]->dimm->nr_pages = row_dct1_pages; nr_pages += row_dct1_pages; } - mtype = amd64_determine_memory_type(pvt, i); + mtype = determine_memory_type(pvt, i); edac_dbg(1, "Total csrow%d pages: %u\n", i, nr_pages); @@ -2307,7 +2322,7 @@ static void get_cpus_on_this_dct_cpumask(struct cpumask *mask, u16 nid) } /* check MCG_CTL on all the cpus on this node */ -static bool amd64_nb_mce_bank_enabled_on_node(u16 nid) +static bool nb_mce_bank_enabled_on_node(u16 nid) { cpumask_var_t mask; int cpu, nbe; @@ -2480,7 +2495,7 @@ static bool ecc_enabled(struct pci_dev *F3, u16 nid) ecc_en = !!(value & NBCFG_ECC_ENABLE); amd64_info("DRAM ECC %s.\n", (ecc_en ? "enabled" : "disabled")); - nb_mce_en = amd64_nb_mce_bank_enabled_on_node(nid); + nb_mce_en = nb_mce_bank_enabled_on_node(nid); if (!nb_mce_en) amd64_notice("NB MCE bank disabled, set MSR " "0x%08x[4] on node %d to enable.\n", @@ -2535,7 +2550,7 @@ static void setup_mci_misc_attrs(struct mem_ctl_info *mci, if (pvt->nbcap & NBCAP_CHIPKILL) mci->edac_ctl_cap |= EDAC_FLAG_S4ECD4ED; - mci->edac_cap = amd64_determine_edac_cap(pvt); + mci->edac_cap = determine_edac_cap(pvt); mci->mod_name = EDAC_MOD_STR; mci->mod_ver = EDAC_AMD64_VERSION; mci->ctl_name = fam->ctl_name; @@ -2543,14 +2558,14 @@ static void setup_mci_misc_attrs(struct mem_ctl_info *mci, mci->ctl_page_to_phys = NULL; /* memory scrubber interface */ - mci->set_sdram_scrub_rate = amd64_set_scrub_rate; - mci->get_sdram_scrub_rate = amd64_get_scrub_rate; + mci->set_sdram_scrub_rate = set_scrub_rate; + mci->get_sdram_scrub_rate = get_scrub_rate; } /* * returns a pointer to the family descriptor on success, NULL otherwise. */ -static struct amd64_family_type *amd64_per_family_init(struct amd64_pvt *pvt) +static struct amd64_family_type *per_family_init(struct amd64_pvt *pvt) { struct amd64_family_type *fam_type = NULL; @@ -2561,29 +2576,34 @@ static struct amd64_family_type *amd64_per_family_init(struct amd64_pvt *pvt) switch (pvt->fam) { case 0xf: - fam_type = &amd64_family_types[K8_CPUS]; - pvt->ops = &amd64_family_types[K8_CPUS].ops; + fam_type = &family_types[K8_CPUS]; + pvt->ops = &family_types[K8_CPUS].ops; break; case 0x10: - fam_type = &amd64_family_types[F10_CPUS]; - pvt->ops = &amd64_family_types[F10_CPUS].ops; + fam_type = &family_types[F10_CPUS]; + pvt->ops = &family_types[F10_CPUS].ops; break; case 0x15: if (pvt->model == 0x30) { - fam_type = &amd64_family_types[F15_M30H_CPUS]; - pvt->ops = &amd64_family_types[F15_M30H_CPUS].ops; + fam_type = &family_types[F15_M30H_CPUS]; + pvt->ops = &family_types[F15_M30H_CPUS].ops; break; } - fam_type = &amd64_family_types[F15_CPUS]; - pvt->ops = &amd64_family_types[F15_CPUS].ops; + fam_type = &family_types[F15_CPUS]; + pvt->ops = &family_types[F15_CPUS].ops; break; case 0x16: - fam_type = &amd64_family_types[F16_CPUS]; - pvt->ops = &amd64_family_types[F16_CPUS].ops; + if (pvt->model == 0x30) { + fam_type = &family_types[F16_M30H_CPUS]; + pvt->ops = &family_types[F16_M30H_CPUS].ops; + break; + } + fam_type = &family_types[F16_CPUS]; + pvt->ops = &family_types[F16_CPUS].ops; break; default: @@ -2599,7 +2619,7 @@ static struct amd64_family_type *amd64_per_family_init(struct amd64_pvt *pvt) return fam_type; } -static int amd64_init_one_instance(struct pci_dev *F2) +static int init_one_instance(struct pci_dev *F2) { struct amd64_pvt *pvt = NULL; struct amd64_family_type *fam_type = NULL; @@ -2617,7 +2637,7 @@ static int amd64_init_one_instance(struct pci_dev *F2) pvt->F2 = F2; ret = -EINVAL; - fam_type = amd64_per_family_init(pvt); + fam_type = per_family_init(pvt); if (!fam_type) goto err_free; @@ -2678,7 +2698,7 @@ static int amd64_init_one_instance(struct pci_dev *F2) if (report_gart_errors) amd_report_gart_errors(true); - amd_register_ecc_decoder(amd64_decode_bus_error); + amd_register_ecc_decoder(decode_bus_error); mcis[nid] = mci; @@ -2701,8 +2721,8 @@ err_ret: return ret; } -static int amd64_probe_one_instance(struct pci_dev *pdev, - const struct pci_device_id *mc_type) +static int probe_one_instance(struct pci_dev *pdev, + const struct pci_device_id *mc_type) { u16 nid = amd_get_node_id(pdev); struct pci_dev *F3 = node_to_amd_nb(nid)->misc; @@ -2734,7 +2754,7 @@ static int amd64_probe_one_instance(struct pci_dev *pdev, goto err_enable; } - ret = amd64_init_one_instance(pdev); + ret = init_one_instance(pdev); if (ret < 0) { amd64_err("Error probing instance: %d\n", nid); restore_ecc_error_reporting(s, nid, F3); @@ -2750,7 +2770,7 @@ err_out: return ret; } -static void amd64_remove_one_instance(struct pci_dev *pdev) +static void remove_one_instance(struct pci_dev *pdev) { struct mem_ctl_info *mci; struct amd64_pvt *pvt; @@ -2775,7 +2795,7 @@ static void amd64_remove_one_instance(struct pci_dev *pdev) /* unregister from EDAC MCE */ amd_report_gart_errors(false); - amd_unregister_ecc_decoder(amd64_decode_bus_error); + amd_unregister_ecc_decoder(decode_bus_error); kfree(ecc_stngs[nid]); ecc_stngs[nid] = NULL; @@ -2793,7 +2813,7 @@ static void amd64_remove_one_instance(struct pci_dev *pdev) * PCI core identifies what devices are on a system during boot, and then * inquiry this table to see if this driver is for a given device found. */ -static DEFINE_PCI_DEVICE_TABLE(amd64_pci_table) = { +static const struct pci_device_id amd64_pci_table[] = { { .vendor = PCI_VENDOR_ID_AMD, .device = PCI_DEVICE_ID_AMD_K8_NB_MEMCTL, @@ -2834,6 +2854,14 @@ static DEFINE_PCI_DEVICE_TABLE(amd64_pci_table) = { .class = 0, .class_mask = 0, }, + { + .vendor = PCI_VENDOR_ID_AMD, + .device = PCI_DEVICE_ID_AMD_16H_M30H_NB_F2, + .subvendor = PCI_ANY_ID, + .subdevice = PCI_ANY_ID, + .class = 0, + .class_mask = 0, + }, {0, } }; @@ -2841,8 +2869,8 @@ MODULE_DEVICE_TABLE(pci, amd64_pci_table); static struct pci_driver amd64_pci_driver = { .name = EDAC_MOD_STR, - .probe = amd64_probe_one_instance, - .remove = amd64_remove_one_instance, + .probe = probe_one_instance, + .remove = remove_one_instance, .id_table = amd64_pci_table, }; @@ -2851,23 +2879,18 @@ static void setup_pci_device(void) struct mem_ctl_info *mci; struct amd64_pvt *pvt; - if (amd64_ctl_pci) + if (pci_ctl) return; mci = mcis[0]; - if (mci) { - - pvt = mci->pvt_info; - amd64_ctl_pci = - edac_pci_create_generic_ctl(&pvt->F2->dev, EDAC_MOD_STR); - - if (!amd64_ctl_pci) { - pr_warning("%s(): Unable to create PCI control\n", - __func__); + if (!mci) + return; - pr_warning("%s(): PCI error report via EDAC not set\n", - __func__); - } + pvt = mci->pvt_info; + pci_ctl = edac_pci_create_generic_ctl(&pvt->F2->dev, EDAC_MOD_STR); + if (!pci_ctl) { + pr_warn("%s(): Unable to create PCI control\n", __func__); + pr_warn("%s(): PCI error report via EDAC not set\n", __func__); } } @@ -2923,8 +2946,8 @@ err_ret: static void __exit amd64_edac_exit(void) { - if (amd64_ctl_pci) - edac_pci_release_generic_ctl(amd64_ctl_pci); + if (pci_ctl) + edac_pci_release_generic_ctl(pci_ctl); pci_unregister_driver(&amd64_pci_driver); diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h index d2443cfa069..d903e0c2114 100644 --- a/drivers/edac/amd64_edac.h +++ b/drivers/edac/amd64_edac.h @@ -160,14 +160,6 @@ #define OFF false /* - * Create a contiguous bitmask starting at bit position @lo and ending at - * position @hi. For example - * - * GENMASK(21, 39) gives us the 64bit vector 0x000000ffffe00000. - */ -#define GENMASK(lo, hi) (((1ULL << ((hi) - (lo) + 1)) - 1) << (lo)) - -/* * PCI-defined configuration space registers */ #define PCI_DEVICE_ID_AMD_15H_M30H_NB_F1 0x141b @@ -176,6 +168,8 @@ #define PCI_DEVICE_ID_AMD_15H_NB_F2 0x1602 #define PCI_DEVICE_ID_AMD_16H_NB_F1 0x1531 #define PCI_DEVICE_ID_AMD_16H_NB_F2 0x1532 +#define PCI_DEVICE_ID_AMD_16H_M30H_NB_F1 0x1581 +#define PCI_DEVICE_ID_AMD_16H_M30H_NB_F2 0x1582 /* * Function 1 - Address Map @@ -308,6 +302,7 @@ enum amd_families { F15_CPUS, F15_M30H_CPUS, F16_CPUS, + F16_M30H_CPUS, NUM_FAMILIES, }; diff --git a/drivers/edac/amd76x_edac.c b/drivers/edac/amd76x_edac.c index 96e3ee3460a..3a501b530e1 100644 --- a/drivers/edac/amd76x_edac.c +++ b/drivers/edac/amd76x_edac.c @@ -333,7 +333,7 @@ static void amd76x_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(amd76x_pci_tbl) = { +static const struct pci_device_id amd76x_pci_tbl[] = { { PCI_VEND_DEV(AMD, FE_GATE_700C), PCI_ANY_ID, PCI_ANY_ID, 0, 0, AMD762}, diff --git a/drivers/edac/amd8111_edac.c b/drivers/edac/amd8111_edac.c index ddd890052ce..2b63f7c2d6d 100644 --- a/drivers/edac/amd8111_edac.c +++ b/drivers/edac/amd8111_edac.c @@ -350,6 +350,7 @@ static int amd8111_dev_probe(struct pci_dev *dev, const struct pci_device_id *id) { struct amd8111_dev_info *dev_info = &amd8111_devices[id->driver_data]; + int ret = -ENODEV; dev_info->dev = pci_get_device(PCI_VENDOR_ID_AMD, dev_info->err_dev, NULL); @@ -359,16 +360,15 @@ static int amd8111_dev_probe(struct pci_dev *dev, "vendor %x, device %x, name %s\n", PCI_VENDOR_ID_AMD, dev_info->err_dev, dev_info->ctl_name); - return -ENODEV; + goto err; } if (pci_enable_device(dev_info->dev)) { - pci_dev_put(dev_info->dev); printk(KERN_ERR "failed to enable:" "vendor %x, device %x, name %s\n", PCI_VENDOR_ID_AMD, dev_info->err_dev, dev_info->ctl_name); - return -ENODEV; + goto err_dev_put; } /* @@ -381,8 +381,10 @@ static int amd8111_dev_probe(struct pci_dev *dev, edac_device_alloc_ctl_info(0, dev_info->ctl_name, 1, NULL, 0, 0, NULL, 0, dev_info->edac_idx); - if (!dev_info->edac_dev) - return -ENOMEM; + if (!dev_info->edac_dev) { + ret = -ENOMEM; + goto err_dev_put; + } dev_info->edac_dev->pvt_info = dev_info; dev_info->edac_dev->dev = &dev_info->dev->dev; @@ -399,8 +401,7 @@ static int amd8111_dev_probe(struct pci_dev *dev, if (edac_device_add_device(dev_info->edac_dev) > 0) { printk(KERN_ERR "failed to add edac_dev for %s\n", dev_info->ctl_name); - edac_device_free_ctl_info(dev_info->edac_dev); - return -ENODEV; + goto err_edac_free_ctl; } printk(KERN_INFO "added one edac_dev on AMD8111 " @@ -409,6 +410,13 @@ static int amd8111_dev_probe(struct pci_dev *dev, dev_info->ctl_name); return 0; + +err_edac_free_ctl: + edac_device_free_ctl_info(dev_info->edac_dev); +err_dev_put: + pci_dev_put(dev_info->dev); +err: + return ret; } static void amd8111_dev_remove(struct pci_dev *dev) @@ -437,6 +445,7 @@ static int amd8111_pci_probe(struct pci_dev *dev, const struct pci_device_id *id) { struct amd8111_pci_info *pci_info = &amd8111_pcis[id->driver_data]; + int ret = -ENODEV; pci_info->dev = pci_get_device(PCI_VENDOR_ID_AMD, pci_info->err_dev, NULL); @@ -446,16 +455,15 @@ static int amd8111_pci_probe(struct pci_dev *dev, "vendor %x, device %x, name %s\n", PCI_VENDOR_ID_AMD, pci_info->err_dev, pci_info->ctl_name); - return -ENODEV; + goto err; } if (pci_enable_device(pci_info->dev)) { - pci_dev_put(pci_info->dev); printk(KERN_ERR "failed to enable:" "vendor %x, device %x, name %s\n", PCI_VENDOR_ID_AMD, pci_info->err_dev, pci_info->ctl_name); - return -ENODEV; + goto err_dev_put; } /* @@ -465,8 +473,10 @@ static int amd8111_pci_probe(struct pci_dev *dev, */ pci_info->edac_idx = edac_pci_alloc_index(); pci_info->edac_dev = edac_pci_alloc_ctl_info(0, pci_info->ctl_name); - if (!pci_info->edac_dev) - return -ENOMEM; + if (!pci_info->edac_dev) { + ret = -ENOMEM; + goto err_dev_put; + } pci_info->edac_dev->pvt_info = pci_info; pci_info->edac_dev->dev = &pci_info->dev->dev; @@ -483,8 +493,7 @@ static int amd8111_pci_probe(struct pci_dev *dev, if (edac_pci_add_device(pci_info->edac_dev, pci_info->edac_idx) > 0) { printk(KERN_ERR "failed to add edac_pci for %s\n", pci_info->ctl_name); - edac_pci_free_ctl_info(pci_info->edac_dev); - return -ENODEV; + goto err_edac_free_ctl; } printk(KERN_INFO "added one edac_pci on AMD8111 " @@ -493,6 +502,13 @@ static int amd8111_pci_probe(struct pci_dev *dev, pci_info->ctl_name); return 0; + +err_edac_free_ctl: + edac_pci_free_ctl_info(pci_info->edac_dev); +err_dev_put: + pci_dev_put(pci_info->dev); +err: + return ret; } static void amd8111_pci_remove(struct pci_dev *dev) diff --git a/drivers/edac/cell_edac.c b/drivers/edac/cell_edac.c index c2eaf334b90..374b57fc596 100644 --- a/drivers/edac/cell_edac.c +++ b/drivers/edac/cell_edac.c @@ -15,6 +15,7 @@ #include <linux/platform_device.h> #include <linux/stop_machine.h> #include <linux/io.h> +#include <linux/of_address.h> #include <asm/machdep.h> #include <asm/cell-regs.h> @@ -162,6 +163,7 @@ static void cell_edac_init_csrows(struct mem_ctl_info *mci) csrow->first_page, nr_pages); break; } + of_node_put(np); } static int cell_edac_probe(struct platform_device *pdev) diff --git a/drivers/edac/e752x_edac.c b/drivers/edac/e752x_edac.c index 644fec54681..b2d71388172 100644 --- a/drivers/edac/e752x_edac.c +++ b/drivers/edac/e752x_edac.c @@ -209,7 +209,6 @@ enum e752x_chips { */ struct e752x_pvt { - struct pci_dev *bridge_ck; struct pci_dev *dev_d0f0; struct pci_dev *dev_d0f1; u32 tolm; @@ -891,7 +890,7 @@ static void e752x_get_error_info(struct mem_ctl_info *mci, info->buf_ferr); if (info->dram_ferr) - pci_write_bits16(pvt->bridge_ck, E752X_DRAM_FERR, + pci_write_bits16(pvt->dev_d0f1, E752X_DRAM_FERR, info->dram_ferr, info->dram_ferr); pci_write_config_dword(dev, E752X_FERR_GLOBAL, @@ -936,7 +935,7 @@ static void e752x_get_error_info(struct mem_ctl_info *mci, info->buf_nerr); if (info->dram_nerr) - pci_write_bits16(pvt->bridge_ck, E752X_DRAM_NERR, + pci_write_bits16(pvt->dev_d0f1, E752X_DRAM_NERR, info->dram_nerr, info->dram_nerr); pci_write_config_dword(dev, E752X_NERR_GLOBAL, @@ -1177,36 +1176,33 @@ static void e752x_init_mem_map_table(struct pci_dev *pdev, static int e752x_get_devs(struct pci_dev *pdev, int dev_idx, struct e752x_pvt *pvt) { - struct pci_dev *dev; - - pvt->bridge_ck = pci_get_device(PCI_VENDOR_ID_INTEL, - pvt->dev_info->err_dev, pvt->bridge_ck); + pvt->dev_d0f1 = pci_get_device(PCI_VENDOR_ID_INTEL, + pvt->dev_info->err_dev, NULL); - if (pvt->bridge_ck == NULL) - pvt->bridge_ck = pci_scan_single_device(pdev->bus, + if (pvt->dev_d0f1 == NULL) { + pvt->dev_d0f1 = pci_scan_single_device(pdev->bus, PCI_DEVFN(0, 1)); + pci_dev_get(pvt->dev_d0f1); + } - if (pvt->bridge_ck == NULL) { + if (pvt->dev_d0f1 == NULL) { e752x_printk(KERN_ERR, "error reporting device not found:" "vendor %x device 0x%x (broken BIOS?)\n", PCI_VENDOR_ID_INTEL, e752x_devs[dev_idx].err_dev); return 1; } - dev = pci_get_device(PCI_VENDOR_ID_INTEL, + pvt->dev_d0f0 = pci_get_device(PCI_VENDOR_ID_INTEL, e752x_devs[dev_idx].ctl_dev, NULL); - if (dev == NULL) + if (pvt->dev_d0f0 == NULL) goto fail; - pvt->dev_d0f0 = dev; - pvt->dev_d0f1 = pci_dev_get(pvt->bridge_ck); - return 0; fail: - pci_dev_put(pvt->bridge_ck); + pci_dev_put(pvt->dev_d0f1); return 1; } @@ -1383,7 +1379,6 @@ static int e752x_probe1(struct pci_dev *pdev, int dev_idx) fail: pci_dev_put(pvt->dev_d0f0); pci_dev_put(pvt->dev_d0f1); - pci_dev_put(pvt->bridge_ck); edac_mc_free(mci); return -ENODEV; @@ -1417,11 +1412,10 @@ static void e752x_remove_one(struct pci_dev *pdev) pvt = (struct e752x_pvt *)mci->pvt_info; pci_dev_put(pvt->dev_d0f0); pci_dev_put(pvt->dev_d0f1); - pci_dev_put(pvt->bridge_ck); edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(e752x_pci_tbl) = { +static const struct pci_device_id e752x_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 7520_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, E7520}, diff --git a/drivers/edac/e7xxx_edac.c b/drivers/edac/e7xxx_edac.c index 1c4056a5038..3cda79bc8b0 100644 --- a/drivers/edac/e7xxx_edac.c +++ b/drivers/edac/e7xxx_edac.c @@ -555,7 +555,7 @@ static void e7xxx_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(e7xxx_pci_tbl) = { +static const struct pci_device_id e7xxx_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 7205_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, E7205}, diff --git a/drivers/edac/edac_device.c b/drivers/edac/edac_device.c index 211021dfec7..592af5f0cf3 100644 --- a/drivers/edac/edac_device.c +++ b/drivers/edac/edac_device.c @@ -437,6 +437,9 @@ void edac_device_workq_teardown(struct edac_device_ctl_info *edac_dev) { int status; + if (!edac_dev->edac_check) + return; + status = cancel_delayed_work(&edac_dev->work); if (status == 0) { /* workq instance might be running, wait for it */ @@ -530,12 +533,9 @@ int edac_device_add_device(struct edac_device_ctl_info *edac_dev) /* Report action taken */ edac_device_printk(edac_dev, KERN_INFO, - "Giving out device to module '%s' controller " - "'%s': DEV '%s' (%s)\n", - edac_dev->mod_name, - edac_dev->ctl_name, - edac_dev_name(edac_dev), - edac_op_state_to_string(edac_dev->op_state)); + "Giving out device to module %s controller %s: DEV %s (%s)\n", + edac_dev->mod_name, edac_dev->ctl_name, edac_dev->dev_name, + edac_op_state_to_string(edac_dev->op_state)); mutex_unlock(&device_ctls_mutex); return 0; diff --git a/drivers/edac/edac_mc.c b/drivers/edac/edac_mc.c index 89e109022d7..2c694b5297c 100644 --- a/drivers/edac/edac_mc.c +++ b/drivers/edac/edac_mc.c @@ -559,7 +559,8 @@ static void edac_mc_workq_function(struct work_struct *work_req) * * called with the mem_ctls_mutex held */ -static void edac_mc_workq_setup(struct mem_ctl_info *mci, unsigned msec) +static void edac_mc_workq_setup(struct mem_ctl_info *mci, unsigned msec, + bool init) { edac_dbg(0, "\n"); @@ -567,7 +568,9 @@ static void edac_mc_workq_setup(struct mem_ctl_info *mci, unsigned msec) if (mci->op_state != OP_RUNNING_POLL) return; - INIT_DELAYED_WORK(&mci->work, edac_mc_workq_function); + if (init) + INIT_DELAYED_WORK(&mci->work, edac_mc_workq_function); + mod_delayed_work(edac_workqueue, &mci->work, msecs_to_jiffies(msec)); } @@ -601,7 +604,7 @@ static void edac_mc_workq_teardown(struct mem_ctl_info *mci) * user space has updated our poll period value, need to * reset our workq delays */ -void edac_mc_reset_delay_period(int value) +void edac_mc_reset_delay_period(unsigned long value) { struct mem_ctl_info *mci; struct list_head *item; @@ -611,7 +614,7 @@ void edac_mc_reset_delay_period(int value) list_for_each(item, &mc_devices) { mci = list_entry(item, struct mem_ctl_info, link); - edac_mc_workq_setup(mci, (unsigned long) value); + edac_mc_workq_setup(mci, value, false); } mutex_unlock(&mem_ctls_mutex); @@ -782,14 +785,16 @@ int edac_mc_add_mc(struct mem_ctl_info *mci) /* This instance is NOW RUNNING */ mci->op_state = OP_RUNNING_POLL; - edac_mc_workq_setup(mci, edac_mc_get_poll_msec()); + edac_mc_workq_setup(mci, edac_mc_get_poll_msec(), true); } else { mci->op_state = OP_RUNNING_INTERRUPT; } /* Report action taken */ - edac_mc_printk(mci, KERN_INFO, "Giving out device to '%s' '%s':" - " DEV %s\n", mci->mod_name, mci->ctl_name, edac_dev_name(mci)); + edac_mc_printk(mci, KERN_INFO, + "Giving out device to module %s controller %s: DEV %s (%s)\n", + mci->mod_name, mci->ctl_name, mci->dev_name, + edac_op_state_to_string(mci->op_state)); edac_mc_owner = mci->mod_name; @@ -1013,7 +1018,7 @@ static void edac_ce_error(struct mem_ctl_info *mci, } edac_inc_ce_error(mci, enable_per_layer_report, pos, error_count); - if (mci->scrub_mode & SCRUB_SW_SRC) { + if (mci->scrub_mode == SCRUB_SW_SRC) { /* * Some memory controllers (called MCs below) can remap * memory so that it is still available at a different diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c index 9f7e0e60951..01fae8289cf 100644 --- a/drivers/edac/edac_mc_sysfs.c +++ b/drivers/edac/edac_mc_sysfs.c @@ -7,7 +7,7 @@ * * Written Doug Thompson <norsk5@xmission.com> www.softwarebitmaker.com * - * (c) 2012-2013 - Mauro Carvalho Chehab <mchehab@redhat.com> + * (c) 2012-2013 - Mauro Carvalho Chehab * The entire API were re-written, and ported to use struct device * */ @@ -52,18 +52,20 @@ int edac_mc_get_poll_msec(void) static int edac_set_poll_msec(const char *val, struct kernel_param *kp) { - long l; + unsigned long l; int ret; if (!val) return -EINVAL; - ret = kstrtol(val, 0, &l); + ret = kstrtoul(val, 0, &l); if (ret) return ret; - if ((int)l != l) + + if (l < 1000) return -EINVAL; - *((int *)kp->arg) = l; + + *((unsigned long *)kp->arg) = l; /* notify edac_mc engine to reset the poll period */ edac_mc_reset_delay_period(l); @@ -914,7 +916,7 @@ void __exit edac_debugfs_exit(void) debugfs_remove(edac_debugfs); } -int edac_create_debug_nodes(struct mem_ctl_info *mci) +static int edac_create_debug_nodes(struct mem_ctl_info *mci) { struct dentry *d, *parent; char name[80]; diff --git a/drivers/edac/edac_module.h b/drivers/edac/edac_module.h index 3d139c6e7fe..f2118bfcf8d 100644 --- a/drivers/edac/edac_module.h +++ b/drivers/edac/edac_module.h @@ -52,7 +52,7 @@ extern void edac_device_workq_setup(struct edac_device_ctl_info *edac_dev, extern void edac_device_workq_teardown(struct edac_device_ctl_info *edac_dev); extern void edac_device_reset_delay_period(struct edac_device_ctl_info *edac_dev, unsigned long value); -extern void edac_mc_reset_delay_period(int value); +extern void edac_mc_reset_delay_period(unsigned long value); extern void *edac_align_ptr(void **p, unsigned size, int n_elems); diff --git a/drivers/edac/edac_pci.c b/drivers/edac/edac_pci.c index dd370f92ace..2cf44b4db80 100644 --- a/drivers/edac/edac_pci.c +++ b/drivers/edac/edac_pci.c @@ -358,11 +358,9 @@ int edac_pci_add_device(struct edac_pci_ctl_info *pci, int edac_idx) } edac_pci_printk(pci, KERN_INFO, - "Giving out device to module '%s' controller '%s':" - " DEV '%s' (%s)\n", - pci->mod_name, - pci->ctl_name, - edac_dev_name(pci), edac_op_state_to_string(pci->op_state)); + "Giving out device to module %s controller %s: DEV %s (%s)\n", + pci->mod_name, pci->ctl_name, pci->dev_name, + edac_op_state_to_string(pci->op_state)); mutex_unlock(&edac_pci_ctls_mutex); return 0; diff --git a/drivers/edac/edac_stub.c b/drivers/edac/edac_stub.c index 351945fa2ec..9d9e18aefaa 100644 --- a/drivers/edac/edac_stub.c +++ b/drivers/edac/edac_stub.c @@ -29,6 +29,25 @@ EXPORT_SYMBOL_GPL(edac_err_assert); static atomic_t edac_subsys_valid = ATOMIC_INIT(0); +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 diff --git a/drivers/edac/ghes_edac.c b/drivers/edac/ghes_edac.c index bb534670ec0..8399b4e16fe 100644 --- a/drivers/edac/ghes_edac.c +++ b/drivers/edac/ghes_edac.c @@ -4,7 +4,7 @@ * This file may be distributed under the terms of the GNU General Public * License version 2. * - * Copyright (c) 2013 by Mauro Carvalho Chehab <mchehab@redhat.com> + * Copyright (c) 2013 by Mauro Carvalho Chehab * * Red Hat Inc. http://www.redhat.com */ @@ -297,15 +297,14 @@ void ghes_edac_report_mem_error(struct ghes *ghes, int sev, } /* Error address */ - if (mem_err->validation_bits & CPER_MEM_VALID_PHYSICAL_ADDRESS) { + if (mem_err->validation_bits & CPER_MEM_VALID_PA) { e->page_frame_number = mem_err->physical_addr >> PAGE_SHIFT; e->offset_in_page = mem_err->physical_addr & ~PAGE_MASK; } /* Error grain */ - if (mem_err->validation_bits & CPER_MEM_VALID_PHYSICAL_ADDRESS_MASK) { + if (mem_err->validation_bits & CPER_MEM_VALID_PA_MASK) e->grain = ~(mem_err->physical_addr_mask & ~PAGE_MASK); - } /* Memory error location, mapped on e->location */ p = e->location; @@ -315,6 +314,8 @@ void ghes_edac_report_mem_error(struct ghes *ghes, int sev, p += sprintf(p, "card:%d ", mem_err->card); if (mem_err->validation_bits & CPER_MEM_VALID_MODULE) p += sprintf(p, "module:%d ", mem_err->module); + if (mem_err->validation_bits & CPER_MEM_VALID_RANK_NUMBER) + p += sprintf(p, "rank:%d ", mem_err->rank); if (mem_err->validation_bits & CPER_MEM_VALID_BANK) p += sprintf(p, "bank:%d ", mem_err->bank); if (mem_err->validation_bits & CPER_MEM_VALID_ROW) @@ -323,6 +324,15 @@ void ghes_edac_report_mem_error(struct ghes *ghes, int sev, p += sprintf(p, "col:%d ", mem_err->column); if (mem_err->validation_bits & CPER_MEM_VALID_BIT_POSITION) p += sprintf(p, "bit_pos:%d ", mem_err->bit_pos); + if (mem_err->validation_bits & CPER_MEM_VALID_MODULE_HANDLE) { + const char *bank = NULL, *device = NULL; + dmi_memdev_name(mem_err->mem_dev_handle, &bank, &device); + if (bank != NULL && device != NULL) + p += sprintf(p, "DIMM location:%s %s ", bank, device); + else + p += sprintf(p, "DIMM DMI handle: 0x%.4x ", + mem_err->mem_dev_handle); + } if (p > e->location) *(p - 1) = '\0'; diff --git a/drivers/edac/highbank_l2_edac.c b/drivers/edac/highbank_l2_edac.c index c2bd8c6a434..2f193668ebc 100644 --- a/drivers/edac/highbank_l2_edac.c +++ b/drivers/edac/highbank_l2_edac.c @@ -50,8 +50,15 @@ static irqreturn_t highbank_l2_err_handler(int irq, void *dev_id) return IRQ_HANDLED; } +static const struct of_device_id hb_l2_err_of_match[] = { + { .compatible = "calxeda,hb-sregs-l2-ecc", }, + {}, +}; +MODULE_DEVICE_TABLE(of, hb_l2_err_of_match); + static int highbank_l2_err_probe(struct platform_device *pdev) { + const struct of_device_id *id; struct edac_device_ctl_info *dci; struct hb_l2_drvdata *drvdata; struct resource *r; @@ -90,28 +97,32 @@ static int highbank_l2_err_probe(struct platform_device *pdev) goto err; } + id = of_match_device(hb_l2_err_of_match, &pdev->dev); + dci->mod_name = pdev->dev.driver->name; + dci->ctl_name = id ? id->compatible : "unknown"; + dci->dev_name = dev_name(&pdev->dev); + + if (edac_device_add_device(dci)) + goto err; + drvdata->db_irq = platform_get_irq(pdev, 0); res = devm_request_irq(&pdev->dev, drvdata->db_irq, highbank_l2_err_handler, 0, dev_name(&pdev->dev), dci); if (res < 0) - goto err; + goto err2; drvdata->sb_irq = platform_get_irq(pdev, 1); res = devm_request_irq(&pdev->dev, drvdata->sb_irq, highbank_l2_err_handler, 0, dev_name(&pdev->dev), dci); if (res < 0) - goto err; - - dci->mod_name = dev_name(&pdev->dev); - dci->dev_name = dev_name(&pdev->dev); - - if (edac_device_add_device(dci)) - goto err; + goto err2; devres_close_group(&pdev->dev, NULL); return 0; +err2: + edac_device_del_device(&pdev->dev); err: devres_release_group(&pdev->dev, NULL); edac_device_free_ctl_info(dci); @@ -127,12 +138,6 @@ static int highbank_l2_err_remove(struct platform_device *pdev) return 0; } -static const struct of_device_id hb_l2_err_of_match[] = { - { .compatible = "calxeda,hb-sregs-l2-ecc", }, - {}, -}; -MODULE_DEVICE_TABLE(of, hb_l2_err_of_match); - static struct platform_driver highbank_l2_edac_driver = { .probe = highbank_l2_err_probe, .remove = highbank_l2_err_remove, diff --git a/drivers/edac/highbank_mc_edac.c b/drivers/edac/highbank_mc_edac.c index 4695dd2d71f..f784de1dc79 100644 --- a/drivers/edac/highbank_mc_edac.c +++ b/drivers/edac/highbank_mc_edac.c @@ -26,31 +26,40 @@ #include "edac_module.h" /* DDR Ctrlr Error Registers */ -#define HB_DDR_ECC_OPT 0x128 -#define HB_DDR_ECC_U_ERR_ADDR 0x130 -#define HB_DDR_ECC_U_ERR_STAT 0x134 -#define HB_DDR_ECC_U_ERR_DATAL 0x138 -#define HB_DDR_ECC_U_ERR_DATAH 0x13c -#define HB_DDR_ECC_C_ERR_ADDR 0x140 -#define HB_DDR_ECC_C_ERR_STAT 0x144 -#define HB_DDR_ECC_C_ERR_DATAL 0x148 -#define HB_DDR_ECC_C_ERR_DATAH 0x14c -#define HB_DDR_ECC_INT_STATUS 0x180 -#define HB_DDR_ECC_INT_ACK 0x184 -#define HB_DDR_ECC_U_ERR_ID 0x424 -#define HB_DDR_ECC_C_ERR_ID 0x428 -#define HB_DDR_ECC_INT_STAT_CE 0x8 -#define HB_DDR_ECC_INT_STAT_DOUBLE_CE 0x10 -#define HB_DDR_ECC_INT_STAT_UE 0x20 -#define HB_DDR_ECC_INT_STAT_DOUBLE_UE 0x40 +#define HB_DDR_ECC_ERR_BASE 0x128 +#define MW_DDR_ECC_ERR_BASE 0x1b4 + +#define HB_DDR_ECC_OPT 0x00 +#define HB_DDR_ECC_U_ERR_ADDR 0x08 +#define HB_DDR_ECC_U_ERR_STAT 0x0c +#define HB_DDR_ECC_U_ERR_DATAL 0x10 +#define HB_DDR_ECC_U_ERR_DATAH 0x14 +#define HB_DDR_ECC_C_ERR_ADDR 0x18 +#define HB_DDR_ECC_C_ERR_STAT 0x1c +#define HB_DDR_ECC_C_ERR_DATAL 0x20 +#define HB_DDR_ECC_C_ERR_DATAH 0x24 #define HB_DDR_ECC_OPT_MODE_MASK 0x3 #define HB_DDR_ECC_OPT_FWC 0x100 #define HB_DDR_ECC_OPT_XOR_SHIFT 16 +/* DDR Ctrlr Interrupt Registers */ + +#define HB_DDR_ECC_INT_BASE 0x180 +#define MW_DDR_ECC_INT_BASE 0x218 + +#define HB_DDR_ECC_INT_STATUS 0x00 +#define HB_DDR_ECC_INT_ACK 0x04 + +#define HB_DDR_ECC_INT_STAT_CE 0x8 +#define HB_DDR_ECC_INT_STAT_DOUBLE_CE 0x10 +#define HB_DDR_ECC_INT_STAT_UE 0x20 +#define HB_DDR_ECC_INT_STAT_DOUBLE_UE 0x40 + struct hb_mc_drvdata { - void __iomem *mc_vbase; + void __iomem *mc_err_base; + void __iomem *mc_int_base; }; static irqreturn_t highbank_mc_err_handler(int irq, void *dev_id) @@ -60,10 +69,10 @@ static irqreturn_t highbank_mc_err_handler(int irq, void *dev_id) u32 status, err_addr; /* Read the interrupt status register */ - status = readl(drvdata->mc_vbase + HB_DDR_ECC_INT_STATUS); + status = readl(drvdata->mc_int_base + HB_DDR_ECC_INT_STATUS); if (status & HB_DDR_ECC_INT_STAT_UE) { - err_addr = readl(drvdata->mc_vbase + HB_DDR_ECC_U_ERR_ADDR); + err_addr = readl(drvdata->mc_err_base + HB_DDR_ECC_U_ERR_ADDR); edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci, 1, err_addr >> PAGE_SHIFT, err_addr & ~PAGE_MASK, 0, @@ -71,9 +80,9 @@ static irqreturn_t highbank_mc_err_handler(int irq, void *dev_id) mci->ctl_name, ""); } if (status & HB_DDR_ECC_INT_STAT_CE) { - u32 syndrome = readl(drvdata->mc_vbase + HB_DDR_ECC_C_ERR_STAT); + u32 syndrome = readl(drvdata->mc_err_base + HB_DDR_ECC_C_ERR_STAT); syndrome = (syndrome >> 8) & 0xff; - err_addr = readl(drvdata->mc_vbase + HB_DDR_ECC_C_ERR_ADDR); + err_addr = readl(drvdata->mc_err_base + HB_DDR_ECC_C_ERR_ADDR); edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, 1, err_addr >> PAGE_SHIFT, err_addr & ~PAGE_MASK, syndrome, @@ -82,66 +91,79 @@ static irqreturn_t highbank_mc_err_handler(int irq, void *dev_id) } /* clear the error, clears the interrupt */ - writel(status, drvdata->mc_vbase + HB_DDR_ECC_INT_ACK); + writel(status, drvdata->mc_int_base + HB_DDR_ECC_INT_ACK); return IRQ_HANDLED; } -#ifdef CONFIG_EDAC_DEBUG -static ssize_t highbank_mc_err_inject_write(struct file *file, - const char __user *data, - size_t count, loff_t *ppos) +static void highbank_mc_err_inject(struct mem_ctl_info *mci, u8 synd) { - struct mem_ctl_info *mci = file->private_data; struct hb_mc_drvdata *pdata = mci->pvt_info; - char buf[32]; - size_t buf_size; u32 reg; + + reg = readl(pdata->mc_err_base + HB_DDR_ECC_OPT); + reg &= HB_DDR_ECC_OPT_MODE_MASK; + reg |= (synd << HB_DDR_ECC_OPT_XOR_SHIFT) | HB_DDR_ECC_OPT_FWC; + writel(reg, pdata->mc_err_base + HB_DDR_ECC_OPT); +} + +#define to_mci(k) container_of(k, struct mem_ctl_info, dev) + +static ssize_t highbank_mc_inject_ctrl(struct device *dev, + struct device_attribute *attr, const char *buf, size_t count) +{ + struct mem_ctl_info *mci = to_mci(dev); u8 synd; - buf_size = min(count, (sizeof(buf)-1)); - if (copy_from_user(buf, data, buf_size)) - return -EFAULT; - buf[buf_size] = 0; + if (kstrtou8(buf, 16, &synd)) + return -EINVAL; - if (!kstrtou8(buf, 16, &synd)) { - reg = readl(pdata->mc_vbase + HB_DDR_ECC_OPT); - reg &= HB_DDR_ECC_OPT_MODE_MASK; - reg |= (synd << HB_DDR_ECC_OPT_XOR_SHIFT) | HB_DDR_ECC_OPT_FWC; - writel(reg, pdata->mc_vbase + HB_DDR_ECC_OPT); - } + highbank_mc_err_inject(mci, synd); return count; } -static const struct file_operations highbank_mc_debug_inject_fops = { - .open = simple_open, - .write = highbank_mc_err_inject_write, - .llseek = generic_file_llseek, +static DEVICE_ATTR(inject_ctrl, S_IWUSR, NULL, highbank_mc_inject_ctrl); + +struct hb_mc_settings { + int err_offset; + int int_offset; }; -static void highbank_mc_create_debugfs_nodes(struct mem_ctl_info *mci) -{ - if (mci->debugfs) - debugfs_create_file("inject_ctrl", S_IWUSR, mci->debugfs, mci, - &highbank_mc_debug_inject_fops); -; -} -#else -static void highbank_mc_create_debugfs_nodes(struct mem_ctl_info *mci) -{} -#endif +static struct hb_mc_settings hb_settings = { + .err_offset = HB_DDR_ECC_ERR_BASE, + .int_offset = HB_DDR_ECC_INT_BASE, +}; + +static struct hb_mc_settings mw_settings = { + .err_offset = MW_DDR_ECC_ERR_BASE, + .int_offset = MW_DDR_ECC_INT_BASE, +}; + +static struct of_device_id hb_ddr_ctrl_of_match[] = { + { .compatible = "calxeda,hb-ddr-ctrl", .data = &hb_settings }, + { .compatible = "calxeda,ecx-2000-ddr-ctrl", .data = &mw_settings }, + {}, +}; +MODULE_DEVICE_TABLE(of, hb_ddr_ctrl_of_match); static int highbank_mc_probe(struct platform_device *pdev) { + const struct of_device_id *id; + const struct hb_mc_settings *settings; struct edac_mc_layer layers[2]; struct mem_ctl_info *mci; struct hb_mc_drvdata *drvdata; struct dimm_info *dimm; struct resource *r; + void __iomem *base; u32 control; int irq; int res = 0; + id = of_match_device(hb_ddr_ctrl_of_match, &pdev->dev); + if (!id) + return -ENODEV; + layers[0].type = EDAC_MC_LAYER_CHIP_SELECT; layers[0].size = 1; layers[0].is_virt_csrow = true; @@ -174,35 +196,31 @@ static int highbank_mc_probe(struct platform_device *pdev) goto err; } - drvdata->mc_vbase = devm_ioremap(&pdev->dev, - r->start, resource_size(r)); - if (!drvdata->mc_vbase) { + base = devm_ioremap(&pdev->dev, r->start, resource_size(r)); + if (!base) { dev_err(&pdev->dev, "Unable to map regs\n"); res = -ENOMEM; goto err; } - control = readl(drvdata->mc_vbase + HB_DDR_ECC_OPT) & 0x3; + settings = id->data; + drvdata->mc_err_base = base + settings->err_offset; + drvdata->mc_int_base = base + settings->int_offset; + + control = readl(drvdata->mc_err_base + HB_DDR_ECC_OPT) & 0x3; if (!control || (control == 0x2)) { dev_err(&pdev->dev, "No ECC present, or ECC disabled\n"); res = -ENODEV; goto err; } - irq = platform_get_irq(pdev, 0); - res = devm_request_irq(&pdev->dev, irq, highbank_mc_err_handler, - 0, dev_name(&pdev->dev), mci); - if (res < 0) { - dev_err(&pdev->dev, "Unable to request irq %d\n", irq); - goto err; - } - mci->mtype_cap = MEM_FLAG_DDR3; mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED; mci->edac_cap = EDAC_FLAG_SECDED; - mci->mod_name = dev_name(&pdev->dev); + mci->mod_name = pdev->dev.driver->name; mci->mod_ver = "1"; - mci->ctl_name = dev_name(&pdev->dev); + mci->ctl_name = id->compatible; + mci->dev_name = dev_name(&pdev->dev); mci->scrub_mode = SCRUB_SW_SRC; /* Only a single 4GB DIMM is supported */ @@ -217,10 +235,20 @@ static int highbank_mc_probe(struct platform_device *pdev) if (res < 0) goto err; - highbank_mc_create_debugfs_nodes(mci); + irq = platform_get_irq(pdev, 0); + res = devm_request_irq(&pdev->dev, irq, highbank_mc_err_handler, + 0, dev_name(&pdev->dev), mci); + if (res < 0) { + dev_err(&pdev->dev, "Unable to request irq %d\n", irq); + goto err2; + } + + device_create_file(&mci->dev, &dev_attr_inject_ctrl); devres_close_group(&pdev->dev, NULL); return 0; +err2: + edac_mc_del_mc(&pdev->dev); err: devres_release_group(&pdev->dev, NULL); edac_mc_free(mci); @@ -231,17 +259,12 @@ static int highbank_mc_remove(struct platform_device *pdev) { struct mem_ctl_info *mci = platform_get_drvdata(pdev); + device_remove_file(&mci->dev, &dev_attr_inject_ctrl); edac_mc_del_mc(&pdev->dev); edac_mc_free(mci); return 0; } -static const struct of_device_id hb_ddr_ctrl_of_match[] = { - { .compatible = "calxeda,hb-ddr-ctrl", }, - {}, -}; -MODULE_DEVICE_TABLE(of, hb_ddr_ctrl_of_match); - static struct platform_driver highbank_mc_edac_driver = { .probe = highbank_mc_probe, .remove = highbank_mc_remove, diff --git a/drivers/edac/i3000_edac.c b/drivers/edac/i3000_edac.c index 694efcbf19c..cd28b968e5c 100644 --- a/drivers/edac/i3000_edac.c +++ b/drivers/edac/i3000_edac.c @@ -487,7 +487,7 @@ static void i3000_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(i3000_pci_tbl) = { +static const struct pci_device_id i3000_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 3000_HB), PCI_ANY_ID, PCI_ANY_ID, 0, 0, I3000}, diff --git a/drivers/edac/i3200_edac.c b/drivers/edac/i3200_edac.c index be10a74b16e..022a70273ad 100644 --- a/drivers/edac/i3200_edac.c +++ b/drivers/edac/i3200_edac.c @@ -464,9 +464,11 @@ static void i3200_remove_one(struct pci_dev *pdev) iounmap(priv->window); edac_mc_free(mci); + + pci_disable_device(pdev); } -static DEFINE_PCI_DEVICE_TABLE(i3200_pci_tbl) = { +static const struct pci_device_id i3200_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 3200_HB), PCI_ANY_ID, PCI_ANY_ID, 0, 0, I3200}, diff --git a/drivers/edac/i5000_edac.c b/drivers/edac/i5000_edac.c index 63b2194e8c2..72e07e3cf71 100644 --- a/drivers/edac/i5000_edac.c +++ b/drivers/edac/i5000_edac.c @@ -1530,7 +1530,7 @@ static void i5000_remove_one(struct pci_dev *pdev) * * The "E500P" device is the first device supported. */ -static DEFINE_PCI_DEVICE_TABLE(i5000_pci_tbl) = { +static const struct pci_device_id i5000_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_I5000_DEV16), .driver_data = I5000P}, diff --git a/drivers/edac/i5100_edac.c b/drivers/edac/i5100_edac.c index 157b934e8ce..6247d186177 100644 --- a/drivers/edac/i5100_edac.c +++ b/drivers/edac/i5100_edac.c @@ -869,16 +869,13 @@ static void i5100_init_csrows(struct mem_ctl_info *mci) chan, rank, 0); dimm->nr_pages = npages; - if (npages) { - dimm->grain = 32; - dimm->dtype = (priv->mtr[chan][rank].width == 4) ? - DEV_X4 : DEV_X8; - dimm->mtype = MEM_RDDR2; - dimm->edac_mode = EDAC_SECDED; - snprintf(dimm->label, sizeof(dimm->label), - "DIMM%u", - i5100_rank_to_slot(mci, chan, rank)); - } + dimm->grain = 32; + dimm->dtype = (priv->mtr[chan][rank].width == 4) ? + DEV_X4 : DEV_X8; + dimm->mtype = MEM_RDDR2; + dimm->edac_mode = EDAC_SECDED; + snprintf(dimm->label, sizeof(dimm->label), "DIMM%u", + i5100_rank_to_slot(mci, chan, rank)); edac_dbg(2, "dimm channel %d, rank %d, size %ld\n", chan, rank, (long)PAGES_TO_MiB(npages)); @@ -1213,7 +1210,7 @@ static void i5100_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(i5100_pci_tbl) = { +static const struct pci_device_id i5100_pci_tbl[] = { /* Device 16, Function 0, Channel 0 Memory Map, Error Flag/Mask, ... */ { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_5100_16) }, { 0, } diff --git a/drivers/edac/i5400_edac.c b/drivers/edac/i5400_edac.c index 0a05bbceb08..6ef6ad1ba16 100644 --- a/drivers/edac/i5400_edac.c +++ b/drivers/edac/i5400_edac.c @@ -6,7 +6,7 @@ * * Copyright (c) 2008 by: * Ben Woodard <woodard@redhat.com> - * Mauro Carvalho Chehab <mchehab@redhat.com> + * Mauro Carvalho Chehab * * Red Hat Inc. http://www.redhat.com * @@ -1408,6 +1408,8 @@ static void i5400_remove_one(struct pci_dev *pdev) /* retrieve references to resources, and free those resources */ i5400_put_devices(mci); + pci_disable_device(pdev); + edac_mc_free(mci); } @@ -1416,7 +1418,7 @@ static void i5400_remove_one(struct pci_dev *pdev) * * The "E500P" device is the first device supported. */ -static DEFINE_PCI_DEVICE_TABLE(i5400_pci_tbl) = { +static const struct pci_device_id i5400_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_5400_ERR)}, {0,} /* 0 terminated list. */ }; @@ -1467,7 +1469,7 @@ module_exit(i5400_exit); MODULE_LICENSE("GPL"); MODULE_AUTHOR("Ben Woodard <woodard@redhat.com>"); -MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@redhat.com>"); +MODULE_AUTHOR("Mauro Carvalho Chehab"); MODULE_AUTHOR("Red Hat Inc. (http://www.redhat.com)"); MODULE_DESCRIPTION("MC Driver for Intel I5400 memory controllers - " I5400_REVISION); diff --git a/drivers/edac/i7300_edac.c b/drivers/edac/i7300_edac.c index 9004c64b169..dcac982fdc7 100644 --- a/drivers/edac/i7300_edac.c +++ b/drivers/edac/i7300_edac.c @@ -5,7 +5,7 @@ * GNU General Public License version 2 only. * * Copyright (c) 2010 by: - * Mauro Carvalho Chehab <mchehab@redhat.com> + * Mauro Carvalho Chehab * * Red Hat Inc. http://www.redhat.com * @@ -943,33 +943,35 @@ static int i7300_get_devices(struct mem_ctl_info *mci) /* Attempt to 'get' the MCH register we want */ pdev = NULL; - while (!pvt->pci_dev_16_1_fsb_addr_map || - !pvt->pci_dev_16_2_fsb_err_regs) { - pdev = pci_get_device(PCI_VENDOR_ID_INTEL, - PCI_DEVICE_ID_INTEL_I7300_MCH_ERR, pdev); - if (!pdev) { - /* End of list, leave */ - i7300_printk(KERN_ERR, - "'system address,Process Bus' " - "device not found:" - "vendor 0x%x device 0x%x ERR funcs " - "(broken BIOS?)\n", - PCI_VENDOR_ID_INTEL, - PCI_DEVICE_ID_INTEL_I7300_MCH_ERR); - goto error; - } - + while ((pdev = pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_I7300_MCH_ERR, + pdev))) { /* Store device 16 funcs 1 and 2 */ switch (PCI_FUNC(pdev->devfn)) { case 1: - pvt->pci_dev_16_1_fsb_addr_map = pdev; + if (!pvt->pci_dev_16_1_fsb_addr_map) + pvt->pci_dev_16_1_fsb_addr_map = + pci_dev_get(pdev); break; case 2: - pvt->pci_dev_16_2_fsb_err_regs = pdev; + if (!pvt->pci_dev_16_2_fsb_err_regs) + pvt->pci_dev_16_2_fsb_err_regs = + pci_dev_get(pdev); break; } } + if (!pvt->pci_dev_16_1_fsb_addr_map || + !pvt->pci_dev_16_2_fsb_err_regs) { + /* At least one device was not found */ + i7300_printk(KERN_ERR, + "'system address,Process Bus' device not found:" + "vendor 0x%x device 0x%x ERR funcs (broken BIOS?)\n", + PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_I7300_MCH_ERR); + goto error; + } + edac_dbg(1, "System Address, processor bus- PCI Bus ID: %s %x:%x\n", pci_name(pvt->pci_dev_16_0_fsb_ctlr), pvt->pci_dev_16_0_fsb_ctlr->vendor, @@ -1160,7 +1162,7 @@ static void i7300_remove_one(struct pci_dev *pdev) * * Has only 8086:360c PCI ID */ -static DEFINE_PCI_DEVICE_TABLE(i7300_pci_tbl) = { +static const struct pci_device_id i7300_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_I7300_MCH_ERR)}, {0,} /* 0 terminated list. */ }; @@ -1207,7 +1209,7 @@ module_init(i7300_init); module_exit(i7300_exit); MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@redhat.com>"); +MODULE_AUTHOR("Mauro Carvalho Chehab"); MODULE_AUTHOR("Red Hat Inc. (http://www.redhat.com)"); MODULE_DESCRIPTION("MC Driver for Intel I7300 memory controllers - " I7300_REVISION); diff --git a/drivers/edac/i7core_edac.c b/drivers/edac/i7core_edac.c index 80a963d64e5..9cd0b301f81 100644 --- a/drivers/edac/i7core_edac.c +++ b/drivers/edac/i7core_edac.c @@ -9,7 +9,7 @@ * GNU General Public License version 2 only. * * Copyright (c) 2009-2010 by: - * Mauro Carvalho Chehab <mchehab@redhat.com> + * Mauro Carvalho Chehab * * Red Hat Inc. http://www.redhat.com * @@ -394,7 +394,7 @@ static const struct pci_id_table pci_dev_table[] = { /* * pci_device_id table for which devices we are looking for */ -static DEFINE_PCI_DEVICE_TABLE(i7core_pci_tbl) = { +static const struct pci_device_id i7core_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_X58_HUB_MGMT)}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_LYNNFIELD_QPI_LINK0)}, {0,} /* 0 terminated list. */ @@ -1334,14 +1334,19 @@ static int i7core_get_onedevice(struct pci_dev **prev, * is at addr 8086:2c40, instead of 8086:2c41. So, we need * to probe for the alternate address in case of failure */ - if (dev_descr->dev_id == PCI_DEVICE_ID_INTEL_I7_NONCORE && !pdev) + if (dev_descr->dev_id == PCI_DEVICE_ID_INTEL_I7_NONCORE && !pdev) { + pci_dev_get(*prev); /* pci_get_device will put it */ pdev = pci_get_device(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_I7_NONCORE_ALT, *prev); + } - if (dev_descr->dev_id == PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE && !pdev) + if (dev_descr->dev_id == PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE && + !pdev) { + pci_dev_get(*prev); /* pci_get_device will put it */ pdev = pci_get_device(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE_ALT, *prev); + } if (!pdev) { if (*prev) { @@ -1703,7 +1708,7 @@ static void i7core_mce_output_error(struct mem_ctl_info *mci, const struct mce *m) { struct i7core_pvt *pvt = mci->pvt_info; - char *type, *optype, *err; + char *optype, *err; enum hw_event_mc_err_type tp_event; unsigned long error = m->status & 0x1ff0000l; bool uncorrected_error = m->mcgstatus & 1ll << 61; @@ -1716,15 +1721,11 @@ static void i7core_mce_output_error(struct mem_ctl_info *mci, u32 errnum = find_first_bit(&error, 32); if (uncorrected_error) { - if (ripv) { - type = "FATAL"; + if (ripv) tp_event = HW_EVENT_ERR_FATAL; - } else { - type = "NON_FATAL"; + else tp_event = HW_EVENT_ERR_UNCORRECTED; - } } else { - type = "CORRECTED"; tp_event = HW_EVENT_ERR_CORRECTED; } @@ -2456,7 +2457,7 @@ module_init(i7core_init); module_exit(i7core_exit); MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@redhat.com>"); +MODULE_AUTHOR("Mauro Carvalho Chehab"); MODULE_AUTHOR("Red Hat Inc. (http://www.redhat.com)"); MODULE_DESCRIPTION("MC Driver for Intel i7 Core memory controllers - " I7CORE_REVISION); diff --git a/drivers/edac/i82443bxgx_edac.c b/drivers/edac/i82443bxgx_edac.c index 57fdb77903b..d730e276d1a 100644 --- a/drivers/edac/i82443bxgx_edac.c +++ b/drivers/edac/i82443bxgx_edac.c @@ -386,7 +386,7 @@ static void i82443bxgx_edacmc_remove_one(struct pci_dev *pdev) EXPORT_SYMBOL_GPL(i82443bxgx_edacmc_remove_one); -static DEFINE_PCI_DEVICE_TABLE(i82443bxgx_pci_tbl) = { +static const struct pci_device_id i82443bxgx_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82443BX_0)}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82443BX_2)}, {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82443GX_0)}, diff --git a/drivers/edac/i82860_edac.c b/drivers/edac/i82860_edac.c index 3e3e431c830..3382f6344e4 100644 --- a/drivers/edac/i82860_edac.c +++ b/drivers/edac/i82860_edac.c @@ -288,7 +288,7 @@ static void i82860_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(i82860_pci_tbl) = { +static const struct pci_device_id i82860_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 82860_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, I82860}, diff --git a/drivers/edac/i82875p_edac.c b/drivers/edac/i82875p_edac.c index 2f8535fc451..64b68320249 100644 --- a/drivers/edac/i82875p_edac.c +++ b/drivers/edac/i82875p_edac.c @@ -275,7 +275,6 @@ static int i82875p_setup_overfl_dev(struct pci_dev *pdev, { struct pci_dev *dev; void __iomem *window; - int err; *ovrfl_pdev = NULL; *ovrfl_window = NULL; @@ -293,13 +292,8 @@ static int i82875p_setup_overfl_dev(struct pci_dev *pdev, if (dev == NULL) return 1; - err = pci_bus_add_device(dev); - if (err) { - i82875p_printk(KERN_ERR, - "%s(): pci_bus_add_device() Failed\n", - __func__); - } pci_bus_assign_resources(dev->bus); + pci_bus_add_device(dev); } *ovrfl_pdev = dev; @@ -406,8 +400,6 @@ static int i82875p_probe1(struct pci_dev *pdev, int dev_idx) edac_dbg(0, "\n"); - ovrfl_pdev = pci_get_device(PCI_VEND_DEV(INTEL, 82875_6), NULL); - if (i82875p_setup_overfl_dev(pdev, &ovrfl_pdev, &ovrfl_window)) return -ENODEV; drc = readl(ovrfl_window + I82875P_DRC); @@ -527,7 +519,7 @@ static void i82875p_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(i82875p_pci_tbl) = { +static const struct pci_device_id i82875p_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 82875_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, I82875P}, diff --git a/drivers/edac/i82975x_edac.c b/drivers/edac/i82975x_edac.c index 0c8d4b0eaa3..10b10521f62 100644 --- a/drivers/edac/i82975x_edac.c +++ b/drivers/edac/i82975x_edac.c @@ -628,7 +628,7 @@ static void i82975x_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(i82975x_pci_tbl) = { +static const struct pci_device_id i82975x_pci_tbl[] = { { PCI_VEND_DEV(INTEL, 82975_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, I82975X diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c index 30f7309446a..5f43620d580 100644 --- a/drivers/edac/mce_amd.c +++ b/drivers/edac/mce_amd.c @@ -6,7 +6,6 @@ static struct amd_decoder_ops *fam_ops; static u8 xec_mask = 0xf; -static u8 nb_err_cpumask = 0xf; static bool report_gart_errors; static void (*nb_bus_decoder)(int node_id, struct mce *m); @@ -741,6 +740,36 @@ int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) if (amd_filter_mce(m)) return NOTIFY_STOP; + pr_emerg(HW_ERR "%s\n", decode_error_status(m)); + + pr_emerg(HW_ERR "CPU:%d (%x:%x:%x) MC%d_STATUS[%s|%s|%s|%s|%s", + m->extcpu, + c->x86, c->x86_model, c->x86_mask, + m->bank, + ((m->status & MCI_STATUS_OVER) ? "Over" : "-"), + ((m->status & MCI_STATUS_UC) ? "UE" : "CE"), + ((m->status & MCI_STATUS_MISCV) ? "MiscV" : "-"), + ((m->status & MCI_STATUS_PCC) ? "PCC" : "-"), + ((m->status & MCI_STATUS_ADDRV) ? "AddrV" : "-")); + + if (c->x86 == 0x15 || c->x86 == 0x16) + pr_cont("|%s|%s", + ((m->status & MCI_STATUS_DEFERRED) ? "Deferred" : "-"), + ((m->status & MCI_STATUS_POISON) ? "Poison" : "-")); + + /* do the two bits[14:13] together */ + ecc = (m->status >> 45) & 0x3; + if (ecc) + pr_cont("|%sECC", ((ecc == 2) ? "C" : "U")); + + pr_cont("]: 0x%016llx\n", m->status); + + if (m->status & MCI_STATUS_ADDRV) + pr_emerg(HW_ERR "MC%d_ADDR: 0x%016llx\n", m->bank, m->addr); + + if (!fam_ops) + goto err_code; + switch (m->bank) { case 0: decode_mc0_mce(m); @@ -774,33 +803,7 @@ int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data) break; } - pr_emerg(HW_ERR "Error Status: %s\n", decode_error_status(m)); - - pr_emerg(HW_ERR "CPU:%d (%x:%x:%x) MC%d_STATUS[%s|%s|%s|%s|%s", - m->extcpu, - c->x86, c->x86_model, c->x86_mask, - m->bank, - ((m->status & MCI_STATUS_OVER) ? "Over" : "-"), - ((m->status & MCI_STATUS_UC) ? "UE" : "CE"), - ((m->status & MCI_STATUS_MISCV) ? "MiscV" : "-"), - ((m->status & MCI_STATUS_PCC) ? "PCC" : "-"), - ((m->status & MCI_STATUS_ADDRV) ? "AddrV" : "-")); - - if (c->x86 == 0x15 || c->x86 == 0x16) - pr_cont("|%s|%s", - ((m->status & MCI_STATUS_DEFERRED) ? "Deferred" : "-"), - ((m->status & MCI_STATUS_POISON) ? "Poison" : "-")); - - /* do the two bits[14:13] together */ - ecc = (m->status >> 45) & 0x3; - if (ecc) - pr_cont("|%sECC", ((ecc == 2) ? "C" : "U")); - - pr_cont("]: 0x%016llx\n", m->status); - - if (m->status & MCI_STATUS_ADDRV) - pr_emerg(HW_ERR "MC%d_ADDR: 0x%016llx\n", m->bank, m->addr); - + err_code: amd_decode_err_code(m->status & 0xffff); return NOTIFY_STOP; @@ -816,10 +819,7 @@ static int __init mce_amd_init(void) struct cpuinfo_x86 *c = &boot_cpu_data; if (c->x86_vendor != X86_VENDOR_AMD) - return 0; - - if (c->x86 < 0xf || c->x86 > 0x16) - return 0; + return -ENODEV; fam_ops = kzalloc(sizeof(struct amd_decoder_ops), GFP_KERNEL); if (!fam_ops) @@ -851,7 +851,6 @@ static int __init mce_amd_init(void) break; case 0x14: - nb_err_cpumask = 0x3; fam_ops->mc0_mce = cat_mc0_mce; fam_ops->mc1_mce = cat_mc1_mce; fam_ops->mc2_mce = k8_mc2_mce; @@ -874,7 +873,7 @@ static int __init mce_amd_init(void) default: printk(KERN_WARNING "Huh? What family is it: 0x%x?!\n", c->x86); kfree(fam_ops); - return -EINVAL; + fam_ops = NULL; } pr_info("MCE: In-kernel MCE decoding enabled.\n"); diff --git a/drivers/edac/mpc85xx_edac.c b/drivers/edac/mpc85xx_edac.c index 3eb32f62d72..f4aec2e6ef5 100644 --- a/drivers/edac/mpc85xx_edac.c +++ b/drivers/edac/mpc85xx_edac.c @@ -1,6 +1,8 @@ /* * Freescale MPC85xx Memory Controller kenel module * + * Parts Copyrighted (c) 2013 by Freescale Semiconductor, Inc. + * * Author: Dave Jiang <djiang@mvista.com> * * 2006-2007 (c) MontaVista Software, Inc. This file is licensed under @@ -196,6 +198,42 @@ static void mpc85xx_pci_check(struct edac_pci_ctl_info *pci) edac_pci_handle_npe(pci, pci->ctl_name); } +static void mpc85xx_pcie_check(struct edac_pci_ctl_info *pci) +{ + struct mpc85xx_pci_pdata *pdata = pci->pvt_info; + u32 err_detect; + + err_detect = in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_DR); + + pr_err("PCIe error(s) detected\n"); + pr_err("PCIe ERR_DR register: 0x%08x\n", err_detect); + pr_err("PCIe ERR_CAP_STAT register: 0x%08x\n", + in_be32(pdata->pci_vbase + MPC85XX_PCI_GAS_TIMR)); + pr_err("PCIe ERR_CAP_R0 register: 0x%08x\n", + in_be32(pdata->pci_vbase + MPC85XX_PCIE_ERR_CAP_R0)); + pr_err("PCIe ERR_CAP_R1 register: 0x%08x\n", + in_be32(pdata->pci_vbase + MPC85XX_PCIE_ERR_CAP_R1)); + pr_err("PCIe ERR_CAP_R2 register: 0x%08x\n", + in_be32(pdata->pci_vbase + MPC85XX_PCIE_ERR_CAP_R2)); + pr_err("PCIe ERR_CAP_R3 register: 0x%08x\n", + in_be32(pdata->pci_vbase + MPC85XX_PCIE_ERR_CAP_R3)); + + /* clear error bits */ + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_DR, err_detect); +} + +static int mpc85xx_pcie_find_capability(struct device_node *np) +{ + struct pci_controller *hose; + + if (!np) + return -EINVAL; + + hose = pci_find_hose_for_OF_device(np); + + return early_find_capability(hose, 0, 0, PCI_CAP_ID_EXP); +} + static irqreturn_t mpc85xx_pci_isr(int irq, void *dev_id) { struct edac_pci_ctl_info *pci = dev_id; @@ -207,7 +245,10 @@ static irqreturn_t mpc85xx_pci_isr(int irq, void *dev_id) if (!err_detect) return IRQ_NONE; - mpc85xx_pci_check(pci); + if (pdata->is_pcie) + mpc85xx_pcie_check(pci); + else + mpc85xx_pci_check(pci); return IRQ_HANDLED; } @@ -239,14 +280,22 @@ int mpc85xx_pci_err_probe(struct platform_device *op) pdata = pci->pvt_info; pdata->name = "mpc85xx_pci_err"; pdata->irq = NO_IRQ; + + if (mpc85xx_pcie_find_capability(op->dev.of_node) > 0) + pdata->is_pcie = true; + dev_set_drvdata(&op->dev, pci); pci->dev = &op->dev; pci->mod_name = EDAC_MOD_STR; pci->ctl_name = pdata->name; pci->dev_name = dev_name(&op->dev); - if (edac_op_state == EDAC_OPSTATE_POLL) - pci->edac_check = mpc85xx_pci_check; + if (edac_op_state == EDAC_OPSTATE_POLL) { + if (pdata->is_pcie) + pci->edac_check = mpc85xx_pcie_check; + else + pci->edac_check = mpc85xx_pci_check; + } pdata->edac_idx = edac_pci_idx++; @@ -275,16 +324,26 @@ int mpc85xx_pci_err_probe(struct platform_device *op) goto err; } - orig_pci_err_cap_dr = - in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_CAP_DR); + if (pdata->is_pcie) { + orig_pci_err_cap_dr = + in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_ADDR); + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_ADDR, ~0); + orig_pci_err_en = + in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN); + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN, 0); + } else { + orig_pci_err_cap_dr = + in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_CAP_DR); - /* PCI master abort is expected during config cycles */ - out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_CAP_DR, 0x40); + /* PCI master abort is expected during config cycles */ + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_CAP_DR, 0x40); - orig_pci_err_en = in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN); + orig_pci_err_en = + in_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN); - /* disable master abort reporting */ - out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN, ~0x40); + /* disable master abort reporting */ + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN, ~0x40); + } /* clear error bits */ out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_DR, ~0); @@ -297,7 +356,8 @@ int mpc85xx_pci_err_probe(struct platform_device *op) if (edac_op_state == EDAC_OPSTATE_INT) { pdata->irq = irq_of_parse_and_map(op->dev.of_node, 0); res = devm_request_irq(&op->dev, pdata->irq, - mpc85xx_pci_isr, IRQF_DISABLED, + mpc85xx_pci_isr, + IRQF_SHARED, "[EDAC] PCI err", pci); if (res < 0) { printk(KERN_ERR @@ -312,6 +372,22 @@ int mpc85xx_pci_err_probe(struct platform_device *op) pdata->irq); } + if (pdata->is_pcie) { + /* + * Enable all PCIe error interrupt & error detect except invalid + * PEX_CONFIG_ADDR/PEX_CONFIG_DATA access interrupt generation + * enable bit and invalid PEX_CONFIG_ADDR/PEX_CONFIG_DATA access + * detection enable bit. Because PCIe bus code to initialize and + * configure these PCIe devices on booting will use some invalid + * PEX_CONFIG_ADDR/PEX_CONFIG_DATA, edac driver prints the much + * notice information. So disable this detect to fix ugly print. + */ + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN, ~0 + & ~PEX_ERR_ICCAIE_EN_BIT); + out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_ADDR, 0 + | PEX_ERR_ICCAD_DISR_BIT); + } + devres_remove_group(&op->dev, mpc85xx_pci_err_probe); edac_dbg(3, "success\n"); printk(KERN_INFO EDAC_MOD_STR " PCI err registered\n"); @@ -327,28 +403,6 @@ err: } EXPORT_SYMBOL(mpc85xx_pci_err_probe); -static int mpc85xx_pci_err_remove(struct platform_device *op) -{ - struct edac_pci_ctl_info *pci = dev_get_drvdata(&op->dev); - struct mpc85xx_pci_pdata *pdata = pci->pvt_info; - - edac_dbg(0, "\n"); - - out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_CAP_DR, - orig_pci_err_cap_dr); - - out_be32(pdata->pci_vbase + MPC85XX_PCI_ERR_EN, orig_pci_err_en); - - edac_pci_del_device(pci->dev); - - if (edac_op_state == EDAC_OPSTATE_INT) - irq_dispose_mapping(pdata->irq); - - edac_pci_free_ctl_info(pci); - - return 0; -} - #endif /* CONFIG_PCI */ /**************************** L2 Err device ***************************/ @@ -579,7 +633,7 @@ static int mpc85xx_l2_err_probe(struct platform_device *op) if (edac_op_state == EDAC_OPSTATE_INT) { pdata->irq = irq_of_parse_and_map(op->dev.of_node, 0); res = devm_request_irq(&op->dev, pdata->irq, - mpc85xx_l2_isr, IRQF_DISABLED, + mpc85xx_l2_isr, 0, "[EDAC] L2 err", edac_dev); if (res < 0) { printk(KERN_ERR @@ -1079,7 +1133,7 @@ static int mpc85xx_mc_err_probe(struct platform_device *op) pdata->irq = irq_of_parse_and_map(op->dev.of_node, 0); res = devm_request_irq(&op->dev, pdata->irq, mpc85xx_mc_isr, - IRQF_DISABLED | IRQF_SHARED, + IRQF_SHARED, "[EDAC] MC err", mci); if (res < 0) { printk(KERN_ERR "%s: Unable to request irq %d for " diff --git a/drivers/edac/mpc85xx_edac.h b/drivers/edac/mpc85xx_edac.h index 932016f2cf0..8c625643622 100644 --- a/drivers/edac/mpc85xx_edac.h +++ b/drivers/edac/mpc85xx_edac.h @@ -134,13 +134,19 @@ #define MPC85XX_PCI_ERR_DR 0x0000 #define MPC85XX_PCI_ERR_CAP_DR 0x0004 #define MPC85XX_PCI_ERR_EN 0x0008 +#define PEX_ERR_ICCAIE_EN_BIT 0x00020000 #define MPC85XX_PCI_ERR_ATTRIB 0x000c #define MPC85XX_PCI_ERR_ADDR 0x0010 +#define PEX_ERR_ICCAD_DISR_BIT 0x00020000 #define MPC85XX_PCI_ERR_EXT_ADDR 0x0014 #define MPC85XX_PCI_ERR_DL 0x0018 #define MPC85XX_PCI_ERR_DH 0x001c #define MPC85XX_PCI_GAS_TIMR 0x0020 #define MPC85XX_PCI_PCIX_TIMR 0x0024 +#define MPC85XX_PCIE_ERR_CAP_R0 0x0028 +#define MPC85XX_PCIE_ERR_CAP_R1 0x002c +#define MPC85XX_PCIE_ERR_CAP_R2 0x0030 +#define MPC85XX_PCIE_ERR_CAP_R3 0x0034 struct mpc85xx_mc_pdata { char *name; @@ -158,6 +164,7 @@ struct mpc85xx_l2_pdata { struct mpc85xx_pci_pdata { char *name; + bool is_pcie; int edac_idx; void __iomem *pci_vbase; int irq; diff --git a/drivers/edac/octeon_edac-lmc.c b/drivers/edac/octeon_edac-lmc.c index 93412d6b3af..4bd10f94f06 100644 --- a/drivers/edac/octeon_edac-lmc.c +++ b/drivers/edac/octeon_edac-lmc.c @@ -5,12 +5,16 @@ * * Copyright (C) 2009 Wind River Systems, * written by Ralf Baechle <ralf@linux-mips.org> + * + * Copyright (c) 2013 by Cisco Systems, Inc. + * All rights reserved. */ #include <linux/module.h> #include <linux/init.h> #include <linux/slab.h> #include <linux/io.h> #include <linux/edac.h> +#include <linux/ctype.h> #include <asm/octeon/octeon.h> #include <asm/octeon/cvmx-lmcx-defs.h> @@ -20,6 +24,18 @@ #define OCTEON_MAX_MC 4 +#define to_mci(k) container_of(k, struct mem_ctl_info, dev) + +struct octeon_lmc_pvt { + unsigned long inject; + unsigned long error_type; + unsigned long dimm; + unsigned long rank; + unsigned long bank; + unsigned long row; + unsigned long col; +}; + static void octeon_lmc_edac_poll(struct mem_ctl_info *mci) { union cvmx_lmcx_mem_cfg0 cfg0; @@ -55,14 +71,31 @@ static void octeon_lmc_edac_poll(struct mem_ctl_info *mci) static void octeon_lmc_edac_poll_o2(struct mem_ctl_info *mci) { + struct octeon_lmc_pvt *pvt = mci->pvt_info; union cvmx_lmcx_int int_reg; bool do_clear = false; char msg[64]; - int_reg.u64 = cvmx_read_csr(CVMX_LMCX_INT(mci->mc_idx)); + if (!pvt->inject) + int_reg.u64 = cvmx_read_csr(CVMX_LMCX_INT(mci->mc_idx)); + else { + if (pvt->error_type == 1) + int_reg.s.sec_err = 1; + if (pvt->error_type == 2) + int_reg.s.ded_err = 1; + } + if (int_reg.s.sec_err || int_reg.s.ded_err) { union cvmx_lmcx_fadr fadr; - fadr.u64 = cvmx_read_csr(CVMX_LMCX_FADR(mci->mc_idx)); + if (likely(!pvt->inject)) + fadr.u64 = cvmx_read_csr(CVMX_LMCX_FADR(mci->mc_idx)); + else { + fadr.cn61xx.fdimm = pvt->dimm; + fadr.cn61xx.fbunk = pvt->rank; + fadr.cn61xx.fbank = pvt->bank; + fadr.cn61xx.frow = pvt->row; + fadr.cn61xx.fcol = pvt->col; + } snprintf(msg, sizeof(msg), "DIMM %d rank %d bank %d row %d col %d", fadr.cn61xx.fdimm, fadr.cn61xx.fbunk, @@ -82,8 +115,128 @@ static void octeon_lmc_edac_poll_o2(struct mem_ctl_info *mci) int_reg.s.ded_err = -1; /* Done, re-arm */ do_clear = true; } - if (do_clear) - cvmx_write_csr(CVMX_LMCX_INT(mci->mc_idx), int_reg.u64); + + if (do_clear) { + if (likely(!pvt->inject)) + cvmx_write_csr(CVMX_LMCX_INT(mci->mc_idx), int_reg.u64); + else + pvt->inject = 0; + } +} + +/************************ MC SYSFS parts ***********************************/ + +/* Only a couple naming differences per template, so very similar */ +#define TEMPLATE_SHOW(reg) \ +static ssize_t octeon_mc_inject_##reg##_show(struct device *dev, \ + struct device_attribute *attr, \ + char *data) \ +{ \ + struct mem_ctl_info *mci = to_mci(dev); \ + struct octeon_lmc_pvt *pvt = mci->pvt_info; \ + return sprintf(data, "%016llu\n", (u64)pvt->reg); \ +} + +#define TEMPLATE_STORE(reg) \ +static ssize_t octeon_mc_inject_##reg##_store(struct device *dev, \ + struct device_attribute *attr, \ + const char *data, size_t count) \ +{ \ + struct mem_ctl_info *mci = to_mci(dev); \ + struct octeon_lmc_pvt *pvt = mci->pvt_info; \ + if (isdigit(*data)) { \ + if (!kstrtoul(data, 0, &pvt->reg)) \ + return count; \ + } \ + return 0; \ +} + +TEMPLATE_SHOW(inject); +TEMPLATE_STORE(inject); +TEMPLATE_SHOW(dimm); +TEMPLATE_STORE(dimm); +TEMPLATE_SHOW(bank); +TEMPLATE_STORE(bank); +TEMPLATE_SHOW(rank); +TEMPLATE_STORE(rank); +TEMPLATE_SHOW(row); +TEMPLATE_STORE(row); +TEMPLATE_SHOW(col); +TEMPLATE_STORE(col); + +static ssize_t octeon_mc_inject_error_type_store(struct device *dev, + struct device_attribute *attr, + const char *data, + size_t count) +{ + struct mem_ctl_info *mci = to_mci(dev); + struct octeon_lmc_pvt *pvt = mci->pvt_info; + + if (!strncmp(data, "single", 6)) + pvt->error_type = 1; + else if (!strncmp(data, "double", 6)) + pvt->error_type = 2; + + return count; +} + +static ssize_t octeon_mc_inject_error_type_show(struct device *dev, + struct device_attribute *attr, + char *data) +{ + struct mem_ctl_info *mci = to_mci(dev); + struct octeon_lmc_pvt *pvt = mci->pvt_info; + if (pvt->error_type == 1) + return sprintf(data, "single"); + else if (pvt->error_type == 2) + return sprintf(data, "double"); + + return 0; +} + +static DEVICE_ATTR(inject, S_IRUGO | S_IWUSR, + octeon_mc_inject_inject_show, octeon_mc_inject_inject_store); +static DEVICE_ATTR(error_type, S_IRUGO | S_IWUSR, + octeon_mc_inject_error_type_show, octeon_mc_inject_error_type_store); +static DEVICE_ATTR(dimm, S_IRUGO | S_IWUSR, + octeon_mc_inject_dimm_show, octeon_mc_inject_dimm_store); +static DEVICE_ATTR(rank, S_IRUGO | S_IWUSR, + octeon_mc_inject_rank_show, octeon_mc_inject_rank_store); +static DEVICE_ATTR(bank, S_IRUGO | S_IWUSR, + octeon_mc_inject_bank_show, octeon_mc_inject_bank_store); +static DEVICE_ATTR(row, S_IRUGO | S_IWUSR, + octeon_mc_inject_row_show, octeon_mc_inject_row_store); +static DEVICE_ATTR(col, S_IRUGO | S_IWUSR, + octeon_mc_inject_col_show, octeon_mc_inject_col_store); + + +static int octeon_set_mc_sysfs_attributes(struct mem_ctl_info *mci) +{ + int rc; + + rc = device_create_file(&mci->dev, &dev_attr_inject); + if (rc < 0) + return rc; + rc = device_create_file(&mci->dev, &dev_attr_error_type); + if (rc < 0) + return rc; + rc = device_create_file(&mci->dev, &dev_attr_dimm); + if (rc < 0) + return rc; + rc = device_create_file(&mci->dev, &dev_attr_rank); + if (rc < 0) + return rc; + rc = device_create_file(&mci->dev, &dev_attr_bank); + if (rc < 0) + return rc; + rc = device_create_file(&mci->dev, &dev_attr_row); + if (rc < 0) + return rc; + rc = device_create_file(&mci->dev, &dev_attr_col); + if (rc < 0) + return rc; + + return 0; } static int octeon_lmc_edac_probe(struct platform_device *pdev) @@ -92,6 +245,8 @@ static int octeon_lmc_edac_probe(struct platform_device *pdev) struct edac_mc_layer layers[1]; int mc = pdev->id; + opstate_init(); + layers[0].type = EDAC_MC_LAYER_CHANNEL; layers[0].size = 1; layers[0].is_virt_csrow = false; @@ -105,7 +260,7 @@ static int octeon_lmc_edac_probe(struct platform_device *pdev) return 0; } - mci = edac_mc_alloc(mc, ARRAY_SIZE(layers), layers, 0); + mci = edac_mc_alloc(mc, ARRAY_SIZE(layers), layers, sizeof(struct octeon_lmc_pvt)); if (!mci) return -ENXIO; @@ -122,6 +277,12 @@ static int octeon_lmc_edac_probe(struct platform_device *pdev) return -ENXIO; } + if (octeon_set_mc_sysfs_attributes(mci)) { + dev_err(&pdev->dev, "octeon_set_mc_sysfs_attributes() failed\n"); + return -ENXIO; + } + + cfg0.u64 = cvmx_read_csr(CVMX_LMCX_MEM_CFG0(mc)); cfg0.s.intr_ded_ena = 0; /* We poll */ cfg0.s.intr_sec_ena = 0; @@ -137,7 +298,7 @@ static int octeon_lmc_edac_probe(struct platform_device *pdev) return 0; } - mci = edac_mc_alloc(mc, ARRAY_SIZE(layers), layers, 0); + mci = edac_mc_alloc(mc, ARRAY_SIZE(layers), layers, sizeof(struct octeon_lmc_pvt)); if (!mci) return -ENXIO; @@ -154,6 +315,12 @@ static int octeon_lmc_edac_probe(struct platform_device *pdev) return -ENXIO; } + if (octeon_set_mc_sysfs_attributes(mci)) { + dev_err(&pdev->dev, "octeon_set_mc_sysfs_attributes() failed\n"); + return -ENXIO; + } + + en.u64 = cvmx_read_csr(CVMX_LMCX_MEM_CFG0(mc)); en.s.intr_ded_ena = 0; /* We poll */ en.s.intr_sec_ena = 0; diff --git a/drivers/edac/r82600_edac.c b/drivers/edac/r82600_edac.c index 2fd6a549090..8f936bc7a01 100644 --- a/drivers/edac/r82600_edac.c +++ b/drivers/edac/r82600_edac.c @@ -383,7 +383,7 @@ static void r82600_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(r82600_pci_tbl) = { +static const struct pci_device_id r82600_pci_tbl[] = { { PCI_DEVICE(PCI_VENDOR_ID_RADISYS, R82600_BRIDGE_ID) }, diff --git a/drivers/edac/sb_edac.c b/drivers/edac/sb_edac.c index e04462b6075..deea0dc9999 100644 --- a/drivers/edac/sb_edac.c +++ b/drivers/edac/sb_edac.c @@ -7,7 +7,7 @@ * GNU General Public License version 2 only. * * Copyright (c) 2011 by: - * Mauro Carvalho Chehab <mchehab@redhat.com> + * Mauro Carvalho Chehab */ #include <linux/module.h> @@ -34,7 +34,7 @@ static int probed; /* * Alter this version for the module when modifications are made */ -#define SBRIDGE_REVISION " Ver: 1.0.0 " +#define SBRIDGE_REVISION " Ver: 1.1.0 " #define EDAC_MOD_STR "sbridge_edac" /* @@ -50,7 +50,7 @@ static int probed; * Get a bit field at register value <v>, from bit <lo> to bit <hi> */ #define GET_BITFIELD(v, lo, hi) \ - (((v) & ((1ULL << ((hi) - (lo) + 1)) - 1) << (lo)) >> (lo)) + (((v) & GENMASK_ULL(hi, lo)) >> (lo)) /* * sbridge Memory Controller Registers @@ -83,11 +83,17 @@ static int probed; #define PCI_DEVICE_ID_INTEL_SBRIDGE_IMC_ERR3 0x3c77 /* 16.7 */ /* Devices 12 Function 6, Offsets 0x80 to 0xcc */ -static const u32 dram_rule[] = { +static const u32 sbridge_dram_rule[] = { 0x80, 0x88, 0x90, 0x98, 0xa0, 0xa8, 0xb0, 0xb8, 0xc0, 0xc8, }; -#define MAX_SAD ARRAY_SIZE(dram_rule) + +static const u32 ibridge_dram_rule[] = { + 0x60, 0x68, 0x70, 0x78, 0x80, + 0x88, 0x90, 0x98, 0xa0, 0xa8, + 0xb0, 0xb8, 0xc0, 0xc8, 0xd0, + 0xd8, 0xe0, 0xe8, 0xf0, 0xf8, +}; #define SAD_LIMIT(reg) ((GET_BITFIELD(reg, 6, 25) << 26) | 0x3ffffff) #define DRAM_ATTR(reg) GET_BITFIELD(reg, 2, 3) @@ -108,43 +114,50 @@ static char *get_dram_attr(u32 reg) } } -static const u32 interleave_list[] = { +static const u32 sbridge_interleave_list[] = { 0x84, 0x8c, 0x94, 0x9c, 0xa4, 0xac, 0xb4, 0xbc, 0xc4, 0xcc, }; -#define MAX_INTERLEAVE ARRAY_SIZE(interleave_list) - -#define SAD_PKG0(reg) GET_BITFIELD(reg, 0, 2) -#define SAD_PKG1(reg) GET_BITFIELD(reg, 3, 5) -#define SAD_PKG2(reg) GET_BITFIELD(reg, 8, 10) -#define SAD_PKG3(reg) GET_BITFIELD(reg, 11, 13) -#define SAD_PKG4(reg) GET_BITFIELD(reg, 16, 18) -#define SAD_PKG5(reg) GET_BITFIELD(reg, 19, 21) -#define SAD_PKG6(reg) GET_BITFIELD(reg, 24, 26) -#define SAD_PKG7(reg) GET_BITFIELD(reg, 27, 29) - -static inline int sad_pkg(u32 reg, int interleave) + +static const u32 ibridge_interleave_list[] = { + 0x64, 0x6c, 0x74, 0x7c, 0x84, + 0x8c, 0x94, 0x9c, 0xa4, 0xac, + 0xb4, 0xbc, 0xc4, 0xcc, 0xd4, + 0xdc, 0xe4, 0xec, 0xf4, 0xfc, +}; + +struct interleave_pkg { + unsigned char start; + unsigned char end; +}; + +static const struct interleave_pkg sbridge_interleave_pkg[] = { + { 0, 2 }, + { 3, 5 }, + { 8, 10 }, + { 11, 13 }, + { 16, 18 }, + { 19, 21 }, + { 24, 26 }, + { 27, 29 }, +}; + +static const struct interleave_pkg ibridge_interleave_pkg[] = { + { 0, 3 }, + { 4, 7 }, + { 8, 11 }, + { 12, 15 }, + { 16, 19 }, + { 20, 23 }, + { 24, 27 }, + { 28, 31 }, +}; + +static inline int sad_pkg(const struct interleave_pkg *table, u32 reg, + int interleave) { - switch (interleave) { - case 0: - return SAD_PKG0(reg); - case 1: - return SAD_PKG1(reg); - case 2: - return SAD_PKG2(reg); - case 3: - return SAD_PKG3(reg); - case 4: - return SAD_PKG4(reg); - case 5: - return SAD_PKG5(reg); - case 6: - return SAD_PKG6(reg); - case 7: - return SAD_PKG7(reg); - default: - return -EINVAL; - } + return GET_BITFIELD(reg, table[interleave].start, + table[interleave].end); } /* Devices 12 Function 7 */ @@ -262,7 +275,9 @@ static const u32 correrrthrsld[] = { /* Device 17, function 0 */ -#define RANK_CFG_A 0x0328 +#define SB_RANK_CFG_A 0x0328 + +#define IB_RANK_CFG_A 0x0320 #define IS_RDIMM_ENABLED(reg) GET_BITFIELD(reg, 11, 11) @@ -273,8 +288,23 @@ static const u32 correrrthrsld[] = { #define NUM_CHANNELS 4 #define MAX_DIMMS 3 /* Max DIMMS per channel */ +enum type { + SANDY_BRIDGE, + IVY_BRIDGE, +}; + +struct sbridge_pvt; struct sbridge_info { - u32 mcmtr; + enum type type; + u32 mcmtr; + u32 rankcfgr; + u64 (*get_tolm)(struct sbridge_pvt *pvt); + u64 (*get_tohm)(struct sbridge_pvt *pvt); + const u32 *dram_rule; + const u32 *interleave_list; + const struct interleave_pkg *interleave_pkg; + u8 max_sad; + u8 max_interleave; }; struct sbridge_channel { @@ -305,8 +335,9 @@ struct sbridge_dev { struct sbridge_pvt { struct pci_dev *pci_ta, *pci_ddrio, *pci_ras; - struct pci_dev *pci_sad0, *pci_sad1, *pci_ha0; - struct pci_dev *pci_br; + struct pci_dev *pci_sad0, *pci_sad1; + struct pci_dev *pci_ha0, *pci_ha1; + struct pci_dev *pci_br0, *pci_br1; struct pci_dev *pci_tad[NUM_CHANNELS]; struct sbridge_dev *sbridge_dev; @@ -364,11 +395,75 @@ static const struct pci_id_table pci_dev_descr_sbridge_table[] = { {0,} /* 0 terminated list. */ }; +/* This changes depending if 1HA or 2HA: + * 1HA: + * 0x0eb8 (17.0) is DDRIO0 + * 2HA: + * 0x0ebc (17.4) is DDRIO0 + */ +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_1HA_DDRIO0 0x0eb8 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_2HA_DDRIO0 0x0ebc + +/* pci ids */ +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0 0x0ea0 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TA 0x0ea8 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_RAS 0x0e71 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD0 0x0eaa +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD1 0x0eab +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD2 0x0eac +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD3 0x0ead +#define PCI_DEVICE_ID_INTEL_IBRIDGE_SAD 0x0ec8 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_BR0 0x0ec9 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_BR1 0x0eca +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1 0x0e60 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_TA 0x0e68 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_RAS 0x0e79 +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_TAD0 0x0e6a +#define PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_TAD1 0x0e6b + +static const struct pci_id_descr pci_dev_descr_ibridge[] = { + /* Processor Home Agent */ + { PCI_DESCR(14, 0, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0, 0) }, + + /* Memory controller */ + { PCI_DESCR(15, 0, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TA, 0) }, + { PCI_DESCR(15, 1, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_RAS, 0) }, + { PCI_DESCR(15, 2, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD0, 0) }, + { PCI_DESCR(15, 3, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD1, 0) }, + { PCI_DESCR(15, 4, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD2, 0) }, + { PCI_DESCR(15, 5, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TAD3, 0) }, + + /* System Address Decoder */ + { PCI_DESCR(22, 0, PCI_DEVICE_ID_INTEL_IBRIDGE_SAD, 0) }, + + /* Broadcast Registers */ + { PCI_DESCR(22, 1, PCI_DEVICE_ID_INTEL_IBRIDGE_BR0, 1) }, + { PCI_DESCR(22, 2, PCI_DEVICE_ID_INTEL_IBRIDGE_BR1, 0) }, + + /* Optional, mode 2HA */ + { PCI_DESCR(28, 0, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1, 1) }, +#if 0 + { PCI_DESCR(29, 0, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_TA, 1) }, + { PCI_DESCR(29, 1, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_RAS, 1) }, +#endif + { PCI_DESCR(29, 2, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_TAD0, 1) }, + { PCI_DESCR(29, 3, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1_TAD1, 1) }, + + { PCI_DESCR(17, 0, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_1HA_DDRIO0, 1) }, + { PCI_DESCR(17, 4, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_2HA_DDRIO0, 1) }, +}; + +static const struct pci_id_table pci_dev_descr_ibridge_table[] = { + PCI_ID_TABLE_ENTRY(pci_dev_descr_ibridge), + {0,} /* 0 terminated list. */ +}; + /* * pci_device_id table for which devices we are looking for */ -static DEFINE_PCI_DEVICE_TABLE(sbridge_pci_tbl) = { +static const struct pci_device_id sbridge_pci_tbl[] = { {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_SBRIDGE_IMC_TA)}, + {PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TA)}, {0,} /* 0 terminated list. */ }; @@ -458,6 +553,52 @@ static void free_sbridge_dev(struct sbridge_dev *sbridge_dev) kfree(sbridge_dev); } +static u64 sbridge_get_tolm(struct sbridge_pvt *pvt) +{ + u32 reg; + + /* Address range is 32:28 */ + pci_read_config_dword(pvt->pci_sad1, TOLM, ®); + return GET_TOLM(reg); +} + +static u64 sbridge_get_tohm(struct sbridge_pvt *pvt) +{ + u32 reg; + + pci_read_config_dword(pvt->pci_sad1, TOHM, ®); + return GET_TOHM(reg); +} + +static u64 ibridge_get_tolm(struct sbridge_pvt *pvt) +{ + u32 reg; + + pci_read_config_dword(pvt->pci_br1, TOLM, ®); + + return GET_TOLM(reg); +} + +static u64 ibridge_get_tohm(struct sbridge_pvt *pvt) +{ + u32 reg; + + pci_read_config_dword(pvt->pci_br1, TOHM, ®); + + return GET_TOHM(reg); +} + +static inline u8 sad_pkg_socket(u8 pkg) +{ + /* on Ivy Bridge, nodeID is SASS, where A is HA and S is node id */ + return (pkg >> 3) | (pkg & 0x3); +} + +static inline u8 sad_pkg_ha(u8 pkg) +{ + return (pkg >> 2) & 0x1; +} + /**************************************************************************** Memory check routines ****************************************************************************/ @@ -520,10 +661,10 @@ static int get_dimm_config(struct mem_ctl_info *mci) enum edac_type mode; enum mem_type mtype; - pci_read_config_dword(pvt->pci_br, SAD_TARGET, ®); + pci_read_config_dword(pvt->pci_br0, SAD_TARGET, ®); pvt->sbridge_dev->source_id = SOURCE_ID(reg); - pci_read_config_dword(pvt->pci_br, SAD_CONTROL, ®); + pci_read_config_dword(pvt->pci_br0, SAD_CONTROL, ®); pvt->sbridge_dev->node_id = NODE_ID(reg); edac_dbg(0, "mc#%d: Node ID: %d, source ID: %d\n", pvt->sbridge_dev->mc, @@ -558,7 +699,8 @@ static int get_dimm_config(struct mem_ctl_info *mci) } if (pvt->pci_ddrio) { - pci_read_config_dword(pvt->pci_ddrio, RANK_CFG_A, ®); + pci_read_config_dword(pvt->pci_ddrio, pvt->info.rankcfgr, + ®); if (IS_RDIMM_ENABLED(reg)) { /* FIXME: Can also be LRDIMM */ edac_dbg(0, "Memory is registered\n"); @@ -629,19 +771,14 @@ static void get_memory_layout(const struct mem_ctl_info *mci) * Step 1) Get TOLM/TOHM ranges */ - /* Address range is 32:28 */ - pci_read_config_dword(pvt->pci_sad1, TOLM, - ®); - pvt->tolm = GET_TOLM(reg); + pvt->tolm = pvt->info.get_tolm(pvt); tmp_mb = (1 + pvt->tolm) >> 20; mb = div_u64_rem(tmp_mb, 1000, &kb); edac_dbg(0, "TOLM: %u.%03u GB (0x%016Lx)\n", mb, kb, (u64)pvt->tolm); /* Address range is already 45:25 */ - pci_read_config_dword(pvt->pci_sad1, TOHM, - ®); - pvt->tohm = GET_TOHM(reg); + pvt->tohm = pvt->info.get_tohm(pvt); tmp_mb = (1 + pvt->tohm) >> 20; mb = div_u64_rem(tmp_mb, 1000, &kb); @@ -654,9 +791,9 @@ static void get_memory_layout(const struct mem_ctl_info *mci) * algorithm bellow. */ prv = 0; - for (n_sads = 0; n_sads < MAX_SAD; n_sads++) { + for (n_sads = 0; n_sads < pvt->info.max_sad; n_sads++) { /* SAD_LIMIT Address range is 45:26 */ - pci_read_config_dword(pvt->pci_sad0, dram_rule[n_sads], + pci_read_config_dword(pvt->pci_sad0, pvt->info.dram_rule[n_sads], ®); limit = SAD_LIMIT(reg); @@ -677,15 +814,16 @@ static void get_memory_layout(const struct mem_ctl_info *mci) reg); prv = limit; - pci_read_config_dword(pvt->pci_sad0, interleave_list[n_sads], + pci_read_config_dword(pvt->pci_sad0, pvt->info.interleave_list[n_sads], ®); - sad_interl = sad_pkg(reg, 0); + sad_interl = sad_pkg(pvt->info.interleave_pkg, reg, 0); for (j = 0; j < 8; j++) { - if (j > 0 && sad_interl == sad_pkg(reg, j)) + u32 pkg = sad_pkg(pvt->info.interleave_pkg, reg, j); + if (j > 0 && sad_interl == pkg) break; edac_dbg(0, "SAD#%d, interleave #%d: %d\n", - n_sads, j, sad_pkg(reg, j)); + n_sads, j, pkg); } } @@ -777,7 +915,7 @@ static void get_memory_layout(const struct mem_ctl_info *mci) } } -struct mem_ctl_info *get_mci_for_node_id(u8 node_id) +static struct mem_ctl_info *get_mci_for_node_id(u8 node_id) { struct sbridge_dev *sbridge_dev; @@ -797,16 +935,17 @@ static int get_memory_error_data(struct mem_ctl_info *mci, { struct mem_ctl_info *new_mci; struct sbridge_pvt *pvt = mci->pvt_info; + struct pci_dev *pci_ha; int n_rir, n_sads, n_tads, sad_way, sck_xch; int sad_interl, idx, base_ch; int interleave_mode; - unsigned sad_interleave[MAX_INTERLEAVE]; + unsigned sad_interleave[pvt->info.max_interleave]; u32 reg; - u8 ch_way,sck_way; + u8 ch_way, sck_way, pkg, sad_ha = 0; u32 tad_offset; u32 rir_way; u32 mb, kb; - u64 ch_addr, offset, limit, prv = 0; + u64 ch_addr, offset, limit = 0, prv = 0; /* @@ -828,8 +967,8 @@ static int get_memory_error_data(struct mem_ctl_info *mci, /* * Step 1) Get socket */ - for (n_sads = 0; n_sads < MAX_SAD; n_sads++) { - pci_read_config_dword(pvt->pci_sad0, dram_rule[n_sads], + for (n_sads = 0; n_sads < pvt->info.max_sad; n_sads++) { + pci_read_config_dword(pvt->pci_sad0, pvt->info.dram_rule[n_sads], ®); if (!DRAM_RULE_ENABLE(reg)) @@ -844,53 +983,65 @@ static int get_memory_error_data(struct mem_ctl_info *mci, break; prv = limit; } - if (n_sads == MAX_SAD) { + if (n_sads == pvt->info.max_sad) { sprintf(msg, "Can't discover the memory socket"); return -EINVAL; } *area_type = get_dram_attr(reg); interleave_mode = INTERLEAVE_MODE(reg); - pci_read_config_dword(pvt->pci_sad0, interleave_list[n_sads], + pci_read_config_dword(pvt->pci_sad0, pvt->info.interleave_list[n_sads], ®); - sad_interl = sad_pkg(reg, 0); - for (sad_way = 0; sad_way < 8; sad_way++) { - if (sad_way > 0 && sad_interl == sad_pkg(reg, sad_way)) + + if (pvt->info.type == SANDY_BRIDGE) { + sad_interl = sad_pkg(pvt->info.interleave_pkg, reg, 0); + for (sad_way = 0; sad_way < 8; sad_way++) { + u32 pkg = sad_pkg(pvt->info.interleave_pkg, reg, sad_way); + if (sad_way > 0 && sad_interl == pkg) + break; + sad_interleave[sad_way] = pkg; + edac_dbg(0, "SAD interleave #%d: %d\n", + sad_way, sad_interleave[sad_way]); + } + edac_dbg(0, "mc#%d: Error detected on SAD#%d: address 0x%016Lx < 0x%016Lx, Interleave [%d:6]%s\n", + pvt->sbridge_dev->mc, + n_sads, + addr, + limit, + sad_way + 7, + !interleave_mode ? "" : "XOR[18:16]"); + if (interleave_mode) + idx = ((addr >> 6) ^ (addr >> 16)) & 7; + else + idx = (addr >> 6) & 7; + switch (sad_way) { + case 1: + idx = 0; break; - sad_interleave[sad_way] = sad_pkg(reg, sad_way); - edac_dbg(0, "SAD interleave #%d: %d\n", - sad_way, sad_interleave[sad_way]); - } - edac_dbg(0, "mc#%d: Error detected on SAD#%d: address 0x%016Lx < 0x%016Lx, Interleave [%d:6]%s\n", - pvt->sbridge_dev->mc, - n_sads, - addr, - limit, - sad_way + 7, - interleave_mode ? "" : "XOR[18:16]"); - if (interleave_mode) - idx = ((addr >> 6) ^ (addr >> 16)) & 7; - else + case 2: + idx = idx & 1; + break; + case 4: + idx = idx & 3; + break; + case 8: + break; + default: + sprintf(msg, "Can't discover socket interleave"); + return -EINVAL; + } + *socket = sad_interleave[idx]; + edac_dbg(0, "SAD interleave index: %d (wayness %d) = CPU socket %d\n", + idx, sad_way, *socket); + } else { + /* Ivy Bridge's SAD mode doesn't support XOR interleave mode */ idx = (addr >> 6) & 7; - switch (sad_way) { - case 1: - idx = 0; - break; - case 2: - idx = idx & 1; - break; - case 4: - idx = idx & 3; - break; - case 8: - break; - default: - sprintf(msg, "Can't discover socket interleave"); - return -EINVAL; + pkg = sad_pkg(pvt->info.interleave_pkg, reg, idx); + *socket = sad_pkg_socket(pkg); + sad_ha = sad_pkg_ha(pkg); + edac_dbg(0, "SAD interleave package: %d = CPU socket %d, HA %d\n", + idx, *socket, sad_ha); } - *socket = sad_interleave[idx]; - edac_dbg(0, "SAD interleave index: %d (wayness %d) = CPU socket %d\n", - idx, sad_way, *socket); /* * Move to the proper node structure, in order to access the @@ -909,9 +1060,16 @@ static int get_memory_error_data(struct mem_ctl_info *mci, * Step 2) Get memory channel */ prv = 0; + if (pvt->info.type == SANDY_BRIDGE) + pci_ha = pvt->pci_ha0; + else { + if (sad_ha) + pci_ha = pvt->pci_ha1; + else + pci_ha = pvt->pci_ha0; + } for (n_tads = 0; n_tads < MAX_TAD; n_tads++) { - pci_read_config_dword(pvt->pci_ha0, tad_dram_rule[n_tads], - ®); + pci_read_config_dword(pci_ha, tad_dram_rule[n_tads], ®); limit = TAD_LIMIT(reg); if (limit <= prv) { sprintf(msg, "Can't discover the memory channel"); @@ -921,14 +1079,13 @@ static int get_memory_error_data(struct mem_ctl_info *mci, break; prv = limit; } + if (n_tads == MAX_TAD) { + sprintf(msg, "Can't discover the memory channel"); + return -EINVAL; + } + ch_way = TAD_CH(reg) + 1; sck_way = TAD_SOCK(reg) + 1; - /* - * FIXME: Is it right to always use channel 0 for offsets? - */ - pci_read_config_dword(pvt->pci_tad[0], - tad_ch_nilv_offset[n_tads], - &tad_offset); if (ch_way == 3) idx = addr >> 6; @@ -958,6 +1115,10 @@ static int get_memory_error_data(struct mem_ctl_info *mci, } *channel_mask = 1 << base_ch; + pci_read_config_dword(pvt->pci_tad[base_ch], + tad_ch_nilv_offset[n_tads], + &tad_offset); + if (pvt->is_mirrored) { *channel_mask |= 1 << ((base_ch + 2) % 4); switch(ch_way) { @@ -1091,12 +1252,6 @@ static void sbridge_put_all_devices(void) } } -/* - * sbridge_get_all_devices Find and perform 'get' operation on the MCH's - * device/functions we want to reference for this driver - * - * Need to 'get' device 16 func 1 and func 2 - */ static int sbridge_get_onedevice(struct pci_dev **prev, u8 *num_mc, const struct pci_id_table *table, @@ -1108,7 +1263,7 @@ static int sbridge_get_onedevice(struct pci_dev **prev, struct pci_dev *pdev = NULL; u8 bus = 0; - sbridge_printk(KERN_INFO, + sbridge_printk(KERN_DEBUG, "Seeking for: dev %02x.%d PCI ID %04x:%04x\n", dev_descr->dev, dev_descr->func, PCI_VENDOR_ID_INTEL, dev_descr->dev_id); @@ -1198,11 +1353,21 @@ static int sbridge_get_onedevice(struct pci_dev **prev, return 0; } -static int sbridge_get_all_devices(u8 *num_mc) +/* + * sbridge_get_all_devices - Find and perform 'get' operation on the MCH's + * device/functions we want to reference for this driver. + * Need to 'get' device 16 func 1 and func 2. + * @num_mc: pointer to the memory controllers count, to be incremented in case + * of success. + * @table: model specific table + * + * returns 0 in case of success or error code + */ +static int sbridge_get_all_devices(u8 *num_mc, + const struct pci_id_table *table) { int i, rc; struct pci_dev *pdev = NULL; - const struct pci_id_table *table = pci_dev_descr_sbridge_table; while (table && table->descr) { for (i = 0; i < table->n_devs; i++) { @@ -1226,8 +1391,8 @@ static int sbridge_get_all_devices(u8 *num_mc) return 0; } -static int mci_bind_devs(struct mem_ctl_info *mci, - struct sbridge_dev *sbridge_dev) +static int sbridge_mci_bind_devs(struct mem_ctl_info *mci, + struct sbridge_dev *sbridge_dev) { struct sbridge_pvt *pvt = mci->pvt_info; struct pci_dev *pdev; @@ -1255,7 +1420,7 @@ static int mci_bind_devs(struct mem_ctl_info *mci, case 13: switch (func) { case 6: - pvt->pci_br = pdev; + pvt->pci_br0 = pdev; break; default: goto error; @@ -1329,6 +1494,131 @@ error: return -EINVAL; } +static int ibridge_mci_bind_devs(struct mem_ctl_info *mci, + struct sbridge_dev *sbridge_dev) +{ + struct sbridge_pvt *pvt = mci->pvt_info; + struct pci_dev *pdev, *tmp; + int i, func, slot; + bool mode_2ha = false; + + tmp = pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA1, NULL); + if (tmp) { + mode_2ha = true; + pci_dev_put(tmp); + } + + for (i = 0; i < sbridge_dev->n_devs; i++) { + pdev = sbridge_dev->pdev[i]; + if (!pdev) + continue; + slot = PCI_SLOT(pdev->devfn); + func = PCI_FUNC(pdev->devfn); + + switch (slot) { + case 14: + if (func == 0) { + pvt->pci_ha0 = pdev; + break; + } + goto error; + case 15: + switch (func) { + case 0: + pvt->pci_ta = pdev; + break; + case 1: + pvt->pci_ras = pdev; + break; + case 4: + case 5: + /* if we have 2 HAs active, channels 2 and 3 + * are in other device */ + if (mode_2ha) + break; + /* fall through */ + case 2: + case 3: + pvt->pci_tad[func - 2] = pdev; + break; + default: + goto error; + } + break; + case 17: + if (func == 4) { + pvt->pci_ddrio = pdev; + break; + } else if (func == 0) { + if (!mode_2ha) + pvt->pci_ddrio = pdev; + break; + } + goto error; + case 22: + switch (func) { + case 0: + pvt->pci_sad0 = pdev; + break; + case 1: + pvt->pci_br0 = pdev; + break; + case 2: + pvt->pci_br1 = pdev; + break; + default: + goto error; + } + break; + case 28: + if (func == 0) { + pvt->pci_ha1 = pdev; + break; + } + goto error; + case 29: + /* we shouldn't have this device if we have just one + * HA present */ + WARN_ON(!mode_2ha); + if (func == 2 || func == 3) { + pvt->pci_tad[func] = pdev; + break; + } + goto error; + default: + goto error; + } + + edac_dbg(0, "Associated PCI %02x.%02d.%d with dev = %p\n", + sbridge_dev->bus, + PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn), + pdev); + } + + /* Check if everything were registered */ + if (!pvt->pci_sad0 || !pvt->pci_ha0 || !pvt->pci_br0 || + !pvt->pci_br1 || !pvt->pci_tad || !pvt->pci_ras || + !pvt->pci_ta) + goto enodev; + + for (i = 0; i < NUM_CHANNELS; i++) { + if (!pvt->pci_tad[i]) + goto enodev; + } + return 0; + +enodev: + sbridge_printk(KERN_ERR, "Some needed devices are missing\n"); + return -ENODEV; + +error: + sbridge_printk(KERN_ERR, + "Device %d, function %d is out of the expected range\n", + slot, func); + return -EINVAL; +} + /**************************************************************************** Error check routines ****************************************************************************/ @@ -1349,7 +1639,7 @@ static void sbridge_mce_output_error(struct mem_ctl_info *mci, bool ripv = GET_BITFIELD(m->mcgstatus, 0, 0); bool overflow = GET_BITFIELD(m->status, 62, 62); bool uncorrected_error = GET_BITFIELD(m->status, 61, 61); - bool recoverable = GET_BITFIELD(m->status, 56, 56); + bool recoverable; u32 core_err_cnt = GET_BITFIELD(m->status, 38, 52); u32 mscod = GET_BITFIELD(m->status, 16, 31); u32 errcode = GET_BITFIELD(m->status, 0, 15); @@ -1360,6 +1650,11 @@ static void sbridge_mce_output_error(struct mem_ctl_info *mci, int rc, dimm; char *area_type = NULL; + if (pvt->info.type == IVY_BRIDGE) + recoverable = true; + else + recoverable = GET_BITFIELD(m->status, 56, 56); + if (uncorrected_error) { if (ripv) { type = "FATAL"; @@ -1409,6 +1704,10 @@ static void sbridge_mce_output_error(struct mem_ctl_info *mci, } } + /* Only decode errors with an valid address (ADDRV) */ + if (!GET_BITFIELD(m->status, 58, 58)) + return; + rc = get_memory_error_data(mci, m->addr, &socket, &channel_mask, &rank, &area_type, msg); if (rc < 0) @@ -1529,6 +1828,10 @@ static int sbridge_mce_check_error(struct notifier_block *nb, unsigned long val, struct mce *mce = (struct mce *)data; struct mem_ctl_info *mci; struct sbridge_pvt *pvt; + char *type; + + if (get_edac_report_status() == EDAC_REPORTING_DISABLED) + return NOTIFY_DONE; mci = get_mci_for_node_id(mce->socketid); if (!mci) @@ -1544,17 +1847,23 @@ static int sbridge_mce_check_error(struct notifier_block *nb, unsigned long val, if ((mce->status & 0xefff) >> 7 != 1) return NOTIFY_DONE; - printk("sbridge: HANDLING MCE MEMORY ERROR\n"); + if (mce->mcgstatus & MCG_STATUS_MCIP) + type = "Exception"; + else + type = "Event"; + + sbridge_mc_printk(mci, KERN_DEBUG, "HANDLING MCE MEMORY ERROR\n"); - printk("CPU %d: Machine Check Exception: %Lx Bank %d: %016Lx\n", - mce->extcpu, mce->mcgstatus, mce->bank, mce->status); - printk("TSC %llx ", mce->tsc); - printk("ADDR %llx ", mce->addr); - printk("MISC %llx ", mce->misc); + sbridge_mc_printk(mci, KERN_DEBUG, "CPU %d: Machine Check %s: %Lx " + "Bank %d: %016Lx\n", mce->extcpu, type, + mce->mcgstatus, mce->bank, mce->status); + sbridge_mc_printk(mci, KERN_DEBUG, "TSC %llx ", mce->tsc); + sbridge_mc_printk(mci, KERN_DEBUG, "ADDR %llx ", mce->addr); + sbridge_mc_printk(mci, KERN_DEBUG, "MISC %llx ", mce->misc); - printk("PROCESSOR %u:%x TIME %llu SOCKET %u APIC %x\n", - mce->cpuvendor, mce->cpuid, mce->time, - mce->socketid, mce->apicid); + sbridge_mc_printk(mci, KERN_DEBUG, "PROCESSOR %u:%x TIME %llu SOCKET " + "%u APIC %x\n", mce->cpuvendor, mce->cpuid, + mce->time, mce->socketid, mce->apicid); /* Only handle if it is the right mc controller */ if (cpu_data(mce->cpu).phys_proc_id != pvt->sbridge_dev->mc) @@ -1614,11 +1923,12 @@ static void sbridge_unregister_mci(struct sbridge_dev *sbridge_dev) sbridge_dev->mci = NULL; } -static int sbridge_register_mci(struct sbridge_dev *sbridge_dev) +static int sbridge_register_mci(struct sbridge_dev *sbridge_dev, enum type type) { struct mem_ctl_info *mci; struct edac_mc_layer layers[2]; struct sbridge_pvt *pvt; + struct pci_dev *pdev = sbridge_dev->pdev[0]; int rc; /* Check the number of active and not disabled channels */ @@ -1640,7 +1950,7 @@ static int sbridge_register_mci(struct sbridge_dev *sbridge_dev) return -ENOMEM; edac_dbg(0, "MC: mci = %p, dev = %p\n", - mci, &sbridge_dev->pdev[0]->dev); + mci, &pdev->dev); pvt = mci->pvt_info; memset(pvt, 0, sizeof(*pvt)); @@ -1654,24 +1964,52 @@ static int sbridge_register_mci(struct sbridge_dev *sbridge_dev) mci->edac_cap = EDAC_FLAG_NONE; mci->mod_name = "sbridge_edac.c"; mci->mod_ver = SBRIDGE_REVISION; - mci->ctl_name = kasprintf(GFP_KERNEL, "Sandy Bridge Socket#%d", mci->mc_idx); - mci->dev_name = pci_name(sbridge_dev->pdev[0]); + mci->dev_name = pci_name(pdev); mci->ctl_page_to_phys = NULL; /* Set the function pointer to an actual operation function */ mci->edac_check = sbridge_check_error; - /* Store pci devices at mci for faster access */ - rc = mci_bind_devs(mci, sbridge_dev); - if (unlikely(rc < 0)) - goto fail0; + pvt->info.type = type; + if (type == IVY_BRIDGE) { + pvt->info.rankcfgr = IB_RANK_CFG_A; + pvt->info.get_tolm = ibridge_get_tolm; + pvt->info.get_tohm = ibridge_get_tohm; + pvt->info.dram_rule = ibridge_dram_rule; + pvt->info.max_sad = ARRAY_SIZE(ibridge_dram_rule); + pvt->info.interleave_list = ibridge_interleave_list; + pvt->info.max_interleave = ARRAY_SIZE(ibridge_interleave_list); + pvt->info.interleave_pkg = ibridge_interleave_pkg; + mci->ctl_name = kasprintf(GFP_KERNEL, "Ivy Bridge Socket#%d", mci->mc_idx); + + /* Store pci devices at mci for faster access */ + rc = ibridge_mci_bind_devs(mci, sbridge_dev); + if (unlikely(rc < 0)) + goto fail0; + } else { + pvt->info.rankcfgr = SB_RANK_CFG_A; + pvt->info.get_tolm = sbridge_get_tolm; + pvt->info.get_tohm = sbridge_get_tohm; + pvt->info.dram_rule = sbridge_dram_rule; + pvt->info.max_sad = ARRAY_SIZE(sbridge_dram_rule); + pvt->info.interleave_list = sbridge_interleave_list; + pvt->info.max_interleave = ARRAY_SIZE(sbridge_interleave_list); + pvt->info.interleave_pkg = sbridge_interleave_pkg; + mci->ctl_name = kasprintf(GFP_KERNEL, "Sandy Bridge Socket#%d", mci->mc_idx); + + /* Store pci devices at mci for faster access */ + rc = sbridge_mci_bind_devs(mci, sbridge_dev); + if (unlikely(rc < 0)) + goto fail0; + } + /* Get dimm basic config and the memory layout */ get_dimm_config(mci); get_memory_layout(mci); /* record ptr to the generic device */ - mci->pdev = &sbridge_dev->pdev[0]->dev; + mci->pdev = &pdev->dev; /* add this new MC control structure to EDAC's list of MCs */ if (unlikely(edac_mc_add_mc(mci))) { @@ -1702,6 +2040,7 @@ static int sbridge_probe(struct pci_dev *pdev, const struct pci_device_id *id) int rc; u8 mc, num_mc = 0; struct sbridge_dev *sbridge_dev; + enum type type; /* get the pci devices we want to reserve for our use */ mutex_lock(&sbridge_edac_lock); @@ -1715,7 +2054,13 @@ static int sbridge_probe(struct pci_dev *pdev, const struct pci_device_id *id) } probed++; - rc = sbridge_get_all_devices(&num_mc); + if (pdev->device == PCI_DEVICE_ID_INTEL_IBRIDGE_IMC_HA0_TA) { + rc = sbridge_get_all_devices(&num_mc, pci_dev_descr_ibridge_table); + type = IVY_BRIDGE; + } else { + rc = sbridge_get_all_devices(&num_mc, pci_dev_descr_sbridge_table); + type = SANDY_BRIDGE; + } if (unlikely(rc < 0)) goto fail0; mc = 0; @@ -1724,7 +2069,7 @@ static int sbridge_probe(struct pci_dev *pdev, const struct pci_device_id *id) edac_dbg(0, "Registering MC#%d (%d of %d)\n", mc, mc + 1, num_mc); sbridge_dev->mc = mc++; - rc = sbridge_register_mci(sbridge_dev); + rc = sbridge_register_mci(sbridge_dev, type); if (unlikely(rc < 0)) goto fail1; } @@ -1807,9 +2152,10 @@ static int __init sbridge_init(void) opstate_init(); pci_rc = pci_register_driver(&sbridge_driver); - if (pci_rc >= 0) { mce_register_decode_chain(&sbridge_mce_dec); + if (get_edac_report_status() == EDAC_REPORTING_DISABLED) + sbridge_printk(KERN_WARNING, "Loading driver, error reporting disabled.\n"); return 0; } @@ -1837,7 +2183,7 @@ module_param(edac_op_state, int, 0444); MODULE_PARM_DESC(edac_op_state, "EDAC Error Reporting state: 0=Poll,1=NMI"); MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@redhat.com>"); +MODULE_AUTHOR("Mauro Carvalho Chehab"); MODULE_AUTHOR("Red Hat Inc. (http://www.redhat.com)"); -MODULE_DESCRIPTION("MC Driver for Intel Sandy Bridge memory controllers - " +MODULE_DESCRIPTION("MC Driver for Intel Sandy Bridge and Ivy Bridge memory controllers - " SBRIDGE_REVISION); diff --git a/drivers/edac/x38_edac.c b/drivers/edac/x38_edac.c index 1a4df82376b..4891b450830 100644 --- a/drivers/edac/x38_edac.c +++ b/drivers/edac/x38_edac.c @@ -448,7 +448,7 @@ static void x38_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static DEFINE_PCI_DEVICE_TABLE(x38_pci_tbl) = { +static const struct pci_device_id x38_pci_tbl[] = { { PCI_VEND_DEV(INTEL, X38_HB), PCI_ANY_ID, PCI_ANY_ID, 0, 0, X38}, |